diff options
Diffstat (limited to 'drivers')
149 files changed, 22974 insertions, 1367 deletions
diff --git a/drivers/Kconfig b/drivers/Kconfig index 9953a42809ec..d27feb5460f3 100644 --- a/drivers/Kconfig +++ b/drivers/Kconfig @@ -166,4 +166,6 @@ source "drivers/ipack/Kconfig" source "drivers/reset/Kconfig" +source "drivers/gator/Kconfig" + endmenu diff --git a/drivers/Makefile b/drivers/Makefile index 130abc1dfd65..092a62e79688 100644 --- a/drivers/Makefile +++ b/drivers/Makefile @@ -152,3 +152,5 @@ obj-$(CONFIG_IIO) += iio/ obj-$(CONFIG_VME_BUS) += vme/ obj-$(CONFIG_IPACK_BUS) += ipack/ obj-$(CONFIG_NTB) += ntb/ + +obj-$(CONFIG_GATOR) += gator/ diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig index 07abd9d76f7f..4fe77b887457 100644 --- a/drivers/base/Kconfig +++ b/drivers/base/Kconfig @@ -187,6 +187,14 @@ config GENERIC_CPU_DEVICES bool default n +config HAVE_CPU_AUTOPROBE + def_bool ARCH_HAS_CPU_AUTOPROBE + +config GENERIC_CPU_AUTOPROBE + bool + depends on !ARCH_HAS_CPU_AUTOPROBE + select HAVE_CPU_AUTOPROBE + config SOC_BUS bool @@ -202,11 +210,9 @@ config DMA_SHARED_BUFFER APIs extension; the file's descriptor can then be passed on to other driver. -config CMA - bool "Contiguous Memory Allocator" - depends on HAVE_DMA_CONTIGUOUS && HAVE_MEMBLOCK - select MIGRATION - select MEMORY_ISOLATION +config DMA_CMA + bool "DMA Contiguous Memory Allocator" + depends on HAVE_DMA_CONTIGUOUS && CMA help This enables the Contiguous Memory Allocator which allows drivers to allocate big physically-contiguous blocks of memory for use with @@ -215,17 +221,7 @@ config CMA For more information see <include/linux/dma-contiguous.h>. If unsure, say "n". -if CMA - -config CMA_DEBUG - bool "CMA debug messages (DEVELOPMENT)" - depends on DEBUG_KERNEL - help - Turns on debug messages in CMA. This produces KERN_DEBUG - messages for every CMA call as well as various messages while - processing calls such as dma_alloc_from_contiguous(). - This option does not affect warning and error messages. - +if DMA_CMA comment "Default contiguous memory area size:" config CMA_SIZE_MBYTES diff --git a/drivers/base/Makefile b/drivers/base/Makefile index 4e22ce3ed73d..5d93bb519753 100644 --- a/drivers/base/Makefile +++ b/drivers/base/Makefile @@ -6,7 +6,7 @@ obj-y := core.o bus.o dd.o syscore.o \ attribute_container.o transport_class.o \ topology.o obj-$(CONFIG_DEVTMPFS) += devtmpfs.o -obj-$(CONFIG_CMA) += dma-contiguous.o +obj-$(CONFIG_DMA_CMA) += dma-contiguous.o obj-y += power/ obj-$(CONFIG_HAS_DMA) += dma-mapping.o obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o diff --git a/drivers/base/cpu.c b/drivers/base/cpu.c index 3d48fc887ef4..607efe6b7dc8 100644 --- a/drivers/base/cpu.c +++ b/drivers/base/cpu.c @@ -13,6 +13,9 @@ #include <linux/gfp.h> #include <linux/slab.h> #include <linux/percpu.h> +#include <linux/acpi.h> +#include <linux/of.h> +#include <linux/cpufeature.h> #include "base.h" @@ -260,6 +263,45 @@ static void cpu_device_release(struct device *dev) */ } +#ifdef CONFIG_HAVE_CPU_AUTOPROBE +#ifdef CONFIG_GENERIC_CPU_AUTOPROBE +static ssize_t print_cpu_modalias(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t n; + u32 i; + + n = sprintf(buf, "cpu:type:" CPU_FEATURE_TYPEFMT ":feature:", + CPU_FEATURE_TYPEVAL); + + for (i = 0; i < MAX_CPU_FEATURES; i++) + if (cpu_have_feature(i)) { + if (PAGE_SIZE < n + sizeof(",XXXX\n")) { + WARN(1, "CPU features overflow page\n"); + break; + } + n += sprintf(&buf[n], ",%04X", i); + } + buf[n++] = '\n'; + return n; +} +#else +#define print_cpu_modalias arch_print_cpu_modalias +#endif + +static int cpu_uevent(struct device *dev, struct kobj_uevent_env *env) +{ + char *buf = kzalloc(PAGE_SIZE, GFP_KERNEL); + if (buf) { + print_cpu_modalias(NULL, NULL, buf); + add_uevent_var(env, "MODALIAS=%s", buf); + kfree(buf); + } + return 0; +} +#endif + /* * register_cpu - Setup a sysfs device for a CPU. * @cpu - cpu->hotpluggable field set to 1 will generate a control file in @@ -277,8 +319,8 @@ int __cpuinit register_cpu(struct cpu *cpu, int num) cpu->dev.id = num; cpu->dev.bus = &cpu_subsys; cpu->dev.release = cpu_device_release; -#ifdef CONFIG_ARCH_HAS_CPU_AUTOPROBE - cpu->dev.bus->uevent = arch_cpu_uevent; +#ifdef CONFIG_HAVE_CPU_AUTOPROBE + cpu->dev.bus->uevent = cpu_uevent; #endif error = device_register(&cpu->dev); if (!error && cpu->hotpluggable) @@ -307,8 +349,8 @@ struct device *get_cpu_device(unsigned cpu) } EXPORT_SYMBOL_GPL(get_cpu_device); -#ifdef CONFIG_ARCH_HAS_CPU_AUTOPROBE -static DEVICE_ATTR(modalias, 0444, arch_print_cpu_modalias, NULL); +#ifdef CONFIG_HAVE_CPU_AUTOPROBE +static DEVICE_ATTR(modalias, 0444, print_cpu_modalias, NULL); #endif static struct attribute *cpu_root_attrs[] = { @@ -321,7 +363,7 @@ static struct attribute *cpu_root_attrs[] = { &cpu_attrs[2].attr.attr, &dev_attr_kernel_max.attr, &dev_attr_offline.attr, -#ifdef CONFIG_ARCH_HAS_CPU_AUTOPROBE +#ifdef CONFIG_HAVE_CPU_AUTOPROBE &dev_attr_modalias.attr, #endif NULL diff --git a/drivers/base/devres.c b/drivers/base/devres.c index 507379e7b763..545c4de412c3 100644 --- a/drivers/base/devres.c +++ b/drivers/base/devres.c @@ -91,7 +91,8 @@ static __always_inline struct devres * alloc_dr(dr_release_t release, if (unlikely(!dr)) return NULL; - memset(dr, 0, tot_size); + memset(dr, 0, offsetof(struct devres, data)); + INIT_LIST_HEAD(&dr->node.entry); dr->node.release = release; return dr; @@ -110,7 +111,7 @@ void * __devres_alloc(dr_release_t release, size_t size, gfp_t gfp, { struct devres *dr; - dr = alloc_dr(release, size, gfp); + dr = alloc_dr(release, size, gfp | __GFP_ZERO); if (unlikely(!dr)) return NULL; set_node_dbginfo(&dr->node, name, size); @@ -135,7 +136,7 @@ void * devres_alloc(dr_release_t release, size_t size, gfp_t gfp) { struct devres *dr; - dr = alloc_dr(release, size, gfp); + dr = alloc_dr(release, size, gfp | __GFP_ZERO); if (unlikely(!dr)) return NULL; return dr->data; @@ -745,58 +746,62 @@ void devm_remove_action(struct device *dev, void (*action)(void *), void *data) EXPORT_SYMBOL_GPL(devm_remove_action); /* - * Managed kzalloc/kfree + * Managed kmalloc/kfree */ -static void devm_kzalloc_release(struct device *dev, void *res) +static void devm_kmalloc_release(struct device *dev, void *res) { /* noop */ } -static int devm_kzalloc_match(struct device *dev, void *res, void *data) +static int devm_kmalloc_match(struct device *dev, void *res, void *data) { return res == data; } /** - * devm_kzalloc - Resource-managed kzalloc + * devm_kmalloc - Resource-managed kmalloc * @dev: Device to allocate memory for * @size: Allocation size * @gfp: Allocation gfp flags * - * Managed kzalloc. Memory allocated with this function is + * Managed kmalloc. Memory allocated with this function is * automatically freed on driver detach. Like all other devres * resources, guaranteed alignment is unsigned long long. * * RETURNS: * Pointer to allocated memory on success, NULL on failure. */ -void * devm_kzalloc(struct device *dev, size_t size, gfp_t gfp) +void * devm_kmalloc(struct device *dev, size_t size, gfp_t gfp) { struct devres *dr; /* use raw alloc_dr for kmalloc caller tracing */ - dr = alloc_dr(devm_kzalloc_release, size, gfp); + dr = alloc_dr(devm_kmalloc_release, size, gfp); if (unlikely(!dr)) return NULL; + /* + * This is named devm_kzalloc_release for historical reasons + * The initial implementation did not support kmalloc, only kzalloc + */ set_node_dbginfo(&dr->node, "devm_kzalloc_release", size); devres_add(dev, dr->data); return dr->data; } -EXPORT_SYMBOL_GPL(devm_kzalloc); +EXPORT_SYMBOL_GPL(devm_kmalloc); /** * devm_kfree - Resource-managed kfree * @dev: Device this memory belongs to * @p: Memory to free * - * Free memory allocated with devm_kzalloc(). + * Free memory allocated with devm_kmalloc(). */ void devm_kfree(struct device *dev, void *p) { int rc; - rc = devres_destroy(dev, devm_kzalloc_release, devm_kzalloc_match, p); + rc = devres_destroy(dev, devm_kmalloc_release, devm_kmalloc_match, p); WARN_ON(rc); } EXPORT_SYMBOL_GPL(devm_kfree); diff --git a/drivers/base/dma-contiguous.c b/drivers/base/dma-contiguous.c index 0ca54421ce97..99802d6f3c60 100644 --- a/drivers/base/dma-contiguous.c +++ b/drivers/base/dma-contiguous.c @@ -96,7 +96,7 @@ static inline __maybe_unused phys_addr_t cma_early_percent_memory(void) #endif /** - * dma_contiguous_reserve() - reserve area for contiguous memory handling + * dma_contiguous_reserve() - reserve area(s) for contiguous memory handling * @limit: End address of the reserved memory (optional, 0 for any). * * This function reserves memory from early allocator. It should be @@ -124,22 +124,29 @@ void __init dma_contiguous_reserve(phys_addr_t limit) #endif } - if (selected_size) { + if (selected_size && !dma_contiguous_default_area) { pr_debug("%s: reserving %ld MiB for global area\n", __func__, (unsigned long)selected_size / SZ_1M); - dma_declare_contiguous(NULL, selected_size, 0, limit); + dma_contiguous_reserve_area(selected_size, 0, limit, + &dma_contiguous_default_area); } }; static DEFINE_MUTEX(cma_mutex); -static __init int cma_activate_area(unsigned long base_pfn, unsigned long count) +static int __init cma_activate_area(struct cma *cma) { - unsigned long pfn = base_pfn; - unsigned i = count >> pageblock_order; + int bitmap_size = BITS_TO_LONGS(cma->count) * sizeof(long); + unsigned long base_pfn = cma->base_pfn, pfn = base_pfn; + unsigned i = cma->count >> pageblock_order; struct zone *zone; + cma->bitmap = kzalloc(bitmap_size, GFP_KERNEL); + + if (!cma->bitmap) + return -ENOMEM; + WARN_ON_ONCE(!pfn_valid(pfn)); zone = page_zone(pfn_to_page(pfn)); @@ -153,92 +160,53 @@ static __init int cma_activate_area(unsigned long base_pfn, unsigned long count) } init_cma_reserved_pageblock(pfn_to_page(base_pfn)); } while (--i); - return 0; -} - -static __init struct cma *cma_create_area(unsigned long base_pfn, - unsigned long count) -{ - int bitmap_size = BITS_TO_LONGS(count) * sizeof(long); - struct cma *cma; - int ret = -ENOMEM; - - pr_debug("%s(base %08lx, count %lx)\n", __func__, base_pfn, count); - - cma = kmalloc(sizeof *cma, GFP_KERNEL); - if (!cma) - return ERR_PTR(-ENOMEM); - - cma->base_pfn = base_pfn; - cma->count = count; - cma->bitmap = kzalloc(bitmap_size, GFP_KERNEL); - if (!cma->bitmap) - goto no_mem; - - ret = cma_activate_area(base_pfn, count); - if (ret) - goto error; - - pr_debug("%s: returned %p\n", __func__, (void *)cma); - return cma; - -error: - kfree(cma->bitmap); -no_mem: - kfree(cma); - return ERR_PTR(ret); + return 0; } -static struct cma_reserved { - phys_addr_t start; - unsigned long size; - struct device *dev; -} cma_reserved[MAX_CMA_AREAS] __initdata; -static unsigned cma_reserved_count __initdata; +static struct cma cma_areas[MAX_CMA_AREAS]; +static unsigned cma_area_count; static int __init cma_init_reserved_areas(void) { - struct cma_reserved *r = cma_reserved; - unsigned i = cma_reserved_count; - - pr_debug("%s()\n", __func__); + int i; - for (; i; --i, ++r) { - struct cma *cma; - cma = cma_create_area(PFN_DOWN(r->start), - r->size >> PAGE_SHIFT); - if (!IS_ERR(cma)) - dev_set_cma_area(r->dev, cma); + for (i = 0; i < cma_area_count; i++) { + int ret = cma_activate_area(&cma_areas[i]); + if (ret) + return ret; } + return 0; } core_initcall(cma_init_reserved_areas); /** - * dma_declare_contiguous() - reserve area for contiguous memory handling - * for particular device - * @dev: Pointer to device structure. - * @size: Size of the reserved memory. - * @base: Start address of the reserved memory (optional, 0 for any). + * dma_contiguous_reserve_area() - reserve custom contiguous area + * @size: Size of the reserved area (in bytes), + * @base: Base address of the reserved area optional, use 0 for any * @limit: End address of the reserved memory (optional, 0 for any). + * @res_cma: Pointer to store the created cma region. * - * This function reserves memory for specified device. It should be - * called by board specific code when early allocator (memblock or bootmem) - * is still activate. + * This function reserves memory from early allocator. It should be + * called by arch specific code once the early allocator (memblock or bootmem) + * has been activated and all other subsystems have already allocated/reserved + * memory. This function allows to create custom reserved areas for specific + * devices. */ -int __init dma_declare_contiguous(struct device *dev, phys_addr_t size, - phys_addr_t base, phys_addr_t limit) +int __init dma_contiguous_reserve_area(phys_addr_t size, phys_addr_t base, + phys_addr_t limit, struct cma **res_cma) { - struct cma_reserved *r = &cma_reserved[cma_reserved_count]; + struct cma *cma = &cma_areas[cma_area_count]; phys_addr_t alignment; + int ret = 0; pr_debug("%s(size %lx, base %08lx, limit %08lx)\n", __func__, (unsigned long)size, (unsigned long)base, (unsigned long)limit); /* Sanity checks */ - if (cma_reserved_count == ARRAY_SIZE(cma_reserved)) { + if (cma_area_count == ARRAY_SIZE(cma_areas)) { pr_err("Not enough slots for CMA reserved regions!\n"); return -ENOSPC; } @@ -256,7 +224,7 @@ int __init dma_declare_contiguous(struct device *dev, phys_addr_t size, if (base) { if (memblock_is_region_reserved(base, size) || memblock_reserve(base, size) < 0) { - base = -EBUSY; + ret = -EBUSY; goto err; } } else { @@ -266,7 +234,7 @@ int __init dma_declare_contiguous(struct device *dev, phys_addr_t size, */ phys_addr_t addr = __memblock_alloc_base(size, alignment, limit); if (!addr) { - base = -ENOMEM; + ret = -ENOMEM; goto err; } else { base = addr; @@ -277,10 +245,11 @@ int __init dma_declare_contiguous(struct device *dev, phys_addr_t size, * Each reserved area must be initialised later, when more kernel * subsystems (like slab allocator) are available. */ - r->start = base; - r->size = size; - r->dev = dev; - cma_reserved_count++; + cma->base_pfn = PFN_DOWN(base); + cma->count = size >> PAGE_SHIFT; + *res_cma = cma; + cma_area_count++; + pr_info("CMA: reserved %ld MiB at %08lx\n", (unsigned long)size / SZ_1M, (unsigned long)base); @@ -289,7 +258,7 @@ int __init dma_declare_contiguous(struct device *dev, phys_addr_t size, return 0; err: pr_err("CMA: failed to reserve %ld MiB\n", (unsigned long)size / SZ_1M); - return base; + return ret; } /** diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index 5a9b6569dd74..220ec3a3ca75 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -28,6 +28,8 @@ #include <linux/sched.h> #include <linux/async.h> #include <linux/suspend.h> +#include <trace/events/power.h> +#include <linux/cpufreq.h> #include <linux/cpuidle.h> #include "../base.h" #include "power.h" @@ -713,6 +715,8 @@ void dpm_resume(pm_message_t state) mutex_unlock(&dpm_list_mtx); async_synchronize_full(); dpm_show_time(starttime, state, NULL); + + cpufreq_resume(); } /** @@ -1177,6 +1181,8 @@ int dpm_suspend(pm_message_t state) might_sleep(); + cpufreq_suspend(); + mutex_lock(&dpm_list_mtx); pm_transition = state; async_error = 0; diff --git a/drivers/bus/Kconfig b/drivers/bus/Kconfig index b05ecab915c4..5286e2d333b0 100644 --- a/drivers/bus/Kconfig +++ b/drivers/bus/Kconfig @@ -26,4 +26,11 @@ config OMAP_INTERCONNECT help Driver to enable OMAP interconnect error handling driver. + +config ARM_CCI + bool "ARM CCI driver support" + depends on ARM + help + Driver supporting the CCI cache coherent interconnect for ARM + platforms. endmenu diff --git a/drivers/bus/Makefile b/drivers/bus/Makefile index 3c7b53c12091..670cea443802 100644 --- a/drivers/bus/Makefile +++ b/drivers/bus/Makefile @@ -7,3 +7,5 @@ obj-$(CONFIG_OMAP_OCP2SCP) += omap-ocp2scp.o # Interconnect bus driver for OMAP SoCs. obj-$(CONFIG_OMAP_INTERCONNECT) += omap_l3_smx.o omap_l3_noc.o +# CCI cache coherent interconnect for ARM platforms +obj-$(CONFIG_ARM_CCI) += arm-cci.o diff --git a/drivers/bus/arm-cci.c b/drivers/bus/arm-cci.c new file mode 100644 index 000000000000..d9d954eb0fa0 --- /dev/null +++ b/drivers/bus/arm-cci.c @@ -0,0 +1,947 @@ +/* + * CCI cache coherent interconnect driver + * + * Copyright (C) 2013 ARM Ltd. + * Author: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * This program is distributed "as is" WITHOUT ANY WARRANTY of any + * kind, whether express or implied; without even the implied warranty + * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#include <linux/arm-cci.h> +#include <linux/io.h> +#include <linux/module.h> +#include <linux/platform_device.h> +#include <linux/of_address.h> +#include <linux/slab.h> + +#include <asm/cacheflush.h> +#include <asm/irq_regs.h> +#include <asm/pmu.h> +#include <asm/smp_plat.h> + +#define DRIVER_NAME "CCI" + +#define CCI_PORT_CTRL 0x0 +#define CCI_CTRL_STATUS 0xc + +#define CCI_ENABLE_SNOOP_REQ 0x1 +#define CCI_ENABLE_DVM_REQ 0x2 +#define CCI_ENABLE_REQ (CCI_ENABLE_SNOOP_REQ | CCI_ENABLE_DVM_REQ) + +struct cci_nb_ports { + unsigned int nb_ace; + unsigned int nb_ace_lite; +}; + +enum cci_ace_port_type { + ACE_INVALID_PORT = 0x0, + ACE_PORT, + ACE_LITE_PORT, +}; + +struct cci_ace_port { + void __iomem *base; + unsigned long phys; + enum cci_ace_port_type type; + struct device_node *dn; +}; + +static struct cci_ace_port *ports; +static unsigned int nb_cci_ports; + +static void __iomem *cci_ctrl_base; +static unsigned long cci_ctrl_phys; + +#ifdef CONFIG_HW_PERF_EVENTS + +static void __iomem *cci_pmu_base; + +#define CCI400_PMCR 0x0100 + +#define CCI400_PMU_CYCLE_CNTR_BASE 0x0000 +#define CCI400_PMU_CNTR_BASE(idx) (CCI400_PMU_CYCLE_CNTR_BASE + (idx) * 0x1000) + +#define CCI400_PMCR_CEN 0x00000001 +#define CCI400_PMCR_RST 0x00000002 +#define CCI400_PMCR_CCR 0x00000004 +#define CCI400_PMCR_CCD 0x00000008 +#define CCI400_PMCR_EX 0x00000010 +#define CCI400_PMCR_DP 0x00000020 +#define CCI400_PMCR_NCNT_MASK 0x0000F800 +#define CCI400_PMCR_NCNT_SHIFT 11 + +#define CCI400_PMU_EVT_SEL 0x000 +#define CCI400_PMU_CNTR 0x004 +#define CCI400_PMU_CNTR_CTRL 0x008 +#define CCI400_PMU_OVERFLOW 0x00C + +#define CCI400_PMU_OVERFLOW_FLAG 1 + +enum cci400_perf_events { + CCI400_PMU_CYCLES = 0xFF +}; + +#define CCI400_PMU_EVENT_MASK 0xff +#define CCI400_PMU_EVENT_SOURCE(event) ((event >> 5) & 0x7) +#define CCI400_PMU_EVENT_CODE(event) (event & 0x1f) + +#define CCI400_PMU_EVENT_SOURCE_S0 0 +#define CCI400_PMU_EVENT_SOURCE_S4 4 +#define CCI400_PMU_EVENT_SOURCE_M0 5 +#define CCI400_PMU_EVENT_SOURCE_M2 7 + +#define CCI400_PMU_EVENT_SLAVE_MIN 0x0 +#define CCI400_PMU_EVENT_SLAVE_MAX 0x13 + +#define CCI400_PMU_EVENT_MASTER_MIN 0x14 +#define CCI400_PMU_EVENT_MASTER_MAX 0x1A + +#define CCI400_PMU_MAX_HW_EVENTS 5 /* CCI PMU has 4 counters + 1 cycle counter */ + +#define CCI400_PMU_CYCLE_COUNTER_IDX 0 +#define CCI400_PMU_COUNTER0_IDX 1 +#define CCI400_PMU_COUNTER_LAST(cci_pmu) (CCI400_PMU_CYCLE_COUNTER_IDX + cci_pmu->num_events - 1) + + +static struct perf_event *events[CCI400_PMU_MAX_HW_EVENTS]; +static unsigned long used_mask[BITS_TO_LONGS(CCI400_PMU_MAX_HW_EVENTS)]; +static struct pmu_hw_events cci_hw_events = { + .events = events, + .used_mask = used_mask, +}; + +static int cci_pmu_validate_hw_event(u8 hw_event) +{ + u8 ev_source = CCI400_PMU_EVENT_SOURCE(hw_event); + u8 ev_code = CCI400_PMU_EVENT_CODE(hw_event); + + if (ev_source <= CCI400_PMU_EVENT_SOURCE_S4 && + ev_code <= CCI400_PMU_EVENT_SLAVE_MAX) + return hw_event; + else if (CCI400_PMU_EVENT_SOURCE_M0 <= ev_source && + ev_source <= CCI400_PMU_EVENT_SOURCE_M2 && + CCI400_PMU_EVENT_MASTER_MIN <= ev_code && + ev_code <= CCI400_PMU_EVENT_MASTER_MAX) + return hw_event; + + return -EINVAL; +} + +static inline int cci_pmu_counter_is_valid(struct arm_pmu *cci_pmu, int idx) +{ + return CCI400_PMU_CYCLE_COUNTER_IDX <= idx && + idx <= CCI400_PMU_COUNTER_LAST(cci_pmu); +} + +static inline u32 cci_pmu_read_register(int idx, unsigned int offset) +{ + return readl_relaxed(cci_pmu_base + CCI400_PMU_CNTR_BASE(idx) + offset); +} + +static inline void cci_pmu_write_register(u32 value, int idx, unsigned int offset) +{ + return writel_relaxed(value, cci_pmu_base + CCI400_PMU_CNTR_BASE(idx) + offset); +} + +static inline void cci_pmu_disable_counter(int idx) +{ + cci_pmu_write_register(0, idx, CCI400_PMU_CNTR_CTRL); +} + +static inline void cci_pmu_enable_counter(int idx) +{ + cci_pmu_write_register(1, idx, CCI400_PMU_CNTR_CTRL); +} + +static inline void cci_pmu_select_event(int idx, unsigned long event) +{ + event &= CCI400_PMU_EVENT_MASK; + cci_pmu_write_register(event, idx, CCI400_PMU_EVT_SEL); +} + +static u32 cci_pmu_get_max_counters(void) +{ + u32 n_cnts = (readl_relaxed(cci_ctrl_base + CCI400_PMCR) & + CCI400_PMCR_NCNT_MASK) >> CCI400_PMCR_NCNT_SHIFT; + + /* add 1 for cycle counter */ + return n_cnts + 1; +} + +static struct pmu_hw_events *cci_pmu_get_hw_events(void) +{ + return &cci_hw_events; +} + +static int cci_pmu_get_event_idx(struct pmu_hw_events *hw, struct perf_event *event) +{ + struct arm_pmu *cci_pmu = to_arm_pmu(event->pmu); + struct hw_perf_event *hw_event = &event->hw; + unsigned long cci_event = hw_event->config_base & CCI400_PMU_EVENT_MASK; + int idx; + + if (cci_event == CCI400_PMU_CYCLES) { + if (test_and_set_bit(CCI400_PMU_CYCLE_COUNTER_IDX, hw->used_mask)) + return -EAGAIN; + + return CCI400_PMU_CYCLE_COUNTER_IDX; + } + + for (idx = CCI400_PMU_COUNTER0_IDX; idx <= CCI400_PMU_COUNTER_LAST(cci_pmu); ++idx) { + if (!test_and_set_bit(idx, hw->used_mask)) + return idx; + } + + /* No counters available */ + return -EAGAIN; +} + +static int cci_pmu_map_event(struct perf_event *event) +{ + int mapping; + u8 config = event->attr.config & CCI400_PMU_EVENT_MASK; + + if (event->attr.type < PERF_TYPE_MAX) + return -ENOENT; + + /* 0xff is used to represent CCI Cycles */ + if (config == 0xff) + mapping = config; + else + mapping = cci_pmu_validate_hw_event(config); + + return mapping; +} + +static int cci_pmu_request_irq(struct arm_pmu *cci_pmu, irq_handler_t handler) +{ + int irq, err, i = 0; + struct platform_device *pmu_device = cci_pmu->plat_device; + + if (unlikely(!pmu_device)) + return -ENODEV; + + /* CCI exports 6 interrupts - 1 nERRORIRQ + 5 nEVNTCNTOVERFLOW (PMU) + nERRORIRQ will be handled by secure firmware on TC2. So we + assume that all CCI interrupts listed in the linux device + tree are PMU interrupts. + + The following code should then be able to handle different routing + of the CCI PMU interrupts. + */ + while ((irq = platform_get_irq(pmu_device, i)) > 0) { + err = request_irq(irq, handler, 0, "arm-cci-pmu", cci_pmu); + if (err) { + dev_err(&pmu_device->dev, "unable to request IRQ%d for ARM CCI PMU counters\n", + irq); + return err; + } + i++; + } + + return 0; +} + +static irqreturn_t cci_pmu_handle_irq(int irq_num, void *dev) +{ + struct arm_pmu *cci_pmu = (struct arm_pmu *)dev; + struct pmu_hw_events *events = cci_pmu->get_hw_events(); + struct perf_sample_data data; + struct pt_regs *regs; + int idx; + + regs = get_irq_regs(); + + /* Iterate over counters and update the corresponding perf events. + This should work regardless of whether we have per-counter overflow + interrupt or a combined overflow interrupt. */ + for (idx = CCI400_PMU_CYCLE_COUNTER_IDX; idx <= CCI400_PMU_COUNTER_LAST(cci_pmu); idx++) { + struct perf_event *event = events->events[idx]; + struct hw_perf_event *hw_counter; + + if (!event) + continue; + + hw_counter = &event->hw; + + /* Did this counter overflow? */ + if (!(cci_pmu_read_register(idx, CCI400_PMU_OVERFLOW) & CCI400_PMU_OVERFLOW_FLAG)) + continue; + cci_pmu_write_register(CCI400_PMU_OVERFLOW_FLAG, idx, CCI400_PMU_OVERFLOW); + + armpmu_event_update(event); + perf_sample_data_init(&data, 0, hw_counter->last_period); + if (!armpmu_event_set_period(event)) + continue; + + if (perf_event_overflow(event, &data, regs)) + cci_pmu->disable(event); + } + + irq_work_run(); + return IRQ_HANDLED; +} + +static void cci_pmu_free_irq(struct arm_pmu *cci_pmu) +{ + int irq, i = 0; + struct platform_device *pmu_device = cci_pmu->plat_device; + + while ((irq = platform_get_irq(pmu_device, i)) > 0) { + free_irq(irq, cci_pmu); + i++; + } +} + +static void cci_pmu_enable_event(struct perf_event *event) +{ + unsigned long flags; + struct arm_pmu *cci_pmu = to_arm_pmu(event->pmu); + struct pmu_hw_events *events = cci_pmu->get_hw_events(); + struct hw_perf_event *hw_counter = &event->hw; + int idx = hw_counter->idx; + + if (unlikely(!cci_pmu_counter_is_valid(cci_pmu, idx))) { + dev_err(&cci_pmu->plat_device->dev, "Invalid CCI PMU counter %d\n", idx); + return; + } + + raw_spin_lock_irqsave(&events->pmu_lock, flags); + + /* Configure the event to count, unless you are counting cycles */ + if (idx != CCI400_PMU_CYCLE_COUNTER_IDX) + cci_pmu_select_event(idx, hw_counter->config_base); + + cci_pmu_enable_counter(idx); + + raw_spin_unlock_irqrestore(&events->pmu_lock, flags); +} + +static void cci_pmu_disable_event(struct perf_event *event) +{ + unsigned long flags; + struct arm_pmu *cci_pmu = to_arm_pmu(event->pmu); + struct pmu_hw_events *events = cci_pmu->get_hw_events(); + struct hw_perf_event *hw_counter = &event->hw; + int idx = hw_counter->idx; + + if (unlikely(!cci_pmu_counter_is_valid(cci_pmu, idx))) { + dev_err(&cci_pmu->plat_device->dev, "Invalid CCI PMU counter %d\n", idx); + return; + } + + raw_spin_lock_irqsave(&events->pmu_lock, flags); + + cci_pmu_disable_counter(idx); + + raw_spin_unlock_irqrestore(&events->pmu_lock, flags); +} + +static void cci_pmu_start(struct arm_pmu *cci_pmu) +{ + u32 val; + unsigned long flags; + struct pmu_hw_events *events = cci_pmu->get_hw_events(); + + raw_spin_lock_irqsave(&events->pmu_lock, flags); + + /* Enable all the PMU counters. */ + val = readl(cci_ctrl_base + CCI400_PMCR) | CCI400_PMCR_CEN; + writel(val, cci_ctrl_base + CCI400_PMCR); + + raw_spin_unlock_irqrestore(&events->pmu_lock, flags); +} + +static void cci_pmu_stop(struct arm_pmu *cci_pmu) +{ + u32 val; + unsigned long flags; + struct pmu_hw_events *events = cci_pmu->get_hw_events(); + + raw_spin_lock_irqsave(&events->pmu_lock, flags); + + /* Disable all the PMU counters. */ + val = readl(cci_ctrl_base + CCI400_PMCR) & ~CCI400_PMCR_CEN; + writel(val, cci_ctrl_base + CCI400_PMCR); + + raw_spin_unlock_irqrestore(&events->pmu_lock, flags); +} + +static u32 cci_pmu_read_counter(struct perf_event *event) +{ + struct arm_pmu *cci_pmu = to_arm_pmu(event->pmu); + struct hw_perf_event *hw_counter = &event->hw; + int idx = hw_counter->idx; + u32 value; + + if (unlikely(!cci_pmu_counter_is_valid(cci_pmu, idx))) { + dev_err(&cci_pmu->plat_device->dev, "Invalid CCI PMU counter %d\n", idx); + return 0; + } + value = cci_pmu_read_register(idx, CCI400_PMU_CNTR); + + return value; +} + +static void cci_pmu_write_counter(struct perf_event *event, u32 value) +{ + struct arm_pmu *cci_pmu = to_arm_pmu(event->pmu); + struct hw_perf_event *hw_counter = &event->hw; + int idx = hw_counter->idx; + + if (unlikely(!cci_pmu_counter_is_valid(cci_pmu, idx))) + dev_err(&cci_pmu->plat_device->dev, "Invalid CCI PMU counter %d\n", idx); + else + cci_pmu_write_register(value, idx, CCI400_PMU_CNTR); +} + +static struct arm_pmu cci_pmu = { + .name = DRIVER_NAME, + .max_period = (1LLU << 32) - 1, + .get_hw_events = cci_pmu_get_hw_events, + .get_event_idx = cci_pmu_get_event_idx, + .map_event = cci_pmu_map_event, + .request_irq = cci_pmu_request_irq, + .handle_irq = cci_pmu_handle_irq, + .free_irq = cci_pmu_free_irq, + .enable = cci_pmu_enable_event, + .disable = cci_pmu_disable_event, + .start = cci_pmu_start, + .stop = cci_pmu_stop, + .read_counter = cci_pmu_read_counter, + .write_counter = cci_pmu_write_counter, +}; + +static int cci_pmu_probe(struct platform_device *pdev) +{ + struct resource *res; + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + cci_pmu_base = devm_ioremap_resource(&pdev->dev, res); + if (IS_ERR(cci_pmu_base)) + return PTR_ERR(cci_pmu_base); + + cci_pmu.plat_device = pdev; + cci_pmu.num_events = cci_pmu_get_max_counters(); + raw_spin_lock_init(&cci_hw_events.pmu_lock); + cpumask_setall(&cci_pmu.valid_cpus); + + return armpmu_register(&cci_pmu, -1); +} + +static const struct of_device_id arm_cci_pmu_matches[] = { + {.compatible = "arm,cci-400-pmu"}, + {}, +}; + +static struct platform_driver cci_pmu_platform_driver = { + .driver = { + .name = DRIVER_NAME, + .of_match_table = arm_cci_pmu_matches, + }, + .probe = cci_pmu_probe, +}; + +static int __init cci_pmu_init(void) +{ + if (platform_driver_register(&cci_pmu_platform_driver)) + WARN(1, "unable to register CCI platform driver\n"); + return 0; +} + +#else + +static int __init cci_pmu_init(void) +{ + return 0; +} + +#endif /* CONFIG_HW_PERF_EVENTS */ + +struct cpu_port { + u64 mpidr; + u32 port; +}; + +/* + * Use the port MSB as valid flag, shift can be made dynamic + * by computing number of bits required for port indexes. + * Code disabling CCI cpu ports runs with D-cache invalidated + * and SCTLR bit clear so data accesses must be kept to a minimum + * to improve performance; for now shift is left static to + * avoid one more data access while disabling the CCI port. + */ +#define PORT_VALID_SHIFT 31 +#define PORT_VALID (0x1 << PORT_VALID_SHIFT) + +static inline void init_cpu_port(struct cpu_port *port, u32 index, u64 mpidr) +{ + port->port = PORT_VALID | index; + port->mpidr = mpidr; +} + +static inline bool cpu_port_is_valid(struct cpu_port *port) +{ + return !!(port->port & PORT_VALID); +} + +static inline bool cpu_port_match(struct cpu_port *port, u64 mpidr) +{ + return port->mpidr == (mpidr & MPIDR_HWID_BITMASK); +} + +static struct cpu_port cpu_port[NR_CPUS]; + +/** + * __cci_ace_get_port - Function to retrieve the port index connected to + * a cpu or device. + * + * @dn: device node of the device to look-up + * @type: port type + * + * Return value: + * - CCI port index if success + * - -ENODEV if failure + */ +static int __cci_ace_get_port(struct device_node *dn, int type) +{ + int i; + bool ace_match; + struct device_node *cci_portn; + + cci_portn = of_parse_phandle(dn, "cci-control-port", 0); + for (i = 0; i < nb_cci_ports; i++) { + ace_match = ports[i].type == type; + if (ace_match && cci_portn == ports[i].dn) + return i; + } + return -ENODEV; +} + +int cci_ace_get_port(struct device_node *dn) +{ + return __cci_ace_get_port(dn, ACE_LITE_PORT); +} +EXPORT_SYMBOL_GPL(cci_ace_get_port); + +static void __init cci_ace_init_ports(void) +{ + int port, ac, cpu; + u64 hwid; + const u32 *cell; + struct device_node *cpun, *cpus; + + cpus = of_find_node_by_path("/cpus"); + if (WARN(!cpus, "Missing cpus node, bailing out\n")) + return; + + if (WARN_ON(of_property_read_u32(cpus, "#address-cells", &ac))) + ac = of_n_addr_cells(cpus); + + /* + * Port index look-up speeds up the function disabling ports by CPU, + * since the logical to port index mapping is done once and does + * not change after system boot. + * The stashed index array is initialized for all possible CPUs + * at probe time. + */ + for_each_child_of_node(cpus, cpun) { + if (of_node_cmp(cpun->type, "cpu")) + continue; + cell = of_get_property(cpun, "reg", NULL); + if (WARN(!cell, "%s: missing reg property\n", cpun->full_name)) + continue; + + hwid = of_read_number(cell, ac); + cpu = get_logical_index(hwid & MPIDR_HWID_BITMASK); + + if (cpu < 0 || !cpu_possible(cpu)) + continue; + port = __cci_ace_get_port(cpun, ACE_PORT); + if (port < 0) + continue; + + init_cpu_port(&cpu_port[cpu], port, cpu_logical_map(cpu)); + } + + for_each_possible_cpu(cpu) { + WARN(!cpu_port_is_valid(&cpu_port[cpu]), + "CPU %u does not have an associated CCI port\n", + cpu); + } +} +/* + * Functions to enable/disable a CCI interconnect slave port + * + * They are called by low-level power management code to disable slave + * interfaces snoops and DVM broadcast. + * Since they may execute with cache data allocation disabled and + * after the caches have been cleaned and invalidated the functions provide + * no explicit locking since they may run with D-cache disabled, so normal + * cacheable kernel locks based on ldrex/strex may not work. + * Locking has to be provided by BSP implementations to ensure proper + * operations. + */ + +/** + * cci_port_control() - function to control a CCI port + * + * @port: index of the port to setup + * @enable: if true enables the port, if false disables it + */ +static void notrace cci_port_control(unsigned int port, bool enable) +{ + void __iomem *base = ports[port].base; + + writel_relaxed(enable ? CCI_ENABLE_REQ : 0, base + CCI_PORT_CTRL); + /* + * This function is called from power down procedures + * and must not execute any instruction that might + * cause the processor to be put in a quiescent state + * (eg wfi). Hence, cpu_relax() can not be added to this + * read loop to optimize power, since it might hide possibly + * disruptive operations. + */ + while (readl_relaxed(cci_ctrl_base + CCI_CTRL_STATUS) & 0x1) + ; +} + +/** + * cci_disable_port_by_cpu() - function to disable a CCI port by CPU + * reference + * + * @mpidr: mpidr of the CPU whose CCI port should be disabled + * + * Disabling a CCI port for a CPU implies disabling the CCI port + * controlling that CPU cluster. Code disabling CPU CCI ports + * must make sure that the CPU running the code is the last active CPU + * in the cluster ie all other CPUs are quiescent in a low power state. + * + * Return: + * 0 on success + * -ENODEV on port look-up failure + */ +int notrace cci_disable_port_by_cpu(u64 mpidr) +{ + int cpu; + bool is_valid; + for (cpu = 0; cpu < nr_cpu_ids; cpu++) { + is_valid = cpu_port_is_valid(&cpu_port[cpu]); + if (is_valid && cpu_port_match(&cpu_port[cpu], mpidr)) { + cci_port_control(cpu_port[cpu].port, false); + return 0; + } + } + return -ENODEV; +} +EXPORT_SYMBOL_GPL(cci_disable_port_by_cpu); + +/** + * cci_enable_port_for_self() - enable a CCI port for calling CPU + * + * Enabling a CCI port for the calling CPU implies enabling the CCI + * port controlling that CPU's cluster. Caller must make sure that the + * CPU running the code is the first active CPU in the cluster and all + * other CPUs are quiescent in a low power state or waiting for this CPU + * to complete the CCI initialization. + * + * Because this is called when the MMU is still off and with no stack, + * the code must be position independent and ideally rely on callee + * clobbered registers only. To achieve this we must code this function + * entirely in assembler. + * + * On success this returns with the proper CCI port enabled. In case of + * any failure this never returns as the inability to enable the CCI is + * fatal and there is no possible recovery at this stage. + */ +asmlinkage void __naked cci_enable_port_for_self(void) +{ + asm volatile ("\n" + +" mrc p15, 0, r0, c0, c0, 5 @ get MPIDR value \n" +" and r0, r0, #"__stringify(MPIDR_HWID_BITMASK)" \n" +" adr r1, 5f \n" +" ldr r2, [r1] \n" +" add r1, r1, r2 @ &cpu_port \n" +" add ip, r1, %[sizeof_cpu_port] \n" + + /* Loop over the cpu_port array looking for a matching MPIDR */ +"1: ldr r2, [r1, %[offsetof_cpu_port_mpidr_lsb]] \n" +" cmp r2, r0 @ compare MPIDR \n" +" bne 2f \n" + + /* Found a match, now test port validity */ +" ldr r3, [r1, %[offsetof_cpu_port_port]] \n" +" tst r3, #"__stringify(PORT_VALID)" \n" +" bne 3f \n" + + /* no match, loop with the next cpu_port entry */ +"2: add r1, r1, %[sizeof_struct_cpu_port] \n" +" cmp r1, ip @ done? \n" +" blo 1b \n" + + /* CCI port not found -- cheaply try to stall this CPU */ +"cci_port_not_found: \n" +" wfi \n" +" wfe \n" +" b cci_port_not_found \n" + + /* Use matched port index to look up the corresponding ports entry */ +"3: bic r3, r3, #"__stringify(PORT_VALID)" \n" +" adr r0, 6f \n" +" ldmia r0, {r1, r2} \n" +" sub r1, r1, r0 @ virt - phys \n" +" ldr r0, [r0, r2] @ *(&ports) \n" +" mov r2, %[sizeof_struct_ace_port] \n" +" mla r0, r2, r3, r0 @ &ports[index] \n" +" sub r0, r0, r1 @ virt_to_phys() \n" + + /* Enable the CCI port */ +" ldr r0, [r0, %[offsetof_port_phys]] \n" +" mov r3, %[cci_enable_req]\n" +" str r3, [r0, #"__stringify(CCI_PORT_CTRL)"] \n" + + /* poll the status reg for completion */ +" adr r1, 7f \n" +" ldr r0, [r1] \n" +" ldr r0, [r0, r1] @ cci_ctrl_base \n" +"4: ldr r1, [r0, #"__stringify(CCI_CTRL_STATUS)"] \n" +" tst r1, %[cci_control_status_bits] \n" +" bne 4b \n" + +" mov r0, #0 \n" +" bx lr \n" + +" .align 2 \n" +"5: .word cpu_port - . \n" +"6: .word . \n" +" .word ports - 6b \n" +"7: .word cci_ctrl_phys - . \n" + : : + [sizeof_cpu_port] "i" (sizeof(cpu_port)), + [cci_enable_req] "i" cpu_to_le32(CCI_ENABLE_REQ), + [cci_control_status_bits] "i" cpu_to_le32(1), +#ifndef __ARMEB__ + [offsetof_cpu_port_mpidr_lsb] "i" (offsetof(struct cpu_port, mpidr)), +#else + [offsetof_cpu_port_mpidr_lsb] "i" (offsetof(struct cpu_port, mpidr)+4), +#endif + [offsetof_cpu_port_port] "i" (offsetof(struct cpu_port, port)), + [sizeof_struct_cpu_port] "i" (sizeof(struct cpu_port)), + [sizeof_struct_ace_port] "i" (sizeof(struct cci_ace_port)), + [offsetof_port_phys] "i" (offsetof(struct cci_ace_port, phys)) ); + + unreachable(); +} + +/** + * __cci_control_port_by_device() - function to control a CCI port by device + * reference + * + * @dn: device node pointer of the device whose CCI port should be + * controlled + * @enable: if true enables the port, if false disables it + * + * Return: + * 0 on success + * -ENODEV on port look-up failure + */ +int notrace __cci_control_port_by_device(struct device_node *dn, bool enable) +{ + int port; + + if (!dn) + return -ENODEV; + + port = __cci_ace_get_port(dn, ACE_LITE_PORT); + if (WARN_ONCE(port < 0, "node %s ACE lite port look-up failure\n", + dn->full_name)) + return -ENODEV; + cci_port_control(port, enable); + return 0; +} +EXPORT_SYMBOL_GPL(__cci_control_port_by_device); + +/** + * __cci_control_port_by_index() - function to control a CCI port by port index + * + * @port: port index previously retrieved with cci_ace_get_port() + * @enable: if true enables the port, if false disables it + * + * Return: + * 0 on success + * -ENODEV on port index out of range + * -EPERM if operation carried out on an ACE PORT + */ +int notrace __cci_control_port_by_index(u32 port, bool enable) +{ + if (port >= nb_cci_ports || ports[port].type == ACE_INVALID_PORT) + return -ENODEV; + /* + * CCI control for ports connected to CPUS is extremely fragile + * and must be made to go through a specific and controlled + * interface (ie cci_disable_port_by_cpu(); control by general purpose + * indexing is therefore disabled for ACE ports. + */ + if (ports[port].type == ACE_PORT) + return -EPERM; + + cci_port_control(port, enable); + return 0; +} +EXPORT_SYMBOL_GPL(__cci_control_port_by_index); + +static const struct cci_nb_ports cci400_ports = { + .nb_ace = 2, + .nb_ace_lite = 3 +}; + +static const struct of_device_id arm_cci_matches[] = { + {.compatible = "arm,cci-400", .data = &cci400_ports }, + {}, +}; + +static const struct of_device_id arm_cci_ctrl_if_matches[] = { + {.compatible = "arm,cci-400-ctrl-if", }, + {}, +}; + +static int __init cci_probe(void) +{ + struct cci_nb_ports const *cci_config; + int ret, i, nb_ace = 0, nb_ace_lite = 0; + struct device_node *np, *cp; + struct resource res; + const char *match_str; + bool is_ace; + + np = of_find_matching_node(NULL, arm_cci_matches); + if (!np) + return -ENODEV; + + cci_config = of_match_node(arm_cci_matches, np)->data; + if (!cci_config) + return -ENODEV; + + nb_cci_ports = cci_config->nb_ace + cci_config->nb_ace_lite; + + ports = kcalloc(sizeof(*ports), nb_cci_ports, GFP_KERNEL); + if (!ports) + return -ENOMEM; + + ret = of_address_to_resource(np, 0, &res); + if (!ret) { + cci_ctrl_base = ioremap(res.start, resource_size(&res)); + cci_ctrl_phys = res.start; + } + if (ret || !cci_ctrl_base) { + WARN(1, "unable to ioremap CCI ctrl\n"); + ret = -ENXIO; + goto memalloc_err; + } + + for_each_child_of_node(np, cp) { + if (!of_match_node(arm_cci_ctrl_if_matches, cp)) + continue; + + i = nb_ace + nb_ace_lite; + + if (i >= nb_cci_ports) + break; + + if (of_property_read_string(cp, "interface-type", + &match_str)) { + WARN(1, "node %s missing interface-type property\n", + cp->full_name); + continue; + } + is_ace = strcmp(match_str, "ace") == 0; + if (!is_ace && strcmp(match_str, "ace-lite")) { + WARN(1, "node %s containing invalid interface-type property, skipping it\n", + cp->full_name); + continue; + } + + ret = of_address_to_resource(cp, 0, &res); + if (!ret) { + ports[i].base = ioremap(res.start, resource_size(&res)); + ports[i].phys = res.start; + } + if (ret || !ports[i].base) { + WARN(1, "unable to ioremap CCI port %d\n", i); + continue; + } + + if (is_ace) { + if (WARN_ON(nb_ace >= cci_config->nb_ace)) + continue; + ports[i].type = ACE_PORT; + ++nb_ace; + } else { + if (WARN_ON(nb_ace_lite >= cci_config->nb_ace_lite)) + continue; + ports[i].type = ACE_LITE_PORT; + ++nb_ace_lite; + } + ports[i].dn = cp; + } + + /* initialize a stashed array of ACE ports to speed-up look-up */ + cci_ace_init_ports(); + + /* + * Multi-cluster systems may need this data when non-coherent, during + * cluster power-up/power-down. Make sure it reaches main memory. + */ + sync_cache_w(&cci_ctrl_base); + sync_cache_w(&cci_ctrl_phys); + sync_cache_w(&ports); + sync_cache_w(&cpu_port); + __sync_cache_range_w(ports, sizeof(*ports) * nb_cci_ports); + pr_info("ARM CCI driver probed\n"); + return 0; + +memalloc_err: + + kfree(ports); + return ret; +} + +static int cci_init_status = -EAGAIN; +static DEFINE_MUTEX(cci_probing); + +static int __init cci_init(void) +{ + if (cci_init_status != -EAGAIN) + return cci_init_status; + + mutex_lock(&cci_probing); + if (cci_init_status == -EAGAIN) + cci_init_status = cci_probe(); + mutex_unlock(&cci_probing); + return cci_init_status; +} + +/* + * To sort out early init calls ordering a helper function is provided to + * check if the CCI driver has beed initialized. Function check if the driver + * has been initialized, if not it calls the init function that probes + * the driver and updates the return value. + */ +bool __init cci_probed(void) +{ + return cci_init() == 0; +} +EXPORT_SYMBOL_GPL(cci_probed); + +early_initcall(cci_init); +core_initcall(cci_pmu_init); +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("ARM CCI support"); diff --git a/drivers/clk/Kconfig b/drivers/clk/Kconfig index 0357ac44638b..d0d9b2124752 100644 --- a/drivers/clk/Kconfig +++ b/drivers/clk/Kconfig @@ -42,7 +42,7 @@ config COMMON_CLK_WM831X config COMMON_CLK_VERSATILE bool "Clock driver for ARM Reference designs" - depends on ARCH_INTEGRATOR || ARCH_REALVIEW || ARCH_VEXPRESS + depends on ARCH_INTEGRATOR || ARCH_REALVIEW || ARCH_VEXPRESS || ARM64 ---help--- Supports clocking on ARM Reference designs: - Integrator/AP and Integrator/CP diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c index 1144e8c7579d..fccac4a41386 100644 --- a/drivers/clk/clk.c +++ b/drivers/clk/clk.c @@ -2084,6 +2084,12 @@ struct clk *of_clk_get_from_provider(struct of_phandle_args *clkspec) return clk; } +int of_clk_get_parent_count(struct device_node *np) +{ + return of_count_phandle_with_args(np, "clocks", "#clock-cells"); +} +EXPORT_SYMBOL_GPL(of_clk_get_parent_count); + const char *of_clk_get_parent_name(struct device_node *np, int index) { struct of_phandle_args clkspec; diff --git a/drivers/clk/versatile/Makefile b/drivers/clk/versatile/Makefile index c16ca787170a..6e76bf87ca87 100644 --- a/drivers/clk/versatile/Makefile +++ b/drivers/clk/versatile/Makefile @@ -4,4 +4,4 @@ obj-$(CONFIG_ARCH_INTEGRATOR) += clk-integrator.o obj-$(CONFIG_INTEGRATOR_IMPD1) += clk-impd1.o obj-$(CONFIG_ARCH_REALVIEW) += clk-realview.o obj-$(CONFIG_ARCH_VEXPRESS) += clk-vexpress.o clk-sp810.o -obj-$(CONFIG_VEXPRESS_CONFIG) += clk-vexpress-osc.o +obj-$(CONFIG_VEXPRESS_CONFIG) += clk-vexpress-osc.o clk-vexpress-spc.o diff --git a/drivers/clk/versatile/clk-vexpress-osc.c b/drivers/clk/versatile/clk-vexpress-osc.c index 8b8798bb93f3..a535c7bf8574 100644 --- a/drivers/clk/versatile/clk-vexpress-osc.c +++ b/drivers/clk/versatile/clk-vexpress-osc.c @@ -107,7 +107,7 @@ void __init vexpress_osc_of_setup(struct device_node *node) osc->func = vexpress_config_func_get_by_node(node); if (!osc->func) { pr_err("Failed to obtain config func for node '%s'!\n", - node->name); + node->full_name); goto error; } @@ -119,7 +119,7 @@ void __init vexpress_osc_of_setup(struct device_node *node) of_property_read_string(node, "clock-output-names", &init.name); if (!init.name) - init.name = node->name; + init.name = node->full_name; init.ops = &vexpress_osc_ops; init.flags = CLK_IS_ROOT; diff --git a/drivers/clk/versatile/clk-vexpress-spc.c b/drivers/clk/versatile/clk-vexpress-spc.c new file mode 100644 index 000000000000..bb566e244b0c --- /dev/null +++ b/drivers/clk/versatile/clk-vexpress-spc.c @@ -0,0 +1,131 @@ +/* + * Copyright (C) 2012 ARM Limited + * Copyright (C) 2012 Linaro + * + * Author: Viresh Kumar <viresh.kumar@linaro.org> + * + * This file is licensed under the terms of the GNU General Public + * License version 2. This program is licensed "as is" without any + * warranty of any kind, whether express or implied. + */ + +/* SPC clock programming interface for Vexpress cpus */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#include <linux/clk-provider.h> +#include <linux/clkdev.h> +#include <linux/err.h> +#include <linux/io.h> +#include <linux/of.h> +#include <linux/slab.h> +#include <linux/types.h> +#include <linux/vexpress.h> + +struct clk_spc { + struct clk_hw hw; + spinlock_t *lock; + int cluster; +}; + +#define to_clk_spc(spc) container_of(spc, struct clk_spc, hw) + +static unsigned long spc_recalc_rate(struct clk_hw *hw, + unsigned long parent_rate) +{ + struct clk_spc *spc = to_clk_spc(hw); + u32 freq; + + if (vexpress_spc_get_performance(spc->cluster, &freq)) { + return -EIO; + pr_err("%s: Failed", __func__); + } + + return freq * 1000; +} + +static long spc_round_rate(struct clk_hw *hw, unsigned long drate, + unsigned long *parent_rate) +{ + return drate; +} + +static int spc_set_rate(struct clk_hw *hw, unsigned long rate, + unsigned long parent_rate) +{ + struct clk_spc *spc = to_clk_spc(hw); + + return vexpress_spc_set_performance(spc->cluster, rate / 1000); +} + +static struct clk_ops clk_spc_ops = { + .recalc_rate = spc_recalc_rate, + .round_rate = spc_round_rate, + .set_rate = spc_set_rate, +}; + +struct clk *vexpress_clk_register_spc(const char *name, int cluster_id) +{ + struct clk_init_data init; + struct clk_spc *spc; + struct clk *clk; + + if (!name) { + pr_err("Invalid name passed"); + return ERR_PTR(-EINVAL); + } + + spc = kzalloc(sizeof(*spc), GFP_KERNEL); + if (!spc) { + pr_err("could not allocate spc clk\n"); + return ERR_PTR(-ENOMEM); + } + + spc->hw.init = &init; + spc->cluster = cluster_id; + + init.name = name; + init.ops = &clk_spc_ops; + init.flags = CLK_IS_ROOT | CLK_GET_RATE_NOCACHE; + init.num_parents = 0; + + clk = clk_register(NULL, &spc->hw); + if (!IS_ERR_OR_NULL(clk)) + return clk; + + pr_err("clk register failed\n"); + kfree(spc); + + return NULL; +} + +#if defined(CONFIG_OF) +void __init vexpress_clk_of_register_spc(void) +{ + char name[14] = "cpu-cluster.X"; + struct device_node *node = NULL; + struct clk *clk; + const u32 *val; + int cluster_id = 0, len; + + if (!of_find_compatible_node(NULL, NULL, "arm,vexpress-spc")) { + pr_debug("%s: No SPC found, Exiting!!\n", __func__); + return; + } + + while ((node = of_find_node_by_name(node, "cluster"))) { + val = of_get_property(node, "reg", &len); + if (val && len == 4) + cluster_id = be32_to_cpup(val); + + name[12] = cluster_id + '0'; + clk = vexpress_clk_register_spc(name, cluster_id); + if (IS_ERR(clk)) + return; + + pr_debug("Registered clock '%s'\n", name); + clk_register_clkdev(clk, NULL, name); + } +} +CLK_OF_DECLARE(spc, "arm,vexpress-spc", vexpress_clk_of_register_spc); +#endif diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig index f151c6cf27c3..593485b02ee5 100644 --- a/drivers/clocksource/Kconfig +++ b/drivers/clocksource/Kconfig @@ -67,6 +67,21 @@ config ARM_ARCH_TIMER bool select CLKSRC_OF if OF +config ARM_ARCH_TIMER_EVTSTREAM + bool "Support for ARM architected timer event stream generation" + default y if ARM_ARCH_TIMER + help + This option enables support for event stream generation based on + the ARM architected timer. It is used for waking up CPUs executing + the wfe instruction at a frequency represented as a power-of-2 + divisor of the clock rate. + The main use of the event stream is wfe-based timeouts of userspace + locking implementations. It might also be useful for imposing timeout + on wfe to safeguard against any programming errors in case an expected + event is not generated. + This must be disabled for hardware validation purposes to detect any + hardware anomalies of missing events. + config CLKSRC_METAG_GENERIC def_bool y if METAG help diff --git a/drivers/clocksource/arm_arch_timer.c b/drivers/clocksource/arm_arch_timer.c index 053d846ab5b1..67bbbd8ae507 100644 --- a/drivers/clocksource/arm_arch_timer.c +++ b/drivers/clocksource/arm_arch_timer.c @@ -13,16 +13,43 @@ #include <linux/device.h> #include <linux/smp.h> #include <linux/cpu.h> +#include <linux/cpu_pm.h> #include <linux/clockchips.h> #include <linux/interrupt.h> #include <linux/of_irq.h> +#include <linux/of_address.h> #include <linux/io.h> +#include <linux/slab.h> #include <asm/arch_timer.h> #include <asm/virt.h> #include <clocksource/arm_arch_timer.h> +#define CNTTIDR 0x08 +#define CNTTIDR_VIRT(n) (BIT(1) << ((n) * 4)) + +#define CNTVCT_LO 0x08 +#define CNTVCT_HI 0x0c +#define CNTFRQ 0x10 +#define CNTP_TVAL 0x28 +#define CNTP_CTL 0x2c +#define CNTV_TVAL 0x38 +#define CNTV_CTL 0x3c + +#define ARCH_CP15_TIMER BIT(0) +#define ARCH_MEM_TIMER BIT(1) +static unsigned arch_timers_present __initdata; + +static void __iomem *arch_counter_base; + +struct arch_timer { + void __iomem *base; + struct clock_event_device evt; +}; + +#define to_arch_timer(e) container_of(e, struct arch_timer, evt) + static u32 arch_timer_rate; enum ppi_nr { @@ -38,19 +65,82 @@ static int arch_timer_ppi[MAX_TIMER_PPI]; static struct clock_event_device __percpu *arch_timer_evt; static bool arch_timer_use_virtual = true; +static bool arch_timer_mem_use_virtual; /* * Architected system timer support. */ -static inline irqreturn_t timer_handler(const int access, +static __always_inline +void arch_timer_reg_write(int access, enum arch_timer_reg reg, u32 val, + struct clock_event_device *clk) +{ + if (access == ARCH_TIMER_MEM_PHYS_ACCESS) { + struct arch_timer *timer = to_arch_timer(clk); + switch (reg) { + case ARCH_TIMER_REG_CTRL: + writel_relaxed(val, timer->base + CNTP_CTL); + break; + case ARCH_TIMER_REG_TVAL: + writel_relaxed(val, timer->base + CNTP_TVAL); + break; + } + } else if (access == ARCH_TIMER_MEM_VIRT_ACCESS) { + struct arch_timer *timer = to_arch_timer(clk); + switch (reg) { + case ARCH_TIMER_REG_CTRL: + writel_relaxed(val, timer->base + CNTV_CTL); + break; + case ARCH_TIMER_REG_TVAL: + writel_relaxed(val, timer->base + CNTV_TVAL); + break; + } + } else { + arch_timer_reg_write_cp15(access, reg, val); + } +} + +static __always_inline +u32 arch_timer_reg_read(int access, enum arch_timer_reg reg, + struct clock_event_device *clk) +{ + u32 val; + + if (access == ARCH_TIMER_MEM_PHYS_ACCESS) { + struct arch_timer *timer = to_arch_timer(clk); + switch (reg) { + case ARCH_TIMER_REG_CTRL: + val = readl_relaxed(timer->base + CNTP_CTL); + break; + case ARCH_TIMER_REG_TVAL: + val = readl_relaxed(timer->base + CNTP_TVAL); + break; + } + } else if (access == ARCH_TIMER_MEM_VIRT_ACCESS) { + struct arch_timer *timer = to_arch_timer(clk); + switch (reg) { + case ARCH_TIMER_REG_CTRL: + val = readl_relaxed(timer->base + CNTV_CTL); + break; + case ARCH_TIMER_REG_TVAL: + val = readl_relaxed(timer->base + CNTV_TVAL); + break; + } + } else { + val = arch_timer_reg_read_cp15(access, reg); + } + + return val; +} + +static __always_inline irqreturn_t timer_handler(const int access, struct clock_event_device *evt) { unsigned long ctrl; - ctrl = arch_timer_reg_read(access, ARCH_TIMER_REG_CTRL); + ctrl = arch_timer_reg_read(access, ARCH_TIMER_REG_CTRL, evt); if (ctrl & ARCH_TIMER_CTRL_IT_STAT) { ctrl |= ARCH_TIMER_CTRL_IT_MASK; - arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl); + arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl, evt); evt->event_handler(evt); return IRQ_HANDLED; } @@ -72,15 +162,30 @@ static irqreturn_t arch_timer_handler_phys(int irq, void *dev_id) return timer_handler(ARCH_TIMER_PHYS_ACCESS, evt); } -static inline void timer_set_mode(const int access, int mode) +static irqreturn_t arch_timer_handler_phys_mem(int irq, void *dev_id) +{ + struct clock_event_device *evt = dev_id; + + return timer_handler(ARCH_TIMER_MEM_PHYS_ACCESS, evt); +} + +static irqreturn_t arch_timer_handler_virt_mem(int irq, void *dev_id) +{ + struct clock_event_device *evt = dev_id; + + return timer_handler(ARCH_TIMER_MEM_VIRT_ACCESS, evt); +} + +static __always_inline void timer_set_mode(const int access, int mode, + struct clock_event_device *clk) { unsigned long ctrl; switch (mode) { case CLOCK_EVT_MODE_UNUSED: case CLOCK_EVT_MODE_SHUTDOWN: - ctrl = arch_timer_reg_read(access, ARCH_TIMER_REG_CTRL); + ctrl = arch_timer_reg_read(access, ARCH_TIMER_REG_CTRL, clk); ctrl &= ~ARCH_TIMER_CTRL_ENABLE; - arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl); + arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl, clk); break; default: break; @@ -90,60 +195,121 @@ static inline void timer_set_mode(const int access, int mode) static void arch_timer_set_mode_virt(enum clock_event_mode mode, struct clock_event_device *clk) { - timer_set_mode(ARCH_TIMER_VIRT_ACCESS, mode); + timer_set_mode(ARCH_TIMER_VIRT_ACCESS, mode, clk); } static void arch_timer_set_mode_phys(enum clock_event_mode mode, struct clock_event_device *clk) { - timer_set_mode(ARCH_TIMER_PHYS_ACCESS, mode); + timer_set_mode(ARCH_TIMER_PHYS_ACCESS, mode, clk); +} + +static void arch_timer_set_mode_virt_mem(enum clock_event_mode mode, + struct clock_event_device *clk) +{ + timer_set_mode(ARCH_TIMER_MEM_VIRT_ACCESS, mode, clk); +} + +static void arch_timer_set_mode_phys_mem(enum clock_event_mode mode, + struct clock_event_device *clk) +{ + timer_set_mode(ARCH_TIMER_MEM_PHYS_ACCESS, mode, clk); } -static inline void set_next_event(const int access, unsigned long evt) +static __always_inline void set_next_event(const int access, unsigned long evt, + struct clock_event_device *clk) { unsigned long ctrl; - ctrl = arch_timer_reg_read(access, ARCH_TIMER_REG_CTRL); + ctrl = arch_timer_reg_read(access, ARCH_TIMER_REG_CTRL, clk); ctrl |= ARCH_TIMER_CTRL_ENABLE; ctrl &= ~ARCH_TIMER_CTRL_IT_MASK; - arch_timer_reg_write(access, ARCH_TIMER_REG_TVAL, evt); - arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl); + arch_timer_reg_write(access, ARCH_TIMER_REG_TVAL, evt, clk); + arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl, clk); } static int arch_timer_set_next_event_virt(unsigned long evt, - struct clock_event_device *unused) + struct clock_event_device *clk) { - set_next_event(ARCH_TIMER_VIRT_ACCESS, evt); + set_next_event(ARCH_TIMER_VIRT_ACCESS, evt, clk); return 0; } static int arch_timer_set_next_event_phys(unsigned long evt, - struct clock_event_device *unused) + struct clock_event_device *clk) { - set_next_event(ARCH_TIMER_PHYS_ACCESS, evt); + set_next_event(ARCH_TIMER_PHYS_ACCESS, evt, clk); return 0; } -static int __cpuinit arch_timer_setup(struct clock_event_device *clk) +static void arch_timer_configure_evtstream(void) { - clk->features = CLOCK_EVT_FEAT_ONESHOT | CLOCK_EVT_FEAT_C3STOP; - clk->name = "arch_sys_timer"; - clk->rating = 450; - if (arch_timer_use_virtual) { - clk->irq = arch_timer_ppi[VIRT_PPI]; - clk->set_mode = arch_timer_set_mode_virt; - clk->set_next_event = arch_timer_set_next_event_virt; + int evt_stream_div, pos; + + /* Find the closest power of two to the divisor */ + evt_stream_div = arch_timer_rate / ARCH_TIMER_EVT_STREAM_FREQ; + pos = fls(evt_stream_div); + if (pos > 1 && !(evt_stream_div & (1 << (pos - 2)))) + pos--; + /* enable event stream */ + arch_timer_evtstrm_enable(min(pos, 15)); +} + +static int arch_timer_set_next_event_virt_mem(unsigned long evt, + struct clock_event_device *clk) +{ + set_next_event(ARCH_TIMER_MEM_VIRT_ACCESS, evt, clk); + return 0; +} + +static int arch_timer_set_next_event_phys_mem(unsigned long evt, + struct clock_event_device *clk) +{ + set_next_event(ARCH_TIMER_MEM_PHYS_ACCESS, evt, clk); + return 0; +} + +static void __cpuinit __arch_timer_setup(unsigned type, + struct clock_event_device *clk) +{ + clk->features = CLOCK_EVT_FEAT_ONESHOT; + + if (type == ARCH_CP15_TIMER) { + clk->features |= CLOCK_EVT_FEAT_C3STOP; + clk->name = "arch_sys_timer"; + clk->rating = 450; + clk->cpumask = cpumask_of(smp_processor_id()); + if (arch_timer_use_virtual) { + clk->irq = arch_timer_ppi[VIRT_PPI]; + clk->set_mode = arch_timer_set_mode_virt; + clk->set_next_event = arch_timer_set_next_event_virt; + } else { + clk->irq = arch_timer_ppi[PHYS_SECURE_PPI]; + clk->set_mode = arch_timer_set_mode_phys; + clk->set_next_event = arch_timer_set_next_event_phys; + } } else { - clk->irq = arch_timer_ppi[PHYS_SECURE_PPI]; - clk->set_mode = arch_timer_set_mode_phys; - clk->set_next_event = arch_timer_set_next_event_phys; + clk->name = "arch_mem_timer"; + clk->rating = 400; + clk->cpumask = cpu_all_mask; + if (arch_timer_mem_use_virtual) { + clk->set_mode = arch_timer_set_mode_virt_mem; + clk->set_next_event = + arch_timer_set_next_event_virt_mem; + } else { + clk->set_mode = arch_timer_set_mode_phys_mem; + clk->set_next_event = + arch_timer_set_next_event_phys_mem; + } } - clk->cpumask = cpumask_of(smp_processor_id()); + clk->set_mode(CLOCK_EVT_MODE_SHUTDOWN, clk); - clk->set_mode(CLOCK_EVT_MODE_SHUTDOWN, NULL); + clockevents_config_and_register(clk, arch_timer_rate, 0xf, 0x7fffffff); +} - clockevents_config_and_register(clk, arch_timer_rate, - 0xf, 0x7fffffff); +static int __cpuinit arch_timer_setup(struct clock_event_device *clk) +{ + __arch_timer_setup(ARCH_CP15_TIMER, clk); if (arch_timer_use_virtual) enable_percpu_irq(arch_timer_ppi[VIRT_PPI], 0); @@ -154,31 +320,47 @@ static int __cpuinit arch_timer_setup(struct clock_event_device *clk) } arch_counter_set_user_access(); + if (IS_ENABLED(CONFIG_ARM_ARCH_TIMER_EVTSTREAM)) + arch_timer_configure_evtstream(); return 0; } -static int arch_timer_available(void) +static void +arch_timer_detect_rate(void __iomem *cntbase, struct device_node *np) { - u32 freq; - - if (arch_timer_rate == 0) { - freq = arch_timer_get_cntfrq(); - - /* Check the timer frequency. */ - if (freq == 0) { - pr_warn("Architected timer frequency not available\n"); - return -EINVAL; - } + /* Who has more than one independent system counter? */ + if (arch_timer_rate) + return; - arch_timer_rate = freq; + /* Try to determine the frequency from the device tree or CNTFRQ */ + if (of_property_read_u32(np, "clock-frequency", &arch_timer_rate)) { + if (cntbase) + arch_timer_rate = readl_relaxed(cntbase + CNTFRQ); + else + arch_timer_rate = arch_timer_get_cntfrq(); } - pr_info_once("Architected local timer running at %lu.%02luMHz (%s).\n", + /* Check the timer frequency. */ + if (arch_timer_rate == 0) + pr_warn("Architected timer frequency not available\n"); +} + +static void arch_timer_banner(unsigned type) +{ + pr_info("Architected %s%s%s timer(s) running at %lu.%02luMHz (%s%s%s).\n", + type & ARCH_CP15_TIMER ? "cp15" : "", + type == (ARCH_CP15_TIMER | ARCH_MEM_TIMER) ? " and " : "", + type & ARCH_MEM_TIMER ? "mmio" : "", (unsigned long)arch_timer_rate / 1000000, (unsigned long)(arch_timer_rate / 10000) % 100, - arch_timer_use_virtual ? "virt" : "phys"); - return 0; + type & ARCH_CP15_TIMER ? + arch_timer_use_virtual ? "virt" : "phys" : + "", + type == (ARCH_CP15_TIMER | ARCH_MEM_TIMER) ? "/" : "", + type & ARCH_MEM_TIMER ? + arch_timer_mem_use_virtual ? "virt" : "phys" : + ""); } u32 arch_timer_get_rate(void) @@ -186,11 +368,27 @@ u32 arch_timer_get_rate(void) return arch_timer_rate; } -u64 arch_timer_read_counter(void) +static u64 arch_counter_get_cntvct_mem(void) { - return arch_counter_get_cntvct(); + u32 vct_lo, vct_hi, tmp_hi; + + do { + vct_hi = readl_relaxed(arch_counter_base + CNTVCT_HI); + vct_lo = readl_relaxed(arch_counter_base + CNTVCT_LO); + tmp_hi = readl_relaxed(arch_counter_base + CNTVCT_HI); + } while (vct_hi != tmp_hi); + + return ((u64) vct_hi << 32) | vct_lo; } +/* + * Default to cp15 based access because arm64 uses this function for + * sched_clock() before DT is probed and the cp15 method is guaranteed + * to exist on arm64. arm doesn't use this before DT is probed so even + * if we don't have the cp15 accessors we won't have a problem. + */ +u64 (*arch_timer_read_counter)(void) = arch_counter_get_cntvct; + static cycle_t arch_counter_read(struct clocksource *cs) { return arch_counter_get_cntvct(); @@ -221,6 +419,23 @@ struct timecounter *arch_timer_get_timecounter(void) return &timecounter; } +static void __init arch_counter_register(unsigned type) +{ + u64 start_count; + + /* Register the CP15 based counter if we have one */ + if (type & ARCH_CP15_TIMER) + arch_timer_read_counter = arch_counter_get_cntvct; + else + arch_timer_read_counter = arch_counter_get_cntvct_mem; + + start_count = arch_timer_read_counter(); + clocksource_register_hz(&clocksource_counter, arch_timer_rate); + cyclecounter.mult = clocksource_counter.mult; + cyclecounter.shift = clocksource_counter.shift; + timecounter_init(&timecounter, &cyclecounter, start_count); +} + static void __cpuinit arch_timer_stop(struct clock_event_device *clk) { pr_debug("arch_timer_teardown disable IRQ%d cpu #%d\n", @@ -260,27 +475,44 @@ static struct notifier_block arch_timer_cpu_nb __cpuinitdata = { .notifier_call = arch_timer_cpu_notify, }; +#ifdef CONFIG_CPU_PM +static unsigned int saved_cntkctl; +static int arch_timer_cpu_pm_notify(struct notifier_block *self, + unsigned long action, void *hcpu) +{ + if (action == CPU_PM_ENTER) + saved_cntkctl = arch_timer_get_cntkctl(); + else if (action == CPU_PM_ENTER_FAILED || action == CPU_PM_EXIT) + arch_timer_set_cntkctl(saved_cntkctl); + return NOTIFY_OK; +} + +static struct notifier_block arch_timer_cpu_pm_notifier = { + .notifier_call = arch_timer_cpu_pm_notify, +}; + +static int __init arch_timer_cpu_pm_init(void) +{ + return cpu_pm_register_notifier(&arch_timer_cpu_pm_notifier); +} +#else +static int __init arch_timer_cpu_pm_init(void) +{ + return 0; +} +#endif + static int __init arch_timer_register(void) { int err; int ppi; - err = arch_timer_available(); - if (err) - goto out; - arch_timer_evt = alloc_percpu(struct clock_event_device); if (!arch_timer_evt) { err = -ENOMEM; goto out; } - clocksource_register_hz(&clocksource_counter, arch_timer_rate); - cyclecounter.mult = clocksource_counter.mult; - cyclecounter.shift = clocksource_counter.shift; - timecounter_init(&timecounter, &cyclecounter, - arch_counter_get_cntvct()); - if (arch_timer_use_virtual) { ppi = arch_timer_ppi[VIRT_PPI]; err = request_percpu_irq(ppi, arch_timer_handler_virt, @@ -309,11 +541,17 @@ static int __init arch_timer_register(void) if (err) goto out_free_irq; + err = arch_timer_cpu_pm_init(); + if (err) + goto out_unreg_notify; + /* Immediately configure the timer on the boot CPU */ arch_timer_setup(this_cpu_ptr(arch_timer_evt)); return 0; +out_unreg_notify: + unregister_cpu_notifier(&arch_timer_cpu_nb); out_free_irq: if (arch_timer_use_virtual) free_percpu_irq(arch_timer_ppi[VIRT_PPI], arch_timer_evt); @@ -331,24 +569,77 @@ out: return err; } +static int __init arch_timer_mem_register(void __iomem *base, unsigned int irq) +{ + int ret; + irq_handler_t func; + struct arch_timer *t; + + t = kzalloc(sizeof(*t), GFP_KERNEL); + if (!t) + return -ENOMEM; + + t->base = base; + t->evt.irq = irq; + __arch_timer_setup(ARCH_MEM_TIMER, &t->evt); + + if (arch_timer_mem_use_virtual) + func = arch_timer_handler_virt_mem; + else + func = arch_timer_handler_phys_mem; + + ret = request_irq(irq, func, IRQF_TIMER, "arch_mem_timer", &t->evt); + if (ret) { + pr_err("arch_timer: Failed to request mem timer irq\n"); + kfree(t); + } + + return ret; +} + +static const struct of_device_id arch_timer_of_match[] __initconst = { + { .compatible = "arm,armv7-timer", }, + { .compatible = "arm,armv8-timer", }, + {}, +}; + +static const struct of_device_id arch_timer_mem_of_match[] __initconst = { + { .compatible = "arm,armv7-timer-mem", }, + {}, +}; + +static void __init arch_timer_common_init(void) +{ + unsigned mask = ARCH_CP15_TIMER | ARCH_MEM_TIMER; + + /* Wait until both nodes are probed if we have two timers */ + if ((arch_timers_present & mask) != mask) { + if (of_find_matching_node(NULL, arch_timer_mem_of_match) && + !(arch_timers_present & ARCH_MEM_TIMER)) + return; + if (of_find_matching_node(NULL, arch_timer_of_match) && + !(arch_timers_present & ARCH_CP15_TIMER)) + return; + } + + arch_timer_banner(arch_timers_present); + arch_counter_register(arch_timers_present); + arch_timer_arch_init(); +} + static void __init arch_timer_init(struct device_node *np) { - u32 freq; int i; - if (arch_timer_get_rate()) { + if (arch_timers_present & ARCH_CP15_TIMER) { pr_warn("arch_timer: multiple nodes in dt, skipping\n"); return; } - /* Try to determine the frequency from the device tree or CNTFRQ */ - if (!of_property_read_u32(np, "clock-frequency", &freq)) - arch_timer_rate = freq; - + arch_timers_present |= ARCH_CP15_TIMER; for (i = PHYS_SECURE_PPI; i < MAX_TIMER_PPI; i++) arch_timer_ppi[i] = irq_of_parse_and_map(np, i); - - of_node_put(np); + arch_timer_detect_rate(NULL, np); /* * If HYP mode is available, we know that the physical timer @@ -369,7 +660,73 @@ static void __init arch_timer_init(struct device_node *np) } arch_timer_register(); - arch_timer_arch_init(); + arch_timer_common_init(); } CLOCKSOURCE_OF_DECLARE(armv7_arch_timer, "arm,armv7-timer", arch_timer_init); CLOCKSOURCE_OF_DECLARE(armv8_arch_timer, "arm,armv8-timer", arch_timer_init); + +static void __init arch_timer_mem_init(struct device_node *np) +{ + struct device_node *frame, *best_frame = NULL; + void __iomem *cntctlbase, *base; + unsigned int irq; + u32 cnttidr; + + arch_timers_present |= ARCH_MEM_TIMER; + cntctlbase = of_iomap(np, 0); + if (!cntctlbase) { + pr_err("arch_timer: Can't find CNTCTLBase\n"); + return; + } + + cnttidr = readl_relaxed(cntctlbase + CNTTIDR); + iounmap(cntctlbase); + + /* + * Try to find a virtual capable frame. Otherwise fall back to a + * physical capable frame. + */ + for_each_available_child_of_node(np, frame) { + int n; + + if (of_property_read_u32(frame, "frame-number", &n)) { + pr_err("arch_timer: Missing frame-number\n"); + of_node_put(best_frame); + of_node_put(frame); + return; + } + + if (cnttidr & CNTTIDR_VIRT(n)) { + of_node_put(best_frame); + best_frame = frame; + arch_timer_mem_use_virtual = true; + break; + } + of_node_put(best_frame); + best_frame = of_node_get(frame); + } + + base = arch_counter_base = of_iomap(best_frame, 0); + if (!base) { + pr_err("arch_timer: Can't map frame's registers\n"); + of_node_put(best_frame); + return; + } + + if (arch_timer_mem_use_virtual) + irq = irq_of_parse_and_map(best_frame, 1); + else + irq = irq_of_parse_and_map(best_frame, 0); + of_node_put(best_frame); + if (!irq) { + pr_err("arch_timer: Frame missing %s irq", + arch_timer_mem_use_virtual ? "virt" : "phys"); + return; + } + + arch_timer_detect_rate(base, np); + arch_timer_mem_register(base, irq); + arch_timer_common_init(); +} +CLOCKSOURCE_OF_DECLARE(armv7_arch_timer_mem, "arm,armv7-timer-mem", + arch_timer_mem_init); diff --git a/drivers/cpufreq/Kconfig b/drivers/cpufreq/Kconfig index 534fcb825153..a9c1324843eb 100644 --- a/drivers/cpufreq/Kconfig +++ b/drivers/cpufreq/Kconfig @@ -201,7 +201,7 @@ source "drivers/cpufreq/Kconfig.x86" endmenu menu "ARM CPU frequency scaling drivers" -depends on ARM +depends on ARM || ARM64 source "drivers/cpufreq/Kconfig.arm" endmenu diff --git a/drivers/cpufreq/Kconfig.arm b/drivers/cpufreq/Kconfig.arm index 6e57543fe0b9..22727792350b 100644 --- a/drivers/cpufreq/Kconfig.arm +++ b/drivers/cpufreq/Kconfig.arm @@ -4,7 +4,8 @@ config ARM_BIG_LITTLE_CPUFREQ tristate "Generic ARM big LITTLE CPUfreq driver" - depends on ARM_CPU_TOPOLOGY && PM_OPP && HAVE_CLK + depends on (ARM_CPU_TOPOLOGY && BIG_LITTLE) || (ARM64 && SMP) + depends on PM_OPP && HAVE_CLK help This enables the Generic CPUfreq driver for ARM big.LITTLE platforms. @@ -15,6 +16,14 @@ config ARM_DT_BL_CPUFREQ This enables probing via DT for Generic CPUfreq driver for ARM big.LITTLE platform. This gets frequency tables from DT. +config ARM_VEXPRESS_BL_CPUFREQ + tristate "ARM Vexpress big LITTLE CPUfreq driver" + select ARM_BIG_LITTLE_CPUFREQ + depends on VEXPRESS_SPC + help + This enables the CPUfreq driver for ARM Vexpress big.LITTLE platform. + If in doubt, say N. + config ARM_EXYNOS_CPUFREQ bool "SAMSUNG EXYNOS SoCs" depends on ARCH_EXYNOS diff --git a/drivers/cpufreq/Makefile b/drivers/cpufreq/Makefile index 3b95322fec5e..505c62bceb9d 100644 --- a/drivers/cpufreq/Makefile +++ b/drivers/cpufreq/Makefile @@ -48,6 +48,7 @@ obj-$(CONFIG_X86_AMD_FREQ_SENSITIVITY) += amd_freq_sensitivity.o obj-$(CONFIG_ARM_BIG_LITTLE_CPUFREQ) += arm_big_little.o # big LITTLE per platform glues. Keep DT_BL_CPUFREQ as the last entry in all big # LITTLE drivers, so that it is probed last. +obj-$(CONFIG_ARM_VEXPRESS_BL_CPUFREQ) += vexpress_big_little.o obj-$(CONFIG_ARM_DT_BL_CPUFREQ) += arm_big_little_dt.o obj-$(CONFIG_ARCH_DAVINCI) += davinci-cpufreq.o diff --git a/drivers/cpufreq/arm_big_little.c b/drivers/cpufreq/arm_big_little.c index 5d7f53fcd6f5..eaee6e222207 100644 --- a/drivers/cpufreq/arm_big_little.c +++ b/drivers/cpufreq/arm_big_little.c @@ -24,27 +24,148 @@ #include <linux/cpufreq.h> #include <linux/cpumask.h> #include <linux/export.h> +#include <linux/mutex.h> #include <linux/of_platform.h> #include <linux/opp.h> #include <linux/slab.h> #include <linux/topology.h> #include <linux/types.h> +#include <asm/bL_switcher.h> #include "arm_big_little.h" -/* Currently we support only two clusters */ -#define MAX_CLUSTERS 2 +#ifdef CONFIG_BL_SWITCHER +bool bL_switching_enabled; +#endif + +#define ACTUAL_FREQ(cluster, freq) ((cluster == A7_CLUSTER) ? freq << 1 : freq) +#define VIRT_FREQ(cluster, freq) ((cluster == A7_CLUSTER) ? freq >> 1 : freq) static struct cpufreq_arm_bL_ops *arm_bL_ops; static struct clk *clk[MAX_CLUSTERS]; -static struct cpufreq_frequency_table *freq_table[MAX_CLUSTERS]; -static atomic_t cluster_usage[MAX_CLUSTERS] = {ATOMIC_INIT(0), ATOMIC_INIT(0)}; +static struct cpufreq_frequency_table *freq_table[MAX_CLUSTERS + 1]; +static atomic_t cluster_usage[MAX_CLUSTERS + 1] = {ATOMIC_INIT(0), + ATOMIC_INIT(0)}; + +static unsigned int clk_big_min; /* (Big) clock frequencies */ +static unsigned int clk_little_max; /* Maximum clock frequency (Little) */ + +static DEFINE_PER_CPU(unsigned int, physical_cluster); +static DEFINE_PER_CPU(unsigned int, cpu_last_req_freq); + +static struct mutex cluster_lock[MAX_CLUSTERS]; + +static unsigned int find_cluster_maxfreq(int cluster) +{ + int j; + u32 max_freq = 0, cpu_freq; + + for_each_online_cpu(j) { + cpu_freq = per_cpu(cpu_last_req_freq, j); + + if ((cluster == per_cpu(physical_cluster, j)) && + (max_freq < cpu_freq)) + max_freq = cpu_freq; + } -static unsigned int bL_cpufreq_get(unsigned int cpu) + pr_debug("%s: cluster: %d, max freq: %d\n", __func__, cluster, + max_freq); + + return max_freq; +} + +static unsigned int clk_get_cpu_rate(unsigned int cpu) { - u32 cur_cluster = cpu_to_cluster(cpu); + u32 cur_cluster = per_cpu(physical_cluster, cpu); + u32 rate = clk_get_rate(clk[cur_cluster]) / 1000; + + /* For switcher we use virtual A15 clock rates */ + if (is_bL_switching_enabled()) + rate = VIRT_FREQ(cur_cluster, rate); + + pr_debug("%s: cpu: %d, cluster: %d, freq: %u\n", __func__, cpu, + cur_cluster, rate); - return clk_get_rate(clk[cur_cluster]) / 1000; + return rate; +} + +static unsigned int bL_cpufreq_get_rate(unsigned int cpu) +{ + if (is_bL_switching_enabled()) { + pr_debug("%s: freq: %d\n", __func__, per_cpu(cpu_last_req_freq, + cpu)); + + return per_cpu(cpu_last_req_freq, cpu); + } else { + return clk_get_cpu_rate(cpu); + } +} + +static unsigned int +bL_cpufreq_set_rate(u32 cpu, u32 old_cluster, u32 new_cluster, u32 rate) +{ + u32 new_rate, prev_rate; + int ret; + bool bLs = is_bL_switching_enabled(); + + mutex_lock(&cluster_lock[new_cluster]); + + if (bLs) { + prev_rate = per_cpu(cpu_last_req_freq, cpu); + per_cpu(cpu_last_req_freq, cpu) = rate; + per_cpu(physical_cluster, cpu) = new_cluster; + + new_rate = find_cluster_maxfreq(new_cluster); + new_rate = ACTUAL_FREQ(new_cluster, new_rate); + } else { + new_rate = rate; + } + + pr_debug("%s: cpu: %d, old cluster: %d, new cluster: %d, freq: %d\n", + __func__, cpu, old_cluster, new_cluster, new_rate); + + ret = clk_set_rate(clk[new_cluster], new_rate * 1000); + if (WARN_ON(ret)) { + pr_err("clk_set_rate failed: %d, new cluster: %d\n", ret, + new_cluster); + if (bLs) { + per_cpu(cpu_last_req_freq, cpu) = prev_rate; + per_cpu(physical_cluster, cpu) = old_cluster; + } + + mutex_unlock(&cluster_lock[new_cluster]); + + return ret; + } + + mutex_unlock(&cluster_lock[new_cluster]); + + /* Recalc freq for old cluster when switching clusters */ + if (old_cluster != new_cluster) { + pr_debug("%s: cpu: %d, old cluster: %d, new cluster: %d\n", + __func__, cpu, old_cluster, new_cluster); + + /* Switch cluster */ + bL_switch_request(cpu, new_cluster); + + mutex_lock(&cluster_lock[old_cluster]); + + /* Set freq of old cluster if there are cpus left on it */ + new_rate = find_cluster_maxfreq(old_cluster); + new_rate = ACTUAL_FREQ(old_cluster, new_rate); + + if (new_rate) { + pr_debug("%s: Updating rate of old cluster: %d, to freq: %d\n", + __func__, old_cluster, new_rate); + + if (clk_set_rate(clk[old_cluster], new_rate * 1000)) + pr_err("%s: clk_set_rate failed: %d, old cluster: %d\n", + __func__, ret, old_cluster); + } + mutex_unlock(&cluster_lock[old_cluster]); + } + + return 0; } /* Validate policy frequency range */ @@ -60,12 +181,14 @@ static int bL_cpufreq_set_target(struct cpufreq_policy *policy, unsigned int target_freq, unsigned int relation) { struct cpufreq_freqs freqs; - u32 cpu = policy->cpu, freq_tab_idx, cur_cluster; + u32 cpu = policy->cpu, freq_tab_idx, cur_cluster, new_cluster, + actual_cluster; int ret = 0; - cur_cluster = cpu_to_cluster(policy->cpu); + cur_cluster = cpu_to_cluster(cpu); + new_cluster = actual_cluster = per_cpu(physical_cluster, cpu); - freqs.old = bL_cpufreq_get(policy->cpu); + freqs.old = bL_cpufreq_get_rate(cpu); /* Determine valid target frequency using freq_table */ cpufreq_frequency_table_target(policy, freq_table[cur_cluster], @@ -79,13 +202,21 @@ static int bL_cpufreq_set_target(struct cpufreq_policy *policy, if (freqs.old == freqs.new) return 0; + if (is_bL_switching_enabled()) { + if ((actual_cluster == A15_CLUSTER) && + (freqs.new < clk_big_min)) { + new_cluster = A7_CLUSTER; + } else if ((actual_cluster == A7_CLUSTER) && + (freqs.new > clk_little_max)) { + new_cluster = A15_CLUSTER; + } + } + cpufreq_notify_transition(policy, &freqs, CPUFREQ_PRECHANGE); - ret = clk_set_rate(clk[cur_cluster], freqs.new * 1000); - if (ret) { - pr_err("clk_set_rate failed: %d\n", ret); + ret = bL_cpufreq_set_rate(cpu, actual_cluster, new_cluster, freqs.new); + if (ret) return ret; - } policy->cur = freqs.new; @@ -94,7 +225,73 @@ static int bL_cpufreq_set_target(struct cpufreq_policy *policy, return ret; } -static void put_cluster_clk_and_freq_table(struct device *cpu_dev) +static inline u32 get_table_count(struct cpufreq_frequency_table *table) +{ + int count; + + for (count = 0; table[count].frequency != CPUFREQ_TABLE_END; count++) + ; + + return count; +} + +/* get the minimum frequency in the cpufreq_frequency_table */ +static inline u32 get_table_min(struct cpufreq_frequency_table *table) +{ + int i; + uint32_t min_freq = ~0; + for (i = 0; (table[i].frequency != CPUFREQ_TABLE_END); i++) + if (table[i].frequency < min_freq) + min_freq = table[i].frequency; + return min_freq; +} + +/* get the maximum frequency in the cpufreq_frequency_table */ +static inline u32 get_table_max(struct cpufreq_frequency_table *table) +{ + int i; + uint32_t max_freq = 0; + for (i = 0; (table[i].frequency != CPUFREQ_TABLE_END); i++) + if (table[i].frequency > max_freq) + max_freq = table[i].frequency; + return max_freq; +} + +static int merge_cluster_tables(void) +{ + int i, j, k = 0, count = 1; + struct cpufreq_frequency_table *table; + + for (i = 0; i < MAX_CLUSTERS; i++) + count += get_table_count(freq_table[i]); + + table = kzalloc(sizeof(*table) * count, GFP_KERNEL); + if (!table) + return -ENOMEM; + + freq_table[MAX_CLUSTERS] = table; + + /* Add in reverse order to get freqs in increasing order */ + for (i = MAX_CLUSTERS - 1; i >= 0; i--) { + for (j = 0; freq_table[i][j].frequency != CPUFREQ_TABLE_END; + j++) { + table[k].frequency = VIRT_FREQ(i, + freq_table[i][j].frequency); + pr_debug("%s: index: %d, freq: %d\n", __func__, k, + table[k].frequency); + k++; + } + } + + table[k].index = k; + table[k].frequency = CPUFREQ_TABLE_END; + + pr_debug("%s: End, table: %p, count: %d\n", __func__, table, k); + + return 0; +} + +static void _put_cluster_clk_and_freq_table(struct device *cpu_dev) { u32 cluster = cpu_to_cluster(cpu_dev->id); @@ -105,10 +302,35 @@ static void put_cluster_clk_and_freq_table(struct device *cpu_dev) } } -static int get_cluster_clk_and_freq_table(struct device *cpu_dev) +static void put_cluster_clk_and_freq_table(struct device *cpu_dev) +{ + u32 cluster = cpu_to_cluster(cpu_dev->id); + int i; + + if (cluster < MAX_CLUSTERS) + return _put_cluster_clk_and_freq_table(cpu_dev); + + if (atomic_dec_return(&cluster_usage[MAX_CLUSTERS])) + return; + + for (i = 0; i < MAX_CLUSTERS; i++) { + struct device *cdev = get_cpu_device(i); + if (!cdev) { + pr_err("%s: failed to get cpu%d device\n", __func__, i); + return; + } + + _put_cluster_clk_and_freq_table(cdev); + } + + /* free virtual table */ + kfree(freq_table[MAX_CLUSTERS]); +} + +static int _get_cluster_clk_and_freq_table(struct device *cpu_dev) { u32 cluster = cpu_to_cluster(cpu_dev->id); - char name[14] = "cpu-cluster."; + char name[14] = "cpu-cluster.X"; int ret; if (atomic_inc_return(&cluster_usage[cluster]) != 1) @@ -149,6 +371,62 @@ atomic_dec: return ret; } +static int get_cluster_clk_and_freq_table(struct device *cpu_dev) +{ + u32 cluster = cpu_to_cluster(cpu_dev->id); + int i, ret; + + if (cluster < MAX_CLUSTERS) + return _get_cluster_clk_and_freq_table(cpu_dev); + + if (atomic_inc_return(&cluster_usage[MAX_CLUSTERS]) != 1) + return 0; + + /* + * Get data for all clusters and fill virtual cluster with a merge of + * both + */ + for (i = 0; i < MAX_CLUSTERS; i++) { + struct device *cdev = get_cpu_device(i); + if (!cdev) { + pr_err("%s: failed to get cpu%d device\n", __func__, i); + return -ENODEV; + } + + ret = _get_cluster_clk_and_freq_table(cdev); + if (ret) + goto put_clusters; + } + + ret = merge_cluster_tables(); + if (ret) + goto put_clusters; + + /* Assuming 2 cluster, set clk_big_min and clk_little_max */ + clk_big_min = get_table_min(freq_table[0]); + clk_little_max = VIRT_FREQ(1, get_table_max(freq_table[1])); + + pr_debug("%s: cluster: %d, clk_big_min: %d, clk_little_max: %d\n", + __func__, cluster, clk_big_min, clk_little_max); + + return 0; + +put_clusters: + while (i--) { + struct device *cdev = get_cpu_device(i); + if (!cdev) { + pr_err("%s: failed to get cpu%d device\n", __func__, i); + return -ENODEV; + } + + _put_cluster_clk_and_freq_table(cdev); + } + + atomic_dec(&cluster_usage[MAX_CLUSTERS]); + + return ret; +} + /* Per-CPU initialization */ static int bL_cpufreq_init(struct cpufreq_policy *policy) { @@ -177,37 +455,33 @@ static int bL_cpufreq_init(struct cpufreq_policy *policy) cpufreq_frequency_table_get_attr(freq_table[cur_cluster], policy->cpu); + if (cur_cluster < MAX_CLUSTERS) { + int cpu; + + cpumask_copy(policy->cpus, topology_core_cpumask(policy->cpu)); + + for_each_cpu(cpu, policy->cpus) + per_cpu(physical_cluster, cpu) = cur_cluster; + } else { + /* Assumption: during init, we are always running on A15 */ + per_cpu(physical_cluster, policy->cpu) = A15_CLUSTER; + } + if (arm_bL_ops->get_transition_latency) policy->cpuinfo.transition_latency = arm_bL_ops->get_transition_latency(cpu_dev); else policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; - policy->cur = bL_cpufreq_get(policy->cpu); + policy->cur = clk_get_cpu_rate(policy->cpu); - cpumask_copy(policy->cpus, topology_core_cpumask(policy->cpu)); + if (is_bL_switching_enabled()) + per_cpu(cpu_last_req_freq, policy->cpu) = policy->cur; dev_info(cpu_dev, "%s: CPU %d initialized\n", __func__, policy->cpu); return 0; } -static int bL_cpufreq_exit(struct cpufreq_policy *policy) -{ - struct device *cpu_dev; - - cpu_dev = get_cpu_device(policy->cpu); - if (!cpu_dev) { - pr_err("%s: failed to get cpu%d device\n", __func__, - policy->cpu); - return -ENODEV; - } - - put_cluster_clk_and_freq_table(cpu_dev); - dev_dbg(cpu_dev, "%s: Exited, cpu: %d\n", __func__, policy->cpu); - - return 0; -} - /* Export freq_table to sysfs */ static struct freq_attr *bL_cpufreq_attr[] = { &cpufreq_freq_attr_scaling_available_freqs, @@ -219,16 +493,47 @@ static struct cpufreq_driver bL_cpufreq_driver = { .flags = CPUFREQ_STICKY, .verify = bL_cpufreq_verify_policy, .target = bL_cpufreq_set_target, - .get = bL_cpufreq_get, + .get = bL_cpufreq_get_rate, .init = bL_cpufreq_init, - .exit = bL_cpufreq_exit, .have_governor_per_policy = true, .attr = bL_cpufreq_attr, }; +static int bL_cpufreq_switcher_notifier(struct notifier_block *nfb, + unsigned long action, void *_arg) +{ + pr_debug("%s: action: %ld\n", __func__, action); + + switch (action) { + case BL_NOTIFY_PRE_ENABLE: + case BL_NOTIFY_PRE_DISABLE: + cpufreq_unregister_driver(&bL_cpufreq_driver); + break; + + case BL_NOTIFY_POST_ENABLE: + set_switching_enabled(true); + cpufreq_register_driver(&bL_cpufreq_driver); + break; + + case BL_NOTIFY_POST_DISABLE: + set_switching_enabled(false); + cpufreq_register_driver(&bL_cpufreq_driver); + break; + + default: + return NOTIFY_DONE; + } + + return NOTIFY_OK; +} + +static struct notifier_block bL_switcher_notifier = { + .notifier_call = bL_cpufreq_switcher_notifier, +}; + int bL_cpufreq_register(struct cpufreq_arm_bL_ops *ops) { - int ret; + int ret, i; if (arm_bL_ops) { pr_debug("%s: Already registered: %s, exiting\n", __func__, @@ -243,16 +548,29 @@ int bL_cpufreq_register(struct cpufreq_arm_bL_ops *ops) arm_bL_ops = ops; + ret = bL_switcher_get_enabled(); + set_switching_enabled(ret); + + for (i = 0; i < MAX_CLUSTERS; i++) + mutex_init(&cluster_lock[i]); + ret = cpufreq_register_driver(&bL_cpufreq_driver); if (ret) { pr_info("%s: Failed registering platform driver: %s, err: %d\n", __func__, ops->name, ret); arm_bL_ops = NULL; } else { - pr_info("%s: Registered platform driver: %s\n", __func__, - ops->name); + ret = bL_switcher_register_notifier(&bL_switcher_notifier); + if (ret) { + cpufreq_unregister_driver(&bL_cpufreq_driver); + arm_bL_ops = NULL; + } else { + pr_info("%s: Registered platform driver: %s\n", + __func__, ops->name); + } } + bL_switcher_put_enabled(); return ret; } EXPORT_SYMBOL_GPL(bL_cpufreq_register); @@ -265,9 +583,31 @@ void bL_cpufreq_unregister(struct cpufreq_arm_bL_ops *ops) return; } + bL_switcher_get_enabled(); + bL_switcher_unregister_notifier(&bL_switcher_notifier); cpufreq_unregister_driver(&bL_cpufreq_driver); + bL_switcher_put_enabled(); pr_info("%s: Un-registered platform driver: %s\n", __func__, arm_bL_ops->name); + + /* For saving table get/put on every cpu in/out */ + if (is_bL_switching_enabled()) { + put_cluster_clk_and_freq_table(get_cpu_device(0)); + } else { + int i; + + for (i = 0; i < MAX_CLUSTERS; i++) { + struct device *cdev = get_cpu_device(i); + if (!cdev) { + pr_err("%s: failed to get cpu%d device\n", + __func__, i); + return; + } + + put_cluster_clk_and_freq_table(cdev); + } + } + arm_bL_ops = NULL; } EXPORT_SYMBOL_GPL(bL_cpufreq_unregister); diff --git a/drivers/cpufreq/arm_big_little.h b/drivers/cpufreq/arm_big_little.h index 79b2ce17884d..4f5a03d3aef6 100644 --- a/drivers/cpufreq/arm_big_little.h +++ b/drivers/cpufreq/arm_big_little.h @@ -23,6 +23,20 @@ #include <linux/device.h> #include <linux/types.h> +/* Currently we support only two clusters */ +#define A15_CLUSTER 0 +#define A7_CLUSTER 1 +#define MAX_CLUSTERS 2 + +#ifdef CONFIG_BL_SWITCHER +extern bool bL_switching_enabled; +#define is_bL_switching_enabled() bL_switching_enabled +#define set_switching_enabled(x) (bL_switching_enabled = (x)) +#else +#define is_bL_switching_enabled() false +#define set_switching_enabled(x) do { } while (0) +#endif + struct cpufreq_arm_bL_ops { char name[CPUFREQ_NAME_LEN]; int (*get_transition_latency)(struct device *cpu_dev); @@ -36,7 +50,8 @@ struct cpufreq_arm_bL_ops { static inline int cpu_to_cluster(int cpu) { - return topology_physical_package_id(cpu); + return is_bL_switching_enabled() ? MAX_CLUSTERS: + topology_physical_package_id(cpu); } int bL_cpufreq_register(struct cpufreq_arm_bL_ops *ops); diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index 66f6cf5064ec..ce94c323fa13 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -31,6 +31,8 @@ #include <linux/completion.h> #include <linux/mutex.h> #include <linux/syscore_ops.h> +#include <linux/suspend.h> +#include <linux/tick.h> #include <trace/events/power.h> @@ -48,6 +50,14 @@ static DEFINE_PER_CPU(char[CPUFREQ_NAME_LEN], cpufreq_cpu_governor); static DEFINE_RWLOCK(cpufreq_driver_lock); static DEFINE_MUTEX(cpufreq_governor_lock); +/* Flag to suspend/resume CPUFreq governors */ +static bool cpufreq_suspended; + +static inline bool has_target(void) +{ + return cpufreq_driver->target; +} + /* * cpu_policy_rwsem is a per CPU reader-writer semaphore designed to cure * all cpufreq/hotplug/workqueue/etc related lock issues. @@ -1275,82 +1285,89 @@ static struct subsys_interface cpufreq_interface = { /** - * cpufreq_bp_suspend - Prepare the boot CPU for system suspend. + * cpufreq_suspend() - Suspend CPUFreq governors * - * This function is only executed for the boot processor. The other CPUs - * have been put offline by means of CPU hotplug. + * Called during system wide Suspend/Hibernate cycles for suspending governors + * as some platforms can't change frequency after this point in suspend cycle. + * Because some of the devices (like: i2c, regulators, etc) they use for + * changing frequency are suspended quickly after this point. */ -static int cpufreq_bp_suspend(void) +void cpufreq_suspend(void) { - int ret = 0; + struct cpufreq_policy *policy; + int cpu; - int cpu = smp_processor_id(); - struct cpufreq_policy *cpu_policy; + if (!cpufreq_driver) + return; - pr_debug("suspending cpu %u\n", cpu); + if (!has_target()) + return; - /* If there's no policy for the boot CPU, we have nothing to do. */ - cpu_policy = cpufreq_cpu_get(cpu); - if (!cpu_policy) - return 0; + pr_debug("%s: Suspending Governors\n", __func__); - if (cpufreq_driver->suspend) { - ret = cpufreq_driver->suspend(cpu_policy); - if (ret) - printk(KERN_ERR "cpufreq: suspend failed in ->suspend " - "step on CPU %u\n", cpu_policy->cpu); + for_each_possible_cpu(cpu) { + if (!cpu_online(cpu)) + continue; + + policy = cpufreq_cpu_get(cpu); + + if (__cpufreq_governor(policy, CPUFREQ_GOV_STOP)) + pr_err("%s: Failed to stop governor for policy: %p\n", + __func__, policy); + else if (cpufreq_driver->suspend + && cpufreq_driver->suspend(policy)) + pr_err("%s: Failed to suspend driver: %p\n", __func__, + policy); } - cpufreq_cpu_put(cpu_policy); - return ret; + cpufreq_suspended = true; } /** - * cpufreq_bp_resume - Restore proper frequency handling of the boot CPU. + * cpufreq_resume() - Resume CPUFreq governors * - * 1.) resume CPUfreq hardware support (cpufreq_driver->resume()) - * 2.) schedule call cpufreq_update_policy() ASAP as interrupts are - * restored. It will verify that the current freq is in sync with - * what we believe it to be. This is a bit later than when it - * should be, but nonethteless it's better than calling - * cpufreq_driver->get() here which might re-enable interrupts... - * - * This function is only executed for the boot CPU. The other CPUs have not - * been turned on yet. + * Called during system wide Suspend/Hibernate cycle for resuming governors that + * are suspended with cpufreq_suspend(). */ -static void cpufreq_bp_resume(void) +void cpufreq_resume(void) { - int ret = 0; - - int cpu = smp_processor_id(); - struct cpufreq_policy *cpu_policy; + struct cpufreq_policy *policy; + int cpu; - pr_debug("resuming cpu %u\n", cpu); + if (!cpufreq_driver) + return; - /* If there's no policy for the boot CPU, we have nothing to do. */ - cpu_policy = cpufreq_cpu_get(cpu); - if (!cpu_policy) + if (!has_target()) return; - if (cpufreq_driver->resume) { - ret = cpufreq_driver->resume(cpu_policy); - if (ret) { - printk(KERN_ERR "cpufreq: resume failed in ->resume " - "step on CPU %u\n", cpu_policy->cpu); - goto fail; - } - } + pr_debug("%s: Resuming Governors\n", __func__); - schedule_work(&cpu_policy->update); + cpufreq_suspended = false; -fail: - cpufreq_cpu_put(cpu_policy); -} + for_each_possible_cpu(cpu) { + if (!cpu_online(cpu)) + continue; -static struct syscore_ops cpufreq_syscore_ops = { - .suspend = cpufreq_bp_suspend, - .resume = cpufreq_bp_resume, -}; + policy = cpufreq_cpu_get(cpu); + + if (__cpufreq_governor(policy, CPUFREQ_GOV_START) + || __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS)) + pr_err("%s: Failed to start governor for policy: %p\n", + __func__, policy); + else if (cpufreq_driver->resume + && cpufreq_driver->resume(policy)) + pr_err("%s: Failed to resume driver: %p\n", __func__, + policy); + + /* + * schedule call cpufreq_update_policy() for boot CPU, i.e. last + * policy in list. It will verify that the current freq is in + * sync with what we believe it to be. + */ + if (cpu == 0) + schedule_work(&policy->update); + } +} /** * cpufreq_get_current_driver - return current driver's name @@ -1544,6 +1561,10 @@ static int __cpufreq_governor(struct cpufreq_policy *policy, struct cpufreq_governor *gov = NULL; #endif + /* Don't start any governor operations if we are entering suspend */ + if (cpufreq_suspended) + return 0; + if (policy->governor->max_transition_latency && policy->cpuinfo.transition_latency > policy->governor->max_transition_latency) { @@ -2001,7 +2022,6 @@ static int __init cpufreq_core_init(void) cpufreq_global_kobject = kobject_create_and_add("cpufreq", &cpu_subsys.dev_root->kobj); BUG_ON(!cpufreq_global_kobject); - register_syscore_ops(&cpufreq_syscore_ops); return 0; } diff --git a/drivers/cpufreq/cpufreq_stats.c b/drivers/cpufreq/cpufreq_stats.c index 7fb600239059..9823037f194b 100644 --- a/drivers/cpufreq/cpufreq_stats.c +++ b/drivers/cpufreq/cpufreq_stats.c @@ -21,6 +21,9 @@ #include <linux/spinlock.h> #include <linux/notifier.h> #include <asm/cputime.h> +#ifdef CONFIG_BL_SWITCHER +#include <asm/bL_switcher.h> +#endif static spinlock_t cpufreq_stats_lock; @@ -378,7 +381,7 @@ static struct notifier_block notifier_trans_block = { .notifier_call = cpufreq_stat_notifier_trans }; -static int __init cpufreq_stats_init(void) +static int cpufreq_stats_setup(void) { int ret; unsigned int cpu; @@ -406,7 +409,8 @@ static int __init cpufreq_stats_init(void) return 0; } -static void __exit cpufreq_stats_exit(void) + +static void cpufreq_stats_cleanup(void) { unsigned int cpu; @@ -421,6 +425,54 @@ static void __exit cpufreq_stats_exit(void) } } +#ifdef CONFIG_BL_SWITCHER +static int cpufreq_stats_switcher_notifier(struct notifier_block *nfb, + unsigned long action, void *_arg) +{ + switch (action) { + case BL_NOTIFY_PRE_ENABLE: + case BL_NOTIFY_PRE_DISABLE: + cpufreq_stats_cleanup(); + break; + + case BL_NOTIFY_POST_ENABLE: + case BL_NOTIFY_POST_DISABLE: + cpufreq_stats_setup(); + break; + + default: + return NOTIFY_DONE; + } + + return NOTIFY_OK; +} + +static struct notifier_block switcher_notifier = { + .notifier_call = cpufreq_stats_switcher_notifier, +}; +#endif + +static int __init cpufreq_stats_init(void) +{ + int ret; + spin_lock_init(&cpufreq_stats_lock); + + ret = cpufreq_stats_setup(); +#ifdef CONFIG_BL_SWITCHER + if (!ret) + bL_switcher_register_notifier(&switcher_notifier); +#endif + return ret; +} + +static void __exit cpufreq_stats_exit(void) +{ +#ifdef CONFIG_BL_SWITCHER + bL_switcher_unregister_notifier(&switcher_notifier); +#endif + cpufreq_stats_cleanup(); +} + MODULE_AUTHOR("Zou Nan hai <nanhai.zou@intel.com>"); MODULE_DESCRIPTION("'cpufreq_stats' - A driver to export cpufreq stats " "through sysfs filesystem"); diff --git a/drivers/cpufreq/highbank-cpufreq.c b/drivers/cpufreq/highbank-cpufreq.c index 8c4771da1b89..9216c2020ec1 100644 --- a/drivers/cpufreq/highbank-cpufreq.c +++ b/drivers/cpufreq/highbank-cpufreq.c @@ -19,7 +19,7 @@ #include <linux/cpu.h> #include <linux/err.h> #include <linux/of.h> -#include <linux/mailbox.h> +#include <linux/pl320-ipc.h> #include <linux/platform_device.h> #define HB_CPUFREQ_CHANGE_NOTE 0x80000001 diff --git a/drivers/cpufreq/vexpress_big_little.c b/drivers/cpufreq/vexpress_big_little.c new file mode 100644 index 000000000000..1abb883c051b --- /dev/null +++ b/drivers/cpufreq/vexpress_big_little.c @@ -0,0 +1,86 @@ +/* + * Vexpress big.LITTLE CPUFreq Interface driver + * + * It provides necessary ops to arm_big_little cpufreq driver and gets + * Frequency information from Device Tree. Freq table in DT must be in KHz. + * + * Copyright (C) 2013 Linaro. + * Viresh Kumar <viresh.kumar@linaro.org> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * This program is distributed "as is" WITHOUT ANY WARRANTY of any + * kind, whether express or implied; without even the implied warranty + * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#include <linux/cpufreq.h> +#include <linux/export.h> +#include <linux/opp.h> +#include <linux/module.h> +#include <linux/types.h> +#include <linux/vexpress.h> +#include "arm_big_little.h" + +static int vexpress_init_opp_table(struct device *cpu_dev) +{ + int i = -1, count, cluster = cpu_to_cluster(cpu_dev->id); + u32 *table; + int ret; + + count = vexpress_spc_get_freq_table(cluster, &table); + if (!table || !count) { + pr_err("SPC controller returned invalid freq table"); + return -EINVAL; + } + + while (++i < count) { + /* FIXME: Voltage value */ + ret = opp_add(cpu_dev, table[i] * 1000, 900000); + if (ret) { + dev_warn(cpu_dev, "%s: Failed to add OPP %d, err: %d\n", + __func__, table[i] * 1000, ret); + return ret; + } + } + + return 0; +} + +static int vexpress_get_transition_latency(struct device *cpu_dev) +{ + /* 1 ms */ + return 1000000; +} + +static struct cpufreq_arm_bL_ops vexpress_bL_ops = { + .name = "vexpress-bL", + .get_transition_latency = vexpress_get_transition_latency, + .init_opp_table = vexpress_init_opp_table, +}; + +static int vexpress_bL_init(void) +{ + if (!vexpress_spc_check_loaded()) { + pr_info("%s: No SPC found\n", __func__); + return -ENOENT; + } + + return bL_cpufreq_register(&vexpress_bL_ops); +} +module_init(vexpress_bL_init); + +static void vexpress_bL_exit(void) +{ + return bL_cpufreq_unregister(&vexpress_bL_ops); +} +module_exit(vexpress_bL_exit); + +MODULE_AUTHOR("Viresh Kumar <viresh.kumar@linaro.org>"); +MODULE_DESCRIPTION("ARM Vexpress big LITTLE cpufreq driver"); +MODULE_LICENSE("GPL"); diff --git a/drivers/cpuidle/Kconfig b/drivers/cpuidle/Kconfig index c4cc27e5c8a5..842d7ba83101 100644 --- a/drivers/cpuidle/Kconfig +++ b/drivers/cpuidle/Kconfig @@ -31,6 +31,15 @@ config CPU_IDLE_GOV_MENU config ARCH_NEEDS_CPU_IDLE_COUPLED def_bool n +config OF_IDLE_STATES + bool "Idle states DT support" + depends on ARM || ARM64 + default n + help + Allows the CPU idle framework to initialize CPU idle drivers + state data by using DT provided nodes compliant with idle states + device tree bindings. + if CPU_IDLE config CPU_IDLE_CALXEDA @@ -39,4 +48,9 @@ config CPU_IDLE_CALXEDA help Select this to enable cpuidle on Calxeda processors. +menu "ARM64 CPU Idle Drivers" +depends on ARM64 +source "drivers/cpuidle/Kconfig.arm64" +endmenu + endif diff --git a/drivers/cpuidle/Kconfig.arm64 b/drivers/cpuidle/Kconfig.arm64 new file mode 100644 index 000000000000..b83612c67e6d --- /dev/null +++ b/drivers/cpuidle/Kconfig.arm64 @@ -0,0 +1,13 @@ +# +# ARM64 CPU Idle drivers +# + +config ARM64_CPUIDLE + bool "Generic ARM64 CPU idle Driver" + select OF_IDLE_STATES + help + Select this to enable generic cpuidle driver for ARM v8. + It provides a generic idle driver whose idle states are configured + at run-time through DT nodes. The CPUidle suspend backend is + initialized by the device tree parsing code on matching the entry + method to the respective CPU operations. diff --git a/drivers/cpuidle/Makefile b/drivers/cpuidle/Makefile index 0d8bd55e776f..2d97bcfecd00 100644 --- a/drivers/cpuidle/Makefile +++ b/drivers/cpuidle/Makefile @@ -4,6 +4,12 @@ obj-y += cpuidle.o driver.o governor.o sysfs.o governors/ obj-$(CONFIG_ARCH_NEEDS_CPU_IDLE_COUPLED) += coupled.o +obj-$(CONFIG_BIG_LITTLE) += arm_big_little.o +obj-$(CONFIG_OF_IDLE_STATES) += of_idle_states.o obj-$(CONFIG_CPU_IDLE_CALXEDA) += cpuidle-calxeda.o obj-$(CONFIG_ARCH_KIRKWOOD) += cpuidle-kirkwood.o + +############################################################################### +# ARM64 drivers +obj-$(CONFIG_ARM64_CPUIDLE) += cpuidle-arm64.o diff --git a/drivers/cpuidle/arm_big_little.c b/drivers/cpuidle/arm_big_little.c new file mode 100644 index 000000000000..e5378896a8cb --- /dev/null +++ b/drivers/cpuidle/arm_big_little.c @@ -0,0 +1,183 @@ +/* + * big.LITTLE CPU idle driver. + * + * Copyright (C) 2012 ARM Ltd. + * Author: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include <linux/arm-cci.h> +#include <linux/bitmap.h> +#include <linux/cpuidle.h> +#include <linux/cpu_pm.h> +#include <linux/clockchips.h> +#include <linux/debugfs.h> +#include <linux/hrtimer.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/tick.h> +#include <linux/vexpress.h> +#include <asm/mcpm.h> +#include <asm/cpuidle.h> +#include <asm/cputype.h> +#include <asm/idmap.h> +#include <asm/proc-fns.h> +#include <asm/suspend.h> +#include <linux/of.h> + +static int bl_cpuidle_simple_enter(struct cpuidle_device *dev, + struct cpuidle_driver *drv, int index) +{ + ktime_t time_start, time_end; + s64 diff; + + time_start = ktime_get(); + + cpu_do_idle(); + + time_end = ktime_get(); + + local_irq_enable(); + + diff = ktime_to_us(ktime_sub(time_end, time_start)); + if (diff > INT_MAX) + diff = INT_MAX; + + dev->last_residency = (int) diff; + + return index; +} + +static int bl_enter_powerdown(struct cpuidle_device *dev, + struct cpuidle_driver *drv, int idx); + +static struct cpuidle_state bl_cpuidle_set[] __initdata = { + [0] = { + .enter = bl_cpuidle_simple_enter, + .exit_latency = 1, + .target_residency = 1, + .power_usage = UINT_MAX, + .flags = CPUIDLE_FLAG_TIME_VALID, + .name = "WFI", + .desc = "ARM WFI", + }, + [1] = { + .enter = bl_enter_powerdown, + .exit_latency = 300, + .target_residency = 1000, + .flags = CPUIDLE_FLAG_TIME_VALID, + .name = "C1", + .desc = "ARM power down", + }, +}; + +struct cpuidle_driver bl_idle_driver = { + .name = "bl_idle", + .owner = THIS_MODULE, + .safe_state_index = 0 +}; + +static DEFINE_PER_CPU(struct cpuidle_device, bl_idle_dev); + +static int notrace bl_powerdown_finisher(unsigned long arg) +{ + unsigned int mpidr = read_cpuid_mpidr(); + unsigned int cluster = (mpidr >> 8) & 0xf; + unsigned int cpu = mpidr & 0xf; + + mcpm_set_entry_vector(cpu, cluster, cpu_resume); + mcpm_cpu_suspend(0); /* 0 should be replaced with better value here */ + return 1; +} + +/* + * bl_enter_powerdown - Programs CPU to enter the specified state + * @dev: cpuidle device + * @drv: The target state to be programmed + * @idx: state index + * + * Called from the CPUidle framework to program the device to the + * specified target state selected by the governor. + */ +static int bl_enter_powerdown(struct cpuidle_device *dev, + struct cpuidle_driver *drv, int idx) +{ + struct timespec ts_preidle, ts_postidle, ts_idle; + int ret; + + /* Used to keep track of the total time in idle */ + getnstimeofday(&ts_preidle); + + BUG_ON(!irqs_disabled()); + + cpu_pm_enter(); + + clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ENTER, &dev->cpu); + + ret = cpu_suspend((unsigned long) dev, bl_powerdown_finisher); + if (ret) + BUG(); + + mcpm_cpu_powered_up(); + + clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_EXIT, &dev->cpu); + + cpu_pm_exit(); + + getnstimeofday(&ts_postidle); + local_irq_enable(); + ts_idle = timespec_sub(ts_postidle, ts_preidle); + + dev->last_residency = ts_idle.tv_nsec / NSEC_PER_USEC + + ts_idle.tv_sec * USEC_PER_SEC; + return idx; +} + +/* + * bl_idle_init + * + * Registers the bl specific cpuidle driver with the cpuidle + * framework with the valid set of states. + */ +int __init bl_idle_init(void) +{ + struct cpuidle_device *dev; + int i, cpu_id; + struct cpuidle_driver *drv = &bl_idle_driver; + + if (!of_find_compatible_node(NULL, NULL, "arm,generic")) { + pr_info("%s: No compatible node found\n", __func__); + return -ENODEV; + } + + drv->state_count = (sizeof(bl_cpuidle_set) / + sizeof(struct cpuidle_state)); + + for (i = 0; i < drv->state_count; i++) { + memcpy(&drv->states[i], &bl_cpuidle_set[i], + sizeof(struct cpuidle_state)); + } + + cpuidle_register_driver(drv); + + for_each_cpu(cpu_id, cpu_online_mask) { + pr_err("CPUidle for CPU%d registered\n", cpu_id); + dev = &per_cpu(bl_idle_dev, cpu_id); + dev->cpu = cpu_id; + + dev->state_count = drv->state_count; + + if (cpuidle_register_device(dev)) { + printk(KERN_ERR "%s: Cpuidle register device failed\n", + __func__); + return -EIO; + } + } + + return 0; +} + +device_initcall(bl_idle_init); diff --git a/drivers/cpuidle/cpuidle-arm64.c b/drivers/cpuidle/cpuidle-arm64.c new file mode 100644 index 000000000000..2cfde6ce3086 --- /dev/null +++ b/drivers/cpuidle/cpuidle-arm64.c @@ -0,0 +1,159 @@ +/* + * ARM64 generic CPU idle driver. + * + * Copyright (C) 2014 ARM Ltd. + * Author: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include <linux/cpuidle.h> +#include <linux/cpumask.h> +#include <linux/cpu_pm.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of.h> + +#include <asm/psci.h> +#include <asm/suspend.h> + +#include "of_idle_states.h" + +typedef int (*suspend_init_fn)(struct cpuidle_driver *, + struct device_node *[]); + +struct cpu_suspend_ops { + const char *id; + suspend_init_fn init_fn; +}; + +static const struct cpu_suspend_ops suspend_operations[] __initconst = { + {"arm,psci", psci_dt_register_idle_states}, + {} +}; + +static __init const struct cpu_suspend_ops *get_suspend_ops(const char *str) +{ + int i; + + if (!str) + return NULL; + + for (i = 0; suspend_operations[i].id; i++) + if (!strcmp(suspend_operations[i].id, str)) + return &suspend_operations[i]; + + return NULL; +} + +/* + * arm_enter_idle_state - Programs CPU to enter the specified state + * + * @dev: cpuidle device + * @drv: cpuidle driver + * @idx: state index + * + * Called from the CPUidle framework to program the device to the + * specified target state selected by the governor. + */ +static int arm_enter_idle_state(struct cpuidle_device *dev, + struct cpuidle_driver *drv, int idx) +{ + int ret; + + if (!idx) { + cpu_do_idle(); + return idx; + } + + cpu_pm_enter(); + /* + * Pass idle state index to cpu_suspend which in turn will call + * the CPU ops suspend protocol with idle index as a parameter. + * + * Some states would not require context to be saved and flushed + * to DRAM, so calling cpu_suspend would not be stricly necessary. + * When power domains specifications for ARM CPUs are finalized then + * this code can be optimized to prevent saving registers if not + * needed. + */ + ret = cpu_suspend(idx); + + cpu_pm_exit(); + + return ret ? -1 : idx; +} + +struct cpuidle_driver arm64_idle_driver = { + .name = "arm64_idle", + .owner = THIS_MODULE, +}; + +static struct device_node *state_nodes[CPUIDLE_STATE_MAX] __initdata; + +/* + * arm64_idle_init + * + * Registers the arm64 specific cpuidle driver with the cpuidle + * framework. It relies on core code to parse the idle states + * and initialize them using driver data structures accordingly. + */ +static int __init arm64_idle_init(void) +{ + int i, ret; + const char *entry_method; + struct device_node *idle_states_node; + const struct cpu_suspend_ops *suspend_init; + struct cpuidle_driver *drv = &arm64_idle_driver; + + idle_states_node = of_find_node_by_path("/cpus/idle-states"); + if (!idle_states_node) + return -ENOENT; + + if (of_property_read_string(idle_states_node, "entry-method", + &entry_method)) { + pr_warn(" * %s missing entry-method property\n", + idle_states_node->full_name); + of_node_put(idle_states_node); + return -EOPNOTSUPP; + } + + suspend_init = get_suspend_ops(entry_method); + if (!suspend_init) { + pr_warn("Missing suspend initializer\n"); + of_node_put(idle_states_node); + return -EOPNOTSUPP; + } + + /* + * State at index 0 is standby wfi and considered standard + * on all ARM platforms. If in some platforms simple wfi + * can't be used as "state 0", DT bindings must be implemented + * to work around this issue and allow installing a special + * handler for idle state index 0. + */ + drv->states[0].exit_latency = 1; + drv->states[0].target_residency = 1; + drv->states[0].flags = CPUIDLE_FLAG_TIME_VALID; + strncpy(drv->states[0].name, "ARM WFI", CPUIDLE_NAME_LEN); + strncpy(drv->states[0].desc, "ARM WFI", CPUIDLE_DESC_LEN); + + drv->cpumask = (struct cpumask *) cpu_possible_mask; + /* + * Start at index 1, request idle state nodes to be filled + */ + ret = of_init_idle_driver(drv, state_nodes, 1, true); + if (ret) + return ret; + + if (suspend_init->init_fn(drv, state_nodes)) + return -EOPNOTSUPP; + + for (i = 0; i < drv->state_count; i++) + drv->states[i].enter = arm_enter_idle_state; + + return cpuidle_register(drv, NULL); +} +device_initcall(arm64_idle_init); diff --git a/drivers/cpuidle/cpuidle-calxeda.c b/drivers/cpuidle/cpuidle-calxeda.c index 223379169cb0..0e6e408c0a63 100644 --- a/drivers/cpuidle/cpuidle-calxeda.c +++ b/drivers/cpuidle/cpuidle-calxeda.c @@ -37,20 +37,6 @@ extern void highbank_set_cpu_jump(int cpu, void *jump_addr); extern void *scu_base_addr; -static inline unsigned int get_auxcr(void) -{ - unsigned int val; - asm("mrc p15, 0, %0, c1, c0, 1 @ get AUXCR" : "=r" (val) : : "cc"); - return val; -} - -static inline void set_auxcr(unsigned int val) -{ - asm volatile("mcr p15, 0, %0, c1, c0, 1 @ set AUXCR" - : : "r" (val) : "cc"); - isb(); -} - static noinline void calxeda_idle_restore(void) { set_cr(get_cr() | CR_C); diff --git a/drivers/cpuidle/of_idle_states.c b/drivers/cpuidle/of_idle_states.c new file mode 100644 index 000000000000..eceb1b4c4657 --- /dev/null +++ b/drivers/cpuidle/of_idle_states.c @@ -0,0 +1,274 @@ +/* + * OF idle states parsing code. + * + * Copyright (C) 2014 ARM Ltd. + * Author: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include <linux/cpuidle.h> +#include <linux/cpumask.h> +#include <linux/errno.h> +#include <linux/kernel.h> +#include <linux/list.h> +#include <linux/list_sort.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/slab.h> + +#include "of_idle_states.h" + +struct state_elem { + struct list_head list; + struct device_node *node; + int val; +}; + +static struct list_head head __initdata = LIST_HEAD_INIT(head); + +static bool __init state_cpu_valid(struct device_node *state_node, + struct device_node *cpu_node) +{ + int i = 0; + struct device_node *cpu_state; + + while ((cpu_state = of_parse_phandle(cpu_node, + "cpu-idle-states", i++))) { + if (cpu_state && state_node == cpu_state) { + of_node_put(cpu_state); + return true; + } + of_node_put(cpu_state); + } + return false; +} + +static bool __init state_cpus_valid(const cpumask_t *cpus, + struct device_node *state_node) +{ + int cpu; + struct device_node *cpu_node; + + /* + * Check if state is valid on driver cpumask cpus + */ + for_each_cpu(cpu, cpus) { + cpu_node = of_get_cpu_node(cpu, NULL); + + if (!cpu_node) { + pr_err("Missing device node for CPU %d\n", cpu); + return false; + } + + if (!state_cpu_valid(state_node, cpu_node)) + return false; + } + + return true; +} + +static int __init state_cmp(void *priv, struct list_head *a, + struct list_head *b) +{ + struct state_elem *ela, *elb; + + ela = container_of(a, struct state_elem, list); + elb = container_of(b, struct state_elem, list); + + return ela->val - elb->val; +} + +static int __init add_state_node(cpumask_t *cpumask, + struct device_node *state_node) +{ + struct state_elem *el; + u32 val; + + pr_debug(" * %s...\n", state_node->full_name); + + if (!state_cpus_valid(cpumask, state_node)) + return -EINVAL; + /* + * Parse just the value required to sort the states. + */ + if (of_property_read_u32(state_node, "min-residency-us", + &val)) { + pr_debug(" * %s missing min-residency-us property\n", + state_node->full_name); + return -EINVAL; + } + + el = kmalloc(sizeof(*el), GFP_KERNEL); + if (!el) { + pr_err("%s failed to allocate memory\n", __func__); + return -ENOMEM; + } + + el->node = state_node; + el->val = val; + list_add_tail(&el->list, &head); + + return 0; +} + +static void __init init_state_node(struct cpuidle_driver *drv, + struct device_node *state_node, + int *cnt) +{ + struct cpuidle_state *idle_state; + + pr_debug(" * %s...\n", state_node->full_name); + + idle_state = &drv->states[*cnt]; + + if (of_property_read_u32(state_node, "exit-latency-us", + &idle_state->exit_latency)) { + pr_debug(" * %s missing exit-latency-us property\n", + state_node->full_name); + return; + } + + if (of_property_read_u32(state_node, "min-residency-us", + &idle_state->target_residency)) { + pr_debug(" * %s missing min-residency-us property\n", + state_node->full_name); + return; + } + /* + * It is unknown to the idle driver if and when the tick_device + * loses context when the CPU enters the idle states. To solve + * this issue the tick device must be linked to a power domain + * so that the idle driver can check on which states the device + * loses its context. Current code takes the conservative choice + * of defining the idle state as one where the tick device always + * loses its context. On platforms where tick device never loses + * its context (ie it is not a C3STOP device) this turns into + * a nop. On platforms where the tick device does lose context in some + * states, this code can be optimized, when power domain specifications + * for ARM CPUs are finalized. + */ + idle_state->flags = CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TIMER_STOP; + + strncpy(idle_state->name, state_node->name, CPUIDLE_NAME_LEN); + strncpy(idle_state->desc, state_node->name, CPUIDLE_NAME_LEN); + + (*cnt)++; +} + +static int __init init_idle_states(struct cpuidle_driver *drv, + struct device_node *state_nodes[], + unsigned int start_idx, bool init_nodes) +{ + struct state_elem *el; + struct list_head *curr, *tmp; + unsigned int cnt = start_idx; + + list_for_each_entry(el, &head, list) { + /* + * Check if the init function has to fill the + * state_nodes array on behalf of the CPUidle driver. + */ + if (init_nodes) + state_nodes[cnt] = el->node; + /* + * cnt is updated on return if a state was added. + */ + init_state_node(drv, el->node, &cnt); + + if (cnt == CPUIDLE_STATE_MAX) { + pr_warn("State index reached static CPU idle state limit\n"); + break; + } + } + + drv->state_count = cnt; + + list_for_each_safe(curr, tmp, &head) { + list_del(curr); + kfree(container_of(curr, struct state_elem, list)); + } + + /* + * If no idle states are detected, return an error and let the idle + * driver initialization fail accordingly. + */ + return (cnt > start_idx) ? 0 : -ENODATA; +} + +static void __init add_idle_states(struct cpuidle_driver *drv, + struct device_node *idle_states) +{ + struct device_node *state_node; + + for_each_child_of_node(idle_states, state_node) { + if ((!of_device_is_compatible(state_node, "arm,idle-state"))) { + pr_warn(" * %s: children of /cpus/idle-states must be \"arm,idle-state\" compatible\n", + state_node->full_name); + continue; + } + /* + * If memory allocation fails, better bail out. + * Initialized nodes are freed at initialization + * completion in of_init_idle_driver(). + */ + if ((add_state_node(drv->cpumask, state_node) == -ENOMEM)) + break; + } + /* + * Sort the states list before initializing the CPUidle driver + * states array. + */ + list_sort(NULL, &head, state_cmp); +} + +/* + * of_init_idle_driver - Parse the DT idle states and initialize the + * idle driver states array + * + * @drv: Pointer to CPU idle driver to be initialized + * @state_nodes: Array of struct device_nodes to be initialized if + * init_nodes == true. Must be sized CPUIDLE_STATE_MAX + * @start_idx: First idle state index to be initialized + * @init_nodes: Boolean to request device nodes initialization + * + * Returns: + * 0 on success + * <0 on failure + * + * On success the states array in the cpuidle driver contains + * initialized entries in the states array, starting from index start_idx. + * If init_nodes == true, on success the state_nodes array is initialized + * with idle state DT node pointers, starting from index start_idx, + * in a 1:1 relation with the idle driver states array. + */ +int __init of_init_idle_driver(struct cpuidle_driver *drv, + struct device_node *state_nodes[], + unsigned int start_idx, bool init_nodes) +{ + struct device_node *idle_states_node; + int ret; + + if (start_idx >= CPUIDLE_STATE_MAX) { + pr_warn("State index exceeds static CPU idle driver states array size\n"); + return -EINVAL; + } + + if (WARN(init_nodes && !state_nodes, + "Requested nodes stashing in an invalid nodes container\n")) + return -EINVAL; + + idle_states_node = of_find_node_by_path("/cpus/idle-states"); + if (!idle_states_node) + return -ENOENT; + + add_idle_states(drv, idle_states_node); + + ret = init_idle_states(drv, state_nodes, start_idx, init_nodes); + + of_node_put(idle_states_node); + + return ret; +} diff --git a/drivers/cpuidle/of_idle_states.h b/drivers/cpuidle/of_idle_states.h new file mode 100644 index 000000000000..049f94ff4428 --- /dev/null +++ b/drivers/cpuidle/of_idle_states.h @@ -0,0 +1,8 @@ +#ifndef __OF_IDLE_STATES +#define __OF_IDLE_STATES + +int __init of_init_idle_driver(struct cpuidle_driver *drv, + struct device_node *state_nodes[], + unsigned int start_idx, + bool init_nodes); +#endif diff --git a/drivers/extcon/extcon-adc-jack.c b/drivers/extcon/extcon-adc-jack.c index d0233cd18ffa..5985807e52c9 100644 --- a/drivers/extcon/extcon-adc-jack.c +++ b/drivers/extcon/extcon-adc-jack.c @@ -87,7 +87,8 @@ static irqreturn_t adc_jack_irq_thread(int irq, void *_data) { struct adc_jack_data *data = _data; - schedule_delayed_work(&data->handler, data->handling_delay); + queue_delayed_work(system_power_efficient_wq, + &data->handler, data->handling_delay); return IRQ_HANDLED; } diff --git a/drivers/extcon/extcon-gpio.c b/drivers/extcon/extcon-gpio.c index 02bec32adde4..f874c30ddbff 100644 --- a/drivers/extcon/extcon-gpio.c +++ b/drivers/extcon/extcon-gpio.c @@ -56,7 +56,7 @@ static irqreturn_t gpio_irq_handler(int irq, void *dev_id) { struct gpio_extcon_data *extcon_data = dev_id; - schedule_delayed_work(&extcon_data->work, + queue_delayed_work(system_power_efficient_wq, &extcon_data->work, extcon_data->debounce_jiffies); return IRQ_HANDLED; } diff --git a/drivers/firmware/efi/Kconfig b/drivers/firmware/efi/Kconfig index b0fc7c79dfbb..0d32596ad092 100644 --- a/drivers/firmware/efi/Kconfig +++ b/drivers/firmware/efi/Kconfig @@ -36,4 +36,11 @@ config EFI_VARS_PSTORE_DEFAULT_DISABLE backend for pstore by default. This setting can be overridden using the efivars module's pstore_disable parameter. +config EFI_PARAMS_FROM_FDT + bool + help + Select this config option from the architecture Kconfig if + the EFI runtime support gets system table address, memory + map address, and other parameters from the device tree. + endmenu diff --git a/drivers/firmware/efi/arm-stub.c b/drivers/firmware/efi/arm-stub.c new file mode 100644 index 000000000000..41114ce03b01 --- /dev/null +++ b/drivers/firmware/efi/arm-stub.c @@ -0,0 +1,278 @@ +/* + * EFI stub implementation that is shared by arm and arm64 architectures. + * This should be #included by the EFI stub implementation files. + * + * Copyright (C) 2013,2014 Linaro Limited + * Roy Franz <roy.franz@linaro.org + * Copyright (C) 2013 Red Hat, Inc. + * Mark Salter <msalter@redhat.com> + * + * This file is part of the Linux kernel, and is made available under the + * terms of the GNU General Public License version 2. + * + */ + +static int __init efi_secureboot_enabled(efi_system_table_t *sys_table_arg) +{ + static efi_guid_t const var_guid __initconst = EFI_GLOBAL_VARIABLE_GUID; + static efi_char16_t const var_name[] __initconst = { + 'S', 'e', 'c', 'u', 'r', 'e', 'B', 'o', 'o', 't', 0 }; + + efi_get_variable_t *f_getvar = sys_table_arg->runtime->get_variable; + unsigned long size = sizeof(u8); + efi_status_t status; + u8 val; + + status = f_getvar((efi_char16_t *)var_name, (efi_guid_t *)&var_guid, + NULL, &size, &val); + + switch (status) { + case EFI_SUCCESS: + return val; + case EFI_NOT_FOUND: + return 0; + default: + return 1; + } +} + +static efi_status_t efi_open_volume(efi_system_table_t *sys_table_arg, + void *__image, void **__fh) +{ + efi_file_io_interface_t *io; + efi_loaded_image_t *image = __image; + efi_file_handle_t *fh; + efi_guid_t fs_proto = EFI_FILE_SYSTEM_GUID; + efi_status_t status; + void *handle = (void *)(unsigned long)image->device_handle; + + status = sys_table_arg->boottime->handle_protocol(handle, + &fs_proto, (void **)&io); + if (status != EFI_SUCCESS) { + efi_printk(sys_table_arg, "Failed to handle fs_proto\n"); + return status; + } + + status = io->open_volume(io, &fh); + if (status != EFI_SUCCESS) + efi_printk(sys_table_arg, "Failed to open volume\n"); + + *__fh = fh; + return status; +} +static efi_status_t efi_file_close(void *handle) +{ + efi_file_handle_t *fh = handle; + + return fh->close(handle); +} + +static efi_status_t +efi_file_read(void *handle, unsigned long *size, void *addr) +{ + efi_file_handle_t *fh = handle; + + return fh->read(handle, size, addr); +} + + +static efi_status_t +efi_file_size(efi_system_table_t *sys_table_arg, void *__fh, + efi_char16_t *filename_16, void **handle, u64 *file_sz) +{ + efi_file_handle_t *h, *fh = __fh; + efi_file_info_t *info; + efi_status_t status; + efi_guid_t info_guid = EFI_FILE_INFO_ID; + unsigned long info_sz; + + status = fh->open(fh, &h, filename_16, EFI_FILE_MODE_READ, (u64)0); + if (status != EFI_SUCCESS) { + efi_printk(sys_table_arg, "Failed to open file: "); + efi_char16_printk(sys_table_arg, filename_16); + efi_printk(sys_table_arg, "\n"); + return status; + } + + *handle = h; + + info_sz = 0; + status = h->get_info(h, &info_guid, &info_sz, NULL); + if (status != EFI_BUFFER_TOO_SMALL) { + efi_printk(sys_table_arg, "Failed to get file info size\n"); + return status; + } + +grow: + status = sys_table_arg->boottime->allocate_pool(EFI_LOADER_DATA, + info_sz, (void **)&info); + if (status != EFI_SUCCESS) { + efi_printk(sys_table_arg, "Failed to alloc mem for file info\n"); + return status; + } + + status = h->get_info(h, &info_guid, &info_sz, + info); + if (status == EFI_BUFFER_TOO_SMALL) { + sys_table_arg->boottime->free_pool(info); + goto grow; + } + + *file_sz = info->file_size; + sys_table_arg->boottime->free_pool(info); + + if (status != EFI_SUCCESS) + efi_printk(sys_table_arg, "Failed to get initrd info\n"); + + return status; +} + + + +static void efi_char16_printk(efi_system_table_t *sys_table_arg, + efi_char16_t *str) +{ + struct efi_simple_text_output_protocol *out; + + out = (struct efi_simple_text_output_protocol *)sys_table_arg->con_out; + out->output_string(out, str); +} + + +/* + * This function handles the architcture specific differences between arm and + * arm64 regarding where the kernel image must be loaded and any memory that + * must be reserved. On failure it is required to free all + * all allocations it has made. + */ +static efi_status_t handle_kernel_image(efi_system_table_t *sys_table, + unsigned long *image_addr, + unsigned long *image_size, + unsigned long *reserve_addr, + unsigned long *reserve_size, + unsigned long dram_base, + efi_loaded_image_t *image); +/* + * EFI entry point for the arm/arm64 EFI stubs. This is the entrypoint + * that is described in the PE/COFF header. Most of the code is the same + * for both archictectures, with the arch-specific code provided in the + * handle_kernel_image() function. + */ +unsigned long __init efi_entry(void *handle, efi_system_table_t *sys_table, + unsigned long *image_addr) +{ + efi_loaded_image_t *image; + efi_status_t status; + unsigned long image_size = 0; + unsigned long dram_base; + /* addr/point and size pairs for memory management*/ + unsigned long initrd_addr; + u64 initrd_size = 0; + unsigned long fdt_addr = 0; /* Original DTB */ + u64 fdt_size = 0; /* We don't get size from configuration table */ + char *cmdline_ptr = NULL; + int cmdline_size = 0; + unsigned long new_fdt_addr; + efi_guid_t loaded_image_proto = LOADED_IMAGE_PROTOCOL_GUID; + unsigned long reserve_addr = 0; + unsigned long reserve_size = 0; + + /* Check if we were booted by the EFI firmware */ + if (sys_table->hdr.signature != EFI_SYSTEM_TABLE_SIGNATURE) + goto fail; + + pr_efi(sys_table, "Booting Linux Kernel...\n"); + + /* + * Get a handle to the loaded image protocol. This is used to get + * information about the running image, such as size and the command + * line. + */ + status = sys_table->boottime->handle_protocol(handle, + &loaded_image_proto, (void *)&image); + if (status != EFI_SUCCESS) { + pr_efi_err(sys_table, "Failed to get loaded image protocol\n"); + goto fail; + } + + dram_base = get_dram_base(sys_table); + if (dram_base == EFI_ERROR) { + pr_efi_err(sys_table, "Failed to find DRAM base\n"); + goto fail; + } + status = handle_kernel_image(sys_table, image_addr, &image_size, + &reserve_addr, + &reserve_size, + dram_base, image); + if (status != EFI_SUCCESS) { + pr_efi_err(sys_table, "Failed to relocate kernel\n"); + goto fail; + } + + /* + * Get the command line from EFI, using the LOADED_IMAGE + * protocol. We are going to copy the command line into the + * device tree, so this can be allocated anywhere. + */ + cmdline_ptr = efi_convert_cmdline(sys_table, image, &cmdline_size); + if (!cmdline_ptr) { + pr_efi_err(sys_table, "getting command line via LOADED_IMAGE_PROTOCOL\n"); + goto fail_free_image; + } + + /* + * Unauthenticated device tree data is a security hazard, so + * ignore 'dtb=' unless UEFI Secure Boot is disabled. + */ + if (efi_secureboot_enabled(sys_table)) { + pr_efi(sys_table, "UEFI Secure Boot is enabled.\n"); + } else { + status = handle_cmdline_files(sys_table, image, cmdline_ptr, + "dtb=", + ~0UL, (unsigned long *)&fdt_addr, + (unsigned long *)&fdt_size); + + if (status != EFI_SUCCESS) { + pr_efi_err(sys_table, "Failed to load device tree!\n"); + goto fail_free_cmdline; + } + } + if (!fdt_addr) + /* Look for a device tree configuration table entry. */ + fdt_addr = (uintptr_t)get_fdt(sys_table); + + status = handle_cmdline_files(sys_table, image, cmdline_ptr, + "initrd=", dram_base + SZ_512M, + (unsigned long *)&initrd_addr, + (unsigned long *)&initrd_size); + if (status != EFI_SUCCESS) + pr_efi_err(sys_table, "Failed initrd from command line!\n"); + + new_fdt_addr = fdt_addr; + status = allocate_new_fdt_and_exit_boot(sys_table, handle, + &new_fdt_addr, dram_base + MAX_FDT_OFFSET, + initrd_addr, initrd_size, cmdline_ptr, + fdt_addr, fdt_size); + + /* + * If all went well, we need to return the FDT address to the + * calling function so it can be passed to kernel as part of + * the kernel boot protocol. + */ + if (status == EFI_SUCCESS) + return new_fdt_addr; + + pr_efi_err(sys_table, "Failed to update FDT and exit boot services\n"); + + efi_free(sys_table, initrd_size, initrd_addr); + efi_free(sys_table, fdt_size, fdt_addr); + +fail_free_cmdline: + efi_free(sys_table, cmdline_size, (unsigned long)cmdline_ptr); + +fail_free_image: + efi_free(sys_table, image_size, *image_addr); + efi_free(sys_table, reserve_size, reserve_addr); +fail: + return EFI_ERROR; +} diff --git a/drivers/firmware/efi/efi-pstore.c b/drivers/firmware/efi/efi-pstore.c index 202d2c85ba2e..bf8dd3d5bee7 100644 --- a/drivers/firmware/efi/efi-pstore.c +++ b/drivers/firmware/efi/efi-pstore.c @@ -18,14 +18,12 @@ module_param_named(pstore_disable, efivars_pstore_disable, bool, 0644); static int efi_pstore_open(struct pstore_info *psi) { - efivar_entry_iter_begin(); psi->data = NULL; return 0; } static int efi_pstore_close(struct pstore_info *psi) { - efivar_entry_iter_end(); psi->data = NULL; return 0; } @@ -78,19 +76,124 @@ static int efi_pstore_read_func(struct efivar_entry *entry, void *data) __efivar_entry_get(entry, &entry->var.Attributes, &entry->var.DataSize, entry->var.Data); size = entry->var.DataSize; + memcpy(*cb_data->buf, entry->var.Data, + (size_t)min_t(unsigned long, EFIVARS_DATA_SIZE_MAX, size)); - *cb_data->buf = kmalloc(size, GFP_KERNEL); - if (*cb_data->buf == NULL) - return -ENOMEM; - memcpy(*cb_data->buf, entry->var.Data, size); return size; } +/** + * efi_pstore_scan_sysfs_enter + * @entry: scanning entry + * @next: next entry + * @head: list head + */ +static void efi_pstore_scan_sysfs_enter(struct efivar_entry *pos, + struct efivar_entry *next, + struct list_head *head) +{ + pos->scanning = true; + if (&next->list != head) + next->scanning = true; +} + +/** + * __efi_pstore_scan_sysfs_exit + * @entry: deleting entry + * @turn_off_scanning: Check if a scanning flag should be turned off + */ +static inline void __efi_pstore_scan_sysfs_exit(struct efivar_entry *entry, + bool turn_off_scanning) +{ + if (entry->deleting) { + list_del(&entry->list); + efivar_entry_iter_end(); + efivar_unregister(entry); + efivar_entry_iter_begin(); + } else if (turn_off_scanning) + entry->scanning = false; +} + +/** + * efi_pstore_scan_sysfs_exit + * @pos: scanning entry + * @next: next entry + * @head: list head + * @stop: a flag checking if scanning will stop + */ +static void efi_pstore_scan_sysfs_exit(struct efivar_entry *pos, + struct efivar_entry *next, + struct list_head *head, bool stop) +{ + __efi_pstore_scan_sysfs_exit(pos, true); + if (stop) + __efi_pstore_scan_sysfs_exit(next, &next->list != head); +} + +/** + * efi_pstore_sysfs_entry_iter + * + * @data: function-specific data to pass to callback + * @pos: entry to begin iterating from + * + * You MUST call efivar_enter_iter_begin() before this function, and + * efivar_entry_iter_end() afterwards. + * + * It is possible to begin iteration from an arbitrary entry within + * the list by passing @pos. @pos is updated on return to point to + * the next entry of the last one passed to efi_pstore_read_func(). + * To begin iterating from the beginning of the list @pos must be %NULL. + */ +static int efi_pstore_sysfs_entry_iter(void *data, struct efivar_entry **pos) +{ + struct efivar_entry *entry, *n; + struct list_head *head = &efivar_sysfs_list; + int size = 0; + + if (!*pos) { + list_for_each_entry_safe(entry, n, head, list) { + efi_pstore_scan_sysfs_enter(entry, n, head); + + size = efi_pstore_read_func(entry, data); + efi_pstore_scan_sysfs_exit(entry, n, head, size < 0); + if (size) + break; + } + *pos = n; + return size; + } + + list_for_each_entry_safe_from((*pos), n, head, list) { + efi_pstore_scan_sysfs_enter((*pos), n, head); + + size = efi_pstore_read_func((*pos), data); + efi_pstore_scan_sysfs_exit((*pos), n, head, size < 0); + if (size) + break; + } + *pos = n; + return size; +} + +/** + * efi_pstore_read + * + * This function returns a size of NVRAM entry logged via efi_pstore_write(). + * The meaning and behavior of efi_pstore/pstore are as below. + * + * size > 0: Got data of an entry logged via efi_pstore_write() successfully, + * and pstore filesystem will continue reading subsequent entries. + * size == 0: Entry was not logged via efi_pstore_write(), + * and efi_pstore driver will continue reading subsequent entries. + * size < 0: Failed to get data of entry logging via efi_pstore_write(), + * and pstore will stop reading entry. + */ static ssize_t efi_pstore_read(u64 *id, enum pstore_type_id *type, int *count, struct timespec *timespec, char **buf, struct pstore_info *psi) { struct pstore_read_data data; + ssize_t size; data.id = id; data.type = type; @@ -98,8 +201,17 @@ static ssize_t efi_pstore_read(u64 *id, enum pstore_type_id *type, data.timespec = timespec; data.buf = buf; - return __efivar_entry_iter(efi_pstore_read_func, &efivar_sysfs_list, &data, - (struct efivar_entry **)&psi->data); + *data.buf = kzalloc(EFIVARS_DATA_SIZE_MAX, GFP_KERNEL); + if (!*data.buf) + return -ENOMEM; + + efivar_entry_iter_begin(); + size = efi_pstore_sysfs_entry_iter(&data, + (struct efivar_entry **)&psi->data); + efivar_entry_iter_end(); + if (size <= 0) + kfree(*data.buf); + return size; } static int efi_pstore_write(enum pstore_type_id type, @@ -170,9 +282,17 @@ static int efi_pstore_erase_func(struct efivar_entry *entry, void *data) return 0; } + if (entry->scanning) { + /* + * Skip deletion because this entry will be deleted + * after scanning is completed. + */ + entry->deleting = true; + } else + list_del(&entry->list); + /* found */ __efivar_entry_delete(entry); - list_del(&entry->list); return 1; } @@ -200,10 +320,12 @@ static int efi_pstore_erase(enum pstore_type_id type, u64 id, int count, efivar_entry_iter_begin(); found = __efivar_entry_iter(efi_pstore_erase_func, &efivar_sysfs_list, &edata, &entry); - efivar_entry_iter_end(); - if (found) + if (found && !entry->scanning) { + efivar_entry_iter_end(); efivar_unregister(entry); + } else + efivar_entry_iter_end(); return 0; } @@ -236,7 +358,11 @@ static __init int efivars_pstore_init(void) efi_pstore_info.bufsize = 1024; spin_lock_init(&efi_pstore_info.buf_lock); - pstore_register(&efi_pstore_info); + if (pstore_register(&efi_pstore_info)) { + kfree(efi_pstore_info.buf); + efi_pstore_info.buf = NULL; + efi_pstore_info.bufsize = 0; + } return 0; } diff --git a/drivers/firmware/efi/efi-stub-helper.c b/drivers/firmware/efi/efi-stub-helper.c new file mode 100644 index 000000000000..eb6d4be9e722 --- /dev/null +++ b/drivers/firmware/efi/efi-stub-helper.c @@ -0,0 +1,634 @@ +/* + * Helper functions used by the EFI stub on multiple + * architectures. This should be #included by the EFI stub + * implementation files. + * + * Copyright 2011 Intel Corporation; author Matt Fleming + * + * This file is part of the Linux kernel, and is made available + * under the terms of the GNU General Public License version 2. + * + */ +#define EFI_READ_CHUNK_SIZE (1024 * 1024) + +/* error code which can't be mistaken for valid address */ +#define EFI_ERROR (~0UL) + + +struct file_info { + efi_file_handle_t *handle; + u64 size; +}; + +static void efi_printk(efi_system_table_t *sys_table_arg, char *str) +{ + char *s8; + + for (s8 = str; *s8; s8++) { + efi_char16_t ch[2] = { 0 }; + + ch[0] = *s8; + if (*s8 == '\n') { + efi_char16_t nl[2] = { '\r', 0 }; + efi_char16_printk(sys_table_arg, nl); + } + + efi_char16_printk(sys_table_arg, ch); + } +} + +#define pr_efi(sys_table, msg) efi_printk(sys_table, "EFI stub: "msg) +#define pr_efi_err(sys_table, msg) efi_printk(sys_table, "EFI stub: ERROR: "msg) + + +static efi_status_t efi_get_memory_map(efi_system_table_t *sys_table_arg, + efi_memory_desc_t **map, + unsigned long *map_size, + unsigned long *desc_size, + u32 *desc_ver, + unsigned long *key_ptr) +{ + efi_memory_desc_t *m = NULL; + efi_status_t status; + unsigned long key; + u32 desc_version; + + *map_size = sizeof(*m) * 32; +again: + /* + * Add an additional efi_memory_desc_t because we're doing an + * allocation which may be in a new descriptor region. + */ + *map_size += sizeof(*m); + status = efi_call_early(allocate_pool, EFI_LOADER_DATA, + *map_size, (void **)&m); + if (status != EFI_SUCCESS) + goto fail; + + *desc_size = 0; + key = 0; + status = efi_call_early(get_memory_map, map_size, m, + &key, desc_size, &desc_version); + if (status == EFI_BUFFER_TOO_SMALL) { + efi_call_early(free_pool, m); + goto again; + } + + if (status != EFI_SUCCESS) + efi_call_early(free_pool, m); + + if (key_ptr && status == EFI_SUCCESS) + *key_ptr = key; + if (desc_ver && status == EFI_SUCCESS) + *desc_ver = desc_version; + +fail: + *map = m; + return status; +} + + +static unsigned long __init get_dram_base(efi_system_table_t *sys_table_arg) +{ + efi_status_t status; + unsigned long map_size; + unsigned long membase = EFI_ERROR; + struct efi_memory_map map; + efi_memory_desc_t *md; + + status = efi_get_memory_map(sys_table_arg, (efi_memory_desc_t **)&map.map, + &map_size, &map.desc_size, NULL, NULL); + if (status != EFI_SUCCESS) + return membase; + + map.map_end = map.map + map_size; + + for_each_efi_memory_desc(&map, md) + if (md->attribute & EFI_MEMORY_WB) + if (membase > md->phys_addr) + membase = md->phys_addr; + + efi_call_early(free_pool, map.map); + + return membase; +} + +/* + * Allocate at the highest possible address that is not above 'max'. + */ +static efi_status_t efi_high_alloc(efi_system_table_t *sys_table_arg, + unsigned long size, unsigned long align, + unsigned long *addr, unsigned long max) +{ + unsigned long map_size, desc_size; + efi_memory_desc_t *map; + efi_status_t status; + unsigned long nr_pages; + u64 max_addr = 0; + int i; + + status = efi_get_memory_map(sys_table_arg, &map, &map_size, &desc_size, + NULL, NULL); + if (status != EFI_SUCCESS) + goto fail; + + /* + * Enforce minimum alignment that EFI requires when requesting + * a specific address. We are doing page-based allocations, + * so we must be aligned to a page. + */ + if (align < EFI_PAGE_SIZE) + align = EFI_PAGE_SIZE; + + nr_pages = round_up(size, EFI_PAGE_SIZE) / EFI_PAGE_SIZE; +again: + for (i = 0; i < map_size / desc_size; i++) { + efi_memory_desc_t *desc; + unsigned long m = (unsigned long)map; + u64 start, end; + + desc = (efi_memory_desc_t *)(m + (i * desc_size)); + if (desc->type != EFI_CONVENTIONAL_MEMORY) + continue; + + if (desc->num_pages < nr_pages) + continue; + + start = desc->phys_addr; + end = start + desc->num_pages * (1UL << EFI_PAGE_SHIFT); + + if ((start + size) > end || (start + size) > max) + continue; + + if (end - size > max) + end = max; + + if (round_down(end - size, align) < start) + continue; + + start = round_down(end - size, align); + + /* + * Don't allocate at 0x0. It will confuse code that + * checks pointers against NULL. + */ + if (start == 0x0) + continue; + + if (start > max_addr) + max_addr = start; + } + + if (!max_addr) + status = EFI_NOT_FOUND; + else { + status = efi_call_early(allocate_pages, + EFI_ALLOCATE_ADDRESS, EFI_LOADER_DATA, + nr_pages, &max_addr); + if (status != EFI_SUCCESS) { + max = max_addr; + max_addr = 0; + goto again; + } + + *addr = max_addr; + } + + efi_call_early(free_pool, map); +fail: + return status; +} + +/* + * Allocate at the lowest possible address. + */ +static efi_status_t efi_low_alloc(efi_system_table_t *sys_table_arg, + unsigned long size, unsigned long align, + unsigned long *addr) +{ + unsigned long map_size, desc_size; + efi_memory_desc_t *map; + efi_status_t status; + unsigned long nr_pages; + int i; + + status = efi_get_memory_map(sys_table_arg, &map, &map_size, &desc_size, + NULL, NULL); + if (status != EFI_SUCCESS) + goto fail; + + /* + * Enforce minimum alignment that EFI requires when requesting + * a specific address. We are doing page-based allocations, + * so we must be aligned to a page. + */ + if (align < EFI_PAGE_SIZE) + align = EFI_PAGE_SIZE; + + nr_pages = round_up(size, EFI_PAGE_SIZE) / EFI_PAGE_SIZE; + for (i = 0; i < map_size / desc_size; i++) { + efi_memory_desc_t *desc; + unsigned long m = (unsigned long)map; + u64 start, end; + + desc = (efi_memory_desc_t *)(m + (i * desc_size)); + + if (desc->type != EFI_CONVENTIONAL_MEMORY) + continue; + + if (desc->num_pages < nr_pages) + continue; + + start = desc->phys_addr; + end = start + desc->num_pages * (1UL << EFI_PAGE_SHIFT); + + /* + * Don't allocate at 0x0. It will confuse code that + * checks pointers against NULL. Skip the first 8 + * bytes so we start at a nice even number. + */ + if (start == 0x0) + start += 8; + + start = round_up(start, align); + if ((start + size) > end) + continue; + + status = efi_call_early(allocate_pages, + EFI_ALLOCATE_ADDRESS, EFI_LOADER_DATA, + nr_pages, &start); + if (status == EFI_SUCCESS) { + *addr = start; + break; + } + } + + if (i == map_size / desc_size) + status = EFI_NOT_FOUND; + + efi_call_early(free_pool, map); +fail: + return status; +} + +static void efi_free(efi_system_table_t *sys_table_arg, unsigned long size, + unsigned long addr) +{ + unsigned long nr_pages; + + if (!size) + return; + + nr_pages = round_up(size, EFI_PAGE_SIZE) / EFI_PAGE_SIZE; + efi_call_early(free_pages, addr, nr_pages); +} + + +/* + * Check the cmdline for a LILO-style file= arguments. + * + * We only support loading a file from the same filesystem as + * the kernel image. + */ +static efi_status_t handle_cmdline_files(efi_system_table_t *sys_table_arg, + efi_loaded_image_t *image, + char *cmd_line, char *option_string, + unsigned long max_addr, + unsigned long *load_addr, + unsigned long *load_size) +{ + struct file_info *files; + unsigned long file_addr; + u64 file_size_total; + efi_file_handle_t *fh = NULL; + efi_status_t status; + int nr_files; + char *str; + int i, j, k; + + file_addr = 0; + file_size_total = 0; + + str = cmd_line; + + j = 0; /* See close_handles */ + + if (!load_addr || !load_size) + return EFI_INVALID_PARAMETER; + + *load_addr = 0; + *load_size = 0; + + if (!str || !*str) + return EFI_SUCCESS; + + for (nr_files = 0; *str; nr_files++) { + str = strstr(str, option_string); + if (!str) + break; + + str += strlen(option_string); + + /* Skip any leading slashes */ + while (*str == '/' || *str == '\\') + str++; + + while (*str && *str != ' ' && *str != '\n') + str++; + } + + if (!nr_files) + return EFI_SUCCESS; + + status = efi_call_early(allocate_pool, EFI_LOADER_DATA, + nr_files * sizeof(*files), (void **)&files); + if (status != EFI_SUCCESS) { + pr_efi_err(sys_table_arg, "Failed to alloc mem for file handle list\n"); + goto fail; + } + + str = cmd_line; + for (i = 0; i < nr_files; i++) { + struct file_info *file; + efi_char16_t filename_16[256]; + efi_char16_t *p; + + str = strstr(str, option_string); + if (!str) + break; + + str += strlen(option_string); + + file = &files[i]; + p = filename_16; + + /* Skip any leading slashes */ + while (*str == '/' || *str == '\\') + str++; + + while (*str && *str != ' ' && *str != '\n') { + if ((u8 *)p >= (u8 *)filename_16 + sizeof(filename_16)) + break; + + if (*str == '/') { + *p++ = '\\'; + str++; + } else { + *p++ = *str++; + } + } + + *p = '\0'; + + /* Only open the volume once. */ + if (!i) { + status = efi_open_volume(sys_table_arg, image, + (void **)&fh); + if (status != EFI_SUCCESS) + goto free_files; + } + + status = efi_file_size(sys_table_arg, fh, filename_16, + (void **)&file->handle, &file->size); + if (status != EFI_SUCCESS) + goto close_handles; + + file_size_total += file->size; + } + + if (file_size_total) { + unsigned long addr; + + /* + * Multiple files need to be at consecutive addresses in memory, + * so allocate enough memory for all the files. This is used + * for loading multiple files. + */ + status = efi_high_alloc(sys_table_arg, file_size_total, 0x1000, + &file_addr, max_addr); + if (status != EFI_SUCCESS) { + pr_efi_err(sys_table_arg, "Failed to alloc highmem for files\n"); + goto close_handles; + } + + /* We've run out of free low memory. */ + if (file_addr > max_addr) { + pr_efi_err(sys_table_arg, "We've run out of free low memory\n"); + status = EFI_INVALID_PARAMETER; + goto free_file_total; + } + + addr = file_addr; + for (j = 0; j < nr_files; j++) { + unsigned long size; + + size = files[j].size; + while (size) { + unsigned long chunksize; + if (size > EFI_READ_CHUNK_SIZE) + chunksize = EFI_READ_CHUNK_SIZE; + else + chunksize = size; + + status = efi_file_read(files[j].handle, + &chunksize, + (void *)addr); + if (status != EFI_SUCCESS) { + pr_efi_err(sys_table_arg, "Failed to read file\n"); + goto free_file_total; + } + addr += chunksize; + size -= chunksize; + } + + efi_file_close(files[j].handle); + } + + } + + efi_call_early(free_pool, files); + + *load_addr = file_addr; + *load_size = file_size_total; + + return status; + +free_file_total: + efi_free(sys_table_arg, file_size_total, file_addr); + +close_handles: + for (k = j; k < i; k++) + efi_file_close(files[k].handle); +free_files: + efi_call_early(free_pool, files); +fail: + *load_addr = 0; + *load_size = 0; + + return status; +} +/* + * Relocate a kernel image, either compressed or uncompressed. + * In the ARM64 case, all kernel images are currently + * uncompressed, and as such when we relocate it we need to + * allocate additional space for the BSS segment. Any low + * memory that this function should avoid needs to be + * unavailable in the EFI memory map, as if the preferred + * address is not available the lowest available address will + * be used. + */ +static efi_status_t efi_relocate_kernel(efi_system_table_t *sys_table_arg, + unsigned long *image_addr, + unsigned long image_size, + unsigned long alloc_size, + unsigned long preferred_addr, + unsigned long alignment) +{ + unsigned long cur_image_addr; + unsigned long new_addr = 0; + efi_status_t status; + unsigned long nr_pages; + efi_physical_addr_t efi_addr = preferred_addr; + + if (!image_addr || !image_size || !alloc_size) + return EFI_INVALID_PARAMETER; + if (alloc_size < image_size) + return EFI_INVALID_PARAMETER; + + cur_image_addr = *image_addr; + + /* + * The EFI firmware loader could have placed the kernel image + * anywhere in memory, but the kernel has restrictions on the + * max physical address it can run at. Some architectures + * also have a prefered address, so first try to relocate + * to the preferred address. If that fails, allocate as low + * as possible while respecting the required alignment. + */ + nr_pages = round_up(alloc_size, EFI_PAGE_SIZE) / EFI_PAGE_SIZE; + status = efi_call_early(allocate_pages, + EFI_ALLOCATE_ADDRESS, EFI_LOADER_DATA, + nr_pages, &efi_addr); + new_addr = efi_addr; + /* + * If preferred address allocation failed allocate as low as + * possible. + */ + if (status != EFI_SUCCESS) { + status = efi_low_alloc(sys_table_arg, alloc_size, alignment, + &new_addr); + } + if (status != EFI_SUCCESS) { + pr_efi_err(sys_table_arg, "Failed to allocate usable memory for kernel.\n"); + return status; + } + + /* + * We know source/dest won't overlap since both memory ranges + * have been allocated by UEFI, so we can safely use memcpy. + */ + memcpy((void *)new_addr, (void *)cur_image_addr, image_size); + + /* Return the new address of the relocated image. */ + *image_addr = new_addr; + + return status; +} + +/* + * Get the number of UTF-8 bytes corresponding to an UTF-16 character. + * This overestimates for surrogates, but that is okay. + */ +static int efi_utf8_bytes(u16 c) +{ + return 1 + (c >= 0x80) + (c >= 0x800); +} + +/* + * Convert an UTF-16 string, not necessarily null terminated, to UTF-8. + */ +static u8 *efi_utf16_to_utf8(u8 *dst, const u16 *src, int n) +{ + unsigned int c; + + while (n--) { + c = *src++; + if (n && c >= 0xd800 && c <= 0xdbff && + *src >= 0xdc00 && *src <= 0xdfff) { + c = 0x10000 + ((c & 0x3ff) << 10) + (*src & 0x3ff); + src++; + n--; + } + if (c >= 0xd800 && c <= 0xdfff) + c = 0xfffd; /* Unmatched surrogate */ + if (c < 0x80) { + *dst++ = c; + continue; + } + if (c < 0x800) { + *dst++ = 0xc0 + (c >> 6); + goto t1; + } + if (c < 0x10000) { + *dst++ = 0xe0 + (c >> 12); + goto t2; + } + *dst++ = 0xf0 + (c >> 18); + *dst++ = 0x80 + ((c >> 12) & 0x3f); + t2: + *dst++ = 0x80 + ((c >> 6) & 0x3f); + t1: + *dst++ = 0x80 + (c & 0x3f); + } + + return dst; +} + +/* + * Convert the unicode UEFI command line to ASCII to pass to kernel. + * Size of memory allocated return in *cmd_line_len. + * Returns NULL on error. + */ +static char *efi_convert_cmdline(efi_system_table_t *sys_table_arg, + efi_loaded_image_t *image, + int *cmd_line_len) +{ + const u16 *s2; + u8 *s1 = NULL; + unsigned long cmdline_addr = 0; + int load_options_chars = image->load_options_size / 2; /* UTF-16 */ + const u16 *options = image->load_options; + int options_bytes = 0; /* UTF-8 bytes */ + int options_chars = 0; /* UTF-16 chars */ + efi_status_t status; + u16 zero = 0; + + if (options) { + s2 = options; + while (*s2 && *s2 != '\n' + && options_chars < load_options_chars) { + options_bytes += efi_utf8_bytes(*s2++); + options_chars++; + } + } + + if (!options_chars) { + /* No command line options, so return empty string*/ + options = &zero; + } + + options_bytes++; /* NUL termination */ + + status = efi_low_alloc(sys_table_arg, options_bytes, 0, &cmdline_addr); + if (status != EFI_SUCCESS) + return NULL; + + s1 = (u8 *)cmdline_addr; + s2 = (const u16 *)options; + + s1 = efi_utf16_to_utf8(s1, s2, options_chars); + *s1 = '\0'; + + *cmd_line_len = options_bytes; + return (char *)cmdline_addr; +} diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c index 5145fa344ad5..87f2890974a4 100644 --- a/drivers/firmware/efi/efi.c +++ b/drivers/firmware/efi/efi.c @@ -13,11 +13,29 @@ * This file is released under the GPLv2. */ +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + #include <linux/kobject.h> #include <linux/module.h> #include <linux/init.h> #include <linux/device.h> #include <linux/efi.h> +#include <linux/of.h> +#include <linux/of_fdt.h> +#include <linux/io.h> + +struct efi __read_mostly efi = { + .mps = EFI_INVALID_TABLE_ADDR, + .acpi = EFI_INVALID_TABLE_ADDR, + .acpi20 = EFI_INVALID_TABLE_ADDR, + .smbios = EFI_INVALID_TABLE_ADDR, + .sal_systab = EFI_INVALID_TABLE_ADDR, + .boot_info = EFI_INVALID_TABLE_ADDR, + .hcdp = EFI_INVALID_TABLE_ADDR, + .uga = EFI_INVALID_TABLE_ADDR, + .uv_systab = EFI_INVALID_TABLE_ADDR, +}; +EXPORT_SYMBOL(efi); static struct kobject *efi_kobj; static struct kobject *efivars_kobj; @@ -132,3 +150,212 @@ err_put: } subsys_initcall(efisubsys_init); + + +/* + * We can't ioremap data in EFI boot services RAM, because we've already mapped + * it as RAM. So, look it up in the existing EFI memory map instead. Only + * callable after efi_enter_virtual_mode and before efi_free_boot_services. + */ +void __iomem *efi_lookup_mapped_addr(u64 phys_addr) +{ + struct efi_memory_map *map; + void *p; + map = efi.memmap; + if (!map) + return NULL; + if (WARN_ON(!map->map)) + return NULL; + for (p = map->map; p < map->map_end; p += map->desc_size) { + efi_memory_desc_t *md = p; + u64 size = md->num_pages << EFI_PAGE_SHIFT; + u64 end = md->phys_addr + size; + if (!(md->attribute & EFI_MEMORY_RUNTIME) && + md->type != EFI_BOOT_SERVICES_CODE && + md->type != EFI_BOOT_SERVICES_DATA) + continue; + if (!md->virt_addr) + continue; + if (phys_addr >= md->phys_addr && phys_addr < end) { + phys_addr += md->virt_addr - md->phys_addr; + return (__force void __iomem *)(unsigned long)phys_addr; + } + } + return NULL; +} + +static __initdata efi_config_table_type_t common_tables[] = { + {ACPI_20_TABLE_GUID, "ACPI 2.0", &efi.acpi20}, + {ACPI_TABLE_GUID, "ACPI", &efi.acpi}, + {HCDP_TABLE_GUID, "HCDP", &efi.hcdp}, + {MPS_TABLE_GUID, "MPS", &efi.mps}, + {SAL_SYSTEM_TABLE_GUID, "SALsystab", &efi.sal_systab}, + {SMBIOS_TABLE_GUID, "SMBIOS", &efi.smbios}, + {UGA_IO_PROTOCOL_GUID, "UGA", &efi.uga}, + {NULL_GUID, NULL, NULL}, +}; + +static __init int match_config_table(efi_guid_t *guid, + unsigned long table, + efi_config_table_type_t *table_types) +{ + u8 str[EFI_VARIABLE_GUID_LEN + 1]; + int i; + + if (table_types) { + efi_guid_unparse(guid, str); + + for (i = 0; efi_guidcmp(table_types[i].guid, NULL_GUID); i++) { + efi_guid_unparse(&table_types[i].guid, str); + + if (!efi_guidcmp(*guid, table_types[i].guid)) { + *(table_types[i].ptr) = table; + pr_cont(" %s=0x%lx ", + table_types[i].name, table); + return 1; + } + } + } + + return 0; +} + +int __init efi_config_init(efi_config_table_type_t *arch_tables) +{ + void *config_tables, *tablep; + int i, sz; + + if (efi_enabled(EFI_64BIT)) + sz = sizeof(efi_config_table_64_t); + else + sz = sizeof(efi_config_table_32_t); + + /* + * Let's see what config tables the firmware passed to us. + */ + config_tables = early_memremap(efi.systab->tables, + efi.systab->nr_tables * sz); + if (config_tables == NULL) { + pr_err("Could not map Configuration table!\n"); + return -ENOMEM; + } + + tablep = config_tables; + pr_info(""); + for (i = 0; i < efi.systab->nr_tables; i++) { + efi_guid_t guid; + unsigned long table; + + if (efi_enabled(EFI_64BIT)) { + u64 table64; + guid = ((efi_config_table_64_t *)tablep)->guid; + table64 = ((efi_config_table_64_t *)tablep)->table; + table = table64; +#ifndef CONFIG_64BIT + if (table64 >> 32) { + pr_cont("\n"); + pr_err("Table located above 4GB, disabling EFI.\n"); + early_iounmap(config_tables, + efi.systab->nr_tables * sz); + return -EINVAL; + } +#endif + } else { + guid = ((efi_config_table_32_t *)tablep)->guid; + table = ((efi_config_table_32_t *)tablep)->table; + } + + if (!match_config_table(&guid, table, common_tables)) + match_config_table(&guid, table, arch_tables); + + tablep += sz; + } + pr_cont("\n"); + early_iounmap(config_tables, efi.systab->nr_tables * sz); + return 0; +} + +#ifdef CONFIG_EFI_PARAMS_FROM_FDT + +#define UEFI_PARAM(name, prop, field) \ + { \ + { name }, \ + { prop }, \ + offsetof(struct efi_fdt_params, field), \ + FIELD_SIZEOF(struct efi_fdt_params, field) \ + } + +static __initdata struct { + const char name[32]; + const char propname[32]; + int offset; + int size; +} dt_params[] = { + UEFI_PARAM("System Table", "linux,uefi-system-table", system_table), + UEFI_PARAM("MemMap Address", "linux,uefi-mmap-start", mmap), + UEFI_PARAM("MemMap Size", "linux,uefi-mmap-size", mmap_size), + UEFI_PARAM("MemMap Desc. Size", "linux,uefi-mmap-desc-size", desc_size), + UEFI_PARAM("MemMap Desc. Version", "linux,uefi-mmap-desc-ver", desc_ver) +}; + +struct param_info { + int verbose; + int found; + void *params; +}; + +static int __init fdt_find_uefi_params(unsigned long node, const char *uname, + int depth, void *data) +{ + struct param_info *info = data; + const void *prop; + void *dest; + u64 val; + int i, len; + + if (depth != 1 || + (strcmp(uname, "chosen") != 0 && strcmp(uname, "chosen@0") != 0)) + return 0; + + for (i = 0; i < ARRAY_SIZE(dt_params); i++) { + prop = of_get_flat_dt_prop(node, dt_params[i].propname, &len); + if (!prop) + return 0; + dest = info->params + dt_params[i].offset; + info->found++; + + val = of_read_number(prop, len / sizeof(u32)); + + if (dt_params[i].size == sizeof(u32)) + *(u32 *)dest = val; + else + *(u64 *)dest = val; + + if (info->verbose) + pr_info(" %s: 0x%0*llx\n", dt_params[i].name, + dt_params[i].size * 2, val); + } + return 1; +} + +int __init efi_get_fdt_params(struct efi_fdt_params *params, int verbose) +{ + struct param_info info; + int ret; + + pr_info("Getting EFI parameters from FDT:\n"); + + info.verbose = verbose; + info.found = 0; + info.params = params; + + ret = of_scan_flat_dt(fdt_find_uefi_params, &info); + if (!info.found) + pr_info("UEFI not found.\n"); + else if (!ret) + pr_err("Can't find '%s' in device tree!\n", + dt_params[info.found].name); + + return ret; +} +#endif /* CONFIG_EFI_PARAMS_FROM_FDT */ diff --git a/drivers/firmware/efi/efivars.c b/drivers/firmware/efi/efivars.c index 8bd1bb6dbe47..463c56545ae8 100644 --- a/drivers/firmware/efi/efivars.c +++ b/drivers/firmware/efi/efivars.c @@ -69,6 +69,7 @@ #include <linux/module.h> #include <linux/slab.h> #include <linux/ucs2_string.h> +#include <linux/compat.h> #define EFIVARS_VERSION "0.08" #define EFIVARS_DATE "2004-May-17" @@ -86,6 +87,15 @@ static struct kset *efivars_kset; static struct bin_attribute *efivars_new_var; static struct bin_attribute *efivars_del_var; +struct compat_efi_variable { + efi_char16_t VariableName[EFI_VAR_NAME_LEN/sizeof(efi_char16_t)]; + efi_guid_t VendorGuid; + __u32 DataSize; + __u8 Data[1024]; + __u32 Status; + __u32 Attributes; +} __packed; + struct efivar_attribute { struct attribute attr; ssize_t (*show) (struct efivar_entry *entry, char *buf); @@ -189,45 +199,107 @@ efivar_data_read(struct efivar_entry *entry, char *buf) memcpy(buf, var->Data, var->DataSize); return var->DataSize; } -/* - * We allow each variable to be edited via rewriting the - * entire efi variable structure. - */ -static ssize_t -efivar_store_raw(struct efivar_entry *entry, const char *buf, size_t count) -{ - struct efi_variable *new_var, *var = &entry->var; - int err; - - if (count != sizeof(struct efi_variable)) - return -EINVAL; - new_var = (struct efi_variable *)buf; +static inline int +sanity_check(struct efi_variable *var, efi_char16_t *name, efi_guid_t vendor, + unsigned long size, u32 attributes, u8 *data) +{ /* * If only updating the variable data, then the name * and guid should remain the same */ - if (memcmp(new_var->VariableName, var->VariableName, sizeof(var->VariableName)) || - efi_guidcmp(new_var->VendorGuid, var->VendorGuid)) { + if (memcmp(name, var->VariableName, sizeof(var->VariableName)) || + efi_guidcmp(vendor, var->VendorGuid)) { printk(KERN_ERR "efivars: Cannot edit the wrong variable!\n"); return -EINVAL; } - if ((new_var->DataSize <= 0) || (new_var->Attributes == 0)){ + if ((size <= 0) || (attributes == 0)){ printk(KERN_ERR "efivars: DataSize & Attributes must be valid!\n"); return -EINVAL; } - if ((new_var->Attributes & ~EFI_VARIABLE_MASK) != 0 || - efivar_validate(new_var, new_var->Data, new_var->DataSize) == false) { + if ((attributes & ~EFI_VARIABLE_MASK) != 0 || + efivar_validate(name, data, size) == false) { printk(KERN_ERR "efivars: Malformed variable content\n"); return -EINVAL; } - memcpy(&entry->var, new_var, count); + return 0; +} - err = efivar_entry_set(entry, new_var->Attributes, - new_var->DataSize, new_var->Data, false); +static inline bool is_compat(void) +{ + if (IS_ENABLED(CONFIG_COMPAT) && is_compat_task()) + return true; + + return false; +} + +static void +copy_out_compat(struct efi_variable *dst, struct compat_efi_variable *src) +{ + memcpy(dst->VariableName, src->VariableName, EFI_VAR_NAME_LEN); + memcpy(dst->Data, src->Data, sizeof(src->Data)); + + dst->VendorGuid = src->VendorGuid; + dst->DataSize = src->DataSize; + dst->Attributes = src->Attributes; +} + +/* + * We allow each variable to be edited via rewriting the + * entire efi variable structure. + */ +static ssize_t +efivar_store_raw(struct efivar_entry *entry, const char *buf, size_t count) +{ + struct efi_variable *new_var, *var = &entry->var; + efi_char16_t *name; + unsigned long size; + efi_guid_t vendor; + u32 attributes; + u8 *data; + int err; + + if (is_compat()) { + struct compat_efi_variable *compat; + + if (count != sizeof(*compat)) + return -EINVAL; + + compat = (struct compat_efi_variable *)buf; + attributes = compat->Attributes; + vendor = compat->VendorGuid; + name = compat->VariableName; + size = compat->DataSize; + data = compat->Data; + + err = sanity_check(var, name, vendor, size, attributes, data); + if (err) + return err; + + copy_out_compat(&entry->var, compat); + } else { + if (count != sizeof(struct efi_variable)) + return -EINVAL; + + new_var = (struct efi_variable *)buf; + + attributes = new_var->Attributes; + vendor = new_var->VendorGuid; + name = new_var->VariableName; + size = new_var->DataSize; + data = new_var->Data; + + err = sanity_check(var, name, vendor, size, attributes, data); + if (err) + return err; + + memcpy(&entry->var, new_var, count); + } + + err = efivar_entry_set(entry, attributes, size, data, NULL); if (err) { printk(KERN_WARNING "efivars: set_variable() failed: status=%d\n", err); return -EIO; @@ -240,6 +312,8 @@ static ssize_t efivar_show_raw(struct efivar_entry *entry, char *buf) { struct efi_variable *var = &entry->var; + struct compat_efi_variable *compat; + size_t size; if (!entry || !buf) return 0; @@ -249,9 +323,23 @@ efivar_show_raw(struct efivar_entry *entry, char *buf) &entry->var.DataSize, entry->var.Data)) return -EIO; - memcpy(buf, var, sizeof(*var)); + if (is_compat()) { + compat = (struct compat_efi_variable *)buf; - return sizeof(*var); + size = sizeof(*compat); + memcpy(compat->VariableName, var->VariableName, + EFI_VAR_NAME_LEN); + memcpy(compat->Data, var->Data, sizeof(compat->Data)); + + compat->VendorGuid = var->VendorGuid; + compat->DataSize = var->DataSize; + compat->Attributes = var->Attributes; + } else { + size = sizeof(*var); + memcpy(buf, var, size); + } + + return size; } /* @@ -326,15 +414,39 @@ static ssize_t efivar_create(struct file *filp, struct kobject *kobj, struct bin_attribute *bin_attr, char *buf, loff_t pos, size_t count) { + struct compat_efi_variable *compat = (struct compat_efi_variable *)buf; struct efi_variable *new_var = (struct efi_variable *)buf; struct efivar_entry *new_entry; + bool need_compat = is_compat(); + efi_char16_t *name; + unsigned long size; + u32 attributes; + u8 *data; int err; if (!capable(CAP_SYS_ADMIN)) return -EACCES; - if ((new_var->Attributes & ~EFI_VARIABLE_MASK) != 0 || - efivar_validate(new_var, new_var->Data, new_var->DataSize) == false) { + if (need_compat) { + if (count != sizeof(*compat)) + return -EINVAL; + + attributes = compat->Attributes; + name = compat->VariableName; + size = compat->DataSize; + data = compat->Data; + } else { + if (count != sizeof(*new_var)) + return -EINVAL; + + attributes = new_var->Attributes; + name = new_var->VariableName; + size = new_var->DataSize; + data = new_var->Data; + } + + if ((attributes & ~EFI_VARIABLE_MASK) != 0 || + efivar_validate(name, data, size) == false) { printk(KERN_ERR "efivars: Malformed variable content\n"); return -EINVAL; } @@ -343,10 +455,13 @@ static ssize_t efivar_create(struct file *filp, struct kobject *kobj, if (!new_entry) return -ENOMEM; - memcpy(&new_entry->var, new_var, sizeof(*new_var)); + if (need_compat) + copy_out_compat(&new_entry->var, compat); + else + memcpy(&new_entry->var, new_var, sizeof(*new_var)); - err = efivar_entry_set(new_entry, new_var->Attributes, new_var->DataSize, - new_var->Data, &efivar_sysfs_list); + err = efivar_entry_set(new_entry, attributes, size, + data, &efivar_sysfs_list); if (err) { if (err == -EEXIST) err = -EINVAL; @@ -369,26 +484,47 @@ static ssize_t efivar_delete(struct file *filp, struct kobject *kobj, char *buf, loff_t pos, size_t count) { struct efi_variable *del_var = (struct efi_variable *)buf; + struct compat_efi_variable *compat; struct efivar_entry *entry; + efi_char16_t *name; + efi_guid_t vendor; int err = 0; if (!capable(CAP_SYS_ADMIN)) return -EACCES; + if (is_compat()) { + if (count != sizeof(*compat)) + return -EINVAL; + + compat = (struct compat_efi_variable *)buf; + name = compat->VariableName; + vendor = compat->VendorGuid; + } else { + if (count != sizeof(*del_var)) + return -EINVAL; + + name = del_var->VariableName; + vendor = del_var->VendorGuid; + } + efivar_entry_iter_begin(); - entry = efivar_entry_find(del_var->VariableName, del_var->VendorGuid, - &efivar_sysfs_list, true); + entry = efivar_entry_find(name, vendor, &efivar_sysfs_list, true); if (!entry) err = -EINVAL; else if (__efivar_entry_delete(entry)) err = -EIO; - efivar_entry_iter_end(); - - if (err) + if (err) { + efivar_entry_iter_end(); return err; + } - efivar_unregister(entry); + if (!entry->scanning) { + efivar_entry_iter_end(); + efivar_unregister(entry); + } else + efivar_entry_iter_end(); /* It's dead Jim.... */ return count; @@ -564,7 +700,7 @@ static int efivar_sysfs_destroy(struct efivar_entry *entry, void *data) return 0; } -void efivars_sysfs_exit(void) +static void efivars_sysfs_exit(void) { /* Remove all entries and destroy */ __efivar_entry_iter(efivar_sysfs_destroy, &efivar_sysfs_list, NULL, NULL); @@ -583,6 +719,9 @@ int efivars_sysfs_init(void) struct kobject *parent_kobj = efivars_kobject(); int error = 0; + if (!efi_enabled(EFI_RUNTIME_SERVICES)) + return -ENODEV; + /* No efivars has been registered yet */ if (!parent_kobj) return 0; diff --git a/drivers/firmware/efi/fdt.c b/drivers/firmware/efi/fdt.c new file mode 100644 index 000000000000..507a3df46a5d --- /dev/null +++ b/drivers/firmware/efi/fdt.c @@ -0,0 +1,275 @@ +/* + * FDT related Helper functions used by the EFI stub on multiple + * architectures. This should be #included by the EFI stub + * implementation files. + * + * Copyright 2013 Linaro Limited; author Roy Franz + * + * This file is part of the Linux kernel, and is made available + * under the terms of the GNU General Public License version 2. + * + */ + +static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt, + unsigned long orig_fdt_size, + void *fdt, int new_fdt_size, char *cmdline_ptr, + u64 initrd_addr, u64 initrd_size, + efi_memory_desc_t *memory_map, + unsigned long map_size, unsigned long desc_size, + u32 desc_ver) +{ + int node, prev; + int status; + u32 fdt_val32; + u64 fdt_val64; + + /* Do some checks on provided FDT, if it exists*/ + if (orig_fdt) { + if (fdt_check_header(orig_fdt)) { + pr_efi_err(sys_table, "Device Tree header not valid!\n"); + return EFI_LOAD_ERROR; + } + /* + * We don't get the size of the FDT if we get if from a + * configuration table. + */ + if (orig_fdt_size && fdt_totalsize(orig_fdt) > orig_fdt_size) { + pr_efi_err(sys_table, "Truncated device tree! foo!\n"); + return EFI_LOAD_ERROR; + } + } + + if (orig_fdt) + status = fdt_open_into(orig_fdt, fdt, new_fdt_size); + else + status = fdt_create_empty_tree(fdt, new_fdt_size); + + if (status != 0) + goto fdt_set_fail; + + /* + * Delete any memory nodes present. We must delete nodes which + * early_init_dt_scan_memory may try to use. + */ + prev = 0; + for (;;) { + const char *type; + int len; + + node = fdt_next_node(fdt, prev, NULL); + if (node < 0) + break; + + type = fdt_getprop(fdt, node, "device_type", &len); + if (type && strncmp(type, "memory", len) == 0) { + fdt_del_node(fdt, node); + continue; + } + + prev = node; + } + + node = fdt_subnode_offset(fdt, 0, "chosen"); + if (node < 0) { + node = fdt_add_subnode(fdt, 0, "chosen"); + if (node < 0) { + status = node; /* node is error code when negative */ + goto fdt_set_fail; + } + } + + if ((cmdline_ptr != NULL) && (strlen(cmdline_ptr) > 0)) { + status = fdt_setprop(fdt, node, "bootargs", cmdline_ptr, + strlen(cmdline_ptr) + 1); + if (status) + goto fdt_set_fail; + } + + /* Set initrd address/end in device tree, if present */ + if (initrd_size != 0) { + u64 initrd_image_end; + u64 initrd_image_start = cpu_to_fdt64(initrd_addr); + + status = fdt_setprop(fdt, node, "linux,initrd-start", + &initrd_image_start, sizeof(u64)); + if (status) + goto fdt_set_fail; + initrd_image_end = cpu_to_fdt64(initrd_addr + initrd_size); + status = fdt_setprop(fdt, node, "linux,initrd-end", + &initrd_image_end, sizeof(u64)); + if (status) + goto fdt_set_fail; + } + + /* Add FDT entries for EFI runtime services in chosen node. */ + node = fdt_subnode_offset(fdt, 0, "chosen"); + fdt_val64 = cpu_to_fdt64((u64)(unsigned long)sys_table); + status = fdt_setprop(fdt, node, "linux,uefi-system-table", + &fdt_val64, sizeof(fdt_val64)); + if (status) + goto fdt_set_fail; + + fdt_val64 = cpu_to_fdt64((u64)(unsigned long)memory_map); + status = fdt_setprop(fdt, node, "linux,uefi-mmap-start", + &fdt_val64, sizeof(fdt_val64)); + if (status) + goto fdt_set_fail; + + fdt_val32 = cpu_to_fdt32(map_size); + status = fdt_setprop(fdt, node, "linux,uefi-mmap-size", + &fdt_val32, sizeof(fdt_val32)); + if (status) + goto fdt_set_fail; + + fdt_val32 = cpu_to_fdt32(desc_size); + status = fdt_setprop(fdt, node, "linux,uefi-mmap-desc-size", + &fdt_val32, sizeof(fdt_val32)); + if (status) + goto fdt_set_fail; + + fdt_val32 = cpu_to_fdt32(desc_ver); + status = fdt_setprop(fdt, node, "linux,uefi-mmap-desc-ver", + &fdt_val32, sizeof(fdt_val32)); + if (status) + goto fdt_set_fail; + + /* + * Add kernel version banner so stub/kernel match can be + * verified. + */ + status = fdt_setprop_string(fdt, node, "linux,uefi-stub-kern-ver", + linux_banner); + if (status) + goto fdt_set_fail; + + return EFI_SUCCESS; + +fdt_set_fail: + if (status == -FDT_ERR_NOSPACE) + return EFI_BUFFER_TOO_SMALL; + + return EFI_LOAD_ERROR; +} + +#ifndef EFI_FDT_ALIGN +#define EFI_FDT_ALIGN EFI_PAGE_SIZE +#endif + +/* + * Allocate memory for a new FDT, then add EFI, commandline, and + * initrd related fields to the FDT. This routine increases the + * FDT allocation size until the allocated memory is large + * enough. EFI allocations are in EFI_PAGE_SIZE granules, + * which are fixed at 4K bytes, so in most cases the first + * allocation should succeed. + * EFI boot services are exited at the end of this function. + * There must be no allocations between the get_memory_map() + * call and the exit_boot_services() call, so the exiting of + * boot services is very tightly tied to the creation of the FDT + * with the final memory map in it. + */ + +efi_status_t allocate_new_fdt_and_exit_boot(efi_system_table_t *sys_table, + void *handle, + unsigned long *new_fdt_addr, + unsigned long max_addr, + u64 initrd_addr, u64 initrd_size, + char *cmdline_ptr, + unsigned long fdt_addr, + unsigned long fdt_size) +{ + unsigned long map_size, desc_size; + u32 desc_ver; + unsigned long mmap_key; + efi_memory_desc_t *memory_map; + unsigned long new_fdt_size; + efi_status_t status; + + /* + * Estimate size of new FDT, and allocate memory for it. We + * will allocate a bigger buffer if this ends up being too + * small, so a rough guess is OK here. + */ + new_fdt_size = fdt_size + EFI_PAGE_SIZE; + while (1) { + status = efi_high_alloc(sys_table, new_fdt_size, EFI_FDT_ALIGN, + new_fdt_addr, max_addr); + if (status != EFI_SUCCESS) { + pr_efi_err(sys_table, "Unable to allocate memory for new device tree.\n"); + goto fail; + } + + /* + * Now that we have done our final memory allocation (and free) + * we can get the memory map key needed for + * exit_boot_services(). + */ + status = efi_get_memory_map(sys_table, &memory_map, &map_size, + &desc_size, &desc_ver, &mmap_key); + if (status != EFI_SUCCESS) + goto fail_free_new_fdt; + + status = update_fdt(sys_table, + (void *)fdt_addr, fdt_size, + (void *)*new_fdt_addr, new_fdt_size, + cmdline_ptr, initrd_addr, initrd_size, + memory_map, map_size, desc_size, desc_ver); + + /* Succeeding the first time is the expected case. */ + if (status == EFI_SUCCESS) + break; + + if (status == EFI_BUFFER_TOO_SMALL) { + /* + * We need to allocate more space for the new + * device tree, so free existing buffer that is + * too small. Also free memory map, as we will need + * to get new one that reflects the free/alloc we do + * on the device tree buffer. + */ + efi_free(sys_table, new_fdt_size, *new_fdt_addr); + sys_table->boottime->free_pool(memory_map); + new_fdt_size += EFI_PAGE_SIZE; + } else { + pr_efi_err(sys_table, "Unable to constuct new device tree.\n"); + goto fail_free_mmap; + } + } + + /* Now we are ready to exit_boot_services.*/ + status = sys_table->boottime->exit_boot_services(handle, mmap_key); + + + if (status == EFI_SUCCESS) + return status; + + pr_efi_err(sys_table, "Exit boot services failed.\n"); + +fail_free_mmap: + sys_table->boottime->free_pool(memory_map); + +fail_free_new_fdt: + efi_free(sys_table, new_fdt_size, *new_fdt_addr); + +fail: + return EFI_LOAD_ERROR; +} + +static void *get_fdt(efi_system_table_t *sys_table) +{ + efi_guid_t fdt_guid = DEVICE_TREE_GUID; + efi_config_table_t *tables; + void *fdt; + int i; + + tables = (efi_config_table_t *) sys_table->tables; + fdt = NULL; + + for (i = 0; i < sys_table->nr_tables; i++) + if (efi_guidcmp(tables[i].guid, fdt_guid) == 0) { + fdt = (void *) tables[i].table; + break; + } + + return fdt; +} diff --git a/drivers/firmware/efi/vars.c b/drivers/firmware/efi/vars.c index 7dbc319e1cf5..5abe943e3404 100644 --- a/drivers/firmware/efi/vars.c +++ b/drivers/firmware/efi/vars.c @@ -42,7 +42,7 @@ DECLARE_WORK(efivar_work, NULL); EXPORT_SYMBOL_GPL(efivar_work); static bool -validate_device_path(struct efi_variable *var, int match, u8 *buffer, +validate_device_path(efi_char16_t *var_name, int match, u8 *buffer, unsigned long len) { struct efi_generic_dev_path *node; @@ -75,7 +75,7 @@ validate_device_path(struct efi_variable *var, int match, u8 *buffer, } static bool -validate_boot_order(struct efi_variable *var, int match, u8 *buffer, +validate_boot_order(efi_char16_t *var_name, int match, u8 *buffer, unsigned long len) { /* An array of 16-bit integers */ @@ -86,18 +86,18 @@ validate_boot_order(struct efi_variable *var, int match, u8 *buffer, } static bool -validate_load_option(struct efi_variable *var, int match, u8 *buffer, +validate_load_option(efi_char16_t *var_name, int match, u8 *buffer, unsigned long len) { u16 filepathlength; int i, desclength = 0, namelen; - namelen = ucs2_strnlen(var->VariableName, sizeof(var->VariableName)); + namelen = ucs2_strnlen(var_name, EFI_VAR_NAME_LEN); /* Either "Boot" or "Driver" followed by four digits of hex */ for (i = match; i < match+4; i++) { - if (var->VariableName[i] > 127 || - hex_to_bin(var->VariableName[i] & 0xff) < 0) + if (var_name[i] > 127 || + hex_to_bin(var_name[i] & 0xff) < 0) return true; } @@ -132,12 +132,12 @@ validate_load_option(struct efi_variable *var, int match, u8 *buffer, /* * And, finally, check the filepath */ - return validate_device_path(var, match, buffer + desclength + 6, + return validate_device_path(var_name, match, buffer + desclength + 6, filepathlength); } static bool -validate_uint16(struct efi_variable *var, int match, u8 *buffer, +validate_uint16(efi_char16_t *var_name, int match, u8 *buffer, unsigned long len) { /* A single 16-bit integer */ @@ -148,7 +148,7 @@ validate_uint16(struct efi_variable *var, int match, u8 *buffer, } static bool -validate_ascii_string(struct efi_variable *var, int match, u8 *buffer, +validate_ascii_string(efi_char16_t *var_name, int match, u8 *buffer, unsigned long len) { int i; @@ -166,7 +166,7 @@ validate_ascii_string(struct efi_variable *var, int match, u8 *buffer, struct variable_validate { char *name; - bool (*validate)(struct efi_variable *var, int match, u8 *data, + bool (*validate)(efi_char16_t *var_name, int match, u8 *data, unsigned long len); }; @@ -189,10 +189,10 @@ static const struct variable_validate variable_validate[] = { }; bool -efivar_validate(struct efi_variable *var, u8 *data, unsigned long len) +efivar_validate(efi_char16_t *var_name, u8 *data, unsigned long len) { int i; - u16 *unicode_name = var->VariableName; + u16 *unicode_name = var_name; for (i = 0; variable_validate[i].validate != NULL; i++) { const char *name = variable_validate[i].name; @@ -208,7 +208,7 @@ efivar_validate(struct efi_variable *var, u8 *data, unsigned long len) /* Wildcard in the matching name means we've matched */ if (c == '*') - return variable_validate[i].validate(var, + return variable_validate[i].validate(var_name, match, data, len); /* Case sensitive match */ @@ -217,7 +217,7 @@ efivar_validate(struct efi_variable *var, u8 *data, unsigned long len) /* Reached the end of the string while matching */ if (!c) - return variable_validate[i].validate(var, + return variable_validate[i].validate(var_name, match, data, len); } } @@ -683,8 +683,16 @@ struct efivar_entry *efivar_entry_find(efi_char16_t *name, efi_guid_t guid, if (!found) return NULL; - if (remove) - list_del(&entry->list); + if (remove) { + if (entry->scanning) { + /* + * The entry will be deleted + * after scanning is completed. + */ + entry->deleting = true; + } else + list_del(&entry->list); + } return entry; } @@ -797,7 +805,7 @@ int efivar_entry_set_get_size(struct efivar_entry *entry, u32 attributes, *set = false; - if (efivar_validate(&entry->var, data, *size) == false) + if (efivar_validate(name, data, *size) == false) return -EINVAL; /* diff --git a/drivers/gator/Kconfig b/drivers/gator/Kconfig new file mode 100644 index 000000000000..b2358bbc1293 --- /dev/null +++ b/drivers/gator/Kconfig @@ -0,0 +1,39 @@ +config GATOR + tristate "Gator module for ARM's Streamline Performance Analyzer" + default m if (ARM || ARM64) + depends on PROFILING + depends on HIGH_RES_TIMERS + depends on LOCAL_TIMERS || !(ARM && SMP) + depends on PERF_EVENTS + depends on HW_PERF_EVENTS || !(ARM || ARM64) + select TRACING + help + Gator module for ARM's Streamline Performance Analyzer + +config GATOR_WITH_MALI_SUPPORT + bool + +choice + prompt "Enable Mali GPU support in Gator" + depends on GATOR + optional + help + Enable Mali GPU support in Gator + +config GATOR_MALI_4XXMP + bool "Mali-400MP or Mali-450MP" + select GATOR_WITH_MALI_SUPPORT + +config GATOR_MALI_MIDGARD + bool "Mali-T60x, Mali-T62x, Mali-T72x or Mali-T76x" + select GATOR_WITH_MALI_SUPPORT + +endchoice + +config GATOR_MALI_PATH + string "Path to Mali driver" + depends on GATOR_WITH_MALI_SUPPORT + default "drivers/gpu/arm/mali400mp" + help + The gator code adds this to its include path so it can get the Mali + trace headers with: #include "linux/mali_linux_trace.h" diff --git a/drivers/gator/LICENSE b/drivers/gator/LICENSE new file mode 100644 index 000000000000..d159169d1050 --- /dev/null +++ b/drivers/gator/LICENSE @@ -0,0 +1,339 @@ + GNU GENERAL PUBLIC LICENSE + Version 2, June 1991 + + Copyright (C) 1989, 1991 Free Software Foundation, Inc., + 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + Everyone is permitted to copy and distribute verbatim copies + of this license document, but changing it is not allowed. + + Preamble + + The licenses for most software are designed to take away your +freedom to share and change it. By contrast, the GNU General Public +License is intended to guarantee your freedom to share and change free +software--to make sure the software is free for all its users. This +General Public License applies to most of the Free Software +Foundation's software and to any other program whose authors commit to +using it. (Some other Free Software Foundation software is covered by +the GNU Lesser General Public License instead.) You can apply it to +your programs, too. + + When we speak of free software, we are referring to freedom, not +price. Our General Public Licenses are designed to make sure that you +have the freedom to distribute copies of free software (and charge for +this service if you wish), that you receive source code or can get it +if you want it, that you can change the software or use pieces of it +in new free programs; and that you know you can do these things. + + To protect your rights, we need to make restrictions that forbid +anyone to deny you these rights or to ask you to surrender the rights. +These restrictions translate to certain responsibilities for you if you +distribute copies of the software, or if you modify it. + + For example, if you distribute copies of such a program, whether +gratis or for a fee, you must give the recipients all the rights that +you have. You must make sure that they, too, receive or can get the +source code. And you must show them these terms so they know their +rights. + + We protect your rights with two steps: (1) copyright the software, and +(2) offer you this license which gives you legal permission to copy, +distribute and/or modify the software. + + Also, for each author's protection and ours, we want to make certain +that everyone understands that there is no warranty for this free +software. If the software is modified by someone else and passed on, we +want its recipients to know that what they have is not the original, so +that any problems introduced by others will not reflect on the original +authors' reputations. + + Finally, any free program is threatened constantly by software +patents. We wish to avoid the danger that redistributors of a free +program will individually obtain patent licenses, in effect making the +program proprietary. To prevent this, we have made it clear that any +patent must be licensed for everyone's free use or not licensed at all. + + The precise terms and conditions for copying, distribution and +modification follow. + + GNU GENERAL PUBLIC LICENSE + TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION + + 0. This License applies to any program or other work which contains +a notice placed by the copyright holder saying it may be distributed +under the terms of this General Public License. The "Program", below, +refers to any such program or work, and a "work based on the Program" +means either the Program or any derivative work under copyright law: +that is to say, a work containing the Program or a portion of it, +either verbatim or with modifications and/or translated into another +language. (Hereinafter, translation is included without limitation in +the term "modification".) Each licensee is addressed as "you". + +Activities other than copying, distribution and modification are not +covered by this License; they are outside its scope. The act of +running the Program is not restricted, and the output from the Program +is covered only if its contents constitute a work based on the +Program (independent of having been made by running the Program). +Whether that is true depends on what the Program does. + + 1. You may copy and distribute verbatim copies of the Program's +source code as you receive it, in any medium, provided that you +conspicuously and appropriately publish on each copy an appropriate +copyright notice and disclaimer of warranty; keep intact all the +notices that refer to this License and to the absence of any warranty; +and give any other recipients of the Program a copy of this License +along with the Program. + +You may charge a fee for the physical act of transferring a copy, and +you may at your option offer warranty protection in exchange for a fee. + + 2. You may modify your copy or copies of the Program or any portion +of it, thus forming a work based on the Program, and copy and +distribute such modifications or work under the terms of Section 1 +above, provided that you also meet all of these conditions: + + a) You must cause the modified files to carry prominent notices + stating that you changed the files and the date of any change. + + b) You must cause any work that you distribute or publish, that in + whole or in part contains or is derived from the Program or any + part thereof, to be licensed as a whole at no charge to all third + parties under the terms of this License. + + c) If the modified program normally reads commands interactively + when run, you must cause it, when started running for such + interactive use in the most ordinary way, to print or display an + announcement including an appropriate copyright notice and a + notice that there is no warranty (or else, saying that you provide + a warranty) and that users may redistribute the program under + these conditions, and telling the user how to view a copy of this + License. (Exception: if the Program itself is interactive but + does not normally print such an announcement, your work based on + the Program is not required to print an announcement.) + +These requirements apply to the modified work as a whole. If +identifiable sections of that work are not derived from the Program, +and can be reasonably considered independent and separate works in +themselves, then this License, and its terms, do not apply to those +sections when you distribute them as separate works. But when you +distribute the same sections as part of a whole which is a work based +on the Program, the distribution of the whole must be on the terms of +this License, whose permissions for other licensees extend to the +entire whole, and thus to each and every part regardless of who wrote it. + +Thus, it is not the intent of this section to claim rights or contest +your rights to work written entirely by you; rather, the intent is to +exercise the right to control the distribution of derivative or +collective works based on the Program. + +In addition, mere aggregation of another work not based on the Program +with the Program (or with a work based on the Program) on a volume of +a storage or distribution medium does not bring the other work under +the scope of this License. + + 3. You may copy and distribute the Program (or a work based on it, +under Section 2) in object code or executable form under the terms of +Sections 1 and 2 above provided that you also do one of the following: + + a) Accompany it with the complete corresponding machine-readable + source code, which must be distributed under the terms of Sections + 1 and 2 above on a medium customarily used for software interchange; or, + + b) Accompany it with a written offer, valid for at least three + years, to give any third party, for a charge no more than your + cost of physically performing source distribution, a complete + machine-readable copy of the corresponding source code, to be + distributed under the terms of Sections 1 and 2 above on a medium + customarily used for software interchange; or, + + c) Accompany it with the information you received as to the offer + to distribute corresponding source code. (This alternative is + allowed only for noncommercial distribution and only if you + received the program in object code or executable form with such + an offer, in accord with Subsection b above.) + +The source code for a work means the preferred form of the work for +making modifications to it. For an executable work, complete source +code means all the source code for all modules it contains, plus any +associated interface definition files, plus the scripts used to +control compilation and installation of the executable. However, as a +special exception, the source code distributed need not include +anything that is normally distributed (in either source or binary +form) with the major components (compiler, kernel, and so on) of the +operating system on which the executable runs, unless that component +itself accompanies the executable. + +If distribution of executable or object code is made by offering +access to copy from a designated place, then offering equivalent +access to copy the source code from the same place counts as +distribution of the source code, even though third parties are not +compelled to copy the source along with the object code. + + 4. You may not copy, modify, sublicense, or distribute the Program +except as expressly provided under this License. Any attempt +otherwise to copy, modify, sublicense or distribute the Program is +void, and will automatically terminate your rights under this License. +However, parties who have received copies, or rights, from you under +this License will not have their licenses terminated so long as such +parties remain in full compliance. + + 5. You are not required to accept this License, since you have not +signed it. However, nothing else grants you permission to modify or +distribute the Program or its derivative works. These actions are +prohibited by law if you do not accept this License. Therefore, by +modifying or distributing the Program (or any work based on the +Program), you indicate your acceptance of this License to do so, and +all its terms and conditions for copying, distributing or modifying +the Program or works based on it. + + 6. Each time you redistribute the Program (or any work based on the +Program), the recipient automatically receives a license from the +original licensor to copy, distribute or modify the Program subject to +these terms and conditions. You may not impose any further +restrictions on the recipients' exercise of the rights granted herein. +You are not responsible for enforcing compliance by third parties to +this License. + + 7. If, as a consequence of a court judgment or allegation of patent +infringement or for any other reason (not limited to patent issues), +conditions are imposed on you (whether by court order, agreement or +otherwise) that contradict the conditions of this License, they do not +excuse you from the conditions of this License. If you cannot +distribute so as to satisfy simultaneously your obligations under this +License and any other pertinent obligations, then as a consequence you +may not distribute the Program at all. For example, if a patent +license would not permit royalty-free redistribution of the Program by +all those who receive copies directly or indirectly through you, then +the only way you could satisfy both it and this License would be to +refrain entirely from distribution of the Program. + +If any portion of this section is held invalid or unenforceable under +any particular circumstance, the balance of the section is intended to +apply and the section as a whole is intended to apply in other +circumstances. + +It is not the purpose of this section to induce you to infringe any +patents or other property right claims or to contest validity of any +such claims; this section has the sole purpose of protecting the +integrity of the free software distribution system, which is +implemented by public license practices. Many people have made +generous contributions to the wide range of software distributed +through that system in reliance on consistent application of that +system; it is up to the author/donor to decide if he or she is willing +to distribute software through any other system and a licensee cannot +impose that choice. + +This section is intended to make thoroughly clear what is believed to +be a consequence of the rest of this License. + + 8. If the distribution and/or use of the Program is restricted in +certain countries either by patents or by copyrighted interfaces, the +original copyright holder who places the Program under this License +may add an explicit geographical distribution limitation excluding +those countries, so that distribution is permitted only in or among +countries not thus excluded. In such case, this License incorporates +the limitation as if written in the body of this License. + + 9. The Free Software Foundation may publish revised and/or new versions +of the General Public License from time to time. Such new versions will +be similar in spirit to the present version, but may differ in detail to +address new problems or concerns. + +Each version is given a distinguishing version number. If the Program +specifies a version number of this License which applies to it and "any +later version", you have the option of following the terms and conditions +either of that version or of any later version published by the Free +Software Foundation. If the Program does not specify a version number of +this License, you may choose any version ever published by the Free Software +Foundation. + + 10. If you wish to incorporate parts of the Program into other free +programs whose distribution conditions are different, write to the author +to ask for permission. For software which is copyrighted by the Free +Software Foundation, write to the Free Software Foundation; we sometimes +make exceptions for this. Our decision will be guided by the two goals +of preserving the free status of all derivatives of our free software and +of promoting the sharing and reuse of software generally. + + NO WARRANTY + + 11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY +FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN +OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES +PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED +OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF +MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS +TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE +PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, +REPAIR OR CORRECTION. + + 12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING +WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR +REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, +INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING +OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED +TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY +YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER +PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE +POSSIBILITY OF SUCH DAMAGES. + + END OF TERMS AND CONDITIONS + + How to Apply These Terms to Your New Programs + + If you develop a new program, and you want it to be of the greatest +possible use to the public, the best way to achieve this is to make it +free software which everyone can redistribute and change under these terms. + + To do so, attach the following notices to the program. It is safest +to attach them to the start of each source file to most effectively +convey the exclusion of warranty; and each file should have at least +the "copyright" line and a pointer to where the full notice is found. + + <one line to give the program's name and a brief idea of what it does.> + Copyright (C) <year> <name of author> + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License along + with this program; if not, write to the Free Software Foundation, Inc., + 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + +Also add information on how to contact you by electronic and paper mail. + +If the program is interactive, make it output a short notice like this +when it starts in an interactive mode: + + Gnomovision version 69, Copyright (C) year name of author + Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. + This is free software, and you are welcome to redistribute it + under certain conditions; type `show c' for details. + +The hypothetical commands `show w' and `show c' should show the appropriate +parts of the General Public License. Of course, the commands you use may +be called something other than `show w' and `show c'; they could even be +mouse-clicks or menu items--whatever suits your program. + +You should also get your employer (if you work as a programmer) or your +school, if any, to sign a "copyright disclaimer" for the program, if +necessary. Here is a sample; alter the names: + + Yoyodyne, Inc., hereby disclaims all copyright interest in the program + `Gnomovision' (which makes passes at compilers) written by James Hacker. + + <signature of Ty Coon>, 1 April 1989 + Ty Coon, President of Vice + +This General Public License does not permit incorporating your program into +proprietary programs. If your program is a subroutine library, you may +consider it more useful to permit linking proprietary applications with the +library. If this is what you want to do, use the GNU Lesser General +Public License instead of this License. diff --git a/drivers/gator/Makefile b/drivers/gator/Makefile new file mode 100644 index 000000000000..28d2070b11d5 --- /dev/null +++ b/drivers/gator/Makefile @@ -0,0 +1,79 @@ +ifneq ($(KERNELRELEASE),) + +# Uncomment the following line to enable kernel stack unwinding within gator, or update gator_backtrace.c +# EXTRA_CFLAGS += -DGATOR_KERNEL_STACK_UNWINDING + +CONFIG_GATOR ?= m +obj-$(CONFIG_GATOR) := gator.o + +gator-y := gator_main.o \ + gator_events_block.o \ + gator_events_irq.o \ + gator_events_meminfo.o \ + gator_events_mmapped.o \ + gator_events_net.o \ + gator_events_perf_pmu.o \ + gator_events_sched.o \ + +# Convert the old GATOR_WITH_MALI_SUPPORT to the new kernel flags +ifneq ($(GATOR_WITH_MALI_SUPPORT),) + CONFIG_GATOR_WITH_MALI_SUPPORT := y + ifeq ($(GATOR_WITH_MALI_SUPPORT),MALI_MIDGARD) + CONFIG_GATOR_MALI_4XXMP := n + CONFIG_GATOR_MALI_MIDGARD := y + else + CONFIG_GATOR_MALI_4XXMP := y + CONFIG_GATOR_MALI_MIDGARD := n + endif + EXTRA_CFLAGS += -DMALI_SUPPORT=$(GATOR_WITH_MALI_SUPPORT) + ifneq ($(GATOR_MALI_INTERFACE_STYLE),) + EXTRA_CFLAGS += -DGATOR_MALI_INTERFACE_STYLE=$(GATOR_MALI_INTERFACE_STYLE) + endif +endif + +ifeq ($(CONFIG_GATOR_WITH_MALI_SUPPORT),y) + ifeq ($(CONFIG_GATOR_MALI_MIDGARD),y) + gator-y += gator_events_mali_midgard.o \ + gator_events_mali_midgard_hw.o + include $(src)/mali_midgard.mk + else + gator-y += gator_events_mali_4xx.o + endif + gator-y += gator_events_mali_common.o + + ifneq ($(CONFIG_GATOR_MALI_PATH),) + ccflags-y += -I$(CONFIG_GATOR_MALI_PATH) + endif + ccflags-$(CONFIG_GATOR_MALI_4XXMP) += -DMALI_SUPPORT=MALI_4xx + ccflags-$(CONFIG_GATOR_MALI_MIDGARD) += -DMALI_SUPPORT=MALI_MIDGARD +endif + +# GATOR_TEST controls whether to include (=1) or exclude (=0) test code. +GATOR_TEST ?= 0 +EXTRA_CFLAGS += -DGATOR_TEST=$(GATOR_TEST) + +# Should the original or new block_rq_complete API be used? +OLD_BLOCK_RQ_COMPLETE := $(shell grep -A3 block_rq_complete $(srctree)/include/trace/events/block.h | grep nr_bytes -q; echo $$?) +EXTRA_CFLAGS += -DOLD_BLOCK_RQ_COMPLETE=$(OLD_BLOCK_RQ_COMPLETE) + +gator-$(CONFIG_ARM) += gator_events_armv6.o \ + gator_events_armv7.o \ + gator_events_l2c-310.o \ + gator_events_scorpion.o + +gator-$(CONFIG_ARM64) += + +else + +all: + @echo + @echo "usage:" + @echo " make -C <kernel_build_dir> M=\`pwd\` ARCH=arm CROSS_COMPILE=<...> modules" + @echo + $(error) + +clean: + rm -f *.o .*.cmd modules.order Module.symvers gator.ko gator.mod.c + rm -rf .tmp_versions + +endif diff --git a/drivers/gator/gator.h b/drivers/gator/gator.h new file mode 100644 index 000000000000..5cc73a388c4f --- /dev/null +++ b/drivers/gator/gator.h @@ -0,0 +1,152 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#ifndef GATOR_H_ +#define GATOR_H_ + +#include <linux/version.h> +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/list.h> + +#define GATOR_PERF_SUPPORT (LINUX_VERSION_CODE >= KERNEL_VERSION(3, 0, 0)) +#define GATOR_PERF_PMU_SUPPORT (GATOR_PERF_SUPPORT && defined(CONFIG_PERF_EVENTS) && (!(defined(__arm__) || defined(__aarch64__)) || defined(CONFIG_HW_PERF_EVENTS))) +#define GATOR_NO_PERF_SUPPORT (!(GATOR_PERF_SUPPORT)) +#define GATOR_CPU_FREQ_SUPPORT ((LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 38)) && defined(CONFIG_CPU_FREQ)) +#define GATOR_IKS_SUPPORT defined(CONFIG_BL_SWITCHER) + +/* cpu ids */ +#define ARM1136 0xb36 +#define ARM1156 0xb56 +#define ARM1176 0xb76 +#define ARM11MPCORE 0xb02 +#define CORTEX_A5 0xc05 +#define CORTEX_A7 0xc07 +#define CORTEX_A8 0xc08 +#define CORTEX_A9 0xc09 +#define CORTEX_A15 0xc0f +#define CORTEX_A17 0xc0e +#define SCORPION 0x00f +#define SCORPIONMP 0x02d +#define KRAITSIM 0x049 +#define KRAIT 0x04d +#define KRAIT_S4_PRO 0x06f +#define CORTEX_A53 0xd03 +#define CORTEX_A57 0xd07 +#define AARCH64 0xd0f +#define OTHER 0xfff + +/* gpu enums */ +#define MALI_4xx 1 +#define MALI_MIDGARD 2 + +#define MAXSIZE_CORE_NAME 32 + +struct gator_cpu { + const int cpuid; + /* Human readable name */ + const char core_name[MAXSIZE_CORE_NAME]; + /* gatorfs event and Perf PMU name */ + const char *const pmnc_name; + /* compatible from Documentation/devicetree/bindings/arm/cpus.txt */ + const char *const dt_name; + const int pmnc_counters; +}; + +const struct gator_cpu *gator_find_cpu_by_cpuid(const u32 cpuid); +const struct gator_cpu *gator_find_cpu_by_pmu_name(const char *const name); + +/****************************************************************************** + * Filesystem + ******************************************************************************/ +struct dentry *gatorfs_mkdir(struct super_block *sb, struct dentry *root, + char const *name); + +int gatorfs_create_ulong(struct super_block *sb, struct dentry *root, + char const *name, unsigned long *val); + +int gatorfs_create_ro_ulong(struct super_block *sb, struct dentry *root, + char const *name, unsigned long *val); + +/****************************************************************************** + * Tracepoints + ******************************************************************************/ +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 32) +# error Kernels prior to 2.6.32 not supported +#elif LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 35) +# define GATOR_DEFINE_PROBE(probe_name, proto) \ + static void probe_##probe_name(PARAMS(proto)) +# define GATOR_REGISTER_TRACE(probe_name) \ + register_trace_##probe_name(probe_##probe_name) +# define GATOR_UNREGISTER_TRACE(probe_name) \ + unregister_trace_##probe_name(probe_##probe_name) +#elif LINUX_VERSION_CODE < KERNEL_VERSION(3, 15, 0) +# define GATOR_DEFINE_PROBE(probe_name, proto) \ + static void probe_##probe_name(void *data, PARAMS(proto)) +# define GATOR_REGISTER_TRACE(probe_name) \ + register_trace_##probe_name(probe_##probe_name, NULL) +# define GATOR_UNREGISTER_TRACE(probe_name) \ + unregister_trace_##probe_name(probe_##probe_name, NULL) +#else +# define GATOR_DEFINE_PROBE(probe_name, proto) \ + extern struct tracepoint *gator_tracepoint_##probe_name; \ + static void probe_##probe_name(void *data, PARAMS(proto)) +# define GATOR_REGISTER_TRACE(probe_name) \ + ((gator_tracepoint_##probe_name == NULL) || tracepoint_probe_register(gator_tracepoint_##probe_name, probe_##probe_name, NULL)) +# define GATOR_UNREGISTER_TRACE(probe_name) \ + tracepoint_probe_unregister(gator_tracepoint_##probe_name, probe_##probe_name, NULL) +#endif + +/****************************************************************************** + * Events + ******************************************************************************/ +struct gator_interface { + /* Complementary function to init */ + void (*shutdown)(void); + int (*create_files)(struct super_block *sb, struct dentry *root); + int (*start)(void); + /* Complementary function to start */ + void (*stop)(void); + int (*online)(int **buffer, bool migrate); + int (*offline)(int **buffer, bool migrate); + /* called in process context but may not be running on core 'cpu' */ + void (*online_dispatch)(int cpu, bool migrate); + /* called in process context but may not be running on core 'cpu' */ + void (*offline_dispatch)(int cpu, bool migrate); + int (*read)(int **buffer, bool sched_switch); + int (*read64)(long long **buffer); + int (*read_proc)(long long **buffer, struct task_struct *); + struct list_head list; +}; + +int gator_events_install(struct gator_interface *interface); +int gator_events_get_key(void); +u32 gator_cpuid(void); + +void gator_backtrace_handler(struct pt_regs *const regs); + +void gator_marshal_activity_switch(int core, int key, int activity, int pid); + +#if !GATOR_IKS_SUPPORT + +#define get_physical_cpu() smp_processor_id() +#define lcpu_to_pcpu(lcpu) lcpu +#define pcpu_to_lcpu(pcpu) pcpu + +#else + +#define get_physical_cpu() lcpu_to_pcpu(get_logical_cpu()) +int lcpu_to_pcpu(const int lcpu); +int pcpu_to_lcpu(const int pcpu); + +#endif + +#define get_logical_cpu() smp_processor_id() +#define on_primary_core() (get_logical_cpu() == 0) + +#endif /* GATOR_H_ */ diff --git a/drivers/gator/gator_annotate.c b/drivers/gator/gator_annotate.c new file mode 100644 index 000000000000..ff9a3cef7b2e --- /dev/null +++ b/drivers/gator/gator_annotate.c @@ -0,0 +1,189 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include <linux/slab.h> +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/sched.h> +#include <linux/uaccess.h> +#include <asm/current.h> +#include <linux/spinlock.h> + +static DEFINE_SPINLOCK(annotate_lock); +static bool collect_annotations; + +static int annotate_copy(struct file *file, char const __user *buf, size_t count) +{ + int cpu = 0; + int write = per_cpu(gator_buffer_write, cpu)[ANNOTATE_BUF]; + + if (file == NULL) { + /* copy from kernel */ + memcpy(&per_cpu(gator_buffer, cpu)[ANNOTATE_BUF][write], buf, count); + } else { + /* copy from user space */ + if (copy_from_user(&per_cpu(gator_buffer, cpu)[ANNOTATE_BUF][write], buf, count) != 0) + return -1; + } + per_cpu(gator_buffer_write, cpu)[ANNOTATE_BUF] = (write + count) & gator_buffer_mask[ANNOTATE_BUF]; + + return 0; +} + +static ssize_t annotate_write(struct file *file, char const __user *buf, size_t count_orig, loff_t *offset) +{ + int pid, cpu, header_size, available, contiguous, length1, length2, size, count = count_orig & 0x7fffffff; + bool interrupt_context; + + if (*offset) + return -EINVAL; + + interrupt_context = in_interrupt(); + /* Annotations are not supported in interrupt context, but may work + * if you comment out the the next four lines of code. By doing so, + * annotations in interrupt context can result in deadlocks and lost + * data. + */ + if (interrupt_context) { + pr_warning("gator: Annotations are not supported in interrupt context. Edit gator_annotate.c in the gator driver to enable annotations in interrupt context.\n"); + return -EINVAL; + } + + retry: + /* synchronize between cores and with collect_annotations */ + spin_lock(&annotate_lock); + + if (!collect_annotations) { + /* Not collecting annotations, tell the caller everything was written */ + size = count_orig; + goto annotate_write_out; + } + + /* Annotation only uses a single per-cpu buffer as the data must be in order to the engine */ + cpu = 0; + + if (current == NULL) + pid = 0; + else + pid = current->pid; + + /* determine total size of the payload */ + header_size = MAXSIZE_PACK32 * 3 + MAXSIZE_PACK64; + available = buffer_bytes_available(cpu, ANNOTATE_BUF) - header_size; + size = count < available ? count : available; + + if (size <= 0) { + /* Buffer is full, wait until space is available */ + spin_unlock(&annotate_lock); + + /* Drop the annotation as blocking is not allowed in interrupt context */ + if (interrupt_context) + return -EINVAL; + + wait_event_interruptible(gator_annotate_wait, buffer_bytes_available(cpu, ANNOTATE_BUF) > header_size || !collect_annotations); + + /* Check to see if a signal is pending */ + if (signal_pending(current)) + return -EINTR; + + goto retry; + } + + /* synchronize shared variables annotateBuf and annotatePos */ + if (per_cpu(gator_buffer, cpu)[ANNOTATE_BUF]) { + u64 time = gator_get_time(); + + gator_buffer_write_packed_int(cpu, ANNOTATE_BUF, get_physical_cpu()); + gator_buffer_write_packed_int(cpu, ANNOTATE_BUF, pid); + gator_buffer_write_packed_int64(cpu, ANNOTATE_BUF, time); + gator_buffer_write_packed_int(cpu, ANNOTATE_BUF, size); + + /* determine the sizes to capture, length1 + length2 will equal size */ + contiguous = contiguous_space_available(cpu, ANNOTATE_BUF); + if (size < contiguous) { + length1 = size; + length2 = 0; + } else { + length1 = contiguous; + length2 = size - contiguous; + } + + if (annotate_copy(file, buf, length1) != 0) { + size = -EINVAL; + goto annotate_write_out; + } + + if (length2 > 0 && annotate_copy(file, &buf[length1], length2) != 0) { + size = -EINVAL; + goto annotate_write_out; + } + + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, ANNOTATE_BUF, time); + } + +annotate_write_out: + spin_unlock(&annotate_lock); + + /* return the number of bytes written */ + return size; +} + +#include "gator_annotate_kernel.c" + +static int annotate_release(struct inode *inode, struct file *file) +{ + int cpu = 0; + + /* synchronize between cores */ + spin_lock(&annotate_lock); + + if (per_cpu(gator_buffer, cpu)[ANNOTATE_BUF] && buffer_check_space(cpu, ANNOTATE_BUF, MAXSIZE_PACK64 + 3 * MAXSIZE_PACK32)) { + uint32_t pid = current->pid; + + gator_buffer_write_packed_int(cpu, ANNOTATE_BUF, get_physical_cpu()); + gator_buffer_write_packed_int(cpu, ANNOTATE_BUF, pid); + /* time */ + gator_buffer_write_packed_int64(cpu, ANNOTATE_BUF, 0); + /* size */ + gator_buffer_write_packed_int(cpu, ANNOTATE_BUF, 0); + } + + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, ANNOTATE_BUF, gator_get_time()); + + spin_unlock(&annotate_lock); + + return 0; +} + +static const struct file_operations annotate_fops = { + .write = annotate_write, + .release = annotate_release +}; + +static int gator_annotate_create_files(struct super_block *sb, struct dentry *root) +{ + return gatorfs_create_file_perm(sb, root, "annotate", &annotate_fops, 0666); +} + +static int gator_annotate_start(void) +{ + collect_annotations = true; + return 0; +} + +static void gator_annotate_stop(void) +{ + /* the spinlock here will ensure that when this function exits, we are not in the middle of an annotation */ + spin_lock(&annotate_lock); + collect_annotations = false; + wake_up(&gator_annotate_wait); + spin_unlock(&annotate_lock); +} diff --git a/drivers/gator/gator_annotate_kernel.c b/drivers/gator/gator_annotate_kernel.c new file mode 100644 index 000000000000..69471f99e5fb --- /dev/null +++ b/drivers/gator/gator_annotate_kernel.c @@ -0,0 +1,200 @@ +/** + * Copyright (C) ARM Limited 2012-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#define ESCAPE_CODE 0x1c +#define STRING_ANNOTATION 0x06 +#define NAME_CHANNEL_ANNOTATION 0x07 +#define NAME_GROUP_ANNOTATION 0x08 +#define VISUAL_ANNOTATION 0x04 +#define MARKER_ANNOTATION 0x05 + +static void kannotate_write(const char *ptr, unsigned int size) +{ + int retval; + int pos = 0; + loff_t offset = 0; + + while (pos < size) { + retval = annotate_write(NULL, &ptr[pos], size - pos, &offset); + if (retval < 0) { + pr_warning("gator: kannotate_write failed with return value %d\n", retval); + return; + } + pos += retval; + } +} + +static void marshal_u16(char *buf, u16 val) +{ + buf[0] = val & 0xff; + buf[1] = (val >> 8) & 0xff; +} + +static void marshal_u32(char *buf, u32 val) +{ + buf[0] = val & 0xff; + buf[1] = (val >> 8) & 0xff; + buf[2] = (val >> 16) & 0xff; + buf[3] = (val >> 24) & 0xff; +} + +void gator_annotate_channel(int channel, const char *str) +{ + const u16 str_size = strlen(str) & 0xffff; + char header[8]; + + header[0] = ESCAPE_CODE; + header[1] = STRING_ANNOTATION; + marshal_u32(header + 2, channel); + marshal_u16(header + 6, str_size); + kannotate_write(header, sizeof(header)); + kannotate_write(str, str_size); +} +EXPORT_SYMBOL(gator_annotate_channel); + +void gator_annotate(const char *str) +{ + gator_annotate_channel(0, str); +} +EXPORT_SYMBOL(gator_annotate); + +void gator_annotate_channel_color(int channel, int color, const char *str) +{ + const u16 str_size = (strlen(str) + 4) & 0xffff; + char header[12]; + + header[0] = ESCAPE_CODE; + header[1] = STRING_ANNOTATION; + marshal_u32(header + 2, channel); + marshal_u16(header + 6, str_size); + marshal_u32(header + 8, color); + kannotate_write(header, sizeof(header)); + kannotate_write(str, str_size - 4); +} +EXPORT_SYMBOL(gator_annotate_channel_color); + +void gator_annotate_color(int color, const char *str) +{ + gator_annotate_channel_color(0, color, str); +} +EXPORT_SYMBOL(gator_annotate_color); + +void gator_annotate_channel_end(int channel) +{ + char header[8]; + + header[0] = ESCAPE_CODE; + header[1] = STRING_ANNOTATION; + marshal_u32(header + 2, channel); + marshal_u16(header + 6, 0); + kannotate_write(header, sizeof(header)); +} +EXPORT_SYMBOL(gator_annotate_channel_end); + +void gator_annotate_end(void) +{ + gator_annotate_channel_end(0); +} +EXPORT_SYMBOL(gator_annotate_end); + +void gator_annotate_name_channel(int channel, int group, const char *str) +{ + const u16 str_size = strlen(str) & 0xffff; + char header[12]; + + header[0] = ESCAPE_CODE; + header[1] = NAME_CHANNEL_ANNOTATION; + marshal_u32(header + 2, channel); + marshal_u32(header + 6, group); + marshal_u16(header + 10, str_size); + kannotate_write(header, sizeof(header)); + kannotate_write(str, str_size); +} +EXPORT_SYMBOL(gator_annotate_name_channel); + +void gator_annotate_name_group(int group, const char *str) +{ + const u16 str_size = strlen(str) & 0xffff; + char header[8]; + + header[0] = ESCAPE_CODE; + header[1] = NAME_GROUP_ANNOTATION; + marshal_u32(header + 2, group); + marshal_u16(header + 6, str_size); + kannotate_write(header, sizeof(header)); + kannotate_write(str, str_size); +} +EXPORT_SYMBOL(gator_annotate_name_group); + +void gator_annotate_visual(const char *data, unsigned int length, const char *str) +{ + const u16 str_size = strlen(str) & 0xffff; + char header[4]; + char header_length[4]; + + header[0] = ESCAPE_CODE; + header[1] = VISUAL_ANNOTATION; + marshal_u16(header + 2, str_size); + marshal_u32(header_length, length); + kannotate_write(header, sizeof(header)); + kannotate_write(str, str_size); + kannotate_write(header_length, sizeof(header_length)); + kannotate_write(data, length); +} +EXPORT_SYMBOL(gator_annotate_visual); + +void gator_annotate_marker(void) +{ + char header[4]; + + header[0] = ESCAPE_CODE; + header[1] = MARKER_ANNOTATION; + marshal_u16(header + 2, 0); + kannotate_write(header, sizeof(header)); +} +EXPORT_SYMBOL(gator_annotate_marker); + +void gator_annotate_marker_str(const char *str) +{ + const u16 str_size = strlen(str) & 0xffff; + char header[4]; + + header[0] = ESCAPE_CODE; + header[1] = MARKER_ANNOTATION; + marshal_u16(header + 2, str_size); + kannotate_write(header, sizeof(header)); + kannotate_write(str, str_size); +} +EXPORT_SYMBOL(gator_annotate_marker_str); + +void gator_annotate_marker_color(int color) +{ + char header[8]; + + header[0] = ESCAPE_CODE; + header[1] = MARKER_ANNOTATION; + marshal_u16(header + 2, 4); + marshal_u32(header + 4, color); + kannotate_write(header, sizeof(header)); +} +EXPORT_SYMBOL(gator_annotate_marker_color); + +void gator_annotate_marker_color_str(int color, const char *str) +{ + const u16 str_size = (strlen(str) + 4) & 0xffff; + char header[8]; + + header[0] = ESCAPE_CODE; + header[1] = MARKER_ANNOTATION; + marshal_u16(header + 2, str_size); + marshal_u32(header + 4, color); + kannotate_write(header, sizeof(header)); + kannotate_write(str, str_size - 4); +} +EXPORT_SYMBOL(gator_annotate_marker_color_str); diff --git a/drivers/gator/gator_backtrace.c b/drivers/gator/gator_backtrace.c new file mode 100644 index 000000000000..76c941d009a9 --- /dev/null +++ b/drivers/gator/gator_backtrace.c @@ -0,0 +1,208 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +/* + * EABI backtrace stores {fp,lr} on the stack. + */ +struct stack_frame_eabi { + union { + struct { + unsigned long fp; + /* May be the fp in the case of a leaf function or clang */ + unsigned long lr; + /* If lr is really the fp, lr2 is the corresponding lr */ + unsigned long lr2; + }; + /* Used to read 32 bit fp/lr from a 64 bit kernel */ + struct { + u32 fp_32; + /* same as lr above */ + u32 lr_32; + /* same as lr2 above */ + u32 lr2_32; + }; + }; +}; + +static void gator_add_trace(int cpu, unsigned long address) +{ + off_t offset = 0; + unsigned long cookie = get_address_cookie(cpu, current, address & ~1, &offset); + + if (cookie == NO_COOKIE || cookie == UNRESOLVED_COOKIE) + offset = address; + + marshal_backtrace(offset & ~1, cookie, 0); +} + +static void arm_backtrace_eabi(int cpu, struct pt_regs *const regs, unsigned int depth) +{ +#if defined(__arm__) || defined(__aarch64__) + struct stack_frame_eabi *curr; + struct stack_frame_eabi bufcurr; +#if defined(__arm__) + const bool is_compat = false; + unsigned long fp = regs->ARM_fp; + unsigned long sp = regs->ARM_sp; + unsigned long lr = regs->ARM_lr; + const int gcc_frame_offset = sizeof(unsigned long); +#else + /* Is userspace aarch32 (32 bit) */ + const bool is_compat = compat_user_mode(regs); + unsigned long fp = (is_compat ? regs->regs[11] : regs->regs[29]); + unsigned long sp = (is_compat ? regs->compat_sp : regs->sp); + unsigned long lr = (is_compat ? regs->compat_lr : regs->regs[30]); + const int gcc_frame_offset = (is_compat ? sizeof(u32) : 0); +#endif + /* clang frame offset is always zero */ + int is_user_mode = user_mode(regs); + + /* pc (current function) has already been added */ + + if (!is_user_mode) + return; + + /* Add the lr (parent function), entry preamble may not have + * executed + */ + gator_add_trace(cpu, lr); + + /* check fp is valid */ + if (fp == 0 || fp < sp) + return; + + /* Get the current stack frame */ + curr = (struct stack_frame_eabi *)(fp - gcc_frame_offset); + if ((unsigned long)curr & 3) + return; + + while (depth-- && curr) { + if (!access_ok(VERIFY_READ, curr, sizeof(struct stack_frame_eabi)) || + __copy_from_user_inatomic(&bufcurr, curr, sizeof(struct stack_frame_eabi))) { + return; + } + + fp = (is_compat ? bufcurr.fp_32 : bufcurr.fp); + lr = (is_compat ? bufcurr.lr_32 : bufcurr.lr); + +#define calc_next(reg) ((reg) - gcc_frame_offset) + /* Returns true if reg is a valid fp */ +#define validate_next(reg, curr) \ + ((reg) != 0 && (calc_next(reg) & 3) == 0 && (unsigned long)(curr) < calc_next(reg)) + + /* Try lr from the stack as the fp because gcc leaf functions do + * not push lr. If gcc_frame_offset is non-zero, the lr will also + * be the clang fp. This assumes code is at a lower address than + * the stack + */ + if (validate_next(lr, curr)) { + fp = lr; + lr = (is_compat ? bufcurr.lr2_32 : bufcurr.lr2); + } + + gator_add_trace(cpu, lr); + + if (!validate_next(fp, curr)) + return; + + /* Move to the next stack frame */ + curr = (struct stack_frame_eabi *)calc_next(fp); + } +#endif +} + +#if defined(__arm__) || defined(__aarch64__) +static int report_trace(struct stackframe *frame, void *d) +{ + unsigned int *depth = d, cookie = NO_COOKIE; + unsigned long addr = frame->pc; + + if (*depth) { +#if defined(MODULE) + unsigned int cpu = get_physical_cpu(); + struct module *mod = __module_address(addr); + + if (mod) { + cookie = get_cookie(cpu, current, mod->name, false); + addr = addr - (unsigned long)mod->module_core; + } +#endif + marshal_backtrace(addr & ~1, cookie, 1); + (*depth)--; + } + + return *depth == 0; +} +#endif + +/* Uncomment the following line to enable kernel stack unwinding within gator, note it can also be defined from the Makefile */ +/* #define GATOR_KERNEL_STACK_UNWINDING */ + +#if (defined(__arm__) || defined(__aarch64__)) && !defined(GATOR_KERNEL_STACK_UNWINDING) +/* Disabled by default */ +MODULE_PARM_DESC(kernel_stack_unwinding, "Allow kernel stack unwinding."); +static bool kernel_stack_unwinding; +module_param(kernel_stack_unwinding, bool, 0644); +#endif + +static void kernel_backtrace(int cpu, struct pt_regs *const regs) +{ +#if defined(__arm__) || defined(__aarch64__) +#ifdef GATOR_KERNEL_STACK_UNWINDING + int depth = gator_backtrace_depth; +#else + int depth = (kernel_stack_unwinding ? gator_backtrace_depth : 1); +#endif + struct stackframe frame; + + if (depth == 0) + depth = 1; +#if defined(__arm__) + frame.fp = regs->ARM_fp; + frame.sp = regs->ARM_sp; + frame.lr = regs->ARM_lr; + frame.pc = regs->ARM_pc; +#else + frame.fp = regs->regs[29]; + frame.sp = regs->sp; + frame.pc = regs->pc; +#endif + walk_stackframe(&frame, report_trace, &depth); +#else + marshal_backtrace(PC_REG & ~1, NO_COOKIE, 1); +#endif +} + +static void gator_add_sample(int cpu, struct pt_regs *const regs, u64 time) +{ + bool in_kernel; + unsigned long exec_cookie; + + if (!regs) + return; + + in_kernel = !user_mode(regs); + exec_cookie = get_exec_cookie(cpu, current); + + if (!marshal_backtrace_header(exec_cookie, current->tgid, current->pid, time)) + return; + + if (in_kernel) { + kernel_backtrace(cpu, regs); + } else { + /* Cookie+PC */ + gator_add_trace(cpu, PC_REG); + + /* Backtrace */ + if (gator_backtrace_depth) + arm_backtrace_eabi(cpu, regs, gator_backtrace_depth); + } + + marshal_backtrace_footer(time); +} diff --git a/drivers/gator/gator_buffer.c b/drivers/gator/gator_buffer.c new file mode 100644 index 000000000000..910d5aa15066 --- /dev/null +++ b/drivers/gator/gator_buffer.c @@ -0,0 +1,171 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +static void marshal_frame(int cpu, int buftype) +{ + int frame; + bool write_cpu; + + if (!per_cpu(gator_buffer, cpu)[buftype]) + return; + + switch (buftype) { + case SUMMARY_BUF: + write_cpu = false; + frame = FRAME_SUMMARY; + break; + case BACKTRACE_BUF: + write_cpu = true; + frame = FRAME_BACKTRACE; + break; + case NAME_BUF: + write_cpu = true; + frame = FRAME_NAME; + break; + case COUNTER_BUF: + write_cpu = false; + frame = FRAME_COUNTER; + break; + case BLOCK_COUNTER_BUF: + write_cpu = true; + frame = FRAME_BLOCK_COUNTER; + break; + case ANNOTATE_BUF: + write_cpu = false; + frame = FRAME_ANNOTATE; + break; + case SCHED_TRACE_BUF: + write_cpu = true; + frame = FRAME_SCHED_TRACE; + break; + case IDLE_BUF: + write_cpu = false; + frame = FRAME_IDLE; + break; + case ACTIVITY_BUF: + write_cpu = false; + frame = FRAME_ACTIVITY; + break; + default: + write_cpu = false; + frame = -1; + break; + } + + /* add response type */ + if (gator_response_type > 0) + gator_buffer_write_packed_int(cpu, buftype, gator_response_type); + + /* leave space for 4-byte unpacked length */ + per_cpu(gator_buffer_write, cpu)[buftype] = (per_cpu(gator_buffer_write, cpu)[buftype] + sizeof(s32)) & gator_buffer_mask[buftype]; + + /* add frame type and core number */ + gator_buffer_write_packed_int(cpu, buftype, frame); + if (write_cpu) + gator_buffer_write_packed_int(cpu, buftype, cpu); +} + +static int buffer_bytes_available(int cpu, int buftype) +{ + int remaining, filled; + + filled = per_cpu(gator_buffer_write, cpu)[buftype] - per_cpu(gator_buffer_read, cpu)[buftype]; + if (filled < 0) + filled += gator_buffer_size[buftype]; + + remaining = gator_buffer_size[buftype] - filled; + + if (per_cpu(buffer_space_available, cpu)[buftype]) + /* Give some extra room; also allows space to insert the overflow error packet */ + remaining -= 200; + else + /* Hysteresis, prevents multiple overflow messages */ + remaining -= 2000; + + return remaining; +} + +static bool buffer_check_space(int cpu, int buftype, int bytes) +{ + int remaining = buffer_bytes_available(cpu, buftype); + + if (remaining < bytes) + per_cpu(buffer_space_available, cpu)[buftype] = false; + else + per_cpu(buffer_space_available, cpu)[buftype] = true; + + return per_cpu(buffer_space_available, cpu)[buftype]; +} + +static int contiguous_space_available(int cpu, int buftype) +{ + int remaining = buffer_bytes_available(cpu, buftype); + int contiguous = gator_buffer_size[buftype] - per_cpu(gator_buffer_write, cpu)[buftype]; + + if (remaining < contiguous) + return remaining; + return contiguous; +} + +static void gator_commit_buffer(int cpu, int buftype, u64 time) +{ + int type_length, commit, length, byte; + unsigned long flags; + + if (!per_cpu(gator_buffer, cpu)[buftype]) + return; + + /* post-populate the length, which does not include the response type length nor the length itself, i.e. only the length of the payload */ + local_irq_save(flags); + type_length = gator_response_type ? 1 : 0; + commit = per_cpu(gator_buffer_commit, cpu)[buftype]; + length = per_cpu(gator_buffer_write, cpu)[buftype] - commit; + if (length < 0) + length += gator_buffer_size[buftype]; + length = length - type_length - sizeof(s32); + + if (length <= FRAME_HEADER_SIZE) { + /* Nothing to write, only the frame header is present */ + local_irq_restore(flags); + return; + } + + for (byte = 0; byte < sizeof(s32); byte++) + per_cpu(gator_buffer, cpu)[buftype][(commit + type_length + byte) & gator_buffer_mask[buftype]] = (length >> byte * 8) & 0xFF; + + per_cpu(gator_buffer_commit, cpu)[buftype] = per_cpu(gator_buffer_write, cpu)[buftype]; + + if (gator_live_rate > 0) { + while (time > per_cpu(gator_buffer_commit_time, cpu)) + per_cpu(gator_buffer_commit_time, cpu) += gator_live_rate; + } + + marshal_frame(cpu, buftype); + local_irq_restore(flags); + + /* had to delay scheduling work as attempting to schedule work during the context switch is illegal in kernel versions 3.5 and greater */ + if (per_cpu(in_scheduler_context, cpu)) { +#ifndef CONFIG_PREEMPT_RT_FULL + /* mod_timer can not be used in interrupt context in RT-Preempt full */ + mod_timer(&gator_buffer_wake_up_timer, jiffies + 1); +#endif + } else { + up(&gator_buffer_wake_sem); + } +} + +static void buffer_check(int cpu, int buftype, u64 time) +{ + int filled = per_cpu(gator_buffer_write, cpu)[buftype] - per_cpu(gator_buffer_commit, cpu)[buftype]; + + if (filled < 0) + filled += gator_buffer_size[buftype]; + if (filled >= ((gator_buffer_size[buftype] * 3) / 4)) + gator_commit_buffer(cpu, buftype, time); +} diff --git a/drivers/gator/gator_buffer_write.c b/drivers/gator/gator_buffer_write.c new file mode 100644 index 000000000000..654ec606cfad --- /dev/null +++ b/drivers/gator/gator_buffer_write.c @@ -0,0 +1,83 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +static void gator_buffer_write_packed_int(int cpu, int buftype, int x) +{ + uint32_t write = per_cpu(gator_buffer_write, cpu)[buftype]; + uint32_t mask = gator_buffer_mask[buftype]; + char *buffer = per_cpu(gator_buffer, cpu)[buftype]; + int packedBytes = 0; + int more = true; + + while (more) { + /* low order 7 bits of x */ + char b = x & 0x7f; + + x >>= 7; + + if ((x == 0 && (b & 0x40) == 0) || (x == -1 && (b & 0x40) != 0)) + more = false; + else + b |= 0x80; + + buffer[(write + packedBytes) & mask] = b; + packedBytes++; + } + + per_cpu(gator_buffer_write, cpu)[buftype] = (write + packedBytes) & mask; +} + +static void gator_buffer_write_packed_int64(int cpu, int buftype, long long x) +{ + uint32_t write = per_cpu(gator_buffer_write, cpu)[buftype]; + uint32_t mask = gator_buffer_mask[buftype]; + char *buffer = per_cpu(gator_buffer, cpu)[buftype]; + int packedBytes = 0; + int more = true; + + while (more) { + /* low order 7 bits of x */ + char b = x & 0x7f; + + x >>= 7; + + if ((x == 0 && (b & 0x40) == 0) || (x == -1 && (b & 0x40) != 0)) + more = false; + else + b |= 0x80; + + buffer[(write + packedBytes) & mask] = b; + packedBytes++; + } + + per_cpu(gator_buffer_write, cpu)[buftype] = (write + packedBytes) & mask; +} + +static void gator_buffer_write_bytes(int cpu, int buftype, const char *x, int len) +{ + int i; + u32 write = per_cpu(gator_buffer_write, cpu)[buftype]; + u32 mask = gator_buffer_mask[buftype]; + char *buffer = per_cpu(gator_buffer, cpu)[buftype]; + + for (i = 0; i < len; i++) { + buffer[write] = x[i]; + write = (write + 1) & mask; + } + + per_cpu(gator_buffer_write, cpu)[buftype] = write; +} + +static void gator_buffer_write_string(int cpu, int buftype, const char *x) +{ + int len = strlen(x); + + gator_buffer_write_packed_int(cpu, buftype, len); + gator_buffer_write_bytes(cpu, buftype, x, len); +} diff --git a/drivers/gator/gator_cookies.c b/drivers/gator/gator_cookies.c new file mode 100644 index 000000000000..c43cce815226 --- /dev/null +++ b/drivers/gator/gator_cookies.c @@ -0,0 +1,446 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +/* must be power of 2 */ +#define COOKIEMAP_ENTRIES 1024 +/* must be a power of 2 - 512/4 = 128 entries */ +#define TRANSLATE_BUFFER_SIZE 512 +#define TRANSLATE_TEXT_SIZE 256 +#define MAX_COLLISIONS 2 + +static uint32_t *gator_crc32_table; +static unsigned int translate_buffer_mask; + +struct cookie_args { + struct task_struct *task; + const char *text; +}; + +static DEFINE_PER_CPU(char *, translate_text); +static DEFINE_PER_CPU(uint32_t, cookie_next_key); +static DEFINE_PER_CPU(uint64_t *, cookie_keys); +static DEFINE_PER_CPU(uint32_t *, cookie_values); +static DEFINE_PER_CPU(int, translate_buffer_read); +static DEFINE_PER_CPU(int, translate_buffer_write); +static DEFINE_PER_CPU(struct cookie_args *, translate_buffer); + +static uint32_t get_cookie(int cpu, struct task_struct *task, const char *text, bool from_wq); +static void wq_cookie_handler(struct work_struct *unused); +static DECLARE_WORK(cookie_work, wq_cookie_handler); +static struct timer_list app_process_wake_up_timer; +static void app_process_wake_up_handler(unsigned long unused_data); + +static uint32_t cookiemap_code(uint64_t value64) +{ + uint32_t value = (uint32_t)((value64 >> 32) + value64); + uint32_t cookiecode = (value >> 24) & 0xff; + + cookiecode = cookiecode * 31 + ((value >> 16) & 0xff); + cookiecode = cookiecode * 31 + ((value >> 8) & 0xff); + cookiecode = cookiecode * 31 + ((value >> 0) & 0xff); + cookiecode &= (COOKIEMAP_ENTRIES - 1); + return cookiecode * MAX_COLLISIONS; +} + +static uint32_t gator_chksum_crc32(const char *data) +{ + register unsigned long crc; + const unsigned char *block = data; + int i, length = strlen(data); + + crc = 0xFFFFFFFF; + for (i = 0; i < length; i++) + crc = ((crc >> 8) & 0x00FFFFFF) ^ gator_crc32_table[(crc ^ *block++) & 0xFF]; + + return (crc ^ 0xFFFFFFFF); +} + +/* + * Exists + * Pre: [0][1][v][3]..[n-1] + * Post: [v][0][1][3]..[n-1] + */ +static uint32_t cookiemap_exists(uint64_t key) +{ + unsigned long x, flags, retval = 0; + int cpu = get_physical_cpu(); + uint32_t cookiecode = cookiemap_code(key); + uint64_t *keys = &(per_cpu(cookie_keys, cpu)[cookiecode]); + uint32_t *values = &(per_cpu(cookie_values, cpu)[cookiecode]); + + /* Can be called from interrupt handler or from work queue */ + local_irq_save(flags); + for (x = 0; x < MAX_COLLISIONS; x++) { + if (keys[x] == key) { + uint32_t value = values[x]; + + for (; x > 0; x--) { + keys[x] = keys[x - 1]; + values[x] = values[x - 1]; + } + keys[0] = key; + values[0] = value; + retval = value; + break; + } + } + local_irq_restore(flags); + + return retval; +} + +/* + * Add + * Pre: [0][1][2][3]..[n-1] + * Post: [v][0][1][2]..[n-2] + */ +static void cookiemap_add(uint64_t key, uint32_t value) +{ + int cpu = get_physical_cpu(); + int cookiecode = cookiemap_code(key); + uint64_t *keys = &(per_cpu(cookie_keys, cpu)[cookiecode]); + uint32_t *values = &(per_cpu(cookie_values, cpu)[cookiecode]); + int x; + + for (x = MAX_COLLISIONS - 1; x > 0; x--) { + keys[x] = keys[x - 1]; + values[x] = values[x - 1]; + } + keys[0] = key; + values[0] = value; +} + +#ifndef CONFIG_PREEMPT_RT_FULL +static void translate_buffer_write_args(int cpu, struct task_struct *task, const char *text) +{ + unsigned long flags; + int write; + int next_write; + struct cookie_args *args; + + local_irq_save(flags); + + write = per_cpu(translate_buffer_write, cpu); + next_write = (write + 1) & translate_buffer_mask; + + /* At least one entry must always remain available as when read == write, the queue is empty not full */ + if (next_write != per_cpu(translate_buffer_read, cpu)) { + args = &per_cpu(translate_buffer, cpu)[write]; + args->task = task; + args->text = text; +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 39) + get_task_struct(task); +#endif + per_cpu(translate_buffer_write, cpu) = next_write; + } + + local_irq_restore(flags); +} +#endif + +static void translate_buffer_read_args(int cpu, struct cookie_args *args) +{ + unsigned long flags; + int read; + + local_irq_save(flags); + + read = per_cpu(translate_buffer_read, cpu); + *args = per_cpu(translate_buffer, cpu)[read]; + per_cpu(translate_buffer_read, cpu) = (read + 1) & translate_buffer_mask; + + local_irq_restore(flags); +} + +static void wq_cookie_handler(struct work_struct *unused) +{ + struct cookie_args args; + int cpu = get_physical_cpu(), cookie; + + mutex_lock(&start_mutex); + + if (gator_started != 0) { + while (per_cpu(translate_buffer_read, cpu) != per_cpu(translate_buffer_write, cpu)) { + translate_buffer_read_args(cpu, &args); + cookie = get_cookie(cpu, args.task, args.text, true); + marshal_link(cookie, args.task->tgid, args.task->pid); +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 39) + put_task_struct(args.task); +#endif + } + } + + mutex_unlock(&start_mutex); +} + +static void app_process_wake_up_handler(unsigned long unused_data) +{ + /* had to delay scheduling work as attempting to schedule work during the context switch is illegal in kernel versions 3.5 and greater */ + schedule_work(&cookie_work); +} + +/* Retrieve full name from proc/pid/cmdline for java processes on Android */ +static int translate_app_process(const char **text, int cpu, struct task_struct *task, bool from_wq) +{ + void *maddr; + unsigned int len; + unsigned long addr; + struct mm_struct *mm; + struct page *page = NULL; + struct vm_area_struct *page_vma; + int bytes, offset, retval = 0; + char *buf = per_cpu(translate_text, cpu); + +#ifndef CONFIG_PREEMPT_RT_FULL + /* Push work into a work queue if in atomic context as the kernel + * functions below might sleep. Rely on the in_interrupt variable + * rather than in_irq() or in_interrupt() kernel functions, as the + * value of these functions seems inconsistent during a context + * switch between android/linux versions + */ + if (!from_wq) { + /* Check if already in buffer */ + int pos = per_cpu(translate_buffer_read, cpu); + + while (pos != per_cpu(translate_buffer_write, cpu)) { + if (per_cpu(translate_buffer, cpu)[pos].task == task) + goto out; + pos = (pos + 1) & translate_buffer_mask; + } + + translate_buffer_write_args(cpu, task, *text); + + /* Not safe to call in RT-Preempt full in schedule switch context */ + mod_timer(&app_process_wake_up_timer, jiffies + 1); + goto out; + } +#endif + + mm = get_task_mm(task); + if (!mm) + goto out; + if (!mm->arg_end) + goto outmm; + addr = mm->arg_start; + len = mm->arg_end - mm->arg_start; + + if (len > TRANSLATE_TEXT_SIZE) + len = TRANSLATE_TEXT_SIZE; + + down_read(&mm->mmap_sem); + while (len) { + if (get_user_pages(task, mm, addr, 1, 0, 1, &page, &page_vma) <= 0) + goto outsem; + + maddr = kmap(page); + offset = addr & (PAGE_SIZE - 1); + bytes = len; + if (bytes > PAGE_SIZE - offset) + bytes = PAGE_SIZE - offset; + + copy_from_user_page(page_vma, page, addr, buf, maddr + offset, bytes); + + /* release page allocated by get_user_pages() */ + kunmap(page); + page_cache_release(page); + + len -= bytes; + buf += bytes; + addr += bytes; + + *text = per_cpu(translate_text, cpu); + retval = 1; + } + + /* On app_process startup, /proc/pid/cmdline is initially "zygote" then "<pre-initialized>" but changes after an initial startup period */ + if (strcmp(*text, "zygote") == 0 || strcmp(*text, "<pre-initialized>") == 0) + retval = 0; + +outsem: + up_read(&mm->mmap_sem); +outmm: + mmput(mm); +out: + return retval; +} + +static const char APP_PROCESS[] = "app_process"; + +static uint32_t get_cookie(int cpu, struct task_struct *task, const char *text, bool from_wq) +{ + unsigned long flags, cookie; + uint64_t key; + + key = gator_chksum_crc32(text); + key = (key << 32) | (uint32_t)task->tgid; + + cookie = cookiemap_exists(key); + if (cookie) + return cookie; + + /* On 64-bit android app_process can be app_process32 or app_process64 */ + if (strncmp(text, APP_PROCESS, sizeof(APP_PROCESS) - 1) == 0) { + if (!translate_app_process(&text, cpu, task, from_wq)) + return UNRESOLVED_COOKIE; + } + + /* Can be called from interrupt handler or from work queue or from scheduler trace */ + local_irq_save(flags); + + cookie = UNRESOLVED_COOKIE; + if (marshal_cookie_header(text)) { + cookie = per_cpu(cookie_next_key, cpu) += nr_cpu_ids; + cookiemap_add(key, cookie); + marshal_cookie(cookie, text); + } + + local_irq_restore(flags); + + return cookie; +} + +static int get_exec_cookie(int cpu, struct task_struct *task) +{ + struct mm_struct *mm = task->mm; + const char *text; + + /* kernel threads have no address space */ + if (!mm) + return NO_COOKIE; + + if (task && task->mm && task->mm->exe_file) { + text = task->mm->exe_file->f_path.dentry->d_name.name; + return get_cookie(cpu, task, text, false); + } + + return UNRESOLVED_COOKIE; +} + +static unsigned long get_address_cookie(int cpu, struct task_struct *task, unsigned long addr, off_t *offset) +{ + unsigned long cookie = NO_COOKIE; + struct mm_struct *mm = task->mm; + struct vm_area_struct *vma; + const char *text; + + if (!mm) + return cookie; + + for (vma = find_vma(mm, addr); vma; vma = vma->vm_next) { + if (addr < vma->vm_start || addr >= vma->vm_end) + continue; + + if (vma->vm_file) { + text = vma->vm_file->f_path.dentry->d_name.name; + cookie = get_cookie(cpu, task, text, false); + *offset = (vma->vm_pgoff << PAGE_SHIFT) + addr - vma->vm_start; + } else { + /* must be an anonymous map */ + *offset = addr; + } + + break; + } + + if (!vma) + cookie = UNRESOLVED_COOKIE; + + return cookie; +} + +static int cookies_initialize(void) +{ + uint32_t crc, poly; + int i, j, cpu, size, err = 0; + + translate_buffer_mask = TRANSLATE_BUFFER_SIZE / sizeof(per_cpu(translate_buffer, 0)[0]) - 1; + + for_each_present_cpu(cpu) { + per_cpu(cookie_next_key, cpu) = nr_cpu_ids + cpu; + + size = COOKIEMAP_ENTRIES * MAX_COLLISIONS * sizeof(uint64_t); + per_cpu(cookie_keys, cpu) = kmalloc(size, GFP_KERNEL); + if (!per_cpu(cookie_keys, cpu)) { + err = -ENOMEM; + goto cookie_setup_error; + } + memset(per_cpu(cookie_keys, cpu), 0, size); + + size = COOKIEMAP_ENTRIES * MAX_COLLISIONS * sizeof(uint32_t); + per_cpu(cookie_values, cpu) = kmalloc(size, GFP_KERNEL); + if (!per_cpu(cookie_values, cpu)) { + err = -ENOMEM; + goto cookie_setup_error; + } + memset(per_cpu(cookie_values, cpu), 0, size); + + per_cpu(translate_buffer, cpu) = kmalloc(TRANSLATE_BUFFER_SIZE, GFP_KERNEL); + if (!per_cpu(translate_buffer, cpu)) { + err = -ENOMEM; + goto cookie_setup_error; + } + + per_cpu(translate_buffer_write, cpu) = 0; + per_cpu(translate_buffer_read, cpu) = 0; + + per_cpu(translate_text, cpu) = kmalloc(TRANSLATE_TEXT_SIZE, GFP_KERNEL); + if (!per_cpu(translate_text, cpu)) { + err = -ENOMEM; + goto cookie_setup_error; + } + } + + /* build CRC32 table */ + poly = 0x04c11db7; + gator_crc32_table = kmalloc(256 * sizeof(*gator_crc32_table), GFP_KERNEL); + if (!gator_crc32_table) { + err = -ENOMEM; + goto cookie_setup_error; + } + for (i = 0; i < 256; i++) { + crc = i; + for (j = 8; j > 0; j--) { + if (crc & 1) + crc = (crc >> 1) ^ poly; + else + crc >>= 1; + } + gator_crc32_table[i] = crc; + } + + setup_timer(&app_process_wake_up_timer, app_process_wake_up_handler, 0); + +cookie_setup_error: + return err; +} + +static void cookies_release(void) +{ + int cpu; + + for_each_present_cpu(cpu) { + kfree(per_cpu(cookie_keys, cpu)); + per_cpu(cookie_keys, cpu) = NULL; + + kfree(per_cpu(cookie_values, cpu)); + per_cpu(cookie_values, cpu) = NULL; + + kfree(per_cpu(translate_buffer, cpu)); + per_cpu(translate_buffer, cpu) = NULL; + per_cpu(translate_buffer_read, cpu) = 0; + per_cpu(translate_buffer_write, cpu) = 0; + + kfree(per_cpu(translate_text, cpu)); + per_cpu(translate_text, cpu) = NULL; + } + + del_timer_sync(&app_process_wake_up_timer); + kfree(gator_crc32_table); + gator_crc32_table = NULL; +} diff --git a/drivers/gator/gator_events_armv6.c b/drivers/gator/gator_events_armv6.c new file mode 100644 index 000000000000..a157a0013302 --- /dev/null +++ b/drivers/gator/gator_events_armv6.c @@ -0,0 +1,234 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include "gator.h" + +/* gator_events_perf_pmu.c is used if perf is supported */ +#if GATOR_NO_PERF_SUPPORT + +static const char *pmnc_name; + +/* + * Per-CPU PMCR + */ +#define PMCR_E (1 << 0) /* Enable */ +#define PMCR_P (1 << 1) /* Count reset */ +#define PMCR_C (1 << 2) /* Cycle counter reset */ +#define PMCR_OFL_PMN0 (1 << 8) /* Count reg 0 overflow */ +#define PMCR_OFL_PMN1 (1 << 9) /* Count reg 1 overflow */ +#define PMCR_OFL_CCNT (1 << 10) /* Cycle counter overflow */ + +#define PMN0 0 +#define PMN1 1 +#define CCNT 2 +#define CNTMAX (CCNT+1) + +static int pmnc_counters; +static unsigned long pmnc_enabled[CNTMAX]; +static unsigned long pmnc_event[CNTMAX]; +static unsigned long pmnc_key[CNTMAX]; + +static DEFINE_PER_CPU(int[CNTMAX * 2], perfCnt); + +static inline void armv6_pmnc_write(u32 val) +{ + /* upper 4bits and 7, 11 are write-as-0 */ + val &= 0x0ffff77f; + asm volatile("mcr p15, 0, %0, c15, c12, 0" : : "r" (val)); +} + +static inline u32 armv6_pmnc_read(void) +{ + u32 val; + + asm volatile("mrc p15, 0, %0, c15, c12, 0" : "=r" (val)); + return val; +} + +static void armv6_pmnc_reset_counter(unsigned int cnt) +{ + u32 val = 0; + + switch (cnt) { + case CCNT: + asm volatile("mcr p15, 0, %0, c15, c12, 1" : : "r" (val)); + break; + case PMN0: + asm volatile("mcr p15, 0, %0, c15, c12, 2" : : "r" (val)); + break; + case PMN1: + asm volatile("mcr p15, 0, %0, c15, c12, 3" : : "r" (val)); + break; + } +} + +int gator_events_armv6_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + int i; + + pmnc_counters = 3; + + for (i = PMN0; i <= CCNT; i++) { + char buf[40]; + + if (i == CCNT) + snprintf(buf, sizeof(buf), "ARM_%s_ccnt", pmnc_name); + else + snprintf(buf, sizeof(buf), "ARM_%s_cnt%d", pmnc_name, i); + dir = gatorfs_mkdir(sb, root, buf); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &pmnc_enabled[i]); + gatorfs_create_ro_ulong(sb, dir, "key", &pmnc_key[i]); + if (i != CCNT) + gatorfs_create_ulong(sb, dir, "event", &pmnc_event[i]); + } + + return 0; +} + +static int gator_events_armv6_online(int **buffer, bool migrate) +{ + unsigned int cnt, len = 0, cpu = smp_processor_id(); + u32 pmnc; + + if (armv6_pmnc_read() & PMCR_E) + armv6_pmnc_write(armv6_pmnc_read() & ~PMCR_E); + + /* initialize PMNC, reset overflow, D bit, C bit and P bit. */ + armv6_pmnc_write(PMCR_OFL_PMN0 | PMCR_OFL_PMN1 | PMCR_OFL_CCNT | + PMCR_C | PMCR_P); + + /* configure control register */ + for (pmnc = 0, cnt = PMN0; cnt <= CCNT; cnt++) { + unsigned long event; + + if (!pmnc_enabled[cnt]) + continue; + + event = pmnc_event[cnt] & 255; + + /* Set event (if destined for PMNx counters) */ + if (cnt == PMN0) + pmnc |= event << 20; + else if (cnt == PMN1) + pmnc |= event << 12; + + /* Reset counter */ + armv6_pmnc_reset_counter(cnt); + } + armv6_pmnc_write(pmnc | PMCR_E); + + /* return zero values, no need to read as the counters were just reset */ + for (cnt = PMN0; cnt <= CCNT; cnt++) { + if (pmnc_enabled[cnt]) { + per_cpu(perfCnt, cpu)[len++] = pmnc_key[cnt]; + per_cpu(perfCnt, cpu)[len++] = 0; + } + } + + if (buffer) + *buffer = per_cpu(perfCnt, cpu); + + return len; +} + +static int gator_events_armv6_offline(int **buffer, bool migrate) +{ + unsigned int cnt; + + armv6_pmnc_write(armv6_pmnc_read() & ~PMCR_E); + for (cnt = PMN0; cnt <= CCNT; cnt++) + armv6_pmnc_reset_counter(cnt); + + return 0; +} + +static void gator_events_armv6_stop(void) +{ + unsigned int cnt; + + for (cnt = PMN0; cnt <= CCNT; cnt++) { + pmnc_enabled[cnt] = 0; + pmnc_event[cnt] = 0; + } +} + +static int gator_events_armv6_read(int **buffer, bool sched_switch) +{ + int cnt, len = 0; + int cpu = smp_processor_id(); + + /* a context switch may occur before the online hotplug event, thus need to check that the pmu is enabled */ + if (!(armv6_pmnc_read() & PMCR_E)) + return 0; + + for (cnt = PMN0; cnt <= CCNT; cnt++) { + if (pmnc_enabled[cnt]) { + u32 value = 0; + + switch (cnt) { + case CCNT: + asm volatile("mrc p15, 0, %0, c15, c12, 1" : "=r" (value)); + break; + case PMN0: + asm volatile("mrc p15, 0, %0, c15, c12, 2" : "=r" (value)); + break; + case PMN1: + asm volatile("mrc p15, 0, %0, c15, c12, 3" : "=r" (value)); + break; + } + armv6_pmnc_reset_counter(cnt); + + per_cpu(perfCnt, cpu)[len++] = pmnc_key[cnt]; + per_cpu(perfCnt, cpu)[len++] = value; + } + } + + if (buffer) + *buffer = per_cpu(perfCnt, cpu); + + return len; +} + +static struct gator_interface gator_events_armv6_interface = { + .create_files = gator_events_armv6_create_files, + .stop = gator_events_armv6_stop, + .online = gator_events_armv6_online, + .offline = gator_events_armv6_offline, + .read = gator_events_armv6_read, +}; + +int gator_events_armv6_init(void) +{ + unsigned int cnt; + + switch (gator_cpuid()) { + case ARM1136: + case ARM1156: + case ARM1176: + pmnc_name = "ARM11"; + break; + case ARM11MPCORE: + pmnc_name = "ARM11MPCore"; + break; + default: + return -1; + } + + for (cnt = PMN0; cnt <= CCNT; cnt++) { + pmnc_enabled[cnt] = 0; + pmnc_event[cnt] = 0; + pmnc_key[cnt] = gator_events_get_key(); + } + + return gator_events_install(&gator_events_armv6_interface); +} + +#endif diff --git a/drivers/gator/gator_events_armv7.c b/drivers/gator/gator_events_armv7.c new file mode 100644 index 000000000000..09c94220114c --- /dev/null +++ b/drivers/gator/gator_events_armv7.c @@ -0,0 +1,314 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +/* Disabling interrupts + * Many of the functions below disable interrupts via local_irq_save(). This disabling of interrupts is done to prevent any race conditions + * between multiple entities (e.g. hrtimer interrupts and event based interrupts) calling the same functions. As accessing the pmu involves + * several steps (disable, select, read, enable), these steps must be performed atomically. Normal synchronization routines cannot be used + * as these functions are being called from interrupt context. + */ + +#include "gator.h" + +/* gator_events_perf_pmu.c is used if perf is supported */ +#if GATOR_NO_PERF_SUPPORT + +/* Per-CPU PMNC: config reg */ +#define PMNC_E (1 << 0) /* Enable all counters */ +#define PMNC_P (1 << 1) /* Reset all counters */ +#define PMNC_C (1 << 2) /* Cycle counter reset */ +#define PMNC_MASK 0x3f /* Mask for writable bits */ + +/* ccnt reg */ +#define CCNT_REG (1 << 31) + +#define CCNT 0 +#define CNT0 1 +#define CNTMAX (6+1) + +static const char *pmnc_name; +static int pmnc_counters; + +static unsigned long pmnc_enabled[CNTMAX]; +static unsigned long pmnc_event[CNTMAX]; +static unsigned long pmnc_key[CNTMAX]; + +static DEFINE_PER_CPU(int[CNTMAX * 2], perfCnt); + +inline void armv7_pmnc_write(u32 val) +{ + val &= PMNC_MASK; + asm volatile("mcr p15, 0, %0, c9, c12, 0" : : "r" (val)); +} + +inline u32 armv7_pmnc_read(void) +{ + u32 val; + + asm volatile("mrc p15, 0, %0, c9, c12, 0" : "=r" (val)); + return val; +} + +inline u32 armv7_ccnt_read(u32 reset_value) +{ + unsigned long flags; + u32 newval = -reset_value; + u32 den = CCNT_REG; + u32 val; + + local_irq_save(flags); + asm volatile("mcr p15, 0, %0, c9, c12, 2" : : "r" (den)); /* disable */ + asm volatile("mrc p15, 0, %0, c9, c13, 0" : "=r" (val)); /* read */ + asm volatile("mcr p15, 0, %0, c9, c13, 0" : : "r" (newval)); /* new value */ + asm volatile("mcr p15, 0, %0, c9, c12, 1" : : "r" (den)); /* enable */ + local_irq_restore(flags); + + return val; +} + +inline u32 armv7_cntn_read(unsigned int cnt, u32 reset_value) +{ + unsigned long flags; + u32 newval = -reset_value; + u32 sel = (cnt - CNT0); + u32 den = 1 << sel; + u32 oldval; + + local_irq_save(flags); + asm volatile("mcr p15, 0, %0, c9, c12, 2" : : "r" (den)); /* disable */ + asm volatile("mcr p15, 0, %0, c9, c12, 5" : : "r" (sel)); /* select */ + asm volatile("mrc p15, 0, %0, c9, c13, 2" : "=r" (oldval)); /* read */ + asm volatile("mcr p15, 0, %0, c9, c13, 2" : : "r" (newval)); /* new value */ + asm volatile("mcr p15, 0, %0, c9, c12, 1" : : "r" (den)); /* enable */ + local_irq_restore(flags); + + return oldval; +} + +static inline void armv7_pmnc_disable_interrupt(unsigned int cnt) +{ + u32 val = cnt ? (1 << (cnt - CNT0)) : (1 << 31); + + asm volatile("mcr p15, 0, %0, c9, c14, 2" : : "r" (val)); +} + +inline u32 armv7_pmnc_reset_interrupt(void) +{ + /* Get and reset overflow status flags */ + u32 flags; + + asm volatile("mrc p15, 0, %0, c9, c12, 3" : "=r" (flags)); + flags &= 0x8000003f; + asm volatile("mcr p15, 0, %0, c9, c12, 3" : : "r" (flags)); + return flags; +} + +static inline u32 armv7_pmnc_enable_counter(unsigned int cnt) +{ + u32 val = cnt ? (1 << (cnt - CNT0)) : CCNT_REG; + + asm volatile("mcr p15, 0, %0, c9, c12, 1" : : "r" (val)); + return cnt; +} + +static inline u32 armv7_pmnc_disable_counter(unsigned int cnt) +{ + u32 val = cnt ? (1 << (cnt - CNT0)) : CCNT_REG; + + asm volatile("mcr p15, 0, %0, c9, c12, 2" : : "r" (val)); + return cnt; +} + +static inline int armv7_pmnc_select_counter(unsigned int cnt) +{ + u32 val = (cnt - CNT0); + + asm volatile("mcr p15, 0, %0, c9, c12, 5" : : "r" (val)); + return cnt; +} + +static inline void armv7_pmnc_write_evtsel(unsigned int cnt, u32 val) +{ + if (armv7_pmnc_select_counter(cnt) == cnt) + asm volatile("mcr p15, 0, %0, c9, c13, 1" : : "r" (val)); +} + +static int gator_events_armv7_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + int i; + + for (i = 0; i < pmnc_counters; i++) { + char buf[40]; + + if (i == 0) + snprintf(buf, sizeof(buf), "%s_ccnt", pmnc_name); + else + snprintf(buf, sizeof(buf), "%s_cnt%d", pmnc_name, i - 1); + dir = gatorfs_mkdir(sb, root, buf); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &pmnc_enabled[i]); + gatorfs_create_ro_ulong(sb, dir, "key", &pmnc_key[i]); + if (i > 0) + gatorfs_create_ulong(sb, dir, "event", &pmnc_event[i]); + } + + return 0; +} + +static int gator_events_armv7_online(int **buffer, bool migrate) +{ + unsigned int cnt, len = 0, cpu = smp_processor_id(); + + if (armv7_pmnc_read() & PMNC_E) + armv7_pmnc_write(armv7_pmnc_read() & ~PMNC_E); + + /* Initialize & Reset PMNC: C bit and P bit */ + armv7_pmnc_write(PMNC_P | PMNC_C); + + /* Reset overflow flags */ + armv7_pmnc_reset_interrupt(); + + for (cnt = CCNT; cnt < CNTMAX; cnt++) { + unsigned long event; + + if (!pmnc_enabled[cnt]) + continue; + + /* Disable counter */ + armv7_pmnc_disable_counter(cnt); + + event = pmnc_event[cnt] & 255; + + /* Set event (if destined for PMNx counters), we don't need to set the event if it's a cycle count */ + if (cnt != CCNT) + armv7_pmnc_write_evtsel(cnt, event); + + armv7_pmnc_disable_interrupt(cnt); + + /* Reset counter */ + cnt ? armv7_cntn_read(cnt, 0) : armv7_ccnt_read(0); + + /* Enable counter */ + armv7_pmnc_enable_counter(cnt); + } + + /* enable */ + armv7_pmnc_write(armv7_pmnc_read() | PMNC_E); + + /* return zero values, no need to read as the counters were just reset */ + for (cnt = 0; cnt < pmnc_counters; cnt++) { + if (pmnc_enabled[cnt]) { + per_cpu(perfCnt, cpu)[len++] = pmnc_key[cnt]; + per_cpu(perfCnt, cpu)[len++] = 0; + } + } + + if (buffer) + *buffer = per_cpu(perfCnt, cpu); + + return len; +} + +static int gator_events_armv7_offline(int **buffer, bool migrate) +{ + /* disable all counters, including PMCCNTR; overflow IRQs will not be signaled */ + armv7_pmnc_write(armv7_pmnc_read() & ~PMNC_E); + + return 0; +} + +static void gator_events_armv7_stop(void) +{ + unsigned int cnt; + + for (cnt = CCNT; cnt < CNTMAX; cnt++) { + pmnc_enabled[cnt] = 0; + pmnc_event[cnt] = 0; + } +} + +static int gator_events_armv7_read(int **buffer, bool sched_switch) +{ + int cnt, len = 0; + int cpu = smp_processor_id(); + + /* a context switch may occur before the online hotplug event, thus need to check that the pmu is enabled */ + if (!(armv7_pmnc_read() & PMNC_E)) + return 0; + + for (cnt = 0; cnt < pmnc_counters; cnt++) { + if (pmnc_enabled[cnt]) { + int value; + + if (cnt == CCNT) + value = armv7_ccnt_read(0); + else + value = armv7_cntn_read(cnt, 0); + per_cpu(perfCnt, cpu)[len++] = pmnc_key[cnt]; + per_cpu(perfCnt, cpu)[len++] = value; + } + } + + if (buffer) + *buffer = per_cpu(perfCnt, cpu); + + return len; +} + +static struct gator_interface gator_events_armv7_interface = { + .create_files = gator_events_armv7_create_files, + .stop = gator_events_armv7_stop, + .online = gator_events_armv7_online, + .offline = gator_events_armv7_offline, + .read = gator_events_armv7_read, +}; + +int gator_events_armv7_init(void) +{ + unsigned int cnt; + + switch (gator_cpuid()) { + case CORTEX_A5: + pmnc_name = "ARMv7_Cortex_A5"; + pmnc_counters = 2; + break; + case CORTEX_A7: + pmnc_name = "ARMv7_Cortex_A7"; + pmnc_counters = 4; + break; + case CORTEX_A8: + pmnc_name = "ARMv7_Cortex_A8"; + pmnc_counters = 4; + break; + case CORTEX_A9: + pmnc_name = "ARMv7_Cortex_A9"; + pmnc_counters = 6; + break; + case CORTEX_A15: + pmnc_name = "ARMv7_Cortex_A15"; + pmnc_counters = 6; + break; + /* ARM Cortex A17 is not supported by version of Linux before 3.0 */ + default: + return -1; + } + + pmnc_counters++; /* CNT[n] + CCNT */ + + for (cnt = CCNT; cnt < CNTMAX; cnt++) { + pmnc_enabled[cnt] = 0; + pmnc_event[cnt] = 0; + pmnc_key[cnt] = gator_events_get_key(); + } + + return gator_events_install(&gator_events_armv7_interface); +} + +#endif diff --git a/drivers/gator/gator_events_block.c b/drivers/gator/gator_events_block.c new file mode 100644 index 000000000000..a352a54afa02 --- /dev/null +++ b/drivers/gator/gator_events_block.c @@ -0,0 +1,160 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include "gator.h" +#include <trace/events/block.h> + +#define BLOCK_RQ_WR 0 +#define BLOCK_RQ_RD 1 + +#define BLOCK_TOTAL (BLOCK_RQ_RD+1) + +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 36) +#define EVENTWRITE REQ_RW +#else +#define EVENTWRITE REQ_WRITE +#endif + +static ulong block_rq_wr_enabled; +static ulong block_rq_rd_enabled; +static ulong block_rq_wr_key; +static ulong block_rq_rd_key; +static atomic_t blockCnt[BLOCK_TOTAL]; +static int blockGet[BLOCK_TOTAL * 4]; + +/* Tracepoint changed in 3.15 backported to older kernels. The Makefile tries to autodetect the correct value, but if it fails change the #if below */ +#if OLD_BLOCK_RQ_COMPLETE +GATOR_DEFINE_PROBE(block_rq_complete, TP_PROTO(struct request_queue *q, struct request *rq)) +#else +GATOR_DEFINE_PROBE(block_rq_complete, TP_PROTO(struct request_queue *q, struct request *rq, unsigned int nr_bytes)) +#endif +{ + int write; + unsigned int size; + + if (!rq) + return; + + write = rq->cmd_flags & EVENTWRITE; +#if OLD_BLOCK_RQ_COMPLETE + size = rq->resid_len; +#else + size = nr_bytes; +#endif + + if (!size) + return; + + if (write) { + if (block_rq_wr_enabled) + atomic_add(size, &blockCnt[BLOCK_RQ_WR]); + } else { + if (block_rq_rd_enabled) + atomic_add(size, &blockCnt[BLOCK_RQ_RD]); + } +} + +static int gator_events_block_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + + /* block_complete_wr */ + dir = gatorfs_mkdir(sb, root, "Linux_block_rq_wr"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &block_rq_wr_enabled); + gatorfs_create_ro_ulong(sb, dir, "key", &block_rq_wr_key); + + /* block_complete_rd */ + dir = gatorfs_mkdir(sb, root, "Linux_block_rq_rd"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &block_rq_rd_enabled); + gatorfs_create_ro_ulong(sb, dir, "key", &block_rq_rd_key); + + return 0; +} + +static int gator_events_block_start(void) +{ + /* register tracepoints */ + if (block_rq_wr_enabled || block_rq_rd_enabled) + if (GATOR_REGISTER_TRACE(block_rq_complete)) + goto fail_block_rq_exit; + pr_debug("gator: registered block event tracepoints\n"); + + return 0; + + /* unregister tracepoints on error */ +fail_block_rq_exit: + pr_err("gator: block event tracepoints failed to activate, please verify that tracepoints are enabled in the linux kernel\n"); + + return -1; +} + +static void gator_events_block_stop(void) +{ + if (block_rq_wr_enabled || block_rq_rd_enabled) + GATOR_UNREGISTER_TRACE(block_rq_complete); + pr_debug("gator: unregistered block event tracepoints\n"); + + block_rq_wr_enabled = 0; + block_rq_rd_enabled = 0; +} + +static int gator_events_block_read(int **buffer, bool sched_switch) +{ + int len, value, data = 0; + + if (!on_primary_core()) + return 0; + + len = 0; + if (block_rq_wr_enabled && (value = atomic_read(&blockCnt[BLOCK_RQ_WR])) > 0) { + atomic_sub(value, &blockCnt[BLOCK_RQ_WR]); + blockGet[len++] = block_rq_wr_key; + /* Indicates to Streamline that value bytes were written now, not since the last message */ + blockGet[len++] = 0; + blockGet[len++] = block_rq_wr_key; + blockGet[len++] = value; + data += value; + } + if (block_rq_rd_enabled && (value = atomic_read(&blockCnt[BLOCK_RQ_RD])) > 0) { + atomic_sub(value, &blockCnt[BLOCK_RQ_RD]); + blockGet[len++] = block_rq_rd_key; + /* Indicates to Streamline that value bytes were read now, not since the last message */ + blockGet[len++] = 0; + blockGet[len++] = block_rq_rd_key; + blockGet[len++] = value; + data += value; + } + + if (buffer) + *buffer = blockGet; + + return len; +} + +static struct gator_interface gator_events_block_interface = { + .create_files = gator_events_block_create_files, + .start = gator_events_block_start, + .stop = gator_events_block_stop, + .read = gator_events_block_read, +}; + +int gator_events_block_init(void) +{ + block_rq_wr_enabled = 0; + block_rq_rd_enabled = 0; + + block_rq_wr_key = gator_events_get_key(); + block_rq_rd_key = gator_events_get_key(); + + return gator_events_install(&gator_events_block_interface); +} diff --git a/drivers/gator/gator_events_irq.c b/drivers/gator/gator_events_irq.c new file mode 100644 index 000000000000..5221aac581b3 --- /dev/null +++ b/drivers/gator/gator_events_irq.c @@ -0,0 +1,163 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include "gator.h" +#include <trace/events/irq.h> + +#define HARDIRQ 0 +#define SOFTIRQ 1 +#define TOTALIRQ (SOFTIRQ+1) + +static ulong hardirq_enabled; +static ulong softirq_enabled; +static ulong hardirq_key; +static ulong softirq_key; +static DEFINE_PER_CPU(atomic_t[TOTALIRQ], irqCnt); +static DEFINE_PER_CPU(int[TOTALIRQ * 2], irqGet); + +GATOR_DEFINE_PROBE(irq_handler_exit, + TP_PROTO(int irq, struct irqaction *action, int ret)) +{ + atomic_inc(&per_cpu(irqCnt, get_physical_cpu())[HARDIRQ]); +} + +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 37) +GATOR_DEFINE_PROBE(softirq_exit, TP_PROTO(struct softirq_action *h, struct softirq_action *vec)) +#else +GATOR_DEFINE_PROBE(softirq_exit, TP_PROTO(unsigned int vec_nr)) +#endif +{ + atomic_inc(&per_cpu(irqCnt, get_physical_cpu())[SOFTIRQ]); +} + +static int gator_events_irq_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + + /* irq */ + dir = gatorfs_mkdir(sb, root, "Linux_irq_irq"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &hardirq_enabled); + gatorfs_create_ro_ulong(sb, dir, "key", &hardirq_key); + + /* soft irq */ + dir = gatorfs_mkdir(sb, root, "Linux_irq_softirq"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &softirq_enabled); + gatorfs_create_ro_ulong(sb, dir, "key", &softirq_key); + + return 0; +} + +static int gator_events_irq_online(int **buffer, bool migrate) +{ + int len = 0, cpu = get_physical_cpu(); + + /* synchronization with the irq_exit functions is not necessary as the values are being reset */ + if (hardirq_enabled) { + atomic_set(&per_cpu(irqCnt, cpu)[HARDIRQ], 0); + per_cpu(irqGet, cpu)[len++] = hardirq_key; + per_cpu(irqGet, cpu)[len++] = 0; + } + + if (softirq_enabled) { + atomic_set(&per_cpu(irqCnt, cpu)[SOFTIRQ], 0); + per_cpu(irqGet, cpu)[len++] = softirq_key; + per_cpu(irqGet, cpu)[len++] = 0; + } + + if (buffer) + *buffer = per_cpu(irqGet, cpu); + + return len; +} + +static int gator_events_irq_start(void) +{ + /* register tracepoints */ + if (hardirq_enabled) + if (GATOR_REGISTER_TRACE(irq_handler_exit)) + goto fail_hardirq_exit; + if (softirq_enabled) + if (GATOR_REGISTER_TRACE(softirq_exit)) + goto fail_softirq_exit; + pr_debug("gator: registered irq tracepoints\n"); + + return 0; + + /* unregister tracepoints on error */ +fail_softirq_exit: + if (hardirq_enabled) + GATOR_UNREGISTER_TRACE(irq_handler_exit); +fail_hardirq_exit: + pr_err("gator: irq tracepoints failed to activate, please verify that tracepoints are enabled in the linux kernel\n"); + + return -1; +} + +static void gator_events_irq_stop(void) +{ + if (hardirq_enabled) + GATOR_UNREGISTER_TRACE(irq_handler_exit); + if (softirq_enabled) + GATOR_UNREGISTER_TRACE(softirq_exit); + pr_debug("gator: unregistered irq tracepoints\n"); + + hardirq_enabled = 0; + softirq_enabled = 0; +} + +static int gator_events_irq_read(int **buffer, bool sched_switch) +{ + int len, value; + int cpu = get_physical_cpu(); + + len = 0; + if (hardirq_enabled) { + value = atomic_read(&per_cpu(irqCnt, cpu)[HARDIRQ]); + atomic_sub(value, &per_cpu(irqCnt, cpu)[HARDIRQ]); + + per_cpu(irqGet, cpu)[len++] = hardirq_key; + per_cpu(irqGet, cpu)[len++] = value; + } + + if (softirq_enabled) { + value = atomic_read(&per_cpu(irqCnt, cpu)[SOFTIRQ]); + atomic_sub(value, &per_cpu(irqCnt, cpu)[SOFTIRQ]); + + per_cpu(irqGet, cpu)[len++] = softirq_key; + per_cpu(irqGet, cpu)[len++] = value; + } + + if (buffer) + *buffer = per_cpu(irqGet, cpu); + + return len; +} + +static struct gator_interface gator_events_irq_interface = { + .create_files = gator_events_irq_create_files, + .online = gator_events_irq_online, + .start = gator_events_irq_start, + .stop = gator_events_irq_stop, + .read = gator_events_irq_read, +}; + +int gator_events_irq_init(void) +{ + hardirq_key = gator_events_get_key(); + softirq_key = gator_events_get_key(); + + hardirq_enabled = 0; + softirq_enabled = 0; + + return gator_events_install(&gator_events_irq_interface); +} diff --git a/drivers/gator/gator_events_l2c-310.c b/drivers/gator/gator_events_l2c-310.c new file mode 100644 index 000000000000..73aaac32327e --- /dev/null +++ b/drivers/gator/gator_events_l2c-310.c @@ -0,0 +1,208 @@ +/** + * l2c310 (L2 Cache Controller) event counters for gator + * + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include <linux/init.h> +#include <linux/io.h> +#include <linux/module.h> +#if defined(CONFIG_OF) +#include <linux/of.h> +#include <linux/of_address.h> +#endif +#include <asm/hardware/cache-l2x0.h> + +#include "gator.h" + +#define L2C310_COUNTERS_NUM 2 + +static struct { + unsigned long enabled; + unsigned long event; + unsigned long key; +} l2c310_counters[L2C310_COUNTERS_NUM]; + +static int l2c310_buffer[L2C310_COUNTERS_NUM * 2]; + +static void __iomem *l2c310_base; + +static void gator_events_l2c310_reset_counters(void) +{ + u32 val = readl(l2c310_base + L2X0_EVENT_CNT_CTRL); + + val |= ((1 << L2C310_COUNTERS_NUM) - 1) << 1; + + writel(val, l2c310_base + L2X0_EVENT_CNT_CTRL); +} + +static int gator_events_l2c310_create_files(struct super_block *sb, + struct dentry *root) +{ + int i; + + for (i = 0; i < L2C310_COUNTERS_NUM; i++) { + char buf[16]; + struct dentry *dir; + + snprintf(buf, sizeof(buf), "L2C-310_cnt%d", i); + dir = gatorfs_mkdir(sb, root, buf); + if (WARN_ON(!dir)) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", + &l2c310_counters[i].enabled); + gatorfs_create_ulong(sb, dir, "event", + &l2c310_counters[i].event); + gatorfs_create_ro_ulong(sb, dir, "key", + &l2c310_counters[i].key); + } + + return 0; +} + +static int gator_events_l2c310_start(void) +{ + static const unsigned long l2x0_event_cntx_cfg[L2C310_COUNTERS_NUM] = { + L2X0_EVENT_CNT0_CFG, + L2X0_EVENT_CNT1_CFG, + }; + int i; + + /* Counter event sources */ + for (i = 0; i < L2C310_COUNTERS_NUM; i++) + writel((l2c310_counters[i].event & 0xf) << 2, + l2c310_base + l2x0_event_cntx_cfg[i]); + + gator_events_l2c310_reset_counters(); + + /* Event counter enable */ + writel(1, l2c310_base + L2X0_EVENT_CNT_CTRL); + + return 0; +} + +static void gator_events_l2c310_stop(void) +{ + /* Event counter disable */ + writel(0, l2c310_base + L2X0_EVENT_CNT_CTRL); +} + +static int gator_events_l2c310_read(int **buffer, bool sched_switch) +{ + static const unsigned long l2x0_event_cntx_val[L2C310_COUNTERS_NUM] = { + L2X0_EVENT_CNT0_VAL, + L2X0_EVENT_CNT1_VAL, + }; + int i; + int len = 0; + + if (!on_primary_core()) + return 0; + + for (i = 0; i < L2C310_COUNTERS_NUM; i++) { + if (l2c310_counters[i].enabled) { + l2c310_buffer[len++] = l2c310_counters[i].key; + l2c310_buffer[len++] = readl(l2c310_base + + l2x0_event_cntx_val[i]); + } + } + + /* l2c310 counters are saturating, not wrapping in case of overflow */ + gator_events_l2c310_reset_counters(); + + if (buffer) + *buffer = l2c310_buffer; + + return len; +} + +static struct gator_interface gator_events_l2c310_interface = { + .create_files = gator_events_l2c310_create_files, + .start = gator_events_l2c310_start, + .stop = gator_events_l2c310_stop, + .read = gator_events_l2c310_read, +}; + +#define L2C310_ADDR_PROBE (~0) + +MODULE_PARM_DESC(l2c310_addr, "L2C310 physical base address (0 to disable)"); +static unsigned long l2c310_addr = L2C310_ADDR_PROBE; +module_param(l2c310_addr, ulong, 0444); + +static void __iomem *gator_events_l2c310_probe(void) +{ + phys_addr_t variants[] = { +#if defined(CONFIG_ARCH_EXYNOS4) || defined(CONFIG_ARCH_S5PV310) + 0x10502000, +#endif +#if defined(CONFIG_ARCH_OMAP4) + 0x48242000, +#endif +#if defined(CONFIG_ARCH_TEGRA) + 0x50043000, +#endif +#if defined(CONFIG_ARCH_U8500) + 0xa0412000, +#endif +#if defined(CONFIG_ARCH_VEXPRESS) + 0x1e00a000, /* A9x4 core tile (HBI-0191) */ + 0x2c0f0000, /* New memory map tiles */ +#endif + }; + int i; + void __iomem *base; +#if defined(CONFIG_OF) + struct device_node *node = of_find_all_nodes(NULL); + + if (node) { + of_node_put(node); + + node = of_find_compatible_node(NULL, NULL, "arm,pl310-cache"); + base = of_iomap(node, 0); + of_node_put(node); + + return base; + } +#endif + + for (i = 0; i < ARRAY_SIZE(variants); i++) { + base = ioremap(variants[i], SZ_4K); + if (base) { + u32 cache_id = readl(base + L2X0_CACHE_ID); + + if ((cache_id & 0xff0003c0) == 0x410000c0) + return base; + + iounmap(base); + } + } + + return NULL; +} + +int gator_events_l2c310_init(void) +{ + int i; + + if (gator_cpuid() != CORTEX_A5 && gator_cpuid() != CORTEX_A9) + return -1; + + if (l2c310_addr == L2C310_ADDR_PROBE) + l2c310_base = gator_events_l2c310_probe(); + else if (l2c310_addr) + l2c310_base = ioremap(l2c310_addr, SZ_4K); + + if (!l2c310_base) + return -1; + + for (i = 0; i < L2C310_COUNTERS_NUM; i++) { + l2c310_counters[i].enabled = 0; + l2c310_counters[i].key = gator_events_get_key(); + } + + return gator_events_install(&gator_events_l2c310_interface); +} diff --git a/drivers/gator/gator_events_mali_4xx.c b/drivers/gator/gator_events_mali_4xx.c new file mode 100644 index 000000000000..9cf43fe2c29b --- /dev/null +++ b/drivers/gator/gator_events_mali_4xx.c @@ -0,0 +1,622 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include "gator.h" + +#include <linux/module.h> +#include <linux/time.h> +#include <linux/math64.h> + +#include "linux/mali_linux_trace.h" + +#include "gator_events_mali_common.h" +#include "gator_events_mali_4xx.h" + +/* +* There have been four different variants of the comms between gator and Mali depending on driver version: +* # | DDK vsn range | Support | Notes +* +* 1 | (obsolete) | No software counter support | Obsolete patches +* 2 | (obsolete) | Tracepoint called for each separate s/w counter value as it appears | Obsolete patches +* 3 | r3p0-04rel0 - r3p2-01rel2 | Single tracepoint for all s/w counters in a bundle. | +* 4 | r3p2-01rel3 - date | As above but with extensions for MP devices (Mali-450) | At least r4p0-00rel1 +*/ + +#if !defined(GATOR_MALI_INTERFACE_STYLE) +#define GATOR_MALI_INTERFACE_STYLE (4) +#endif + +#if GATOR_MALI_INTERFACE_STYLE == 1 +#error GATOR_MALI_INTERFACE_STYLE 1 is obsolete +#elif GATOR_MALI_INTERFACE_STYLE == 2 +#error GATOR_MALI_INTERFACE_STYLE 2 is obsolete +#elif GATOR_MALI_INTERFACE_STYLE >= 3 +/* Valid GATOR_MALI_INTERFACE_STYLE */ +#else +#error Unknown GATOR_MALI_INTERFACE_STYLE option. +#endif + +#if GATOR_MALI_INTERFACE_STYLE < 4 +#include "mali/mali_mjollnir_profiling_gator_api.h" +#else +#include "mali/mali_utgard_profiling_gator_api.h" +#endif + +/* + * Check that the MALI_SUPPORT define is set to one of the allowable device codes. + */ +#if (MALI_SUPPORT != MALI_4xx) +#error MALI_SUPPORT set to an invalid device code: expecting MALI_4xx +#endif + +static const char mali_name[] = "4xx"; + +/* gatorfs variables for counter enable state, + * the event the counter should count and the + * 'key' (a unique id set by gatord and returned + * by gator.ko) + */ +static unsigned long counter_enabled[NUMBER_OF_EVENTS]; +static unsigned long counter_event[NUMBER_OF_EVENTS]; +static unsigned long counter_key[NUMBER_OF_EVENTS]; + +/* The data we have recorded */ +static u32 counter_data[NUMBER_OF_EVENTS]; +/* The address to sample (or 0 if samples are sent to us) */ +static u32 *counter_address[NUMBER_OF_EVENTS]; + +/* An array used to return the data we recorded + * as key,value pairs hence the *2 + */ +static int counter_dump[NUMBER_OF_EVENTS * 2]; +static int counter_prev[NUMBER_OF_EVENTS]; +static bool prev_set[NUMBER_OF_EVENTS]; + +/* Note whether tracepoints have been registered */ +static int trace_registered; + +/* + * These numbers define the actual numbers of each block type that exist in the system. Initially + * these are set to the maxima defined above; if the driver is capable of being queried (newer + * drivers only) then the values may be revised. + */ +static unsigned int n_vp_cores = MAX_NUM_VP_CORES; +static unsigned int n_l2_cores = MAX_NUM_L2_CACHE_CORES; +static unsigned int n_fp_cores = MAX_NUM_FP_CORES; + +extern struct mali_counter mali_activity[2]; +static const char *const mali_activity_names[] = { + "fragment", + "vertex", +}; + +/** + * Returns non-zero if the given counter ID is an activity counter. + */ +static inline int is_activity_counter(unsigned int event_id) +{ + return (event_id >= FIRST_ACTIVITY_EVENT && + event_id <= LAST_ACTIVITY_EVENT); +} + +/** + * Returns non-zero if the given counter ID is a hardware counter. + */ +static inline int is_hw_counter(unsigned int event_id) +{ + return (event_id >= FIRST_HW_COUNTER && event_id <= LAST_HW_COUNTER); +} + +/* Probe for hardware counter events */ +GATOR_DEFINE_PROBE(mali_hw_counter, TP_PROTO(unsigned int event_id, unsigned int value)) +{ + if (is_hw_counter(event_id)) + counter_data[event_id] = value; +} + +GATOR_DEFINE_PROBE(mali_sw_counters, TP_PROTO(pid_t pid, pid_t tid, void *surface_id, unsigned int *counters)) +{ + u32 i; + + /* Copy over the values for those counters which are enabled. */ + for (i = FIRST_SW_COUNTER; i <= LAST_SW_COUNTER; i++) { + if (counter_enabled[i]) + counter_data[i] = (u32)(counters[i - FIRST_SW_COUNTER]); + } +} + +/** + * Create a single filesystem entry for a specified event. + * @param sb the superblock + * @param root Filesystem root + * @param name The name of the entry to create + * @param event The ID of the event + * @param create_event_item boolean indicating whether to create an 'event' filesystem entry. True to create. + * + * @return 0 if ok, non-zero if the create failed. + */ +static int create_fs_entry(struct super_block *sb, struct dentry *root, const char *name, int event, int create_event_item) +{ + struct dentry *dir; + + dir = gatorfs_mkdir(sb, root, name); + + if (!dir) + return -1; + + if (create_event_item) + gatorfs_create_ulong(sb, dir, "event", &counter_event[event]); + + gatorfs_create_ulong(sb, dir, "enabled", &counter_enabled[event]); + gatorfs_create_ro_ulong(sb, dir, "key", &counter_key[event]); + + return 0; +} + +#if GATOR_MALI_INTERFACE_STYLE > 3 +/* + * Read the version info structure if available + */ +static void initialise_version_info(void) +{ + void (*mali_profiling_get_mali_version_symbol)(struct _mali_profiling_mali_version *values); + + mali_profiling_get_mali_version_symbol = symbol_get(_mali_profiling_get_mali_version); + + if (mali_profiling_get_mali_version_symbol) { + struct _mali_profiling_mali_version version_info; + + pr_debug("gator: mali online _mali_profiling_get_mali_version symbol @ %p\n", + mali_profiling_get_mali_version_symbol); + + /* + * Revise the number of each different core type using information derived from the DDK. + */ + mali_profiling_get_mali_version_symbol(&version_info); + + n_fp_cores = version_info.num_of_fp_cores; + n_vp_cores = version_info.num_of_vp_cores; + n_l2_cores = version_info.num_of_l2_cores; + + /* Release the function - we're done with it. */ + symbol_put(_mali_profiling_get_mali_version); + } else { + pr_err("gator: mali online _mali_profiling_get_mali_version symbol not found\n"); + pr_err("gator: check your Mali DDK version versus the GATOR_MALI_INTERFACE_STYLE setting\n"); + } +} +#endif + +static int create_files(struct super_block *sb, struct dentry *root) +{ + int event; + + char buf[40]; + int core_id; + int counter_number; + + pr_debug("gator: Initialising counters with style = %d\n", GATOR_MALI_INTERFACE_STYLE); + +#if GATOR_MALI_INTERFACE_STYLE > 3 + /* + * Initialise first: this sets up the number of cores available (on compatible DDK versions). + * Ideally this would not need guarding but other parts of the code depend on the interface style being set + * correctly; if it is not then the system can enter an inconsistent state. + */ + initialise_version_info(); +#endif + + mali_activity[0].cores = n_fp_cores; + mali_activity[1].cores = n_vp_cores; + for (event = 0; event < ARRAY_SIZE(mali_activity); event++) { + if (gator_mali_create_file_system(mali_name, mali_activity_names[event], sb, root, &mali_activity[event], NULL) != 0) + return -1; + } + + /* Vertex processor counters */ + for (core_id = 0; core_id < n_vp_cores; core_id++) { + int activity_counter_id = ACTIVITY_VP_0; + + snprintf(buf, sizeof(buf), "ARM_Mali-%s_VP_%d_active", mali_name, core_id); + if (create_fs_entry(sb, root, buf, activity_counter_id, 0) != 0) + return -1; + + for (counter_number = 0; counter_number < 2; counter_number++) { + int counter_id = COUNTER_VP_0_C0 + (2 * core_id) + counter_number; + + snprintf(buf, sizeof(buf), "ARM_Mali-%s_VP_%d_cnt%d", mali_name, core_id, counter_number); + if (create_fs_entry(sb, root, buf, counter_id, 1) != 0) + return -1; + } + } + + /* Fragment processors' counters */ + for (core_id = 0; core_id < n_fp_cores; core_id++) { + int activity_counter_id = ACTIVITY_FP_0 + core_id; + + snprintf(buf, sizeof(buf), "ARM_Mali-%s_FP_%d_active", mali_name, core_id); + if (create_fs_entry(sb, root, buf, activity_counter_id, 0) != 0) + return -1; + + for (counter_number = 0; counter_number < 2; counter_number++) { + int counter_id = COUNTER_FP_0_C0 + (2 * core_id) + counter_number; + + snprintf(buf, sizeof(buf), "ARM_Mali-%s_FP_%d_cnt%d", mali_name, core_id, counter_number); + if (create_fs_entry(sb, root, buf, counter_id, 1) != 0) + return -1; + } + } + + /* L2 Cache counters */ + for (core_id = 0; core_id < n_l2_cores; core_id++) { + for (counter_number = 0; counter_number < 2; counter_number++) { + int counter_id = COUNTER_L2_0_C0 + (2 * core_id) + counter_number; + + snprintf(buf, sizeof(buf), "ARM_Mali-%s_L2_%d_cnt%d", mali_name, core_id, counter_number); + if (create_fs_entry(sb, root, buf, counter_id, 1) != 0) + return -1; + } + } + + /* Now set up the software counter entries */ + for (event = FIRST_SW_COUNTER; event <= LAST_SW_COUNTER; event++) { + snprintf(buf, sizeof(buf), "ARM_Mali-%s_SW_%d", mali_name, event - FIRST_SW_COUNTER); + + if (create_fs_entry(sb, root, buf, event, 0) != 0) + return -1; + } + + /* Now set up the special counter entries */ + snprintf(buf, sizeof(buf), "ARM_Mali-%s_Filmstrip_cnt0", mali_name); + if (create_fs_entry(sb, root, buf, COUNTER_FILMSTRIP, 1) != 0) + return -1; + +#ifdef DVFS_REPORTED_BY_DDK + snprintf(buf, sizeof(buf), "ARM_Mali-%s_Frequency", mali_name); + if (create_fs_entry(sb, root, buf, COUNTER_FREQUENCY, 1) != 0) + return -1; + + snprintf(buf, sizeof(buf), "ARM_Mali-%s_Voltage", mali_name); + if (create_fs_entry(sb, root, buf, COUNTER_VOLTAGE, 1) != 0) + return -1; +#endif + + return 0; +} + +/* + * Local store for the get_counters entry point into the DDK. + * This is stored here since it is used very regularly. + */ +static void (*mali_get_counters)(unsigned int *, unsigned int *, unsigned int *, unsigned int *); +static u32 (*mali_get_l2_counters)(struct _mali_profiling_l2_counter_values *values); + +/* + * Examine list of counters between two index limits and determine if any one is enabled. + * Returns 1 if any counter is enabled, 0 if none is. + */ +static int is_any_counter_enabled(unsigned int first_counter, unsigned int last_counter) +{ + unsigned int i; + + for (i = first_counter; i <= last_counter; i++) { + if (counter_enabled[i]) + return 1; /* At least one counter is enabled */ + } + + return 0; /* No s/w counters enabled */ +} + +static void init_counters(unsigned int from_counter, unsigned int to_counter) +{ + unsigned int counter_id; + + /* If a Mali driver is present and exporting the appropriate symbol + * then we can request the HW counters (of which there are only 2) + * be configured to count the desired events + */ + mali_profiling_set_event_type *mali_set_hw_event; + + mali_set_hw_event = symbol_get(_mali_profiling_set_event); + + if (mali_set_hw_event) { + pr_debug("gator: mali online _mali_profiling_set_event symbol @ %p\n", mali_set_hw_event); + + for (counter_id = from_counter; counter_id <= to_counter; counter_id++) { + if (counter_enabled[counter_id]) + mali_set_hw_event(counter_id, counter_event[counter_id]); + else + mali_set_hw_event(counter_id, 0xFFFFFFFF); + } + + symbol_put(_mali_profiling_set_event); + } else { + pr_err("gator: mali online _mali_profiling_set_event symbol not found\n"); + } +} + +static void mali_counter_initialize(void) +{ + int i; + + mali_profiling_control_type *mali_control; + + init_counters(COUNTER_L2_0_C0, COUNTER_L2_0_C0 + (2 * n_l2_cores) - 1); + init_counters(COUNTER_VP_0_C0, COUNTER_VP_0_C0 + (2 * n_vp_cores) - 1); + init_counters(COUNTER_FP_0_C0, COUNTER_FP_0_C0 + (2 * n_fp_cores) - 1); + + /* Generic control interface for Mali DDK. */ + mali_control = symbol_get(_mali_profiling_control); + if (mali_control) { + /* The event attribute in the XML file keeps the actual frame rate. */ + unsigned int rate = counter_event[COUNTER_FILMSTRIP] & 0xff; + unsigned int resize_factor = (counter_event[COUNTER_FILMSTRIP] >> 8) & 0xff; + + pr_debug("gator: mali online _mali_profiling_control symbol @ %p\n", mali_control); + + mali_control(SW_COUNTER_ENABLE, (is_any_counter_enabled(FIRST_SW_COUNTER, LAST_SW_COUNTER) ? 1 : 0)); + mali_control(FBDUMP_CONTROL_ENABLE, (counter_enabled[COUNTER_FILMSTRIP] ? 1 : 0)); + mali_control(FBDUMP_CONTROL_RATE, rate); + mali_control(FBDUMP_CONTROL_RESIZE_FACTOR, resize_factor); + + pr_debug("gator: sent mali_control enabled=%d, rate=%d\n", (counter_enabled[COUNTER_FILMSTRIP] ? 1 : 0), rate); + + symbol_put(_mali_profiling_control); + } else { + pr_err("gator: mali online _mali_profiling_control symbol not found\n"); + } + + mali_get_counters = symbol_get(_mali_profiling_get_counters); + if (mali_get_counters) + pr_debug("gator: mali online _mali_profiling_get_counters symbol @ %p\n", mali_get_counters); + else + pr_debug("gator WARNING: mali _mali_profiling_get_counters symbol not defined\n"); + + mali_get_l2_counters = symbol_get(_mali_profiling_get_l2_counters); + if (mali_get_l2_counters) + pr_debug("gator: mali online _mali_profiling_get_l2_counters symbol @ %p\n", mali_get_l2_counters); + else + pr_debug("gator WARNING: mali _mali_profiling_get_l2_counters symbol not defined\n"); + + if (!mali_get_counters && !mali_get_l2_counters) { + pr_debug("gator: WARNING: no L2 counters available\n"); + n_l2_cores = 0; + } + + /* Clear counters in the start */ + for (i = 0; i < NUMBER_OF_EVENTS; i++) { + counter_data[i] = 0; + prev_set[i] = false; + } +} + +static void mali_counter_deinitialize(void) +{ + mali_profiling_set_event_type *mali_set_hw_event; + mali_profiling_control_type *mali_control; + + mali_set_hw_event = symbol_get(_mali_profiling_set_event); + + if (mali_set_hw_event) { + int i; + + pr_debug("gator: mali offline _mali_profiling_set_event symbol @ %p\n", mali_set_hw_event); + for (i = FIRST_HW_COUNTER; i <= LAST_HW_COUNTER; i++) + mali_set_hw_event(i, 0xFFFFFFFF); + + symbol_put(_mali_profiling_set_event); + } else { + pr_err("gator: mali offline _mali_profiling_set_event symbol not found\n"); + } + + /* Generic control interface for Mali DDK. */ + mali_control = symbol_get(_mali_profiling_control); + + if (mali_control) { + pr_debug("gator: mali offline _mali_profiling_control symbol @ %p\n", mali_control); + + /* Reset the DDK state - disable counter collection */ + mali_control(SW_COUNTER_ENABLE, 0); + + mali_control(FBDUMP_CONTROL_ENABLE, 0); + + symbol_put(_mali_profiling_control); + } else { + pr_err("gator: mali offline _mali_profiling_control symbol not found\n"); + } + + if (mali_get_counters) + symbol_put(_mali_profiling_get_counters); + + if (mali_get_l2_counters) + symbol_put(_mali_profiling_get_l2_counters); +} + +static int start(void) +{ + /* register tracepoints */ + if (GATOR_REGISTER_TRACE(mali_hw_counter)) { + pr_err("gator: mali_hw_counter tracepoint failed to activate\n"); + return -1; + } + + /* For Mali drivers with built-in support. */ + if (GATOR_REGISTER_TRACE(mali_sw_counters)) { + pr_err("gator: mali_sw_counters tracepoint failed to activate\n"); + return -1; + } + + trace_registered = 1; + + mali_counter_initialize(); + return 0; +} + +static void stop(void) +{ + unsigned int cnt; + + pr_debug("gator: mali stop\n"); + + if (trace_registered) { + GATOR_UNREGISTER_TRACE(mali_hw_counter); + + /* For Mali drivers with built-in support. */ + GATOR_UNREGISTER_TRACE(mali_sw_counters); + + pr_debug("gator: mali timeline tracepoint deactivated\n"); + + trace_registered = 0; + } + + for (cnt = 0; cnt < NUMBER_OF_EVENTS; cnt++) { + counter_enabled[cnt] = 0; + counter_event[cnt] = 0; + counter_address[cnt] = NULL; + } + + mali_counter_deinitialize(); +} + +static void dump_counters(unsigned int from_counter, unsigned int to_counter, unsigned int *len) +{ + unsigned int counter_id; + + for (counter_id = from_counter; counter_id <= to_counter; counter_id++) { + if (counter_enabled[counter_id]) { + counter_dump[(*len)++] = counter_key[counter_id]; + counter_dump[(*len)++] = counter_data[counter_id]; + + counter_data[counter_id] = 0; + } + } +} + +static int read(int **buffer, bool sched_switch) +{ + int len = 0; + + if (!on_primary_core()) + return 0; + + /* Read the L2 C0 and C1 here. */ + if (n_l2_cores > 0 && is_any_counter_enabled(COUNTER_L2_0_C0, COUNTER_L2_0_C0 + (2 * n_l2_cores))) { + unsigned int unavailable_l2_caches = 0; + struct _mali_profiling_l2_counter_values cache_values; + unsigned int cache_id; + struct _mali_profiling_core_counters *per_core; + + /* Poke the driver to get the counter values - older style; only one L2 cache */ + if (mali_get_l2_counters) { + unavailable_l2_caches = mali_get_l2_counters(&cache_values); + } else if (mali_get_counters) { + per_core = &cache_values.cores[0]; + mali_get_counters(&per_core->source0, &per_core->value0, &per_core->source1, &per_core->value1); + } else { + /* This should never happen, as n_l2_caches is only set > 0 if one of the above functions is found. */ + } + + /* Fill in the two cache counter values for each cache block. */ + for (cache_id = 0; cache_id < n_l2_cores; cache_id++) { + unsigned int counter_id_0 = COUNTER_L2_0_C0 + (2 * cache_id); + unsigned int counter_id_1 = counter_id_0 + 1; + + if ((1 << cache_id) & unavailable_l2_caches) + continue; /* This cache is unavailable (powered-off, possibly). */ + + per_core = &cache_values.cores[cache_id]; + + if (counter_enabled[counter_id_0] && prev_set[counter_id_0]) { + /* Calculate and save src0's counter val0 */ + counter_dump[len++] = counter_key[counter_id_0]; + counter_dump[len++] = per_core->value0 - counter_prev[counter_id_0]; + } + + if (counter_enabled[counter_id_1] && prev_set[counter_id_1]) { + /* Calculate and save src1's counter val1 */ + counter_dump[len++] = counter_key[counter_id_1]; + counter_dump[len++] = per_core->value1 - counter_prev[counter_id_1]; + } + + /* Save the previous values for the counters. */ + counter_prev[counter_id_0] = per_core->value0; + prev_set[counter_id_0] = true; + counter_prev[counter_id_1] = per_core->value1; + prev_set[counter_id_1] = true; + } + } + + /* Process other (non-timeline) counters. */ + dump_counters(COUNTER_VP_0_C0, COUNTER_VP_0_C0 + (2 * n_vp_cores) - 1, &len); + dump_counters(COUNTER_FP_0_C0, COUNTER_FP_0_C0 + (2 * n_fp_cores) - 1, &len); + + dump_counters(FIRST_SW_COUNTER, LAST_SW_COUNTER, &len); + +#ifdef DVFS_REPORTED_BY_DDK + { + int cnt; + /* + * Add in the voltage and frequency counters if enabled. Note + * that, since these are actually passed as events, the counter + * value should not be cleared. + */ + cnt = COUNTER_FREQUENCY; + if (counter_enabled[cnt]) { + counter_dump[len++] = counter_key[cnt]; + counter_dump[len++] = counter_data[cnt]; + } + + cnt = COUNTER_VOLTAGE; + if (counter_enabled[cnt]) { + counter_dump[len++] = counter_key[cnt]; + counter_dump[len++] = counter_data[cnt]; + } + } +#endif + + if (buffer) + *buffer = counter_dump; + + return len; +} + +static struct gator_interface gator_events_mali_interface = { + .create_files = create_files, + .start = start, + .stop = stop, + .read = read, +}; + +extern void gator_events_mali_log_dvfs_event(unsigned int frequency_mhz, unsigned int voltage_mv) +{ +#ifdef DVFS_REPORTED_BY_DDK + counter_data[COUNTER_FREQUENCY] = frequency_mhz; + counter_data[COUNTER_VOLTAGE] = voltage_mv; +#endif +} + +int gator_events_mali_init(void) +{ + unsigned int cnt; + + pr_debug("gator: mali init\n"); + + gator_mali_initialise_counters(mali_activity, ARRAY_SIZE(mali_activity)); + + for (cnt = 0; cnt < NUMBER_OF_EVENTS; cnt++) { + counter_enabled[cnt] = 0; + counter_event[cnt] = 0; + counter_key[cnt] = gator_events_get_key(); + counter_address[cnt] = NULL; + counter_data[cnt] = 0; + } + + trace_registered = 0; + + return gator_events_install(&gator_events_mali_interface); +} diff --git a/drivers/gator/gator_events_mali_4xx.h b/drivers/gator/gator_events_mali_4xx.h new file mode 100644 index 000000000000..976ca8c4cfa1 --- /dev/null +++ b/drivers/gator/gator_events_mali_4xx.h @@ -0,0 +1,18 @@ +/** + * Copyright (C) ARM Limited 2011-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +/* + * Header contains common definitions for the Mali-4xx processors. + */ +#if !defined(GATOR_EVENTS_MALI_4xx_H) +#define GATOR_EVENTS_MALI_4xx_H + +extern void gator_events_mali_log_dvfs_event(unsigned int d0, unsigned int d1); + +#endif /* GATOR_EVENTS_MALI_4xx_H */ diff --git a/drivers/gator/gator_events_mali_common.c b/drivers/gator/gator_events_mali_common.c new file mode 100644 index 000000000000..1af87d649afe --- /dev/null +++ b/drivers/gator/gator_events_mali_common.c @@ -0,0 +1,72 @@ +/** + * Copyright (C) ARM Limited 2012-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ +#include "gator_events_mali_common.h" + +extern int gator_mali_create_file_system(const char *mali_name, const char *event_name, struct super_block *sb, struct dentry *root, struct mali_counter *counter, unsigned long *event) +{ + int err; + char buf[255]; + struct dentry *dir; + + /* If the counter name is empty ignore it */ + if (strlen(event_name) != 0) { + /* Set up the filesystem entry for this event. */ + if (mali_name == NULL) + snprintf(buf, sizeof(buf), "ARM_Mali-%s", event_name); + else + snprintf(buf, sizeof(buf), "ARM_Mali-%s_%s", mali_name, event_name); + + dir = gatorfs_mkdir(sb, root, buf); + + if (dir == NULL) { + pr_debug("gator: %s: error creating file system for: %s (%s)\n", mali_name, event_name, buf); + return -1; + } + + err = gatorfs_create_ulong(sb, dir, "enabled", &counter->enabled); + if (err != 0) { + pr_debug("gator: %s: error calling gatorfs_create_ulong for: %s (%s)\n", mali_name, event_name, buf); + return -1; + } + err = gatorfs_create_ro_ulong(sb, dir, "key", &counter->key); + if (err != 0) { + pr_debug("gator: %s: error calling gatorfs_create_ro_ulong for: %s (%s)\n", mali_name, event_name, buf); + return -1; + } + if (counter->cores != -1) { + err = gatorfs_create_ro_ulong(sb, dir, "cores", &counter->cores); + if (err != 0) { + pr_debug("gator: %s: error calling gatorfs_create_ro_ulong for: %s (%s)\n", mali_name, event_name, buf); + return -1; + } + } + if (event != NULL) { + err = gatorfs_create_ulong(sb, dir, "event", event); + if (err != 0) { + pr_debug("gator: %s: error calling gatorfs_create_ro_ulong for: %s (%s)\n", mali_name, event_name, buf); + return -1; + } + } + } + + return 0; +} + +extern void gator_mali_initialise_counters(struct mali_counter counters[], unsigned int n_counters) +{ + unsigned int cnt; + + for (cnt = 0; cnt < n_counters; cnt++) { + struct mali_counter *counter = &counters[cnt]; + + counter->key = gator_events_get_key(); + counter->enabled = 0; + counter->cores = -1; + } +} diff --git a/drivers/gator/gator_events_mali_common.h b/drivers/gator/gator_events_mali_common.h new file mode 100644 index 000000000000..e7082e62fe88 --- /dev/null +++ b/drivers/gator/gator_events_mali_common.h @@ -0,0 +1,77 @@ +/** + * Copyright (C) ARM Limited 2012-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#if !defined(GATOR_EVENTS_MALI_COMMON_H) +#define GATOR_EVENTS_MALI_COMMON_H + +#include "gator.h" + +#include <linux/module.h> +#include <linux/time.h> +#include <linux/math64.h> +#include <linux/slab.h> +#include <linux/io.h> + +/* Ensure that MALI_SUPPORT has been defined to something. */ +#ifndef MALI_SUPPORT +#error MALI_SUPPORT not defined! +#endif + +/* Values for the supported activity event types */ +#define ACTIVITY_START (1) +#define ACTIVITY_STOP (2) + +/* + * Runtime state information for a counter. + */ +struct mali_counter { + /* 'key' (a unique id set by gatord and returned by gator.ko) */ + unsigned long key; + /* counter enable state */ + unsigned long enabled; + /* for activity counters, the number of cores, otherwise -1 */ + unsigned long cores; +}; + +/* + * Mali-4xx + */ +typedef int mali_profiling_set_event_type(unsigned int, int); +typedef void mali_profiling_control_type(unsigned int, unsigned int); + +/* + * Driver entry points for functions called directly by gator. + */ +extern int _mali_profiling_set_event(unsigned int, int); +extern void _mali_profiling_control(unsigned int, unsigned int); +extern void _mali_profiling_get_counters(unsigned int *, unsigned int *, unsigned int *, unsigned int *); + +/** + * Creates a filesystem entry under /dev/gator relating to the specified event name and key, and + * associate the key/enable values with this entry point. + * + * @param event_name The name of the event. + * @param sb Linux super block + * @param root Directory under which the entry will be created. + * @param counter_key Ptr to location which will be associated with the counter key. + * @param counter_enabled Ptr to location which will be associated with the counter enable state. + * + * @return 0 if entry point was created, non-zero if not. + */ +extern int gator_mali_create_file_system(const char *mali_name, const char *event_name, struct super_block *sb, struct dentry *root, struct mali_counter *counter, unsigned long *event); + +/** + * Initializes the counter array. + * + * @param keys The array of counters + * @param n_counters The number of entries in each of the arrays. + */ +extern void gator_mali_initialise_counters(struct mali_counter counters[], unsigned int n_counters); + +#endif /* GATOR_EVENTS_MALI_COMMON_H */ diff --git a/drivers/gator/gator_events_mali_midgard.c b/drivers/gator/gator_events_mali_midgard.c new file mode 100644 index 000000000000..0aec906d7ae5 --- /dev/null +++ b/drivers/gator/gator_events_mali_midgard.c @@ -0,0 +1,562 @@ +/** + * Copyright (C) ARM Limited 2011-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include "gator.h" + +#include <linux/module.h> +#include <linux/time.h> +#include <linux/math64.h> +#include <linux/slab.h> +#include <linux/io.h> + +#ifdef MALI_DIR_MIDGARD +/* New DDK Directory structure with kernel/drivers/gpu/arm/midgard*/ +#include "mali_linux_trace.h" +#else +/* Old DDK Directory structure with kernel/drivers/gpu/arm/t6xx*/ +#include "linux/mali_linux_trace.h" +#endif + +#include "gator_events_mali_common.h" + +/* + * Check that the MALI_SUPPORT define is set to one of the allowable device codes. + */ +#if (MALI_SUPPORT != MALI_MIDGARD) +#error MALI_SUPPORT set to an invalid device code: expecting MALI_MIDGARD +#endif + +static const char mali_name[] = "Midgard"; + +/* Counters for Mali-Midgard: + * + * - Timeline events + * They are tracepoints, but instead of reporting a number they report a START/STOP event. + * They are reported in Streamline as number of microseconds while that particular counter was active. + * + * - SW counters + * They are tracepoints reporting a particular number. + * They are accumulated in sw_counter_data array until they are passed to Streamline, then they are zeroed. + * + * - Accumulators + * They are the same as software counters but their value is not zeroed. + */ + +/* Timeline (start/stop) activity */ +static const char *const timeline_event_names[] = { + "PM_SHADER_0", + "PM_SHADER_1", + "PM_SHADER_2", + "PM_SHADER_3", + "PM_SHADER_4", + "PM_SHADER_5", + "PM_SHADER_6", + "PM_SHADER_7", + "PM_TILER_0", + "PM_L2_0", + "PM_L2_1", + "MMU_AS_0", + "MMU_AS_1", + "MMU_AS_2", + "MMU_AS_3" +}; + +enum { + PM_SHADER_0 = 0, + PM_SHADER_1, + PM_SHADER_2, + PM_SHADER_3, + PM_SHADER_4, + PM_SHADER_5, + PM_SHADER_6, + PM_SHADER_7, + PM_TILER_0, + PM_L2_0, + PM_L2_1, + MMU_AS_0, + MMU_AS_1, + MMU_AS_2, + MMU_AS_3 +}; +/* The number of shader blocks in the enum above */ +#define NUM_PM_SHADER (8) + +/* Software Counters */ +static const char *const software_counter_names[] = { + "MMU_PAGE_FAULT_0", + "MMU_PAGE_FAULT_1", + "MMU_PAGE_FAULT_2", + "MMU_PAGE_FAULT_3" +}; + +enum { + MMU_PAGE_FAULT_0 = 0, + MMU_PAGE_FAULT_1, + MMU_PAGE_FAULT_2, + MMU_PAGE_FAULT_3 +}; + +/* Software Counters */ +static const char *const accumulators_names[] = { + "TOTAL_ALLOC_PAGES" +}; + +enum { + TOTAL_ALLOC_PAGES = 0 +}; + +#define FIRST_TIMELINE_EVENT (0) +#define NUMBER_OF_TIMELINE_EVENTS (sizeof(timeline_event_names) / sizeof(timeline_event_names[0])) +#define FIRST_SOFTWARE_COUNTER (FIRST_TIMELINE_EVENT + NUMBER_OF_TIMELINE_EVENTS) +#define NUMBER_OF_SOFTWARE_COUNTERS (sizeof(software_counter_names) / sizeof(software_counter_names[0])) +#define FIRST_ACCUMULATOR (FIRST_SOFTWARE_COUNTER + NUMBER_OF_SOFTWARE_COUNTERS) +#define NUMBER_OF_ACCUMULATORS (sizeof(accumulators_names) / sizeof(accumulators_names[0])) +#define FILMSTRIP (NUMBER_OF_TIMELINE_EVENTS + NUMBER_OF_SOFTWARE_COUNTERS + NUMBER_OF_ACCUMULATORS) +#define NUMBER_OF_EVENTS (NUMBER_OF_TIMELINE_EVENTS + NUMBER_OF_SOFTWARE_COUNTERS + NUMBER_OF_ACCUMULATORS + 1) + +/* + * gatorfs variables for counter enable state + */ +static struct mali_counter counters[NUMBER_OF_EVENTS]; +static unsigned long filmstrip_event; + +/* An array used to return the data we recorded + * as key,value pairs hence the *2 + */ +static int counter_dump[NUMBER_OF_EVENTS * 2]; + +/* + * Array holding counter start times (in ns) for each counter. A zero + * here indicates that the activity monitored by this counter is not + * running. + */ +static struct timespec timeline_event_starttime[NUMBER_OF_TIMELINE_EVENTS]; + +/* The data we have recorded */ +static unsigned int timeline_data[NUMBER_OF_TIMELINE_EVENTS]; +static unsigned int sw_counter_data[NUMBER_OF_SOFTWARE_COUNTERS]; +static unsigned int accumulators_data[NUMBER_OF_ACCUMULATORS]; + +/* Hold the previous timestamp, used to calculate the sample interval. */ +static struct timespec prev_timestamp; + +/** + * Returns the timespan (in microseconds) between the two specified timestamps. + * + * @param start Ptr to the start timestamp + * @param end Ptr to the end timestamp + * + * @return Number of microseconds between the two timestamps (can be negative if start follows end). + */ +static inline long get_duration_us(const struct timespec *start, const struct timespec *end) +{ + long event_duration_us = (end->tv_nsec - start->tv_nsec) / 1000; + + event_duration_us += (end->tv_sec - start->tv_sec) * 1000000; + + return event_duration_us; +} + +static void record_timeline_event(unsigned int timeline_index, unsigned int type) +{ + struct timespec event_timestamp; + struct timespec *event_start = &timeline_event_starttime[timeline_index]; + + switch (type) { + case ACTIVITY_START: + /* Get the event time... */ + getnstimeofday(&event_timestamp); + + /* Remember the start time if the activity is not already started */ + if (event_start->tv_sec == 0) + *event_start = event_timestamp; /* Structure copy */ + break; + + case ACTIVITY_STOP: + /* if the counter was started... */ + if (event_start->tv_sec != 0) { + /* Get the event time... */ + getnstimeofday(&event_timestamp); + + /* Accumulate the duration in us */ + timeline_data[timeline_index] += get_duration_us(event_start, &event_timestamp); + + /* Reset the start time to indicate the activity is stopped. */ + event_start->tv_sec = 0; + } + break; + + default: + /* Other activity events are ignored. */ + break; + } +} + +/* + * Documentation about the following tracepoints is in mali_linux_trace.h + */ + +GATOR_DEFINE_PROBE(mali_pm_status, TP_PROTO(unsigned int event_id, unsigned long long value)) +{ +#define SHADER_PRESENT_LO 0x100 /* (RO) Shader core present bitmap, low word */ +#define TILER_PRESENT_LO 0x110 /* (RO) Tiler core present bitmap, low word */ +#define L2_PRESENT_LO 0x120 /* (RO) Level 2 cache present bitmap, low word */ +#define BIT_AT(value, pos) ((value >> pos) & 1) + + static unsigned long long previous_shader_bitmask; + static unsigned long long previous_tiler_bitmask; + static unsigned long long previous_l2_bitmask; + + switch (event_id) { + case SHADER_PRESENT_LO: + { + unsigned long long changed_bitmask = previous_shader_bitmask ^ value; + int pos; + + for (pos = 0; pos < NUM_PM_SHADER; ++pos) { + if (BIT_AT(changed_bitmask, pos)) + record_timeline_event(PM_SHADER_0 + pos, BIT_AT(value, pos) ? ACTIVITY_START : ACTIVITY_STOP); + } + + previous_shader_bitmask = value; + break; + } + + case TILER_PRESENT_LO: + { + unsigned long long changed = previous_tiler_bitmask ^ value; + + if (BIT_AT(changed, 0)) + record_timeline_event(PM_TILER_0, BIT_AT(value, 0) ? ACTIVITY_START : ACTIVITY_STOP); + + previous_tiler_bitmask = value; + break; + } + + case L2_PRESENT_LO: + { + unsigned long long changed = previous_l2_bitmask ^ value; + + if (BIT_AT(changed, 0)) + record_timeline_event(PM_L2_0, BIT_AT(value, 0) ? ACTIVITY_START : ACTIVITY_STOP); + if (BIT_AT(changed, 4)) + record_timeline_event(PM_L2_1, BIT_AT(value, 4) ? ACTIVITY_START : ACTIVITY_STOP); + + previous_l2_bitmask = value; + break; + } + + default: + /* No other blocks are supported at present */ + break; + } + +#undef SHADER_PRESENT_LO +#undef TILER_PRESENT_LO +#undef L2_PRESENT_LO +#undef BIT_AT +} + +GATOR_DEFINE_PROBE(mali_page_fault_insert_pages, TP_PROTO(int event_id, unsigned long value)) +{ + /* We add to the previous since we may receive many tracepoints in one sample period */ + sw_counter_data[MMU_PAGE_FAULT_0 + event_id] += value; +} + +GATOR_DEFINE_PROBE(mali_mmu_as_in_use, TP_PROTO(int event_id)) +{ + record_timeline_event(MMU_AS_0 + event_id, ACTIVITY_START); +} + +GATOR_DEFINE_PROBE(mali_mmu_as_released, TP_PROTO(int event_id)) +{ + record_timeline_event(MMU_AS_0 + event_id, ACTIVITY_STOP); +} + +GATOR_DEFINE_PROBE(mali_total_alloc_pages_change, TP_PROTO(long long int event_id)) +{ + accumulators_data[TOTAL_ALLOC_PAGES] = event_id; +} + +static int create_files(struct super_block *sb, struct dentry *root) +{ + int event; + /* + * Create the filesystem for all events + */ + int counter_index = 0; + mali_profiling_control_type *mali_control; + + for (event = FIRST_TIMELINE_EVENT; event < FIRST_TIMELINE_EVENT + NUMBER_OF_TIMELINE_EVENTS; event++) { + if (gator_mali_create_file_system(mali_name, timeline_event_names[counter_index], sb, root, &counters[event], NULL) != 0) + return -1; + counter_index++; + } + counter_index = 0; + for (event = FIRST_SOFTWARE_COUNTER; event < FIRST_SOFTWARE_COUNTER + NUMBER_OF_SOFTWARE_COUNTERS; event++) { + if (gator_mali_create_file_system(mali_name, software_counter_names[counter_index], sb, root, &counters[event], NULL) != 0) + return -1; + counter_index++; + } + counter_index = 0; + for (event = FIRST_ACCUMULATOR; event < FIRST_ACCUMULATOR + NUMBER_OF_ACCUMULATORS; event++) { + if (gator_mali_create_file_system(mali_name, accumulators_names[counter_index], sb, root, &counters[event], NULL) != 0) + return -1; + counter_index++; + } + + mali_control = symbol_get(_mali_profiling_control); + if (mali_control) { + if (gator_mali_create_file_system(mali_name, "Filmstrip_cnt0", sb, root, &counters[FILMSTRIP], &filmstrip_event) != 0) + return -1; + symbol_put(_mali_profiling_control); + } + + return 0; +} + +static int register_tracepoints(void) +{ + if (GATOR_REGISTER_TRACE(mali_pm_status)) { + pr_debug("gator: Mali-Midgard: mali_pm_status tracepoint failed to activate\n"); + return 0; + } + + if (GATOR_REGISTER_TRACE(mali_page_fault_insert_pages)) { + pr_debug("gator: Mali-Midgard: mali_page_fault_insert_pages tracepoint failed to activate\n"); + return 0; + } + + if (GATOR_REGISTER_TRACE(mali_mmu_as_in_use)) { + pr_debug("gator: Mali-Midgard: mali_mmu_as_in_use tracepoint failed to activate\n"); + return 0; + } + + if (GATOR_REGISTER_TRACE(mali_mmu_as_released)) { + pr_debug("gator: Mali-Midgard: mali_mmu_as_released tracepoint failed to activate\n"); + return 0; + } + + if (GATOR_REGISTER_TRACE(mali_total_alloc_pages_change)) { + pr_debug("gator: Mali-Midgard: mali_total_alloc_pages_change tracepoint failed to activate\n"); + return 0; + } + + pr_debug("gator: Mali-Midgard: start\n"); + pr_debug("gator: Mali-Midgard: mali_pm_status probe is at %p\n", &probe_mali_pm_status); + pr_debug("gator: Mali-Midgard: mali_page_fault_insert_pages probe is at %p\n", &probe_mali_page_fault_insert_pages); + pr_debug("gator: Mali-Midgard: mali_mmu_as_in_use probe is at %p\n", &probe_mali_mmu_as_in_use); + pr_debug("gator: Mali-Midgard: mali_mmu_as_released probe is at %p\n", &probe_mali_mmu_as_released); + pr_debug("gator: Mali-Midgard: mali_total_alloc_pages_change probe is at %p\n", &probe_mali_total_alloc_pages_change); + + return 1; +} + +static int start(void) +{ + unsigned int cnt; + mali_profiling_control_type *mali_control; + + /* Clean all data for the next capture */ + for (cnt = 0; cnt < NUMBER_OF_TIMELINE_EVENTS; cnt++) { + timeline_event_starttime[cnt].tv_sec = timeline_event_starttime[cnt].tv_nsec = 0; + timeline_data[cnt] = 0; + } + + for (cnt = 0; cnt < NUMBER_OF_SOFTWARE_COUNTERS; cnt++) + sw_counter_data[cnt] = 0; + + for (cnt = 0; cnt < NUMBER_OF_ACCUMULATORS; cnt++) + accumulators_data[cnt] = 0; + + /* Register tracepoints */ + if (register_tracepoints() == 0) + return -1; + + /* Generic control interface for Mali DDK. */ + mali_control = symbol_get(_mali_profiling_control); + if (mali_control) { + /* The event attribute in the XML file keeps the actual frame rate. */ + unsigned int enabled = counters[FILMSTRIP].enabled ? 1 : 0; + unsigned int rate = filmstrip_event & 0xff; + unsigned int resize_factor = (filmstrip_event >> 8) & 0xff; + + pr_debug("gator: mali online _mali_profiling_control symbol @ %p\n", mali_control); + +#define FBDUMP_CONTROL_ENABLE (1) +#define FBDUMP_CONTROL_RATE (2) +#define FBDUMP_CONTROL_RESIZE_FACTOR (4) + mali_control(FBDUMP_CONTROL_ENABLE, enabled); + mali_control(FBDUMP_CONTROL_RATE, rate); + mali_control(FBDUMP_CONTROL_RESIZE_FACTOR, resize_factor); + + pr_debug("gator: sent mali_control enabled=%d, rate=%d, resize_factor=%d\n", enabled, rate, resize_factor); + + symbol_put(_mali_profiling_control); + } else { + pr_err("gator: mali online _mali_profiling_control symbol not found\n"); + } + + /* + * Set the first timestamp for calculating the sample interval. The first interval could be quite long, + * since it will be the time between 'start' and the first 'read'. + * This means that timeline values will be divided by a big number for the first sample. + */ + getnstimeofday(&prev_timestamp); + + return 0; +} + +static void stop(void) +{ + mali_profiling_control_type *mali_control; + + pr_debug("gator: Mali-Midgard: stop\n"); + + /* + * It is safe to unregister traces even if they were not successfully + * registered, so no need to check. + */ + GATOR_UNREGISTER_TRACE(mali_pm_status); + pr_debug("gator: Mali-Midgard: mali_pm_status tracepoint deactivated\n"); + + GATOR_UNREGISTER_TRACE(mali_page_fault_insert_pages); + pr_debug("gator: Mali-Midgard: mali_page_fault_insert_pages tracepoint deactivated\n"); + + GATOR_UNREGISTER_TRACE(mali_mmu_as_in_use); + pr_debug("gator: Mali-Midgard: mali_mmu_as_in_use tracepoint deactivated\n"); + + GATOR_UNREGISTER_TRACE(mali_mmu_as_released); + pr_debug("gator: Mali-Midgard: mali_mmu_as_released tracepoint deactivated\n"); + + GATOR_UNREGISTER_TRACE(mali_total_alloc_pages_change); + pr_debug("gator: Mali-Midgard: mali_total_alloc_pages_change tracepoint deactivated\n"); + + /* Generic control interface for Mali DDK. */ + mali_control = symbol_get(_mali_profiling_control); + if (mali_control) { + pr_debug("gator: mali offline _mali_profiling_control symbol @ %p\n", mali_control); + + mali_control(FBDUMP_CONTROL_ENABLE, 0); + + symbol_put(_mali_profiling_control); + } else { + pr_err("gator: mali offline _mali_profiling_control symbol not found\n"); + } +} + +static int read(int **buffer, bool sched_switch) +{ + int cnt; + int len = 0; + long sample_interval_us = 0; + struct timespec read_timestamp; + + if (!on_primary_core()) + return 0; + + /* Get the start of this sample period. */ + getnstimeofday(&read_timestamp); + + /* + * Calculate the sample interval if the previous sample time is valid. + * We use tv_sec since it will not be 0. + */ + if (prev_timestamp.tv_sec != 0) + sample_interval_us = get_duration_us(&prev_timestamp, &read_timestamp); + + /* Structure copy. Update the previous timestamp. */ + prev_timestamp = read_timestamp; + + /* + * Report the timeline counters (ACTIVITY_START/STOP) + */ + for (cnt = FIRST_TIMELINE_EVENT; cnt < (FIRST_TIMELINE_EVENT + NUMBER_OF_TIMELINE_EVENTS); cnt++) { + struct mali_counter *counter = &counters[cnt]; + + if (counter->enabled) { + const int index = cnt - FIRST_TIMELINE_EVENT; + unsigned int value; + + /* If the activity is still running, reset its start time to the + * start of this sample period to correct the count. Add the + * time up to the end of the sample onto the count. + */ + if (timeline_event_starttime[index].tv_sec != 0) { + const long event_duration = get_duration_us(&timeline_event_starttime[index], &read_timestamp); + + timeline_data[index] += event_duration; + timeline_event_starttime[index] = read_timestamp; /* Activity is still running. */ + } + + if (sample_interval_us != 0) { + /* Convert the counter to a percent-of-sample value */ + value = (timeline_data[index] * 100) / sample_interval_us; + } else { + pr_debug("gator: Mali-Midgard: setting value to zero\n"); + value = 0; + } + + /* Clear the counter value ready for the next sample. */ + timeline_data[index] = 0; + + counter_dump[len++] = counter->key; + counter_dump[len++] = value; + } + } + + /* Report the software counters */ + for (cnt = FIRST_SOFTWARE_COUNTER; cnt < (FIRST_SOFTWARE_COUNTER + NUMBER_OF_SOFTWARE_COUNTERS); cnt++) { + const struct mali_counter *counter = &counters[cnt]; + + if (counter->enabled) { + const int index = cnt - FIRST_SOFTWARE_COUNTER; + + counter_dump[len++] = counter->key; + counter_dump[len++] = sw_counter_data[index]; + /* Set the value to zero for the next time */ + sw_counter_data[index] = 0; + } + } + + /* Report the accumulators */ + for (cnt = FIRST_ACCUMULATOR; cnt < (FIRST_ACCUMULATOR + NUMBER_OF_ACCUMULATORS); cnt++) { + const struct mali_counter *counter = &counters[cnt]; + + if (counter->enabled) { + const int index = cnt - FIRST_ACCUMULATOR; + + counter_dump[len++] = counter->key; + counter_dump[len++] = accumulators_data[index]; + /* Do not zero the accumulator */ + } + } + + /* Update the buffer */ + if (buffer) + *buffer = counter_dump; + + return len; +} + +static struct gator_interface gator_events_mali_midgard_interface = { + .create_files = create_files, + .start = start, + .stop = stop, + .read = read +}; + +extern int gator_events_mali_midgard_init(void) +{ + pr_debug("gator: Mali-Midgard: sw_counters init\n"); + + gator_mali_initialise_counters(counters, NUMBER_OF_EVENTS); + + return gator_events_install(&gator_events_mali_midgard_interface); +} diff --git a/drivers/gator/gator_events_mali_midgard_hw.c b/drivers/gator/gator_events_mali_midgard_hw.c new file mode 100644 index 000000000000..c8065da56815 --- /dev/null +++ b/drivers/gator/gator_events_mali_midgard_hw.c @@ -0,0 +1,977 @@ +/** + * Copyright (C) ARM Limited 2012-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include "gator.h" + +#include <linux/module.h> +#include <linux/time.h> +#include <linux/math64.h> +#include <linux/slab.h> +#include <linux/io.h> + +/* Mali Midgard DDK includes */ +#if defined(MALI_SIMPLE_API) +/* Header with wrapper functions to kbase structures and functions */ +#include "mali/mali_kbase_gator_api.h" +#elif defined(MALI_DIR_MIDGARD) +/* New DDK Directory structure with kernel/drivers/gpu/arm/midgard */ +#include "mali_linux_trace.h" +#include "mali_kbase.h" +#include "mali_kbase_mem_linux.h" +#else +/* Old DDK Directory structure with kernel/drivers/gpu/arm/t6xx */ +#include "linux/mali_linux_trace.h" +#include "kbase/src/common/mali_kbase.h" +#include "kbase/src/linux/mali_kbase_mem_linux.h" +#endif + +/* If API version is not specified then assume API version 1. */ +#ifndef MALI_DDK_GATOR_API_VERSION +#define MALI_DDK_GATOR_API_VERSION 1 +#endif + +#if (MALI_DDK_GATOR_API_VERSION != 1) && (MALI_DDK_GATOR_API_VERSION != 2) && (MALI_DDK_GATOR_API_VERSION != 3) +#error MALI_DDK_GATOR_API_VERSION is invalid (must be 1 for r1/r2 DDK, or 2 for r3/r4 DDK, or 3 for r5 and later DDK). +#endif + +#include "gator_events_mali_common.h" + +/* + * Mali-Midgard + */ +#if MALI_DDK_GATOR_API_VERSION == 3 +static uint32_t (*kbase_gator_instr_hwcnt_dump_irq_symbol)(struct kbase_gator_hwcnt_handles *); +static uint32_t (*kbase_gator_instr_hwcnt_dump_complete_symbol)(struct kbase_gator_hwcnt_handles *, uint32_t *const); +static struct kbase_gator_hwcnt_handles *(*kbase_gator_hwcnt_init_symbol)(struct kbase_gator_hwcnt_info *); +static void (*kbase_gator_hwcnt_term_symbol)(struct kbase_gator_hwcnt_info *, struct kbase_gator_hwcnt_handles *); + +#else +static struct kbase_device *(*kbase_find_device_symbol)(int); +static struct kbase_context *(*kbase_create_context_symbol)(struct kbase_device *); +static void (*kbase_destroy_context_symbol)(struct kbase_context *); + +#if MALI_DDK_GATOR_API_VERSION == 1 +static void *(*kbase_va_alloc_symbol)(struct kbase_context *, u32); +static void (*kbase_va_free_symbol)(struct kbase_context *, void *); +#elif MALI_DDK_GATOR_API_VERSION == 2 +static void *(*kbase_va_alloc_symbol)(struct kbase_context *, u32, struct kbase_hwc_dma_mapping *); +static void (*kbase_va_free_symbol)(struct kbase_context *, struct kbase_hwc_dma_mapping *); +#endif + +static mali_error (*kbase_instr_hwcnt_enable_symbol)(struct kbase_context *, struct kbase_uk_hwcnt_setup *); +static mali_error (*kbase_instr_hwcnt_disable_symbol)(struct kbase_context *); +static mali_error (*kbase_instr_hwcnt_clear_symbol)(struct kbase_context *); +static mali_error (*kbase_instr_hwcnt_dump_irq_symbol)(struct kbase_context *); +static mali_bool (*kbase_instr_hwcnt_dump_complete_symbol)(struct kbase_context *, mali_bool *); + +static long shader_present_low; +#endif + +/** The interval between reads, in ns. + * + * Earlier we introduced a 'hold off for 1ms after last read' to + * resolve MIDBASE-2178 and MALINE-724. However, the 1ms hold off is + * too long if no context switches occur as there is a race between + * this value and the tick of the read clock in gator which is also + * 1ms. If we 'miss' the current read, the counter values are + * effectively 'spread' over 2ms and the values seen are half what + * they should be (since Streamline averages over sample time). In the + * presence of context switches this spread can vary and markedly + * affect the counters. Currently there is no 'proper' solution to + * this, but empirically we have found that reducing the minimum read + * interval to 950us causes the counts to be much more stable. + */ +static const int READ_INTERVAL_NSEC = 950000; + +#if GATOR_TEST +#include "gator_events_mali_midgard_hw_test.c" +#endif + +#if MALI_DDK_GATOR_API_VERSION != 3 +/* Blocks for HW counters */ +enum { + JM_BLOCK = 0, + TILER_BLOCK, + SHADER_BLOCK, + MMU_BLOCK +}; +#endif + +static const char *mali_name; + +/* Counters for Mali-Midgard: + * + * For HW counters we need strings to create /dev/gator/events files. + * Enums are not needed because the position of the HW name in the array is the same + * of the corresponding value in the received block of memory. + * HW counters are requested by calculating a bitmask, passed then to the driver. + * Every millisecond a HW counters dump is requested, and if the previous has been completed they are read. + */ + +/* Hardware Counters */ +#if MALI_DDK_GATOR_API_VERSION == 3 + +static const char *const *hardware_counter_names; +static int number_of_hardware_counters; + +#else + +static const char *const hardware_counter_names[] = { + /* Job Manager */ + "", + "", + "", + "", + "MESSAGES_SENT", + "MESSAGES_RECEIVED", + "GPU_ACTIVE", /* 6 */ + "IRQ_ACTIVE", + "JS0_JOBS", + "JS0_TASKS", + "JS0_ACTIVE", + "", + "JS0_WAIT_READ", + "JS0_WAIT_ISSUE", + "JS0_WAIT_DEPEND", + "JS0_WAIT_FINISH", + "JS1_JOBS", + "JS1_TASKS", + "JS1_ACTIVE", + "", + "JS1_WAIT_READ", + "JS1_WAIT_ISSUE", + "JS1_WAIT_DEPEND", + "JS1_WAIT_FINISH", + "JS2_JOBS", + "JS2_TASKS", + "JS2_ACTIVE", + "", + "JS2_WAIT_READ", + "JS2_WAIT_ISSUE", + "JS2_WAIT_DEPEND", + "JS2_WAIT_FINISH", + "JS3_JOBS", + "JS3_TASKS", + "JS3_ACTIVE", + "", + "JS3_WAIT_READ", + "JS3_WAIT_ISSUE", + "JS3_WAIT_DEPEND", + "JS3_WAIT_FINISH", + "JS4_JOBS", + "JS4_TASKS", + "JS4_ACTIVE", + "", + "JS4_WAIT_READ", + "JS4_WAIT_ISSUE", + "JS4_WAIT_DEPEND", + "JS4_WAIT_FINISH", + "JS5_JOBS", + "JS5_TASKS", + "JS5_ACTIVE", + "", + "JS5_WAIT_READ", + "JS5_WAIT_ISSUE", + "JS5_WAIT_DEPEND", + "JS5_WAIT_FINISH", + "JS6_JOBS", + "JS6_TASKS", + "JS6_ACTIVE", + "", + "JS6_WAIT_READ", + "JS6_WAIT_ISSUE", + "JS6_WAIT_DEPEND", + "JS6_WAIT_FINISH", + + /*Tiler */ + "", + "", + "", + "JOBS_PROCESSED", + "TRIANGLES", + "QUADS", + "POLYGONS", + "POINTS", + "LINES", + "VCACHE_HIT", + "VCACHE_MISS", + "FRONT_FACING", + "BACK_FACING", + "PRIM_VISIBLE", + "PRIM_CULLED", + "PRIM_CLIPPED", + "LEVEL0", + "LEVEL1", + "LEVEL2", + "LEVEL3", + "LEVEL4", + "LEVEL5", + "LEVEL6", + "LEVEL7", + "COMMAND_1", + "COMMAND_2", + "COMMAND_3", + "COMMAND_4", + "COMMAND_4_7", + "COMMAND_8_15", + "COMMAND_16_63", + "COMMAND_64", + "COMPRESS_IN", + "COMPRESS_OUT", + "COMPRESS_FLUSH", + "TIMESTAMPS", + "PCACHE_HIT", + "PCACHE_MISS", + "PCACHE_LINE", + "PCACHE_STALL", + "WRBUF_HIT", + "WRBUF_MISS", + "WRBUF_LINE", + "WRBUF_PARTIAL", + "WRBUF_STALL", + "ACTIVE", + "LOADING_DESC", + "INDEX_WAIT", + "INDEX_RANGE_WAIT", + "VERTEX_WAIT", + "PCACHE_WAIT", + "WRBUF_WAIT", + "BUS_READ", + "BUS_WRITE", + "", + "", + "", + "", + "", + "UTLB_STALL", + "UTLB_REPLAY_MISS", + "UTLB_REPLAY_FULL", + "UTLB_NEW_MISS", + "UTLB_HIT", + + /* Shader Core */ + "", + "", + "", + "SHADER_CORE_ACTIVE", + "FRAG_ACTIVE", + "FRAG_PRIMATIVES", + "FRAG_PRIMATIVES_DROPPED", + "FRAG_CYCLE_DESC", + "FRAG_CYCLES_PLR", + "FRAG_CYCLES_VERT", + "FRAG_CYCLES_TRISETUP", + "FRAG_CYCLES_RAST", + "FRAG_THREADS", + "FRAG_DUMMY_THREADS", + "FRAG_QUADS_RAST", + "FRAG_QUADS_EZS_TEST", + "FRAG_QUADS_EZS_KILLED", + "FRAG_QUADS_LZS_TEST", + "FRAG_QUADS_LZS_KILLED", + "FRAG_CYCLE_NO_TILE", + "FRAG_NUM_TILES", + "FRAG_TRANS_ELIM", + "COMPUTE_ACTIVE", + "COMPUTE_TASKS", + "COMPUTE_THREADS", + "COMPUTE_CYCLES_DESC", + "TRIPIPE_ACTIVE", + "ARITH_WORDS", + "ARITH_CYCLES_REG", + "ARITH_CYCLES_L0", + "ARITH_FRAG_DEPEND", + "LS_WORDS", + "LS_ISSUES", + "LS_RESTARTS", + "LS_REISSUES_MISS", + "LS_REISSUES_VD", + "LS_REISSUE_ATTRIB_MISS", + "LS_NO_WB", + "TEX_WORDS", + "TEX_BUBBLES", + "TEX_WORDS_L0", + "TEX_WORDS_DESC", + "TEX_THREADS", + "TEX_RECIRC_FMISS", + "TEX_RECIRC_DESC", + "TEX_RECIRC_MULTI", + "TEX_RECIRC_PMISS", + "TEX_RECIRC_CONF", + "LSC_READ_HITS", + "LSC_READ_MISSES", + "LSC_WRITE_HITS", + "LSC_WRITE_MISSES", + "LSC_ATOMIC_HITS", + "LSC_ATOMIC_MISSES", + "LSC_LINE_FETCHES", + "LSC_DIRTY_LINE", + "LSC_SNOOPS", + "AXI_TLB_STALL", + "AXI_TLB_MIESS", + "AXI_TLB_TRANSACTION", + "LS_TLB_MISS", + "LS_TLB_HIT", + "AXI_BEATS_READ", + "AXI_BEATS_WRITTEN", + + /*L2 and MMU */ + "", + "", + "", + "", + "MMU_HIT", + "MMU_NEW_MISS", + "MMU_REPLAY_FULL", + "MMU_REPLAY_MISS", + "MMU_TABLE_WALK", + "", + "", + "", + "", + "", + "", + "", + "UTLB_HIT", + "UTLB_NEW_MISS", + "UTLB_REPLAY_FULL", + "UTLB_REPLAY_MISS", + "UTLB_STALL", + "", + "", + "", + "", + "", + "", + "", + "", + "", + "L2_WRITE_BEATS", + "L2_READ_BEATS", + "L2_ANY_LOOKUP", + "L2_READ_LOOKUP", + "L2_SREAD_LOOKUP", + "L2_READ_REPLAY", + "L2_READ_SNOOP", + "L2_READ_HIT", + "L2_CLEAN_MISS", + "L2_WRITE_LOOKUP", + "L2_SWRITE_LOOKUP", + "L2_WRITE_REPLAY", + "L2_WRITE_SNOOP", + "L2_WRITE_HIT", + "L2_EXT_READ_FULL", + "L2_EXT_READ_HALF", + "L2_EXT_WRITE_FULL", + "L2_EXT_WRITE_HALF", + "L2_EXT_READ", + "L2_EXT_READ_LINE", + "L2_EXT_WRITE", + "L2_EXT_WRITE_LINE", + "L2_EXT_WRITE_SMALL", + "L2_EXT_BARRIER", + "L2_EXT_AR_STALL", + "L2_EXT_R_BUF_FULL", + "L2_EXT_RD_BUF_FULL", + "L2_EXT_R_RAW", + "L2_EXT_W_STALL", + "L2_EXT_W_BUF_FULL", + "L2_EXT_R_W_HAZARD", + "L2_TAG_HAZARD", + "L2_SNOOP_FULL", + "L2_REPLAY_FULL" +}; + +static const int number_of_hardware_counters = ARRAY_SIZE(hardware_counter_names); + +#endif + +#define GET_HW_BLOCK(c) (((c) >> 6) & 0x3) +#define GET_COUNTER_OFFSET(c) ((c) & 0x3f) + +#if MALI_DDK_GATOR_API_VERSION == 3 +/* Opaque handles for kbase_context and kbase_hwc_dma_mapping */ +static struct kbase_gator_hwcnt_handles *handles; + +/* Information about hardware counters */ +static struct kbase_gator_hwcnt_info *in_out_info; + +#else +/* Memory to dump hardware counters into */ +static void *kernel_dump_buffer; + +#if MALI_DDK_GATOR_API_VERSION == 2 +/* DMA state used to manage lifetime of the buffer */ +struct kbase_hwc_dma_mapping kernel_dump_buffer_handle; +#endif + +/* kbase context and device */ +static struct kbase_context *kbcontext; +static struct kbase_device *kbdevice; + +/* + * The following function has no external prototype in older DDK + * revisions. When the DDK is updated then this should be removed. + */ +struct kbase_device *kbase_find_device(int minor); +#endif + +static volatile bool kbase_device_busy; +static unsigned int num_hardware_counters_enabled; + +/* gatorfs variables for counter enable state */ +static struct mali_counter *counters; + +/* An array used to return the data we recorded as key,value pairs */ +static int *counter_dump; + +extern struct mali_counter mali_activity[3]; + +static const char *const mali_activity_names[] = { + "fragment", + "vertex", + "opencl", +}; + +#define SYMBOL_GET(FUNCTION, ERROR_COUNT) \ + do { \ + if (FUNCTION ## _symbol) { \ + pr_err("gator: mali " #FUNCTION " symbol was already registered\n"); \ + (ERROR_COUNT)++; \ + } else { \ + FUNCTION ## _symbol = symbol_get(FUNCTION); \ + if (!FUNCTION ## _symbol) { \ + pr_err("gator: mali online " #FUNCTION " symbol not found\n"); \ + (ERROR_COUNT)++; \ + } \ + } \ + } while (0) + +#define SYMBOL_CLEANUP(FUNCTION) \ + do { \ + if (FUNCTION ## _symbol) { \ + symbol_put(FUNCTION); \ + FUNCTION ## _symbol = NULL; \ + } \ + } while (0) + +/** + * Execute symbol_get for all the Mali symbols and check for success. + * @return the number of symbols not loaded. + */ +static int init_symbols(void) +{ + int error_count = 0; +#if MALI_DDK_GATOR_API_VERSION == 3 + SYMBOL_GET(kbase_gator_instr_hwcnt_dump_irq, error_count); + SYMBOL_GET(kbase_gator_instr_hwcnt_dump_complete, error_count); + SYMBOL_GET(kbase_gator_hwcnt_init, error_count); + SYMBOL_GET(kbase_gator_hwcnt_term, error_count); +#else + SYMBOL_GET(kbase_find_device, error_count); + SYMBOL_GET(kbase_create_context, error_count); + SYMBOL_GET(kbase_va_alloc, error_count); + SYMBOL_GET(kbase_instr_hwcnt_enable, error_count); + SYMBOL_GET(kbase_instr_hwcnt_clear, error_count); + SYMBOL_GET(kbase_instr_hwcnt_dump_irq, error_count); + SYMBOL_GET(kbase_instr_hwcnt_dump_complete, error_count); + SYMBOL_GET(kbase_instr_hwcnt_disable, error_count); + SYMBOL_GET(kbase_va_free, error_count); + SYMBOL_GET(kbase_destroy_context, error_count); +#endif + + return error_count; +} + +/** + * Execute symbol_put for all the registered Mali symbols. + */ +static void clean_symbols(void) +{ +#if MALI_DDK_GATOR_API_VERSION == 3 + SYMBOL_CLEANUP(kbase_gator_instr_hwcnt_dump_irq); + SYMBOL_CLEANUP(kbase_gator_instr_hwcnt_dump_complete); + SYMBOL_CLEANUP(kbase_gator_hwcnt_init); + SYMBOL_CLEANUP(kbase_gator_hwcnt_term); +#else + SYMBOL_CLEANUP(kbase_find_device); + SYMBOL_CLEANUP(kbase_create_context); + SYMBOL_CLEANUP(kbase_va_alloc); + SYMBOL_CLEANUP(kbase_instr_hwcnt_enable); + SYMBOL_CLEANUP(kbase_instr_hwcnt_clear); + SYMBOL_CLEANUP(kbase_instr_hwcnt_dump_irq); + SYMBOL_CLEANUP(kbase_instr_hwcnt_dump_complete); + SYMBOL_CLEANUP(kbase_instr_hwcnt_disable); + SYMBOL_CLEANUP(kbase_va_free); + SYMBOL_CLEANUP(kbase_destroy_context); +#endif +} + +/** + * Determines whether a read should take place + * @param current_time The current time, obtained from getnstimeofday() + * @param prev_time_s The number of seconds at the previous read attempt. + * @param next_read_time_ns The time (in ns) when the next read should be allowed. + * + * Note that this function has been separated out here to allow it to be tested. + */ +static int is_read_scheduled(const struct timespec *current_time, u32 *prev_time_s, s32 *next_read_time_ns) +{ + /* If the current ns count rolls over a second, roll the next read time too. */ + if (current_time->tv_sec != *prev_time_s) + *next_read_time_ns = *next_read_time_ns - NSEC_PER_SEC; + + /* Abort the read if the next read time has not arrived. */ + if (current_time->tv_nsec < *next_read_time_ns) + return 0; + + /* Set the next read some fixed time after this one, and update the read timestamp. */ + *next_read_time_ns = current_time->tv_nsec + READ_INTERVAL_NSEC; + + *prev_time_s = current_time->tv_sec; + return 1; +} + +static int start(void) +{ +#if MALI_DDK_GATOR_API_VERSION != 3 + struct kbase_uk_hwcnt_setup setup; + unsigned long long shadersPresent = 0; + u16 bitmask[] = { 0, 0, 0, 0 }; + mali_error err; +#endif + int cnt; + +#if MALI_DDK_GATOR_API_VERSION == 3 + /* Setup HW counters */ + num_hardware_counters_enabled = 0; + + /* Declare and initialise kbase_gator_hwcnt_info structure */ + in_out_info = kmalloc(sizeof(*in_out_info), GFP_KERNEL); + for (cnt = 0; cnt < ARRAY_SIZE(in_out_info->bitmask); cnt++) + in_out_info->bitmask[cnt] = 0; + + /* Calculate enable bitmasks based on counters_enabled array */ + for (cnt = 0; cnt < number_of_hardware_counters; cnt++) { + if (counters[cnt].enabled) { + int block = GET_HW_BLOCK(cnt); + int enable_bit = GET_COUNTER_OFFSET(cnt) / 4; + + in_out_info->bitmask[block] |= (1 << enable_bit); + pr_debug("gator: Mali-Midgard: hardware counter %s selected [%d]\n", hardware_counter_names[cnt], cnt); + num_hardware_counters_enabled++; + } + } + + /* Create a kbase context for HW counters */ + if (num_hardware_counters_enabled > 0) { + if (init_symbols() > 0) { + clean_symbols(); + /* No Mali driver code entrypoints found - not a fault. */ + return 0; + } + + handles = kbase_gator_hwcnt_init_symbol(in_out_info); + + if (handles == NULL) + goto out; + + kbase_device_busy = false; + } + + return 0; +#else + /* Setup HW counters */ + num_hardware_counters_enabled = 0; + + /* Calculate enable bitmasks based on counters_enabled array */ + for (cnt = 0; cnt < number_of_hardware_counters; cnt++) { + const struct mali_counter *counter = &counters[cnt]; + + if (counter->enabled) { + int block = GET_HW_BLOCK(cnt); + int enable_bit = GET_COUNTER_OFFSET(cnt) / 4; + + bitmask[block] |= (1 << enable_bit); + pr_debug("gator: Mali-Midgard: hardware counter %s selected [%d]\n", hardware_counter_names[cnt], cnt); + num_hardware_counters_enabled++; + } + } + + /* Create a kbase context for HW counters */ + if (num_hardware_counters_enabled > 0) { + if (init_symbols() > 0) { + clean_symbols(); + /* No Mali driver code entrypoints found - not a fault. */ + return 0; + } + + kbdevice = kbase_find_device_symbol(-1); + + /* If we already got a context, fail */ + if (kbcontext) { + pr_debug("gator: Mali-Midgard: error context already present\n"); + goto out; + } + + /* kbcontext will only be valid after all the Mali symbols are loaded successfully */ + kbcontext = kbase_create_context_symbol(kbdevice); + if (!kbcontext) { + pr_debug("gator: Mali-Midgard: error creating kbase context\n"); + goto out; + } + + /* See if we can get the number of shader cores */ + shadersPresent = kbdevice->shader_present_bitmap; + shader_present_low = (unsigned long)shadersPresent; + + /* + * The amount of memory needed to store the dump (bytes) + * DUMP_SIZE = number of core groups + * * number of blocks (always 8 for midgard) + * * number of counters per block (always 64 for midgard) + * * number of bytes per counter (always 4 in midgard) + * For a Mali-Midgard with a single core group = 1 * 8 * 64 * 4 = 2048 + * For a Mali-Midgard with a dual core group = 2 * 8 * 64 * 4 = 4096 + */ +#if MALI_DDK_GATOR_API_VERSION == 1 + kernel_dump_buffer = kbase_va_alloc_symbol(kbcontext, 4096); +#elif MALI_DDK_GATOR_API_VERSION == 2 + kernel_dump_buffer = kbase_va_alloc_symbol(kbcontext, 4096, &kernel_dump_buffer_handle); +#endif + if (!kernel_dump_buffer) { + pr_debug("gator: Mali-Midgard: error trying to allocate va\n"); + goto destroy_context; + } + + setup.dump_buffer = (uintptr_t)kernel_dump_buffer; + setup.jm_bm = bitmask[JM_BLOCK]; + setup.tiler_bm = bitmask[TILER_BLOCK]; + setup.shader_bm = bitmask[SHADER_BLOCK]; + setup.mmu_l2_bm = bitmask[MMU_BLOCK]; + /* These counters do not exist on Mali-T60x */ + setup.l3_cache_bm = 0; + + /* Use kbase API to enable hardware counters and provide dump buffer */ + err = kbase_instr_hwcnt_enable_symbol(kbcontext, &setup); + if (err != MALI_ERROR_NONE) { + pr_debug("gator: Mali-Midgard: can't setup hardware counters\n"); + goto free_buffer; + } + pr_debug("gator: Mali-Midgard: hardware counters enabled\n"); + kbase_instr_hwcnt_clear_symbol(kbcontext); + pr_debug("gator: Mali-Midgard: hardware counters cleared\n"); + + kbase_device_busy = false; + } + + return 0; + +free_buffer: +#if MALI_DDK_GATOR_API_VERSION == 1 + kbase_va_free_symbol(kbcontext, kernel_dump_buffer); +#elif MALI_DDK_GATOR_API_VERSION == 2 + kbase_va_free_symbol(kbcontext, &kernel_dump_buffer_handle); +#endif + +destroy_context: + kbase_destroy_context_symbol(kbcontext); +#endif + +out: + clean_symbols(); + return -1; +} + +static void stop(void) +{ + unsigned int cnt; +#if MALI_DDK_GATOR_API_VERSION == 3 + struct kbase_gator_hwcnt_handles *temp_hand; +#else + struct kbase_context *temp_kbcontext; +#endif + + pr_debug("gator: Mali-Midgard: stop\n"); + + /* Set all counters as disabled */ + for (cnt = 0; cnt < number_of_hardware_counters; cnt++) + counters[cnt].enabled = 0; + + /* Destroy the context for HW counters */ +#if MALI_DDK_GATOR_API_VERSION == 3 + if (num_hardware_counters_enabled > 0 && handles != NULL) { + /* + * Set the global variable to NULL before destroying it, because + * other function will check this before using it. + */ + temp_hand = handles; + handles = NULL; + + kbase_gator_hwcnt_term_symbol(in_out_info, temp_hand); + + kfree(in_out_info); + +#else + if (num_hardware_counters_enabled > 0 && kbcontext != NULL) { + /* + * Set the global variable to NULL before destroying it, because + * other function will check this before using it. + */ + temp_kbcontext = kbcontext; + kbcontext = NULL; + + kbase_instr_hwcnt_disable_symbol(temp_kbcontext); + +#if MALI_DDK_GATOR_API_VERSION == 1 + kbase_va_free_symbol(temp_kbcontext, kernel_dump_buffer); +#elif MALI_DDK_GATOR_API_VERSION == 2 + kbase_va_free_symbol(temp_kbcontext, &kernel_dump_buffer_handle); +#endif + + kbase_destroy_context_symbol(temp_kbcontext); +#endif + + pr_debug("gator: Mali-Midgard: hardware counters stopped\n"); + + clean_symbols(); + } +} + +static int read_counter(const int cnt, const int len, const struct mali_counter *counter) +{ + const int block = GET_HW_BLOCK(cnt); + const int counter_offset = GET_COUNTER_OFFSET(cnt); + +#if MALI_DDK_GATOR_API_VERSION == 3 + const char *block_base_address = (char *)in_out_info->kernel_dump_buffer; + int i; + int shader_core_count = 0; + u32 value = 0; + + for (i = 0; i < in_out_info->nr_hwc_blocks; i++) { + if (block == in_out_info->hwc_layout[i]) { + value += *((u32 *)(block_base_address + (0x100 * i)) + counter_offset); + if (block == SHADER_BLOCK) + ++shader_core_count; + } + } + + if (shader_core_count > 1) + value /= shader_core_count; +#else + const char *block_base_address = (char *)kernel_dump_buffer + vithar_blocks[block]; + + /* If counter belongs to shader block need to take into account all cores */ + if (block == SHADER_BLOCK) { + int i = 0; + int shader_core_count = 0; + + value = 0; + + for (i = 0; i < 4; i++) { + if ((shader_present_low >> i) & 1) { + value += *((u32 *)(block_base_address + (0x100 * i)) + counter_offset); + shader_core_count++; + } + } + + for (i = 0; i < 4; i++) { + if ((shader_present_low >> (i+4)) & 1) { + value += *((u32 *)(block_base_address + (0x100 * i) + 0x800) + counter_offset); + shader_core_count++; + } + } + + /* Need to total by number of cores to produce an average */ + if (shader_core_count != 0) + value /= shader_core_count; + } else { + value = *((u32 *)block_base_address + counter_offset); + } +#endif + + counter_dump[len + 0] = counter->key; + counter_dump[len + 1] = value; + + return 2; +} + +static int read(int **buffer, bool sched_switch) +{ + int cnt; + int len = 0; + uint32_t success; + + struct timespec current_time; + static u32 prev_time_s; + static s32 next_read_time_ns; + + if (!on_primary_core() || sched_switch) + return 0; + + getnstimeofday(¤t_time); + + /* + * Discard reads unless a respectable time has passed. This + * reduces the load on the GPU without sacrificing accuracy on + * the Streamline display. + */ + if (!is_read_scheduled(¤t_time, &prev_time_s, &next_read_time_ns)) + return 0; + + /* + * Report the HW counters + * Only process hardware counters if at least one of the hardware counters is enabled. + */ + if (num_hardware_counters_enabled > 0) { +#if MALI_DDK_GATOR_API_VERSION != 3 + const unsigned int vithar_blocks[] = { + 0x700, /* VITHAR_JOB_MANAGER, Block 0 */ + 0x400, /* VITHAR_TILER, Block 1 */ + 0x000, /* VITHAR_SHADER_CORE, Block 2 */ + 0x500 /* VITHAR_MEMORY_SYSTEM, Block 3 */ + }; +#endif + +#if MALI_DDK_GATOR_API_VERSION == 3 + if (!handles) + return -1; + + /* Mali symbols can be called safely since a kbcontext is valid */ + if (kbase_gator_instr_hwcnt_dump_complete_symbol(handles, &success) == MALI_TRUE) { +#else + if (!kbcontext) + return -1; + + /* Mali symbols can be called safely since a kbcontext is valid */ + if (kbase_instr_hwcnt_dump_complete_symbol(kbcontext, &success) == MALI_TRUE) { +#endif + kbase_device_busy = false; + + if (success == MALI_TRUE) { + /* Cycle through hardware counters and accumulate totals */ + for (cnt = 0; cnt < number_of_hardware_counters; cnt++) { + const struct mali_counter *counter = &counters[cnt]; + + if (counter->enabled) + len += read_counter(cnt, len, counter); + } + } + } + + if (!kbase_device_busy) { + kbase_device_busy = true; +#if MALI_DDK_GATOR_API_VERSION == 3 + kbase_gator_instr_hwcnt_dump_irq_symbol(handles); +#else + kbase_instr_hwcnt_dump_irq_symbol(kbcontext); +#endif + } + } + + /* Update the buffer */ + if (buffer) + *buffer = counter_dump; + + return len; +} + +static int create_files(struct super_block *sb, struct dentry *root) +{ + unsigned int event; + /* + * Create the filesystem for all events + */ + for (event = 0; event < ARRAY_SIZE(mali_activity); event++) { + if (gator_mali_create_file_system("Midgard", mali_activity_names[event], sb, root, &mali_activity[event], NULL) != 0) + return -1; + } + + for (event = 0; event < number_of_hardware_counters; event++) { + if (gator_mali_create_file_system(mali_name, hardware_counter_names[event], sb, root, &counters[event], NULL) != 0) + return -1; + } + + return 0; +} + +static void shutdown(void) +{ +#if MALI_DDK_GATOR_API_VERSION == 3 + void (*kbase_gator_hwcnt_term_names_symbol)(void) = NULL; + int error_count = 0; +#endif + + kfree(counters); + kfree(counter_dump); + +#if MALI_DDK_GATOR_API_VERSION == 3 + SYMBOL_GET(kbase_gator_hwcnt_term_names, error_count); + + number_of_hardware_counters = -1; + hardware_counter_names = NULL; + if (kbase_gator_hwcnt_term_names_symbol != NULL) { + kbase_gator_hwcnt_term_names_symbol(); + pr_err("Released symbols\n"); + } + + SYMBOL_CLEANUP(kbase_gator_hwcnt_term_names); +#endif +} + +static struct gator_interface gator_events_mali_midgard_interface = { + .shutdown = shutdown, + .create_files = create_files, + .start = start, + .stop = stop, + .read = read +}; + +int gator_events_mali_midgard_hw_init(void) +{ +#if MALI_DDK_GATOR_API_VERSION == 3 + const char *const *(*kbase_gator_hwcnt_init_names_symbol)(uint32_t *) = NULL; + int error_count = 0; +#endif + + pr_debug("gator: Mali-Midgard: sw_counters init\n"); + +#if GATOR_TEST + test_all_is_read_scheduled(); +#endif + +#if MALI_DDK_GATOR_API_VERSION == 3 + SYMBOL_GET(kbase_gator_hwcnt_init_names, error_count); + if (error_count > 0) { + SYMBOL_CLEANUP(kbase_gator_hwcnt_init_names); + return 1; + } + + number_of_hardware_counters = -1; + hardware_counter_names = kbase_gator_hwcnt_init_names_symbol(&number_of_hardware_counters); + + SYMBOL_CLEANUP(kbase_gator_hwcnt_init_names); + + if ((hardware_counter_names == NULL) || (number_of_hardware_counters <= 0)) { + pr_err("gator: Error reading hardware counters names: got %d names\n", number_of_hardware_counters); + return -1; + } +#else + mali_name = "Midgard"; +#endif + + counters = kmalloc(sizeof(*counters)*number_of_hardware_counters, GFP_KERNEL); + counter_dump = kmalloc(sizeof(*counter_dump)*number_of_hardware_counters*2, GFP_KERNEL); + + gator_mali_initialise_counters(mali_activity, ARRAY_SIZE(mali_activity)); + gator_mali_initialise_counters(counters, number_of_hardware_counters); + + return gator_events_install(&gator_events_mali_midgard_interface); +} diff --git a/drivers/gator/gator_events_mali_midgard_hw_test.c b/drivers/gator/gator_events_mali_midgard_hw_test.c new file mode 100644 index 000000000000..31a91e1c72b2 --- /dev/null +++ b/drivers/gator/gator_events_mali_midgard_hw_test.c @@ -0,0 +1,55 @@ +/** + * Copyright (C) ARM Limited 2012-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +/** + * Test functions for mali_t600_hw code. + */ + +static int is_read_scheduled(const struct timespec *current_time, u32 *prev_time_s, s32 *next_read_time_ns); + +static int test_is_read_scheduled(u32 s, u32 ns, u32 prev_s, s32 next_ns, int expected_result, s32 expected_next_ns) +{ + struct timespec current_time; + u32 prev_time_s = prev_s; + s32 next_read_time_ns = next_ns; + + current_time.tv_sec = s; + current_time.tv_nsec = ns; + + if (is_read_scheduled(¤t_time, &prev_time_s, &next_read_time_ns) != expected_result) { + pr_err("Failed do_read(%u, %u, %u, %d): expected %d\n", s, ns, prev_s, next_ns, expected_result); + return 0; + } + + if (next_read_time_ns != expected_next_ns) { + pr_err("Failed: next_read_ns expected=%d, actual=%d\n", expected_next_ns, next_read_time_ns); + return 0; + } + + return 1; +} + +static void test_all_is_read_scheduled(void) +{ + const int HIGHEST_NS = 999999999; + int n_tests_passed = 0; + + pr_err("gator: running tests on %s\n", __FILE__); + + n_tests_passed += test_is_read_scheduled(0, 0, 0, 0, 1, READ_INTERVAL_NSEC); /* Null time */ + n_tests_passed += test_is_read_scheduled(100, 1000, 0, 0, 1, READ_INTERVAL_NSEC + 1000); /* Initial values */ + + n_tests_passed += test_is_read_scheduled(100, HIGHEST_NS, 100, HIGHEST_NS + 500, 0, HIGHEST_NS + 500); + n_tests_passed += test_is_read_scheduled(101, 0001, 100, HIGHEST_NS + 500, 0, HIGHEST_NS + 500 - NSEC_PER_SEC); + n_tests_passed += test_is_read_scheduled(101, 600, 100, HIGHEST_NS + 500 - NSEC_PER_SEC, 1, 600 + READ_INTERVAL_NSEC); + + n_tests_passed += test_is_read_scheduled(101, 600, 100, HIGHEST_NS + 500, 1, 600 + READ_INTERVAL_NSEC); + + pr_err("gator: %d tests passed\n", n_tests_passed); +} diff --git a/drivers/gator/gator_events_meminfo.c b/drivers/gator/gator_events_meminfo.c new file mode 100644 index 000000000000..c625ac5af9cd --- /dev/null +++ b/drivers/gator/gator_events_meminfo.c @@ -0,0 +1,470 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include "gator.h" + +#include <linux/hardirq.h> +#include <linux/kthread.h> +#include <linux/sched.h> +#include <linux/semaphore.h> +#include <linux/workqueue.h> +#include <trace/events/kmem.h> + +#define USE_THREAD defined(CONFIG_PREEMPT_RT_FULL) + +enum { + MEMINFO_MEMFREE, + MEMINFO_MEMUSED, + MEMINFO_BUFFERRAM, + MEMINFO_TOTAL, +}; + +enum { + PROC_SIZE, + PROC_SHARE, + PROC_TEXT, + PROC_DATA, + PROC_COUNT, +}; + +static const char * const meminfo_names[] = { + "Linux_meminfo_memfree", + "Linux_meminfo_memused", + "Linux_meminfo_bufferram", +}; + +static const char * const proc_names[] = { + "Linux_proc_statm_size", + "Linux_proc_statm_share", + "Linux_proc_statm_text", + "Linux_proc_statm_data", +}; + +static bool meminfo_global_enabled; +static ulong meminfo_enabled[MEMINFO_TOTAL]; +static ulong meminfo_keys[MEMINFO_TOTAL]; +static long long meminfo_buffer[2 * (MEMINFO_TOTAL + 2)]; +static int meminfo_length; +static bool new_data_avail; + +static bool proc_global_enabled; +static ulong proc_enabled[PROC_COUNT]; +static ulong proc_keys[PROC_COUNT]; +static DEFINE_PER_CPU(long long, proc_buffer[2 * (PROC_COUNT + 3)]); + +#if USE_THREAD + +static int gator_meminfo_func(void *data); +static bool gator_meminfo_run; +/* Initialize semaphore unlocked to initialize memory values */ +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 36) +static DECLARE_MUTEX(gator_meminfo_sem); +#else +static DEFINE_SEMAPHORE(gator_meminfo_sem); +#endif + +static void notify(void) +{ + up(&gator_meminfo_sem); +} + +#else + +static unsigned int mem_event; +static void wq_sched_handler(struct work_struct *wsptr); +DECLARE_WORK(work, wq_sched_handler); +static struct timer_list meminfo_wake_up_timer; +static void meminfo_wake_up_handler(unsigned long unused_data); + +static void notify(void) +{ + mem_event++; +} + +#endif + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 3, 0) +GATOR_DEFINE_PROBE(mm_page_free_direct, TP_PROTO(struct page *page, unsigned int order)) +#else +GATOR_DEFINE_PROBE(mm_page_free, TP_PROTO(struct page *page, unsigned int order)) +#endif +{ + notify(); +} + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 3, 0) +GATOR_DEFINE_PROBE(mm_pagevec_free, TP_PROTO(struct page *page, int cold)) +#else +GATOR_DEFINE_PROBE(mm_page_free_batched, TP_PROTO(struct page *page, int cold)) +#endif +{ + notify(); +} + +GATOR_DEFINE_PROBE(mm_page_alloc, TP_PROTO(struct page *page, unsigned int order, gfp_t gfp_flags, int migratetype)) +{ + notify(); +} + +static int gator_events_meminfo_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + int i; + + for (i = 0; i < MEMINFO_TOTAL; i++) { + dir = gatorfs_mkdir(sb, root, meminfo_names[i]); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &meminfo_enabled[i]); + gatorfs_create_ro_ulong(sb, dir, "key", &meminfo_keys[i]); + } + + for (i = 0; i < PROC_COUNT; ++i) { + dir = gatorfs_mkdir(sb, root, proc_names[i]); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &proc_enabled[i]); + gatorfs_create_ro_ulong(sb, dir, "key", &proc_keys[i]); + } + + return 0; +} + +static int gator_events_meminfo_start(void) +{ + int i; + + new_data_avail = false; + meminfo_global_enabled = 0; + for (i = 0; i < MEMINFO_TOTAL; i++) { + if (meminfo_enabled[i]) { + meminfo_global_enabled = 1; + break; + } + } + + proc_global_enabled = 0; + for (i = 0; i < PROC_COUNT; ++i) { + if (proc_enabled[i]) { + proc_global_enabled = 1; + break; + } + } + if (meminfo_enabled[MEMINFO_MEMUSED]) + proc_global_enabled = 1; + + if (meminfo_global_enabled == 0) + return 0; + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 3, 0) + if (GATOR_REGISTER_TRACE(mm_page_free_direct)) +#else + if (GATOR_REGISTER_TRACE(mm_page_free)) +#endif + goto mm_page_free_exit; +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 3, 0) + if (GATOR_REGISTER_TRACE(mm_pagevec_free)) +#else + if (GATOR_REGISTER_TRACE(mm_page_free_batched)) +#endif + goto mm_page_free_batched_exit; + if (GATOR_REGISTER_TRACE(mm_page_alloc)) + goto mm_page_alloc_exit; + +#if USE_THREAD + /* Start worker thread */ + gator_meminfo_run = true; + /* Since the mutex starts unlocked, memory values will be initialized */ + if (IS_ERR(kthread_run(gator_meminfo_func, NULL, "gator_meminfo"))) + goto kthread_run_exit; +#else + setup_timer(&meminfo_wake_up_timer, meminfo_wake_up_handler, 0); +#endif + + return 0; + +#if USE_THREAD +kthread_run_exit: + GATOR_UNREGISTER_TRACE(mm_page_alloc); +#endif +mm_page_alloc_exit: +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 3, 0) + GATOR_UNREGISTER_TRACE(mm_pagevec_free); +#else + GATOR_UNREGISTER_TRACE(mm_page_free_batched); +#endif +mm_page_free_batched_exit: +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 3, 0) + GATOR_UNREGISTER_TRACE(mm_page_free_direct); +#else + GATOR_UNREGISTER_TRACE(mm_page_free); +#endif +mm_page_free_exit: + return -1; +} + +static void gator_events_meminfo_stop(void) +{ + if (meminfo_global_enabled) { +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 3, 0) + GATOR_UNREGISTER_TRACE(mm_page_free_direct); + GATOR_UNREGISTER_TRACE(mm_pagevec_free); +#else + GATOR_UNREGISTER_TRACE(mm_page_free); + GATOR_UNREGISTER_TRACE(mm_page_free_batched); +#endif + GATOR_UNREGISTER_TRACE(mm_page_alloc); + +#if USE_THREAD + /* Stop worker thread */ + gator_meminfo_run = false; + up(&gator_meminfo_sem); +#else + del_timer_sync(&meminfo_wake_up_timer); +#endif + } +} + +static void do_read(void) +{ + struct sysinfo info; + int i, len; + unsigned long long value; + + meminfo_length = len = 0; + + si_meminfo(&info); + for (i = 0; i < MEMINFO_TOTAL; i++) { + if (meminfo_enabled[i]) { + switch (i) { + case MEMINFO_MEMFREE: + value = info.freeram * PAGE_SIZE; + break; + case MEMINFO_MEMUSED: + /* pid -1 means system wide */ + meminfo_buffer[len++] = 1; + meminfo_buffer[len++] = -1; + /* Emit value */ + meminfo_buffer[len++] = meminfo_keys[MEMINFO_MEMUSED]; + meminfo_buffer[len++] = (info.totalram - info.freeram) * PAGE_SIZE; + /* Clear pid */ + meminfo_buffer[len++] = 1; + meminfo_buffer[len++] = 0; + continue; + case MEMINFO_BUFFERRAM: + value = info.bufferram * PAGE_SIZE; + break; + default: + value = 0; + break; + } + meminfo_buffer[len++] = meminfo_keys[i]; + meminfo_buffer[len++] = value; + } + } + + meminfo_length = len; + new_data_avail = true; +} + +#if USE_THREAD + +static int gator_meminfo_func(void *data) +{ + for (;;) { + if (down_killable(&gator_meminfo_sem)) + break; + + /* Eat up any pending events */ + while (!down_trylock(&gator_meminfo_sem)) + ; + + if (!gator_meminfo_run) + break; + + do_read(); + } + + return 0; +} + +#else + +/* Must be run in process context as the kernel function si_meminfo() can sleep */ +static void wq_sched_handler(struct work_struct *wsptr) +{ + do_read(); +} + +static void meminfo_wake_up_handler(unsigned long unused_data) +{ + /* had to delay scheduling work as attempting to schedule work during the context switch is illegal in kernel versions 3.5 and greater */ + schedule_work(&work); +} + +#endif + +static int gator_events_meminfo_read(long long **buffer) +{ +#if !USE_THREAD + static unsigned int last_mem_event; +#endif + + if (!on_primary_core() || !meminfo_global_enabled) + return 0; + +#if !USE_THREAD + if (last_mem_event != mem_event) { + last_mem_event = mem_event; + mod_timer(&meminfo_wake_up_timer, jiffies + 1); + } +#endif + + if (!new_data_avail) + return 0; + + new_data_avail = false; + + if (buffer) + *buffer = meminfo_buffer; + + return meminfo_length; +} + +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 34) && LINUX_VERSION_CODE < KERNEL_VERSION(3, 4, 0) + +static inline unsigned long gator_get_mm_counter(struct mm_struct *mm, int member) +{ +#ifdef SPLIT_RSS_COUNTING + long val = atomic_long_read(&mm->rss_stat.count[member]); + + if (val < 0) + val = 0; + return (unsigned long)val; +#else +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 0, 0) + return mm->rss_stat.count[member]; +#else + return atomic_long_read(&mm->rss_stat.count[member]); +#endif +#endif +} + +#define get_mm_counter(mm, member) gator_get_mm_counter(mm, member) + +#endif + +static int gator_events_meminfo_read_proc(long long **buffer, struct task_struct *task) +{ + struct mm_struct *mm; + u64 share = 0; + int i; + long long value; + int len = 0; + int cpu = get_physical_cpu(); + long long *buf = per_cpu(proc_buffer, cpu); + + if (!proc_global_enabled) + return 0; + + /* Collect the memory stats of the process instead of the thread */ + if (task->group_leader != NULL) + task = task->group_leader; + + /* get_task_mm/mmput is not needed in this context because the task and it's mm are required as part of the sched_switch */ + mm = task->mm; + if (mm == NULL) + return 0; + + /* Derived from task_statm in fs/proc/task_mmu.c */ + if (meminfo_enabled[MEMINFO_MEMUSED] || proc_enabled[PROC_SHARE]) { + share = get_mm_counter(mm, +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 32) && LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 34) + file_rss +#else + MM_FILEPAGES +#endif + ); + } + + /* key of 1 indicates a pid */ + buf[len++] = 1; + buf[len++] = task->pid; + + for (i = 0; i < PROC_COUNT; ++i) { + if (proc_enabled[i]) { + switch (i) { + case PROC_SIZE: + value = mm->total_vm; + break; + case PROC_SHARE: + value = share; + break; + case PROC_TEXT: + value = (PAGE_ALIGN(mm->end_code) - (mm->start_code & PAGE_MASK)) >> PAGE_SHIFT; + break; + case PROC_DATA: + value = mm->total_vm - mm->shared_vm; + break; + } + + buf[len++] = proc_keys[i]; + buf[len++] = value * PAGE_SIZE; + } + } + + if (meminfo_enabled[MEMINFO_MEMUSED]) { + value = share + get_mm_counter(mm, +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 32) && LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 34) + anon_rss +#else + MM_ANONPAGES +#endif + ); + /* Send resident for this pid */ + buf[len++] = meminfo_keys[MEMINFO_MEMUSED]; + buf[len++] = value * PAGE_SIZE; + } + + /* Clear pid */ + buf[len++] = 1; + buf[len++] = 0; + + if (buffer) + *buffer = buf; + + return len; +} + +static struct gator_interface gator_events_meminfo_interface = { + .create_files = gator_events_meminfo_create_files, + .start = gator_events_meminfo_start, + .stop = gator_events_meminfo_stop, + .read64 = gator_events_meminfo_read, + .read_proc = gator_events_meminfo_read_proc, +}; + +int gator_events_meminfo_init(void) +{ + int i; + + meminfo_global_enabled = 0; + for (i = 0; i < MEMINFO_TOTAL; i++) { + meminfo_enabled[i] = 0; + meminfo_keys[i] = gator_events_get_key(); + } + + proc_global_enabled = 0; + for (i = 0; i < PROC_COUNT; ++i) { + proc_enabled[i] = 0; + proc_keys[i] = gator_events_get_key(); + } + + return gator_events_install(&gator_events_meminfo_interface); +} diff --git a/drivers/gator/gator_events_mmapped.c b/drivers/gator/gator_events_mmapped.c new file mode 100644 index 000000000000..6b2af995ed41 --- /dev/null +++ b/drivers/gator/gator_events_mmapped.c @@ -0,0 +1,209 @@ +/* + * Example events provider + * + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * Similar entries to those below must be present in the events.xml file. + * To add them to the events.xml, create an events-mmap.xml with the + * following contents and rebuild gatord: + * + * <category name="mmapped"> + * <event counter="mmapped_cnt0" title="Simulated1" name="Sine" display="maximum" class="absolute" description="Sort-of-sine"/> + * <event counter="mmapped_cnt1" title="Simulated2" name="Triangle" display="maximum" class="absolute" description="Triangular wave"/> + * <event counter="mmapped_cnt2" title="Simulated3" name="PWM" display="maximum" class="absolute" description="PWM Signal"/> + * </category> + * + * When adding custom events, be sure to do the following: + * - add any needed .c files to the gator driver Makefile + * - call gator_events_install in the events init function + * - add the init function to GATOR_EVENTS_LIST in gator_main.c + * - add a new events-*.xml file to the gator daemon and rebuild + * + * Troubleshooting: + * - verify the new events are part of events.xml, which is created when building the daemon + * - verify the new events exist at /dev/gator/events/ once gatord is launched + * - verify the counter name in the XML matches the name at /dev/gator/events + */ + +#include <linux/init.h> +#include <linux/io.h> +#include <linux/ratelimit.h> + +#include "gator.h" + +#define MMAPPED_COUNTERS_NUM 3 + +static int mmapped_global_enabled; + +static struct { + unsigned long enabled; + unsigned long key; +} mmapped_counters[MMAPPED_COUNTERS_NUM]; + +static int mmapped_buffer[MMAPPED_COUNTERS_NUM * 2]; + +static s64 prev_time; + +/* Adds mmapped_cntX directories and enabled, event, and key files to /dev/gator/events */ +static int gator_events_mmapped_create_files(struct super_block *sb, + struct dentry *root) +{ + int i; + + for (i = 0; i < MMAPPED_COUNTERS_NUM; i++) { + char buf[16]; + struct dentry *dir; + + snprintf(buf, sizeof(buf), "mmapped_cnt%d", i); + dir = gatorfs_mkdir(sb, root, buf); + if (WARN_ON(!dir)) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", + &mmapped_counters[i].enabled); + gatorfs_create_ro_ulong(sb, dir, "key", + &mmapped_counters[i].key); + } + + return 0; +} + +static int gator_events_mmapped_start(void) +{ + int i; + struct timespec ts; + + getnstimeofday(&ts); + prev_time = timespec_to_ns(&ts); + + mmapped_global_enabled = 0; + for (i = 0; i < MMAPPED_COUNTERS_NUM; i++) { + if (mmapped_counters[i].enabled) { + mmapped_global_enabled = 1; + break; + } + } + + return 0; +} + +static void gator_events_mmapped_stop(void) +{ +} + +/* This function "simulates" counters, generating values of fancy + * functions like sine or triangle... */ +static int mmapped_simulate(int counter, int delta_in_us) +{ + int result = 0; + + switch (counter) { + case 0: /* sort-of-sine */ + { + static int t; + int x; + + t += delta_in_us; + if (t > 2048000) + t = 0; + + if (t % 1024000 < 512000) + x = 512000 - (t % 512000); + else + x = t % 512000; + + result = 32 * x / 512000; + result = result * result; + + if (t < 1024000) + result = 1922 - result; + } + break; + case 1: /* triangle */ + { + static int v, d = 1; + + v = v + d * delta_in_us; + if (v < 0) { + v = 0; + d = 1; + } else if (v > 1000000) { + v = 1000000; + d = -1; + } + + result = v; + } + break; + case 2: /* PWM signal */ + { + static int dc, x, t; + + t += delta_in_us; + if (t > 1000000) + t = 0; + if (x / 1000000 != (x + delta_in_us) / 1000000) + dc = (dc + 100000) % 1000000; + x += delta_in_us; + + result = t < dc ? 0 : 10; + } + break; + } + + return result; +} + +static int gator_events_mmapped_read(int **buffer, bool sched_switch) +{ + int i; + int len = 0; + int delta_in_us; + struct timespec ts; + s64 time; + + /* System wide counters - read from one core only */ + if (!on_primary_core() || !mmapped_global_enabled) + return 0; + + getnstimeofday(&ts); + time = timespec_to_ns(&ts); + delta_in_us = (int)(time - prev_time) / 1000; + prev_time = time; + + for (i = 0; i < MMAPPED_COUNTERS_NUM; i++) { + if (mmapped_counters[i].enabled) { + mmapped_buffer[len++] = mmapped_counters[i].key; + mmapped_buffer[len++] = + mmapped_simulate(i, delta_in_us); + } + } + + if (buffer) + *buffer = mmapped_buffer; + + return len; +} + +static struct gator_interface gator_events_mmapped_interface = { + .create_files = gator_events_mmapped_create_files, + .start = gator_events_mmapped_start, + .stop = gator_events_mmapped_stop, + .read = gator_events_mmapped_read, +}; + +/* Must not be static! */ +int __init gator_events_mmapped_init(void) +{ + int i; + + for (i = 0; i < MMAPPED_COUNTERS_NUM; i++) { + mmapped_counters[i].enabled = 0; + mmapped_counters[i].key = gator_events_get_key(); + } + + return gator_events_install(&gator_events_mmapped_interface); +} diff --git a/drivers/gator/gator_events_net.c b/drivers/gator/gator_events_net.c new file mode 100644 index 000000000000..d21b4db7b77c --- /dev/null +++ b/drivers/gator/gator_events_net.c @@ -0,0 +1,172 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include "gator.h" +#include <linux/netdevice.h> +#include <linux/hardirq.h> + +#define NETRX 0 +#define NETTX 1 +#define TOTALNET 2 + +static ulong netrx_enabled; +static ulong nettx_enabled; +static ulong netrx_key; +static ulong nettx_key; +static int rx_total, tx_total; +static ulong netPrev[TOTALNET]; +static int netGet[TOTALNET * 4]; + +static struct timer_list net_wake_up_timer; + +/* Must be run in process context as the kernel function dev_get_stats() can sleep */ +static void get_network_stats(struct work_struct *wsptr) +{ + int rx = 0, tx = 0; + struct net_device *dev; + + for_each_netdev(&init_net, dev) { +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 36) + const struct net_device_stats *stats = dev_get_stats(dev); +#else + struct rtnl_link_stats64 temp; + const struct rtnl_link_stats64 *stats = dev_get_stats(dev, &temp); +#endif + rx += stats->rx_bytes; + tx += stats->tx_bytes; + } + rx_total = rx; + tx_total = tx; +} + +DECLARE_WORK(wq_get_stats, get_network_stats); + +static void net_wake_up_handler(unsigned long unused_data) +{ + /* had to delay scheduling work as attempting to schedule work during the context switch is illegal in kernel versions 3.5 and greater */ + schedule_work(&wq_get_stats); +} + +static void calculate_delta(int *rx, int *tx) +{ + int rx_calc, tx_calc; + + rx_calc = (int)(rx_total - netPrev[NETRX]); + if (rx_calc < 0) + rx_calc = 0; + netPrev[NETRX] += rx_calc; + + tx_calc = (int)(tx_total - netPrev[NETTX]); + if (tx_calc < 0) + tx_calc = 0; + netPrev[NETTX] += tx_calc; + + *rx = rx_calc; + *tx = tx_calc; +} + +static int gator_events_net_create_files(struct super_block *sb, struct dentry *root) +{ + /* Network counters are not currently supported in RT-Preempt full because mod_timer is used */ +#ifndef CONFIG_PREEMPT_RT_FULL + struct dentry *dir; + + dir = gatorfs_mkdir(sb, root, "Linux_net_rx"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &netrx_enabled); + gatorfs_create_ro_ulong(sb, dir, "key", &netrx_key); + + dir = gatorfs_mkdir(sb, root, "Linux_net_tx"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &nettx_enabled); + gatorfs_create_ro_ulong(sb, dir, "key", &nettx_key); +#endif + + return 0; +} + +static int gator_events_net_start(void) +{ + get_network_stats(0); + netPrev[NETRX] = rx_total; + netPrev[NETTX] = tx_total; +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 36) + setup_timer(&net_wake_up_timer, net_wake_up_handler, 0); +#else + setup_deferrable_timer_on_stack(&net_wake_up_timer, net_wake_up_handler, 0); +#endif + return 0; +} + +static void gator_events_net_stop(void) +{ + del_timer_sync(&net_wake_up_timer); + netrx_enabled = 0; + nettx_enabled = 0; +} + +static int gator_events_net_read(int **buffer, bool sched_switch) +{ + int len, rx_delta, tx_delta; + static int last_rx_delta, last_tx_delta; + + if (!on_primary_core()) + return 0; + + if (!netrx_enabled && !nettx_enabled) + return 0; + + mod_timer(&net_wake_up_timer, jiffies + 1); + + calculate_delta(&rx_delta, &tx_delta); + + len = 0; + if (netrx_enabled && last_rx_delta != rx_delta) { + last_rx_delta = rx_delta; + netGet[len++] = netrx_key; + /* indicates to Streamline that rx_delta bytes were transmitted now, not since the last message */ + netGet[len++] = 0; + netGet[len++] = netrx_key; + netGet[len++] = rx_delta; + } + + if (nettx_enabled && last_tx_delta != tx_delta) { + last_tx_delta = tx_delta; + netGet[len++] = nettx_key; + /* indicates to Streamline that tx_delta bytes were transmitted now, not since the last message */ + netGet[len++] = 0; + netGet[len++] = nettx_key; + netGet[len++] = tx_delta; + } + + if (buffer) + *buffer = netGet; + + return len; +} + +static struct gator_interface gator_events_net_interface = { + .create_files = gator_events_net_create_files, + .start = gator_events_net_start, + .stop = gator_events_net_stop, + .read = gator_events_net_read, +}; + +int gator_events_net_init(void) +{ + netrx_key = gator_events_get_key(); + nettx_key = gator_events_get_key(); + + netrx_enabled = 0; + nettx_enabled = 0; + + return gator_events_install(&gator_events_net_interface); +} diff --git a/drivers/gator/gator_events_perf_pmu.c b/drivers/gator/gator_events_perf_pmu.c new file mode 100644 index 000000000000..47cf278e508b --- /dev/null +++ b/drivers/gator/gator_events_perf_pmu.c @@ -0,0 +1,574 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include "gator.h" + +/* gator_events_armvX.c is used for Linux 2.6.x */ +#if GATOR_PERF_PMU_SUPPORT + +#include <linux/io.h> +#ifdef CONFIG_OF +#include <linux/of_address.h> +#endif +#include <linux/perf_event.h> +#include <linux/slab.h> + +extern bool event_based_sampling; + +/* Maximum number of per-core counters - currently reserves enough space for two full hardware PMUs for big.LITTLE */ +#define CNTMAX 16 +#define CCI_400 4 +#define CCN_5XX 8 +/* Maximum number of uncore counters */ +/* + 1 for the cci-400 cycles counter */ +/* + 1 for the CCN-5xx cycles counter */ +#define UCCNT (CCI_400 + 1 + CCN_5XX + 1) + +/* Default to 0 if unable to probe the revision which was the previous behavior */ +#define DEFAULT_CCI_REVISION 0 + +/* A gator_attr is needed for every counter */ +struct gator_attr { + /* Set once in gator_events_perf_pmu_*_init - the name of the event in the gatorfs */ + char name[40]; + /* Exposed in gatorfs - set by gatord to enable this counter */ + unsigned long enabled; + /* Set once in gator_events_perf_pmu_*_init - the perf type to use, see perf_type_id in the perf_event.h header file. */ + unsigned long type; + /* Exposed in gatorfs - set by gatord to select the event to collect */ + unsigned long event; + /* Exposed in gatorfs - set by gatord with the sample period to use and enable EBS for this counter */ + unsigned long count; + /* Exposed as read only in gatorfs - set once in __attr_init as the key to use in the APC data */ + unsigned long key; +}; + +/* Per-core counter attributes */ +static struct gator_attr attrs[CNTMAX]; +/* Number of initialized per-core counters */ +static int attr_count; +/* Uncore counter attributes */ +static struct gator_attr uc_attrs[UCCNT]; +/* Number of initialized uncore counters */ +static int uc_attr_count; + +struct gator_event { + int curr; + int prev; + int prev_delta; + bool zero; + struct perf_event *pevent; + struct perf_event_attr *pevent_attr; +}; + +static DEFINE_PER_CPU(struct gator_event[CNTMAX], events); +static struct gator_event uc_events[UCCNT]; +static DEFINE_PER_CPU(int[(CNTMAX + UCCNT)*2], perf_cnt); + +static void gator_events_perf_pmu_stop(void); + +static int __create_files(struct super_block *sb, struct dentry *root, struct gator_attr *const attr) +{ + struct dentry *dir; + + if (attr->name[0] == '\0') + return 0; + dir = gatorfs_mkdir(sb, root, attr->name); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &attr->enabled); + gatorfs_create_ulong(sb, dir, "count", &attr->count); + gatorfs_create_ro_ulong(sb, dir, "key", &attr->key); + gatorfs_create_ulong(sb, dir, "event", &attr->event); + + return 0; +} + +static int gator_events_perf_pmu_create_files(struct super_block *sb, struct dentry *root) +{ + int cnt; + + for (cnt = 0; cnt < attr_count; cnt++) { + if (__create_files(sb, root, &attrs[cnt]) != 0) + return -1; + } + + for (cnt = 0; cnt < uc_attr_count; cnt++) { + if (__create_files(sb, root, &uc_attrs[cnt]) != 0) + return -1; + } + + return 0; +} + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 1, 0) +static void ebs_overflow_handler(struct perf_event *event, int unused, struct perf_sample_data *data, struct pt_regs *regs) +#else +static void ebs_overflow_handler(struct perf_event *event, struct perf_sample_data *data, struct pt_regs *regs) +#endif +{ + gator_backtrace_handler(regs); +} + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 1, 0) +static void dummy_handler(struct perf_event *event, int unused, struct perf_sample_data *data, struct pt_regs *regs) +#else +static void dummy_handler(struct perf_event *event, struct perf_sample_data *data, struct pt_regs *regs) +#endif +{ + /* Required as perf_event_create_kernel_counter() requires an overflow handler, even though all we do is poll */ +} + +static int gator_events_perf_pmu_read(int **buffer, bool sched_switch); + +static int gator_events_perf_pmu_online(int **buffer, bool migrate) +{ + return gator_events_perf_pmu_read(buffer, false); +} + +static void __online_dispatch(int cpu, bool migrate, struct gator_attr *const attr, struct gator_event *const event) +{ + perf_overflow_handler_t handler; + + event->zero = true; + + if (event->pevent != NULL || event->pevent_attr == 0 || migrate) + return; + + if (attr->count > 0) + handler = ebs_overflow_handler; + else + handler = dummy_handler; + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 1, 0) + event->pevent = perf_event_create_kernel_counter(event->pevent_attr, cpu, 0, handler); +#else + event->pevent = perf_event_create_kernel_counter(event->pevent_attr, cpu, 0, handler, 0); +#endif + if (IS_ERR(event->pevent)) { + pr_debug("gator: unable to online a counter on cpu %d\n", cpu); + event->pevent = NULL; + return; + } + + if (event->pevent->state != PERF_EVENT_STATE_ACTIVE) { + pr_debug("gator: inactive counter on cpu %d\n", cpu); + perf_event_release_kernel(event->pevent); + event->pevent = NULL; + return; + } +} + +static void gator_events_perf_pmu_online_dispatch(int cpu, bool migrate) +{ + int cnt; + + cpu = pcpu_to_lcpu(cpu); + + for (cnt = 0; cnt < attr_count; cnt++) + __online_dispatch(cpu, migrate, &attrs[cnt], &per_cpu(events, cpu)[cnt]); + + if (cpu == 0) { + for (cnt = 0; cnt < uc_attr_count; cnt++) + __online_dispatch(cpu, migrate, &uc_attrs[cnt], &uc_events[cnt]); + } +} + +static void __offline_dispatch(int cpu, struct gator_event *const event) +{ + struct perf_event *pe = NULL; + + if (event->pevent) { + pe = event->pevent; + event->pevent = NULL; + } + + if (pe) + perf_event_release_kernel(pe); +} + +static void gator_events_perf_pmu_offline_dispatch(int cpu, bool migrate) +{ + int cnt; + + if (migrate) + return; + cpu = pcpu_to_lcpu(cpu); + + for (cnt = 0; cnt < attr_count; cnt++) + __offline_dispatch(cpu, &per_cpu(events, cpu)[cnt]); + + if (cpu == 0) { + for (cnt = 0; cnt < uc_attr_count; cnt++) + __offline_dispatch(cpu, &uc_events[cnt]); + } +} + +static int __check_ebs(struct gator_attr *const attr) +{ + if (attr->count > 0) { + if (!event_based_sampling) { + event_based_sampling = true; + } else { + pr_warning("gator: Only one ebs counter is allowed\n"); + return -1; + } + } + + return 0; +} + +static int __start(struct gator_attr *const attr, struct gator_event *const event) +{ + u32 size = sizeof(struct perf_event_attr); + + event->pevent = NULL; + /* Skip disabled counters */ + if (!attr->enabled) + return 0; + + event->prev = 0; + event->curr = 0; + event->prev_delta = 0; + event->pevent_attr = kmalloc(size, GFP_KERNEL); + if (!event->pevent_attr) { + gator_events_perf_pmu_stop(); + return -1; + } + + memset(event->pevent_attr, 0, size); + event->pevent_attr->type = attr->type; + event->pevent_attr->size = size; + event->pevent_attr->config = attr->event; + event->pevent_attr->sample_period = attr->count; + event->pevent_attr->pinned = 1; + + return 0; +} + +static int gator_events_perf_pmu_start(void) +{ + int cnt, cpu; + + event_based_sampling = false; + for (cnt = 0; cnt < attr_count; cnt++) { + if (__check_ebs(&attrs[cnt]) != 0) + return -1; + } + + for (cnt = 0; cnt < uc_attr_count; cnt++) { + if (__check_ebs(&uc_attrs[cnt]) != 0) + return -1; + } + + for_each_present_cpu(cpu) { + for (cnt = 0; cnt < attr_count; cnt++) { + if (__start(&attrs[cnt], &per_cpu(events, cpu)[cnt]) != 0) + return -1; + } + } + + for (cnt = 0; cnt < uc_attr_count; cnt++) { + if (__start(&uc_attrs[cnt], &uc_events[cnt]) != 0) + return -1; + } + + return 0; +} + +static void __event_stop(struct gator_event *const event) +{ + kfree(event->pevent_attr); + event->pevent_attr = NULL; +} + +static void __attr_stop(struct gator_attr *const attr) +{ + attr->enabled = 0; + attr->event = 0; + attr->count = 0; +} + +static void gator_events_perf_pmu_stop(void) +{ + unsigned int cnt, cpu; + + for_each_present_cpu(cpu) { + for (cnt = 0; cnt < attr_count; cnt++) + __event_stop(&per_cpu(events, cpu)[cnt]); + } + + for (cnt = 0; cnt < uc_attr_count; cnt++) + __event_stop(&uc_events[cnt]); + + for (cnt = 0; cnt < attr_count; cnt++) + __attr_stop(&attrs[cnt]); + + for (cnt = 0; cnt < uc_attr_count; cnt++) + __attr_stop(&uc_attrs[cnt]); +} + +static void __read(int *const len, int cpu, struct gator_attr *const attr, struct gator_event *const event) +{ + int delta; + struct perf_event *const ev = event->pevent; + + if (ev != NULL && ev->state == PERF_EVENT_STATE_ACTIVE) { + /* After creating the perf counter in __online_dispatch, there + * is a race condition between gator_events_perf_pmu_online and + * gator_events_perf_pmu_read. So have + * gator_events_perf_pmu_online call gator_events_perf_pmu_read + * and in __read check to see if it's the first call after + * __online_dispatch and if so, run the online code. + */ + if (event->zero) { + ev->pmu->read(ev); + event->prev = event->curr = local64_read(&ev->count); + event->prev_delta = 0; + per_cpu(perf_cnt, cpu)[(*len)++] = attr->key; + per_cpu(perf_cnt, cpu)[(*len)++] = 0; + event->zero = false; + } else { + ev->pmu->read(ev); + event->curr = local64_read(&ev->count); + delta = event->curr - event->prev; + if (delta != 0 || delta != event->prev_delta) { + event->prev_delta = delta; + event->prev = event->curr; + per_cpu(perf_cnt, cpu)[(*len)++] = attr->key; + if (delta < 0) + delta *= -1; + per_cpu(perf_cnt, cpu)[(*len)++] = delta; + } + } + } +} + +static int gator_events_perf_pmu_read(int **buffer, bool sched_switch) +{ + int cnt, len = 0; + const int cpu = get_logical_cpu(); + + for (cnt = 0; cnt < attr_count; cnt++) + __read(&len, cpu, &attrs[cnt], &per_cpu(events, cpu)[cnt]); + + if (cpu == 0) { + for (cnt = 0; cnt < uc_attr_count; cnt++) + __read(&len, cpu, &uc_attrs[cnt], &uc_events[cnt]); + } + + if (buffer) + *buffer = per_cpu(perf_cnt, cpu); + + return len; +} + +static struct gator_interface gator_events_perf_pmu_interface = { + .create_files = gator_events_perf_pmu_create_files, + .start = gator_events_perf_pmu_start, + .stop = gator_events_perf_pmu_stop, + .online = gator_events_perf_pmu_online, + .online_dispatch = gator_events_perf_pmu_online_dispatch, + .offline_dispatch = gator_events_perf_pmu_offline_dispatch, + .read = gator_events_perf_pmu_read, +}; + +static void __attr_init(struct gator_attr *const attr) +{ + attr->name[0] = '\0'; + attr->enabled = 0; + attr->type = 0; + attr->event = 0; + attr->count = 0; + attr->key = gator_events_get_key(); +} + +#ifdef CONFIG_OF + +static const struct of_device_id arm_cci_matches[] = { + {.compatible = "arm,cci-400" }, + {}, +}; + +static int probe_cci_revision(void) +{ + struct device_node *np; + struct resource res; + void __iomem *cci_ctrl_base; + int rev; + int ret = DEFAULT_CCI_REVISION; + + np = of_find_matching_node(NULL, arm_cci_matches); + if (!np) + return ret; + + if (of_address_to_resource(np, 0, &res)) + goto node_put; + + cci_ctrl_base = ioremap(res.start, resource_size(&res)); + + rev = (readl_relaxed(cci_ctrl_base + 0xfe8) >> 4) & 0xf; + + if (rev <= 4) + ret = 0; + else if (rev <= 6) + ret = 1; + + iounmap(cci_ctrl_base); + + node_put: + of_node_put(np); + + return ret; +} + +#else + +static int probe_cci_revision(void) +{ + return DEFAULT_CCI_REVISION; +} + +#endif + +static void gator_events_perf_pmu_uncore_init(const char *const name, const int type, const int count) +{ + int cnt; + + snprintf(uc_attrs[uc_attr_count].name, sizeof(uc_attrs[uc_attr_count].name), "%s_ccnt", name); + uc_attrs[uc_attr_count].type = type; + ++uc_attr_count; + + for (cnt = 0; cnt < count; ++cnt, ++uc_attr_count) { + struct gator_attr *const attr = &uc_attrs[uc_attr_count]; + + snprintf(attr->name, sizeof(attr->name), "%s_cnt%d", name, cnt); + attr->type = type; + } +} + +static void gator_events_perf_pmu_cci_init(const int type) +{ + const char *cci_name; + + switch (probe_cci_revision()) { + case 0: + cci_name = "CCI_400"; + break; + case 1: + cci_name = "CCI_400-r1"; + break; + default: + pr_debug("gator: unrecognized cci-400 revision\n"); + return; + } + + gator_events_perf_pmu_uncore_init(cci_name, type, CCI_400); +} + +static void gator_events_perf_pmu_cpu_init(const struct gator_cpu *const gator_cpu, const int type) +{ + int cnt; + + snprintf(attrs[attr_count].name, sizeof(attrs[attr_count].name), "%s_ccnt", gator_cpu->pmnc_name); + attrs[attr_count].type = type; + ++attr_count; + + for (cnt = 0; cnt < gator_cpu->pmnc_counters; ++cnt, ++attr_count) { + struct gator_attr *const attr = &attrs[attr_count]; + + snprintf(attr->name, sizeof(attr->name), "%s_cnt%d", gator_cpu->pmnc_name, cnt); + attr->type = type; + } +} + +int gator_events_perf_pmu_init(void) +{ + struct perf_event_attr pea; + struct perf_event *pe; + const struct gator_cpu *gator_cpu; + int type; + int cpu; + int cnt; + bool found_cpu = false; + + for (cnt = 0; cnt < CNTMAX; cnt++) + __attr_init(&attrs[cnt]); + for (cnt = 0; cnt < UCCNT; cnt++) + __attr_init(&uc_attrs[cnt]); + + memset(&pea, 0, sizeof(pea)); + pea.size = sizeof(pea); + pea.config = 0xFF; + attr_count = 0; + uc_attr_count = 0; + for (type = PERF_TYPE_MAX; type < 0x20; ++type) { + pea.type = type; + + /* A particular PMU may work on some but not all cores, so try on each core */ + pe = NULL; + for_each_present_cpu(cpu) { +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 1, 0) + pe = perf_event_create_kernel_counter(&pea, cpu, 0, dummy_handler); +#else + pe = perf_event_create_kernel_counter(&pea, cpu, 0, dummy_handler, 0); +#endif + if (!IS_ERR(pe)) + break; + } + /* Assume that valid PMUs are contiguous */ + if (IS_ERR(pe)) { + pea.config = 0xff00; +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 1, 0) + pe = perf_event_create_kernel_counter(&pea, 0, 0, dummy_handler); +#else + pe = perf_event_create_kernel_counter(&pea, 0, 0, dummy_handler, 0); +#endif + if (IS_ERR(pe)) + break; + } + + if (pe->pmu != NULL && type == pe->pmu->type) { + if (strcmp("CCI", pe->pmu->name) == 0 || strcmp("CCI_400", pe->pmu->name) == 0 || strcmp("CCI_400-r1", pe->pmu->name) == 0) { + gator_events_perf_pmu_cci_init(type); + } else if (strcmp("ccn", pe->pmu->name) == 0) { + gator_events_perf_pmu_uncore_init("ARM_CCN_5XX", type, CCN_5XX); + } else if ((gator_cpu = gator_find_cpu_by_pmu_name(pe->pmu->name)) != NULL) { + found_cpu = true; + gator_events_perf_pmu_cpu_init(gator_cpu, type); + } + /* Initialize gator_attrs for dynamic PMUs here */ + } + + perf_event_release_kernel(pe); + } + + if (!found_cpu) { + const struct gator_cpu *const gator_cpu = gator_find_cpu_by_cpuid(gator_cpuid()); + + if (gator_cpu == NULL) + return -1; + gator_events_perf_pmu_cpu_init(gator_cpu, PERF_TYPE_RAW); + } + + /* Initialize gator_attrs for non-dynamic PMUs here */ + + if (attr_count > CNTMAX) { + pr_err("gator: Too many perf counters\n"); + return -1; + } + + if (uc_attr_count > UCCNT) { + pr_err("gator: Too many perf uncore counters\n"); + return -1; + } + + return gator_events_install(&gator_events_perf_pmu_interface); +} + +#endif diff --git a/drivers/gator/gator_events_sched.c b/drivers/gator/gator_events_sched.c new file mode 100644 index 000000000000..637107d6af1d --- /dev/null +++ b/drivers/gator/gator_events_sched.c @@ -0,0 +1,113 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include "gator.h" +#include <trace/events/sched.h> + +#define SCHED_SWITCH 0 +#define SCHED_TOTAL (SCHED_SWITCH+1) + +static ulong sched_switch_enabled; +static ulong sched_switch_key; +static DEFINE_PER_CPU(int[SCHED_TOTAL], schedCnt); +static DEFINE_PER_CPU(int[SCHED_TOTAL * 2], schedGet); + +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 35) +GATOR_DEFINE_PROBE(sched_switch, TP_PROTO(struct rq *rq, struct task_struct *prev, struct task_struct *next)) +#else +GATOR_DEFINE_PROBE(sched_switch, TP_PROTO(struct task_struct *prev, struct task_struct *next)) +#endif +{ + unsigned long flags; + + /* disable interrupts to synchronize with gator_events_sched_read() + * spinlocks not needed since percpu buffers are used + */ + local_irq_save(flags); + per_cpu(schedCnt, get_physical_cpu())[SCHED_SWITCH]++; + local_irq_restore(flags); +} + +static int gator_events_sched_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + + /* switch */ + dir = gatorfs_mkdir(sb, root, "Linux_sched_switch"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &sched_switch_enabled); + gatorfs_create_ro_ulong(sb, dir, "key", &sched_switch_key); + + return 0; +} + +static int gator_events_sched_start(void) +{ + /* register tracepoints */ + if (sched_switch_enabled) + if (GATOR_REGISTER_TRACE(sched_switch)) + goto sched_switch_exit; + pr_debug("gator: registered scheduler event tracepoints\n"); + + return 0; + + /* unregister tracepoints on error */ +sched_switch_exit: + pr_err("gator: scheduler event tracepoints failed to activate, please verify that tracepoints are enabled in the linux kernel\n"); + + return -1; +} + +static void gator_events_sched_stop(void) +{ + if (sched_switch_enabled) + GATOR_UNREGISTER_TRACE(sched_switch); + pr_debug("gator: unregistered scheduler event tracepoints\n"); + + sched_switch_enabled = 0; +} + +static int gator_events_sched_read(int **buffer, bool sched_switch) +{ + unsigned long flags; + int len, value; + int cpu = get_physical_cpu(); + + len = 0; + if (sched_switch_enabled) { + local_irq_save(flags); + value = per_cpu(schedCnt, cpu)[SCHED_SWITCH]; + per_cpu(schedCnt, cpu)[SCHED_SWITCH] = 0; + local_irq_restore(flags); + per_cpu(schedGet, cpu)[len++] = sched_switch_key; + per_cpu(schedGet, cpu)[len++] = value; + } + + if (buffer) + *buffer = per_cpu(schedGet, cpu); + + return len; +} + +static struct gator_interface gator_events_sched_interface = { + .create_files = gator_events_sched_create_files, + .start = gator_events_sched_start, + .stop = gator_events_sched_stop, + .read = gator_events_sched_read, +}; + +int gator_events_sched_init(void) +{ + sched_switch_enabled = 0; + + sched_switch_key = gator_events_get_key(); + + return gator_events_install(&gator_events_sched_interface); +} diff --git a/drivers/gator/gator_events_scorpion.c b/drivers/gator/gator_events_scorpion.c new file mode 100644 index 000000000000..49219362db09 --- /dev/null +++ b/drivers/gator/gator_events_scorpion.c @@ -0,0 +1,674 @@ +/** + * Copyright (C) ARM Limited 2011-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include "gator.h" + +/* gator_events_perf_pmu.c is used if perf is supported */ +#if GATOR_NO_PERF_SUPPORT + +static const char *pmnc_name; +static int pmnc_counters; + +/* Per-CPU PMNC: config reg */ +#define PMNC_E (1 << 0) /* Enable all counters */ +#define PMNC_P (1 << 1) /* Reset all counters */ +#define PMNC_C (1 << 2) /* Cycle counter reset */ +#define PMNC_D (1 << 3) /* CCNT counts every 64th cpu cycle */ +#define PMNC_X (1 << 4) /* Export to ETM */ +#define PMNC_DP (1 << 5) /* Disable CCNT if non-invasive debug */ +#define PMNC_MASK 0x3f /* Mask for writable bits */ + +/* ccnt reg */ +#define CCNT_REG (1 << 31) + +#define CCNT 0 +#define CNT0 1 +#define CNTMAX (4+1) + +static unsigned long pmnc_enabled[CNTMAX]; +static unsigned long pmnc_event[CNTMAX]; +static unsigned long pmnc_key[CNTMAX]; + +static DEFINE_PER_CPU(int[CNTMAX * 2], perfCnt); + +enum scorpion_perf_types { + SCORPION_ICACHE_EXPL_INV = 0x4c, + SCORPION_ICACHE_MISS = 0x4d, + SCORPION_ICACHE_ACCESS = 0x4e, + SCORPION_ICACHE_CACHEREQ_L2 = 0x4f, + SCORPION_ICACHE_NOCACHE_L2 = 0x50, + SCORPION_HIQUP_NOPED = 0x51, + SCORPION_DATA_ABORT = 0x52, + SCORPION_IRQ = 0x53, + SCORPION_FIQ = 0x54, + SCORPION_ALL_EXCPT = 0x55, + SCORPION_UNDEF = 0x56, + SCORPION_SVC = 0x57, + SCORPION_SMC = 0x58, + SCORPION_PREFETCH_ABORT = 0x59, + SCORPION_INDEX_CHECK = 0x5a, + SCORPION_NULL_CHECK = 0x5b, + SCORPION_EXPL_ICIALLU = 0x5c, + SCORPION_IMPL_ICIALLU = 0x5d, + SCORPION_NONICIALLU_BTAC_INV = 0x5e, + SCORPION_ICIMVAU_IMPL_ICIALLU = 0x5f, + SCORPION_SPIPE_ONLY_CYCLES = 0x60, + SCORPION_XPIPE_ONLY_CYCLES = 0x61, + SCORPION_DUAL_CYCLES = 0x62, + SCORPION_DISPATCH_ANY_CYCLES = 0x63, + SCORPION_FIFO_FULLBLK_CMT = 0x64, + SCORPION_FAIL_COND_INST = 0x65, + SCORPION_PASS_COND_INST = 0x66, + SCORPION_ALLOW_VU_CLK = 0x67, + SCORPION_VU_IDLE = 0x68, + SCORPION_ALLOW_L2_CLK = 0x69, + SCORPION_L2_IDLE = 0x6a, + SCORPION_DTLB_IMPL_INV_SCTLR_DACR = 0x6b, + SCORPION_DTLB_EXPL_INV = 0x6c, + SCORPION_DTLB_MISS = 0x6d, + SCORPION_DTLB_ACCESS = 0x6e, + SCORPION_ITLB_MISS = 0x6f, + SCORPION_ITLB_IMPL_INV = 0x70, + SCORPION_ITLB_EXPL_INV = 0x71, + SCORPION_UTLB_D_MISS = 0x72, + SCORPION_UTLB_D_ACCESS = 0x73, + SCORPION_UTLB_I_MISS = 0x74, + SCORPION_UTLB_I_ACCESS = 0x75, + SCORPION_UTLB_INV_ASID = 0x76, + SCORPION_UTLB_INV_MVA = 0x77, + SCORPION_UTLB_INV_ALL = 0x78, + SCORPION_S2_HOLD_RDQ_UNAVAIL = 0x79, + SCORPION_S2_HOLD = 0x7a, + SCORPION_S2_HOLD_DEV_OP = 0x7b, + SCORPION_S2_HOLD_ORDER = 0x7c, + SCORPION_S2_HOLD_BARRIER = 0x7d, + SCORPION_VIU_DUAL_CYCLE = 0x7e, + SCORPION_VIU_SINGLE_CYCLE = 0x7f, + SCORPION_VX_PIPE_WAR_STALL_CYCLES = 0x80, + SCORPION_VX_PIPE_WAW_STALL_CYCLES = 0x81, + SCORPION_VX_PIPE_RAW_STALL_CYCLES = 0x82, + SCORPION_VX_PIPE_LOAD_USE_STALL = 0x83, + SCORPION_VS_PIPE_WAR_STALL_CYCLES = 0x84, + SCORPION_VS_PIPE_WAW_STALL_CYCLES = 0x85, + SCORPION_VS_PIPE_RAW_STALL_CYCLES = 0x86, + SCORPION_EXCEPTIONS_INV_OPERATION = 0x87, + SCORPION_EXCEPTIONS_DIV_BY_ZERO = 0x88, + SCORPION_COND_INST_FAIL_VX_PIPE = 0x89, + SCORPION_COND_INST_FAIL_VS_PIPE = 0x8a, + SCORPION_EXCEPTIONS_OVERFLOW = 0x8b, + SCORPION_EXCEPTIONS_UNDERFLOW = 0x8c, + SCORPION_EXCEPTIONS_DENORM = 0x8d, +#ifdef CONFIG_ARCH_MSM_SCORPIONMP + SCORPIONMP_NUM_BARRIERS = 0x8e, + SCORPIONMP_BARRIER_CYCLES = 0x8f, +#else + SCORPION_BANK_AB_HIT = 0x8e, + SCORPION_BANK_AB_ACCESS = 0x8f, + SCORPION_BANK_CD_HIT = 0x90, + SCORPION_BANK_CD_ACCESS = 0x91, + SCORPION_BANK_AB_DSIDE_HIT = 0x92, + SCORPION_BANK_AB_DSIDE_ACCESS = 0x93, + SCORPION_BANK_CD_DSIDE_HIT = 0x94, + SCORPION_BANK_CD_DSIDE_ACCESS = 0x95, + SCORPION_BANK_AB_ISIDE_HIT = 0x96, + SCORPION_BANK_AB_ISIDE_ACCESS = 0x97, + SCORPION_BANK_CD_ISIDE_HIT = 0x98, + SCORPION_BANK_CD_ISIDE_ACCESS = 0x99, + SCORPION_ISIDE_RD_WAIT = 0x9a, + SCORPION_DSIDE_RD_WAIT = 0x9b, + SCORPION_BANK_BYPASS_WRITE = 0x9c, + SCORPION_BANK_AB_NON_CASTOUT = 0x9d, + SCORPION_BANK_AB_L2_CASTOUT = 0x9e, + SCORPION_BANK_CD_NON_CASTOUT = 0x9f, + SCORPION_BANK_CD_L2_CASTOUT = 0xa0, +#endif + MSM_MAX_EVT +}; + +struct scorp_evt { + u32 evt_type; + u32 val; + u8 grp; + u32 evt_type_act; +}; + +static const struct scorp_evt sc_evt[] = { + {SCORPION_ICACHE_EXPL_INV, 0x80000500, 0, 0x4d}, + {SCORPION_ICACHE_MISS, 0x80050000, 0, 0x4e}, + {SCORPION_ICACHE_ACCESS, 0x85000000, 0, 0x4f}, + {SCORPION_ICACHE_CACHEREQ_L2, 0x86000000, 0, 0x4f}, + {SCORPION_ICACHE_NOCACHE_L2, 0x87000000, 0, 0x4f}, + {SCORPION_HIQUP_NOPED, 0x80080000, 0, 0x4e}, + {SCORPION_DATA_ABORT, 0x8000000a, 0, 0x4c}, + {SCORPION_IRQ, 0x80000a00, 0, 0x4d}, + {SCORPION_FIQ, 0x800a0000, 0, 0x4e}, + {SCORPION_ALL_EXCPT, 0x8a000000, 0, 0x4f}, + {SCORPION_UNDEF, 0x8000000b, 0, 0x4c}, + {SCORPION_SVC, 0x80000b00, 0, 0x4d}, + {SCORPION_SMC, 0x800b0000, 0, 0x4e}, + {SCORPION_PREFETCH_ABORT, 0x8b000000, 0, 0x4f}, + {SCORPION_INDEX_CHECK, 0x8000000c, 0, 0x4c}, + {SCORPION_NULL_CHECK, 0x80000c00, 0, 0x4d}, + {SCORPION_EXPL_ICIALLU, 0x8000000d, 0, 0x4c}, + {SCORPION_IMPL_ICIALLU, 0x80000d00, 0, 0x4d}, + {SCORPION_NONICIALLU_BTAC_INV, 0x800d0000, 0, 0x4e}, + {SCORPION_ICIMVAU_IMPL_ICIALLU, 0x8d000000, 0, 0x4f}, + + {SCORPION_SPIPE_ONLY_CYCLES, 0x80000600, 1, 0x51}, + {SCORPION_XPIPE_ONLY_CYCLES, 0x80060000, 1, 0x52}, + {SCORPION_DUAL_CYCLES, 0x86000000, 1, 0x53}, + {SCORPION_DISPATCH_ANY_CYCLES, 0x89000000, 1, 0x53}, + {SCORPION_FIFO_FULLBLK_CMT, 0x8000000d, 1, 0x50}, + {SCORPION_FAIL_COND_INST, 0x800d0000, 1, 0x52}, + {SCORPION_PASS_COND_INST, 0x8d000000, 1, 0x53}, + {SCORPION_ALLOW_VU_CLK, 0x8000000e, 1, 0x50}, + {SCORPION_VU_IDLE, 0x80000e00, 1, 0x51}, + {SCORPION_ALLOW_L2_CLK, 0x800e0000, 1, 0x52}, + {SCORPION_L2_IDLE, 0x8e000000, 1, 0x53}, + + {SCORPION_DTLB_IMPL_INV_SCTLR_DACR, 0x80000001, 2, 0x54}, + {SCORPION_DTLB_EXPL_INV, 0x80000100, 2, 0x55}, + {SCORPION_DTLB_MISS, 0x80010000, 2, 0x56}, + {SCORPION_DTLB_ACCESS, 0x81000000, 2, 0x57}, + {SCORPION_ITLB_MISS, 0x80000200, 2, 0x55}, + {SCORPION_ITLB_IMPL_INV, 0x80020000, 2, 0x56}, + {SCORPION_ITLB_EXPL_INV, 0x82000000, 2, 0x57}, + {SCORPION_UTLB_D_MISS, 0x80000003, 2, 0x54}, + {SCORPION_UTLB_D_ACCESS, 0x80000300, 2, 0x55}, + {SCORPION_UTLB_I_MISS, 0x80030000, 2, 0x56}, + {SCORPION_UTLB_I_ACCESS, 0x83000000, 2, 0x57}, + {SCORPION_UTLB_INV_ASID, 0x80000400, 2, 0x55}, + {SCORPION_UTLB_INV_MVA, 0x80040000, 2, 0x56}, + {SCORPION_UTLB_INV_ALL, 0x84000000, 2, 0x57}, + {SCORPION_S2_HOLD_RDQ_UNAVAIL, 0x80000800, 2, 0x55}, + {SCORPION_S2_HOLD, 0x88000000, 2, 0x57}, + {SCORPION_S2_HOLD_DEV_OP, 0x80000900, 2, 0x55}, + {SCORPION_S2_HOLD_ORDER, 0x80090000, 2, 0x56}, + {SCORPION_S2_HOLD_BARRIER, 0x89000000, 2, 0x57}, + + {SCORPION_VIU_DUAL_CYCLE, 0x80000001, 4, 0x5c}, + {SCORPION_VIU_SINGLE_CYCLE, 0x80000100, 4, 0x5d}, + {SCORPION_VX_PIPE_WAR_STALL_CYCLES, 0x80000005, 4, 0x5c}, + {SCORPION_VX_PIPE_WAW_STALL_CYCLES, 0x80000500, 4, 0x5d}, + {SCORPION_VX_PIPE_RAW_STALL_CYCLES, 0x80050000, 4, 0x5e}, + {SCORPION_VX_PIPE_LOAD_USE_STALL, 0x80000007, 4, 0x5c}, + {SCORPION_VS_PIPE_WAR_STALL_CYCLES, 0x80000008, 4, 0x5c}, + {SCORPION_VS_PIPE_WAW_STALL_CYCLES, 0x80000800, 4, 0x5d}, + {SCORPION_VS_PIPE_RAW_STALL_CYCLES, 0x80080000, 4, 0x5e}, + {SCORPION_EXCEPTIONS_INV_OPERATION, 0x8000000b, 4, 0x5c}, + {SCORPION_EXCEPTIONS_DIV_BY_ZERO, 0x80000b00, 4, 0x5d}, + {SCORPION_COND_INST_FAIL_VX_PIPE, 0x800b0000, 4, 0x5e}, + {SCORPION_COND_INST_FAIL_VS_PIPE, 0x8b000000, 4, 0x5f}, + {SCORPION_EXCEPTIONS_OVERFLOW, 0x8000000c, 4, 0x5c}, + {SCORPION_EXCEPTIONS_UNDERFLOW, 0x80000c00, 4, 0x5d}, + {SCORPION_EXCEPTIONS_DENORM, 0x8c000000, 4, 0x5f}, + +#ifdef CONFIG_ARCH_MSM_SCORPIONMP + {SCORPIONMP_NUM_BARRIERS, 0x80000e00, 3, 0x59}, + {SCORPIONMP_BARRIER_CYCLES, 0x800e0000, 3, 0x5a}, +#else + {SCORPION_BANK_AB_HIT, 0x80000001, 3, 0x58}, + {SCORPION_BANK_AB_ACCESS, 0x80000100, 3, 0x59}, + {SCORPION_BANK_CD_HIT, 0x80010000, 3, 0x5a}, + {SCORPION_BANK_CD_ACCESS, 0x81000000, 3, 0x5b}, + {SCORPION_BANK_AB_DSIDE_HIT, 0x80000002, 3, 0x58}, + {SCORPION_BANK_AB_DSIDE_ACCESS, 0x80000200, 3, 0x59}, + {SCORPION_BANK_CD_DSIDE_HIT, 0x80020000, 3, 0x5a}, + {SCORPION_BANK_CD_DSIDE_ACCESS, 0x82000000, 3, 0x5b}, + {SCORPION_BANK_AB_ISIDE_HIT, 0x80000003, 3, 0x58}, + {SCORPION_BANK_AB_ISIDE_ACCESS, 0x80000300, 3, 0x59}, + {SCORPION_BANK_CD_ISIDE_HIT, 0x80030000, 3, 0x5a}, + {SCORPION_BANK_CD_ISIDE_ACCESS, 0x83000000, 3, 0x5b}, + {SCORPION_ISIDE_RD_WAIT, 0x80000009, 3, 0x58}, + {SCORPION_DSIDE_RD_WAIT, 0x80090000, 3, 0x5a}, + {SCORPION_BANK_BYPASS_WRITE, 0x8000000a, 3, 0x58}, + {SCORPION_BANK_AB_NON_CASTOUT, 0x8000000c, 3, 0x58}, + {SCORPION_BANK_AB_L2_CASTOUT, 0x80000c00, 3, 0x59}, + {SCORPION_BANK_CD_NON_CASTOUT, 0x800c0000, 3, 0x5a}, + {SCORPION_BANK_CD_L2_CASTOUT, 0x8c000000, 3, 0x5b}, +#endif +}; + +static inline void scorpion_pmnc_write(u32 val) +{ + val &= PMNC_MASK; + asm volatile("mcr p15, 0, %0, c9, c12, 0" : : "r" (val)); +} + +static inline u32 scorpion_pmnc_read(void) +{ + u32 val; + + asm volatile("mrc p15, 0, %0, c9, c12, 0" : "=r" (val)); + return val; +} + +static inline u32 scorpion_ccnt_read(void) +{ + u32 val; + + asm volatile("mrc p15, 0, %0, c9, c13, 0" : "=r" (val)); + return val; +} + +static inline u32 scorpion_cntn_read(void) +{ + u32 val; + + asm volatile("mrc p15, 0, %0, c9, c13, 2" : "=r" (val)); + return val; +} + +static inline u32 scorpion_pmnc_enable_counter(unsigned int cnt) +{ + u32 val; + + if (cnt >= CNTMAX) { + pr_err("gator: CPU%u enabling wrong PMNC counter %d\n", smp_processor_id(), cnt); + return -1; + } + + if (cnt == CCNT) + val = CCNT_REG; + else + val = (1 << (cnt - CNT0)); + + asm volatile("mcr p15, 0, %0, c9, c12, 1" : : "r" (val)); + + return cnt; +} + +static inline u32 scorpion_pmnc_disable_counter(unsigned int cnt) +{ + u32 val; + + if (cnt >= CNTMAX) { + pr_err("gator: CPU%u disabling wrong PMNC counter %d\n", smp_processor_id(), cnt); + return -1; + } + + if (cnt == CCNT) + val = CCNT_REG; + else + val = (1 << (cnt - CNT0)); + + asm volatile("mcr p15, 0, %0, c9, c12, 2" : : "r" (val)); + + return cnt; +} + +static inline int scorpion_pmnc_select_counter(unsigned int cnt) +{ + u32 val; + + if ((cnt == CCNT) || (cnt >= CNTMAX)) { + pr_err("gator: CPU%u selecting wrong PMNC counter %d\n", smp_processor_id(), cnt); + return -1; + } + + val = (cnt - CNT0); + asm volatile("mcr p15, 0, %0, c9, c12, 5" : : "r" (val)); + + return cnt; +} + +static u32 scorpion_read_lpm0(void) +{ + u32 val; + + asm volatile("mrc p15, 0, %0, c15, c0, 0" : "=r" (val)); + return val; +} + +static void scorpion_write_lpm0(u32 val) +{ + asm volatile("mcr p15, 0, %0, c15, c0, 0" : : "r" (val)); +} + +static u32 scorpion_read_lpm1(void) +{ + u32 val; + + asm volatile("mrc p15, 1, %0, c15, c0, 0" : "=r" (val)); + return val; +} + +static void scorpion_write_lpm1(u32 val) +{ + asm volatile("mcr p15, 1, %0, c15, c0, 0" : : "r" (val)); +} + +static u32 scorpion_read_lpm2(void) +{ + u32 val; + + asm volatile("mrc p15, 2, %0, c15, c0, 0" : "=r" (val)); + return val; +} + +static void scorpion_write_lpm2(u32 val) +{ + asm volatile("mcr p15, 2, %0, c15, c0, 0" : : "r" (val)); +} + +static u32 scorpion_read_l2lpm(void) +{ + u32 val; + + asm volatile("mrc p15, 3, %0, c15, c2, 0" : "=r" (val)); + return val; +} + +static void scorpion_write_l2lpm(u32 val) +{ + asm volatile("mcr p15, 3, %0, c15, c2, 0" : : "r" (val)); +} + +static u32 scorpion_read_vlpm(void) +{ + u32 val; + + asm volatile("mrc p10, 7, %0, c11, c0, 0" : "=r" (val)); + return val; +} + +static void scorpion_write_vlpm(u32 val) +{ + asm volatile("mcr p10, 7, %0, c11, c0, 0" : : "r" (val)); +} + +struct scorpion_access_funcs { + u32 (*read)(void); + void (*write)(u32); +}; + +struct scorpion_access_funcs scor_func[] = { + {scorpion_read_lpm0, scorpion_write_lpm0}, + {scorpion_read_lpm1, scorpion_write_lpm1}, + {scorpion_read_lpm2, scorpion_write_lpm2}, + {scorpion_read_l2lpm, scorpion_write_l2lpm}, + {scorpion_read_vlpm, scorpion_write_vlpm}, +}; + +u32 venum_orig_val; +u32 fp_orig_val; + +static void scorpion_pre_vlpm(void) +{ + u32 venum_new_val; + u32 fp_new_val; + + /* CPACR Enable CP10 access */ + asm volatile("mrc p15, 0, %0, c1, c0, 2" : "=r" (venum_orig_val)); + venum_new_val = venum_orig_val | 0x00300000; + asm volatile("mcr p15, 0, %0, c1, c0, 2" : : "r" (venum_new_val)); + /* Enable FPEXC */ + asm volatile("mrc p10, 7, %0, c8, c0, 0" : "=r" (fp_orig_val)); + fp_new_val = fp_orig_val | 0x40000000; + asm volatile("mcr p10, 7, %0, c8, c0, 0" : : "r" (fp_new_val)); +} + +static void scorpion_post_vlpm(void) +{ + /* Restore FPEXC */ + asm volatile("mcr p10, 7, %0, c8, c0, 0" : : "r" (fp_orig_val)); + /* Restore CPACR */ + asm volatile("mcr p15, 0, %0, c1, c0, 2" : : "r" (venum_orig_val)); +} + +#define COLMN0MASK 0x000000ff +#define COLMN1MASK 0x0000ff00 +#define COLMN2MASK 0x00ff0000 +static u32 scorpion_get_columnmask(u32 setval) +{ + if (setval & COLMN0MASK) + return 0xffffff00; + if (setval & COLMN1MASK) + return 0xffff00ff; + if (setval & COLMN2MASK) + return 0xff00ffff; + return 0x80ffffff; +} + +static void scorpion_evt_setup(u32 gr, u32 setval) +{ + u32 val; + + if (gr == 4) + scorpion_pre_vlpm(); + val = scorpion_get_columnmask(setval) & scor_func[gr].read(); + val = val | setval; + scor_func[gr].write(val); + if (gr == 4) + scorpion_post_vlpm(); +} + +static int get_scorpion_evtinfo(unsigned int evt_type, struct scorp_evt *evtinfo) +{ + u32 idx; + + if ((evt_type < 0x4c) || (evt_type >= MSM_MAX_EVT)) + return 0; + idx = evt_type - 0x4c; + if (sc_evt[idx].evt_type == evt_type) { + evtinfo->val = sc_evt[idx].val; + evtinfo->grp = sc_evt[idx].grp; + evtinfo->evt_type_act = sc_evt[idx].evt_type_act; + return 1; + } + return 0; +} + +static inline void scorpion_pmnc_write_evtsel(unsigned int cnt, u32 val) +{ + if (scorpion_pmnc_select_counter(cnt) == cnt) { + if (val < 0x40) { + asm volatile("mcr p15, 0, %0, c9, c13, 1" : : "r" (val)); + } else { + u32 zero = 0; + struct scorp_evt evtinfo; + /* extract evtinfo.grp and evtinfo.tevt_type_act from val */ + if (get_scorpion_evtinfo(val, &evtinfo) == 0) + return; + asm volatile("mcr p15, 0, %0, c9, c13, 1" : : "r" (evtinfo.evt_type_act)); + asm volatile("mcr p15, 0, %0, c9, c15, 0" : : "r" (zero)); + scorpion_evt_setup(evtinfo.grp, val); + } + } +} + +static void scorpion_pmnc_reset_counter(unsigned int cnt) +{ + u32 val = 0; + + if (cnt == CCNT) { + scorpion_pmnc_disable_counter(cnt); + + asm volatile("mcr p15, 0, %0, c9, c13, 0" : : "r" (val)); + + if (pmnc_enabled[cnt] != 0) + scorpion_pmnc_enable_counter(cnt); + + } else if (cnt >= CNTMAX) { + pr_err("gator: CPU%u resetting wrong PMNC counter %d\n", smp_processor_id(), cnt); + } else { + scorpion_pmnc_disable_counter(cnt); + + if (scorpion_pmnc_select_counter(cnt) == cnt) + asm volatile("mcr p15, 0, %0, c9, c13, 2" : : "r" (val)); + + if (pmnc_enabled[cnt] != 0) + scorpion_pmnc_enable_counter(cnt); + } +} + +static int gator_events_scorpion_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + int i; + + for (i = 0; i < pmnc_counters; i++) { + char buf[40]; + + if (i == 0) + snprintf(buf, sizeof(buf), "%s_ccnt", pmnc_name); + else + snprintf(buf, sizeof(buf), "%s_cnt%d", pmnc_name, i - 1); + dir = gatorfs_mkdir(sb, root, buf); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &pmnc_enabled[i]); + gatorfs_create_ro_ulong(sb, dir, "key", &pmnc_key[i]); + if (i > 0) + gatorfs_create_ulong(sb, dir, "event", &pmnc_event[i]); + } + + return 0; +} + +static int gator_events_scorpion_online(int **buffer, bool migrate) +{ + unsigned int cnt, len = 0, cpu = smp_processor_id(); + + if (scorpion_pmnc_read() & PMNC_E) + scorpion_pmnc_write(scorpion_pmnc_read() & ~PMNC_E); + + /* Initialize & Reset PMNC: C bit and P bit */ + scorpion_pmnc_write(PMNC_P | PMNC_C); + + for (cnt = CCNT; cnt < CNTMAX; cnt++) { + unsigned long event; + + if (!pmnc_enabled[cnt]) + continue; + + /* disable counter */ + scorpion_pmnc_disable_counter(cnt); + + event = pmnc_event[cnt] & 255; + + /* Set event (if destined for PMNx counters), We don't need to set the event if it's a cycle count */ + if (cnt != CCNT) + scorpion_pmnc_write_evtsel(cnt, event); + + /* reset counter */ + scorpion_pmnc_reset_counter(cnt); + + /* Enable counter, do not enable interrupt for this counter */ + scorpion_pmnc_enable_counter(cnt); + } + + /* enable */ + scorpion_pmnc_write(scorpion_pmnc_read() | PMNC_E); + + /* read the counters and toss the invalid data, return zero instead */ + for (cnt = 0; cnt < pmnc_counters; cnt++) { + if (pmnc_enabled[cnt]) { + if (cnt == CCNT) + scorpion_ccnt_read(); + else if (scorpion_pmnc_select_counter(cnt) == cnt) + scorpion_cntn_read(); + scorpion_pmnc_reset_counter(cnt); + + per_cpu(perfCnt, cpu)[len++] = pmnc_key[cnt]; + per_cpu(perfCnt, cpu)[len++] = 0; + } + } + + if (buffer) + *buffer = per_cpu(perfCnt, cpu); + + return len; +} + +static int gator_events_scorpion_offline(int **buffer, bool migrate) +{ + scorpion_pmnc_write(scorpion_pmnc_read() & ~PMNC_E); + return 0; +} + +static void gator_events_scorpion_stop(void) +{ + unsigned int cnt; + + for (cnt = CCNT; cnt < CNTMAX; cnt++) { + pmnc_enabled[cnt] = 0; + pmnc_event[cnt] = 0; + } +} + +static int gator_events_scorpion_read(int **buffer, bool sched_switch) +{ + int cnt, len = 0; + int cpu = smp_processor_id(); + + /* a context switch may occur before the online hotplug event, thus need to check that the pmu is enabled */ + if (!(scorpion_pmnc_read() & PMNC_E)) + return 0; + + for (cnt = 0; cnt < pmnc_counters; cnt++) { + if (pmnc_enabled[cnt]) { + int value; + + if (cnt == CCNT) + value = scorpion_ccnt_read(); + else if (scorpion_pmnc_select_counter(cnt) == cnt) + value = scorpion_cntn_read(); + else + value = 0; + scorpion_pmnc_reset_counter(cnt); + + per_cpu(perfCnt, cpu)[len++] = pmnc_key[cnt]; + per_cpu(perfCnt, cpu)[len++] = value; + } + } + + if (buffer) + *buffer = per_cpu(perfCnt, cpu); + + return len; +} + +static struct gator_interface gator_events_scorpion_interface = { + .create_files = gator_events_scorpion_create_files, + .stop = gator_events_scorpion_stop, + .online = gator_events_scorpion_online, + .offline = gator_events_scorpion_offline, + .read = gator_events_scorpion_read, +}; + +int gator_events_scorpion_init(void) +{ + unsigned int cnt; + + switch (gator_cpuid()) { + case SCORPION: + pmnc_name = "Scorpion"; + pmnc_counters = 4; + break; + case SCORPIONMP: + pmnc_name = "ScorpionMP"; + pmnc_counters = 4; + break; + default: + return -1; + } + + /* CNT[n] + CCNT */ + pmnc_counters++; + + for (cnt = CCNT; cnt < CNTMAX; cnt++) { + pmnc_enabled[cnt] = 0; + pmnc_event[cnt] = 0; + pmnc_key[cnt] = gator_events_get_key(); + } + + return gator_events_install(&gator_events_scorpion_interface); +} + +#endif diff --git a/drivers/gator/gator_fs.c b/drivers/gator/gator_fs.c new file mode 100644 index 000000000000..d8fb357b9eda --- /dev/null +++ b/drivers/gator/gator_fs.c @@ -0,0 +1,370 @@ +/** + * @file gatorfs.c + * + * @remark Copyright 2002 OProfile authors + * @remark Read the file COPYING + * + * @author John Levon + * + * A simple filesystem for configuration and + * access of oprofile. + */ + +#include <linux/init.h> +#include <linux/module.h> +#include <linux/fs.h> +#include <linux/pagemap.h> +#include <linux/uaccess.h> + +#define gatorfs_MAGIC 0x24051020 +#define TMPBUFSIZE 50 +DEFINE_SPINLOCK(gatorfs_lock); + +static struct inode *gatorfs_get_inode(struct super_block *sb, int mode) +{ + struct inode *inode = new_inode(sb); + + if (inode) { +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 37) + inode->i_ino = get_next_ino(); +#endif + inode->i_mode = mode; + inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; + } + return inode; +} + +static const struct super_operations s_ops = { + .statfs = simple_statfs, + .drop_inode = generic_delete_inode, +}; + +static ssize_t gatorfs_ulong_to_user(unsigned long val, char __user *buf, size_t count, loff_t *offset) +{ + char tmpbuf[TMPBUFSIZE]; + size_t maxlen = snprintf(tmpbuf, TMPBUFSIZE, "%lu\n", val); + + if (maxlen > TMPBUFSIZE) + maxlen = TMPBUFSIZE; + return simple_read_from_buffer(buf, count, offset, tmpbuf, maxlen); +} + +static ssize_t gatorfs_u64_to_user(u64 val, char __user *buf, size_t count, loff_t *offset) +{ + char tmpbuf[TMPBUFSIZE]; + size_t maxlen = snprintf(tmpbuf, TMPBUFSIZE, "%llu\n", val); + + if (maxlen > TMPBUFSIZE) + maxlen = TMPBUFSIZE; + return simple_read_from_buffer(buf, count, offset, tmpbuf, maxlen); +} + +static int gatorfs_ulong_from_user(unsigned long *val, char const __user *buf, size_t count) +{ + char tmpbuf[TMPBUFSIZE]; + unsigned long flags; + + if (!count) + return 0; + + if (count > TMPBUFSIZE - 1) + return -EINVAL; + + memset(tmpbuf, 0x0, TMPBUFSIZE); + + if (copy_from_user(tmpbuf, buf, count)) + return -EFAULT; + + spin_lock_irqsave(&gatorfs_lock, flags); + *val = simple_strtoul(tmpbuf, NULL, 0); + spin_unlock_irqrestore(&gatorfs_lock, flags); + return 0; +} + +static int gatorfs_u64_from_user(u64 *val, char const __user *buf, size_t count) +{ + char tmpbuf[TMPBUFSIZE]; + unsigned long flags; + + if (!count) + return 0; + + if (count > TMPBUFSIZE - 1) + return -EINVAL; + + memset(tmpbuf, 0x0, TMPBUFSIZE); + + if (copy_from_user(tmpbuf, buf, count)) + return -EFAULT; + + spin_lock_irqsave(&gatorfs_lock, flags); + *val = simple_strtoull(tmpbuf, NULL, 0); + spin_unlock_irqrestore(&gatorfs_lock, flags); + return 0; +} + +static ssize_t ulong_read_file(struct file *file, char __user *buf, size_t count, loff_t *offset) +{ + unsigned long *val = file->private_data; + + return gatorfs_ulong_to_user(*val, buf, count, offset); +} + +static ssize_t u64_read_file(struct file *file, char __user *buf, size_t count, loff_t *offset) +{ + u64 *val = file->private_data; + + return gatorfs_u64_to_user(*val, buf, count, offset); +} + +static ssize_t ulong_write_file(struct file *file, char const __user *buf, size_t count, loff_t *offset) +{ + unsigned long *value = file->private_data; + int retval; + + if (*offset) + return -EINVAL; + + retval = gatorfs_ulong_from_user(value, buf, count); + + if (retval) + return retval; + return count; +} + +static ssize_t u64_write_file(struct file *file, char const __user *buf, size_t count, loff_t *offset) +{ + u64 *value = file->private_data; + int retval; + + if (*offset) + return -EINVAL; + + retval = gatorfs_u64_from_user(value, buf, count); + + if (retval) + return retval; + return count; +} + +static int default_open(struct inode *inode, struct file *filp) +{ + if (inode->i_private) + filp->private_data = inode->i_private; + return 0; +} + +static const struct file_operations ulong_fops = { + .read = ulong_read_file, + .write = ulong_write_file, + .open = default_open, +}; + +static const struct file_operations u64_fops = { + .read = u64_read_file, + .write = u64_write_file, + .open = default_open, +}; + +static const struct file_operations ulong_ro_fops = { + .read = ulong_read_file, + .open = default_open, +}; + +static const struct file_operations u64_ro_fops = { + .read = u64_read_file, + .open = default_open, +}; + +static struct dentry *__gatorfs_create_file(struct super_block *sb, + struct dentry *root, + char const *name, + const struct file_operations *fops, + int perm) +{ + struct dentry *dentry; + struct inode *inode; + + dentry = d_alloc_name(root, name); + if (!dentry) + return NULL; + inode = gatorfs_get_inode(sb, S_IFREG | perm); + if (!inode) { + dput(dentry); + return NULL; + } + inode->i_fop = fops; + d_add(dentry, inode); + return dentry; +} + +int gatorfs_create_ulong(struct super_block *sb, struct dentry *root, + char const *name, unsigned long *val) +{ + struct dentry *d = __gatorfs_create_file(sb, root, name, + &ulong_fops, 0644); + if (!d) + return -EFAULT; + + d->d_inode->i_private = val; + return 0; +} + +static int gatorfs_create_u64(struct super_block *sb, struct dentry *root, + char const *name, u64 *val) +{ + struct dentry *d = __gatorfs_create_file(sb, root, name, + &u64_fops, 0644); + if (!d) + return -EFAULT; + + d->d_inode->i_private = val; + return 0; +} + +int gatorfs_create_ro_ulong(struct super_block *sb, struct dentry *root, + char const *name, unsigned long *val) +{ + struct dentry *d = __gatorfs_create_file(sb, root, name, + &ulong_ro_fops, 0444); + if (!d) + return -EFAULT; + + d->d_inode->i_private = val; + return 0; +} + +static int gatorfs_create_ro_u64(struct super_block *sb, struct dentry *root, + char const *name, u64 *val) +{ + struct dentry *d = + __gatorfs_create_file(sb, root, name, &u64_ro_fops, 0444); + if (!d) + return -EFAULT; + + d->d_inode->i_private = val; + return 0; +} + +static ssize_t atomic_read_file(struct file *file, char __user *buf, size_t count, loff_t *offset) +{ + atomic_t *val = file->private_data; + + return gatorfs_ulong_to_user(atomic_read(val), buf, count, offset); +} + +static const struct file_operations atomic_ro_fops = { + .read = atomic_read_file, + .open = default_open, +}; + +static int gatorfs_create_file(struct super_block *sb, struct dentry *root, + char const *name, const struct file_operations *fops) +{ + if (!__gatorfs_create_file(sb, root, name, fops, 0644)) + return -EFAULT; + return 0; +} + +static int gatorfs_create_file_perm(struct super_block *sb, struct dentry *root, + char const *name, + const struct file_operations *fops, int perm) +{ + if (!__gatorfs_create_file(sb, root, name, fops, perm)) + return -EFAULT; + return 0; +} + +struct dentry *gatorfs_mkdir(struct super_block *sb, + struct dentry *root, char const *name) +{ + struct dentry *dentry; + struct inode *inode; + + dentry = d_alloc_name(root, name); + if (!dentry) + return NULL; + inode = gatorfs_get_inode(sb, S_IFDIR | 0755); + if (!inode) { + dput(dentry); + return NULL; + } + inode->i_op = &simple_dir_inode_operations; + inode->i_fop = &simple_dir_operations; + d_add(dentry, inode); + return dentry; +} + +static int gatorfs_fill_super(struct super_block *sb, void *data, int silent) +{ + struct inode *root_inode; + struct dentry *root_dentry; + + sb->s_blocksize = PAGE_CACHE_SIZE; + sb->s_blocksize_bits = PAGE_CACHE_SHIFT; + sb->s_magic = gatorfs_MAGIC; + sb->s_op = &s_ops; + sb->s_time_gran = 1; + + root_inode = gatorfs_get_inode(sb, S_IFDIR | 0755); + if (!root_inode) + return -ENOMEM; + root_inode->i_op = &simple_dir_inode_operations; + root_inode->i_fop = &simple_dir_operations; + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 4, 0) + root_dentry = d_alloc_root(root_inode); +#else + root_dentry = d_make_root(root_inode); +#endif + + if (!root_dentry) { +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 4, 0) + iput(root_inode); +#endif + return -ENOMEM; + } + + sb->s_root = root_dentry; + + gator_op_create_files(sb, root_dentry); + + return 0; +} + +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 39) +static int gatorfs_get_sb(struct file_system_type *fs_type, + int flags, const char *dev_name, void *data, + struct vfsmount *mnt) +{ + return get_sb_single(fs_type, flags, data, gatorfs_fill_super, mnt); +} +#else +static struct dentry *gatorfs_mount(struct file_system_type *fs_type, + int flags, const char *dev_name, void *data) +{ + return mount_nodev(fs_type, flags, data, gatorfs_fill_super); +} +#endif + +static struct file_system_type gatorfs_type = { + .owner = THIS_MODULE, + .name = "gatorfs", +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 39) + .get_sb = gatorfs_get_sb, +#else + .mount = gatorfs_mount, +#endif + + .kill_sb = kill_litter_super, +}; + +static int __init gatorfs_register(void) +{ + return register_filesystem(&gatorfs_type); +} + +static void gatorfs_unregister(void) +{ + unregister_filesystem(&gatorfs_type); +} diff --git a/drivers/gator/gator_hrtimer_gator.c b/drivers/gator/gator_hrtimer_gator.c new file mode 100644 index 000000000000..c1525e10a8da --- /dev/null +++ b/drivers/gator/gator_hrtimer_gator.c @@ -0,0 +1,80 @@ +/** + * Copyright (C) ARM Limited 2011-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +void (*callback)(void); +DEFINE_PER_CPU(struct hrtimer, percpu_hrtimer); +DEFINE_PER_CPU(ktime_t, hrtimer_expire); +DEFINE_PER_CPU(int, hrtimer_is_active); +static ktime_t profiling_interval; +static void gator_hrtimer_online(void); +static void gator_hrtimer_offline(void); + +static enum hrtimer_restart gator_hrtimer_notify(struct hrtimer *hrtimer) +{ + int cpu = get_logical_cpu(); + + hrtimer_forward(hrtimer, per_cpu(hrtimer_expire, cpu), profiling_interval); + per_cpu(hrtimer_expire, cpu) = ktime_add(per_cpu(hrtimer_expire, cpu), profiling_interval); + (*callback)(); + return HRTIMER_RESTART; +} + +static void gator_hrtimer_online(void) +{ + int cpu = get_logical_cpu(); + struct hrtimer *hrtimer = &per_cpu(percpu_hrtimer, cpu); + + if (per_cpu(hrtimer_is_active, cpu) || profiling_interval.tv64 == 0) + return; + + per_cpu(hrtimer_is_active, cpu) = 1; + hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS); + hrtimer->function = gator_hrtimer_notify; +#ifdef CONFIG_PREEMPT_RT_BASE + hrtimer->irqsafe = 1; +#endif + per_cpu(hrtimer_expire, cpu) = ktime_add(hrtimer->base->get_time(), profiling_interval); + hrtimer_start(hrtimer, per_cpu(hrtimer_expire, cpu), HRTIMER_MODE_ABS_PINNED); +} + +static void gator_hrtimer_offline(void) +{ + int cpu = get_logical_cpu(); + struct hrtimer *hrtimer = &per_cpu(percpu_hrtimer, cpu); + + if (!per_cpu(hrtimer_is_active, cpu)) + return; + + per_cpu(hrtimer_is_active, cpu) = 0; + hrtimer_cancel(hrtimer); +} + +static int gator_hrtimer_init(int interval, void (*func)(void)) +{ + int cpu; + + (callback) = (func); + + for_each_present_cpu(cpu) { + per_cpu(hrtimer_is_active, cpu) = 0; + } + + /* calculate profiling interval */ + if (interval > 0) + profiling_interval = ns_to_ktime(1000000000UL / interval); + else + profiling_interval.tv64 = 0; + + return 0; +} + +static void gator_hrtimer_shutdown(void) +{ + /* empty */ +} diff --git a/drivers/gator/gator_iks.c b/drivers/gator/gator_iks.c new file mode 100644 index 000000000000..fb78c10fd987 --- /dev/null +++ b/drivers/gator/gator_iks.c @@ -0,0 +1,197 @@ +/** + * Copyright (C) ARM Limited 2013-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#if GATOR_IKS_SUPPORT + +#include <linux/of.h> +#include <asm/bL_switcher.h> +#include <asm/smp_plat.h> +#include <trace/events/power_cpu_migrate.h> + +static bool map_cpuids; +static int mpidr_cpuids[NR_CPUS]; +static const struct gator_cpu *mpidr_cpus[NR_CPUS]; +static int __lcpu_to_pcpu[NR_CPUS]; + +static const struct gator_cpu *gator_find_cpu_by_dt_name(const char *const name) +{ + int i; + + for (i = 0; gator_cpus[i].cpuid != 0; ++i) { + const struct gator_cpu *const gator_cpu = &gator_cpus[i]; + + if (gator_cpu->dt_name != NULL && strcmp(gator_cpu->dt_name, name) == 0) + return gator_cpu; + } + + return NULL; +} + +static void calc_first_cluster_size(void) +{ + int len; + const u32 *val; + const char *compatible; + struct device_node *cn = NULL; + int mpidr_cpuids_count = 0; + + /* Zero is a valid cpuid, so initialize the array to 0xff's */ + memset(&mpidr_cpuids, 0xff, sizeof(mpidr_cpuids)); + memset(&mpidr_cpus, 0, sizeof(mpidr_cpus)); + + while ((cn = of_find_node_by_type(cn, "cpu"))) { + BUG_ON(mpidr_cpuids_count >= NR_CPUS); + + val = of_get_property(cn, "reg", &len); + if (!val || len != 4) { + pr_err("%s missing reg property\n", cn->full_name); + continue; + } + compatible = of_get_property(cn, "compatible", NULL); + if (compatible == NULL) { + pr_err("%s missing compatible property\n", cn->full_name); + continue; + } + + mpidr_cpuids[mpidr_cpuids_count] = be32_to_cpup(val); + mpidr_cpus[mpidr_cpuids_count] = gator_find_cpu_by_dt_name(compatible); + ++mpidr_cpuids_count; + } + + map_cpuids = (mpidr_cpuids_count == nr_cpu_ids); +} + +static int linearize_mpidr(int mpidr) +{ + int i; + + for (i = 0; i < nr_cpu_ids; ++i) { + if (mpidr_cpuids[i] == mpidr) + return i; + } + + BUG(); +} + +int lcpu_to_pcpu(const int lcpu) +{ + int pcpu; + + if (!map_cpuids) + return lcpu; + + BUG_ON(lcpu >= nr_cpu_ids || lcpu < 0); + pcpu = __lcpu_to_pcpu[lcpu]; + BUG_ON(pcpu >= nr_cpu_ids || pcpu < 0); + return pcpu; +} + +int pcpu_to_lcpu(const int pcpu) +{ + int lcpu; + + if (!map_cpuids) + return pcpu; + + BUG_ON(pcpu >= nr_cpu_ids || pcpu < 0); + for (lcpu = 0; lcpu < nr_cpu_ids; ++lcpu) { + if (__lcpu_to_pcpu[lcpu] == pcpu) { + BUG_ON(lcpu >= nr_cpu_ids || lcpu < 0); + return lcpu; + } + } + BUG(); +} + +static void gator_update_cpu_mapping(u32 cpu_hwid) +{ + int lcpu = smp_processor_id(); + int pcpu = linearize_mpidr(cpu_hwid & MPIDR_HWID_BITMASK); + + BUG_ON(lcpu >= nr_cpu_ids || lcpu < 0); + BUG_ON(pcpu >= nr_cpu_ids || pcpu < 0); + __lcpu_to_pcpu[lcpu] = pcpu; +} + +GATOR_DEFINE_PROBE(cpu_migrate_begin, TP_PROTO(u64 timestamp, u32 cpu_hwid)) +{ + const int cpu = get_physical_cpu(); + + gator_timer_offline((void *)1); + gator_timer_offline_dispatch(cpu, true); +} + +GATOR_DEFINE_PROBE(cpu_migrate_finish, TP_PROTO(u64 timestamp, u32 cpu_hwid)) +{ + int cpu; + + gator_update_cpu_mapping(cpu_hwid); + + /* get_physical_cpu must be called after gator_update_cpu_mapping */ + cpu = get_physical_cpu(); + gator_timer_online_dispatch(cpu, true); + gator_timer_online((void *)1); +} + +GATOR_DEFINE_PROBE(cpu_migrate_current, TP_PROTO(u64 timestamp, u32 cpu_hwid)) +{ + gator_update_cpu_mapping(cpu_hwid); +} + +static void gator_send_iks_core_names(void) +{ + int cpu; + /* Send the cpu names */ + preempt_disable(); + for (cpu = 0; cpu < nr_cpu_ids; ++cpu) { + if (mpidr_cpus[cpu] != NULL) + gator_send_core_name(cpu, mpidr_cpus[cpu]->cpuid); + } + preempt_enable(); +} + +static int gator_migrate_start(void) +{ + int retval = 0; + + if (!map_cpuids) + return retval; + + if (retval == 0) + retval = GATOR_REGISTER_TRACE(cpu_migrate_begin); + if (retval == 0) + retval = GATOR_REGISTER_TRACE(cpu_migrate_finish); + if (retval == 0) + retval = GATOR_REGISTER_TRACE(cpu_migrate_current); + if (retval == 0) { + /* Initialize the logical to physical cpu mapping */ + memset(&__lcpu_to_pcpu, 0xff, sizeof(__lcpu_to_pcpu)); + bL_switcher_trace_trigger(); + } + return retval; +} + +static void gator_migrate_stop(void) +{ + if (!map_cpuids) + return; + + GATOR_UNREGISTER_TRACE(cpu_migrate_current); + GATOR_UNREGISTER_TRACE(cpu_migrate_finish); + GATOR_UNREGISTER_TRACE(cpu_migrate_begin); +} + +#else + +#define calc_first_cluster_size() +#define gator_send_iks_core_names() +#define gator_migrate_start() 0 +#define gator_migrate_stop() + +#endif diff --git a/drivers/gator/gator_main.c b/drivers/gator/gator_main.c new file mode 100644 index 000000000000..30bf60d95286 --- /dev/null +++ b/drivers/gator/gator_main.c @@ -0,0 +1,1491 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +/* This version must match the gator daemon version */ +#define PROTOCOL_VERSION 20 +static unsigned long gator_protocol_version = PROTOCOL_VERSION; + +#include <linux/slab.h> +#include <linux/cpu.h> +#include <linux/sched.h> +#include <linux/irq.h> +#include <linux/vmalloc.h> +#include <linux/hardirq.h> +#include <linux/highmem.h> +#include <linux/pagemap.h> +#include <linux/suspend.h> +#include <linux/module.h> +#include <linux/perf_event.h> +#include <linux/utsname.h> +#include <linux/kthread.h> +#include <asm/stacktrace.h> +#include <linux/uaccess.h> + +#include "gator.h" + +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 32) +#error kernels prior to 2.6.32 are not supported +#endif + +#if defined(MODULE) && !defined(CONFIG_MODULES) +#error Cannot build a module against a kernel that does not support modules. To resolve, either rebuild the kernel to support modules or build gator as part of the kernel. +#endif + +#if !defined(CONFIG_GENERIC_TRACER) && !defined(CONFIG_TRACING) +#error gator requires the kernel to have CONFIG_GENERIC_TRACER or CONFIG_TRACING defined +#endif + +#ifndef CONFIG_PROFILING +#error gator requires the kernel to have CONFIG_PROFILING defined +#endif + +#ifndef CONFIG_HIGH_RES_TIMERS +#error gator requires the kernel to have CONFIG_HIGH_RES_TIMERS defined to support PC sampling +#endif + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 12, 0) && defined(__arm__) && defined(CONFIG_SMP) && !defined(CONFIG_LOCAL_TIMERS) +#error gator requires the kernel to have CONFIG_LOCAL_TIMERS defined on SMP systems +#endif + +#if (GATOR_PERF_SUPPORT) && (!(GATOR_PERF_PMU_SUPPORT)) +#ifndef CONFIG_PERF_EVENTS +#error gator requires the kernel to have CONFIG_PERF_EVENTS defined to support pmu hardware counters +#elif !defined CONFIG_HW_PERF_EVENTS +#error gator requires the kernel to have CONFIG_HW_PERF_EVENTS defined to support pmu hardware counters +#endif +#endif + +/****************************************************************************** + * DEFINES + ******************************************************************************/ +#define SUMMARY_BUFFER_SIZE (1*1024) +#define BACKTRACE_BUFFER_SIZE (128*1024) +#define NAME_BUFFER_SIZE (64*1024) +#define COUNTER_BUFFER_SIZE (64*1024) /* counters have the core as part of the data and the core value in the frame header may be discarded */ +#define BLOCK_COUNTER_BUFFER_SIZE (128*1024) +#define ANNOTATE_BUFFER_SIZE (128*1024) /* annotate counters have the core as part of the data and the core value in the frame header may be discarded */ +#define SCHED_TRACE_BUFFER_SIZE (128*1024) +#define IDLE_BUFFER_SIZE (32*1024) /* idle counters have the core as part of the data and the core value in the frame header may be discarded */ +#define ACTIVITY_BUFFER_SIZE (128*1024) + +#define NO_COOKIE 0U +#define UNRESOLVED_COOKIE ~0U + +#define FRAME_SUMMARY 1 +#define FRAME_BACKTRACE 2 +#define FRAME_NAME 3 +#define FRAME_COUNTER 4 +#define FRAME_BLOCK_COUNTER 5 +#define FRAME_ANNOTATE 6 +#define FRAME_SCHED_TRACE 7 +#define FRAME_IDLE 9 +#define FRAME_ACTIVITY 13 + +#define MESSAGE_END_BACKTRACE 1 + +/* Name Frame Messages */ +#define MESSAGE_COOKIE 1 +#define MESSAGE_THREAD_NAME 2 +#define MESSAGE_LINK 4 + +/* Scheduler Trace Frame Messages */ +#define MESSAGE_SCHED_SWITCH 1 +#define MESSAGE_SCHED_EXIT 2 + +/* Idle Frame Messages */ +#define MESSAGE_IDLE_ENTER 1 +#define MESSAGE_IDLE_EXIT 2 + +/* Summary Frame Messages */ +#define MESSAGE_SUMMARY 1 +#define MESSAGE_CORE_NAME 3 + +/* Activity Frame Messages */ +#define MESSAGE_SWITCH 2 +#define MESSAGE_EXIT 3 + +#define MAXSIZE_PACK32 5 +#define MAXSIZE_PACK64 10 + +#define FRAME_HEADER_SIZE 3 + +#if defined(__arm__) +#define PC_REG regs->ARM_pc +#elif defined(__aarch64__) +#define PC_REG regs->pc +#else +#define PC_REG regs->ip +#endif + +enum { + SUMMARY_BUF, + BACKTRACE_BUF, + NAME_BUF, + COUNTER_BUF, + BLOCK_COUNTER_BUF, + ANNOTATE_BUF, + SCHED_TRACE_BUF, + IDLE_BUF, + ACTIVITY_BUF, + NUM_GATOR_BUFS +}; + +/****************************************************************************** + * Globals + ******************************************************************************/ +static unsigned long gator_cpu_cores; +/* Size of the largest buffer. Effectively constant, set in gator_op_create_files */ +static unsigned long userspace_buffer_size; +static unsigned long gator_backtrace_depth; +/* How often to commit the buffers for live in nanoseconds */ +static u64 gator_live_rate; + +static unsigned long gator_started; +static u64 gator_monotonic_started; +static u64 gator_sync_time; +static u64 gator_hibernate_time; +static unsigned long gator_buffer_opened; +static unsigned long gator_timer_count; +static unsigned long gator_response_type; +static DEFINE_MUTEX(start_mutex); +static DEFINE_MUTEX(gator_buffer_mutex); + +bool event_based_sampling; + +static DECLARE_WAIT_QUEUE_HEAD(gator_buffer_wait); +static DECLARE_WAIT_QUEUE_HEAD(gator_annotate_wait); +static struct timer_list gator_buffer_wake_up_timer; +static bool gator_buffer_wake_run; +/* Initialize semaphore unlocked to initialize memory values */ +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 36) +static DECLARE_MUTEX(gator_buffer_wake_sem); +#else +static DEFINE_SEMAPHORE(gator_buffer_wake_sem); +#endif +static struct task_struct *gator_buffer_wake_thread; +static LIST_HEAD(gator_events); + +static DEFINE_PER_CPU(u64, last_timestamp); + +static bool printed_monotonic_warning; + +static u32 gator_cpuids[NR_CPUS]; +static bool sent_core_name[NR_CPUS]; + +static DEFINE_PER_CPU(bool, in_scheduler_context); + +/****************************************************************************** + * Prototypes + ******************************************************************************/ +static u64 gator_get_time(void); +static void gator_emit_perf_time(u64 time); +static void gator_op_create_files(struct super_block *sb, struct dentry *root); + +/* gator_buffer is protected by being per_cpu and by having IRQs + * disabled when writing to it. Most marshal_* calls take care of this + * except for marshal_cookie*, marshal_backtrace* and marshal_frame + * where the caller is responsible for doing so. No synchronization is + * needed with the backtrace buffer as it is per cpu and is only used + * from the hrtimer. The annotate_lock must be held when using the + * annotation buffer as it is not per cpu. collect_counters which is + * the sole writer to the block counter frame is additionally + * protected by the per cpu collecting flag. + */ + +/* Size of the buffer, must be a power of 2. Effectively constant, set in gator_op_setup. */ +static uint32_t gator_buffer_size[NUM_GATOR_BUFS]; +/* gator_buffer_size - 1, bitwise and with pos to get offset into the array. Effectively constant, set in gator_op_setup. */ +static uint32_t gator_buffer_mask[NUM_GATOR_BUFS]; +/* Read position in the buffer. Initialized to zero in gator_op_setup and incremented after bytes are read by userspace in userspace_buffer_read */ +static DEFINE_PER_CPU(int[NUM_GATOR_BUFS], gator_buffer_read); +/* Write position in the buffer. Initialized to zero in gator_op_setup and incremented after bytes are written to the buffer */ +static DEFINE_PER_CPU(int[NUM_GATOR_BUFS], gator_buffer_write); +/* Commit position in the buffer. Initialized to zero in gator_op_setup and incremented after a frame is ready to be read by userspace */ +static DEFINE_PER_CPU(int[NUM_GATOR_BUFS], gator_buffer_commit); +/* If set to false, decreases the number of bytes returned by + * buffer_bytes_available. Set in buffer_check_space if no space is + * remaining. Initialized to true in gator_op_setup. This means that + * if we run out of space, continue to report that no space is + * available until bytes are read by userspace + */ +static DEFINE_PER_CPU(int[NUM_GATOR_BUFS], buffer_space_available); +/* The buffer. Allocated in gator_op_setup */ +static DEFINE_PER_CPU(char *[NUM_GATOR_BUFS], gator_buffer); +/* The time after which the buffer should be committed for live display */ +static DEFINE_PER_CPU(u64, gator_buffer_commit_time); + +/* List of all gator events - new events must be added to this list */ +#define GATOR_EVENTS_LIST \ + GATOR_EVENT(gator_events_armv6_init) \ + GATOR_EVENT(gator_events_armv7_init) \ + GATOR_EVENT(gator_events_block_init) \ + GATOR_EVENT(gator_events_ccn504_init) \ + GATOR_EVENT(gator_events_irq_init) \ + GATOR_EVENT(gator_events_l2c310_init) \ + GATOR_EVENT(gator_events_mali_init) \ + GATOR_EVENT(gator_events_mali_midgard_hw_init) \ + GATOR_EVENT(gator_events_mali_midgard_init) \ + GATOR_EVENT(gator_events_meminfo_init) \ + GATOR_EVENT(gator_events_mmapped_init) \ + GATOR_EVENT(gator_events_net_init) \ + GATOR_EVENT(gator_events_perf_pmu_init) \ + GATOR_EVENT(gator_events_sched_init) \ + GATOR_EVENT(gator_events_scorpion_init) \ + +#define GATOR_EVENT(EVENT_INIT) __weak int EVENT_INIT(void); +GATOR_EVENTS_LIST +#undef GATOR_EVENT + +static int (*gator_events_list[])(void) = { +#define GATOR_EVENT(EVENT_INIT) EVENT_INIT, +GATOR_EVENTS_LIST +#undef GATOR_EVENT +}; + +/****************************************************************************** + * Application Includes + ******************************************************************************/ +#include "gator_fs.c" +#include "gator_buffer_write.c" +#include "gator_buffer.c" +#include "gator_marshaling.c" +#include "gator_hrtimer_gator.c" +#include "gator_cookies.c" +#include "gator_annotate.c" +#include "gator_trace_sched.c" +#include "gator_trace_power.c" +#include "gator_trace_gpu.c" +#include "gator_backtrace.c" + +/****************************************************************************** + * Misc + ******************************************************************************/ + +static const struct gator_cpu gator_cpus[] = { + { + .cpuid = ARM1136, + .core_name = "ARM1136", + .pmnc_name = "ARM_ARM11", + .dt_name = "arm,arm1136", + .pmnc_counters = 3, + }, + { + .cpuid = ARM1156, + .core_name = "ARM1156", + .pmnc_name = "ARM_ARM11", + .dt_name = "arm,arm1156", + .pmnc_counters = 3, + }, + { + .cpuid = ARM1176, + .core_name = "ARM1176", + .pmnc_name = "ARM_ARM11", + .dt_name = "arm,arm1176", + .pmnc_counters = 3, + }, + { + .cpuid = ARM11MPCORE, + .core_name = "ARM11MPCore", + .pmnc_name = "ARM_ARM11MPCore", + .dt_name = "arm,arm11mpcore", + .pmnc_counters = 3, + }, + { + .cpuid = CORTEX_A5, + .core_name = "Cortex-A5", + .pmnc_name = "ARMv7_Cortex_A5", + .dt_name = "arm,cortex-a5", + .pmnc_counters = 2, + }, + { + .cpuid = CORTEX_A7, + .core_name = "Cortex-A7", + .pmnc_name = "ARMv7_Cortex_A7", + .dt_name = "arm,cortex-a7", + .pmnc_counters = 4, + }, + { + .cpuid = CORTEX_A8, + .core_name = "Cortex-A8", + .pmnc_name = "ARMv7_Cortex_A8", + .dt_name = "arm,cortex-a8", + .pmnc_counters = 4, + }, + { + .cpuid = CORTEX_A9, + .core_name = "Cortex-A9", + .pmnc_name = "ARMv7_Cortex_A9", + .dt_name = "arm,cortex-a9", + .pmnc_counters = 6, + }, + { + .cpuid = CORTEX_A15, + .core_name = "Cortex-A15", + .pmnc_name = "ARMv7_Cortex_A15", + .dt_name = "arm,cortex-a15", + .pmnc_counters = 6, + }, + { + .cpuid = CORTEX_A17, + .core_name = "Cortex-A17", + .pmnc_name = "ARMv7_Cortex_A17", + .dt_name = "arm,cortex-a17", + .pmnc_counters = 6, + }, + { + .cpuid = SCORPION, + .core_name = "Scorpion", + .pmnc_name = "Scorpion", + .pmnc_counters = 4, + }, + { + .cpuid = SCORPIONMP, + .core_name = "ScorpionMP", + .pmnc_name = "ScorpionMP", + .pmnc_counters = 4, + }, + { + .cpuid = KRAITSIM, + .core_name = "KraitSIM", + .pmnc_name = "Krait", + .pmnc_counters = 4, + }, + { + .cpuid = KRAIT, + .core_name = "Krait", + .pmnc_name = "Krait", + .pmnc_counters = 4, + }, + { + .cpuid = KRAIT_S4_PRO, + .core_name = "Krait S4 Pro", + .pmnc_name = "Krait", + .pmnc_counters = 4, + }, + { + .cpuid = CORTEX_A53, + .core_name = "Cortex-A53", + .pmnc_name = "ARM_Cortex-A53", + .dt_name = "arm,cortex-a53", + .pmnc_counters = 6, + }, + { + .cpuid = CORTEX_A57, + .core_name = "Cortex-A57", + .pmnc_name = "ARM_Cortex-A57", + .dt_name = "arm,cortex-a57", + .pmnc_counters = 6, + }, + { + .cpuid = AARCH64, + .core_name = "AArch64", + .pmnc_name = "ARM_AArch64", + .pmnc_counters = 6, + }, + { + .cpuid = OTHER, + .core_name = "Other", + .pmnc_name = "Other", + .pmnc_counters = 6, + }, + {} +}; + +const struct gator_cpu *gator_find_cpu_by_cpuid(const u32 cpuid) +{ + int i; + + for (i = 0; gator_cpus[i].cpuid != 0; ++i) { + const struct gator_cpu *const gator_cpu = &gator_cpus[i]; + + if (gator_cpu->cpuid == cpuid) + return gator_cpu; + } + + return NULL; +} + +static const char OLD_PMU_PREFIX[] = "ARMv7 Cortex-"; +static const char NEW_PMU_PREFIX[] = "ARMv7_Cortex_"; + +const struct gator_cpu *gator_find_cpu_by_pmu_name(const char *const name) +{ + int i; + + for (i = 0; gator_cpus[i].cpuid != 0; ++i) { + const struct gator_cpu *const gator_cpu = &gator_cpus[i]; + + if (gator_cpu->pmnc_name != NULL && + /* Do the names match exactly? */ + (strcasecmp(gator_cpu->pmnc_name, name) == 0 || + /* Do these names match but have the old vs new prefix? */ + ((strncasecmp(name, OLD_PMU_PREFIX, sizeof(OLD_PMU_PREFIX) - 1) == 0 && + strncasecmp(gator_cpu->pmnc_name, NEW_PMU_PREFIX, sizeof(NEW_PMU_PREFIX) - 1) == 0 && + strcasecmp(name + sizeof(OLD_PMU_PREFIX) - 1, gator_cpu->pmnc_name + sizeof(NEW_PMU_PREFIX) - 1) == 0)))) + return gator_cpu; + } + + return NULL; +} + +u32 gator_cpuid(void) +{ +#if defined(__arm__) || defined(__aarch64__) + u32 val; +#if !defined(__aarch64__) + asm volatile("mrc p15, 0, %0, c0, c0, 0" : "=r" (val)); +#else + asm volatile("mrs %0, midr_el1" : "=r" (val)); +#endif + return (val >> 4) & 0xfff; +#else + return OTHER; +#endif +} + +static void gator_buffer_wake_up(unsigned long data) +{ + wake_up(&gator_buffer_wait); +} + +static int gator_buffer_wake_func(void *data) +{ + for (;;) { + if (down_killable(&gator_buffer_wake_sem)) + break; + + /* Eat up any pending events */ + while (!down_trylock(&gator_buffer_wake_sem)) + ; + + if (!gator_buffer_wake_run) + break; + + gator_buffer_wake_up(0); + } + + return 0; +} + +/****************************************************************************** + * Commit interface + ******************************************************************************/ +static bool buffer_commit_ready(int *cpu, int *buftype) +{ + int cpu_x, x; + + for_each_present_cpu(cpu_x) { + for (x = 0; x < NUM_GATOR_BUFS; x++) + if (per_cpu(gator_buffer_commit, cpu_x)[x] != per_cpu(gator_buffer_read, cpu_x)[x]) { + *cpu = cpu_x; + *buftype = x; + return true; + } + } + *cpu = -1; + *buftype = -1; + return false; +} + +/****************************************************************************** + * hrtimer interrupt processing + ******************************************************************************/ +static void gator_timer_interrupt(void) +{ + struct pt_regs *const regs = get_irq_regs(); + + gator_backtrace_handler(regs); +} + +void gator_backtrace_handler(struct pt_regs *const regs) +{ + u64 time = gator_get_time(); + int cpu = get_physical_cpu(); + + /* Output backtrace */ + gator_add_sample(cpu, regs, time); + + /* Collect counters */ + if (!per_cpu(collecting, cpu)) + collect_counters(time, current, false); + + /* No buffer flushing occurs during sched switch for RT-Preempt full. The block counter frame will be flushed by collect_counters, but the sched buffer needs to be explicitly flushed */ +#ifdef CONFIG_PREEMPT_RT_FULL + buffer_check(cpu, SCHED_TRACE_BUF, time); +#endif +} + +static int gator_running; + +/* This function runs in interrupt context and on the appropriate core */ +static void gator_timer_offline(void *migrate) +{ + struct gator_interface *gi; + int i, len, cpu = get_physical_cpu(); + int *buffer; + u64 time; + + gator_trace_sched_offline(); + gator_trace_power_offline(); + + if (!migrate) + gator_hrtimer_offline(); + + /* Offline any events and output counters */ + time = gator_get_time(); + if (marshal_event_header(time)) { + list_for_each_entry(gi, &gator_events, list) { + if (gi->offline) { + len = gi->offline(&buffer, migrate); + marshal_event(len, buffer); + } + } + /* Only check after writing all counters so that time and corresponding counters appear in the same frame */ + buffer_check(cpu, BLOCK_COUNTER_BUF, time); + } + + /* Flush all buffers on this core */ + for (i = 0; i < NUM_GATOR_BUFS; i++) + gator_commit_buffer(cpu, i, time); +} + +/* This function runs in interrupt context and may be running on a core other than core 'cpu' */ +static void gator_timer_offline_dispatch(int cpu, bool migrate) +{ + struct gator_interface *gi; + + list_for_each_entry(gi, &gator_events, list) { + if (gi->offline_dispatch) + gi->offline_dispatch(cpu, migrate); + } +} + +static void gator_timer_stop(void) +{ + int cpu; + + if (gator_running) { + on_each_cpu(gator_timer_offline, NULL, 1); + for_each_online_cpu(cpu) { + gator_timer_offline_dispatch(lcpu_to_pcpu(cpu), false); + } + + gator_running = 0; + gator_hrtimer_shutdown(); + } +} + +static void gator_send_core_name(const int cpu, const u32 cpuid) +{ +#if defined(__arm__) || defined(__aarch64__) + if (!sent_core_name[cpu] || (cpuid != gator_cpuids[cpu])) { + const struct gator_cpu *const gator_cpu = gator_find_cpu_by_cpuid(cpuid); + const char *core_name = NULL; + char core_name_buf[32]; + + /* Save off this cpuid */ + gator_cpuids[cpu] = cpuid; + if (gator_cpu != NULL) { + core_name = gator_cpu->core_name; + } else { + if (cpuid == -1) + snprintf(core_name_buf, sizeof(core_name_buf), "Unknown"); + else + snprintf(core_name_buf, sizeof(core_name_buf), "Unknown (0x%.3x)", cpuid); + core_name = core_name_buf; + } + + marshal_core_name(cpu, cpuid, core_name); + sent_core_name[cpu] = true; + } +#endif +} + +static void gator_read_cpuid(void *arg) +{ + gator_cpuids[get_physical_cpu()] = gator_cpuid(); +} + +/* This function runs in interrupt context and on the appropriate core */ +static void gator_timer_online(void *migrate) +{ + struct gator_interface *gi; + int len, cpu = get_physical_cpu(); + int *buffer; + u64 time; + + /* Send what is currently running on this core */ + marshal_sched_trace_switch(current->pid, 0); + + gator_trace_power_online(); + + /* online any events and output counters */ + time = gator_get_time(); + if (marshal_event_header(time)) { + list_for_each_entry(gi, &gator_events, list) { + if (gi->online) { + len = gi->online(&buffer, migrate); + marshal_event(len, buffer); + } + } + /* Only check after writing all counters so that time and corresponding counters appear in the same frame */ + buffer_check(cpu, BLOCK_COUNTER_BUF, time); + } + + if (!migrate) + gator_hrtimer_online(); + + gator_send_core_name(cpu, gator_cpuid()); +} + +/* This function runs in interrupt context and may be running on a core other than core 'cpu' */ +static void gator_timer_online_dispatch(int cpu, bool migrate) +{ + struct gator_interface *gi; + + list_for_each_entry(gi, &gator_events, list) { + if (gi->online_dispatch) + gi->online_dispatch(cpu, migrate); + } +} + +#include "gator_iks.c" + +static int gator_timer_start(unsigned long sample_rate) +{ + int cpu; + + if (gator_running) { + pr_notice("gator: already running\n"); + return 0; + } + + gator_running = 1; + + /* event based sampling trumps hr timer based sampling */ + if (event_based_sampling) + sample_rate = 0; + + if (gator_hrtimer_init(sample_rate, gator_timer_interrupt) == -1) + return -1; + + /* Send off the previously saved cpuids */ + for_each_present_cpu(cpu) { + preempt_disable(); + gator_send_core_name(cpu, gator_cpuids[cpu]); + preempt_enable(); + } + + gator_send_iks_core_names(); + for_each_online_cpu(cpu) { + gator_timer_online_dispatch(lcpu_to_pcpu(cpu), false); + } + on_each_cpu(gator_timer_online, NULL, 1); + + return 0; +} + +static u64 gator_get_time(void) +{ + struct timespec ts; + u64 timestamp; + u64 prev_timestamp; + u64 delta; + int cpu = smp_processor_id(); + + /* Match clock_gettime(CLOCK_MONOTONIC_RAW, &ts) from userspace */ + getrawmonotonic(&ts); + timestamp = timespec_to_ns(&ts); + + /* getrawmonotonic is not monotonic on all systems. Detect and + * attempt to correct these cases. up to 0.5ms delta has been seen + * on some systems, which can skew Streamline data when viewing at + * high resolution. This doesn't work well with interrupts, but that + * it's OK - the real concern is to catch big jumps in time + */ + prev_timestamp = per_cpu(last_timestamp, cpu); + if (prev_timestamp <= timestamp) { + per_cpu(last_timestamp, cpu) = timestamp; + } else { + delta = prev_timestamp - timestamp; + /* Log the error once */ + if (!printed_monotonic_warning && delta > 500000) { + pr_err("%s: getrawmonotonic is not monotonic cpu: %i delta: %lli\nSkew in Streamline data may be present at the fine zoom levels\n", __func__, cpu, delta); + printed_monotonic_warning = true; + } + timestamp = prev_timestamp; + } + + return timestamp - gator_monotonic_started; +} + +static void gator_emit_perf_time(u64 time) +{ +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 10, 0) + if (time >= gator_sync_time) { + int cpu = get_physical_cpu(); + + marshal_event_single64(0, -1, local_clock()); + gator_sync_time += NSEC_PER_SEC; + gator_commit_buffer(cpu, COUNTER_BUF, time); + } +#endif +} + +/****************************************************************************** + * cpu hotplug and pm notifiers + ******************************************************************************/ +static int __cpuinit gator_hotcpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) +{ + int cpu = lcpu_to_pcpu((long)hcpu); + + switch (action) { + case CPU_DOWN_PREPARE: + case CPU_DOWN_PREPARE_FROZEN: + smp_call_function_single(cpu, gator_timer_offline, NULL, 1); + gator_timer_offline_dispatch(cpu, false); + break; + case CPU_ONLINE: + case CPU_ONLINE_FROZEN: + gator_timer_online_dispatch(cpu, false); + smp_call_function_single(cpu, gator_timer_online, NULL, 1); + break; + } + + return NOTIFY_OK; +} + +static struct notifier_block __refdata gator_hotcpu_notifier = { + .notifier_call = gator_hotcpu_notify, +}; + +/* n.b. calling "on_each_cpu" only runs on those that are online. + * Registered linux events are not disabled, so their counters will + * continue to collect + */ +static int gator_pm_notify(struct notifier_block *nb, unsigned long event, void *dummy) +{ + int cpu; + struct timespec ts; + + switch (event) { + case PM_HIBERNATION_PREPARE: + case PM_SUSPEND_PREPARE: + unregister_hotcpu_notifier(&gator_hotcpu_notifier); + unregister_scheduler_tracepoints(); + on_each_cpu(gator_timer_offline, NULL, 1); + for_each_online_cpu(cpu) { + gator_timer_offline_dispatch(lcpu_to_pcpu(cpu), false); + } + + /* Record the wallclock hibernate time */ + getnstimeofday(&ts); + gator_hibernate_time = timespec_to_ns(&ts) - gator_get_time(); + break; + case PM_POST_HIBERNATION: + case PM_POST_SUSPEND: + /* Adjust gator_monotonic_started for the time spent sleeping, as gator_get_time does not account for it */ + if (gator_hibernate_time > 0) { + getnstimeofday(&ts); + gator_monotonic_started += gator_hibernate_time + gator_get_time() - timespec_to_ns(&ts); + gator_hibernate_time = 0; + } + + for_each_online_cpu(cpu) { + gator_timer_online_dispatch(lcpu_to_pcpu(cpu), false); + } + on_each_cpu(gator_timer_online, NULL, 1); + register_scheduler_tracepoints(); + register_hotcpu_notifier(&gator_hotcpu_notifier); + break; + } + + return NOTIFY_OK; +} + +static struct notifier_block gator_pm_notifier = { + .notifier_call = gator_pm_notify, +}; + +static int gator_notifier_start(void) +{ + int retval; + + retval = register_hotcpu_notifier(&gator_hotcpu_notifier); + if (retval == 0) + retval = register_pm_notifier(&gator_pm_notifier); + return retval; +} + +static void gator_notifier_stop(void) +{ + unregister_pm_notifier(&gator_pm_notifier); + unregister_hotcpu_notifier(&gator_hotcpu_notifier); +} + +/****************************************************************************** + * Main + ******************************************************************************/ +static void gator_summary(void) +{ + u64 timestamp, uptime; + struct timespec ts; + char uname_buf[512]; + + snprintf(uname_buf, sizeof(uname_buf), "%s %s %s %s %s GNU/Linux", utsname()->sysname, utsname()->nodename, utsname()->release, utsname()->version, utsname()->machine); + + getnstimeofday(&ts); + timestamp = timespec_to_ns(&ts); + + /* Similar to reading /proc/uptime from fs/proc/uptime.c, calculate uptime */ +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 11, 0) + { + void (*m2b)(struct timespec *ts); + + do_posix_clock_monotonic_gettime(&ts); + /* monotonic_to_bootbased is not defined for some versions of Android */ + m2b = symbol_get(monotonic_to_bootbased); + if (m2b) + m2b(&ts); + } +#else + get_monotonic_boottime(&ts); +#endif + uptime = timespec_to_ns(&ts); + + /* Disable preemption as gator_get_time calls smp_processor_id to verify time is monotonic */ + preempt_disable(); + /* Set monotonic_started to zero as gator_get_time is uptime minus monotonic_started */ + gator_monotonic_started = 0; + gator_monotonic_started = gator_get_time(); + + marshal_summary(timestamp, uptime, gator_monotonic_started, uname_buf); + gator_sync_time = 0; + gator_emit_perf_time(gator_monotonic_started); + preempt_enable(); +} + +int gator_events_install(struct gator_interface *interface) +{ + list_add_tail(&interface->list, &gator_events); + + return 0; +} + +int gator_events_get_key(void) +{ + /* key 0 is reserved as a timestamp. key 1 is reserved as the marker + * for thread specific counters. key 2 is reserved as the marker for + * core. Odd keys are assigned by the driver, even keys by the + * daemon. + */ + static int key = 3; + const int ret = key; + + key += 2; + return ret; +} + +static int gator_init(void) +{ + int i; + + calc_first_cluster_size(); + + /* events sources */ + for (i = 0; i < ARRAY_SIZE(gator_events_list); i++) + if (gator_events_list[i]) + gator_events_list[i](); + + gator_trace_sched_init(); + gator_trace_power_init(); + + return 0; +} + +static void gator_exit(void) +{ + struct gator_interface *gi; + + list_for_each_entry(gi, &gator_events, list) + if (gi->shutdown) + gi->shutdown(); +} + +static int gator_start(void) +{ + unsigned long cpu, i; + struct gator_interface *gi; + + gator_buffer_wake_run = true; + gator_buffer_wake_thread = kthread_run(gator_buffer_wake_func, NULL, "gator_bwake"); + if (IS_ERR(gator_buffer_wake_thread)) + goto bwake_failure; + + if (gator_migrate_start()) + goto migrate_failure; + + /* Initialize the buffer with the frame type and core */ + for_each_present_cpu(cpu) { + for (i = 0; i < NUM_GATOR_BUFS; i++) + marshal_frame(cpu, i); + per_cpu(last_timestamp, cpu) = 0; + } + printed_monotonic_warning = false; + + /* Capture the start time */ + gator_summary(); + + /* start all events */ + list_for_each_entry(gi, &gator_events, list) { + if (gi->start && gi->start() != 0) { + struct list_head *ptr = gi->list.prev; + + while (ptr != &gator_events) { + gi = list_entry(ptr, struct gator_interface, list); + + if (gi->stop) + gi->stop(); + + ptr = ptr->prev; + } + goto events_failure; + } + } + + /* cookies shall be initialized before trace_sched_start() and gator_timer_start() */ + if (cookies_initialize()) + goto cookies_failure; + if (gator_annotate_start()) + goto annotate_failure; + if (gator_trace_sched_start()) + goto sched_failure; + if (gator_trace_power_start()) + goto power_failure; + if (gator_trace_gpu_start()) + goto gpu_failure; + if (gator_timer_start(gator_timer_count)) + goto timer_failure; + if (gator_notifier_start()) + goto notifier_failure; + + return 0; + +notifier_failure: + gator_timer_stop(); +timer_failure: + gator_trace_gpu_stop(); +gpu_failure: + gator_trace_power_stop(); +power_failure: + gator_trace_sched_stop(); +sched_failure: + gator_annotate_stop(); +annotate_failure: + cookies_release(); +cookies_failure: + /* stop all events */ + list_for_each_entry(gi, &gator_events, list) + if (gi->stop) + gi->stop(); +events_failure: + gator_migrate_stop(); +migrate_failure: + gator_buffer_wake_run = false; + up(&gator_buffer_wake_sem); + gator_buffer_wake_thread = NULL; +bwake_failure: + + return -1; +} + +static void gator_stop(void) +{ + struct gator_interface *gi; + + gator_annotate_stop(); + gator_trace_sched_stop(); + gator_trace_power_stop(); + gator_trace_gpu_stop(); + + /* stop all interrupt callback reads before tearing down other interfaces */ + gator_notifier_stop(); /* should be called before gator_timer_stop to avoid re-enabling the hrtimer after it has been offlined */ + gator_timer_stop(); + + /* stop all events */ + list_for_each_entry(gi, &gator_events, list) + if (gi->stop) + gi->stop(); + + gator_migrate_stop(); + + gator_buffer_wake_run = false; + up(&gator_buffer_wake_sem); + gator_buffer_wake_thread = NULL; +} + +/****************************************************************************** + * Filesystem + ******************************************************************************/ +/* fopen("buffer") */ +static int gator_op_setup(void) +{ + int err = 0; + int cpu, i; + + mutex_lock(&start_mutex); + + gator_buffer_size[SUMMARY_BUF] = SUMMARY_BUFFER_SIZE; + gator_buffer_mask[SUMMARY_BUF] = SUMMARY_BUFFER_SIZE - 1; + + gator_buffer_size[BACKTRACE_BUF] = BACKTRACE_BUFFER_SIZE; + gator_buffer_mask[BACKTRACE_BUF] = BACKTRACE_BUFFER_SIZE - 1; + + gator_buffer_size[NAME_BUF] = NAME_BUFFER_SIZE; + gator_buffer_mask[NAME_BUF] = NAME_BUFFER_SIZE - 1; + + gator_buffer_size[COUNTER_BUF] = COUNTER_BUFFER_SIZE; + gator_buffer_mask[COUNTER_BUF] = COUNTER_BUFFER_SIZE - 1; + + gator_buffer_size[BLOCK_COUNTER_BUF] = BLOCK_COUNTER_BUFFER_SIZE; + gator_buffer_mask[BLOCK_COUNTER_BUF] = BLOCK_COUNTER_BUFFER_SIZE - 1; + + gator_buffer_size[ANNOTATE_BUF] = ANNOTATE_BUFFER_SIZE; + gator_buffer_mask[ANNOTATE_BUF] = ANNOTATE_BUFFER_SIZE - 1; + + gator_buffer_size[SCHED_TRACE_BUF] = SCHED_TRACE_BUFFER_SIZE; + gator_buffer_mask[SCHED_TRACE_BUF] = SCHED_TRACE_BUFFER_SIZE - 1; + + gator_buffer_size[IDLE_BUF] = IDLE_BUFFER_SIZE; + gator_buffer_mask[IDLE_BUF] = IDLE_BUFFER_SIZE - 1; + + gator_buffer_size[ACTIVITY_BUF] = ACTIVITY_BUFFER_SIZE; + gator_buffer_mask[ACTIVITY_BUF] = ACTIVITY_BUFFER_SIZE - 1; + + /* Initialize percpu per buffer variables */ + for (i = 0; i < NUM_GATOR_BUFS; i++) { + /* Verify buffers are a power of 2 */ + if (gator_buffer_size[i] & (gator_buffer_size[i] - 1)) { + err = -ENOEXEC; + goto setup_error; + } + + for_each_present_cpu(cpu) { + per_cpu(gator_buffer_read, cpu)[i] = 0; + per_cpu(gator_buffer_write, cpu)[i] = 0; + per_cpu(gator_buffer_commit, cpu)[i] = 0; + per_cpu(buffer_space_available, cpu)[i] = true; + per_cpu(gator_buffer_commit_time, cpu) = gator_live_rate; + + /* Annotation is a special case that only uses a single buffer */ + if (cpu > 0 && i == ANNOTATE_BUF) { + per_cpu(gator_buffer, cpu)[i] = NULL; + continue; + } + + per_cpu(gator_buffer, cpu)[i] = vmalloc(gator_buffer_size[i]); + if (!per_cpu(gator_buffer, cpu)[i]) { + err = -ENOMEM; + goto setup_error; + } + } + } + +setup_error: + mutex_unlock(&start_mutex); + return err; +} + +/* Actually start profiling (echo 1>/dev/gator/enable) */ +static int gator_op_start(void) +{ + int err = 0; + + mutex_lock(&start_mutex); + + if (gator_started || gator_start()) + err = -EINVAL; + else + gator_started = 1; + + mutex_unlock(&start_mutex); + + return err; +} + +/* echo 0>/dev/gator/enable */ +static void gator_op_stop(void) +{ + mutex_lock(&start_mutex); + + if (gator_started) { + gator_stop(); + + mutex_lock(&gator_buffer_mutex); + + gator_started = 0; + gator_monotonic_started = 0; + cookies_release(); + wake_up(&gator_buffer_wait); + + mutex_unlock(&gator_buffer_mutex); + } + + mutex_unlock(&start_mutex); +} + +static void gator_shutdown(void) +{ + int cpu, i; + + mutex_lock(&start_mutex); + + for_each_present_cpu(cpu) { + mutex_lock(&gator_buffer_mutex); + for (i = 0; i < NUM_GATOR_BUFS; i++) { + vfree(per_cpu(gator_buffer, cpu)[i]); + per_cpu(gator_buffer, cpu)[i] = NULL; + per_cpu(gator_buffer_read, cpu)[i] = 0; + per_cpu(gator_buffer_write, cpu)[i] = 0; + per_cpu(gator_buffer_commit, cpu)[i] = 0; + per_cpu(buffer_space_available, cpu)[i] = true; + per_cpu(gator_buffer_commit_time, cpu) = 0; + } + mutex_unlock(&gator_buffer_mutex); + } + + memset(&sent_core_name, 0, sizeof(sent_core_name)); + + mutex_unlock(&start_mutex); +} + +static int gator_set_backtrace(unsigned long val) +{ + int err = 0; + + mutex_lock(&start_mutex); + + if (gator_started) + err = -EBUSY; + else + gator_backtrace_depth = val; + + mutex_unlock(&start_mutex); + + return err; +} + +static ssize_t enable_read(struct file *file, char __user *buf, size_t count, loff_t *offset) +{ + return gatorfs_ulong_to_user(gator_started, buf, count, offset); +} + +static ssize_t enable_write(struct file *file, char const __user *buf, size_t count, loff_t *offset) +{ + unsigned long val; + int retval; + + if (*offset) + return -EINVAL; + + retval = gatorfs_ulong_from_user(&val, buf, count); + if (retval) + return retval; + + if (val) + retval = gator_op_start(); + else + gator_op_stop(); + + if (retval) + return retval; + return count; +} + +static const struct file_operations enable_fops = { + .read = enable_read, + .write = enable_write, +}; + +static int userspace_buffer_open(struct inode *inode, struct file *file) +{ + int err = -EPERM; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (test_and_set_bit_lock(0, &gator_buffer_opened)) + return -EBUSY; + + err = gator_op_setup(); + if (err) + goto fail; + + /* NB: the actual start happens from userspace + * echo 1 >/dev/gator/enable + */ + + return 0; + +fail: + __clear_bit_unlock(0, &gator_buffer_opened); + return err; +} + +static int userspace_buffer_release(struct inode *inode, struct file *file) +{ + gator_op_stop(); + gator_shutdown(); + __clear_bit_unlock(0, &gator_buffer_opened); + return 0; +} + +static ssize_t userspace_buffer_read(struct file *file, char __user *buf, size_t count, loff_t *offset) +{ + int commit, length1, length2, read; + char *buffer1; + char *buffer2; + int cpu, buftype; + int written = 0; + + /* ensure there is enough space for a whole frame */ + if (count < userspace_buffer_size || *offset) + return -EINVAL; + + /* sleep until the condition is true or a signal is received the + * condition is checked each time gator_buffer_wait is woken up + */ + wait_event_interruptible(gator_buffer_wait, buffer_commit_ready(&cpu, &buftype) || !gator_started); + + if (signal_pending(current)) + return -EINTR; + + if (buftype == -1 || cpu == -1) + return 0; + + mutex_lock(&gator_buffer_mutex); + + do { + read = per_cpu(gator_buffer_read, cpu)[buftype]; + commit = per_cpu(gator_buffer_commit, cpu)[buftype]; + + /* May happen if the buffer is freed during pending reads. */ + if (!per_cpu(gator_buffer, cpu)[buftype]) + break; + + /* determine the size of two halves */ + length1 = commit - read; + length2 = 0; + buffer1 = &(per_cpu(gator_buffer, cpu)[buftype][read]); + buffer2 = &(per_cpu(gator_buffer, cpu)[buftype][0]); + if (length1 < 0) { + length1 = gator_buffer_size[buftype] - read; + length2 = commit; + } + + if (length1 + length2 > count - written) + break; + + /* start, middle or end */ + if (length1 > 0 && copy_to_user(&buf[written], buffer1, length1)) + break; + + /* possible wrap around */ + if (length2 > 0 && copy_to_user(&buf[written + length1], buffer2, length2)) + break; + + per_cpu(gator_buffer_read, cpu)[buftype] = commit; + written += length1 + length2; + + /* Wake up annotate_write if more space is available */ + if (buftype == ANNOTATE_BUF) + wake_up(&gator_annotate_wait); + } while (buffer_commit_ready(&cpu, &buftype)); + + mutex_unlock(&gator_buffer_mutex); + + /* kick just in case we've lost an SMP event */ + wake_up(&gator_buffer_wait); + + return written > 0 ? written : -EFAULT; +} + +static const struct file_operations gator_event_buffer_fops = { + .open = userspace_buffer_open, + .release = userspace_buffer_release, + .read = userspace_buffer_read, +}; + +static ssize_t depth_read(struct file *file, char __user *buf, size_t count, loff_t *offset) +{ + return gatorfs_ulong_to_user(gator_backtrace_depth, buf, count, offset); +} + +static ssize_t depth_write(struct file *file, char const __user *buf, size_t count, loff_t *offset) +{ + unsigned long val; + int retval; + + if (*offset) + return -EINVAL; + + retval = gatorfs_ulong_from_user(&val, buf, count); + if (retval) + return retval; + + retval = gator_set_backtrace(val); + + if (retval) + return retval; + return count; +} + +static const struct file_operations depth_fops = { + .read = depth_read, + .write = depth_write +}; + +static void gator_op_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + struct gator_interface *gi; + int cpu; + + /* reinitialize default values */ + gator_cpu_cores = 0; + for_each_present_cpu(cpu) { + gator_cpu_cores++; + } + userspace_buffer_size = BACKTRACE_BUFFER_SIZE; + gator_response_type = 1; + gator_live_rate = 0; + + gatorfs_create_file(sb, root, "enable", &enable_fops); + gatorfs_create_file(sb, root, "buffer", &gator_event_buffer_fops); + gatorfs_create_file(sb, root, "backtrace_depth", &depth_fops); + gatorfs_create_ro_ulong(sb, root, "cpu_cores", &gator_cpu_cores); + gatorfs_create_ro_ulong(sb, root, "buffer_size", &userspace_buffer_size); + gatorfs_create_ulong(sb, root, "tick", &gator_timer_count); + gatorfs_create_ulong(sb, root, "response_type", &gator_response_type); + gatorfs_create_ro_ulong(sb, root, "version", &gator_protocol_version); + gatorfs_create_ro_u64(sb, root, "started", &gator_monotonic_started); + gatorfs_create_u64(sb, root, "live_rate", &gator_live_rate); + + /* Annotate interface */ + gator_annotate_create_files(sb, root); + + /* Linux Events */ + dir = gatorfs_mkdir(sb, root, "events"); + list_for_each_entry(gi, &gator_events, list) + if (gi->create_files) + gi->create_files(sb, dir); + + /* Sched Events */ + sched_trace_create_files(sb, dir); + + /* Power interface */ + gator_trace_power_create_files(sb, dir); +} + +/****************************************************************************** + * Module + ******************************************************************************/ + +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 15, 0) + +#define GATOR_TRACEPOINTS \ + GATOR_HANDLE_TRACEPOINT(block_rq_complete); \ + GATOR_HANDLE_TRACEPOINT(cpu_frequency); \ + GATOR_HANDLE_TRACEPOINT(cpu_idle); \ + GATOR_HANDLE_TRACEPOINT(cpu_migrate_begin); \ + GATOR_HANDLE_TRACEPOINT(cpu_migrate_current); \ + GATOR_HANDLE_TRACEPOINT(cpu_migrate_finish); \ + GATOR_HANDLE_TRACEPOINT(irq_handler_exit); \ + GATOR_HANDLE_TRACEPOINT(mali_hw_counter); \ + GATOR_HANDLE_TRACEPOINT(mali_job_slots_event); \ + GATOR_HANDLE_TRACEPOINT(mali_mmu_as_in_use); \ + GATOR_HANDLE_TRACEPOINT(mali_mmu_as_released); \ + GATOR_HANDLE_TRACEPOINT(mali_page_fault_insert_pages); \ + GATOR_HANDLE_TRACEPOINT(mali_pm_status); \ + GATOR_HANDLE_TRACEPOINT(mali_sw_counter); \ + GATOR_HANDLE_TRACEPOINT(mali_sw_counters); \ + GATOR_HANDLE_TRACEPOINT(mali_timeline_event); \ + GATOR_HANDLE_TRACEPOINT(mali_total_alloc_pages_change); \ + GATOR_HANDLE_TRACEPOINT(mm_page_alloc); \ + GATOR_HANDLE_TRACEPOINT(mm_page_free); \ + GATOR_HANDLE_TRACEPOINT(mm_page_free_batched); \ + GATOR_HANDLE_TRACEPOINT(sched_process_exec); \ + GATOR_HANDLE_TRACEPOINT(sched_process_fork); \ + GATOR_HANDLE_TRACEPOINT(sched_process_free); \ + GATOR_HANDLE_TRACEPOINT(sched_switch); \ + GATOR_HANDLE_TRACEPOINT(softirq_exit); \ + GATOR_HANDLE_TRACEPOINT(task_rename); \ + +#define GATOR_HANDLE_TRACEPOINT(probe_name) \ + struct tracepoint *gator_tracepoint_##probe_name +GATOR_TRACEPOINTS; +#undef GATOR_HANDLE_TRACEPOINT + +static void gator_save_tracepoint(struct tracepoint *tp, void *priv) +{ +#define GATOR_HANDLE_TRACEPOINT(probe_name) \ + do { \ + if (strcmp(tp->name, #probe_name) == 0) { \ + gator_tracepoint_##probe_name = tp; \ + return; \ + } \ + } while (0) +GATOR_TRACEPOINTS; +#undef GATOR_HANDLE_TRACEPOINT +} + +#else + +#define for_each_kernel_tracepoint(fct, priv) + +#endif + +static int __init gator_module_init(void) +{ + for_each_kernel_tracepoint(gator_save_tracepoint, NULL); + + if (gatorfs_register()) + return -1; + + if (gator_init()) { + gatorfs_unregister(); + return -1; + } + + setup_timer(&gator_buffer_wake_up_timer, gator_buffer_wake_up, 0); + + /* Initialize the list of cpuids */ + memset(gator_cpuids, -1, sizeof(gator_cpuids)); + on_each_cpu(gator_read_cpuid, NULL, 1); + + return 0; +} + +static void __exit gator_module_exit(void) +{ + del_timer_sync(&gator_buffer_wake_up_timer); + tracepoint_synchronize_unregister(); + gator_exit(); + gatorfs_unregister(); +} + +module_init(gator_module_init); +module_exit(gator_module_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("ARM Ltd"); +MODULE_DESCRIPTION("Gator system profiler"); +#define STRIFY2(ARG) #ARG +#define STRIFY(ARG) STRIFY2(ARG) +MODULE_VERSION(STRIFY(PROTOCOL_VERSION)); diff --git a/drivers/gator/gator_marshaling.c b/drivers/gator/gator_marshaling.c new file mode 100644 index 000000000000..0d1167643642 --- /dev/null +++ b/drivers/gator/gator_marshaling.c @@ -0,0 +1,371 @@ +/** + * Copyright (C) ARM Limited 2012-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#define NEWLINE_CANARY \ + /* Unix */ \ + "1\n" \ + /* Windows */ \ + "2\r\n" \ + /* Mac OS */ \ + "3\r" \ + /* RISC OS */ \ + "4\n\r" \ + /* Add another character so the length isn't 0x0a bytes */ \ + "5" + +#ifdef MALI_SUPPORT +#include "gator_events_mali_common.h" +#endif + +static void marshal_summary(long long timestamp, long long uptime, long long monotonic_delta, const char *uname) +{ + unsigned long flags; + int cpu = 0; + + local_irq_save(flags); + gator_buffer_write_packed_int(cpu, SUMMARY_BUF, MESSAGE_SUMMARY); + gator_buffer_write_string(cpu, SUMMARY_BUF, NEWLINE_CANARY); + gator_buffer_write_packed_int64(cpu, SUMMARY_BUF, timestamp); + gator_buffer_write_packed_int64(cpu, SUMMARY_BUF, uptime); + gator_buffer_write_packed_int64(cpu, SUMMARY_BUF, monotonic_delta); + gator_buffer_write_string(cpu, SUMMARY_BUF, "uname"); + gator_buffer_write_string(cpu, SUMMARY_BUF, uname); +#if GATOR_IKS_SUPPORT + gator_buffer_write_string(cpu, SUMMARY_BUF, "iks"); + gator_buffer_write_string(cpu, SUMMARY_BUF, ""); +#endif +#ifdef CONFIG_PREEMPT_RTB + gator_buffer_write_string(cpu, SUMMARY_BUF, "preempt_rtb"); + gator_buffer_write_string(cpu, SUMMARY_BUF, ""); +#endif +#ifdef CONFIG_PREEMPT_RT_FULL + gator_buffer_write_string(cpu, SUMMARY_BUF, "preempt_rt_full"); + gator_buffer_write_string(cpu, SUMMARY_BUF, ""); +#endif + /* Let Streamline know which GPU is used so that it can label the GPU Activity appropriately. This is a temporary fix, to be improved in a future release. */ +#ifdef MALI_SUPPORT + gator_buffer_write_string(cpu, SUMMARY_BUF, "mali_type"); +#if (MALI_SUPPORT == MALI_4xx) + gator_buffer_write_string(cpu, SUMMARY_BUF, "4xx"); +#elif (MALI_SUPPORT == MALI_MIDGARD) + gator_buffer_write_string(cpu, SUMMARY_BUF, "6xx"); +#else + gator_buffer_write_string(cpu, SUMMARY_BUF, "unknown"); +#endif +#endif + gator_buffer_write_string(cpu, SUMMARY_BUF, ""); + /* Commit the buffer now so it can be one of the first frames read by Streamline */ + local_irq_restore(flags); + gator_commit_buffer(cpu, SUMMARY_BUF, gator_get_time()); +} + +static bool marshal_cookie_header(const char *text) +{ + int cpu = get_physical_cpu(); + + return buffer_check_space(cpu, NAME_BUF, strlen(text) + 3 * MAXSIZE_PACK32); +} + +static void marshal_cookie(int cookie, const char *text) +{ + int cpu = get_physical_cpu(); + /* buffer_check_space already called by marshal_cookie_header */ + gator_buffer_write_packed_int(cpu, NAME_BUF, MESSAGE_COOKIE); + gator_buffer_write_packed_int(cpu, NAME_BUF, cookie); + gator_buffer_write_string(cpu, NAME_BUF, text); + buffer_check(cpu, NAME_BUF, gator_get_time()); +} + +static void marshal_thread_name(int pid, char *name) +{ + unsigned long flags, cpu; + u64 time; + + local_irq_save(flags); + cpu = get_physical_cpu(); + time = gator_get_time(); + if (buffer_check_space(cpu, NAME_BUF, TASK_COMM_LEN + 3 * MAXSIZE_PACK32 + MAXSIZE_PACK64)) { + gator_buffer_write_packed_int(cpu, NAME_BUF, MESSAGE_THREAD_NAME); + gator_buffer_write_packed_int64(cpu, NAME_BUF, time); + gator_buffer_write_packed_int(cpu, NAME_BUF, pid); + gator_buffer_write_string(cpu, NAME_BUF, name); + } + local_irq_restore(flags); + buffer_check(cpu, NAME_BUF, time); +} + +static void marshal_link(int cookie, int tgid, int pid) +{ + unsigned long cpu = get_physical_cpu(), flags; + u64 time; + + local_irq_save(flags); + time = gator_get_time(); + if (buffer_check_space(cpu, NAME_BUF, MAXSIZE_PACK64 + 5 * MAXSIZE_PACK32)) { + gator_buffer_write_packed_int(cpu, NAME_BUF, MESSAGE_LINK); + gator_buffer_write_packed_int64(cpu, NAME_BUF, time); + gator_buffer_write_packed_int(cpu, NAME_BUF, cookie); + gator_buffer_write_packed_int(cpu, NAME_BUF, tgid); + gator_buffer_write_packed_int(cpu, NAME_BUF, pid); + } + local_irq_restore(flags); + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, NAME_BUF, time); +} + +static bool marshal_backtrace_header(int exec_cookie, int tgid, int pid, u64 time) +{ + int cpu = get_physical_cpu(); + + if (!buffer_check_space(cpu, BACKTRACE_BUF, MAXSIZE_PACK64 + 5 * MAXSIZE_PACK32 + gator_backtrace_depth * 2 * MAXSIZE_PACK32)) { + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, BACKTRACE_BUF, time); + + return false; + } + + gator_buffer_write_packed_int64(cpu, BACKTRACE_BUF, time); + gator_buffer_write_packed_int(cpu, BACKTRACE_BUF, exec_cookie); + gator_buffer_write_packed_int(cpu, BACKTRACE_BUF, tgid); + gator_buffer_write_packed_int(cpu, BACKTRACE_BUF, pid); + + return true; +} + +static void marshal_backtrace(unsigned long address, int cookie, int in_kernel) +{ + int cpu = get_physical_cpu(); + + if (cookie == 0 && !in_kernel) + cookie = UNRESOLVED_COOKIE; + gator_buffer_write_packed_int(cpu, BACKTRACE_BUF, cookie); + gator_buffer_write_packed_int64(cpu, BACKTRACE_BUF, address); +} + +static void marshal_backtrace_footer(u64 time) +{ + int cpu = get_physical_cpu(); + + gator_buffer_write_packed_int(cpu, BACKTRACE_BUF, MESSAGE_END_BACKTRACE); + + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, BACKTRACE_BUF, time); +} + +static bool marshal_event_header(u64 time) +{ + unsigned long flags, cpu = get_physical_cpu(); + bool retval = false; + + local_irq_save(flags); + if (buffer_check_space(cpu, BLOCK_COUNTER_BUF, MAXSIZE_PACK32 + MAXSIZE_PACK64)) { + gator_buffer_write_packed_int(cpu, BLOCK_COUNTER_BUF, 0); /* key of zero indicates a timestamp */ + gator_buffer_write_packed_int64(cpu, BLOCK_COUNTER_BUF, time); + retval = true; + } + local_irq_restore(flags); + + return retval; +} + +static void marshal_event(int len, int *buffer) +{ + unsigned long i, flags, cpu = get_physical_cpu(); + + if (len <= 0) + return; + + /* length must be even since all data is a (key, value) pair */ + if (len & 0x1) { + pr_err("gator: invalid counter data detected and discarded\n"); + return; + } + + /* events must be written in key,value pairs */ + local_irq_save(flags); + for (i = 0; i < len; i += 2) { + if (!buffer_check_space(cpu, BLOCK_COUNTER_BUF, 2 * MAXSIZE_PACK32)) + break; + gator_buffer_write_packed_int(cpu, BLOCK_COUNTER_BUF, buffer[i]); + gator_buffer_write_packed_int(cpu, BLOCK_COUNTER_BUF, buffer[i + 1]); + } + local_irq_restore(flags); +} + +static void marshal_event64(int len, long long *buffer64) +{ + unsigned long i, flags, cpu = get_physical_cpu(); + + if (len <= 0) + return; + + /* length must be even since all data is a (key, value) pair */ + if (len & 0x1) { + pr_err("gator: invalid counter data detected and discarded\n"); + return; + } + + /* events must be written in key,value pairs */ + local_irq_save(flags); + for (i = 0; i < len; i += 2) { + if (!buffer_check_space(cpu, BLOCK_COUNTER_BUF, 2 * MAXSIZE_PACK64)) + break; + gator_buffer_write_packed_int64(cpu, BLOCK_COUNTER_BUF, buffer64[i]); + gator_buffer_write_packed_int64(cpu, BLOCK_COUNTER_BUF, buffer64[i + 1]); + } + local_irq_restore(flags); +} + +static void __maybe_unused marshal_event_single(int core, int key, int value) +{ + unsigned long flags, cpu; + u64 time; + + local_irq_save(flags); + cpu = get_physical_cpu(); + time = gator_get_time(); + if (buffer_check_space(cpu, COUNTER_BUF, MAXSIZE_PACK64 + 3 * MAXSIZE_PACK32)) { + gator_buffer_write_packed_int64(cpu, COUNTER_BUF, time); + gator_buffer_write_packed_int(cpu, COUNTER_BUF, core); + gator_buffer_write_packed_int(cpu, COUNTER_BUF, key); + gator_buffer_write_packed_int(cpu, COUNTER_BUF, value); + } + local_irq_restore(flags); + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, COUNTER_BUF, time); +} + +static void __maybe_unused marshal_event_single64(int core, int key, long long value) +{ + unsigned long flags, cpu; + u64 time; + + local_irq_save(flags); + cpu = get_physical_cpu(); + time = gator_get_time(); + if (buffer_check_space(cpu, COUNTER_BUF, 2 * MAXSIZE_PACK64 + 2 * MAXSIZE_PACK32)) { + gator_buffer_write_packed_int64(cpu, COUNTER_BUF, time); + gator_buffer_write_packed_int(cpu, COUNTER_BUF, core); + gator_buffer_write_packed_int(cpu, COUNTER_BUF, key); + gator_buffer_write_packed_int64(cpu, COUNTER_BUF, value); + } + local_irq_restore(flags); + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, COUNTER_BUF, time); +} + +static void marshal_sched_trace_switch(int pid, int state) +{ + unsigned long cpu = get_physical_cpu(), flags; + u64 time; + + if (!per_cpu(gator_buffer, cpu)[SCHED_TRACE_BUF]) + return; + + local_irq_save(flags); + time = gator_get_time(); + if (buffer_check_space(cpu, SCHED_TRACE_BUF, MAXSIZE_PACK64 + 5 * MAXSIZE_PACK32)) { + gator_buffer_write_packed_int(cpu, SCHED_TRACE_BUF, MESSAGE_SCHED_SWITCH); + gator_buffer_write_packed_int64(cpu, SCHED_TRACE_BUF, time); + gator_buffer_write_packed_int(cpu, SCHED_TRACE_BUF, pid); + gator_buffer_write_packed_int(cpu, SCHED_TRACE_BUF, state); + } + local_irq_restore(flags); + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, SCHED_TRACE_BUF, time); +} + +static void marshal_sched_trace_exit(int tgid, int pid) +{ + unsigned long cpu = get_physical_cpu(), flags; + u64 time; + + if (!per_cpu(gator_buffer, cpu)[SCHED_TRACE_BUF]) + return; + + local_irq_save(flags); + time = gator_get_time(); + if (buffer_check_space(cpu, SCHED_TRACE_BUF, MAXSIZE_PACK64 + 2 * MAXSIZE_PACK32)) { + gator_buffer_write_packed_int(cpu, SCHED_TRACE_BUF, MESSAGE_SCHED_EXIT); + gator_buffer_write_packed_int64(cpu, SCHED_TRACE_BUF, time); + gator_buffer_write_packed_int(cpu, SCHED_TRACE_BUF, pid); + } + local_irq_restore(flags); + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, SCHED_TRACE_BUF, time); +} + +#if GATOR_CPU_FREQ_SUPPORT +static void marshal_idle(int core, int state) +{ + unsigned long flags, cpu; + u64 time; + + local_irq_save(flags); + cpu = get_physical_cpu(); + time = gator_get_time(); + if (buffer_check_space(cpu, IDLE_BUF, MAXSIZE_PACK64 + 2 * MAXSIZE_PACK32)) { + gator_buffer_write_packed_int(cpu, IDLE_BUF, state); + gator_buffer_write_packed_int64(cpu, IDLE_BUF, time); + gator_buffer_write_packed_int(cpu, IDLE_BUF, core); + } + local_irq_restore(flags); + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, IDLE_BUF, time); +} +#endif + +#if defined(__arm__) || defined(__aarch64__) +static void marshal_core_name(const int core, const int cpuid, const char *name) +{ + int cpu = get_physical_cpu(); + unsigned long flags; + + local_irq_save(flags); + if (buffer_check_space(cpu, SUMMARY_BUF, MAXSIZE_PACK32 + MAXSIZE_CORE_NAME)) { + gator_buffer_write_packed_int(cpu, SUMMARY_BUF, MESSAGE_CORE_NAME); + gator_buffer_write_packed_int(cpu, SUMMARY_BUF, core); + gator_buffer_write_packed_int(cpu, SUMMARY_BUF, cpuid); + gator_buffer_write_string(cpu, SUMMARY_BUF, name); + } + /* Commit core names now so that they can show up in live */ + local_irq_restore(flags); + gator_commit_buffer(cpu, SUMMARY_BUF, gator_get_time()); +} +#endif + +static void marshal_activity_switch(int core, int key, int activity, int pid, int state) +{ + unsigned long cpu = get_physical_cpu(), flags; + u64 time; + + if (!per_cpu(gator_buffer, cpu)[ACTIVITY_BUF]) + return; + + local_irq_save(flags); + time = gator_get_time(); + if (buffer_check_space(cpu, ACTIVITY_BUF, MAXSIZE_PACK64 + 5 * MAXSIZE_PACK32)) { + gator_buffer_write_packed_int(cpu, ACTIVITY_BUF, MESSAGE_SWITCH); + gator_buffer_write_packed_int64(cpu, ACTIVITY_BUF, time); + gator_buffer_write_packed_int(cpu, ACTIVITY_BUF, core); + gator_buffer_write_packed_int(cpu, ACTIVITY_BUF, key); + gator_buffer_write_packed_int(cpu, ACTIVITY_BUF, activity); + gator_buffer_write_packed_int(cpu, ACTIVITY_BUF, pid); + gator_buffer_write_packed_int(cpu, ACTIVITY_BUF, state); + } + local_irq_restore(flags); + /* Check and commit; commit is set to occur once buffer is 3/4 full */ + buffer_check(cpu, ACTIVITY_BUF, time); +} + +void gator_marshal_activity_switch(int core, int key, int activity, int pid) +{ + /* state is reserved for cpu use only */ + marshal_activity_switch(core, key, activity, pid, 0); +} diff --git a/drivers/gator/gator_trace_gpu.c b/drivers/gator/gator_trace_gpu.c new file mode 100644 index 000000000000..5de9152e365a --- /dev/null +++ b/drivers/gator/gator_trace_gpu.c @@ -0,0 +1,321 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include "gator.h" + +#include <linux/slab.h> +#include <linux/module.h> +#include <linux/time.h> +#include <linux/math64.h> + +#ifdef MALI_SUPPORT +#ifdef MALI_DIR_MIDGARD +/* New DDK Directory structure with kernel/drivers/gpu/arm/midgard*/ +#include "mali_linux_trace.h" +#else +/* Old DDK Directory structure with kernel/drivers/gpu/arm/t6xx*/ +#include "linux/mali_linux_trace.h" +#endif +#endif + +/* + * Taken from MALI_PROFILING_EVENT_TYPE_* items in Mali DDK. + */ +#define EVENT_TYPE_SINGLE 0 +#define EVENT_TYPE_START 1 +#define EVENT_TYPE_STOP 2 +#define EVENT_TYPE_SUSPEND 3 +#define EVENT_TYPE_RESUME 4 + +/* Note whether tracepoints have been registered */ +static int mali_timeline_trace_registered; +static int mali_job_slots_trace_registered; + +enum { + GPU_UNIT_NONE = 0, + GPU_UNIT_VP, + GPU_UNIT_FP, + GPU_UNIT_CL, + NUMBER_OF_GPU_UNITS +}; + +#if defined(MALI_SUPPORT) + +struct mali_activity { + int core; + int key; + int count; + int last_activity; + int last_pid; +}; + +#define NUMBER_OF_GPU_CORES 16 +static struct mali_activity mali_activities[NUMBER_OF_GPU_UNITS*NUMBER_OF_GPU_CORES]; +static DEFINE_SPINLOCK(mali_activities_lock); + +/* Only one event should be running on a unit and core at a time (ie, + * a start event can only be followed by a stop and vice versa), but + * because the kernel only knows when a job is enqueued and not + * started, it is possible for a start1, start2, stop1, stop2. Change + * it back into start1, stop1, start2, stop2 by queueing up start2 and + * releasing it when stop1 is received. + */ + +static int mali_activity_index(int core, int key) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(mali_activities); ++i) { + if ((mali_activities[i].core == core) && (mali_activities[i].key == key)) + break; + if ((mali_activities[i].core == 0) && (mali_activities[i].key == 0)) { + mali_activities[i].core = core; + mali_activities[i].key = key; + break; + } + } + BUG_ON(i >= ARRAY_SIZE(mali_activities)); + + return i; +} + +static void mali_activity_enqueue(int core, int key, int activity, int pid) +{ + int i; + int count; + + spin_lock(&mali_activities_lock); + i = mali_activity_index(core, key); + + count = mali_activities[i].count; + BUG_ON(count < 0); + ++mali_activities[i].count; + if (count) { + mali_activities[i].last_activity = activity; + mali_activities[i].last_pid = pid; + } + spin_unlock(&mali_activities_lock); + + if (!count) + gator_marshal_activity_switch(core, key, activity, pid); +} + +static void mali_activity_stop(int core, int key) +{ + int i; + int count; + int last_activity = 0; + int last_pid = 0; + + spin_lock(&mali_activities_lock); + i = mali_activity_index(core, key); + + if (mali_activities[i].count == 0) { + spin_unlock(&mali_activities_lock); + return; + } + --mali_activities[i].count; + count = mali_activities[i].count; + if (count) { + last_activity = mali_activities[i].last_activity; + last_pid = mali_activities[i].last_pid; + } + spin_unlock(&mali_activities_lock); + + gator_marshal_activity_switch(core, key, 0, 0); + if (count) + gator_marshal_activity_switch(core, key, last_activity, last_pid); +} + +void mali_activity_clear(struct mali_counter mali_activity[], size_t mali_activity_size) +{ + int activity; + int cores; + int core; + + for (activity = 0; activity < mali_activity_size; ++activity) { + cores = mali_activity[activity].cores; + if (cores < 0) + cores = 1; + for (core = 0; core < cores; ++core) { + if (mali_activity[activity].enabled) { + preempt_disable(); + gator_marshal_activity_switch(core, mali_activity[activity].key, 0, 0); + preempt_enable(); + } + } + } +} + +#endif + +#if defined(MALI_SUPPORT) && (MALI_SUPPORT != MALI_MIDGARD) +#include "gator_events_mali_4xx.h" + +/* + * Taken from MALI_PROFILING_EVENT_CHANNEL_* in Mali DDK. + */ +enum { + EVENT_CHANNEL_SOFTWARE = 0, + EVENT_CHANNEL_VP0 = 1, + EVENT_CHANNEL_FP0 = 5, + EVENT_CHANNEL_FP1, + EVENT_CHANNEL_FP2, + EVENT_CHANNEL_FP3, + EVENT_CHANNEL_FP4, + EVENT_CHANNEL_FP5, + EVENT_CHANNEL_FP6, + EVENT_CHANNEL_FP7, + EVENT_CHANNEL_GPU = 21 +}; + +/** + * These events are applicable when the type MALI_PROFILING_EVENT_TYPE_SINGLE is used from the GPU channel + */ +enum { + EVENT_REASON_SINGLE_GPU_NONE = 0, + EVENT_REASON_SINGLE_GPU_FREQ_VOLT_CHANGE = 1, +}; + +struct mali_counter mali_activity[2]; + +GATOR_DEFINE_PROBE(mali_timeline_event, TP_PROTO(unsigned int event_id, unsigned int d0, unsigned int d1, unsigned int d2, unsigned int d3, unsigned int d4)) +{ + unsigned int component, state; + + /* do as much work as possible before disabling interrupts */ + component = (event_id >> 16) & 0xFF; /* component is an 8-bit field */ + state = (event_id >> 24) & 0xF; /* state is a 4-bit field */ + + switch (state) { + case EVENT_TYPE_START: + if (component == EVENT_CHANNEL_VP0) { + /* tgid = d0; pid = d1; */ + if (mali_activity[1].enabled) + mali_activity_enqueue(0, mali_activity[1].key, 1, d1); + } else if (component >= EVENT_CHANNEL_FP0 && component <= EVENT_CHANNEL_FP7) { + /* tgid = d0; pid = d1; */ + if (mali_activity[0].enabled) + mali_activity_enqueue(component - EVENT_CHANNEL_FP0, mali_activity[0].key, 1, d1); + } + break; + + case EVENT_TYPE_STOP: + if (component == EVENT_CHANNEL_VP0) { + if (mali_activity[1].enabled) + mali_activity_stop(0, mali_activity[1].key); + } else if (component >= EVENT_CHANNEL_FP0 && component <= EVENT_CHANNEL_FP7) { + if (mali_activity[0].enabled) + mali_activity_stop(component - EVENT_CHANNEL_FP0, mali_activity[0].key); + } + break; + + case EVENT_TYPE_SINGLE: + if (component == EVENT_CHANNEL_GPU) { + unsigned int reason = (event_id & 0xffff); + + if (reason == EVENT_REASON_SINGLE_GPU_FREQ_VOLT_CHANGE) + gator_events_mali_log_dvfs_event(d0, d1); + } + break; + + default: + break; + } +} +#endif + +#if defined(MALI_SUPPORT) && (MALI_SUPPORT == MALI_MIDGARD) + +struct mali_counter mali_activity[3]; + +#if defined(MALI_JOB_SLOTS_EVENT_CHANGED) +GATOR_DEFINE_PROBE(mali_job_slots_event, TP_PROTO(unsigned int event_id, unsigned int tgid, unsigned int pid, unsigned char job_id)) +#else +GATOR_DEFINE_PROBE(mali_job_slots_event, TP_PROTO(unsigned int event_id, unsigned int tgid, unsigned int pid)) +#endif +{ + unsigned int component, state, unit; +#if !defined(MALI_JOB_SLOTS_EVENT_CHANGED) + unsigned char job_id = 0; +#endif + + component = (event_id >> 16) & 0xFF; /* component is an 8-bit field */ + state = (event_id >> 24) & 0xF; /* state is a 4-bit field */ + + switch (component) { + case 0: + unit = GPU_UNIT_FP; + break; + case 1: + unit = GPU_UNIT_VP; + break; + case 2: + unit = GPU_UNIT_CL; + break; + default: + unit = GPU_UNIT_NONE; + } + + if (unit != GPU_UNIT_NONE) { + switch (state) { + case EVENT_TYPE_START: + if (mali_activity[component].enabled) + mali_activity_enqueue(0, mali_activity[component].key, 1, (pid != 0 ? pid : tgid)); + break; + case EVENT_TYPE_STOP: + default: /* Some jobs can be soft-stopped, so ensure that this terminates the activity trace. */ + if (mali_activity[component].enabled) + mali_activity_stop(0, mali_activity[component].key); + break; + } + } +} +#endif + +static int gator_trace_gpu_start(void) +{ + /* + * Returns nonzero for installation failed + * Absence of gpu trace points is not an error + */ + +#if defined(MALI_SUPPORT) + memset(&mali_activities, 0, sizeof(mali_activities)); +#endif + mali_timeline_trace_registered = mali_job_slots_trace_registered = 0; + +#if defined(MALI_SUPPORT) && (MALI_SUPPORT != MALI_MIDGARD) + mali_activity_clear(mali_activity, ARRAY_SIZE(mali_activity)); + if (!GATOR_REGISTER_TRACE(mali_timeline_event)) + mali_timeline_trace_registered = 1; +#endif + +#if defined(MALI_SUPPORT) && (MALI_SUPPORT == MALI_MIDGARD) + mali_activity_clear(mali_activity, ARRAY_SIZE(mali_activity)); + if (!GATOR_REGISTER_TRACE(mali_job_slots_event)) + mali_job_slots_trace_registered = 1; +#endif + + return 0; +} + +static void gator_trace_gpu_stop(void) +{ +#if defined(MALI_SUPPORT) && (MALI_SUPPORT != MALI_MIDGARD) + if (mali_timeline_trace_registered) + GATOR_UNREGISTER_TRACE(mali_timeline_event); +#endif + +#if defined(MALI_SUPPORT) && (MALI_SUPPORT == MALI_MIDGARD) + if (mali_job_slots_trace_registered) + GATOR_UNREGISTER_TRACE(mali_job_slots_event); +#endif + + mali_timeline_trace_registered = mali_job_slots_trace_registered = 0; +} diff --git a/drivers/gator/gator_trace_power.c b/drivers/gator/gator_trace_power.c new file mode 100644 index 000000000000..46e04b29a187 --- /dev/null +++ b/drivers/gator/gator_trace_power.c @@ -0,0 +1,192 @@ +/** + * Copyright (C) ARM Limited 2011-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include <linux/cpufreq.h> +#include <trace/events/power.h> + +#if defined(__arm__) + +#include <asm/mach-types.h> + +#define implements_wfi() (!machine_is_omap3_beagle()) + +#else + +#define implements_wfi() false + +#endif + +/* cpu_frequency and cpu_idle trace points were introduced in Linux + * kernel v2.6.38 the now deprecated power_frequency trace point was + * available prior to 2.6.38, but only for x86 + */ +#if GATOR_CPU_FREQ_SUPPORT +enum { + POWER_CPU_FREQ, + POWER_TOTAL +}; + +static DEFINE_PER_CPU(ulong, idle_prev_state); +static ulong power_cpu_enabled[POWER_TOTAL]; +static ulong power_cpu_key[POWER_TOTAL]; +static ulong power_cpu_cores; + +static int gator_trace_power_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + int cpu; + bool found_nonzero_freq = false; + + /* Even if CONFIG_CPU_FREQ is defined, it still may not be + * used. Check for non-zero values from cpufreq_quick_get + */ + for_each_online_cpu(cpu) { + if (cpufreq_quick_get(cpu) > 0) { + found_nonzero_freq = true; + break; + } + } + + if (found_nonzero_freq) { + /* cpu_frequency */ + dir = gatorfs_mkdir(sb, root, "Linux_power_cpu_freq"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &power_cpu_enabled[POWER_CPU_FREQ]); + gatorfs_create_ro_ulong(sb, dir, "key", &power_cpu_key[POWER_CPU_FREQ]); + } + + return 0; +} + +/* 'cpu' may not equal smp_processor_id(), i.e. may not be running on the core that is having the freq/idle state change */ +GATOR_DEFINE_PROBE(cpu_frequency, TP_PROTO(unsigned int frequency, unsigned int cpu)) +{ + cpu = lcpu_to_pcpu(cpu); + marshal_event_single64(cpu, power_cpu_key[POWER_CPU_FREQ], frequency * 1000L); +} + +GATOR_DEFINE_PROBE(cpu_idle, TP_PROTO(unsigned int state, unsigned int cpu)) +{ + cpu = lcpu_to_pcpu(cpu); + + if (state == per_cpu(idle_prev_state, cpu)) + return; + + if (implements_wfi()) { + if (state == PWR_EVENT_EXIT) { + /* transition from wfi to non-wfi */ + marshal_idle(cpu, MESSAGE_IDLE_EXIT); + } else { + /* transition from non-wfi to wfi */ + marshal_idle(cpu, MESSAGE_IDLE_ENTER); + } + } + + per_cpu(idle_prev_state, cpu) = state; +} + +static void gator_trace_power_online(void) +{ + int pcpu = get_physical_cpu(); + int lcpu = get_logical_cpu(); + + if (power_cpu_enabled[POWER_CPU_FREQ]) + marshal_event_single64(pcpu, power_cpu_key[POWER_CPU_FREQ], cpufreq_quick_get(lcpu) * 1000L); +} + +static void gator_trace_power_offline(void) +{ + /* Set frequency to zero on an offline */ + int cpu = get_physical_cpu(); + + if (power_cpu_enabled[POWER_CPU_FREQ]) + marshal_event_single(cpu, power_cpu_key[POWER_CPU_FREQ], 0); +} + +static int gator_trace_power_start(void) +{ + int cpu; + + /* register tracepoints */ + if (power_cpu_enabled[POWER_CPU_FREQ]) + if (GATOR_REGISTER_TRACE(cpu_frequency)) + goto fail_cpu_frequency_exit; + + /* Always register for cpu_idle for detecting WFI */ + if (GATOR_REGISTER_TRACE(cpu_idle)) + goto fail_cpu_idle_exit; + pr_debug("gator: registered power event tracepoints\n"); + + for_each_present_cpu(cpu) { + per_cpu(idle_prev_state, cpu) = 0; + } + + return 0; + + /* unregister tracepoints on error */ +fail_cpu_idle_exit: + if (power_cpu_enabled[POWER_CPU_FREQ]) + GATOR_UNREGISTER_TRACE(cpu_frequency); +fail_cpu_frequency_exit: + pr_err("gator: power event tracepoints failed to activate, please verify that tracepoints are enabled in the linux kernel\n"); + + return -1; +} + +static void gator_trace_power_stop(void) +{ + int i; + + if (power_cpu_enabled[POWER_CPU_FREQ]) + GATOR_UNREGISTER_TRACE(cpu_frequency); + GATOR_UNREGISTER_TRACE(cpu_idle); + pr_debug("gator: unregistered power event tracepoints\n"); + + for (i = 0; i < POWER_TOTAL; i++) + power_cpu_enabled[i] = 0; +} + +static void gator_trace_power_init(void) +{ + int i; + + power_cpu_cores = nr_cpu_ids; + for (i = 0; i < POWER_TOTAL; i++) { + power_cpu_enabled[i] = 0; + power_cpu_key[i] = gator_events_get_key(); + } +} +#else +static int gator_trace_power_create_files(struct super_block *sb, struct dentry *root) +{ + return 0; +} + +static void gator_trace_power_online(void) +{ +} + +static void gator_trace_power_offline(void) +{ +} + +static int gator_trace_power_start(void) +{ + return 0; +} + +static void gator_trace_power_stop(void) +{ +} + +static void gator_trace_power_init(void) +{ +} +#endif diff --git a/drivers/gator/gator_trace_sched.c b/drivers/gator/gator_trace_sched.c new file mode 100644 index 000000000000..6d7cbd7348e1 --- /dev/null +++ b/drivers/gator/gator_trace_sched.c @@ -0,0 +1,321 @@ +/** + * Copyright (C) ARM Limited 2010-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include <trace/events/sched.h> +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 4, 0) +#include <trace/events/task.h> +#endif + +#include "gator.h" + +#define TASK_MAP_ENTRIES 1024 /* must be power of 2 */ +#define TASK_MAX_COLLISIONS 2 + +enum { + STATE_WAIT_ON_OTHER = 0, + STATE_CONTENTION, + STATE_WAIT_ON_IO, + CPU_WAIT_TOTAL +}; + +static DEFINE_PER_CPU(uint64_t *, taskname_keys); +static DEFINE_PER_CPU(int, collecting); + +/* this array is never read as the cpu wait charts are derived + * counters the files are needed, nonetheless, to show that these + * counters are available + */ +static ulong cpu_wait_enabled[CPU_WAIT_TOTAL]; +static ulong sched_cpu_key[CPU_WAIT_TOTAL]; + +static int sched_trace_create_files(struct super_block *sb, struct dentry *root) +{ + struct dentry *dir; + + /* CPU Wait - Contention */ + dir = gatorfs_mkdir(sb, root, "Linux_cpu_wait_contention"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &cpu_wait_enabled[STATE_CONTENTION]); + gatorfs_create_ro_ulong(sb, dir, "key", &sched_cpu_key[STATE_CONTENTION]); + + /* CPU Wait - I/O */ + dir = gatorfs_mkdir(sb, root, "Linux_cpu_wait_io"); + if (!dir) + return -1; + gatorfs_create_ulong(sb, dir, "enabled", &cpu_wait_enabled[STATE_WAIT_ON_IO]); + gatorfs_create_ro_ulong(sb, dir, "key", &sched_cpu_key[STATE_WAIT_ON_IO]); + + return 0; +} + +static void emit_pid_name(const char *comm, struct task_struct *task) +{ + bool found = false; + char taskcomm[TASK_COMM_LEN + 3]; + unsigned long x, cpu = get_physical_cpu(); + uint64_t *keys = &(per_cpu(taskname_keys, cpu)[(task->pid & 0xFF) * TASK_MAX_COLLISIONS]); + uint64_t value; + + value = gator_chksum_crc32(comm); + value = (value << 32) | (uint32_t)task->pid; + + /* determine if the thread name was emitted already */ + for (x = 0; x < TASK_MAX_COLLISIONS; x++) { + if (keys[x] == value) { + found = true; + break; + } + } + + if (!found) { + /* shift values, new value always in front */ + uint64_t oldv, newv = value; + + for (x = 0; x < TASK_MAX_COLLISIONS; x++) { + oldv = keys[x]; + keys[x] = newv; + newv = oldv; + } + + /* emit pid names, cannot use get_task_comm, as it's not exported on all kernel versions */ + if (strlcpy(taskcomm, comm, TASK_COMM_LEN) == TASK_COMM_LEN - 1) { + /* append ellipses if comm has length of TASK_COMM_LEN - 1 */ + strcat(taskcomm, "..."); + } + + marshal_thread_name(task->pid, taskcomm); + } +} + +static void collect_counters(u64 time, struct task_struct *task, bool sched_switch) +{ + int *buffer, len, cpu = get_physical_cpu(); + long long *buffer64; + struct gator_interface *gi; + + if (marshal_event_header(time)) { + list_for_each_entry(gi, &gator_events, list) { + if (gi->read) { + len = gi->read(&buffer, sched_switch); + marshal_event(len, buffer); + } else if (gi->read64) { + len = gi->read64(&buffer64); + marshal_event64(len, buffer64); + } + if (gi->read_proc && task != NULL) { + len = gi->read_proc(&buffer64, task); + marshal_event64(len, buffer64); + } + } + if (cpu == 0) + gator_emit_perf_time(time); + /* Only check after writing all counters so that time and corresponding counters appear in the same frame */ + buffer_check(cpu, BLOCK_COUNTER_BUF, time); + + /* Commit buffers on timeout */ + if (gator_live_rate > 0 && time >= per_cpu(gator_buffer_commit_time, cpu)) { + static const int buftypes[] = { NAME_BUF, COUNTER_BUF, BLOCK_COUNTER_BUF, SCHED_TRACE_BUF, ACTIVITY_BUF }; + int i; + + for (i = 0; i < ARRAY_SIZE(buftypes); ++i) + gator_commit_buffer(cpu, buftypes[i], time); + + /* spinlocks are noops on uniprocessor machines and mutexes do + * not work in sched_switch context in RT-Preempt full, so + * disable proactive flushing of the annotate frame on + * uniprocessor machines. + */ +#ifdef CONFIG_SMP + /* Try to preemptively flush the annotate buffer to reduce the chance of the buffer being full */ + if (on_primary_core() && spin_trylock(&annotate_lock)) { + gator_commit_buffer(0, ANNOTATE_BUF, time); + spin_unlock(&annotate_lock); + } +#endif + } + } +} + +/* special case used during a suspend of the system */ +static void trace_sched_insert_idle(void) +{ + marshal_sched_trace_switch(0, 0); +} + +static void gator_trace_emit_link(struct task_struct *p) +{ + int cookie; + int cpu = get_physical_cpu(); + + cookie = get_exec_cookie(cpu, p); + emit_pid_name(p->comm, p); + + marshal_link(cookie, p->tgid, p->pid); +} + +GATOR_DEFINE_PROBE(sched_process_fork, TP_PROTO(struct task_struct *parent, struct task_struct *child)) +{ + gator_trace_emit_link(child); +} + +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 4, 0) +GATOR_DEFINE_PROBE(sched_process_exec, TP_PROTO(struct task_struct *p, pid_t old_pid, struct linux_binprm *bprm)) +{ + gator_trace_emit_link(p); +} + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 15, 0) +GATOR_DEFINE_PROBE(task_rename, TP_PROTO(struct task_struct *task, char *comm)) +#else +GATOR_DEFINE_PROBE(task_rename, TP_PROTO(struct task_struct *task, const char *comm)) +#endif +{ + emit_pid_name(comm, task); +} +#endif + +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 35) +GATOR_DEFINE_PROBE(sched_switch, TP_PROTO(struct rq *rq, struct task_struct *prev, struct task_struct *next)) +#else +GATOR_DEFINE_PROBE(sched_switch, TP_PROTO(struct task_struct *prev, struct task_struct *next)) +#endif +{ + int state; + int cpu = get_physical_cpu(); + + per_cpu(in_scheduler_context, cpu) = true; + + /* do as much work as possible before disabling interrupts */ + if (prev->state == TASK_RUNNING) + state = STATE_CONTENTION; + else if (prev->in_iowait) + state = STATE_WAIT_ON_IO; + else + state = STATE_WAIT_ON_OTHER; + + per_cpu(collecting, cpu) = 1; + collect_counters(gator_get_time(), prev, true); + per_cpu(collecting, cpu) = 0; + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 4, 0) + gator_trace_emit_link(next); +#endif + marshal_sched_trace_switch(next->pid, state); + + per_cpu(in_scheduler_context, cpu) = false; +} + +GATOR_DEFINE_PROBE(sched_process_free, TP_PROTO(struct task_struct *p)) +{ + marshal_sched_trace_exit(p->tgid, p->pid); +} + +static void do_nothing(void *info) +{ + /* Intentionally do nothing */ + (void)info; +} + +static int register_scheduler_tracepoints(void) +{ + /* register tracepoints */ + if (GATOR_REGISTER_TRACE(sched_process_fork)) + goto fail_sched_process_fork; +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 4, 0) + if (GATOR_REGISTER_TRACE(sched_process_exec)) + goto fail_sched_process_exec; + if (GATOR_REGISTER_TRACE(task_rename)) + goto fail_task_rename; +#endif + if (GATOR_REGISTER_TRACE(sched_switch)) + goto fail_sched_switch; + if (GATOR_REGISTER_TRACE(sched_process_free)) + goto fail_sched_process_free; + pr_debug("gator: registered tracepoints\n"); + + /* Now that the scheduler tracepoint is registered, force a context + * switch on all cpus to capture what is currently running. + */ + on_each_cpu(do_nothing, NULL, 0); + + return 0; + + /* unregister tracepoints on error */ +fail_sched_process_free: + GATOR_UNREGISTER_TRACE(sched_switch); +fail_sched_switch: +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 4, 0) + GATOR_UNREGISTER_TRACE(task_rename); +fail_task_rename: + GATOR_UNREGISTER_TRACE(sched_process_exec); +fail_sched_process_exec: +#endif + GATOR_UNREGISTER_TRACE(sched_process_fork); +fail_sched_process_fork: + pr_err("gator: tracepoints failed to activate, please verify that tracepoints are enabled in the linux kernel\n"); + + return -1; +} + +static void unregister_scheduler_tracepoints(void) +{ + GATOR_UNREGISTER_TRACE(sched_process_fork); +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 4, 0) + GATOR_UNREGISTER_TRACE(sched_process_exec); + GATOR_UNREGISTER_TRACE(task_rename); +#endif + GATOR_UNREGISTER_TRACE(sched_switch); + GATOR_UNREGISTER_TRACE(sched_process_free); + pr_debug("gator: unregistered tracepoints\n"); +} + +static void gator_trace_sched_stop(void) +{ + int cpu; + + unregister_scheduler_tracepoints(); + + for_each_present_cpu(cpu) { + kfree(per_cpu(taskname_keys, cpu)); + } +} + +static int gator_trace_sched_start(void) +{ + int cpu, size; + int ret; + + for_each_present_cpu(cpu) { + size = TASK_MAP_ENTRIES * TASK_MAX_COLLISIONS * sizeof(uint64_t); + per_cpu(taskname_keys, cpu) = kmalloc(size, GFP_KERNEL); + if (!per_cpu(taskname_keys, cpu)) + return -1; + memset(per_cpu(taskname_keys, cpu), 0, size); + } + + ret = register_scheduler_tracepoints(); + + return ret; +} + +static void gator_trace_sched_offline(void) +{ + trace_sched_insert_idle(); +} + +static void gator_trace_sched_init(void) +{ + int i; + + for (i = 0; i < CPU_WAIT_TOTAL; i++) { + cpu_wait_enabled[i] = 0; + sched_cpu_key[i] = gator_events_get_key(); + } +} diff --git a/drivers/gator/mali/mali_kbase_gator_api.h b/drivers/gator/mali/mali_kbase_gator_api.h new file mode 100644 index 000000000000..5ed069797e36 --- /dev/null +++ b/drivers/gator/mali/mali_kbase_gator_api.h @@ -0,0 +1,219 @@ +/** + * Copyright (C) ARM Limited 2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#ifndef _KBASE_GATOR_API_H_ +#define _KBASE_GATOR_API_H_ + +/** + * @brief This file describes the API used by Gator to collect hardware counters data from a Mali device. + */ + +/* This define is used by the gator kernel module compile to select which DDK + * API calling convention to use. If not defined (legacy DDK) gator assumes + * version 1. The version to DDK release mapping is: + * Version 1 API: DDK versions r1px, r2px + * Version 2 API: DDK versions r3px, r4px + * Version 3 API: DDK version r5p0 and newer + * + * API Usage + * ========= + * + * 1] Call kbase_gator_hwcnt_init_names() to return the list of short counter + * names for the GPU present in this device. + * + * 2] Create a kbase_gator_hwcnt_info structure and set the counter enables for + * the counters you want enabled. The enables can all be set for simplicity in + * most use cases, but disabling some will let you minimize bandwidth impact. + * + * 3] Call kbase_gator_hwcnt_init() using the above structure, to create a + * counter context. On successful return the DDK will have populated the + * structure with a variety of useful information. + * + * 4] Call kbase_gator_hwcnt_dump_irq() to queue a non-blocking request for a + * counter dump. If this returns a non-zero value the request has been queued, + * otherwise the driver has been unable to do so (typically because of another + * user of the instrumentation exists concurrently). + * + * 5] Call kbase_gator_hwcnt_dump_complete() to test whether the previously + * requested dump has been succesful. If this returns non-zero the counter dump + * has resolved, but the value of *success must also be tested as the dump + * may have not been successful. If it returns zero the counter dump was + * abandoned due to the device being busy (typically because of another + * user of the instrumentation exists concurrently). + * + * 6] Process the counters stored in the buffer pointed to by ... + * + * kbase_gator_hwcnt_info->kernel_dump_buffer + * + * In pseudo code you can find all of the counters via this approach: + * + * + * hwcnt_info # pointer to kbase_gator_hwcnt_info structure + * hwcnt_name # pointer to name list + * + * u32 * hwcnt_data = (u32*)hwcnt_info->kernel_dump_buffer + * + * # Iterate over each 64-counter block in this GPU configuration + * for( i = 0; i < hwcnt_info->nr_hwc_blocks; i++) { + * hwc_type type = hwcnt_info->hwc_layout[i]; + * + * # Skip reserved type blocks - they contain no counters at all + * if( type == RESERVED_BLOCK ) { + * continue; + * } + * + * size_t name_offset = type * 64; + * size_t data_offset = i * 64; + * + * # Iterate over the names of the counters in this block type + * for( j = 0; j < 64; j++) { + * const char * name = hwcnt_name[name_offset+j]; + * + * # Skip empty name strings - there is no counter here + * if( name[0] == '\0' ) { + * continue; + * } + * + * u32 data = hwcnt_data[data_offset+j]; + * + * printk( "COUNTER: %s DATA: %u\n", name, data ); + * } + * } + * + * + * Note that in most implementations you typically want to either SUM or + * AVERAGE multiple instances of the same counter if, for example, you have + * multiple shader cores or multiple L2 caches. The most sensible view for + * analysis is to AVERAGE shader core counters, but SUM L2 cache and MMU + * counters. + * + * 7] Goto 4, repeating until you want to stop collecting counters. + * + * 8] Release the dump resources by calling kbase_gator_hwcnt_term(). + * + * 9] Release the name table resources by calling kbase_gator_hwcnt_term_names(). + * This function must only be called if init_names() returned a non-NULL value. + **/ + +#define MALI_DDK_GATOR_API_VERSION 3 + +#if !defined(MALI_TRUE) + #define MALI_TRUE ((uint32_t)1) +#endif + +#if !defined(MALI_FALSE) + #define MALI_FALSE ((uint32_t)0) +#endif + +enum hwc_type { + JM_BLOCK = 0, + TILER_BLOCK, + SHADER_BLOCK, + MMU_L2_BLOCK, + RESERVED_BLOCK +}; + +struct kbase_gator_hwcnt_info { + + /* Passed from Gator to kbase */ + + /* the bitmask of enabled hardware counters for each counter block */ + uint16_t bitmask[4]; + + /* Passed from kbase to Gator */ + + /* ptr to counter dump memory */ + void *kernel_dump_buffer; + + /* size of counter dump memory */ + uint32_t size; + + /* the ID of the Mali device */ + uint32_t gpu_id; + + /* the number of shader cores in the GPU */ + uint32_t nr_cores; + + /* the number of core groups */ + uint32_t nr_core_groups; + + /* the memory layout of the performance counters */ + enum hwc_type *hwc_layout; + + /* the total number of hardware couter blocks */ + uint32_t nr_hwc_blocks; +}; + +/** + * @brief Opaque block of Mali data which Gator needs to return to the API later. + */ +struct kbase_gator_hwcnt_handles; + +/** + * @brief Initialize the resources Gator needs for performance profiling. + * + * @param in_out_info A pointer to a structure containing the enabled counters passed from Gator and all the Mali + * specific information that will be returned to Gator. On entry Gator must have populated the + * 'bitmask' field with the counters it wishes to enable for each class of counter block. + * Each entry in the array corresponds to a single counter class based on the "hwc_type" + * enumeration, and each bit corresponds to an enable for 4 sequential counters (LSB enables + * the first 4 counters in the block, and so on). See the GPU counter array as returned by + * kbase_gator_hwcnt_get_names() for the index values of each counter for the curernt GPU. + * + * @return Pointer to an opaque handle block on success, NULL on error. + */ +extern struct kbase_gator_hwcnt_handles *kbase_gator_hwcnt_init(struct kbase_gator_hwcnt_info *in_out_info); + +/** + * @brief Free all resources once Gator has finished using performance counters. + * + * @param in_out_info A pointer to a structure containing the enabled counters passed from Gator and all the + * Mali specific information that will be returned to Gator. + * @param opaque_handles A wrapper structure for kbase structures. + */ +extern void kbase_gator_hwcnt_term(struct kbase_gator_hwcnt_info *in_out_info, struct kbase_gator_hwcnt_handles *opaque_handles); + +/** + * @brief Poll whether a counter dump is successful. + * + * @param opaque_handles A wrapper structure for kbase structures. + * @param[out] success Non-zero on success, zero on failure. + * + * @return Zero if the dump is still pending, non-zero if the dump has completed. Note that a + * completed dump may not have dumped succesfully, so the caller must test for both + * a completed and successful dump before processing counters. + */ +extern uint32_t kbase_gator_instr_hwcnt_dump_complete(struct kbase_gator_hwcnt_handles *opaque_handles, uint32_t * const success); + +/** + * @brief Request the generation of a new counter dump. + * + * @param opaque_handles A wrapper structure for kbase structures. + * + * @return Zero if the hardware device is busy and cannot handle the request, non-zero otherwise. + */ +extern uint32_t kbase_gator_instr_hwcnt_dump_irq(struct kbase_gator_hwcnt_handles *opaque_handles); + +/** + * @brief This function is used to fetch the names table based on the Mali device in use. + * + * @param[out] total_number_of_counters The total number of counters short names in the Mali devices' list. + * + * @return Pointer to an array of strings of length *total_number_of_counters. + */ +extern const char * const *kbase_gator_hwcnt_init_names(uint32_t *total_number_of_counters); + +/** + * @brief This function is used to terminate the use of the names table. + * + * This function must only be called if the initial call to kbase_gator_hwcnt_init_names returned a non-NULL value. + */ +extern void kbase_gator_hwcnt_term_names(void); + +#endif diff --git a/drivers/gator/mali/mali_mjollnir_profiling_gator_api.h b/drivers/gator/mali/mali_mjollnir_profiling_gator_api.h new file mode 100644 index 000000000000..2bc0b037eee6 --- /dev/null +++ b/drivers/gator/mali/mali_mjollnir_profiling_gator_api.h @@ -0,0 +1,159 @@ +/** + * Copyright (C) ARM Limited 2013-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#ifndef __MALI_MJOLLNIR_PROFILING_GATOR_API_H__ +#define __MALI_MJOLLNIR_PROFILING_GATOR_API_H__ + +#ifdef __cplusplus +extern "C" +{ +#endif + + +/* + * The number of processor cores. Update to suit your hardware implementation. + */ +#define MAX_NUM_FP_CORES (4) +#define MAX_NUM_VP_CORES (1) +#define MAX_NUM_L2_CACHE_CORES (1) + +enum counters { + /* Timeline activity */ + ACTIVITY_VP_0 = 0, + ACTIVITY_FP_0, + ACTIVITY_FP_1, + ACTIVITY_FP_2, + ACTIVITY_FP_3, + + /* L2 cache counters */ + COUNTER_L2_0_C0, + COUNTER_L2_0_C1, + + /* Vertex processor counters */ + COUNTER_VP_0_C0, + COUNTER_VP_0_C1, + + /* Fragment processor counters */ + COUNTER_FP_0_C0, + COUNTER_FP_0_C1, + COUNTER_FP_1_C0, + COUNTER_FP_1_C1, + COUNTER_FP_2_C0, + COUNTER_FP_2_C1, + COUNTER_FP_3_C0, + COUNTER_FP_3_C1, + + /* EGL Software Counters */ + COUNTER_EGL_BLIT_TIME, + + /* GLES Software Counters */ + COUNTER_GLES_DRAW_ELEMENTS_CALLS, + COUNTER_GLES_DRAW_ELEMENTS_NUM_INDICES, + COUNTER_GLES_DRAW_ELEMENTS_NUM_TRANSFORMED, + COUNTER_GLES_DRAW_ARRAYS_CALLS, + COUNTER_GLES_DRAW_ARRAYS_NUM_TRANSFORMED, + COUNTER_GLES_DRAW_POINTS, + COUNTER_GLES_DRAW_LINES, + COUNTER_GLES_DRAW_LINE_LOOP, + COUNTER_GLES_DRAW_LINE_STRIP, + COUNTER_GLES_DRAW_TRIANGLES, + COUNTER_GLES_DRAW_TRIANGLE_STRIP, + COUNTER_GLES_DRAW_TRIANGLE_FAN, + COUNTER_GLES_NON_VBO_DATA_COPY_TIME, + COUNTER_GLES_UNIFORM_BYTES_COPIED_TO_MALI, + COUNTER_GLES_UPLOAD_TEXTURE_TIME, + COUNTER_GLES_UPLOAD_VBO_TIME, + COUNTER_GLES_NUM_FLUSHES, + COUNTER_GLES_NUM_VSHADERS_GENERATED, + COUNTER_GLES_NUM_FSHADERS_GENERATED, + COUNTER_GLES_VSHADER_GEN_TIME, + COUNTER_GLES_FSHADER_GEN_TIME, + COUNTER_GLES_INPUT_TRIANGLES, + COUNTER_GLES_VXCACHE_HIT, + COUNTER_GLES_VXCACHE_MISS, + COUNTER_GLES_VXCACHE_COLLISION, + COUNTER_GLES_CULLED_TRIANGLES, + COUNTER_GLES_CULLED_LINES, + COUNTER_GLES_BACKFACE_TRIANGLES, + COUNTER_GLES_GBCLIP_TRIANGLES, + COUNTER_GLES_GBCLIP_LINES, + COUNTER_GLES_TRIANGLES_DRAWN, + COUNTER_GLES_DRAWCALL_TIME, + COUNTER_GLES_TRIANGLES_COUNT, + COUNTER_GLES_INDEPENDENT_TRIANGLES_COUNT, + COUNTER_GLES_STRIP_TRIANGLES_COUNT, + COUNTER_GLES_FAN_TRIANGLES_COUNT, + COUNTER_GLES_LINES_COUNT, + COUNTER_GLES_INDEPENDENT_LINES_COUNT, + COUNTER_GLES_STRIP_LINES_COUNT, + COUNTER_GLES_LOOP_LINES_COUNT, + + COUNTER_FILMSTRIP, + COUNTER_FREQUENCY, + COUNTER_VOLTAGE, + + NUMBER_OF_EVENTS +}; + +#define FIRST_ACTIVITY_EVENT ACTIVITY_VP_0 +#define LAST_ACTIVITY_EVENT ACTIVITY_FP_3 + +#define FIRST_HW_COUNTER COUNTER_L2_0_C0 +#define LAST_HW_COUNTER COUNTER_FP_3_C1 + +#define FIRST_SW_COUNTER COUNTER_EGL_BLIT_TIME +#define LAST_SW_COUNTER COUNTER_GLES_LOOP_LINES_COUNT + +/* Signifies that the system is able to report voltage and frequency numbers. */ +#define DVFS_REPORTED_BY_DDK 1 + +/** + * Structure to pass performance counter data of a Mali core + */ +struct _mali_profiling_core_counters { + u32 source0; + u32 value0; + u32 source1; + u32 value1; +}; + +/* + * For compatibility with utgard. + */ +struct _mali_profiling_l2_counter_values { + struct _mali_profiling_core_counters cores[MAX_NUM_L2_CACHE_CORES]; +}; + +struct _mali_profiling_mali_version { + u32 mali_product_id; + u32 mali_version_major; + u32 mali_version_minor; + u32 num_of_l2_cores; + u32 num_of_fp_cores; + u32 num_of_vp_cores; +}; + +extern void _mali_profiling_get_mali_version(struct _mali_profiling_mali_version *values); +extern u32 _mali_profiling_get_l2_counters(struct _mali_profiling_l2_counter_values *values); + +/* + * List of possible actions allowing DDK to be controlled by Streamline. + * The following numbers are used by DDK to control the frame buffer dumping. + */ +#define FBDUMP_CONTROL_ENABLE (1) +#define FBDUMP_CONTROL_RATE (2) +#define SW_COUNTER_ENABLE (3) +#define FBDUMP_CONTROL_RESIZE_FACTOR (4) + + +#ifdef __cplusplus +} +#endif + +#endif /* __MALI_MJOLLNIR_PROFILING_GATOR_API_H__ */ diff --git a/drivers/gator/mali/mali_utgard_profiling_gator_api.h b/drivers/gator/mali/mali_utgard_profiling_gator_api.h new file mode 100644 index 000000000000..d6465312628e --- /dev/null +++ b/drivers/gator/mali/mali_utgard_profiling_gator_api.h @@ -0,0 +1,197 @@ +/** + * Copyright (C) ARM Limited 2013-2014. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#ifndef __MALI_UTGARD_PROFILING_GATOR_API_H__ +#define __MALI_UTGARD_PROFILING_GATOR_API_H__ + +#ifdef __cplusplus +extern "C" +{ +#endif + +#define MALI_PROFILING_API_VERSION 4 + +#define MAX_NUM_L2_CACHE_CORES 3 +#define MAX_NUM_FP_CORES 8 +#define MAX_NUM_VP_CORES 1 + +/** The list of events supported by the Mali DDK. */ +enum { + /* Vertex processor activity */ + ACTIVITY_VP_0 = 0, + + /* Fragment processor activity */ + ACTIVITY_FP_0, /* 1 */ + ACTIVITY_FP_1, + ACTIVITY_FP_2, + ACTIVITY_FP_3, + ACTIVITY_FP_4, + ACTIVITY_FP_5, + ACTIVITY_FP_6, + ACTIVITY_FP_7, + + /* L2 cache counters */ + COUNTER_L2_0_C0, + COUNTER_L2_0_C1, + COUNTER_L2_1_C0, + COUNTER_L2_1_C1, + COUNTER_L2_2_C0, + COUNTER_L2_2_C1, + + /* Vertex processor counters */ + COUNTER_VP_0_C0, /*15*/ + COUNTER_VP_0_C1, + + /* Fragment processor counters */ + COUNTER_FP_0_C0, + COUNTER_FP_0_C1, + COUNTER_FP_1_C0, + COUNTER_FP_1_C1, + COUNTER_FP_2_C0, + COUNTER_FP_2_C1, + COUNTER_FP_3_C0, + COUNTER_FP_3_C1, + COUNTER_FP_4_C0, + COUNTER_FP_4_C1, + COUNTER_FP_5_C0, + COUNTER_FP_5_C1, + COUNTER_FP_6_C0, + COUNTER_FP_6_C1, + COUNTER_FP_7_C0, + COUNTER_FP_7_C1, /* 32 */ + + /* + * If more hardware counters are added, the _mali_osk_hw_counter_table + * below should also be updated. + */ + + /* EGL software counters */ + COUNTER_EGL_BLIT_TIME, + + /* GLES software counters */ + COUNTER_GLES_DRAW_ELEMENTS_CALLS, + COUNTER_GLES_DRAW_ELEMENTS_NUM_INDICES, + COUNTER_GLES_DRAW_ELEMENTS_NUM_TRANSFORMED, + COUNTER_GLES_DRAW_ARRAYS_CALLS, + COUNTER_GLES_DRAW_ARRAYS_NUM_TRANSFORMED, + COUNTER_GLES_DRAW_POINTS, + COUNTER_GLES_DRAW_LINES, + COUNTER_GLES_DRAW_LINE_LOOP, + COUNTER_GLES_DRAW_LINE_STRIP, + COUNTER_GLES_DRAW_TRIANGLES, + COUNTER_GLES_DRAW_TRIANGLE_STRIP, + COUNTER_GLES_DRAW_TRIANGLE_FAN, + COUNTER_GLES_NON_VBO_DATA_COPY_TIME, + COUNTER_GLES_UNIFORM_BYTES_COPIED_TO_MALI, + COUNTER_GLES_UPLOAD_TEXTURE_TIME, + COUNTER_GLES_UPLOAD_VBO_TIME, + COUNTER_GLES_NUM_FLUSHES, + COUNTER_GLES_NUM_VSHADERS_GENERATED, + COUNTER_GLES_NUM_FSHADERS_GENERATED, + COUNTER_GLES_VSHADER_GEN_TIME, + COUNTER_GLES_FSHADER_GEN_TIME, + COUNTER_GLES_INPUT_TRIANGLES, + COUNTER_GLES_VXCACHE_HIT, + COUNTER_GLES_VXCACHE_MISS, + COUNTER_GLES_VXCACHE_COLLISION, + COUNTER_GLES_CULLED_TRIANGLES, + COUNTER_GLES_CULLED_LINES, + COUNTER_GLES_BACKFACE_TRIANGLES, + COUNTER_GLES_GBCLIP_TRIANGLES, + COUNTER_GLES_GBCLIP_LINES, + COUNTER_GLES_TRIANGLES_DRAWN, + COUNTER_GLES_DRAWCALL_TIME, + COUNTER_GLES_TRIANGLES_COUNT, + COUNTER_GLES_INDEPENDENT_TRIANGLES_COUNT, + COUNTER_GLES_STRIP_TRIANGLES_COUNT, + COUNTER_GLES_FAN_TRIANGLES_COUNT, + COUNTER_GLES_LINES_COUNT, + COUNTER_GLES_INDEPENDENT_LINES_COUNT, + COUNTER_GLES_STRIP_LINES_COUNT, + COUNTER_GLES_LOOP_LINES_COUNT, + + /* Framebuffer capture pseudo-counter */ + COUNTER_FILMSTRIP, + + NUMBER_OF_EVENTS +} _mali_osk_counter_id; + +#define FIRST_ACTIVITY_EVENT ACTIVITY_VP_0 +#define LAST_ACTIVITY_EVENT ACTIVITY_FP_7 + +#define FIRST_HW_COUNTER COUNTER_L2_0_C0 +#define LAST_HW_COUNTER COUNTER_FP_7_C1 + +#define FIRST_SW_COUNTER COUNTER_EGL_BLIT_TIME +#define LAST_SW_COUNTER COUNTER_GLES_LOOP_LINES_COUNT + +#define FIRST_SPECIAL_COUNTER COUNTER_FILMSTRIP +#define LAST_SPECIAL_COUNTER COUNTER_FILMSTRIP + +/** + * Structure to pass performance counter data of a Mali core + */ +struct _mali_profiling_core_counters { + u32 source0; + u32 value0; + u32 source1; + u32 value1; +}; + +/** + * Structure to pass performance counter data of Mali L2 cache cores + */ +struct _mali_profiling_l2_counter_values { + struct _mali_profiling_core_counters cores[MAX_NUM_L2_CACHE_CORES]; +}; + +/** + * Structure to pass data defining Mali instance in use: + * + * mali_product_id - Mali product id + * mali_version_major - Mali version major number + * mali_version_minor - Mali version minor number + * num_of_l2_cores - number of L2 cache cores + * num_of_fp_cores - number of fragment processor cores + * num_of_vp_cores - number of vertex processor cores + */ +struct _mali_profiling_mali_version { + u32 mali_product_id; + u32 mali_version_major; + u32 mali_version_minor; + u32 num_of_l2_cores; + u32 num_of_fp_cores; + u32 num_of_vp_cores; +}; + +/* + * List of possible actions to be controlled by Streamline. + * The following numbers are used by gator to control the frame buffer dumping and s/w counter reporting. + * We cannot use the enums in mali_uk_types.h because they are unknown inside gator. + */ +#define FBDUMP_CONTROL_ENABLE (1) +#define FBDUMP_CONTROL_RATE (2) +#define SW_COUNTER_ENABLE (3) +#define FBDUMP_CONTROL_RESIZE_FACTOR (4) + +void _mali_profiling_control(u32 action, u32 value); + +u32 _mali_profiling_get_l2_counters(struct _mali_profiling_l2_counter_values *values); + +int _mali_profiling_set_event(u32 counter_id, s32 event_id); + +u32 _mali_profiling_get_api_version(void); + +void _mali_profiling_get_mali_version(struct _mali_profiling_mali_version *values); + +#ifdef __cplusplus +} +#endif + +#endif /* __MALI_UTGARD_PROFILING_GATOR_API_H__ */ diff --git a/drivers/gator/mali_midgard.mk b/drivers/gator/mali_midgard.mk new file mode 100644 index 000000000000..1b784d5c3d58 --- /dev/null +++ b/drivers/gator/mali_midgard.mk @@ -0,0 +1,39 @@ +# Defines for Mali-Midgard driver +EXTRA_CFLAGS += -DMALI_USE_UMP=1 \ + -DMALI_LICENSE_IS_GPL=1 \ + -DMALI_BASE_TRACK_MEMLEAK=0 \ + -DMALI_DEBUG=0 \ + -DMALI_ERROR_INJECT_ON=0 \ + -DMALI_CUSTOMER_RELEASE=1 \ + -DMALI_UNIT_TEST=0 \ + -DMALI_BACKEND_KERNEL=1 \ + -DMALI_NO_MALI=0 + +DDK_DIR ?= . +ifneq ($(wildcard $(DDK_DIR)/drivers/gpu/arm/t6xx),) +KBASE_DIR = $(DDK_DIR)/drivers/gpu/arm/t6xx/kbase +OSK_DIR = $(DDK_DIR)/drivers/gpu/arm/t6xx/kbase/osk +endif + +ifneq ($(wildcard $(DDK_DIR)/drivers/gpu/arm/midgard),) +KBASE_DIR = $(DDK_DIR)/drivers/gpu/arm/midgard +OSK_DIR = $(DDK_DIR)/drivers/gpu/arm/midgard/osk +EXTRA_CFLAGS += -DMALI_DIR_MIDGARD=1 +endif + +ifneq ($(wildcard $(DDK_DIR)/drivers/gpu/arm/midgard/mali_kbase_gator_api.h),) +EXTRA_CFLAGS += -DMALI_SIMPLE_API=1 +endif + +UMP_DIR = $(DDK_DIR)/include/linux + +# Include directories in the DDK +EXTRA_CFLAGS += -I$(KBASE_DIR)/ \ + -I$(KBASE_DIR)/.. \ + -I$(OSK_DIR)/.. \ + -I$(UMP_DIR)/.. \ + -I$(DDK_DIR)/include \ + -I$(KBASE_DIR)/osk/src/linux/include \ + -I$(KBASE_DIR)/platform_dummy \ + -I$(KBASE_DIR)/src + diff --git a/drivers/hwmon/lm75.c b/drivers/hwmon/lm75.c index c03b490bba81..1d3600aa4ca5 100644 --- a/drivers/hwmon/lm75.c +++ b/drivers/hwmon/lm75.c @@ -27,6 +27,8 @@ #include <linux/hwmon-sysfs.h> #include <linux/err.h> #include <linux/mutex.h> +#include <linux/of.h> +#include <linux/thermal.h> #include "lm75.h" @@ -70,6 +72,7 @@ static const u8 LM75_REG_TEMP[3] = { /* Each client has this additional data */ struct lm75_data { struct device *hwmon_dev; + struct thermal_zone_device *tz; struct mutex update_lock; u8 orig_conf; u8 resolution; /* In bits, between 9 and 12 */ @@ -90,22 +93,36 @@ static struct lm75_data *lm75_update_device(struct device *dev); /*-----------------------------------------------------------------------*/ +static inline long lm75_reg_to_mc(s16 temp, u8 resolution) +{ + return ((temp >> (16 - resolution)) * 1000) >> (resolution - 8); +} + /* sysfs attributes for hwmon */ +static int lm75_read_temp(void *dev, long *temp) +{ + struct lm75_data *data = lm75_update_device(dev); + + if (IS_ERR(data)) + return PTR_ERR(data); + + *temp = lm75_reg_to_mc(data->temp[0], data->resolution); + + return 0; +} + static ssize_t show_temp(struct device *dev, struct device_attribute *da, char *buf) { struct sensor_device_attribute *attr = to_sensor_dev_attr(da); struct lm75_data *data = lm75_update_device(dev); - long temp; if (IS_ERR(data)) return PTR_ERR(data); - temp = ((data->temp[attr->index] >> (16 - data->resolution)) * 1000) - >> (data->resolution - 8); - - return sprintf(buf, "%ld\n", temp); + return sprintf(buf, "%ld\n", lm75_reg_to_mc(data->temp[attr->index], + data->resolution)); } static ssize_t set_temp(struct device *dev, struct device_attribute *da, @@ -271,6 +288,13 @@ lm75_probe(struct i2c_client *client, const struct i2c_device_id *id) goto exit_remove; } + data->tz = thermal_zone_of_sensor_register(&client->dev, + 0, + &client->dev, + lm75_read_temp, NULL); + if (IS_ERR(data->tz)) + data->tz = NULL; + dev_info(&client->dev, "%s: sensor '%s'\n", dev_name(data->hwmon_dev), client->name); @@ -285,6 +309,7 @@ static int lm75_remove(struct i2c_client *client) { struct lm75_data *data = i2c_get_clientdata(client); + thermal_zone_of_sensor_unregister(&client->dev, data->tz); hwmon_device_unregister(data->hwmon_dev); sysfs_remove_group(&client->dev.kobj, &lm75_group); lm75_write_value(client, LM75_REG_CONF, data->orig_conf); diff --git a/drivers/hwmon/tmp102.c b/drivers/hwmon/tmp102.c index d7b47abf37fe..6748b4583e7b 100644 --- a/drivers/hwmon/tmp102.c +++ b/drivers/hwmon/tmp102.c @@ -27,6 +27,8 @@ #include <linux/mutex.h> #include <linux/device.h> #include <linux/jiffies.h> +#include <linux/thermal.h> +#include <linux/of.h> #define DRIVER_NAME "tmp102" @@ -50,6 +52,7 @@ struct tmp102 { struct device *hwmon_dev; + struct thermal_zone_device *tz; struct mutex lock; u16 config_orig; unsigned long last_update; @@ -93,6 +96,15 @@ static struct tmp102 *tmp102_update_device(struct i2c_client *client) return tmp102; } +static int tmp102_read_temp(void *dev, long *temp) +{ + struct tmp102 *tmp102 = tmp102_update_device(to_i2c_client(dev)); + + *temp = tmp102->temp[0]; + + return 0; +} + static ssize_t tmp102_show_temp(struct device *dev, struct device_attribute *attr, char *buf) @@ -204,6 +216,12 @@ static int tmp102_probe(struct i2c_client *client, goto fail_remove_sysfs; } + tmp102->tz = thermal_zone_of_sensor_register(&client->dev, 0, + &client->dev, + tmp102_read_temp, NULL); + if (IS_ERR(tmp102->tz)) + tmp102->tz = NULL; + dev_info(&client->dev, "initialized\n"); return 0; @@ -220,6 +238,7 @@ static int tmp102_remove(struct i2c_client *client) { struct tmp102 *tmp102 = i2c_get_clientdata(client); + thermal_zone_of_sensor_unregister(&client->dev, tmp102->tz); hwmon_device_unregister(tmp102->hwmon_dev); sysfs_remove_group(&client->dev.kobj, &tmp102_attr_group); diff --git a/drivers/irqchip/Kconfig b/drivers/irqchip/Kconfig index 4a33351c25dc..7709b1dbf6cc 100644 --- a/drivers/irqchip/Kconfig +++ b/drivers/irqchip/Kconfig @@ -10,6 +10,11 @@ config ARM_GIC config GIC_NON_BANKED bool +config ARM_GIC_V3 + bool + select IRQ_DOMAIN + select MULTI_IRQ_HANDLER + config ARM_VIC bool select IRQ_DOMAIN diff --git a/drivers/irqchip/Makefile b/drivers/irqchip/Makefile index cda4cb5f7327..bf4667b34306 100644 --- a/drivers/irqchip/Makefile +++ b/drivers/irqchip/Makefile @@ -9,7 +9,8 @@ obj-$(CONFIG_METAG) += irq-metag-ext.o obj-$(CONFIG_METAG_PERFCOUNTER_IRQS) += irq-metag.o obj-$(CONFIG_ARCH_SUNXI) += irq-sun4i.o obj-$(CONFIG_ARCH_SPEAR3XX) += spear-shirq.o -obj-$(CONFIG_ARM_GIC) += irq-gic.o +obj-$(CONFIG_ARM_GIC) += irq-gic.o irq-gic-common.o +obj-$(CONFIG_ARM_GIC_V3) += irq-gic-v3.o irq-gic-common.o obj-$(CONFIG_ARM_VIC) += irq-vic.o obj-$(CONFIG_SIRF_IRQ) += irq-sirfsoc.o obj-$(CONFIG_RENESAS_INTC_IRQPIN) += irq-renesas-intc-irqpin.o diff --git a/drivers/irqchip/irq-gic-common.c b/drivers/irqchip/irq-gic-common.c new file mode 100644 index 000000000000..60ac704d2090 --- /dev/null +++ b/drivers/irqchip/irq-gic-common.c @@ -0,0 +1,115 @@ +/* + * Copyright (C) 2002 ARM Limited, All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#include <linux/interrupt.h> +#include <linux/io.h> +#include <linux/irq.h> +#include <linux/irqchip/arm-gic.h> + +#include "irq-gic-common.h" + +void gic_configure_irq(unsigned int irq, unsigned int type, + void __iomem *base, void (*sync_access)(void)) +{ + u32 enablemask = 1 << (irq % 32); + u32 enableoff = (irq / 32) * 4; + u32 confmask = 0x2 << ((irq % 16) * 2); + u32 confoff = (irq / 16) * 4; + bool enabled = false; + u32 val; + + /* + * Read current configuration register, and insert the config + * for "irq", depending on "type". + */ + val = readl_relaxed(base + GIC_DIST_CONFIG + confoff); + if (type == IRQ_TYPE_LEVEL_HIGH) + val &= ~confmask; + else if (type == IRQ_TYPE_EDGE_RISING) + val |= confmask; + + /* + * As recommended by the spec, disable the interrupt before changing + * the configuration + */ + if (readl_relaxed(base + GIC_DIST_ENABLE_SET + enableoff) & enablemask) { + writel_relaxed(enablemask, base + GIC_DIST_ENABLE_CLEAR + enableoff); + if (sync_access) + sync_access(); + enabled = true; + } + + /* + * Write back the new configuration, and possibly re-enable + * the interrupt. + */ + writel_relaxed(val, base + GIC_DIST_CONFIG + confoff); + + if (enabled) + writel_relaxed(enablemask, base + GIC_DIST_ENABLE_SET + enableoff); + + if (sync_access) + sync_access(); +} + +void __init gic_dist_config(void __iomem *base, int gic_irqs, + void (*sync_access)(void)) +{ + unsigned int i; + + /* + * Set all global interrupts to be level triggered, active low. + */ + for (i = 32; i < gic_irqs; i += 16) + writel_relaxed(0, base + GIC_DIST_CONFIG + i / 4); + + /* + * Set priority on all global interrupts. + */ + for (i = 32; i < gic_irqs; i += 4) + writel_relaxed(0xa0a0a0a0, base + GIC_DIST_PRI + i); + + /* + * Disable all interrupts. Leave the PPI and SGIs alone + * as they are enabled by redistributor registers. + */ + for (i = 32; i < gic_irqs; i += 32) + writel_relaxed(0xffffffff, base + GIC_DIST_ENABLE_CLEAR + i / 8); + + if (sync_access) + sync_access(); +} + +void gic_cpu_config(void __iomem *base, void (*sync_access)(void)) +{ + int i; + + /* + * Deal with the banked PPI and SGI interrupts - disable all + * PPI interrupts, ensure all SGI interrupts are enabled. + */ + writel_relaxed(0xffff0000, base + GIC_DIST_ENABLE_CLEAR); + writel_relaxed(0x0000ffff, base + GIC_DIST_ENABLE_SET); + + /* + * Set priority on PPI and SGI interrupts + */ + for (i = 0; i < 32; i += 4) + writel_relaxed(0xa0a0a0a0, base + GIC_DIST_PRI + i * 4 / 4); + + if (sync_access) + sync_access(); +} diff --git a/drivers/irqchip/irq-gic-common.h b/drivers/irqchip/irq-gic-common.h new file mode 100644 index 000000000000..b41f02481c3a --- /dev/null +++ b/drivers/irqchip/irq-gic-common.h @@ -0,0 +1,29 @@ +/* + * Copyright (C) 2002 ARM Limited, All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#ifndef _IRQ_GIC_COMMON_H +#define _IRQ_GIC_COMMON_H + +#include <linux/of.h> +#include <linux/irqdomain.h> + +void gic_configure_irq(unsigned int irq, unsigned int type, + void __iomem *base, void (*sync_access)(void)); +void gic_dist_config(void __iomem *base, int gic_irqs, + void (*sync_access)(void)); +void gic_cpu_config(void __iomem *base, void (*sync_access)(void)); + +#endif /* _IRQ_GIC_COMMON_H */ diff --git a/drivers/irqchip/irq-gic-v3.c b/drivers/irqchip/irq-gic-v3.c new file mode 100644 index 000000000000..57eaa5a0b1e3 --- /dev/null +++ b/drivers/irqchip/irq-gic-v3.c @@ -0,0 +1,692 @@ +/* + * Copyright (C) 2013, 2014 ARM Limited, All Rights Reserved. + * Author: Marc Zyngier <marc.zyngier@arm.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#include <linux/cpu.h> +#include <linux/delay.h> +#include <linux/interrupt.h> +#include <linux/of.h> +#include <linux/of_address.h> +#include <linux/of_irq.h> +#include <linux/percpu.h> +#include <linux/slab.h> + +#include <linux/irqchip/arm-gic-v3.h> + +#include <asm/cputype.h> +#include <asm/exception.h> +#include <asm/smp_plat.h> + +#include "irq-gic-common.h" +#include "irqchip.h" + +struct gic_chip_data { + void __iomem *dist_base; + void __iomem **redist_base; + void __percpu __iomem **rdist; + struct irq_domain *domain; + u64 redist_stride; + u32 redist_regions; + unsigned int irq_nr; +}; + +static struct gic_chip_data gic_data __read_mostly; + +#define gic_data_rdist() (this_cpu_ptr(gic_data.rdist)) +#define gic_data_rdist_rd_base() (*gic_data_rdist()) +#define gic_data_rdist_sgi_base() (gic_data_rdist_rd_base() + SZ_64K) + +/* Our default, arbitrary priority value. Linux only uses one anyway. */ +#define DEFAULT_PMR_VALUE 0xf0 + +static inline unsigned int gic_irq(struct irq_data *d) +{ + return d->hwirq; +} + +static inline int gic_irq_in_rdist(struct irq_data *d) +{ + return gic_irq(d) < 32; +} + +static inline void __iomem *gic_dist_base(struct irq_data *d) +{ + if (gic_irq_in_rdist(d)) /* SGI+PPI -> SGI_base for this CPU */ + return gic_data_rdist_sgi_base(); + + if (d->hwirq <= 1023) /* SPI -> dist_base */ + return gic_data.dist_base; + + if (d->hwirq >= 8192) + BUG(); /* LPI Detected!!! */ + + return NULL; +} + +static void gic_do_wait_for_rwp(void __iomem *base) +{ + u32 count = 1000000; /* 1s! */ + + while (readl_relaxed(base + GICD_CTLR) & GICD_CTLR_RWP) { + count--; + if (!count) { + pr_err_ratelimited("RWP timeout, gone fishing\n"); + return; + } + cpu_relax(); + udelay(1); + }; +} + +/* Wait for completion of a distributor change */ +static void gic_dist_wait_for_rwp(void) +{ + gic_do_wait_for_rwp(gic_data.dist_base); +} + +/* Wait for completion of a redistributor change */ +static void gic_redist_wait_for_rwp(void) +{ + gic_do_wait_for_rwp(gic_data_rdist_rd_base()); +} + +/* Low level accessors */ +static u64 gic_read_iar(void) +{ + u64 irqstat; + + asm volatile("mrs_s %0, " __stringify(ICC_IAR1_EL1) : "=r" (irqstat)); + return irqstat; +} + +static void gic_write_pmr(u64 val) +{ + asm volatile("msr_s " __stringify(ICC_PMR_EL1) ", %0" : : "r" (val)); +} + +static void gic_write_ctlr(u64 val) +{ + asm volatile("msr_s " __stringify(ICC_CTLR_EL1) ", %0" : : "r" (val)); + isb(); +} + +static void gic_write_grpen1(u64 val) +{ + asm volatile("msr_s " __stringify(ICC_GRPEN1_EL1) ", %0" : : "r" (val)); + isb(); +} + +static void gic_write_sgi1r(u64 val) +{ + asm volatile("msr_s " __stringify(ICC_SGI1R_EL1) ", %0" : : "r" (val)); +} + +static void gic_enable_sre(void) +{ + u64 val; + + asm volatile("mrs_s %0, " __stringify(ICC_SRE_EL1) : "=r" (val)); + val |= ICC_SRE_EL1_SRE; + asm volatile("msr_s " __stringify(ICC_SRE_EL1) ", %0" : : "r" (val)); + isb(); + + /* + * Need to check that the SRE bit has actually been set. If + * not, it means that SRE is disabled at EL2. We're going to + * die painfully, and there is nothing we can do about it. + * + * Kindly inform the luser. + */ + asm volatile("mrs_s %0, " __stringify(ICC_SRE_EL1) : "=r" (val)); + if (!(val & ICC_SRE_EL1_SRE)) + pr_err("GIC: unable to set SRE (disabled at EL2), panic ahead\n"); +} + +static void gic_enable_redist(void) +{ + void __iomem *rbase; + u32 count = 1000000; /* 1s! */ + u32 val; + + rbase = gic_data_rdist_rd_base(); + + /* Wake up this CPU redistributor */ + val = readl_relaxed(rbase + GICR_WAKER); + val &= ~GICR_WAKER_ProcessorSleep; + writel_relaxed(val, rbase + GICR_WAKER); + + while (readl_relaxed(rbase + GICR_WAKER) & GICR_WAKER_ChildrenAsleep) { + count--; + if (!count) { + pr_err_ratelimited("redist didn't wake up...\n"); + return; + } + cpu_relax(); + udelay(1); + }; +} + +/* + * Routines to disable, enable, EOI and route interrupts + */ +static void gic_poke_irq(struct irq_data *d, u32 offset) +{ + u32 mask = 1 << (gic_irq(d) % 32); + void (*rwp_wait)(void); + void __iomem *base; + + if (gic_irq_in_rdist(d)) { + base = gic_data_rdist_sgi_base(); + rwp_wait = gic_redist_wait_for_rwp; + } else { + base = gic_data.dist_base; + rwp_wait = gic_dist_wait_for_rwp; + } + + writel_relaxed(mask, base + offset + (gic_irq(d) / 32) * 4); + rwp_wait(); +} + +static int gic_peek_irq(struct irq_data *d, u32 offset) +{ + u32 mask = 1 << (gic_irq(d) % 32); + void __iomem *base; + + if (gic_irq_in_rdist(d)) + base = gic_data_rdist_sgi_base(); + else + base = gic_data.dist_base; + + return !!(readl_relaxed(base + offset + (gic_irq(d) / 32) * 4) & mask); +} + +static void gic_mask_irq(struct irq_data *d) +{ + gic_poke_irq(d, GICD_ICENABLER); +} + +static void gic_unmask_irq(struct irq_data *d) +{ + gic_poke_irq(d, GICD_ISENABLER); +} + +static void gic_eoi_irq(struct irq_data *d) +{ + gic_write_eoir(gic_irq(d)); +} + +static int gic_set_type(struct irq_data *d, unsigned int type) +{ + unsigned int irq = gic_irq(d); + void (*rwp_wait)(void); + void __iomem *base; + + /* Interrupt configuration for SGIs can't be changed */ + if (irq < 16) + return -EINVAL; + + if (type != IRQ_TYPE_LEVEL_HIGH && type != IRQ_TYPE_EDGE_RISING) + return -EINVAL; + + if (gic_irq_in_rdist(d)) { + base = gic_data_rdist_sgi_base(); + rwp_wait = gic_redist_wait_for_rwp; + } else { + base = gic_data.dist_base; + rwp_wait = gic_dist_wait_for_rwp; + } + + gic_configure_irq(irq, type, base, rwp_wait); + + return 0; +} + +static u64 gic_mpidr_to_affinity(u64 mpidr) +{ + u64 aff; + + aff = (MPIDR_AFFINITY_LEVEL(mpidr, 3) << 32 | + MPIDR_AFFINITY_LEVEL(mpidr, 2) << 16 | + MPIDR_AFFINITY_LEVEL(mpidr, 1) << 8 | + MPIDR_AFFINITY_LEVEL(mpidr, 0)); + + return aff; +} + +static asmlinkage void __exception_irq_entry gic_handle_irq(struct pt_regs *regs) +{ + u64 irqnr; + + do { + irqnr = gic_read_iar(); + + if (likely(irqnr > 15 && irqnr < 1020)) { + u64 irq = irq_find_mapping(gic_data.domain, irqnr); + if (likely(irq)) { + handle_IRQ(irq, regs); + continue; + } + + WARN_ONCE(true, "Unexpected SPI received!\n"); + gic_write_eoir(irqnr); + } + if (irqnr < 16) { + gic_write_eoir(irqnr); +#ifdef CONFIG_SMP + handle_IPI(irqnr, regs); +#else + WARN_ONCE(true, "Unexpected SGI received!\n"); +#endif + continue; + } + } while (irqnr != ICC_IAR1_EL1_SPURIOUS); +} + +static void __init gic_dist_init(void) +{ + unsigned int i; + u64 affinity; + void __iomem *base = gic_data.dist_base; + + /* Disable the distributor */ + writel_relaxed(0, base + GICD_CTLR); + gic_dist_wait_for_rwp(); + + gic_dist_config(base, gic_data.irq_nr, gic_dist_wait_for_rwp); + + /* Enable distributor with ARE, Group1 */ + writel_relaxed(GICD_CTLR_ARE_NS | GICD_CTLR_ENABLE_G1A | GICD_CTLR_ENABLE_G1, + base + GICD_CTLR); + + /* + * Set all global interrupts to the boot CPU only. ARE must be + * enabled. + */ + affinity = gic_mpidr_to_affinity(cpu_logical_map(smp_processor_id())); + for (i = 32; i < gic_data.irq_nr; i++) + writeq_relaxed(affinity, base + GICD_IROUTER + i * 8); +} + +static int gic_populate_rdist(void) +{ + u64 mpidr = cpu_logical_map(smp_processor_id()); + u64 typer; + u32 aff; + int i; + + /* + * Convert affinity to a 32bit value that can be matched to + * GICR_TYPER bits [63:32]. + */ + aff = (MPIDR_AFFINITY_LEVEL(mpidr, 3) << 24 | + MPIDR_AFFINITY_LEVEL(mpidr, 2) << 16 | + MPIDR_AFFINITY_LEVEL(mpidr, 1) << 8 | + MPIDR_AFFINITY_LEVEL(mpidr, 0)); + + for (i = 0; i < gic_data.redist_regions; i++) { + void __iomem *ptr = gic_data.redist_base[i]; + u32 reg; + + reg = readl_relaxed(ptr + GICR_PIDR2) & GIC_PIDR2_ARCH_MASK; + if (reg != GIC_PIDR2_ARCH_GICv3 && + reg != GIC_PIDR2_ARCH_GICv4) { /* We're in trouble... */ + pr_warn("No redistributor present @%p\n", ptr); + break; + } + + do { + typer = readq_relaxed(ptr + GICR_TYPER); + if ((typer >> 32) == aff) { + gic_data_rdist_rd_base() = ptr; + pr_info("CPU%d: found redistributor %llx @%p\n", + smp_processor_id(), + (unsigned long long)mpidr, ptr); + return 0; + } + + if (gic_data.redist_stride) { + ptr += gic_data.redist_stride; + } else { + ptr += SZ_64K * 2; /* Skip RD_base + SGI_base */ + if (typer & GICR_TYPER_VLPIS) + ptr += SZ_64K * 2; /* Skip VLPI_base + reserved page */ + } + } while (!(typer & GICR_TYPER_LAST)); + } + + /* We couldn't even deal with ourselves... */ + WARN(true, "CPU%d: mpidr %llx has no re-distributor!\n", + smp_processor_id(), (unsigned long long)mpidr); + return -ENODEV; +} + +static void gic_cpu_init(void) +{ + void __iomem *rbase; + + /* Register ourselves with the rest of the world */ + if (gic_populate_rdist()) + return; + + gic_enable_redist(); + + rbase = gic_data_rdist_sgi_base(); + + gic_cpu_config(rbase, gic_redist_wait_for_rwp); + + /* Enable system registers */ + gic_enable_sre(); + + /* Set priority mask register */ + gic_write_pmr(DEFAULT_PMR_VALUE); + + /* EOI deactivates interrupt too (mode 0) */ + gic_write_ctlr(ICC_CTLR_EL1_EOImode_drop_dir); + + /* ... and let's hit the road... */ + gic_write_grpen1(1); +} + +#ifdef CONFIG_SMP +static int gic_secondary_init(struct notifier_block *nfb, + unsigned long action, void *hcpu) +{ + if (action == CPU_STARTING || action == CPU_STARTING_FROZEN) + gic_cpu_init(); + return NOTIFY_OK; +} + +/* + * Notifier for enabling the GIC CPU interface. Set an arbitrarily high + * priority because the GIC needs to be up before the ARM generic timers. + */ +static struct notifier_block gic_cpu_notifier = { + .notifier_call = gic_secondary_init, + .priority = 100, +}; + +static u16 gic_compute_target_list(int *base_cpu, const struct cpumask *mask, + u64 cluster_id) +{ + int cpu = *base_cpu; + u64 mpidr = cpu_logical_map(cpu); + u16 tlist = 0; + + while (cpu < nr_cpu_ids) { + /* + * If we ever get a cluster of more than 16 CPUs, just + * scream and skip that CPU. + */ + if (WARN_ON((mpidr & 0xff) >= 16)) + goto out; + + tlist |= 1 << (mpidr & 0xf); + + cpu = cpumask_next(cpu, mask); + if (cpu == nr_cpu_ids) + goto out; + + mpidr = cpu_logical_map(cpu); + + if (cluster_id != (mpidr & ~0xffUL)) { + cpu--; + goto out; + } + } +out: + *base_cpu = cpu; + return tlist; +} + +static void gic_send_sgi(u64 cluster_id, u16 tlist, unsigned int irq) +{ + u64 val; + + val = (MPIDR_AFFINITY_LEVEL(cluster_id, 3) << 48 | + MPIDR_AFFINITY_LEVEL(cluster_id, 2) << 32 | + irq << 24 | + MPIDR_AFFINITY_LEVEL(cluster_id, 1) << 16 | + tlist); + + pr_debug("CPU%d: ICC_SGI1R_EL1 %llx\n", smp_processor_id(), val); + gic_write_sgi1r(val); +} + +static void gic_raise_softirq(const struct cpumask *mask, unsigned int irq) +{ + int cpu; + + if (WARN_ON(irq >= 16)) + return; + + /* + * Ensure that stores to Normal memory are visible to the + * other CPUs before issuing the IPI. + */ + smp_wmb(); + + for_each_cpu_mask(cpu, *mask) { + u64 cluster_id = cpu_logical_map(cpu) & ~0xffUL; + u16 tlist; + + tlist = gic_compute_target_list(&cpu, mask, cluster_id); + gic_send_sgi(cluster_id, tlist, irq); + } + + /* Force the above writes to ICC_SGI1R_EL1 to be executed */ + isb(); +} + +static void gic_smp_init(void) +{ + set_smp_cross_call(gic_raise_softirq); + register_cpu_notifier(&gic_cpu_notifier); +} + +static int gic_set_affinity(struct irq_data *d, const struct cpumask *mask_val, + bool force) +{ + unsigned int cpu = cpumask_any_and(mask_val, cpu_online_mask); + void __iomem *reg; + int enabled; + u64 val; + + if (gic_irq_in_rdist(d)) + return -EINVAL; + + /* If interrupt was enabled, disable it first */ + enabled = gic_peek_irq(d, GICD_ISENABLER); + if (enabled) + gic_mask_irq(d); + + reg = gic_dist_base(d) + GICD_IROUTER + (gic_irq(d) * 8); + val = gic_mpidr_to_affinity(cpu_logical_map(cpu)); + + writeq_relaxed(val, reg); + + /* + * If the interrupt was enabled, enabled it again. Otherwise, + * just wait for the distributor to have digested our changes. + */ + if (enabled) + gic_unmask_irq(d); + else + gic_dist_wait_for_rwp(); + + return IRQ_SET_MASK_OK; +} +#else +#define gic_set_affinity NULL +#define gic_smp_init() do { } while(0) +#endif + +static struct irq_chip gic_chip = { + .name = "GICv3", + .irq_mask = gic_mask_irq, + .irq_unmask = gic_unmask_irq, + .irq_eoi = gic_eoi_irq, + .irq_set_type = gic_set_type, + .irq_set_affinity = gic_set_affinity, +}; + +static int gic_irq_domain_map(struct irq_domain *d, unsigned int irq, + irq_hw_number_t hw) +{ + /* SGIs are private to the core kernel */ + if (hw < 16) + return -EPERM; + /* PPIs */ + if (hw < 32) { + irq_set_percpu_devid(irq); + irq_set_chip_and_handler(irq, &gic_chip, + handle_percpu_devid_irq); + set_irq_flags(irq, IRQF_VALID | IRQF_NOAUTOEN); + } + /* SPIs */ + if (hw >= 32 && hw < gic_data.irq_nr) { + irq_set_chip_and_handler(irq, &gic_chip, + handle_fasteoi_irq); + set_irq_flags(irq, IRQF_VALID | IRQF_PROBE); + } + irq_set_chip_data(irq, d->host_data); + return 0; +} + +static int gic_irq_domain_xlate(struct irq_domain *d, + struct device_node *controller, + const u32 *intspec, unsigned int intsize, + unsigned long *out_hwirq, unsigned int *out_type) +{ + if (d->of_node != controller) + return -EINVAL; + if (intsize < 3) + return -EINVAL; + + switch(intspec[0]) { + case 0: /* SPI */ + *out_hwirq = intspec[1] + 32; + break; + case 1: /* PPI */ + *out_hwirq = intspec[1] + 16; + break; + default: + return -EINVAL; + } + + *out_type = intspec[2] & IRQ_TYPE_SENSE_MASK; + return 0; +} + +static const struct irq_domain_ops gic_irq_domain_ops = { + .map = gic_irq_domain_map, + .xlate = gic_irq_domain_xlate, +}; + +static int __init gic_of_init(struct device_node *node, struct device_node *parent) +{ + void __iomem *dist_base; + void __iomem **redist_base; + u64 redist_stride; + u32 redist_regions; + u32 reg; + int gic_irqs; + int err; + int i; + + dist_base = of_iomap(node, 0); + if (!dist_base) { + pr_err("%s: unable to map gic dist registers\n", + node->full_name); + return -ENXIO; + } + + reg = readl_relaxed(dist_base + GICD_PIDR2) & GIC_PIDR2_ARCH_MASK; + if (reg != GIC_PIDR2_ARCH_GICv3 && reg != GIC_PIDR2_ARCH_GICv4) { + pr_err("%s: no distributor detected, giving up\n", + node->full_name); + err = -ENODEV; + goto out_unmap_dist; + } + + if (of_property_read_u32(node, "#redistributor-regions", &redist_regions)) + redist_regions = 1; + + redist_base = kzalloc(sizeof(*redist_base) * redist_regions, GFP_KERNEL); + if (!redist_base) { + err = -ENOMEM; + goto out_unmap_dist; + } + + for (i = 0; i < redist_regions; i++) { + redist_base[i] = of_iomap(node, 1 + i); + if (!redist_base[i]) { + pr_err("%s: couldn't map region %d\n", + node->full_name, i); + err = -ENODEV; + goto out_unmap_rdist; + } + } + + if (of_property_read_u64(node, "redistributor-stride", &redist_stride)) + redist_stride = 0; + + gic_data.dist_base = dist_base; + gic_data.redist_base = redist_base; + gic_data.redist_regions = redist_regions; + gic_data.redist_stride = redist_stride; + + /* + * Find out how many interrupts are supported. + * The GIC only supports up to 1020 interrupt sources (SGI+PPI+SPI) + */ + gic_irqs = readl_relaxed(gic_data.dist_base + GICD_TYPER) & 0x1f; + gic_irqs = (gic_irqs + 1) * 32; + if (gic_irqs > 1020) + gic_irqs = 1020; + gic_data.irq_nr = gic_irqs; + + gic_data.domain = irq_domain_add_tree(node, &gic_irq_domain_ops, + &gic_data); + gic_data.rdist = alloc_percpu(typeof(*gic_data.rdist)); + + if (WARN_ON(!gic_data.domain) || WARN_ON(!gic_data.rdist)) { + err = -ENOMEM; + goto out_free; + } + + set_handle_irq(gic_handle_irq); + + gic_smp_init(); + gic_dist_init(); + gic_cpu_init(); + + return 0; + +out_free: + if (gic_data.domain) + irq_domain_remove(gic_data.domain); + free_percpu(gic_data.rdist); +out_unmap_rdist: + for (i = 0; i < redist_regions; i++) + if (redist_base[i]) + iounmap(redist_base[i]); + kfree(redist_base); +out_unmap_dist: + iounmap(dist_base); + return err; +} + +IRQCHIP_DECLARE(gic_v3, "arm,gic-v3", gic_of_init); diff --git a/drivers/irqchip/irq-gic.c b/drivers/irqchip/irq-gic.c index c8ee1cb023b8..4cb670e61707 100644 --- a/drivers/irqchip/irq-gic.c +++ b/drivers/irqchip/irq-gic.c @@ -1,6 +1,4 @@ /* - * linux/arch/arm/common/gic.c - * * Copyright (C) 2002 ARM Limited, All Rights Reserved. * * This program is free software; you can redistribute it and/or modify @@ -41,17 +39,19 @@ #include <linux/slab.h> #include <linux/irqchip/chained_irq.h> #include <linux/irqchip/arm-gic.h> +#include <trace/events/arm-ipi.h> #include <asm/cputype.h> #include <asm/irq.h> #include <asm/exception.h> #include <asm/smp_plat.h> +#include "irq-gic-common.h" #include "irqchip.h" union gic_base { void __iomem *common_base; - void __percpu __iomem **percpu_base; + void __percpu * __iomem *percpu_base; }; struct gic_chip_data { @@ -189,12 +189,6 @@ static int gic_set_type(struct irq_data *d, unsigned int type) { void __iomem *base = gic_dist_base(d); unsigned int gicirq = gic_irq(d); - u32 enablemask = 1 << (gicirq % 32); - u32 enableoff = (gicirq / 32) * 4; - u32 confmask = 0x2 << ((gicirq % 16) * 2); - u32 confoff = (gicirq / 16) * 4; - bool enabled = false; - u32 val; /* Interrupt configuration for SGIs can't be changed */ if (gicirq < 16) @@ -208,25 +202,7 @@ static int gic_set_type(struct irq_data *d, unsigned int type) if (gic_arch_extn.irq_set_type) gic_arch_extn.irq_set_type(d, type); - val = readl_relaxed(base + GIC_DIST_CONFIG + confoff); - if (type == IRQ_TYPE_LEVEL_HIGH) - val &= ~confmask; - else if (type == IRQ_TYPE_EDGE_RISING) - val |= confmask; - - /* - * As recommended by the spec, disable the interrupt before changing - * the configuration - */ - if (readl_relaxed(base + GIC_DIST_ENABLE_SET + enableoff) & enablemask) { - writel_relaxed(enablemask, base + GIC_DIST_ENABLE_CLEAR + enableoff); - enabled = true; - } - - writel_relaxed(val, base + GIC_DIST_CONFIG + confoff); - - if (enabled) - writel_relaxed(enablemask, base + GIC_DIST_ENABLE_SET + enableoff); + gic_configure_irq(gicirq, type, base, NULL); raw_spin_unlock(&irq_controller_lock); @@ -247,21 +223,16 @@ static int gic_set_affinity(struct irq_data *d, const struct cpumask *mask_val, bool force) { void __iomem *reg = gic_dist_base(d) + GIC_DIST_TARGET + (gic_irq(d) & ~3); - unsigned int cpu, shift = (gic_irq(d) % 4) * 8; + unsigned int shift = (gic_irq(d) % 4) * 8; + unsigned int cpu = cpumask_any_and(mask_val, cpu_online_mask); u32 val, mask, bit; - if (!force) - cpu = cpumask_any_and(mask_val, cpu_online_mask); - else - cpu = cpumask_first(mask_val); - if (cpu >= NR_GIC_CPU_IF || cpu >= nr_cpu_ids) return -EINVAL; + raw_spin_lock(&irq_controller_lock); mask = 0xff << shift; bit = gic_cpu_map[cpu] << shift; - - raw_spin_lock(&irq_controller_lock); val = readl_relaxed(reg) & ~mask; writel_relaxed(val | bit, reg); raw_spin_unlock(&irq_controller_lock); @@ -285,7 +256,7 @@ static int gic_set_wake(struct irq_data *d, unsigned int on) #define gic_set_wake NULL #endif -static asmlinkage void __exception_irq_entry gic_handle_irq(struct pt_regs *regs) +static void __exception_irq_entry gic_handle_irq(struct pt_regs *regs) { u32 irqstat, irqnr; struct gic_chip_data *gic = &gic_data[0]; @@ -293,7 +264,7 @@ static asmlinkage void __exception_irq_entry gic_handle_irq(struct pt_regs *regs do { irqstat = readl_relaxed(cpu_base + GIC_CPU_INTACK); - irqnr = irqstat & ~0x1c00; + irqnr = irqstat & GICC_IAR_INT_ID_MASK; if (likely(irqnr > 15 && irqnr < 1021)) { irqnr = irq_find_mapping(gic->domain, irqnr); @@ -389,12 +360,6 @@ static void __init gic_dist_init(struct gic_chip_data *gic) writel_relaxed(0, base + GIC_DIST_CTRL); /* - * Set all global interrupts to be level triggered, active low. - */ - for (i = 32; i < gic_irqs; i += 16) - writel_relaxed(0, base + GIC_DIST_CONFIG + i * 4 / 16); - - /* * Set all global interrupts to this CPU only. */ cpumask = gic_get_cpumask(gic); @@ -403,18 +368,7 @@ static void __init gic_dist_init(struct gic_chip_data *gic) for (i = 32; i < gic_irqs; i += 4) writel_relaxed(cpumask, base + GIC_DIST_TARGET + i * 4 / 4); - /* - * Set priority on all global interrupts. - */ - for (i = 32; i < gic_irqs; i += 4) - writel_relaxed(0xa0a0a0a0, base + GIC_DIST_PRI + i * 4 / 4); - - /* - * Disable all interrupts. Leave the PPI and SGIs alone - * as these enables are banked registers. - */ - for (i = 32; i < gic_irqs; i += 32) - writel_relaxed(0xffffffff, base + GIC_DIST_ENABLE_CLEAR + i * 4 / 32); + gic_dist_config(base, gic_irqs, NULL); writel_relaxed(1, base + GIC_DIST_CTRL); } @@ -441,23 +395,18 @@ static void __cpuinit gic_cpu_init(struct gic_chip_data *gic) if (i != cpu) gic_cpu_map[i] &= ~cpu_mask; - /* - * Deal with the banked PPI and SGI interrupts - disable all - * PPI interrupts, ensure all SGI interrupts are enabled. - */ - writel_relaxed(0xffff0000, dist_base + GIC_DIST_ENABLE_CLEAR); - writel_relaxed(0x0000ffff, dist_base + GIC_DIST_ENABLE_SET); - - /* - * Set priority on PPI and SGI interrupts - */ - for (i = 0; i < 32; i += 4) - writel_relaxed(0xa0a0a0a0, dist_base + GIC_DIST_PRI + i * 4 / 4); + gic_cpu_config(dist_base, NULL); writel_relaxed(0xf0, base + GIC_CPU_PRIMASK); writel_relaxed(1, base + GIC_CPU_CTRL); } +void gic_cpu_if_down(void) +{ + void __iomem *cpu_base = gic_data_cpu_base(&gic_data[0]); + writel_relaxed(0, cpu_base + GIC_CPU_CTRL); +} + #ifdef CONFIG_CPU_PM /* * Saves the GIC distributor registers during suspend or idle. Must be called @@ -648,26 +597,172 @@ static void __init gic_pm_init(struct gic_chip_data *gic) #endif #ifdef CONFIG_SMP -void gic_raise_softirq(const struct cpumask *mask, unsigned int irq) +static void gic_raise_softirq(const struct cpumask *mask, unsigned int irq) { int cpu; - unsigned long map = 0; + unsigned long flags, map = 0; + + raw_spin_lock_irqsave(&irq_controller_lock, flags); /* Convert our logical CPU mask into a physical one. */ - for_each_cpu(cpu, mask) + for_each_cpu(cpu, mask) { + trace_arm_ipi_send(irq, cpu); map |= gic_cpu_map[cpu]; + } /* * Ensure that stores to Normal memory are visible to the - * other CPUs before issuing the IPI. + * other CPUs before they observe us issuing the IPI. */ - dsb(); + dmb(ishst); /* this always happens on GIC0 */ writel_relaxed(map << 16 | irq, gic_data_dist_base(&gic_data[0]) + GIC_DIST_SOFTINT); + + raw_spin_unlock_irqrestore(&irq_controller_lock, flags); } #endif +#ifdef CONFIG_BL_SWITCHER +/* + * gic_send_sgi - send a SGI directly to given CPU interface number + * + * cpu_id: the ID for the destination CPU interface + * irq: the IPI number to send a SGI for + */ +void gic_send_sgi(unsigned int cpu_id, unsigned int irq) +{ + BUG_ON(cpu_id >= NR_GIC_CPU_IF); + cpu_id = 1 << cpu_id; + /* this always happens on GIC0 */ + writel_relaxed((cpu_id << 16) | irq, gic_data_dist_base(&gic_data[0]) + GIC_DIST_SOFTINT); +} + +/* + * gic_get_cpu_id - get the CPU interface ID for the specified CPU + * + * @cpu: the logical CPU number to get the GIC ID for. + * + * Return the CPU interface ID for the given logical CPU number, + * or -1 if the CPU number is too large or the interface ID is + * unknown (more than one bit set). + */ +int gic_get_cpu_id(unsigned int cpu) +{ + unsigned int cpu_bit; + + if (cpu >= NR_GIC_CPU_IF) + return -1; + cpu_bit = gic_cpu_map[cpu]; + if (cpu_bit & (cpu_bit - 1)) + return -1; + return __ffs(cpu_bit); +} + +/* + * gic_migrate_target - migrate IRQs to another CPU interface + * + * @new_cpu_id: the CPU target ID to migrate IRQs to + * + * Migrate all peripheral interrupts with a target matching the current CPU + * to the interface corresponding to @new_cpu_id. The CPU interface mapping + * is also updated. Targets to other CPU interfaces are unchanged. + * This must be called with IRQs locally disabled. + */ +void gic_migrate_target(unsigned int new_cpu_id) +{ + unsigned int cur_cpu_id, gic_irqs, gic_nr = 0; + void __iomem *dist_base; + int i, ror_val, cpu = smp_processor_id(); + u32 val, cur_target_mask, active_mask; + + if (gic_nr >= MAX_GIC_NR) + BUG(); + + dist_base = gic_data_dist_base(&gic_data[gic_nr]); + if (!dist_base) + return; + gic_irqs = gic_data[gic_nr].gic_irqs; + + cur_cpu_id = __ffs(gic_cpu_map[cpu]); + cur_target_mask = 0x01010101 << cur_cpu_id; + ror_val = (cur_cpu_id - new_cpu_id) & 31; + + raw_spin_lock(&irq_controller_lock); + + /* Update the target interface for this logical CPU */ + gic_cpu_map[cpu] = 1 << new_cpu_id; + + /* + * Find all the peripheral interrupts targetting the current + * CPU interface and migrate them to the new CPU interface. + * We skip DIST_TARGET 0 to 7 as they are read-only. + */ + for (i = 8; i < DIV_ROUND_UP(gic_irqs, 4); i++) { + val = readl_relaxed(dist_base + GIC_DIST_TARGET + i * 4); + active_mask = val & cur_target_mask; + if (active_mask) { + val &= ~active_mask; + val |= ror32(active_mask, ror_val); + writel_relaxed(val, dist_base + GIC_DIST_TARGET + i*4); + } + } + + raw_spin_unlock(&irq_controller_lock); + + /* + * Now let's migrate and clear any potential SGIs that might be + * pending for us (cur_cpu_id). Since GIC_DIST_SGI_PENDING_SET + * is a banked register, we can only forward the SGI using + * GIC_DIST_SOFTINT. The original SGI source is lost but Linux + * doesn't use that information anyway. + * + * For the same reason we do not adjust SGI source information + * for previously sent SGIs by us to other CPUs either. + */ + for (i = 0; i < 16; i += 4) { + int j; + val = readl_relaxed(dist_base + GIC_DIST_SGI_PENDING_SET + i); + if (!val) + continue; + writel_relaxed(val, dist_base + GIC_DIST_SGI_PENDING_CLEAR + i); + for (j = i; j < i + 4; j++) { + if (val & 0xff) + writel_relaxed((1 << (new_cpu_id + 16)) | j, + dist_base + GIC_DIST_SOFTINT); + val >>= 8; + } + } +} + +/* + * gic_get_sgir_physaddr - get the physical address for the SGI register + * + * REturn the physical address of the SGI register to be used + * by some early assembly code when the kernel is not yet available. + */ +static unsigned long gic_dist_physaddr; + +unsigned long gic_get_sgir_physaddr(void) +{ + if (!gic_dist_physaddr) + return 0; + return gic_dist_physaddr + GIC_DIST_SOFTINT; +} + +void __init gic_init_physaddr(struct device_node *node) +{ + struct resource res; + if (of_address_to_resource(node, 0, &res) == 0) { + gic_dist_physaddr = res.start; + pr_info("GIC physical location is %#lx\n", gic_dist_physaddr); + } +} + +#else +#define gic_init_physaddr(node) do { } while (0) +#endif + static int gic_irq_domain_map(struct irq_domain *d, unsigned int irq, irq_hw_number_t hw) { @@ -680,16 +775,25 @@ static int gic_irq_domain_map(struct irq_domain *d, unsigned int irq, irq_set_chip_and_handler(irq, &gic_chip, handle_fasteoi_irq); set_irq_flags(irq, IRQF_VALID | IRQF_PROBE); + + gic_routable_irq_domain_ops->map(d, irq, hw); } irq_set_chip_data(irq, d->host_data); return 0; } +static void gic_irq_domain_unmap(struct irq_domain *d, unsigned int irq) +{ + gic_routable_irq_domain_ops->unmap(d, irq); +} + static int gic_irq_domain_xlate(struct irq_domain *d, struct device_node *controller, const u32 *intspec, unsigned int intsize, unsigned long *out_hwirq, unsigned int *out_type) { + unsigned long ret = 0; + if (d->of_node != controller) return -EINVAL; if (intsize < 3) @@ -699,11 +803,20 @@ static int gic_irq_domain_xlate(struct irq_domain *d, *out_hwirq = intspec[1] + 16; /* For SPIs, we need to add 16 more to get the GIC irq ID number */ - if (!intspec[0]) - *out_hwirq += 16; + if (!intspec[0]) { + ret = gic_routable_irq_domain_ops->xlate(d, controller, + intspec, + intsize, + out_hwirq, + out_type); + + if (IS_ERR_VALUE(ret)) + return ret; + } *out_type = intspec[2] & IRQ_TYPE_SENSE_MASK; - return 0; + + return ret; } #ifdef CONFIG_SMP @@ -725,11 +838,43 @@ static struct notifier_block __cpuinitdata gic_cpu_notifier = { }; #endif -const struct irq_domain_ops gic_irq_domain_ops = { +static const struct irq_domain_ops gic_irq_domain_ops = { .map = gic_irq_domain_map, + .unmap = gic_irq_domain_unmap, .xlate = gic_irq_domain_xlate, }; +/* Default functions for routable irq domain */ +static int gic_routable_irq_domain_map(struct irq_domain *d, unsigned int irq, + irq_hw_number_t hw) +{ + return 0; +} + +static void gic_routable_irq_domain_unmap(struct irq_domain *d, + unsigned int irq) +{ +} + +static int gic_routable_irq_domain_xlate(struct irq_domain *d, + struct device_node *controller, + const u32 *intspec, unsigned int intsize, + unsigned long *out_hwirq, + unsigned int *out_type) +{ + *out_hwirq += 16; + return 0; +} + +const struct irq_domain_ops gic_default_routable_irq_domain_ops = { + .map = gic_routable_irq_domain_map, + .unmap = gic_routable_irq_domain_unmap, + .xlate = gic_routable_irq_domain_xlate, +}; + +const struct irq_domain_ops *gic_routable_irq_domain_ops = + &gic_default_routable_irq_domain_ops; + void __init gic_init_bases(unsigned int gic_nr, int irq_start, void __iomem *dist_base, void __iomem *cpu_base, u32 percpu_offset, struct device_node *node) @@ -737,6 +882,7 @@ void __init gic_init_bases(unsigned int gic_nr, int irq_start, irq_hw_number_t hwirq_base; struct gic_chip_data *gic; int gic_irqs, irq_base, i; + int nr_routable_irqs; BUG_ON(gic_nr >= MAX_GIC_NR); @@ -804,23 +950,35 @@ void __init gic_init_bases(unsigned int gic_nr, int irq_start, gic->gic_irqs = gic_irqs; gic_irqs -= hwirq_base; /* calculate # of irqs to allocate */ - irq_base = irq_alloc_descs(irq_start, 16, gic_irqs, numa_node_id()); - if (IS_ERR_VALUE(irq_base)) { - WARN(1, "Cannot allocate irq_descs @ IRQ%d, assuming pre-allocated\n", - irq_start); - irq_base = irq_start; + + if (of_property_read_u32(node, "arm,routable-irqs", + &nr_routable_irqs)) { + irq_base = irq_alloc_descs(irq_start, 16, gic_irqs, + numa_node_id()); + if (IS_ERR_VALUE(irq_base)) { + WARN(1, "Cannot allocate irq_descs @ IRQ%d, assuming pre-allocated\n", + irq_start); + irq_base = irq_start; + } + + gic->domain = irq_domain_add_legacy(node, gic_irqs, irq_base, + hwirq_base, &gic_irq_domain_ops, gic); + } else { + gic->domain = irq_domain_add_linear(node, nr_routable_irqs, + &gic_irq_domain_ops, + gic); } - gic->domain = irq_domain_add_legacy(node, gic_irqs, irq_base, - hwirq_base, &gic_irq_domain_ops, gic); + if (WARN_ON(!gic->domain)) return; + if (gic_nr == 0) { #ifdef CONFIG_SMP - set_smp_cross_call(gic_raise_softirq); - register_cpu_notifier(&gic_cpu_notifier); + set_smp_cross_call(gic_raise_softirq); + register_cpu_notifier(&gic_cpu_notifier); #endif - - set_handle_irq(gic_handle_irq); + set_handle_irq(gic_handle_irq); + } gic_chip.flags |= gic_arch_extn.flags; gic_dist_init(gic); @@ -831,7 +989,8 @@ void __init gic_init_bases(unsigned int gic_nr, int irq_start, #ifdef CONFIG_OF static int gic_cnt __initdata; -int __init gic_of_init(struct device_node *node, struct device_node *parent) +static int __init +gic_of_init(struct device_node *node, struct device_node *parent) { void __iomem *cpu_base; void __iomem *dist_base; @@ -851,6 +1010,8 @@ int __init gic_of_init(struct device_node *node, struct device_node *parent) percpu_offset = 0; gic_init_bases(gic_cnt, -1, dist_base, cpu_base, percpu_offset, node); + if (!gic_cnt) + gic_init_physaddr(node); if (parent) { irq = irq_of_parse_and_map(node, 0); @@ -859,6 +1020,7 @@ int __init gic_of_init(struct device_node *node, struct device_node *parent) gic_cnt++; return 0; } +IRQCHIP_DECLARE(gic_400, "arm,gic-400", gic_of_init); IRQCHIP_DECLARE(cortex_a15_gic, "arm,cortex-a15-gic", gic_of_init); IRQCHIP_DECLARE(cortex_a9_gic, "arm,cortex-a9-gic", gic_of_init); IRQCHIP_DECLARE(cortex_a7_gic, "arm,cortex-a7-gic", gic_of_init); diff --git a/drivers/mailbox/Makefile b/drivers/mailbox/Makefile index 543ad6a79505..fefef7ebcbec 100644 --- a/drivers/mailbox/Makefile +++ b/drivers/mailbox/Makefile @@ -1 +1,5 @@ +# Generic MAILBOX API + +obj-$(CONFIG_MAILBOX) += mailbox.o + obj-$(CONFIG_PL320_MBOX) += pl320-ipc.o diff --git a/drivers/mailbox/mailbox.c b/drivers/mailbox/mailbox.c new file mode 100644 index 000000000000..9a937ef35068 --- /dev/null +++ b/drivers/mailbox/mailbox.c @@ -0,0 +1,467 @@ +/* + * Mailbox: Common code for Mailbox controllers and users + * + * Copyright (C) 2013-2014 Linaro Ltd. + * Author: Jassi Brar <jassisinghbrar@gmail.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include <linux/interrupt.h> +#include <linux/spinlock.h> +#include <linux/mutex.h> +#include <linux/delay.h> +#include <linux/slab.h> +#include <linux/err.h> +#include <linux/module.h> +#include <linux/device.h> +#include <linux/bitops.h> +#include <linux/mailbox_client.h> +#include <linux/mailbox_controller.h> + +#define TXDONE_BY_IRQ BIT(0) /* controller has remote RTR irq */ +#define TXDONE_BY_POLL BIT(1) /* controller can read status of last TX */ +#define TXDONE_BY_ACK BIT(2) /* S/W ACK recevied by Client ticks the TX */ + +static LIST_HEAD(mbox_cons); +static DEFINE_MUTEX(con_mutex); + +static int add_to_rbuf(struct mbox_chan *chan, void *mssg) +{ + int idx; + unsigned long flags; + + spin_lock_irqsave(&chan->lock, flags); + + /* See if there is any space left */ + if (chan->msg_count == MBOX_TX_QUEUE_LEN) { + spin_unlock_irqrestore(&chan->lock, flags); + return -ENOBUFS; + } + + idx = chan->msg_free; + chan->msg_data[idx] = mssg; + chan->msg_count++; + + if (idx == MBOX_TX_QUEUE_LEN - 1) + chan->msg_free = 0; + else + chan->msg_free++; + + spin_unlock_irqrestore(&chan->lock, flags); + + return idx; +} + +static void msg_submit(struct mbox_chan *chan) +{ + unsigned count, idx; + unsigned long flags; + void *data; + int err; + + spin_lock_irqsave(&chan->lock, flags); + + if (!chan->msg_count || chan->active_req) + goto exit; + + count = chan->msg_count; + idx = chan->msg_free; + if (idx >= count) + idx -= count; + else + idx += MBOX_TX_QUEUE_LEN - count; + + data = chan->msg_data[idx]; + + /* Try to submit a message to the MBOX controller */ + err = chan->mbox->ops->send_data(chan, data); + if (!err) { + chan->active_req = data; + chan->msg_count--; + } +exit: + spin_unlock_irqrestore(&chan->lock, flags); +} + +static void tx_tick(struct mbox_chan *chan, int r) +{ + unsigned long flags; + void *mssg; + + spin_lock_irqsave(&chan->lock, flags); + mssg = chan->active_req; + chan->active_req = NULL; + spin_unlock_irqrestore(&chan->lock, flags); + + /* Submit next message */ + msg_submit(chan); + + /* Notify the client */ + if (mssg && chan->cl->tx_done) + chan->cl->tx_done(chan->cl, mssg, r); + + if (chan->cl->tx_block) + complete(&chan->tx_complete); +} + +static void poll_txdone(unsigned long data) +{ + struct mbox_controller *mbox = (struct mbox_controller *)data; + bool txdone, resched = false; + int i; + + for (i = 0; i < mbox->num_chans; i++) { + struct mbox_chan *chan = &mbox->chans[i]; + + if (chan->active_req && chan->cl) { + resched = true; + txdone = chan->mbox->ops->last_tx_done(chan); + if (txdone) + tx_tick(chan, 0); + } + } + + if (resched) + mod_timer(&mbox->poll, jiffies + + msecs_to_jiffies(mbox->period)); +} + +/** + * mbox_chan_received_data - A way for controller driver to push data + * received from remote to the upper layer. + * @chan: Pointer to the mailbox channel on which RX happened. + * @mssg: Client specific message typecasted as void * + * + * After startup and before shutdown any data received on the chan + * is passed on to the API via atomic mbox_chan_received_data(). + * The controller should ACK the RX only after this call returns. + */ +void mbox_chan_received_data(struct mbox_chan *chan, void *mssg) +{ + /* No buffering the received data */ + if (chan->cl->rx_callback) + chan->cl->rx_callback(chan->cl, mssg); +} +EXPORT_SYMBOL_GPL(mbox_chan_received_data); + +/** + * mbox_chan_txdone - A way for controller driver to notify the + * framework that the last TX has completed. + * @chan: Pointer to the mailbox chan on which TX happened. + * @r: Status of last TX - OK or ERROR + * + * The controller that has IRQ for TX ACK calls this atomic API + * to tick the TX state machine. It works only if txdone_irq + * is set by the controller. + */ +void mbox_chan_txdone(struct mbox_chan *chan, int r) +{ + if (unlikely(!(chan->txdone_method & TXDONE_BY_IRQ))) { + dev_err(chan->mbox->dev, + "Controller can't run the TX ticker\n"); + return; + } + + tx_tick(chan, r); +} +EXPORT_SYMBOL_GPL(mbox_chan_txdone); + +/** + * mbox_client_txdone - The way for a client to run the TX state machine. + * @chan: Mailbox channel assigned to this client. + * @r: Success status of last transmission. + * + * The client/protocol had received some 'ACK' packet and it notifies + * the API that the last packet was sent successfully. This only works + * if the controller can't sense TX-Done. + */ +void mbox_client_txdone(struct mbox_chan *chan, int r) +{ + if (unlikely(!(chan->txdone_method & TXDONE_BY_ACK))) { + dev_err(chan->mbox->dev, "Client can't run the TX ticker\n"); + return; + } + + tx_tick(chan, r); +} +EXPORT_SYMBOL_GPL(mbox_client_txdone); + +/** + * mbox_client_peek_data - A way for client driver to pull data + * received from remote by the controller. + * @chan: Mailbox channel assigned to this client. + * + * A poke to controller driver for any received data. + * The data is actually passed onto client via the + * mbox_chan_received_data() + * The call can be made from atomic context, so the controller's + * implementation of peek_data() must not sleep. + * + * Return: True, if controller has, and is going to push after this, + * some data. + * False, if controller doesn't have any data to be read. + */ +bool mbox_client_peek_data(struct mbox_chan *chan) +{ + if (chan->mbox->ops->peek_data) + return chan->mbox->ops->peek_data(chan); + + return false; +} +EXPORT_SYMBOL_GPL(mbox_client_peek_data); + +/** + * mbox_send_message - For client to submit a message to be + * sent to the remote. + * @chan: Mailbox channel assigned to this client. + * @mssg: Client specific message typecasted. + * + * For client to submit data to the controller destined for a remote + * processor. If the client had set 'tx_block', the call will return + * either when the remote receives the data or when 'tx_tout' millisecs + * run out. + * In non-blocking mode, the requests are buffered by the API and a + * non-negative token is returned for each queued request. If the request + * is not queued, a negative token is returned. Upon failure or successful + * TX, the API calls 'tx_done' from atomic context, from which the client + * could submit yet another request. + * The pointer to message should be preserved until it is sent + * over the chan, i.e, tx_done() is made. + * This function could be called from atomic context as it simply + * queues the data and returns a token against the request. + * + * Return: Non-negative integer for successful submission (non-blocking mode) + * or transmission over chan (blocking mode). + * Negative value denotes failure. + */ +int mbox_send_message(struct mbox_chan *chan, void *mssg) +{ + int t; + + if (!chan || !chan->cl) + return -EINVAL; + + t = add_to_rbuf(chan, mssg); + if (t < 0) { + dev_err(chan->mbox->dev, "Try increasing MBOX_TX_QUEUE_LEN\n"); + return t; + } + + msg_submit(chan); + + INIT_COMPLETION(chan->tx_complete); + + if (chan->txdone_method == TXDONE_BY_POLL) + poll_txdone((unsigned long)chan->mbox); + + if (chan->cl->tx_block && chan->active_req) { + unsigned long wait; + int ret; + + if (!chan->cl->tx_tout) /* wait forever */ + wait = msecs_to_jiffies(3600000); + else + wait = msecs_to_jiffies(chan->cl->tx_tout); + + ret = wait_for_completion_timeout(&chan->tx_complete, wait); + if (ret == 0) { + t = -EIO; + tx_tick(chan, -EIO); + } + } + + return t; +} +EXPORT_SYMBOL_GPL(mbox_send_message); + +/** + * mbox_request_channel - Request a mailbox channel. + * @cl: Identity of the client requesting the channel. + * @index: Index of mailbox specifier in 'mboxes' property. + * + * The Client specifies its requirements and capabilities while asking for + * a mailbox channel. It can't be called from atomic context. + * The channel is exclusively allocated and can't be used by another + * client before the owner calls mbox_free_channel. + * After assignment, any packet received on this channel will be + * handed over to the client via the 'rx_callback'. + * The framework holds reference to the client, so the mbox_client + * structure shouldn't be modified until the mbox_free_channel returns. + * + * Return: Pointer to the channel assigned to the client if successful. + * ERR_PTR for request failure. + */ +struct mbox_chan *mbox_request_channel(struct mbox_client *cl, int index) +{ + struct device *dev = cl->dev; + struct mbox_controller *mbox; + struct of_phandle_args spec; + struct mbox_chan *chan; + unsigned long flags; + int ret; + + if (!dev || !dev->of_node) { + pr_debug("%s: No owner device node\n", __func__); + return ERR_PTR(-ENODEV); + } + + mutex_lock(&con_mutex); + + if (of_parse_phandle_with_args(dev->of_node, "mboxes", + "#mbox-cells", index, &spec)) { + dev_dbg(dev, "%s: can't parse \"mboxes\" property\n", __func__); + mutex_unlock(&con_mutex); + return ERR_PTR(-ENODEV); + } + + chan = NULL; + list_for_each_entry(mbox, &mbox_cons, node) + if (mbox->dev->of_node == spec.np) { + chan = mbox->of_xlate(mbox, &spec); + break; + } + + of_node_put(spec.np); + + if (!chan || chan->cl || !try_module_get(mbox->dev->driver->owner)) { + dev_dbg(dev, "%s: mailbox not free\n", __func__); + mutex_unlock(&con_mutex); + return ERR_PTR(-EBUSY); + } + + spin_lock_irqsave(&chan->lock, flags); + chan->msg_free = 0; + chan->msg_count = 0; + chan->active_req = NULL; + chan->cl = cl; + init_completion(&chan->tx_complete); + + if (chan->txdone_method == TXDONE_BY_POLL && cl->knows_txdone) + chan->txdone_method |= TXDONE_BY_ACK; + + spin_unlock_irqrestore(&chan->lock, flags); + + ret = chan->mbox->ops->startup(chan); + if (ret) { + dev_err(dev, "Unable to startup the chan (%d)\n", ret); + mbox_free_channel(chan); + chan = ERR_PTR(ret); + } + + mutex_unlock(&con_mutex); + return chan; +} +EXPORT_SYMBOL_GPL(mbox_request_channel); + +/** + * mbox_free_channel - The client relinquishes control of a mailbox + * channel by this call. + * @chan: The mailbox channel to be freed. + */ +void mbox_free_channel(struct mbox_chan *chan) +{ + unsigned long flags; + + if (!chan || !chan->cl) + return; + + chan->mbox->ops->shutdown(chan); + + /* The queued TX requests are simply aborted, no callbacks are made */ + spin_lock_irqsave(&chan->lock, flags); + chan->cl = NULL; + chan->active_req = NULL; + if (chan->txdone_method == (TXDONE_BY_POLL | TXDONE_BY_ACK)) + chan->txdone_method = TXDONE_BY_POLL; + + module_put(chan->mbox->dev->driver->owner); + spin_unlock_irqrestore(&chan->lock, flags); +} +EXPORT_SYMBOL_GPL(mbox_free_channel); + +static struct mbox_chan * +of_mbox_index_xlate(struct mbox_controller *mbox, + const struct of_phandle_args *sp) +{ + int ind = sp->args[0]; + + if (ind >= mbox->num_chans) + return NULL; + + return &mbox->chans[ind]; +} + +/** + * mbox_controller_register - Register the mailbox controller + * @mbox: Pointer to the mailbox controller. + * + * The controller driver registers its communication channels + */ +int mbox_controller_register(struct mbox_controller *mbox) +{ + int i, txdone; + + /* Sanity check */ + if (!mbox || !mbox->dev || !mbox->ops || !mbox->num_chans) + return -EINVAL; + + if (mbox->txdone_irq) + txdone = TXDONE_BY_IRQ; + else if (mbox->txdone_poll) + txdone = TXDONE_BY_POLL; + else /* It has to be ACK then */ + txdone = TXDONE_BY_ACK; + + if (txdone == TXDONE_BY_POLL) { + mbox->poll.function = &poll_txdone; + mbox->poll.data = (unsigned long)mbox; + init_timer(&mbox->poll); + } + + for (i = 0; i < mbox->num_chans; i++) { + struct mbox_chan *chan = &mbox->chans[i]; + + chan->cl = NULL; + chan->mbox = mbox; + chan->txdone_method = txdone; + spin_lock_init(&chan->lock); + } + + if (!mbox->of_xlate) + mbox->of_xlate = of_mbox_index_xlate; + + mutex_lock(&con_mutex); + list_add_tail(&mbox->node, &mbox_cons); + mutex_unlock(&con_mutex); + + return 0; +} +EXPORT_SYMBOL_GPL(mbox_controller_register); + +/** + * mbox_controller_unregister - Unregister the mailbox controller + * @mbox: Pointer to the mailbox controller. + */ +void mbox_controller_unregister(struct mbox_controller *mbox) +{ + int i; + + if (!mbox) + return; + + mutex_lock(&con_mutex); + + list_del(&mbox->node); + + for (i = 0; i < mbox->num_chans; i++) + mbox_free_channel(&mbox->chans[i]); + + if (mbox->txdone_poll) + del_timer_sync(&mbox->poll); + + mutex_unlock(&con_mutex); +} +EXPORT_SYMBOL_GPL(mbox_controller_unregister); diff --git a/drivers/mailbox/pl320-ipc.c b/drivers/mailbox/pl320-ipc.c index d873cbae2fbb..f3755e0aa935 100644 --- a/drivers/mailbox/pl320-ipc.c +++ b/drivers/mailbox/pl320-ipc.c @@ -26,7 +26,7 @@ #include <linux/device.h> #include <linux/amba/bus.h> -#include <linux/mailbox.h> +#include <linux/pl320-ipc.h> #define IPCMxSOURCE(m) ((m) * 0x40) #define IPCMxDSET(m) (((m) * 0x40) + 0x004) diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig index d54e985748b7..a5e54f0d6a73 100644 --- a/drivers/mfd/Kconfig +++ b/drivers/mfd/Kconfig @@ -1144,7 +1144,15 @@ config MCP_UCB1200_TS endmenu config VEXPRESS_CONFIG - bool + bool "ARM Versatile Express platform infrastructure" + depends on ARM || ARM64 help Platform configuration infrastructure for the ARM Ltd. Versatile Express. + +config VEXPRESS_SPC + bool "Versatile Express SPC driver support" + depends on ARM + depends on VEXPRESS_CONFIG + help + Serial Power Controller driver for ARM Ltd. test chips. diff --git a/drivers/mfd/Makefile b/drivers/mfd/Makefile index 718e94a2a9a7..3a0120315aa3 100644 --- a/drivers/mfd/Makefile +++ b/drivers/mfd/Makefile @@ -153,5 +153,6 @@ obj-$(CONFIG_MFD_SEC_CORE) += sec-core.o sec-irq.o obj-$(CONFIG_MFD_SYSCON) += syscon.o obj-$(CONFIG_MFD_LM3533) += lm3533-core.o lm3533-ctrlbank.o obj-$(CONFIG_VEXPRESS_CONFIG) += vexpress-config.o vexpress-sysreg.o +obj-$(CONFIG_VEXPRESS_SPC) += vexpress-spc.o obj-$(CONFIG_MFD_RETU) += retu-mfd.o obj-$(CONFIG_MFD_AS3711) += as3711.o diff --git a/drivers/mfd/vexpress-config.c b/drivers/mfd/vexpress-config.c index 84ce6b9daa3d..1af2b0e0182f 100644 --- a/drivers/mfd/vexpress-config.c +++ b/drivers/mfd/vexpress-config.c @@ -86,29 +86,13 @@ void vexpress_config_bridge_unregister(struct vexpress_config_bridge *bridge) } EXPORT_SYMBOL(vexpress_config_bridge_unregister); - -struct vexpress_config_func { - struct vexpress_config_bridge *bridge; - void *func; -}; - -struct vexpress_config_func *__vexpress_config_func_get(struct device *dev, - struct device_node *node) +static struct vexpress_config_bridge * + vexpress_config_bridge_find(struct device_node *node) { - struct device_node *bridge_node; - struct vexpress_config_func *func; int i; + struct vexpress_config_bridge *res = NULL; + struct device_node *bridge_node = of_node_get(node); - if (WARN_ON(dev && node && dev->of_node != node)) - return NULL; - if (dev && !node) - node = dev->of_node; - - func = kzalloc(sizeof(*func), GFP_KERNEL); - if (!func) - return NULL; - - bridge_node = of_node_get(node); while (bridge_node) { const __be32 *prop = of_get_property(bridge_node, "arm,vexpress,config-bridge", NULL); @@ -129,13 +113,46 @@ struct vexpress_config_func *__vexpress_config_func_get(struct device *dev, if (test_bit(i, vexpress_config_bridges_map) && bridge->node == bridge_node) { - func->bridge = bridge; - func->func = bridge->info->func_get(dev, node); + res = bridge; break; } } mutex_unlock(&vexpress_config_bridges_mutex); + return res; +} + + +struct vexpress_config_func { + struct vexpress_config_bridge *bridge; + void *func; +}; + +struct vexpress_config_func *__vexpress_config_func_get( + struct vexpress_config_bridge *bridge, + struct device *dev, + struct device_node *node, + const char *id) +{ + struct vexpress_config_func *func; + + if (WARN_ON(dev && node && dev->of_node != node)) + return NULL; + if (dev && !node) + node = dev->of_node; + + if (!bridge) + bridge = vexpress_config_bridge_find(node); + if (!bridge) + return NULL; + + func = kzalloc(sizeof(*func), GFP_KERNEL); + if (!func) + return NULL; + + func->bridge = bridge; + func->func = bridge->info->func_get(dev, node, id); + if (!func->func) { of_node_put(node); kfree(func); diff --git a/drivers/mfd/vexpress-spc.c b/drivers/mfd/vexpress-spc.c new file mode 100644 index 000000000000..0c6718abf1ba --- /dev/null +++ b/drivers/mfd/vexpress-spc.c @@ -0,0 +1,633 @@ +/* + * Versatile Express Serial Power Controller (SPC) support + * + * Copyright (C) 2013 ARM Ltd. + * + * Authors: Sudeep KarkadaNagesha <sudeep.karkadanagesha@arm.com> + * Achin Gupta <achin.gupta@arm.com> + * Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * This program is distributed "as is" WITHOUT ANY WARRANTY of any + * kind, whether express or implied; without even the implied warranty + * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#include <linux/device.h> +#include <linux/err.h> +#include <linux/io.h> +#include <linux/interrupt.h> +#include <linux/module.h> +#include <linux/of_address.h> +#include <linux/of_irq.h> +#include <linux/slab.h> +#include <linux/vexpress.h> + +#include <asm/cacheflush.h> + +#define SCC_CFGREG19 0x120 +#define SCC_CFGREG20 0x124 +#define A15_CONF 0x400 +#define A7_CONF 0x500 +#define SYS_INFO 0x700 +#define PERF_LVL_A15 0xB00 +#define PERF_REQ_A15 0xB04 +#define PERF_LVL_A7 0xB08 +#define PERF_REQ_A7 0xB0c +#define SYS_CFGCTRL 0xB10 +#define SYS_CFGCTRL_REQ 0xB14 +#define PWC_STATUS 0xB18 +#define PWC_FLAG 0xB1c +#define WAKE_INT_MASK 0xB24 +#define WAKE_INT_RAW 0xB28 +#define WAKE_INT_STAT 0xB2c +#define A15_PWRDN_EN 0xB30 +#define A7_PWRDN_EN 0xB34 +#define A7_PWRDNACK 0xB54 +#define A15_BX_ADDR0 0xB68 +#define SYS_CFG_WDATA 0xB70 +#define SYS_CFG_RDATA 0xB74 +#define A7_BX_ADDR0 0xB78 + +#define GBL_WAKEUP_INT_MSK (0x3 << 10) + +#define CLKF_SHIFT 16 +#define CLKF_MASK 0x1FFF +#define CLKR_SHIFT 0 +#define CLKR_MASK 0x3F +#define CLKOD_SHIFT 8 +#define CLKOD_MASK 0xF + +#define OPP_FUNCTION 6 +#define OPP_BASE_DEVICE 0x300 +#define OPP_A15_OFFSET 0x4 +#define OPP_A7_OFFSET 0xc + +#define SYS_CFGCTRL_START (1 << 31) +#define SYS_CFGCTRL_WRITE (1 << 30) +#define SYS_CFGCTRL_FUNC(n) (((n) & 0x3f) << 20) +#define SYS_CFGCTRL_DEVICE(n) (((n) & 0xfff) << 0) + +#define MAX_OPPS 8 +#define MAX_CLUSTERS 2 + +enum { + A15_OPP_TYPE = 0, + A7_OPP_TYPE = 1, + SYS_CFGCTRL_TYPE = 2, + INVALID_TYPE +}; + +#define STAT_COMPLETE(type) ((1 << 0) << (type << 2)) +#define STAT_ERR(type) ((1 << 1) << (type << 2)) +#define RESPONSE_MASK(type) (STAT_COMPLETE(type) | STAT_ERR(type)) + +struct vexpress_spc_drvdata { + void __iomem *baseaddr; + u32 a15_clusid; + int irq; + u32 cur_req_type; + u32 freqs[MAX_CLUSTERS][MAX_OPPS]; + int freqs_cnt[MAX_CLUSTERS]; +}; + +enum spc_func_type { + CONFIG_FUNC = 0, + PERF_FUNC = 1, +}; + +struct vexpress_spc_func { + enum spc_func_type type; + u32 function; + u32 device; +}; + +static struct vexpress_spc_drvdata *info; +static u32 *vexpress_spc_config_data; +static struct vexpress_config_bridge *vexpress_spc_config_bridge; +static struct vexpress_config_func *opp_func, *perf_func; + +static int vexpress_spc_load_result = -EAGAIN; + +static bool vexpress_spc_initialized(void) +{ + return vexpress_spc_load_result == 0; +} + +/** + * vexpress_spc_write_resume_reg() - set the jump address used for warm boot + * + * @cluster: mpidr[15:8] bitfield describing cluster affinity level + * @cpu: mpidr[7:0] bitfield describing cpu affinity level + * @addr: physical resume address + */ +void vexpress_spc_write_resume_reg(u32 cluster, u32 cpu, u32 addr) +{ + void __iomem *baseaddr; + + if (WARN_ON_ONCE(cluster >= MAX_CLUSTERS)) + return; + + if (cluster != info->a15_clusid) + baseaddr = info->baseaddr + A7_BX_ADDR0 + (cpu << 2); + else + baseaddr = info->baseaddr + A15_BX_ADDR0 + (cpu << 2); + + writel_relaxed(addr, baseaddr); +} + +/** + * vexpress_spc_get_nb_cpus() - get number of cpus in a cluster + * + * @cluster: mpidr[15:8] bitfield describing cluster affinity level + * + * Return: number of cpus in the cluster + * -EINVAL if cluster number invalid + */ +int vexpress_spc_get_nb_cpus(u32 cluster) +{ + u32 val; + + if (WARN_ON_ONCE(cluster >= MAX_CLUSTERS)) + return -EINVAL; + + val = readl_relaxed(info->baseaddr + SYS_INFO); + val = (cluster != info->a15_clusid) ? (val >> 20) : (val >> 16); + return val & 0xf; +} +EXPORT_SYMBOL_GPL(vexpress_spc_get_nb_cpus); + +/** + * vexpress_spc_get_performance - get current performance level of cluster + * @cluster: mpidr[15:8] bitfield describing cluster affinity level + * @freq: pointer to the performance level to be assigned + * + * Return: 0 on success + * < 0 on read error + */ +int vexpress_spc_get_performance(u32 cluster, u32 *freq) +{ + u32 perf_cfg_reg; + int perf, ret; + + if (!vexpress_spc_initialized() || (cluster >= MAX_CLUSTERS)) + return -EINVAL; + + perf_cfg_reg = cluster != info->a15_clusid ? PERF_LVL_A7 : PERF_LVL_A15; + ret = vexpress_config_read(perf_func, perf_cfg_reg, &perf); + + if (!ret) + *freq = info->freqs[cluster][perf]; + + return ret; +} +EXPORT_SYMBOL_GPL(vexpress_spc_get_performance); + +/** + * vexpress_spc_get_perf_index - get performance level corresponding to + * a frequency + * @cluster: mpidr[15:8] bitfield describing cluster affinity level + * @freq: frequency to be looked-up + * + * Return: perf level index on success + * -EINVAL on error + */ +static int vexpress_spc_find_perf_index(u32 cluster, u32 freq) +{ + int idx; + + for (idx = 0; idx < info->freqs_cnt[cluster]; idx++) + if (info->freqs[cluster][idx] == freq) + break; + return (idx == info->freqs_cnt[cluster]) ? -EINVAL : idx; +} + +/** + * vexpress_spc_set_performance - set current performance level of cluster + * + * @cluster: mpidr[15:8] bitfield describing cluster affinity level + * @freq: performance level to be programmed + * + * Returns: 0 on success + * < 0 on write error + */ +int vexpress_spc_set_performance(u32 cluster, u32 freq) +{ + int ret, perf, offset; + + if (!vexpress_spc_initialized() || (cluster >= MAX_CLUSTERS)) + return -EINVAL; + + offset = (cluster != info->a15_clusid) ? PERF_LVL_A7 : PERF_LVL_A15; + + perf = vexpress_spc_find_perf_index(cluster, freq); + + if (perf < 0 || perf >= MAX_OPPS) + return -EINVAL; + + ret = vexpress_config_write(perf_func, offset, perf); + + return ret; +} +EXPORT_SYMBOL_GPL(vexpress_spc_set_performance); + +static void vexpress_spc_set_wake_intr(u32 mask) +{ + writel_relaxed(mask & VEXPRESS_SPC_WAKE_INTR_MASK, + info->baseaddr + WAKE_INT_MASK); +} + +static inline void reg_bitmask(u32 *reg, u32 mask, bool set) +{ + if (set) + *reg |= mask; + else + *reg &= ~mask; +} + +/** + * vexpress_spc_set_global_wakeup_intr() + * + * Function to set/clear global wakeup IRQs. Not protected by locking since + * it might be used in code paths where normal cacheable locks are not + * working. Locking must be provided by the caller to ensure atomicity. + * + * @set: if true, global wake-up IRQs are set, if false they are cleared + */ +void vexpress_spc_set_global_wakeup_intr(bool set) +{ + u32 wake_int_mask_reg = 0; + + wake_int_mask_reg = readl_relaxed(info->baseaddr + WAKE_INT_MASK); + reg_bitmask(&wake_int_mask_reg, GBL_WAKEUP_INT_MSK, set); + vexpress_spc_set_wake_intr(wake_int_mask_reg); +} + +/** + * vexpress_spc_set_cpu_wakeup_irq() + * + * Function to set/clear per-CPU wake-up IRQs. Not protected by locking since + * it might be used in code paths where normal cacheable locks are not + * working. Locking must be provided by the caller to ensure atomicity. + * + * @cpu: mpidr[7:0] bitfield describing cpu affinity level + * @cluster: mpidr[15:8] bitfield describing cluster affinity level + * @set: if true, wake-up IRQs are set, if false they are cleared + */ +void vexpress_spc_set_cpu_wakeup_irq(u32 cpu, u32 cluster, bool set) +{ + u32 mask = 0; + u32 wake_int_mask_reg = 0; + + mask = 1 << cpu; + if (info->a15_clusid != cluster) + mask <<= 4; + + wake_int_mask_reg = readl_relaxed(info->baseaddr + WAKE_INT_MASK); + reg_bitmask(&wake_int_mask_reg, mask, set); + vexpress_spc_set_wake_intr(wake_int_mask_reg); +} + +/** + * vexpress_spc_powerdown_enable() + * + * Function to enable/disable cluster powerdown. Not protected by locking + * since it might be used in code paths where normal cacheable locks are not + * working. Locking must be provided by the caller to ensure atomicity. + * + * @cluster: mpidr[15:8] bitfield describing cluster affinity level + * @enable: if true enables powerdown, if false disables it + */ +void vexpress_spc_powerdown_enable(u32 cluster, bool enable) +{ + u32 pwdrn_reg = 0; + + if (cluster >= MAX_CLUSTERS) + return; + pwdrn_reg = cluster != info->a15_clusid ? A7_PWRDN_EN : A15_PWRDN_EN; + writel_relaxed(enable, info->baseaddr + pwdrn_reg); +} + +irqreturn_t vexpress_spc_irq_handler(int irq, void *data) +{ + int ret; + u32 status = readl_relaxed(info->baseaddr + PWC_STATUS); + + if (!(status & RESPONSE_MASK(info->cur_req_type))) + return IRQ_NONE; + + if ((status == STAT_COMPLETE(SYS_CFGCTRL_TYPE)) + && vexpress_spc_config_data) { + *vexpress_spc_config_data = + readl_relaxed(info->baseaddr + SYS_CFG_RDATA); + vexpress_spc_config_data = NULL; + } + + ret = STAT_COMPLETE(info->cur_req_type) ? 0 : -EIO; + info->cur_req_type = INVALID_TYPE; + vexpress_config_complete(vexpress_spc_config_bridge, ret); + return IRQ_HANDLED; +} + +/** + * Based on the firmware documentation, this is always fixed to 20 + * All the 4 OSC: A15 PLL0/1, A7 PLL0/1 must be programmed same + * values for both control and value registers. + * This function uses A15 PLL 0 registers to compute multiple factor + * F out = F in * (CLKF + 1) / ((CLKOD + 1) * (CLKR + 1)) + */ +static inline int __get_mult_factor(void) +{ + int i_div, o_div, f_div; + u32 tmp; + + tmp = readl(info->baseaddr + SCC_CFGREG19); + f_div = (tmp >> CLKF_SHIFT) & CLKF_MASK; + + tmp = readl(info->baseaddr + SCC_CFGREG20); + o_div = (tmp >> CLKOD_SHIFT) & CLKOD_MASK; + i_div = (tmp >> CLKR_SHIFT) & CLKR_MASK; + + return (f_div + 1) / ((o_div + 1) * (i_div + 1)); +} + +/** + * vexpress_spc_populate_opps() - initialize opp tables from microcontroller + * + * @cluster: mpidr[15:8] bitfield describing cluster affinity level + * + * Return: 0 on success + * < 0 on error + */ +static int vexpress_spc_populate_opps(u32 cluster) +{ + u32 data = 0, ret, i, offset; + int mult_fact = __get_mult_factor(); + + if (WARN_ON_ONCE(cluster >= MAX_CLUSTERS)) + return -EINVAL; + + offset = cluster != info->a15_clusid ? OPP_A7_OFFSET : OPP_A15_OFFSET; + for (i = 0; i < MAX_OPPS; i++) { + ret = vexpress_config_read(opp_func, i + offset, &data); + if (!ret) + info->freqs[cluster][i] = (data & 0xFFFFF) * mult_fact; + else + break; + } + + info->freqs_cnt[cluster] = i; + return ret; +} + +/** + * vexpress_spc_get_freq_table() - Retrieve a pointer to the frequency + * table for a given cluster + * + * @cluster: mpidr[15:8] bitfield describing cluster affinity level + * @fptr: pointer to be initialized + * Return: operating points count on success + * -EINVAL on pointer error + */ +int vexpress_spc_get_freq_table(u32 cluster, u32 **fptr) +{ + if (WARN_ON_ONCE(!fptr || cluster >= MAX_CLUSTERS)) + return -EINVAL; + *fptr = info->freqs[cluster]; + return info->freqs_cnt[cluster]; +} +EXPORT_SYMBOL_GPL(vexpress_spc_get_freq_table); + +static void *vexpress_spc_func_get(struct device *dev, + struct device_node *node, const char *id) +{ + struct vexpress_spc_func *spc_func; + u32 func_device[2]; + int err = 0; + + spc_func = kzalloc(sizeof(*spc_func), GFP_KERNEL); + if (!spc_func) + return NULL; + + if (strcmp(id, "opp") == 0) { + spc_func->type = CONFIG_FUNC; + spc_func->function = OPP_FUNCTION; + spc_func->device = OPP_BASE_DEVICE; + } else if (strcmp(id, "perf") == 0) { + spc_func->type = PERF_FUNC; + } else if (node) { + of_node_get(node); + err = of_property_read_u32_array(node, + "arm,vexpress-sysreg,func", func_device, + ARRAY_SIZE(func_device)); + of_node_put(node); + spc_func->type = CONFIG_FUNC; + spc_func->function = func_device[0]; + spc_func->device = func_device[1]; + } + + if (WARN_ON(err)) { + kfree(spc_func); + return NULL; + } + + pr_debug("func 0x%p = 0x%x, %d %d\n", spc_func, + spc_func->function, + spc_func->device, + spc_func->type); + + return spc_func; +} + +static void vexpress_spc_func_put(void *func) +{ + kfree(func); +} + +static int vexpress_spc_func_exec(void *func, int offset, bool write, + u32 *data) +{ + struct vexpress_spc_func *spc_func = func; + u32 command; + + if (!data) + return -EINVAL; + /* + * Setting and retrieval of operating points is not part of + * DCC config interface. It was made to go through the same + * code path so that requests to the M3 can be serialized + * properly with config reads/writes through the common + * vexpress config interface + */ + switch (spc_func->type) { + case PERF_FUNC: + if (write) { + info->cur_req_type = (offset == PERF_LVL_A15) ? + A15_OPP_TYPE : A7_OPP_TYPE; + writel_relaxed(*data, info->baseaddr + offset); + return VEXPRESS_CONFIG_STATUS_WAIT; + } else { + *data = readl_relaxed(info->baseaddr + offset); + return VEXPRESS_CONFIG_STATUS_DONE; + } + case CONFIG_FUNC: + info->cur_req_type = SYS_CFGCTRL_TYPE; + + command = SYS_CFGCTRL_START; + command |= write ? SYS_CFGCTRL_WRITE : 0; + command |= SYS_CFGCTRL_FUNC(spc_func->function); + command |= SYS_CFGCTRL_DEVICE(spc_func->device + offset); + + pr_debug("command %x\n", command); + + if (!write) + vexpress_spc_config_data = data; + else + writel_relaxed(*data, info->baseaddr + SYS_CFG_WDATA); + writel_relaxed(command, info->baseaddr + SYS_CFGCTRL); + + return VEXPRESS_CONFIG_STATUS_WAIT; + default: + return -EINVAL; + } +} + +struct vexpress_config_bridge_info vexpress_spc_config_bridge_info = { + .name = "vexpress-spc", + .func_get = vexpress_spc_func_get, + .func_put = vexpress_spc_func_put, + .func_exec = vexpress_spc_func_exec, +}; + +static const struct of_device_id vexpress_spc_ids[] __initconst = { + { .compatible = "arm,vexpress-spc,v2p-ca15_a7" }, + { .compatible = "arm,vexpress-spc" }, + {}, +}; + +static int __init vexpress_spc_init(void) +{ + int ret; + struct device_node *node = of_find_matching_node(NULL, + vexpress_spc_ids); + + if (!node) + return -ENODEV; + + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) { + pr_err("%s: unable to allocate mem\n", __func__); + return -ENOMEM; + } + info->cur_req_type = INVALID_TYPE; + + info->baseaddr = of_iomap(node, 0); + if (WARN_ON(!info->baseaddr)) { + ret = -ENXIO; + goto mem_free; + } + + info->irq = irq_of_parse_and_map(node, 0); + + if (WARN_ON(!info->irq)) { + ret = -ENXIO; + goto unmap; + } + + readl_relaxed(info->baseaddr + PWC_STATUS); + + ret = request_irq(info->irq, vexpress_spc_irq_handler, + IRQF_DISABLED | IRQF_TRIGGER_HIGH | IRQF_ONESHOT, + "arm-spc", info); + + if (ret) { + pr_err("IRQ %d request failed\n", info->irq); + ret = -ENODEV; + goto unmap; + } + + info->a15_clusid = readl_relaxed(info->baseaddr + A15_CONF) & 0xf; + + vexpress_spc_config_bridge = vexpress_config_bridge_register( + node, &vexpress_spc_config_bridge_info); + + if (WARN_ON(!vexpress_spc_config_bridge)) { + ret = -ENODEV; + goto unmap; + } + + opp_func = vexpress_config_func_get(vexpress_spc_config_bridge, "opp"); + perf_func = + vexpress_config_func_get(vexpress_spc_config_bridge, "perf"); + + if (!opp_func || !perf_func) { + ret = -ENODEV; + goto unmap; + } + + if (vexpress_spc_populate_opps(0) || vexpress_spc_populate_opps(1)) { + if (info->irq) + free_irq(info->irq, info); + pr_err("failed to build OPP table\n"); + ret = -ENODEV; + goto unmap; + } + /* + * Multi-cluster systems may need this data when non-coherent, during + * cluster power-up/power-down. Make sure it reaches main memory: + */ + sync_cache_w(info); + sync_cache_w(&info); + pr_info("vexpress-spc loaded at %p\n", info->baseaddr); + return 0; + +unmap: + iounmap(info->baseaddr); + +mem_free: + kfree(info); + return ret; +} + +static bool __init __vexpress_spc_check_loaded(void); +/* + * Pointer spc_check_loaded is swapped after init hence it is safe + * to initialize it to a function in the __init section + */ +static bool (*spc_check_loaded)(void) __refdata = &__vexpress_spc_check_loaded; + +static bool __init __vexpress_spc_check_loaded(void) +{ + if (vexpress_spc_load_result == -EAGAIN) + vexpress_spc_load_result = vexpress_spc_init(); + spc_check_loaded = &vexpress_spc_initialized; + return vexpress_spc_initialized(); +} + +/* + * Function exported to manage early_initcall ordering. + * SPC code is needed very early in the boot process + * to bring CPUs out of reset and initialize power + * management back-end. After boot swap pointers to + * make the functionality check available to loadable + * modules, when early boot init functions have been + * already freed from kernel address space. + */ +bool vexpress_spc_check_loaded(void) +{ + return spc_check_loaded(); +} +EXPORT_SYMBOL_GPL(vexpress_spc_check_loaded); + +static int __init vexpress_spc_early_init(void) +{ + __vexpress_spc_check_loaded(); + return vexpress_spc_load_result; +} +early_initcall(vexpress_spc_early_init); +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("Serial Power Controller (SPC) support"); diff --git a/drivers/mfd/vexpress-sysreg.c b/drivers/mfd/vexpress-sysreg.c index 96a020b1dcd1..7f429afce112 100644 --- a/drivers/mfd/vexpress-sysreg.c +++ b/drivers/mfd/vexpress-sysreg.c @@ -165,7 +165,7 @@ static u32 *vexpress_sysreg_config_data; static int vexpress_sysreg_config_tries; static void *vexpress_sysreg_config_func_get(struct device *dev, - struct device_node *node) + struct device_node *node, const char *id) { struct vexpress_sysreg_config_func *config_func; u32 site; @@ -351,6 +351,8 @@ void __init vexpress_sysreg_of_early_init(void) } +#ifdef CONFIG_GPIOLIB + #define VEXPRESS_SYSREG_GPIO(_name, _reg, _value) \ [VEXPRESS_GPIO_##_name] = { \ .reg = _reg, \ @@ -445,6 +447,8 @@ struct gpio_led_platform_data vexpress_sysreg_leds_pdata = { .leds = vexpress_sysreg_leds, }; +#endif + static ssize_t vexpress_sysreg_sys_id_show(struct device *dev, struct device_attribute *attr, char *buf) @@ -480,6 +484,9 @@ static int vexpress_sysreg_probe(struct platform_device *pdev) setup_timer(&vexpress_sysreg_config_timer, vexpress_sysreg_config_complete, 0); + vexpress_sysreg_dev = &pdev->dev; + +#ifdef CONFIG_GPIOLIB vexpress_sysreg_gpio_chip.dev = &pdev->dev; err = gpiochip_add(&vexpress_sysreg_gpio_chip); if (err) { @@ -490,11 +497,10 @@ static int vexpress_sysreg_probe(struct platform_device *pdev) return err; } - vexpress_sysreg_dev = &pdev->dev; - platform_device_register_data(vexpress_sysreg_dev, "leds-gpio", PLATFORM_DEVID_AUTO, &vexpress_sysreg_leds_pdata, sizeof(vexpress_sysreg_leds_pdata)); +#endif device_create_file(vexpress_sysreg_dev, &dev_attr_sys_id); diff --git a/drivers/net/ethernet/smsc/Kconfig b/drivers/net/ethernet/smsc/Kconfig index bb4c1674ff99..ff9e99474039 100644 --- a/drivers/net/ethernet/smsc/Kconfig +++ b/drivers/net/ethernet/smsc/Kconfig @@ -97,7 +97,7 @@ config SMC911X config SMSC911X tristate "SMSC LAN911x/LAN921x families embedded ethernet support" - depends on (ARM || SUPERH || BLACKFIN || MIPS || MN10300) + depends on HAS_IOMEM select CRC32 select NET_CORE select MII diff --git a/drivers/net/ethernet/smsc/smc91x.c b/drivers/net/ethernet/smsc/smc91x.c index dfbf978315df..bdd703c6bf16 100644 --- a/drivers/net/ethernet/smsc/smc91x.c +++ b/drivers/net/ethernet/smsc/smc91x.c @@ -1896,7 +1896,7 @@ static int smc_probe(struct net_device *dev, void __iomem *ioaddr, SMC_SELECT_BANK(lp, 1); val = SMC_GET_BASE(lp); val = ((val & 0x1F00) >> 3) << SMC_IO_SHIFT; - if (((unsigned int)ioaddr & (0x3e0 << SMC_IO_SHIFT)) != val) { + if (((unsigned long)ioaddr & (0x3e0 << SMC_IO_SHIFT)) != val) { printk("%s: IOADDR %p doesn't match configuration (%x).\n", CARDNAME, ioaddr, val); } diff --git a/drivers/net/ethernet/smsc/smc91x.h b/drivers/net/ethernet/smsc/smc91x.h index fcb94806f4e9..51285872481c 100644 --- a/drivers/net/ethernet/smsc/smc91x.h +++ b/drivers/net/ethernet/smsc/smc91x.h @@ -1110,8 +1110,7 @@ static const char * chip_ids[ 16 ] = { void __iomem *__ioaddr = ioaddr; \ if (__len >= 2 && (unsigned long)__ptr & 2) { \ __len -= 2; \ - SMC_outw(*(u16 *)__ptr, ioaddr, \ - DATA_REG(lp)); \ + SMC_outsw(ioaddr, DATA_REG(lp), __ptr, 1); \ __ptr += 2; \ } \ if (SMC_CAN_USE_DATACS && lp->datacs) \ @@ -1119,8 +1118,7 @@ static const char * chip_ids[ 16 ] = { SMC_outsl(__ioaddr, DATA_REG(lp), __ptr, __len>>2); \ if (__len & 2) { \ __ptr += (__len & ~3); \ - SMC_outw(*((u16 *)__ptr), ioaddr, \ - DATA_REG(lp)); \ + SMC_outsw(ioaddr, DATA_REG(lp), __ptr, 1); \ } \ } else if (SMC_16BIT(lp)) \ SMC_outsw(ioaddr, DATA_REG(lp), p, (l) >> 1); \ diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c index 38f0b312ff85..663d2d0448b7 100644 --- a/drivers/net/phy/phy.c +++ b/drivers/net/phy/phy.c @@ -439,7 +439,7 @@ void phy_start_machine(struct phy_device *phydev, { phydev->adjust_state = handler; - schedule_delayed_work(&phydev->state_queue, HZ); + queue_delayed_work(system_power_efficient_wq, &phydev->state_queue, HZ); } /** @@ -500,7 +500,7 @@ static irqreturn_t phy_interrupt(int irq, void *phy_dat) disable_irq_nosync(irq); atomic_inc(&phydev->irq_disable); - schedule_work(&phydev->phy_queue); + queue_work(system_power_efficient_wq, &phydev->phy_queue); return IRQ_HANDLED; } @@ -655,7 +655,7 @@ static void phy_change(struct work_struct *work) /* reschedule state queue work to run as soon as possible */ cancel_delayed_work_sync(&phydev->state_queue); - schedule_delayed_work(&phydev->state_queue, 0); + queue_delayed_work(system_power_efficient_wq, &phydev->state_queue, 0); return; @@ -918,7 +918,8 @@ void phy_state_machine(struct work_struct *work) if (err < 0) phy_error(phydev); - schedule_delayed_work(&phydev->state_queue, PHY_STATE_TIME * HZ); + queue_delayed_work(system_power_efficient_wq, &phydev->state_queue, + PHY_STATE_TIME * HZ); } static inline void mmd_phy_indirect(struct mii_bus *bus, int prtad, int devad, diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig index d37bfcf5a3a2..a7bb5da6a96b 100644 --- a/drivers/of/Kconfig +++ b/drivers/of/Kconfig @@ -27,6 +27,7 @@ config OF_SELFTEST config OF_FLATTREE bool select DTC + select LIBFDT config OF_EARLY_FLATTREE bool @@ -83,4 +84,10 @@ config OF_MTD depends on MTD def_bool y +config OF_RESERVED_MEM + depends on OF_EARLY_FLATTREE + bool + help + Helpers to allow for reservation of memory regions + endmenu # OF diff --git a/drivers/of/Makefile b/drivers/of/Makefile index e027f444d10c..2aaa7b90fc14 100644 --- a/drivers/of/Makefile +++ b/drivers/of/Makefile @@ -11,3 +11,6 @@ obj-$(CONFIG_OF_MDIO) += of_mdio.o obj-$(CONFIG_OF_PCI) += of_pci.o obj-$(CONFIG_OF_PCI_IRQ) += of_pci_irq.o obj-$(CONFIG_OF_MTD) += of_mtd.o +obj-$(CONFIG_OF_RESERVED_MEM) += of_reserved_mem.o + +CFLAGS_fdt.o = -I$(src)/../../scripts/dtc/libfdt diff --git a/drivers/of/base.c b/drivers/of/base.c index b60f9a77ab03..eda24f54b863 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -18,6 +18,7 @@ * 2 of the License, or (at your option) any later version. */ #include <linux/ctype.h> +#include <linux/cpu.h> #include <linux/module.h> #include <linux/of.h> #include <linux/spinlock.h> @@ -32,6 +33,7 @@ struct device_node *of_allnodes; EXPORT_SYMBOL(of_allnodes); struct device_node *of_chosen; struct device_node *of_aliases; +static struct device_node *of_stdout; DEFINE_MUTEX(of_aliases_mutex); @@ -230,6 +232,100 @@ const void *of_get_property(const struct device_node *np, const char *name, } EXPORT_SYMBOL(of_get_property); +/* + * arch_match_cpu_phys_id - Match the given logical CPU and physical id + * + * @cpu: logical cpu index of a core/thread + * @phys_id: physical identifier of a core/thread + * + * CPU logical to physical index mapping is architecture specific. + * However this __weak function provides a default match of physical + * id to logical cpu index. phys_id provided here is usually values read + * from the device tree which must match the hardware internal registers. + * + * Returns true if the physical identifier and the logical cpu index + * correspond to the same core/thread, false otherwise. + */ +bool __weak arch_match_cpu_phys_id(int cpu, u64 phys_id) +{ + return (u32)phys_id == cpu; +} + +/** + * Checks if the given "prop_name" property holds the physical id of the + * core/thread corresponding to the logical cpu 'cpu'. If 'thread' is not + * NULL, local thread number within the core is returned in it. + */ +static bool __of_find_n_match_cpu_property(struct device_node *cpun, + const char *prop_name, int cpu, unsigned int *thread) +{ + const __be32 *cell; + int ac, prop_len, tid; + u64 hwid; + + ac = of_n_addr_cells(cpun); + cell = of_get_property(cpun, prop_name, &prop_len); + if (!cell) + return false; + prop_len /= sizeof(*cell); + for (tid = 0; tid < prop_len; tid++) { + hwid = of_read_number(cell, ac); + if (arch_match_cpu_phys_id(cpu, hwid)) { + if (thread) + *thread = tid; + return true; + } + cell += ac; + } + return false; +} + +/** + * of_get_cpu_node - Get device node associated with the given logical CPU + * + * @cpu: CPU number(logical index) for which device node is required + * @thread: if not NULL, local thread number within the physical core is + * returned + * + * The main purpose of this function is to retrieve the device node for the + * given logical CPU index. It should be used to initialize the of_node in + * cpu device. Once of_node in cpu device is populated, all the further + * references can use that instead. + * + * CPU logical to physical index mapping is architecture specific and is built + * before booting secondary cores. This function uses arch_match_cpu_phys_id + * which can be overridden by architecture specific implementation. + * + * Returns a node pointer for the logical cpu if found, else NULL. + */ +struct device_node *of_get_cpu_node(int cpu, unsigned int *thread) +{ + struct device_node *cpun, *cpus; + + cpus = of_find_node_by_path("/cpus"); + if (!cpus) { + pr_warn("Missing cpus node, bailing out\n"); + return NULL; + } + + for_each_child_of_node(cpus, cpun) { + if (of_node_cmp(cpun->type, "cpu")) + continue; + /* Check for non-standard "ibm,ppc-interrupt-server#s" property + * for thread ids on PowerPC. If it doesn't exist fallback to + * standard "reg" property. + */ + if (IS_ENABLED(CONFIG_PPC) && + __of_find_n_match_cpu_property(cpun, + "ibm,ppc-interrupt-server#s", cpu, thread)) + return cpun; + if (__of_find_n_match_cpu_property(cpun, "reg", cpu, thread)) + return cpun; + } + return NULL; +} +EXPORT_SYMBOL(of_get_cpu_node); + /** Checks if the given "compat" string matches one of the strings in * the device's "compatible" property */ @@ -1551,6 +1647,15 @@ void of_alias_scan(void * (*dt_alloc)(u64 size, u64 align)) of_chosen = of_find_node_by_path("/chosen"); if (of_chosen == NULL) of_chosen = of_find_node_by_path("/chosen@0"); + + if (of_chosen) { + const char *name; + + name = of_get_property(of_chosen, "linux,stdout-path", NULL); + if (name) + of_stdout = of_find_node_by_path(name); + } + of_aliases = of_find_node_by_path("/aliases"); if (!of_aliases) return; @@ -1660,3 +1765,19 @@ const char *of_prop_next_string(struct property *prop, const char *cur) return curv; } EXPORT_SYMBOL_GPL(of_prop_next_string); + +/** + * of_device_is_stdout_path - check if a device node matches the + * linux,stdout-path property + * + * Check if this device node matches the linux,stdout-path property + * in the chosen node. return true if yes, false otherwise. + */ +int of_device_is_stdout_path(struct device_node *dn) +{ + if (!of_stdout) + return false; + + return of_stdout == dn; +} +EXPORT_SYMBOL_GPL(of_device_is_stdout_path); diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 118773751ea4..4911158cba8a 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -14,9 +14,13 @@ #include <linux/module.h> #include <linux/of.h> #include <linux/of_fdt.h> +#include <linux/of_reserved_mem.h> +#include <linux/sizes.h> #include <linux/string.h> #include <linux/errno.h> #include <linux/slab.h> +#include <linux/memblock.h> +#include <linux/libfdt.h> #include <asm/setup.h> /* for COMMAND_LINE_SIZE */ #ifdef CONFIG_PPC @@ -25,54 +29,6 @@ #include <asm/page.h> -char *of_fdt_get_string(struct boot_param_header *blob, u32 offset) -{ - return ((char *)blob) + - be32_to_cpu(blob->off_dt_strings) + offset; -} - -/** - * of_fdt_get_property - Given a node in the given flat blob, return - * the property ptr - */ -void *of_fdt_get_property(struct boot_param_header *blob, - unsigned long node, const char *name, - unsigned long *size) -{ - unsigned long p = node; - - do { - u32 tag = be32_to_cpup((__be32 *)p); - u32 sz, noff; - const char *nstr; - - p += 4; - if (tag == OF_DT_NOP) - continue; - if (tag != OF_DT_PROP) - return NULL; - - sz = be32_to_cpup((__be32 *)p); - noff = be32_to_cpup((__be32 *)(p + 4)); - p += 8; - if (be32_to_cpu(blob->version) < 0x10) - p = ALIGN(p, sz >= 8 ? 8 : 4); - - nstr = of_fdt_get_string(blob, noff); - if (nstr == NULL) { - pr_warning("Can't find property index name !\n"); - return NULL; - } - if (strcmp(name, nstr) == 0) { - if (size) - *size = sz; - return (void *)p; - } - p += sz; - p = ALIGN(p, 4); - } while (1); -} - /** * of_fdt_is_compatible - Return true if given node from the given blob has * compat in its compatible list @@ -87,9 +43,10 @@ int of_fdt_is_compatible(struct boot_param_header *blob, unsigned long node, const char *compat) { const char *cp; - unsigned long cplen, l, score = 0; + int cplen; + unsigned long l, score = 0; - cp = of_fdt_get_property(blob, node, "compatible", &cplen); + cp = fdt_getprop(blob, node, "compatible", &cplen); if (cp == NULL) return 0; while (cplen > 0) { @@ -125,12 +82,12 @@ int of_fdt_match(struct boot_param_header *blob, unsigned long node, return score; } -static void *unflatten_dt_alloc(unsigned long *mem, unsigned long size, +static void *unflatten_dt_alloc(void **mem, unsigned long size, unsigned long align) { void *res; - *mem = ALIGN(*mem, align); + *mem = PTR_ALIGN(*mem, align); res = (void *)*mem; *mem += size; @@ -146,30 +103,29 @@ static void *unflatten_dt_alloc(unsigned long *mem, unsigned long size, * @allnextpp: pointer to ->allnext from last allocated device_node * @fpsize: Size of the node path up at the current depth. */ -static unsigned long unflatten_dt_node(struct boot_param_header *blob, - unsigned long mem, - unsigned long *p, +static void * unflatten_dt_node(struct boot_param_header *blob, + void *mem, + int *poffset, struct device_node *dad, struct device_node ***allnextpp, unsigned long fpsize) { + const __be32 *p; struct device_node *np; struct property *pp, **prev_pp = NULL; - char *pathp; - u32 tag; + const char *pathp; unsigned int l, allocl; + static int depth = 0; + int old_depth; + int offset; int has_name = 0; int new_format = 0; - tag = be32_to_cpup((__be32 *)(*p)); - if (tag != OF_DT_BEGIN_NODE) { - pr_err("Weird tag at start of node: %x\n", tag); + pathp = fdt_get_name(blob, *poffset, &l); + if (!pathp) return mem; - } - *p += 4; - pathp = (char *)*p; - l = allocl = strlen(pathp) + 1; - *p = ALIGN(*p + l, 4); + + allocl = l++; /* version 0x10 has a more compact unit name here instead of the full * path. we accumulate the full path size using "fpsize", we'll rebuild @@ -187,7 +143,7 @@ static unsigned long unflatten_dt_node(struct boot_param_header *blob, fpsize = 1; allocl = 2; l = 1; - *pathp = '\0'; + pathp = ""; } else { /* account for '/' and path size minus terminal 0 * already in 'l' @@ -235,32 +191,23 @@ static unsigned long unflatten_dt_node(struct boot_param_header *blob, kref_init(&np->kref); } /* process properties */ - while (1) { - u32 sz, noff; - char *pname; - - tag = be32_to_cpup((__be32 *)(*p)); - if (tag == OF_DT_NOP) { - *p += 4; - continue; - } - if (tag != OF_DT_PROP) + for (offset = fdt_first_property_offset(blob, *poffset); + (offset >= 0); + (offset = fdt_next_property_offset(blob, offset))) { + const char *pname; + u32 sz; + + if (!(p = fdt_getprop_by_offset(blob, offset, &pname, &sz))) { + offset = -FDT_ERR_INTERNAL; break; - *p += 4; - sz = be32_to_cpup((__be32 *)(*p)); - noff = be32_to_cpup((__be32 *)((*p) + 4)); - *p += 8; - if (be32_to_cpu(blob->version) < 0x10) - *p = ALIGN(*p, sz >= 8 ? 8 : 4); - - pname = of_fdt_get_string(blob, noff); + } + if (pname == NULL) { pr_info("Can't find property name in list !\n"); break; } if (strcmp(pname, "name") == 0) has_name = 1; - l = strlen(pname) + 1; pp = unflatten_dt_alloc(&mem, sizeof(struct property), __alignof__(struct property)); if (allnextpp) { @@ -272,26 +219,25 @@ static unsigned long unflatten_dt_node(struct boot_param_header *blob, if ((strcmp(pname, "phandle") == 0) || (strcmp(pname, "linux,phandle") == 0)) { if (np->phandle == 0) - np->phandle = be32_to_cpup((__be32*)*p); + np->phandle = be32_to_cpup(p); } /* And we process the "ibm,phandle" property * used in pSeries dynamic device tree * stuff */ if (strcmp(pname, "ibm,phandle") == 0) - np->phandle = be32_to_cpup((__be32 *)*p); - pp->name = pname; + np->phandle = be32_to_cpup(p); + pp->name = (char *)pname; pp->length = sz; - pp->value = (void *)*p; + pp->value = (__be32 *)p; *prev_pp = pp; prev_pp = &pp->next; } - *p = ALIGN((*p) + sz, 4); } /* with version 0x10 we may not have the name property, recreate * it here from the unit name if absent */ if (!has_name) { - char *p1 = pathp, *ps = pathp, *pa = NULL; + const char *p1 = pathp, *ps = pathp, *pa = NULL; int sz; while (*p1) { @@ -328,19 +274,18 @@ static unsigned long unflatten_dt_node(struct boot_param_header *blob, if (!np->type) np->type = "<NULL>"; } - while (tag == OF_DT_BEGIN_NODE || tag == OF_DT_NOP) { - if (tag == OF_DT_NOP) - *p += 4; - else - mem = unflatten_dt_node(blob, mem, p, np, allnextpp, - fpsize); - tag = be32_to_cpup((__be32 *)(*p)); - } - if (tag != OF_DT_END_NODE) { - pr_err("Weird tag at end of node: %x\n", tag); - return mem; - } - *p += 4; + + old_depth = depth; + *poffset = fdt_next_node(blob, *poffset, &depth); + if (depth < 0) + depth = 0; + while (*poffset > 0 && depth > old_depth) + mem = unflatten_dt_node(blob, mem, poffset, np, allnextpp, + fpsize); + + if (*poffset < 0 && *poffset != -FDT_ERR_NOTFOUND) + pr_err("unflatten: error %d processing FDT\n", *poffset); + return mem; } @@ -360,7 +305,9 @@ static void __unflatten_device_tree(struct boot_param_header *blob, struct device_node **mynodes, void * (*dt_alloc)(u64 size, u64 align)) { - unsigned long start, mem, size; + unsigned long size; + int start; + void *mem; struct device_node **allnextp = mynodes; pr_debug(" -> unflatten_device_tree()\n"); @@ -381,30 +328,25 @@ static void __unflatten_device_tree(struct boot_param_header *blob, } /* First pass, scan for size */ - start = ((unsigned long)blob) + - be32_to_cpu(blob->off_dt_struct); - size = unflatten_dt_node(blob, 0, &start, NULL, NULL, 0); - size = (size | 3) + 1; + start = 0; + size = (unsigned long)unflatten_dt_node(blob, 0, &start, NULL, NULL, 0); + size = ALIGN(size, 4); pr_debug(" size is %lx, allocating...\n", size); /* Allocate memory for the expanded device tree */ - mem = (unsigned long) - dt_alloc(size + 4, __alignof__(struct device_node)); + mem = dt_alloc(size + 4, __alignof__(struct device_node)); memset((void *)mem, 0, size); ((__be32 *)mem)[size / 4] = cpu_to_be32(0xdeadbeef); - pr_debug(" unflattening %lx...\n", mem); + pr_debug(" unflattening %p...\n", mem); /* Second pass, do actual unflattening */ - start = ((unsigned long)blob) + - be32_to_cpu(blob->off_dt_struct); + start = 0; unflatten_dt_node(blob, mem, &start, NULL, &allnextp, 0); - if (be32_to_cpup((__be32 *)start) != OF_DT_END) - pr_warning("Weird tag at end of tree: %08x\n", *((u32 *)start)); - if (be32_to_cpu(((__be32 *)mem)[size / 4]) != 0xdeadbeef) + if (be32_to_cpup(mem + size) != 0xdeadbeef) pr_warning("End of tree marker overwritten: %08x\n", be32_to_cpu(((__be32 *)mem)[size / 4])); *allnextp = NULL; @@ -443,6 +385,129 @@ struct boot_param_header *initial_boot_params; #ifdef CONFIG_OF_EARLY_FLATTREE /** + * res_mem_reserve_reg() - reserve all memory described in 'reg' property + */ +static int __init __reserved_mem_reserve_reg(unsigned long node, + const char *uname) +{ + int t_len = (dt_root_addr_cells + dt_root_size_cells) * sizeof(__be32); + phys_addr_t base, size; + int len; + const __be32 *prop; + int nomap, first = 1; + + prop = of_get_flat_dt_prop(node, "reg", &len); + if (!prop) + return -ENOENT; + + if (len && len % t_len != 0) { + pr_err("Reserved memory: invalid reg property in '%s', skipping node.\n", + uname); + return -EINVAL; + } + + nomap = of_get_flat_dt_prop(node, "no-map", NULL) != NULL; + + while (len >= t_len) { + base = dt_mem_next_cell(dt_root_addr_cells, &prop); + size = dt_mem_next_cell(dt_root_size_cells, &prop); + + if (base && size && + early_init_dt_reserve_memory_arch(base, size, nomap) == 0) + pr_debug("Reserved memory: reserved region for node '%s': base %pa, size %ld MiB\n", + uname, &base, (unsigned long)size / SZ_1M); + else + pr_info("Reserved memory: failed to reserve memory for node '%s': base %pa, size %ld MiB\n", + uname, &base, (unsigned long)size / SZ_1M); + + len -= t_len; + if (first) { + fdt_reserved_mem_save_node(node, uname, base, size); + first = 0; + } + } + return 0; +} + +/** + * __reserved_mem_check_root() - check if #size-cells, #address-cells provided + * in /reserved-memory matches the values supported by the current implementation, + * also check if ranges property has been provided + */ +static int __init __reserved_mem_check_root(unsigned long node) +{ + const __be32 *prop; + + prop = of_get_flat_dt_prop(node, "#size-cells", NULL); + if (!prop || be32_to_cpup(prop) != dt_root_size_cells) + return -EINVAL; + + prop = of_get_flat_dt_prop(node, "#address-cells", NULL); + if (!prop || be32_to_cpup(prop) != dt_root_addr_cells) + return -EINVAL; + + prop = of_get_flat_dt_prop(node, "ranges", NULL); + if (!prop) + return -EINVAL; + return 0; +} + +/** + * fdt_scan_reserved_mem() - scan a single FDT node for reserved memory + */ +static int __init __fdt_scan_reserved_mem(unsigned long node, const char *uname, + int depth, void *data) +{ + static int found; + const char *status; + int err; + + if (!found && depth == 1 && strcmp(uname, "reserved-memory") == 0) { + if (__reserved_mem_check_root(node) != 0) { + pr_err("Reserved memory: unsupported node format, ignoring\n"); + /* break scan */ + return 1; + } + found = 1; + /* scan next node */ + return 0; + } else if (!found) { + /* scan next node */ + return 0; + } else if (found && depth < 2) { + /* scanning of /reserved-memory has been finished */ + return 1; + } + + status = of_get_flat_dt_prop(node, "status", NULL); + if (status && strcmp(status, "okay") != 0 && strcmp(status, "ok") != 0) + return 0; + + err = __reserved_mem_reserve_reg(node, uname); + if (err == -ENOENT && of_get_flat_dt_prop(node, "size", NULL)) + fdt_reserved_mem_save_node(node, uname, 0, 0); + + /* scan next node */ + return 0; +} + +/** + * early_init_fdt_scan_reserved_mem() - create reserved memory regions + * + * This function grabs memory from early allocator for device exclusive use + * defined in device tree structures. It should be called by arch specific code + * once the early allocator (i.e. memblock) has been fully activated. + */ +void __init early_init_fdt_scan_reserved_mem(void) +{ + if (!initial_boot_params) + return; + + of_scan_flat_dt(__fdt_scan_reserved_mem, NULL); + fdt_init_reserved_mem(); +} + +/** * of_scan_flat_dt - scan flattened tree blob and call callback on each. * @it: callback function * @data: context data pointer @@ -456,47 +521,19 @@ int __init of_scan_flat_dt(int (*it)(unsigned long node, void *data), void *data) { - unsigned long p = ((unsigned long)initial_boot_params) + - be32_to_cpu(initial_boot_params->off_dt_struct); - int rc = 0; - int depth = -1; - - do { - u32 tag = be32_to_cpup((__be32 *)p); - const char *pathp; - - p += 4; - if (tag == OF_DT_END_NODE) { - depth--; - continue; - } - if (tag == OF_DT_NOP) - continue; - if (tag == OF_DT_END) - break; - if (tag == OF_DT_PROP) { - u32 sz = be32_to_cpup((__be32 *)p); - p += 8; - if (be32_to_cpu(initial_boot_params->version) < 0x10) - p = ALIGN(p, sz >= 8 ? 8 : 4); - p += sz; - p = ALIGN(p, 4); - continue; - } - if (tag != OF_DT_BEGIN_NODE) { - pr_err("Invalid tag %x in flat device tree!\n", tag); - return -EINVAL; - } - depth++; - pathp = (char *)p; - p = ALIGN(p + strlen(pathp) + 1, 4); + const void *blob = initial_boot_params; + const char *pathp; + int offset, rc = 0, depth = -1; + + for (offset = fdt_next_node(blob, -1, &depth); + offset >= 0 && depth >= 0 && !rc; + offset = fdt_next_node(blob, offset, &depth)) { + + pathp = fdt_get_name(blob, offset, NULL); if (*pathp == '/') pathp = kbasename(pathp); - rc = it(p, pathp, depth, data); - if (rc != 0) - break; - } while (1); - + rc = it(offset, pathp, depth, data); + } return rc; } @@ -505,14 +542,7 @@ int __init of_scan_flat_dt(int (*it)(unsigned long node, */ unsigned long __init of_get_flat_dt_root(void) { - unsigned long p = ((unsigned long)initial_boot_params) + - be32_to_cpu(initial_boot_params->off_dt_struct); - - while (be32_to_cpup((__be32 *)p) == OF_DT_NOP) - p += 4; - BUG_ON(be32_to_cpup((__be32 *)p) != OF_DT_BEGIN_NODE); - p += 4; - return ALIGN(p + strlen((char *)p) + 1, 4); + return 0; } /** @@ -521,10 +551,10 @@ unsigned long __init of_get_flat_dt_root(void) * This function can be used within scan_flattened_dt callback to get * access to properties */ -void *__init of_get_flat_dt_prop(unsigned long node, const char *name, - unsigned long *size) +const void *__init of_get_flat_dt_prop(unsigned long node, const char *name, + int *size) { - return of_fdt_get_property(initial_boot_params, node, name, size); + return fdt_getprop(initial_boot_params, node, name, size); } /** @@ -545,6 +575,15 @@ int __init of_flat_dt_match(unsigned long node, const char *const *compat) return of_fdt_match(initial_boot_params, node, compat); } +struct fdt_scan_status { + const char *name; + int namelen; + int depth; + int found; + int (*iterator)(unsigned long node, const char *uname, int depth, void *data); + void *data; +}; + #ifdef CONFIG_BLK_DEV_INITRD /** * early_init_dt_check_for_initrd - Decode initrd location from flat tree @@ -552,23 +591,25 @@ int __init of_flat_dt_match(unsigned long node, const char *const *compat) */ void __init early_init_dt_check_for_initrd(unsigned long node) { - unsigned long start, end, len; - __be32 *prop; + u64 start, end; + int len; + const __be32 *prop; pr_debug("Looking for initrd properties... "); prop = of_get_flat_dt_prop(node, "linux,initrd-start", &len); if (!prop) return; - start = of_read_ulong(prop, len/4); + start = of_read_number(prop, len/4); prop = of_get_flat_dt_prop(node, "linux,initrd-end", &len); if (!prop) return; - end = of_read_ulong(prop, len/4); + end = of_read_number(prop, len/4); early_init_dt_setup_initrd_arch(start, end); - pr_debug("initrd_start=0x%lx initrd_end=0x%lx\n", start, end); + pr_debug("initrd_start=0x%llx initrd_end=0x%llx\n", + (unsigned long long)start, (unsigned long long)end); } #else inline void early_init_dt_check_for_initrd(unsigned long node) @@ -582,7 +623,7 @@ inline void early_init_dt_check_for_initrd(unsigned long node) int __init early_init_dt_scan_root(unsigned long node, const char *uname, int depth, void *data) { - __be32 *prop; + const __be32 *prop; if (depth != 0) return 0; @@ -604,9 +645,9 @@ int __init early_init_dt_scan_root(unsigned long node, const char *uname, return 1; } -u64 __init dt_mem_next_cell(int s, __be32 **cellp) +u64 __init dt_mem_next_cell(int s, const __be32 **cellp) { - __be32 *p = *cellp; + const __be32 *p = *cellp; *cellp = p + s; return of_read_number(p, s); @@ -618,9 +659,9 @@ u64 __init dt_mem_next_cell(int s, __be32 **cellp) int __init early_init_dt_scan_memory(unsigned long node, const char *uname, int depth, void *data) { - char *type = of_get_flat_dt_prop(node, "device_type", NULL); - __be32 *reg, *endp; - unsigned long l; + const char *type = of_get_flat_dt_prop(node, "device_type", NULL); + const __be32 *reg, *endp; + int l; /* We are scanning "memory" nodes only */ if (type == NULL) { @@ -641,7 +682,7 @@ int __init early_init_dt_scan_memory(unsigned long node, const char *uname, endp = reg + (l / sizeof(__be32)); - pr_debug("memory scan node %s, reg size %ld, data: %x %x %x %x,\n", + pr_debug("memory scan node %s, reg size %d, data: %x %x %x %x,\n", uname, l, reg[0], reg[1], reg[2], reg[3]); while ((endp - reg) >= (dt_root_addr_cells + dt_root_size_cells)) { @@ -664,8 +705,8 @@ int __init early_init_dt_scan_memory(unsigned long node, const char *uname, int __init early_init_dt_scan_chosen(unsigned long node, const char *uname, int depth, void *data) { - unsigned long l; - char *p; + int l; + const char *p; pr_debug("search \"chosen\", depth: %d, uname: %s\n", depth, uname); @@ -698,6 +739,80 @@ int __init early_init_dt_scan_chosen(unsigned long node, const char *uname, return 1; } +#ifdef CONFIG_HAVE_MEMBLOCK +void __init __weak early_init_dt_add_memory_arch(u64 base, u64 size) +{ + const u64 phys_offset = __pa(PAGE_OFFSET); + base &= PAGE_MASK; + size &= PAGE_MASK; + if (base + size < phys_offset) { + pr_warning("Ignoring memory block 0x%llx - 0x%llx\n", + base, base + size); + return; + } + if (base < phys_offset) { + pr_warning("Ignoring memory range 0x%llx - 0x%llx\n", + base, phys_offset); + size -= phys_offset - base; + base = phys_offset; + } + memblock_add(base, size); +} + +int __init __weak early_init_dt_reserve_memory_arch(phys_addr_t base, + phys_addr_t size, bool nomap) +{ + if (memblock_is_region_reserved(base, size)) + return -EBUSY; + if (nomap) + return memblock_remove(base, size); + return memblock_reserve(base, size); +} + +/* + * called from unflatten_device_tree() to bootstrap devicetree itself + * Architectures can override this definition if memblock isn't used + */ +void * __init __weak early_init_dt_alloc_memory_arch(u64 size, u64 align) +{ + return __va(memblock_alloc(size, align)); +} +#else +int __init __weak early_init_dt_reserve_memory_arch(phys_addr_t base, + phys_addr_t size, bool nomap) +{ + pr_err("Reserved memory not supported, ignoring range 0x%llx - 0x%llx%s\n", + base, size, nomap ? " (nomap)" : ""); + return -ENOSYS; +} +#endif + +bool __init early_init_dt_scan(void *params) +{ + if (!params) + return false; + + /* Setup flat device-tree pointer */ + initial_boot_params = params; + + /* check device tree validity */ + if (be32_to_cpu(initial_boot_params->magic) != OF_DT_HEADER) { + initial_boot_params = NULL; + return false; + } + + /* Retrieve various information from the /chosen node */ + of_scan_flat_dt(early_init_dt_scan_chosen, boot_command_line); + + /* Initialize {size,address}-cells info */ + of_scan_flat_dt(early_init_dt_scan_root, NULL); + + /* Setup memory, calling early_init_dt_add_memory_arch */ + of_scan_flat_dt(early_init_dt_scan_memory, NULL); + + return true; +} + /** * unflatten_device_tree - create tree of device_nodes from flat blob * diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c new file mode 100644 index 000000000000..e420eb52e5c9 --- /dev/null +++ b/drivers/of/of_reserved_mem.c @@ -0,0 +1,217 @@ +/* + * Device tree based initialization code for reserved memory. + * + * Copyright (c) 2013, The Linux Foundation. All Rights Reserved. + * Copyright (c) 2013,2014 Samsung Electronics Co., Ltd. + * http://www.samsung.com + * Author: Marek Szyprowski <m.szyprowski@samsung.com> + * Author: Josh Cartwright <joshc@codeaurora.org> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of the + * License or (at your optional) any later version of the license. + */ + +#include <linux/err.h> +#include <linux/of.h> +#include <linux/of_fdt.h> +#include <linux/of_platform.h> +#include <linux/mm.h> +#include <linux/sizes.h> +#include <linux/of_reserved_mem.h> + +#define MAX_RESERVED_REGIONS 16 +static struct reserved_mem reserved_mem[MAX_RESERVED_REGIONS]; +static int reserved_mem_count; + +#if defined(CONFIG_HAVE_MEMBLOCK) +#include <linux/memblock.h> +int __init __weak early_init_dt_alloc_reserved_memory_arch(phys_addr_t size, + phys_addr_t align, phys_addr_t start, phys_addr_t end, bool nomap, + phys_addr_t *res_base) +{ + /* + * We use __memblock_alloc_base() because memblock_alloc_base() + * panic()s on allocation failure. + */ + phys_addr_t base = __memblock_alloc_base(size, align, end); + if (!base) + return -ENOMEM; + + /* + * Check if the allocated region fits in to start..end window + */ + if (base < start) { + memblock_free(base, size); + return -ENOMEM; + } + + *res_base = base; + if (nomap) + return memblock_remove(base, size); + return 0; +} +#else +int __init __weak early_init_dt_alloc_reserved_memory_arch(phys_addr_t size, + phys_addr_t align, phys_addr_t start, phys_addr_t end, bool nomap, + phys_addr_t *res_base) +{ + pr_err("Reserved memory not supported, ignoring region 0x%llx%s\n", + size, nomap ? " (nomap)" : ""); + return -ENOSYS; +} +#endif + +/** + * res_mem_save_node() - save fdt node for second pass initialization + */ +void __init fdt_reserved_mem_save_node(unsigned long node, const char *uname, + phys_addr_t base, phys_addr_t size) +{ + struct reserved_mem *rmem = &reserved_mem[reserved_mem_count]; + + if (reserved_mem_count == ARRAY_SIZE(reserved_mem)) { + pr_err("Reserved memory: not enough space all defined regions.\n"); + return; + } + + rmem->fdt_node = node; + rmem->name = uname; + rmem->base = base; + rmem->size = size; + + reserved_mem_count++; + return; +} + +/** + * res_mem_alloc_size() - allocate reserved memory described by 'size', 'align' + * and 'alloc-ranges' properties + */ +static int __init __reserved_mem_alloc_size(unsigned long node, + const char *uname, phys_addr_t *res_base, phys_addr_t *res_size) +{ + int t_len = (dt_root_addr_cells + dt_root_size_cells) * sizeof(__be32); + phys_addr_t start = 0, end = 0; + phys_addr_t base = 0, align = 0, size; + int len; + const __be32 *prop; + int nomap; + int ret; + + prop = of_get_flat_dt_prop(node, "size", &len); + if (!prop) + return -EINVAL; + + if (len != dt_root_size_cells * sizeof(__be32)) { + pr_err("Reserved memory: invalid size property in '%s' node.\n", + uname); + return -EINVAL; + } + size = dt_mem_next_cell(dt_root_size_cells, &prop); + + nomap = of_get_flat_dt_prop(node, "no-map", NULL) != NULL; + + prop = of_get_flat_dt_prop(node, "alignment", &len); + if (prop) { + if (len != dt_root_addr_cells * sizeof(__be32)) { + pr_err("Reserved memory: invalid alignment property in '%s' node.\n", + uname); + return -EINVAL; + } + align = dt_mem_next_cell(dt_root_addr_cells, &prop); + } + + prop = of_get_flat_dt_prop(node, "alloc-ranges", &len); + if (prop) { + + if (len % t_len != 0) { + pr_err("Reserved memory: invalid alloc-ranges property in '%s', skipping node.\n", + uname); + return -EINVAL; + } + + base = 0; + + while (len > 0) { + start = dt_mem_next_cell(dt_root_addr_cells, &prop); + end = start + dt_mem_next_cell(dt_root_size_cells, + &prop); + + ret = early_init_dt_alloc_reserved_memory_arch(size, + align, start, end, nomap, &base); + if (ret == 0) { + pr_debug("Reserved memory: allocated memory for '%s' node: base %pa, size %ld MiB\n", + uname, &base, + (unsigned long)size / SZ_1M); + break; + } + len -= t_len; + } + + } else { + ret = early_init_dt_alloc_reserved_memory_arch(size, align, + 0, 0, nomap, &base); + if (ret == 0) + pr_debug("Reserved memory: allocated memory for '%s' node: base %pa, size %ld MiB\n", + uname, &base, (unsigned long)size / SZ_1M); + } + + if (base == 0) { + pr_info("Reserved memory: failed to allocate memory for node '%s'\n", + uname); + return -ENOMEM; + } + + *res_base = base; + *res_size = size; + + return 0; +} + +static const struct of_device_id __rmem_of_table_sentinel + __used __section(__reservedmem_of_table_end); + +/** + * res_mem_init_node() - call region specific reserved memory init code + */ +static int __init __reserved_mem_init_node(struct reserved_mem *rmem) +{ + extern const struct of_device_id __reservedmem_of_table[]; + const struct of_device_id *i; + + for (i = __reservedmem_of_table; i < &__rmem_of_table_sentinel; i++) { + reservedmem_of_init_fn initfn = i->data; + const char *compat = i->compatible; + + if (!of_flat_dt_is_compatible(rmem->fdt_node, compat)) + continue; + + if (initfn(rmem, rmem->fdt_node, rmem->name) == 0) { + pr_info("Reserved memory: initialized node %s, compatible id %s\n", + rmem->name, compat); + return 0; + } + } + return -ENOENT; +} + +/** + * fdt_init_reserved_mem - allocate and init all saved reserved memory regions + */ +void __init fdt_init_reserved_mem(void) +{ + int i; + for (i = 0; i < reserved_mem_count; i++) { + struct reserved_mem *rmem = &reserved_mem[i]; + unsigned long node = rmem->fdt_node; + int err = 0; + + if (rmem->size == 0) + err = __reserved_mem_alloc_size(node, rmem->name, + &rmem->base, &rmem->size); + if (err == 0) + __reserved_mem_init_node(rmem); + } +} diff --git a/drivers/pinctrl/Makefile b/drivers/pinctrl/Makefile index 9bdaeb8785ce..b32d6c798743 100644 --- a/drivers/pinctrl/Makefile +++ b/drivers/pinctrl/Makefile @@ -2,7 +2,7 @@ ccflags-$(CONFIG_DEBUG_PINCTRL) += -DDEBUG -obj-$(CONFIG_PINCTRL) += core.o +obj-$(CONFIG_PINCTRL) += core.o pinctrl-utils.o obj-$(CONFIG_PINMUX) += pinmux.o obj-$(CONFIG_PINCONF) += pinconf.o ifeq ($(CONFIG_OF),y) diff --git a/drivers/pinctrl/pinconf-generic.c b/drivers/pinctrl/pinconf-generic.c index 2ad5a8d337b5..872e74a23788 100644 --- a/drivers/pinctrl/pinconf-generic.c +++ b/drivers/pinctrl/pinconf-generic.c @@ -21,8 +21,10 @@ #include <linux/pinctrl/pinctrl.h> #include <linux/pinctrl/pinconf.h> #include <linux/pinctrl/pinconf-generic.h> +#include <linux/of.h> #include "core.h" #include "pinconf.h" +#include "pinctrl-utils.h" #ifdef CONFIG_DEBUG_FS @@ -37,14 +39,19 @@ struct pin_config_item { static struct pin_config_item conf_items[] = { PCONFDUMP(PIN_CONFIG_BIAS_DISABLE, "input bias disabled", NULL), PCONFDUMP(PIN_CONFIG_BIAS_HIGH_IMPEDANCE, "input bias high impedance", NULL), + PCONFDUMP(PIN_CONFIG_BIAS_BUS_HOLD, "input bias bus hold", NULL), PCONFDUMP(PIN_CONFIG_BIAS_PULL_UP, "input bias pull up", NULL), PCONFDUMP(PIN_CONFIG_BIAS_PULL_DOWN, "input bias pull down", NULL), + PCONFDUMP(PIN_CONFIG_BIAS_PULL_PIN_DEFAULT, + "input bias pull to pin specific state", NULL), PCONFDUMP(PIN_CONFIG_DRIVE_PUSH_PULL, "output drive push pull", NULL), PCONFDUMP(PIN_CONFIG_DRIVE_OPEN_DRAIN, "output drive open drain", NULL), PCONFDUMP(PIN_CONFIG_DRIVE_OPEN_SOURCE, "output drive open source", NULL), + PCONFDUMP(PIN_CONFIG_DRIVE_STRENGTH, "output drive strength", "mA"), + PCONFDUMP(PIN_CONFIG_INPUT_ENABLE, "input enabled", NULL), PCONFDUMP(PIN_CONFIG_INPUT_SCHMITT_ENABLE, "input schmitt enabled", NULL), PCONFDUMP(PIN_CONFIG_INPUT_SCHMITT, "input schmitt trigger", NULL), - PCONFDUMP(PIN_CONFIG_INPUT_DEBOUNCE, "input debounce", "time units"), + PCONFDUMP(PIN_CONFIG_INPUT_DEBOUNCE, "input debounce", "usec"), PCONFDUMP(PIN_CONFIG_POWER_SOURCE, "pin power source", "selector"), PCONFDUMP(PIN_CONFIG_SLEW_RATE, "slew rate", NULL), PCONFDUMP(PIN_CONFIG_LOW_POWER_MODE, "pin low power", "mode"), @@ -135,3 +142,198 @@ void pinconf_generic_dump_config(struct pinctrl_dev *pctldev, } EXPORT_SYMBOL_GPL(pinconf_generic_dump_config); #endif + +#ifdef CONFIG_OF +struct pinconf_generic_dt_params { + const char * const property; + enum pin_config_param param; + u32 default_value; +}; + +static struct pinconf_generic_dt_params dt_params[] = { + { "bias-disable", PIN_CONFIG_BIAS_DISABLE, 0 }, + { "bias-high-impedance", PIN_CONFIG_BIAS_HIGH_IMPEDANCE, 0 }, + { "bias-bus-hold", PIN_CONFIG_BIAS_BUS_HOLD, 0 }, + { "bias-pull-up", PIN_CONFIG_BIAS_PULL_UP, 1 }, + { "bias-pull-down", PIN_CONFIG_BIAS_PULL_DOWN, 1 }, + { "bias-pull-pin-default", PIN_CONFIG_BIAS_PULL_PIN_DEFAULT, 1 }, + { "drive-push-pull", PIN_CONFIG_DRIVE_PUSH_PULL, 0 }, + { "drive-open-drain", PIN_CONFIG_DRIVE_OPEN_DRAIN, 0 }, + { "drive-open-source", PIN_CONFIG_DRIVE_OPEN_SOURCE, 0 }, + { "drive-strength", PIN_CONFIG_DRIVE_STRENGTH, 0 }, + { "input-enable", PIN_CONFIG_INPUT_ENABLE, 1 }, + { "input-disable", PIN_CONFIG_INPUT_ENABLE, 0 }, + { "input-schmitt-enable", PIN_CONFIG_INPUT_SCHMITT_ENABLE, 1 }, + { "input-schmitt-disable", PIN_CONFIG_INPUT_SCHMITT_ENABLE, 0 }, + { "input-debounce", PIN_CONFIG_INPUT_DEBOUNCE, 0 }, + { "low-power-enable", PIN_CONFIG_LOW_POWER_MODE, 1 }, + { "low-power-disable", PIN_CONFIG_LOW_POWER_MODE, 0 }, + { "output-low", PIN_CONFIG_OUTPUT, 0, }, + { "output-high", PIN_CONFIG_OUTPUT, 1, }, + { "slew-rate", PIN_CONFIG_SLEW_RATE, 0}, +}; + +/** + * pinconf_generic_parse_dt_config() + * parse the config properties into generic pinconfig values. + * @np: node containing the pinconfig properties + * @configs: array with nconfigs entries containing the generic pinconf values + * @nconfigs: umber of configurations + */ +int pinconf_generic_parse_dt_config(struct device_node *np, + unsigned long **configs, + unsigned int *nconfigs) +{ + unsigned long *cfg; + unsigned int ncfg = 0; + int ret; + int i; + u32 val; + + if (!np) + return -EINVAL; + + /* allocate a temporary array big enough to hold one of each option */ + cfg = kzalloc(sizeof(*cfg) * ARRAY_SIZE(dt_params), GFP_KERNEL); + if (!cfg) + return -ENOMEM; + + for (i = 0; i < ARRAY_SIZE(dt_params); i++) { + struct pinconf_generic_dt_params *par = &dt_params[i]; + ret = of_property_read_u32(np, par->property, &val); + + /* property not found */ + if (ret == -EINVAL) + continue; + + /* use default value, when no value is specified */ + if (ret) + val = par->default_value; + + pr_debug("found %s with value %u\n", par->property, val); + cfg[ncfg] = pinconf_to_config_packed(par->param, val); + ncfg++; + } + + ret = 0; + + /* no configs found at all */ + if (ncfg == 0) { + *configs = NULL; + *nconfigs = 0; + goto out; + } + + /* + * Now limit the number of configs to the real number of + * found properties. + */ + *configs = kzalloc(ncfg * sizeof(unsigned long), GFP_KERNEL); + if (!*configs) { + ret = -ENOMEM; + goto out; + } + + memcpy(*configs, cfg, ncfg * sizeof(unsigned long)); + *nconfigs = ncfg; + +out: + kfree(cfg); + return ret; +} + +int pinconf_generic_dt_subnode_to_map_new(struct pinctrl_dev *pctldev, + struct device_node *np, struct pinctrl_map **map, + unsigned *reserved_maps, unsigned *num_maps, + enum pinctrl_map_type type) +{ + int ret; + const char *function; + struct device *dev = pctldev->dev; + unsigned long *configs = NULL; + unsigned num_configs = 0; + unsigned reserve; + struct property *prop; + const char *group; + + ret = of_property_read_string(np, "function", &function); + if (ret < 0) { + /* EINVAL=missing, which is fine since it's optional */ + if (ret != -EINVAL) + dev_err(dev, "could not parse property function\n"); + function = NULL; + } + + ret = pinconf_generic_parse_dt_config(np, &configs, &num_configs); + if (ret < 0) { + dev_err(dev, "could not parse node property\n"); + return ret; + } + + reserve = 0; + if (function != NULL) + reserve++; + if (num_configs) + reserve++; + ret = of_property_count_strings(np, "pins"); + if (ret < 0) { + dev_err(dev, "could not parse property pins\n"); + goto exit; + } + reserve *= ret; + + ret = pinctrl_utils_reserve_map(pctldev, map, reserved_maps, + num_maps, reserve); + if (ret < 0) + goto exit; + + of_property_for_each_string(np, "pins", prop, group) { + if (function) { + ret = pinctrl_utils_add_map_mux(pctldev, map, + reserved_maps, num_maps, group, + function); + if (ret < 0) + goto exit; + } + + if (num_configs) { + ret = pinctrl_utils_add_map_configs(pctldev, map, + reserved_maps, num_maps, group, configs, + num_configs, type); + if (ret < 0) + goto exit; + } + } + ret = 0; + +exit: + kfree(configs); + return ret; +} +EXPORT_SYMBOL_GPL(pinconf_generic_dt_subnode_to_map_new); + +int pinconf_generic_dt_node_to_map_new(struct pinctrl_dev *pctldev, + struct device_node *np_config, struct pinctrl_map **map, + unsigned *num_maps, enum pinctrl_map_type type) +{ + unsigned reserved_maps; + struct device_node *np; + int ret; + + reserved_maps = 0; + *map = NULL; + *num_maps = 0; + + for_each_child_of_node(np_config, np) { + ret = pinconf_generic_dt_subnode_to_map_new(pctldev, np, map, + &reserved_maps, num_maps, type); + if (ret < 0) { + pinctrl_utils_dt_free_map(pctldev, *map, *num_maps); + return ret; + } + } + return 0; +} +EXPORT_SYMBOL_GPL(pinconf_generic_dt_node_to_map_new); + +#endif diff --git a/drivers/pinctrl/pinconf.c b/drivers/pinctrl/pinconf.c index 694c3ace4520..596a2522a6b1 100644 --- a/drivers/pinctrl/pinconf.c +++ b/drivers/pinctrl/pinconf.c @@ -35,7 +35,9 @@ int pinconf_check_ops(struct pinctrl_dev *pctldev) return -EINVAL; } /* We have to be able to config the pins in SOME way */ - if (!ops->pin_config_set && !ops->pin_config_group_set) { + if (!ops->pin_config_set && !ops->pin_config_group_set + && !ops->pin_config_set_bulk + && !ops->pin_config_group_set_bulk) { dev_err(pctldev->dev, "pinconf has to be able to set a pins config\n"); return -EINVAL; @@ -75,98 +77,6 @@ int pin_config_get_for_pin(struct pinctrl_dev *pctldev, unsigned pin, return ops->pin_config_get(pctldev, pin, config); } -/** - * pin_config_get() - get the configuration of a single pin parameter - * @dev_name: name of the pin controller device for this pin - * @name: name of the pin to get the config for - * @config: the config pointed to by this argument will be filled in with the - * current pin state, it can be used directly by drivers as a numeral, or - * it can be dereferenced to any struct. - */ -int pin_config_get(const char *dev_name, const char *name, - unsigned long *config) -{ - struct pinctrl_dev *pctldev; - int pin; - - pctldev = get_pinctrl_dev_from_devname(dev_name); - if (!pctldev) { - pin = -EINVAL; - return pin; - } - - mutex_lock(&pctldev->mutex); - - pin = pin_get_from_name(pctldev, name); - if (pin < 0) - goto unlock; - - pin = pin_config_get_for_pin(pctldev, pin, config); - -unlock: - mutex_unlock(&pctldev->mutex); - return pin; -} -EXPORT_SYMBOL(pin_config_get); - -static int pin_config_set_for_pin(struct pinctrl_dev *pctldev, unsigned pin, - unsigned long config) -{ - const struct pinconf_ops *ops = pctldev->desc->confops; - int ret; - - if (!ops || !ops->pin_config_set) { - dev_err(pctldev->dev, "cannot configure pin, missing " - "config function in driver\n"); - return -EINVAL; - } - - ret = ops->pin_config_set(pctldev, pin, config); - if (ret) { - dev_err(pctldev->dev, - "unable to set pin configuration on pin %d\n", pin); - return ret; - } - - return 0; -} - -/** - * pin_config_set() - set the configuration of a single pin parameter - * @dev_name: name of pin controller device for this pin - * @name: name of the pin to set the config for - * @config: the config in this argument will contain the desired pin state, it - * can be used directly by drivers as a numeral, or it can be dereferenced - * to any struct. - */ -int pin_config_set(const char *dev_name, const char *name, - unsigned long config) -{ - struct pinctrl_dev *pctldev; - int pin, ret; - - pctldev = get_pinctrl_dev_from_devname(dev_name); - if (!pctldev) { - ret = -EINVAL; - return ret; - } - - mutex_lock(&pctldev->mutex); - - pin = pin_get_from_name(pctldev, name); - if (pin < 0) { - ret = pin; - goto unlock; - } - - ret = pin_config_set_for_pin(pctldev, pin, config); - -unlock: - mutex_unlock(&pctldev->mutex); - return ret; -} -EXPORT_SYMBOL(pin_config_set); - int pin_config_group_get(const char *dev_name, const char *pin_group, unsigned long *config) { @@ -204,88 +114,6 @@ unlock: mutex_unlock(&pctldev->mutex); return ret; } -EXPORT_SYMBOL(pin_config_group_get); - -int pin_config_group_set(const char *dev_name, const char *pin_group, - unsigned long config) -{ - struct pinctrl_dev *pctldev; - const struct pinconf_ops *ops; - const struct pinctrl_ops *pctlops; - int selector; - const unsigned *pins; - unsigned num_pins; - int ret; - int i; - - pctldev = get_pinctrl_dev_from_devname(dev_name); - if (!pctldev) { - ret = -EINVAL; - return ret; - } - - mutex_lock(&pctldev->mutex); - - ops = pctldev->desc->confops; - pctlops = pctldev->desc->pctlops; - - if (!ops || (!ops->pin_config_group_set && !ops->pin_config_set)) { - dev_err(pctldev->dev, "cannot configure pin group, missing " - "config function in driver\n"); - ret = -EINVAL; - goto unlock; - } - - selector = pinctrl_get_group_selector(pctldev, pin_group); - if (selector < 0) { - ret = selector; - goto unlock; - } - - ret = pctlops->get_group_pins(pctldev, selector, &pins, &num_pins); - if (ret) { - dev_err(pctldev->dev, "cannot configure pin group, error " - "getting pins\n"); - goto unlock; - } - - /* - * If the pin controller supports handling entire groups we use that - * capability. - */ - if (ops->pin_config_group_set) { - ret = ops->pin_config_group_set(pctldev, selector, config); - /* - * If the pin controller prefer that a certain group be handled - * pin-by-pin as well, it returns -EAGAIN. - */ - if (ret != -EAGAIN) - goto unlock; - } - - /* - * If the controller cannot handle entire groups, we configure each pin - * individually. - */ - if (!ops->pin_config_set) { - ret = 0; - goto unlock; - } - - for (i = 0; i < num_pins; i++) { - ret = ops->pin_config_set(pctldev, pins[i], config); - if (ret < 0) - goto unlock; - } - - ret = 0; - -unlock: - mutex_unlock(&pctldev->mutex); - - return ret; -} -EXPORT_SYMBOL(pin_config_group_set); int pinconf_map_to_setting(struct pinctrl_map const *map, struct pinctrl_setting *setting) @@ -332,7 +160,7 @@ int pinconf_apply_setting(struct pinctrl_setting const *setting) { struct pinctrl_dev *pctldev = setting->pctldev; const struct pinconf_ops *ops = pctldev->desc->confops; - int i, ret; + int ret, i; if (!ops) { dev_err(pctldev->dev, "missing confops\n"); @@ -341,39 +169,66 @@ int pinconf_apply_setting(struct pinctrl_setting const *setting) switch (setting->type) { case PIN_MAP_TYPE_CONFIGS_PIN: - if (!ops->pin_config_set) { + if (!ops->pin_config_set && !ops->pin_config_set_bulk) { dev_err(pctldev->dev, "missing pin_config_set op\n"); return -EINVAL; } - for (i = 0; i < setting->data.configs.num_configs; i++) { - ret = ops->pin_config_set(pctldev, + if (ops->pin_config_set_bulk) { + ret = ops->pin_config_set_bulk(pctldev, setting->data.configs.group_or_pin, - setting->data.configs.configs[i]); + setting->data.configs.configs, + setting->data.configs.num_configs); if (ret < 0) { dev_err(pctldev->dev, - "pin_config_set op failed for pin %d config %08lx\n", - setting->data.configs.group_or_pin, - setting->data.configs.configs[i]); + "pin_config_set_bulk op failed for pin %d\n", + setting->data.configs.group_or_pin); return ret; } + } else if (ops->pin_config_set) { + for (i = 0; i < setting->data.configs.num_configs; i++) { + ret = ops->pin_config_set(pctldev, + setting->data.configs.group_or_pin, + setting->data.configs.configs[i]); + if (ret < 0) { + dev_err(pctldev->dev, + "pin_config_set op failed for pin %d config %08lx\n", + setting->data.configs.group_or_pin, + setting->data.configs.configs[i]); + return ret; + } + } } break; case PIN_MAP_TYPE_CONFIGS_GROUP: - if (!ops->pin_config_group_set) { + if (!ops->pin_config_group_set && + !ops->pin_config_group_set_bulk) { dev_err(pctldev->dev, "missing pin_config_group_set op\n"); return -EINVAL; } - for (i = 0; i < setting->data.configs.num_configs; i++) { - ret = ops->pin_config_group_set(pctldev, + if (ops->pin_config_group_set_bulk) { + ret = ops->pin_config_group_set_bulk(pctldev, setting->data.configs.group_or_pin, - setting->data.configs.configs[i]); + setting->data.configs.configs, + setting->data.configs.num_configs); if (ret < 0) { dev_err(pctldev->dev, - "pin_config_group_set op failed for group %d config %08lx\n", + "pin_config_group_set_bulk op failed for group %d\n", + setting->data.configs.group_or_pin); + return ret; + } + } else if (ops->pin_config_group_set) { + for (i = 0; i < setting->data.configs.num_configs; i++) { + ret = ops->pin_config_group_set(pctldev, setting->data.configs.group_or_pin, setting->data.configs.configs[i]); - return ret; + if (ret < 0) { + dev_err(pctldev->dev, + "pin_config_group_set op failed for group %d config %08lx\n", + setting->data.configs.group_or_pin, + setting->data.configs.configs[i]); + return ret; + } } } break; diff --git a/drivers/pinctrl/pinconf.h b/drivers/pinctrl/pinconf.h index 92c7267244d2..a4a5417e1413 100644 --- a/drivers/pinctrl/pinconf.h +++ b/drivers/pinctrl/pinconf.h @@ -123,3 +123,9 @@ static inline void pinconf_generic_dump_config(struct pinctrl_dev *pctldev, return; } #endif + +#if defined(CONFIG_GENERIC_PINCONF) && defined(CONFIG_OF) +int pinconf_generic_parse_dt_config(struct device_node *np, + unsigned long **configs, + unsigned int *nconfigs); +#endif diff --git a/drivers/pinctrl/pinctrl-utils.c b/drivers/pinctrl/pinctrl-utils.c new file mode 100644 index 000000000000..48277e025f84 --- /dev/null +++ b/drivers/pinctrl/pinctrl-utils.c @@ -0,0 +1,141 @@ +/* + * Utils functions to implement the pincontrol driver. + * + * Copyright (c) 2013, NVIDIA Corporation. + * + * Author: Laxman Dewangan <ldewangan@nvidia.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation version 2. + * + * This program is distributed "as is" WITHOUT ANY WARRANTY of any kind, + * whether express or implied; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + * 02111-1307, USA + */ +#include <linux/device.h> +#include <linux/kernel.h> +#include <linux/pinctrl/pinctrl.h> +#include <linux/of.h> +#include <linux/slab.h> +#include "core.h" +#include "pinctrl-utils.h" + +int pinctrl_utils_reserve_map(struct pinctrl_dev *pctldev, + struct pinctrl_map **map, unsigned *reserved_maps, + unsigned *num_maps, unsigned reserve) +{ + unsigned old_num = *reserved_maps; + unsigned new_num = *num_maps + reserve; + struct pinctrl_map *new_map; + + if (old_num >= new_num) + return 0; + + new_map = krealloc(*map, sizeof(*new_map) * new_num, GFP_KERNEL); + if (!new_map) { + dev_err(pctldev->dev, "krealloc(map) failed\n"); + return -ENOMEM; + } + + memset(new_map + old_num, 0, (new_num - old_num) * sizeof(*new_map)); + + *map = new_map; + *reserved_maps = new_num; + return 0; +} +EXPORT_SYMBOL_GPL(pinctrl_utils_reserve_map); + +int pinctrl_utils_add_map_mux(struct pinctrl_dev *pctldev, + struct pinctrl_map **map, unsigned *reserved_maps, + unsigned *num_maps, const char *group, + const char *function) +{ + if (WARN_ON(*num_maps == *reserved_maps)) + return -ENOSPC; + + (*map)[*num_maps].type = PIN_MAP_TYPE_MUX_GROUP; + (*map)[*num_maps].data.mux.group = group; + (*map)[*num_maps].data.mux.function = function; + (*num_maps)++; + + return 0; +} +EXPORT_SYMBOL_GPL(pinctrl_utils_add_map_mux); + +int pinctrl_utils_add_map_configs(struct pinctrl_dev *pctldev, + struct pinctrl_map **map, unsigned *reserved_maps, + unsigned *num_maps, const char *group, + unsigned long *configs, unsigned num_configs, + enum pinctrl_map_type type) +{ + unsigned long *dup_configs; + + if (WARN_ON(*num_maps == *reserved_maps)) + return -ENOSPC; + + dup_configs = kmemdup(configs, num_configs * sizeof(*dup_configs), + GFP_KERNEL); + if (!dup_configs) { + dev_err(pctldev->dev, "kmemdup(configs) failed\n"); + return -ENOMEM; + } + + (*map)[*num_maps].type = type; + (*map)[*num_maps].data.configs.group_or_pin = group; + (*map)[*num_maps].data.configs.configs = dup_configs; + (*map)[*num_maps].data.configs.num_configs = num_configs; + (*num_maps)++; + + return 0; +} +EXPORT_SYMBOL_GPL(pinctrl_utils_add_map_configs); + +int pinctrl_utils_add_config(struct pinctrl_dev *pctldev, + unsigned long **configs, unsigned *num_configs, + unsigned long config) +{ + unsigned old_num = *num_configs; + unsigned new_num = old_num + 1; + unsigned long *new_configs; + + new_configs = krealloc(*configs, sizeof(*new_configs) * new_num, + GFP_KERNEL); + if (!new_configs) { + dev_err(pctldev->dev, "krealloc(configs) failed\n"); + return -ENOMEM; + } + + new_configs[old_num] = config; + + *configs = new_configs; + *num_configs = new_num; + + return 0; +} +EXPORT_SYMBOL_GPL(pinctrl_utils_add_config); + +void pinctrl_utils_dt_free_map(struct pinctrl_dev *pctldev, + struct pinctrl_map *map, unsigned num_maps) +{ + int i; + + for (i = 0; i < num_maps; i++) { + switch (map[i].type) { + case PIN_MAP_TYPE_CONFIGS_GROUP: + case PIN_MAP_TYPE_CONFIGS_PIN: + kfree(map[i].data.configs.configs); + break; + default: + break; + } + } + kfree(map); +} +EXPORT_SYMBOL_GPL(pinctrl_utils_dt_free_map); diff --git a/drivers/pinctrl/pinctrl-utils.h b/drivers/pinctrl/pinctrl-utils.h new file mode 100644 index 000000000000..d0ffe1ce200f --- /dev/null +++ b/drivers/pinctrl/pinctrl-utils.h @@ -0,0 +1,43 @@ +/* + * Utils functions to implement the pincontrol driver. + * + * Copyright (c) 2013, NVIDIA Corporation. + * + * Author: Laxman Dewangan <ldewangan@nvidia.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation version 2. + * + * This program is distributed "as is" WITHOUT ANY WARRANTY of any kind, + * whether express or implied; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + * 02111-1307, USA + */ +#ifndef __PINCTRL_UTILS_H__ +#define __PINCTRL_UTILS_H__ + +int pinctrl_utils_reserve_map(struct pinctrl_dev *pctldev, + struct pinctrl_map **map, unsigned *reserved_maps, + unsigned *num_maps, unsigned reserve); +int pinctrl_utils_add_map_mux(struct pinctrl_dev *pctldev, + struct pinctrl_map **map, unsigned *reserved_maps, + unsigned *num_maps, const char *group, + const char *function); +int pinctrl_utils_add_map_configs(struct pinctrl_dev *pctldev, + struct pinctrl_map **map, unsigned *reserved_maps, + unsigned *num_maps, const char *group, + unsigned long *configs, unsigned num_configs, + enum pinctrl_map_type type); +int pinctrl_utils_add_config(struct pinctrl_dev *pctldev, + unsigned long **configs, unsigned *num_configs, + unsigned long config); +void pinctrl_utils_dt_free_map(struct pinctrl_dev *pctldev, + struct pinctrl_map *map, unsigned num_maps); + +#endif /* __PINCTRL_UTILS_H__ */ diff --git a/drivers/power/reset/Kconfig b/drivers/power/reset/Kconfig index 349e9ae8090a..ee039dcead04 100644 --- a/drivers/power/reset/Kconfig +++ b/drivers/power/reset/Kconfig @@ -32,7 +32,8 @@ config POWER_RESET_RESTART user presses a key. u-boot then boots into Linux. config POWER_RESET_VEXPRESS - bool + bool "ARM Versatile Express power-off and reset driver" + depends on ARM || ARM64 depends on POWER_RESET help Power off and reset support for the ARM Ltd. Versatile diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c index a86d12326137..874dd261b248 100644 --- a/drivers/regulator/core.c +++ b/drivers/regulator/core.c @@ -1887,8 +1887,9 @@ int regulator_disable_deferred(struct regulator *regulator, int ms) rdev->deferred_disables++; mutex_unlock(&rdev->mutex); - ret = schedule_delayed_work(&rdev->disable_work, - msecs_to_jiffies(ms)); + ret = queue_delayed_work(system_power_efficient_wq, + &rdev->disable_work, + msecs_to_jiffies(ms)); if (ret < 0) return ret; else diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig index 5e3c02554d99..e92e1dcb6e3f 100644 --- a/drivers/thermal/Kconfig +++ b/drivers/thermal/Kconfig @@ -17,8 +17,30 @@ if THERMAL config THERMAL_HWMON bool + prompt "Expose thermal sensors as hwmon device" depends on HWMON=y || HWMON=THERMAL default y + help + In case a sensor is registered with the thermal + framework, this option will also register it + as a hwmon. The sensor will then have the common + hwmon sysfs interface. + + Say 'Y' here if you want all thermal sensors to + have hwmon sysfs interface too. + +config THERMAL_OF + bool + prompt "APIs to parse thermal data out of device tree" + depends on OF + default y + help + This options provides helpers to add the support to + read and parse thermal data definitions out of the + device tree blob. + + Say 'Y' here if you need to build thermal infrastructure + based on device tree. choice prompt "Default Thermal governor" @@ -69,6 +91,7 @@ config THERMAL_GOV_USER_SPACE config CPU_THERMAL bool "generic cpu cooling support" depends on CPU_FREQ + depends on THERMAL_OF select CPU_FREQ_TABLE help This implements the generic cpu cooling mechanism through frequency diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile index c054d410ac3f..71f8eef49c1c 100644 --- a/drivers/thermal/Makefile +++ b/drivers/thermal/Makefile @@ -5,6 +5,10 @@ obj-$(CONFIG_THERMAL) += thermal_sys.o thermal_sys-y += thermal_core.o +# interface to/from other layers providing sensors +thermal_sys-$(CONFIG_THERMAL_HWMON) += thermal_hwmon.o +thermal_sys-$(CONFIG_THERMAL_OF) += of-thermal.o + # governors thermal_sys-$(CONFIG_THERMAL_GOV_FAIR_SHARE) += fair_share.o thermal_sys-$(CONFIG_THERMAL_GOV_STEP_WISE) += step_wise.o diff --git a/drivers/thermal/cpu_cooling.c b/drivers/thermal/cpu_cooling.c index c94bf2e5de62..4246262c4bd2 100644 --- a/drivers/thermal/cpu_cooling.c +++ b/drivers/thermal/cpu_cooling.c @@ -167,13 +167,20 @@ static int get_property(unsigned int cpu, unsigned long input, continue; /* get the frequency order */ - if (freq != CPUFREQ_ENTRY_INVALID && descend != -1) + if (freq != CPUFREQ_ENTRY_INVALID && descend == -1) descend = !!(freq > table[i].frequency); freq = table[i].frequency; max_level++; } + /* No valid cpu frequency entry */ + if (max_level == 0) + return -EINVAL; + + /* max_level is an index, not a counter */ + max_level--; + /* get max level */ if (property == GET_MAXL) { *output = (unsigned int)max_level; @@ -181,7 +188,7 @@ static int get_property(unsigned int cpu, unsigned long input, } if (property == GET_FREQ) - level = descend ? input : (max_level - input - 1); + level = descend ? input : (max_level - input); for (i = 0, j = 0; table[i].frequency != CPUFREQ_TABLE_END; i++) { /* ignore invalid entry */ @@ -197,7 +204,7 @@ static int get_property(unsigned int cpu, unsigned long input, if (property == GET_LEVEL && (unsigned int)input == freq) { /* get level by frequency */ - *output = descend ? j : (max_level - j - 1); + *output = descend ? j : (max_level - j); return 0; } if (property == GET_FREQ && level == j) { @@ -322,6 +329,8 @@ static int cpufreq_thermal_notifier(struct notifier_block *nb, if (cpumask_test_cpu(policy->cpu, ¬ify_device->allowed_cpus)) max_freq = notify_device->cpufreq_val; + else + return 0; /* Never exceed user_policy.max */ if (max_freq > policy->user_policy.max) @@ -415,18 +424,21 @@ static struct notifier_block thermal_cpufreq_notifier_block = { }; /** - * cpufreq_cooling_register - function to create cpufreq cooling device. + * __cpufreq_cooling_register - helper function to create cpufreq cooling device + * @np: a valid struct device_node to the cooling device device tree node * @clip_cpus: cpumask of cpus where the frequency constraints will happen. * * This interface function registers the cpufreq cooling device with the name * "thermal-cpufreq-%x". This api can support multiple instances of cpufreq - * cooling devices. + * cooling devices. It also gives the opportunity to link the cooling device + * with a device tree node, in order to bind it via the thermal DT code. * * Return: a valid struct thermal_cooling_device pointer on success, * on failure, it returns a corresponding ERR_PTR(). */ -struct thermal_cooling_device * -cpufreq_cooling_register(const struct cpumask *clip_cpus) +static struct thermal_cooling_device * +__cpufreq_cooling_register(struct device_node *np, + const struct cpumask *clip_cpus) { struct thermal_cooling_device *cool_dev; struct cpufreq_cooling_device *cpufreq_dev = NULL; @@ -465,12 +477,12 @@ cpufreq_cooling_register(const struct cpumask *clip_cpus) snprintf(dev_name, sizeof(dev_name), "thermal-cpufreq-%d", cpufreq_dev->id); - cool_dev = thermal_cooling_device_register(dev_name, cpufreq_dev, - &cpufreq_cooling_ops); - if (!cool_dev) { + cool_dev = thermal_of_cooling_device_register(np, dev_name, cpufreq_dev, + &cpufreq_cooling_ops); + if (IS_ERR(cool_dev)) { release_idr(&cpufreq_idr, cpufreq_dev->id); kfree(cpufreq_dev); - return ERR_PTR(-EINVAL); + return cool_dev; } cpufreq_dev->cool_dev = cool_dev; cpufreq_dev->cpufreq_state = 0; @@ -486,9 +498,50 @@ cpufreq_cooling_register(const struct cpumask *clip_cpus) return cool_dev; } + +/** + * cpufreq_cooling_register - function to create cpufreq cooling device. + * @clip_cpus: cpumask of cpus where the frequency constraints will happen. + * + * This interface function registers the cpufreq cooling device with the name + * "thermal-cpufreq-%x". This api can support multiple instances of cpufreq + * cooling devices. + * + * Return: a valid struct thermal_cooling_device pointer on success, + * on failure, it returns a corresponding ERR_PTR(). + */ +struct thermal_cooling_device * +cpufreq_cooling_register(const struct cpumask *clip_cpus) +{ + return __cpufreq_cooling_register(NULL, clip_cpus); +} EXPORT_SYMBOL_GPL(cpufreq_cooling_register); /** + * of_cpufreq_cooling_register - function to create cpufreq cooling device. + * @np: a valid struct device_node to the cooling device device tree node + * @clip_cpus: cpumask of cpus where the frequency constraints will happen. + * + * This interface function registers the cpufreq cooling device with the name + * "thermal-cpufreq-%x". This api can support multiple instances of cpufreq + * cooling devices. Using this API, the cpufreq cooling device will be + * linked to the device tree node provided. + * + * Return: a valid struct thermal_cooling_device pointer on success, + * on failure, it returns a corresponding ERR_PTR(). + */ +struct thermal_cooling_device * +of_cpufreq_cooling_register(struct device_node *np, + const struct cpumask *clip_cpus) +{ + if (!np) + return ERR_PTR(-EINVAL); + + return __cpufreq_cooling_register(np, clip_cpus); +} +EXPORT_SYMBOL_GPL(of_cpufreq_cooling_register); + +/** * cpufreq_cooling_unregister - function to remove cpufreq cooling device. * @cdev: thermal cooling device pointer. * @@ -496,8 +549,12 @@ EXPORT_SYMBOL_GPL(cpufreq_cooling_register); */ void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev) { - struct cpufreq_cooling_device *cpufreq_dev = cdev->devdata; + struct cpufreq_cooling_device *cpufreq_dev; + + if (!cdev) + return; + cpufreq_dev = cdev->devdata; mutex_lock(&cooling_cpufreq_lock); cpufreq_dev_count--; diff --git a/drivers/thermal/of-thermal.c b/drivers/thermal/of-thermal.c new file mode 100644 index 000000000000..4b2b999b7611 --- /dev/null +++ b/drivers/thermal/of-thermal.c @@ -0,0 +1,850 @@ +/* + * of-thermal.c - Generic Thermal Management device tree support. + * + * Copyright (C) 2013 Texas Instruments + * Copyright (C) 2013 Eduardo Valentin <eduardo.valentin@ti.com> + * + * + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; version 2 of the License. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. + * + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + */ +#include <linux/thermal.h> +#include <linux/slab.h> +#include <linux/types.h> +#include <linux/of_device.h> +#include <linux/of_platform.h> +#include <linux/err.h> +#include <linux/export.h> +#include <linux/string.h> + +#include "thermal_core.h" + +/*** Private data structures to represent thermal device tree data ***/ + +/** + * struct __thermal_trip - representation of a point in temperature domain + * @np: pointer to struct device_node that this trip point was created from + * @temperature: temperature value in miliCelsius + * @hysteresis: relative hysteresis in miliCelsius + * @type: trip point type + */ + +struct __thermal_trip { + struct device_node *np; + unsigned long int temperature; + unsigned long int hysteresis; + enum thermal_trip_type type; +}; + +/** + * struct __thermal_bind_param - a match between trip and cooling device + * @cooling_device: a pointer to identify the referred cooling device + * @trip_id: the trip point index + * @usage: the percentage (from 0 to 100) of cooling contribution + * @min: minimum cooling state used at this trip point + * @max: maximum cooling state used at this trip point + */ + +struct __thermal_bind_params { + struct device_node *cooling_device; + unsigned int trip_id; + unsigned int usage; + unsigned long min; + unsigned long max; +}; + +/** + * struct __thermal_zone - internal representation of a thermal zone + * @mode: current thermal zone device mode (enabled/disabled) + * @passive_delay: polling interval while passive cooling is activated + * @polling_delay: zone polling interval + * @ntrips: number of trip points + * @trips: an array of trip points (0..ntrips - 1) + * @num_tbps: number of thermal bind params + * @tbps: an array of thermal bind params (0..num_tbps - 1) + * @sensor_data: sensor private data used while reading temperature and trend + * @get_temp: sensor callback to read temperature + * @get_trend: sensor callback to read temperature trend + */ + +struct __thermal_zone { + enum thermal_device_mode mode; + int passive_delay; + int polling_delay; + + /* trip data */ + int ntrips; + struct __thermal_trip *trips; + + /* cooling binding data */ + int num_tbps; + struct __thermal_bind_params *tbps; + + /* sensor interface */ + void *sensor_data; + int (*get_temp)(void *, long *); + int (*get_trend)(void *, long *); +}; + +/*** DT thermal zone device callbacks ***/ + +static int of_thermal_get_temp(struct thermal_zone_device *tz, + unsigned long *temp) +{ + struct __thermal_zone *data = tz->devdata; + + if (!data->get_temp) + return -EINVAL; + + return data->get_temp(data->sensor_data, temp); +} + +static int of_thermal_get_trend(struct thermal_zone_device *tz, int trip, + enum thermal_trend *trend) +{ + struct __thermal_zone *data = tz->devdata; + long dev_trend; + int r; + + if (!data->get_trend) + return -EINVAL; + + r = data->get_trend(data->sensor_data, &dev_trend); + if (r) + return r; + + /* TODO: These intervals might have some thresholds, but in core code */ + if (dev_trend > 0) + *trend = THERMAL_TREND_RAISING; + else if (dev_trend < 0) + *trend = THERMAL_TREND_DROPPING; + else + *trend = THERMAL_TREND_STABLE; + + return 0; +} + +static int of_thermal_bind(struct thermal_zone_device *thermal, + struct thermal_cooling_device *cdev) +{ + struct __thermal_zone *data = thermal->devdata; + int i; + + if (!data || IS_ERR(data)) + return -ENODEV; + + /* find where to bind */ + for (i = 0; i < data->num_tbps; i++) { + struct __thermal_bind_params *tbp = data->tbps + i; + + if (tbp->cooling_device == cdev->np) { + int ret; + + ret = thermal_zone_bind_cooling_device(thermal, + tbp->trip_id, cdev, + tbp->max, + tbp->min); + if (ret) + return ret; + } + } + + return 0; +} + +static int of_thermal_unbind(struct thermal_zone_device *thermal, + struct thermal_cooling_device *cdev) +{ + struct __thermal_zone *data = thermal->devdata; + int i; + + if (!data || IS_ERR(data)) + return -ENODEV; + + /* find where to unbind */ + for (i = 0; i < data->num_tbps; i++) { + struct __thermal_bind_params *tbp = data->tbps + i; + + if (tbp->cooling_device == cdev->np) { + int ret; + + ret = thermal_zone_unbind_cooling_device(thermal, + tbp->trip_id, cdev); + if (ret) + return ret; + } + } + + return 0; +} + +static int of_thermal_get_mode(struct thermal_zone_device *tz, + enum thermal_device_mode *mode) +{ + struct __thermal_zone *data = tz->devdata; + + *mode = data->mode; + + return 0; +} + +static int of_thermal_set_mode(struct thermal_zone_device *tz, + enum thermal_device_mode mode) +{ + struct __thermal_zone *data = tz->devdata; + + mutex_lock(&tz->lock); + + if (mode == THERMAL_DEVICE_ENABLED) + tz->polling_delay = data->polling_delay; + else + tz->polling_delay = 0; + + mutex_unlock(&tz->lock); + + data->mode = mode; + thermal_zone_device_update(tz); + + return 0; +} + +static int of_thermal_get_trip_type(struct thermal_zone_device *tz, int trip, + enum thermal_trip_type *type) +{ + struct __thermal_zone *data = tz->devdata; + + if (trip >= data->ntrips || trip < 0) + return -EDOM; + + *type = data->trips[trip].type; + + return 0; +} + +static int of_thermal_get_trip_temp(struct thermal_zone_device *tz, int trip, + unsigned long *temp) +{ + struct __thermal_zone *data = tz->devdata; + + if (trip >= data->ntrips || trip < 0) + return -EDOM; + + *temp = data->trips[trip].temperature; + + return 0; +} + +static int of_thermal_set_trip_temp(struct thermal_zone_device *tz, int trip, + unsigned long temp) +{ + struct __thermal_zone *data = tz->devdata; + + if (trip >= data->ntrips || trip < 0) + return -EDOM; + + /* thermal framework should take care of data->mask & (1 << trip) */ + data->trips[trip].temperature = temp; + + return 0; +} + +static int of_thermal_get_trip_hyst(struct thermal_zone_device *tz, int trip, + unsigned long *hyst) +{ + struct __thermal_zone *data = tz->devdata; + + if (trip >= data->ntrips || trip < 0) + return -EDOM; + + *hyst = data->trips[trip].hysteresis; + + return 0; +} + +static int of_thermal_set_trip_hyst(struct thermal_zone_device *tz, int trip, + unsigned long hyst) +{ + struct __thermal_zone *data = tz->devdata; + + if (trip >= data->ntrips || trip < 0) + return -EDOM; + + /* thermal framework should take care of data->mask & (1 << trip) */ + data->trips[trip].hysteresis = hyst; + + return 0; +} + +static int of_thermal_get_crit_temp(struct thermal_zone_device *tz, + unsigned long *temp) +{ + struct __thermal_zone *data = tz->devdata; + int i; + + for (i = 0; i < data->ntrips; i++) + if (data->trips[i].type == THERMAL_TRIP_CRITICAL) { + *temp = data->trips[i].temperature; + return 0; + } + + return -EINVAL; +} + +static struct thermal_zone_device_ops of_thermal_ops = { + .get_mode = of_thermal_get_mode, + .set_mode = of_thermal_set_mode, + + .get_trip_type = of_thermal_get_trip_type, + .get_trip_temp = of_thermal_get_trip_temp, + .set_trip_temp = of_thermal_set_trip_temp, + .get_trip_hyst = of_thermal_get_trip_hyst, + .set_trip_hyst = of_thermal_set_trip_hyst, + .get_crit_temp = of_thermal_get_crit_temp, + + .bind = of_thermal_bind, + .unbind = of_thermal_unbind, +}; + +/*** sensor API ***/ + +static struct thermal_zone_device * +thermal_zone_of_add_sensor(struct device_node *zone, + struct device_node *sensor, void *data, + int (*get_temp)(void *, long *), + int (*get_trend)(void *, long *)) +{ + struct thermal_zone_device *tzd; + struct __thermal_zone *tz; + + tzd = thermal_zone_get_zone_by_name(zone->name); + if (IS_ERR(tzd)) + return ERR_PTR(-EPROBE_DEFER); + + tz = tzd->devdata; + + mutex_lock(&tzd->lock); + tz->get_temp = get_temp; + tz->get_trend = get_trend; + tz->sensor_data = data; + + tzd->ops->get_temp = of_thermal_get_temp; + tzd->ops->get_trend = of_thermal_get_trend; + mutex_unlock(&tzd->lock); + + return tzd; +} + +/** + * thermal_zone_of_sensor_register - registers a sensor to a DT thermal zone + * @dev: a valid struct device pointer of a sensor device. Must contain + * a valid .of_node, for the sensor node. + * @sensor_id: a sensor identifier, in case the sensor IP has more + * than one sensors + * @data: a private pointer (owned by the caller) that will be passed + * back, when a temperature reading is needed. + * @get_temp: a pointer to a function that reads the sensor temperature. + * @get_trend: a pointer to a function that reads the sensor temperature trend. + * + * This function will search the list of thermal zones described in device + * tree and look for the zone that refer to the sensor device pointed by + * @dev->of_node as temperature providers. For the zone pointing to the + * sensor node, the sensor will be added to the DT thermal zone device. + * + * The thermal zone temperature is provided by the @get_temp function + * pointer. When called, it will have the private pointer @data back. + * + * The thermal zone temperature trend is provided by the @get_trend function + * pointer. When called, it will have the private pointer @data back. + * + * TODO: + * 01 - This function must enqueue the new sensor instead of using + * it as the only source of temperature values. + * + * 02 - There must be a way to match the sensor with all thermal zones + * that refer to it. + * + * Return: On success returns a valid struct thermal_zone_device, + * otherwise, it returns a corresponding ERR_PTR(). Caller must + * check the return value with help of IS_ERR() helper. + */ +struct thermal_zone_device * +thermal_zone_of_sensor_register(struct device *dev, int sensor_id, + void *data, int (*get_temp)(void *, long *), + int (*get_trend)(void *, long *)) +{ + struct device_node *np, *child, *sensor_np; + + np = of_find_node_by_name(NULL, "thermal-zones"); + if (!np) + return ERR_PTR(-ENODEV); + + if (!dev || !dev->of_node) + return ERR_PTR(-EINVAL); + + sensor_np = dev->of_node; + + for_each_child_of_node(np, child) { + struct of_phandle_args sensor_specs; + int ret, id; + + /* For now, thermal framework supports only 1 sensor per zone */ + ret = of_parse_phandle_with_args(child, "thermal-sensors", + "#thermal-sensor-cells", + 0, &sensor_specs); + if (ret) + continue; + + if (sensor_specs.args_count >= 1) { + id = sensor_specs.args[0]; + WARN(sensor_specs.args_count > 1, + "%s: too many cells in sensor specifier %d\n", + sensor_specs.np->name, sensor_specs.args_count); + } else { + id = 0; + } + + if (sensor_specs.np == sensor_np && id == sensor_id) { + of_node_put(np); + return thermal_zone_of_add_sensor(child, sensor_np, + data, + get_temp, + get_trend); + } + } + of_node_put(np); + + return ERR_PTR(-ENODEV); +} +EXPORT_SYMBOL_GPL(thermal_zone_of_sensor_register); + +/** + * thermal_zone_of_sensor_unregister - unregisters a sensor from a DT thermal zone + * @dev: a valid struct device pointer of a sensor device. Must contain + * a valid .of_node, for the sensor node. + * @tzd: a pointer to struct thermal_zone_device where the sensor is registered. + * + * This function removes the sensor callbacks and private data from the + * thermal zone device registered with thermal_zone_of_sensor_register() + * API. It will also silent the zone by remove the .get_temp() and .get_trend() + * thermal zone device callbacks. + * + * TODO: When the support to several sensors per zone is added, this + * function must search the sensor list based on @dev parameter. + * + */ +void thermal_zone_of_sensor_unregister(struct device *dev, + struct thermal_zone_device *tzd) +{ + struct __thermal_zone *tz; + + if (!dev || !tzd || !tzd->devdata) + return; + + tz = tzd->devdata; + + /* no __thermal_zone, nothing to be done */ + if (!tz) + return; + + mutex_lock(&tzd->lock); + tzd->ops->get_temp = NULL; + tzd->ops->get_trend = NULL; + + tz->get_temp = NULL; + tz->get_trend = NULL; + tz->sensor_data = NULL; + mutex_unlock(&tzd->lock); +} +EXPORT_SYMBOL_GPL(thermal_zone_of_sensor_unregister); + +/*** functions parsing device tree nodes ***/ + +/** + * thermal_of_populate_bind_params - parse and fill cooling map data + * @np: DT node containing a cooling-map node + * @__tbp: data structure to be filled with cooling map info + * @trips: array of thermal zone trip points + * @ntrips: number of trip points inside trips. + * + * This function parses a cooling-map type of node represented by + * @np parameter and fills the read data into @__tbp data structure. + * It needs the already parsed array of trip points of the thermal zone + * in consideration. + * + * Return: 0 on success, proper error code otherwise + */ +static int thermal_of_populate_bind_params(struct device_node *np, + struct __thermal_bind_params *__tbp, + struct __thermal_trip *trips, + int ntrips) +{ + struct of_phandle_args cooling_spec; + struct device_node *trip; + int ret, i; + u32 prop; + + /* Default weight. Usage is optional */ + __tbp->usage = 0; + ret = of_property_read_u32(np, "contribution", &prop); + if (ret == 0) + __tbp->usage = prop; + + trip = of_parse_phandle(np, "trip", 0); + if (!trip) { + pr_err("missing trip property\n"); + return -ENODEV; + } + + /* match using device_node */ + for (i = 0; i < ntrips; i++) + if (trip == trips[i].np) { + __tbp->trip_id = i; + break; + } + + if (i == ntrips) { + ret = -ENODEV; + goto end; + } + + ret = of_parse_phandle_with_args(np, "cooling-device", "#cooling-cells", + 0, &cooling_spec); + if (ret < 0) { + pr_err("missing cooling_device property\n"); + goto end; + } + __tbp->cooling_device = cooling_spec.np; + if (cooling_spec.args_count >= 2) { /* at least min and max */ + __tbp->min = cooling_spec.args[0]; + __tbp->max = cooling_spec.args[1]; + } else { + pr_err("wrong reference to cooling device, missing limits\n"); + } + +end: + of_node_put(trip); + + return ret; +} + +/** + * It maps 'enum thermal_trip_type' found in include/linux/thermal.h + * into the device tree binding of 'trip', property type. + */ +static const char * const trip_types[] = { + [THERMAL_TRIP_ACTIVE] = "active", + [THERMAL_TRIP_PASSIVE] = "passive", + [THERMAL_TRIP_HOT] = "hot", + [THERMAL_TRIP_CRITICAL] = "critical", +}; + +/** + * thermal_of_get_trip_type - Get phy mode for given device_node + * @np: Pointer to the given device_node + * @type: Pointer to resulting trip type + * + * The function gets trip type string from property 'type', + * and store its index in trip_types table in @type, + * + * Return: 0 on success, or errno in error case. + */ +static int thermal_of_get_trip_type(struct device_node *np, + enum thermal_trip_type *type) +{ + const char *t; + int err, i; + + err = of_property_read_string(np, "type", &t); + if (err < 0) + return err; + + for (i = 0; i < ARRAY_SIZE(trip_types); i++) + if (!strcasecmp(t, trip_types[i])) { + *type = i; + return 0; + } + + return -ENODEV; +} + +/** + * thermal_of_populate_trip - parse and fill one trip point data + * @np: DT node containing a trip point node + * @trip: trip point data structure to be filled up + * + * This function parses a trip point type of node represented by + * @np parameter and fills the read data into @trip data structure. + * + * Return: 0 on success, proper error code otherwise + */ +static int thermal_of_populate_trip(struct device_node *np, + struct __thermal_trip *trip) +{ + int prop; + int ret; + + ret = of_property_read_u32(np, "temperature", &prop); + if (ret < 0) { + pr_err("missing temperature property\n"); + return ret; + } + trip->temperature = prop; + + ret = of_property_read_u32(np, "hysteresis", &prop); + if (ret < 0) { + pr_err("missing hysteresis property\n"); + return ret; + } + trip->hysteresis = prop; + + ret = thermal_of_get_trip_type(np, &trip->type); + if (ret < 0) { + pr_err("wrong trip type property\n"); + return ret; + } + + /* Required for cooling map matching */ + trip->np = np; + + return 0; +} + +/** + * thermal_of_build_thermal_zone - parse and fill one thermal zone data + * @np: DT node containing a thermal zone node + * + * This function parses a thermal zone type of node represented by + * @np parameter and fills the read data into a __thermal_zone data structure + * and return this pointer. + * + * TODO: Missing properties to parse: thermal-sensor-names and coefficients + * + * Return: On success returns a valid struct __thermal_zone, + * otherwise, it returns a corresponding ERR_PTR(). Caller must + * check the return value with help of IS_ERR() helper. + */ +static struct __thermal_zone * +thermal_of_build_thermal_zone(struct device_node *np) +{ + struct device_node *child = NULL, *gchild; + struct __thermal_zone *tz; + int ret, i; + u32 prop; + + if (!np) { + pr_err("no thermal zone np\n"); + return ERR_PTR(-EINVAL); + } + + tz = kzalloc(sizeof(*tz), GFP_KERNEL); + if (!tz) + return ERR_PTR(-ENOMEM); + + ret = of_property_read_u32(np, "polling-delay-passive", &prop); + if (ret < 0) { + pr_err("missing polling-delay-passive property\n"); + goto free_tz; + } + tz->passive_delay = prop; + + ret = of_property_read_u32(np, "polling-delay", &prop); + if (ret < 0) { + pr_err("missing polling-delay property\n"); + goto free_tz; + } + tz->polling_delay = prop; + + /* trips */ + child = of_get_child_by_name(np, "trips"); + + /* No trips provided */ + if (!child) + goto finish; + + tz->ntrips = of_get_child_count(child); + if (tz->ntrips == 0) /* must have at least one child */ + goto finish; + + tz->trips = kzalloc(tz->ntrips * sizeof(*tz->trips), GFP_KERNEL); + if (!tz->trips) { + ret = -ENOMEM; + goto free_tz; + } + + i = 0; + for_each_child_of_node(child, gchild) { + ret = thermal_of_populate_trip(gchild, &tz->trips[i++]); + if (ret) + goto free_trips; + } + + of_node_put(child); + + /* cooling-maps */ + child = of_get_child_by_name(np, "cooling-maps"); + + /* cooling-maps not provided */ + if (!child) + goto finish; + + tz->num_tbps = of_get_child_count(child); + if (tz->num_tbps == 0) + goto finish; + + tz->tbps = kzalloc(tz->num_tbps * sizeof(*tz->tbps), GFP_KERNEL); + if (!tz->tbps) { + ret = -ENOMEM; + goto free_trips; + } + + i = 0; + for_each_child_of_node(child, gchild) { + ret = thermal_of_populate_bind_params(gchild, &tz->tbps[i++], + tz->trips, tz->ntrips); + if (ret) + goto free_tbps; + } + +finish: + of_node_put(child); + tz->mode = THERMAL_DEVICE_DISABLED; + + return tz; + +free_tbps: + kfree(tz->tbps); +free_trips: + kfree(tz->trips); +free_tz: + kfree(tz); + of_node_put(child); + + return ERR_PTR(ret); +} + +static inline void of_thermal_free_zone(struct __thermal_zone *tz) +{ + kfree(tz->tbps); + kfree(tz->trips); + kfree(tz); +} + +/** + * of_parse_thermal_zones - parse device tree thermal data + * + * Initialization function that can be called by machine initialization + * code to parse thermal data and populate the thermal framework + * with hardware thermal zones info. This function only parses thermal zones. + * Cooling devices and sensor devices nodes are supposed to be parsed + * by their respective drivers. + * + * Return: 0 on success, proper error code otherwise + * + */ +int __init of_parse_thermal_zones(void) +{ + struct device_node *np, *child; + struct __thermal_zone *tz; + struct thermal_zone_device_ops *ops; + + np = of_find_node_by_name(NULL, "thermal-zones"); + if (!np) { + pr_debug("unable to find thermal zones\n"); + return 0; /* Run successfully on systems without thermal DT */ + } + + for_each_child_of_node(np, child) { + struct thermal_zone_device *zone; + struct thermal_zone_params *tzp; + + tz = thermal_of_build_thermal_zone(child); + if (IS_ERR(tz)) { + pr_err("failed to build thermal zone %s: %ld\n", + child->name, + PTR_ERR(tz)); + continue; + } + + ops = kmemdup(&of_thermal_ops, sizeof(*ops), GFP_KERNEL); + if (!ops) + goto exit_free; + + tzp = kzalloc(sizeof(*tzp), GFP_KERNEL); + if (!tzp) { + kfree(ops); + goto exit_free; + } + + /* No hwmon because there might be hwmon drivers registering */ + tzp->no_hwmon = true; + + zone = thermal_zone_device_register(child->name, tz->ntrips, + 0, tz, + ops, tzp, + tz->passive_delay, + tz->polling_delay); + if (IS_ERR(zone)) { + pr_err("Failed to build %s zone %ld\n", child->name, + PTR_ERR(zone)); + kfree(tzp); + kfree(ops); + of_thermal_free_zone(tz); + /* attempting to build remaining zones still */ + } + } + + return 0; + +exit_free: + of_thermal_free_zone(tz); + + /* no memory available, so free what we have built */ + of_thermal_destroy_zones(); + + return -ENOMEM; +} + +/** + * of_thermal_destroy_zones - remove all zones parsed and allocated resources + * + * Finds all zones parsed and added to the thermal framework and remove them + * from the system, together with their resources. + * + */ +void of_thermal_destroy_zones(void) +{ + struct device_node *np, *child; + + np = of_find_node_by_name(NULL, "thermal-zones"); + if (!np) { + pr_err("unable to find thermal zones\n"); + return; + } + + for_each_child_of_node(np, child) { + struct thermal_zone_device *zone; + + zone = thermal_zone_get_zone_by_name(child->name); + if (IS_ERR(zone)) + continue; + + thermal_zone_device_unregister(zone); + kfree(zone->tzp); + kfree(zone->ops); + of_thermal_free_zone(zone->devdata); + } +} diff --git a/drivers/thermal/step_wise.c b/drivers/thermal/step_wise.c index 4d4ddae1a991..c83a3097ef90 100644 --- a/drivers/thermal/step_wise.c +++ b/drivers/thermal/step_wise.c @@ -53,6 +53,7 @@ static unsigned long get_target_state(struct thermal_instance *instance, unsigned long cur_state; cdev->ops->get_cur_state(cdev, &cur_state); + dev_dbg(&cdev->device, "cur_state=%ld\n", cur_state); switch (trend) { case THERMAL_TREND_RAISING: @@ -124,6 +125,9 @@ static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) if (tz->temperature >= trip_temp) throttle = true; + dev_dbg(&tz->device, "Trip%d[type=%d,temp=%ld]:trend=%d,throttle=%d\n", + trip, trip_type, trip_temp, trend, throttle); + mutex_lock(&tz->lock); list_for_each_entry(instance, &tz->thermal_instances, tz_node) { @@ -132,6 +136,8 @@ static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) old_target = instance->target; instance->target = get_target_state(instance, trend, throttle); + dev_dbg(&instance->cdev->device, "old_target=%d, target=%d\n", + old_target, (int)instance->target); /* Activate a passive thermal instance */ if (old_target == THERMAL_NO_TARGET && diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c index d755440791b7..edc0cb88f1d0 100644 --- a/drivers/thermal/thermal_core.c +++ b/drivers/thermal/thermal_core.c @@ -33,10 +33,13 @@ #include <linux/idr.h> #include <linux/thermal.h> #include <linux/reboot.h> +#include <linux/string.h> +#include <linux/of.h> #include <net/netlink.h> #include <net/genetlink.h> #include "thermal_core.h" +#include "thermal_hwmon.h" MODULE_AUTHOR("Zhang Rui"); MODULE_DESCRIPTION("Generic thermal management sysfs support"); @@ -53,10 +56,15 @@ static LIST_HEAD(thermal_governor_list); static DEFINE_MUTEX(thermal_list_lock); static DEFINE_MUTEX(thermal_governor_lock); +static struct thermal_governor *def_governor; + static struct thermal_governor *__find_governor(const char *name) { struct thermal_governor *pos; + if (!name || !name[0]) + return def_governor; + list_for_each_entry(pos, &thermal_governor_list, governor_list) if (!strnicmp(name, pos->name, THERMAL_NAME_LENGTH)) return pos; @@ -79,17 +87,23 @@ int thermal_register_governor(struct thermal_governor *governor) if (__find_governor(governor->name) == NULL) { err = 0; list_add(&governor->governor_list, &thermal_governor_list); + if (!def_governor && !strncmp(governor->name, + DEFAULT_THERMAL_GOVERNOR, THERMAL_NAME_LENGTH)) + def_governor = governor; } mutex_lock(&thermal_list_lock); list_for_each_entry(pos, &thermal_tz_list, node) { + /* + * only thermal zones with specified tz->tzp->governor_name + * may run with tz->govenor unset + */ if (pos->governor) continue; - if (pos->tzp) - name = pos->tzp->governor_name; - else - name = DEFAULT_THERMAL_GOVERNOR; + + name = pos->tzp->governor_name; + if (!strnicmp(name, governor->name, THERMAL_NAME_LENGTH)) pos->governor = governor; } @@ -235,10 +249,11 @@ static void bind_cdev(struct thermal_cooling_device *cdev) if (!pos->tzp && !pos->ops->bind) continue; - if (!pos->tzp && pos->ops->bind) { + if (pos->ops->bind) { ret = pos->ops->bind(pos, cdev); if (ret) print_bind_err_msg(pos, cdev, ret); + continue; } tzp = pos->tzp; @@ -269,8 +284,8 @@ static void bind_tz(struct thermal_zone_device *tz) mutex_lock(&thermal_list_lock); - /* If there is no platform data, try to use ops->bind */ - if (!tzp && tz->ops->bind) { + /* If there is ops->bind, try to use ops->bind */ + if (tz->ops->bind) { list_for_each_entry(pos, &thermal_cdev_list, node) { ret = tz->ops->bind(tz, pos); if (ret) @@ -326,8 +341,8 @@ static void monitor_thermal_zone(struct thermal_zone_device *tz) static void handle_non_critical_trips(struct thermal_zone_device *tz, int trip, enum thermal_trip_type trip_type) { - if (tz->governor) - tz->governor->throttle(tz, trip); + tz->governor ? tz->governor->throttle(tz, trip) : + def_governor->throttle(tz, trip); } static void handle_critical_trips(struct thermal_zone_device *tz, @@ -388,7 +403,7 @@ int thermal_zone_get_temp(struct thermal_zone_device *tz, unsigned long *temp) enum thermal_trip_type type; #endif - if (!tz || IS_ERR(tz)) + if (!tz || IS_ERR(tz) || !tz->ops->get_temp) goto exit; mutex_lock(&tz->lock); @@ -435,12 +450,18 @@ static void update_temperature(struct thermal_zone_device *tz) tz->last_temperature = tz->temperature; tz->temperature = temp; mutex_unlock(&tz->lock); + + dev_dbg(&tz->device, "last_temperature=%d, current_temperature=%d\n", + tz->last_temperature, tz->temperature); } void thermal_zone_device_update(struct thermal_zone_device *tz) { int count; + if (!tz->ops->get_temp) + return; + update_temperature(tz); for (count = 0; count < tz->trips; count++) @@ -756,6 +777,9 @@ emul_temp_store(struct device *dev, struct device_attribute *attr, ret = tz->ops->set_emul_temp(tz, temperature); } + if (!ret) + thermal_zone_device_update(tz); + return ret ? ret : count; } static DEVICE_ATTR(emul_temp, S_IWUSR, NULL, emul_temp_store); @@ -854,260 +878,6 @@ thermal_cooling_device_trip_point_show(struct device *dev, /* Device management */ -#if defined(CONFIG_THERMAL_HWMON) - -/* hwmon sys I/F */ -#include <linux/hwmon.h> - -/* thermal zone devices with the same type share one hwmon device */ -struct thermal_hwmon_device { - char type[THERMAL_NAME_LENGTH]; - struct device *device; - int count; - struct list_head tz_list; - struct list_head node; -}; - -struct thermal_hwmon_attr { - struct device_attribute attr; - char name[16]; -}; - -/* one temperature input for each thermal zone */ -struct thermal_hwmon_temp { - struct list_head hwmon_node; - struct thermal_zone_device *tz; - struct thermal_hwmon_attr temp_input; /* hwmon sys attr */ - struct thermal_hwmon_attr temp_crit; /* hwmon sys attr */ -}; - -static LIST_HEAD(thermal_hwmon_list); - -static ssize_t -name_show(struct device *dev, struct device_attribute *attr, char *buf) -{ - struct thermal_hwmon_device *hwmon = dev_get_drvdata(dev); - return sprintf(buf, "%s\n", hwmon->type); -} -static DEVICE_ATTR(name, 0444, name_show, NULL); - -static ssize_t -temp_input_show(struct device *dev, struct device_attribute *attr, char *buf) -{ - long temperature; - int ret; - struct thermal_hwmon_attr *hwmon_attr - = container_of(attr, struct thermal_hwmon_attr, attr); - struct thermal_hwmon_temp *temp - = container_of(hwmon_attr, struct thermal_hwmon_temp, - temp_input); - struct thermal_zone_device *tz = temp->tz; - - ret = thermal_zone_get_temp(tz, &temperature); - - if (ret) - return ret; - - return sprintf(buf, "%ld\n", temperature); -} - -static ssize_t -temp_crit_show(struct device *dev, struct device_attribute *attr, - char *buf) -{ - struct thermal_hwmon_attr *hwmon_attr - = container_of(attr, struct thermal_hwmon_attr, attr); - struct thermal_hwmon_temp *temp - = container_of(hwmon_attr, struct thermal_hwmon_temp, - temp_crit); - struct thermal_zone_device *tz = temp->tz; - long temperature; - int ret; - - ret = tz->ops->get_trip_temp(tz, 0, &temperature); - if (ret) - return ret; - - return sprintf(buf, "%ld\n", temperature); -} - - -static struct thermal_hwmon_device * -thermal_hwmon_lookup_by_type(const struct thermal_zone_device *tz) -{ - struct thermal_hwmon_device *hwmon; - - mutex_lock(&thermal_list_lock); - list_for_each_entry(hwmon, &thermal_hwmon_list, node) - if (!strcmp(hwmon->type, tz->type)) { - mutex_unlock(&thermal_list_lock); - return hwmon; - } - mutex_unlock(&thermal_list_lock); - - return NULL; -} - -/* Find the temperature input matching a given thermal zone */ -static struct thermal_hwmon_temp * -thermal_hwmon_lookup_temp(const struct thermal_hwmon_device *hwmon, - const struct thermal_zone_device *tz) -{ - struct thermal_hwmon_temp *temp; - - mutex_lock(&thermal_list_lock); - list_for_each_entry(temp, &hwmon->tz_list, hwmon_node) - if (temp->tz == tz) { - mutex_unlock(&thermal_list_lock); - return temp; - } - mutex_unlock(&thermal_list_lock); - - return NULL; -} - -static int -thermal_add_hwmon_sysfs(struct thermal_zone_device *tz) -{ - struct thermal_hwmon_device *hwmon; - struct thermal_hwmon_temp *temp; - int new_hwmon_device = 1; - int result; - - hwmon = thermal_hwmon_lookup_by_type(tz); - if (hwmon) { - new_hwmon_device = 0; - goto register_sys_interface; - } - - hwmon = kzalloc(sizeof(struct thermal_hwmon_device), GFP_KERNEL); - if (!hwmon) - return -ENOMEM; - - INIT_LIST_HEAD(&hwmon->tz_list); - strlcpy(hwmon->type, tz->type, THERMAL_NAME_LENGTH); - hwmon->device = hwmon_device_register(NULL); - if (IS_ERR(hwmon->device)) { - result = PTR_ERR(hwmon->device); - goto free_mem; - } - dev_set_drvdata(hwmon->device, hwmon); - result = device_create_file(hwmon->device, &dev_attr_name); - if (result) - goto free_mem; - - register_sys_interface: - temp = kzalloc(sizeof(struct thermal_hwmon_temp), GFP_KERNEL); - if (!temp) { - result = -ENOMEM; - goto unregister_name; - } - - temp->tz = tz; - hwmon->count++; - - snprintf(temp->temp_input.name, sizeof(temp->temp_input.name), - "temp%d_input", hwmon->count); - temp->temp_input.attr.attr.name = temp->temp_input.name; - temp->temp_input.attr.attr.mode = 0444; - temp->temp_input.attr.show = temp_input_show; - sysfs_attr_init(&temp->temp_input.attr.attr); - result = device_create_file(hwmon->device, &temp->temp_input.attr); - if (result) - goto free_temp_mem; - - if (tz->ops->get_crit_temp) { - unsigned long temperature; - if (!tz->ops->get_crit_temp(tz, &temperature)) { - snprintf(temp->temp_crit.name, - sizeof(temp->temp_crit.name), - "temp%d_crit", hwmon->count); - temp->temp_crit.attr.attr.name = temp->temp_crit.name; - temp->temp_crit.attr.attr.mode = 0444; - temp->temp_crit.attr.show = temp_crit_show; - sysfs_attr_init(&temp->temp_crit.attr.attr); - result = device_create_file(hwmon->device, - &temp->temp_crit.attr); - if (result) - goto unregister_input; - } - } - - mutex_lock(&thermal_list_lock); - if (new_hwmon_device) - list_add_tail(&hwmon->node, &thermal_hwmon_list); - list_add_tail(&temp->hwmon_node, &hwmon->tz_list); - mutex_unlock(&thermal_list_lock); - - return 0; - - unregister_input: - device_remove_file(hwmon->device, &temp->temp_input.attr); - free_temp_mem: - kfree(temp); - unregister_name: - if (new_hwmon_device) { - device_remove_file(hwmon->device, &dev_attr_name); - hwmon_device_unregister(hwmon->device); - } - free_mem: - if (new_hwmon_device) - kfree(hwmon); - - return result; -} - -static void -thermal_remove_hwmon_sysfs(struct thermal_zone_device *tz) -{ - struct thermal_hwmon_device *hwmon; - struct thermal_hwmon_temp *temp; - - hwmon = thermal_hwmon_lookup_by_type(tz); - if (unlikely(!hwmon)) { - /* Should never happen... */ - dev_dbg(&tz->device, "hwmon device lookup failed!\n"); - return; - } - - temp = thermal_hwmon_lookup_temp(hwmon, tz); - if (unlikely(!temp)) { - /* Should never happen... */ - dev_dbg(&tz->device, "temperature input lookup failed!\n"); - return; - } - - device_remove_file(hwmon->device, &temp->temp_input.attr); - if (tz->ops->get_crit_temp) - device_remove_file(hwmon->device, &temp->temp_crit.attr); - - mutex_lock(&thermal_list_lock); - list_del(&temp->hwmon_node); - kfree(temp); - if (!list_empty(&hwmon->tz_list)) { - mutex_unlock(&thermal_list_lock); - return; - } - list_del(&hwmon->node); - mutex_unlock(&thermal_list_lock); - - device_remove_file(hwmon->device, &dev_attr_name); - hwmon_device_unregister(hwmon->device); - kfree(hwmon); -} -#else -static int -thermal_add_hwmon_sysfs(struct thermal_zone_device *tz) -{ - return 0; -} - -static void -thermal_remove_hwmon_sysfs(struct thermal_zone_device *tz) -{ -} -#endif - /** * thermal_zone_bind_cooling_device() - bind a cooling device to a thermal zone * @tz: pointer to struct thermal_zone_device @@ -1287,7 +1057,8 @@ static struct class thermal_class = { }; /** - * thermal_cooling_device_register() - register a new thermal cooling device + * __thermal_cooling_device_register() - register a new thermal cooling device + * @np: a pointer to a device tree node. * @type: the thermal cooling device type. * @devdata: device private data. * @ops: standard thermal cooling devices callbacks. @@ -1295,13 +1066,16 @@ static struct class thermal_class = { * This interface function adds a new thermal cooling device (fan/processor/...) * to /sys/class/thermal/ folder as cooling_device[0-*]. It tries to bind itself * to all the thermal zone devices registered at the same time. + * It also gives the opportunity to link the cooling device to a device tree + * node, so that it can be bound to a thermal zone created out of device tree. * * Return: a pointer to the created struct thermal_cooling_device or an * ERR_PTR. Caller must check return value with IS_ERR*() helpers. */ -struct thermal_cooling_device * -thermal_cooling_device_register(char *type, void *devdata, - const struct thermal_cooling_device_ops *ops) +static struct thermal_cooling_device * +__thermal_cooling_device_register(struct device_node *np, + char *type, void *devdata, + const struct thermal_cooling_device_ops *ops) { struct thermal_cooling_device *cdev; int result; @@ -1326,8 +1100,9 @@ thermal_cooling_device_register(char *type, void *devdata, strlcpy(cdev->type, type ? : "", sizeof(cdev->type)); mutex_init(&cdev->lock); INIT_LIST_HEAD(&cdev->thermal_instances); + cdev->np = np; cdev->ops = ops; - cdev->updated = true; + cdev->updated = false; cdev->device.class = &thermal_class; cdev->devdata = devdata; dev_set_name(&cdev->device, "cooling_device%d", cdev->id); @@ -1368,9 +1143,53 @@ unregister: device_unregister(&cdev->device); return ERR_PTR(result); } + +/** + * thermal_cooling_device_register() - register a new thermal cooling device + * @type: the thermal cooling device type. + * @devdata: device private data. + * @ops: standard thermal cooling devices callbacks. + * + * This interface function adds a new thermal cooling device (fan/processor/...) + * to /sys/class/thermal/ folder as cooling_device[0-*]. It tries to bind itself + * to all the thermal zone devices registered at the same time. + * + * Return: a pointer to the created struct thermal_cooling_device or an + * ERR_PTR. Caller must check return value with IS_ERR*() helpers. + */ +struct thermal_cooling_device * +thermal_cooling_device_register(char *type, void *devdata, + const struct thermal_cooling_device_ops *ops) +{ + return __thermal_cooling_device_register(NULL, type, devdata, ops); +} EXPORT_SYMBOL_GPL(thermal_cooling_device_register); /** + * thermal_of_cooling_device_register() - register an OF thermal cooling device + * @np: a pointer to a device tree node. + * @type: the thermal cooling device type. + * @devdata: device private data. + * @ops: standard thermal cooling devices callbacks. + * + * This function will register a cooling device with device tree node reference. + * This interface function adds a new thermal cooling device (fan/processor/...) + * to /sys/class/thermal/ folder as cooling_device[0-*]. It tries to bind itself + * to all the thermal zone devices registered at the same time. + * + * Return: a pointer to the created struct thermal_cooling_device or an + * ERR_PTR. Caller must check return value with IS_ERR*() helpers. + */ +struct thermal_cooling_device * +thermal_of_cooling_device_register(struct device_node *np, + char *type, void *devdata, + const struct thermal_cooling_device_ops *ops) +{ + return __thermal_cooling_device_register(np, type, devdata, ops); +} +EXPORT_SYMBOL_GPL(thermal_of_cooling_device_register); + +/** * thermal_cooling_device_unregister - removes the registered thermal cooling device * @cdev: the thermal cooling device to remove. * @@ -1442,6 +1261,8 @@ void thermal_cdev_update(struct thermal_cooling_device *cdev) mutex_lock(&cdev->lock); /* Make sure cdev enters the deepest cooling state */ list_for_each_entry(instance, &cdev->thermal_instances, cdev_node) { + dev_dbg(&cdev->device, "zone%d->target=%lu\n", + instance->tz->id, instance->target); if (instance->target == THERMAL_NO_TARGET) continue; if (instance->target > target) @@ -1450,6 +1271,7 @@ void thermal_cdev_update(struct thermal_cooling_device *cdev) mutex_unlock(&cdev->lock); cdev->ops->set_cur_state(cdev, target); cdev->updated = true; + dev_dbg(&cdev->device, "set to state %lu\n", target); } EXPORT_SYMBOL(thermal_cdev_update); @@ -1605,7 +1427,7 @@ static void remove_trip_attrs(struct thermal_zone_device *tz) */ struct thermal_zone_device *thermal_zone_device_register(const char *type, int trips, int mask, void *devdata, - const struct thermal_zone_device_ops *ops, + struct thermal_zone_device_ops *ops, const struct thermal_zone_params *tzp, int passive_delay, int polling_delay) { @@ -1621,7 +1443,7 @@ struct thermal_zone_device *thermal_zone_device_register(const char *type, if (trips > THERMAL_MAX_TRIPS || trips < 0 || mask >> trips) return ERR_PTR(-EINVAL); - if (!ops || !ops->get_temp) + if (!ops) return ERR_PTR(-EINVAL); if (trips > 0 && !ops->get_trip_type) @@ -1706,13 +1528,15 @@ struct thermal_zone_device *thermal_zone_device_register(const char *type, if (tz->tzp) tz->governor = __find_governor(tz->tzp->governor_name); else - tz->governor = __find_governor(DEFAULT_THERMAL_GOVERNOR); + tz->governor = def_governor; mutex_unlock(&thermal_governor_lock); - result = thermal_add_hwmon_sysfs(tz); - if (result) - goto unregister; + if (!tz->tzp || !tz->tzp->no_hwmon) { + result = thermal_add_hwmon_sysfs(tz); + if (result) + goto unregister; + } mutex_lock(&thermal_list_lock); list_add_tail(&tz->node, &thermal_tz_list); @@ -1723,6 +1547,9 @@ struct thermal_zone_device *thermal_zone_device_register(const char *type, INIT_DELAYED_WORK(&(tz->poll_queue), thermal_zone_device_check); + if (!tz->ops->get_temp) + thermal_zone_device_set_polling(tz, 0); + thermal_zone_device_update(tz); if (!result) @@ -1980,8 +1807,14 @@ static int __init thermal_init(void) if (result) goto unregister_class; + result = of_parse_thermal_zones(); + if (result) + goto exit_netlink; + return 0; +exit_netlink: + genetlink_exit(); unregister_governors: thermal_unregister_governors(); unregister_class: @@ -1997,6 +1830,7 @@ error: static void __exit thermal_exit(void) { + of_thermal_destroy_zones(); genetlink_exit(); class_unregister(&thermal_class); thermal_unregister_governors(); diff --git a/drivers/thermal/thermal_core.h b/drivers/thermal/thermal_core.h index 7cf2f6626251..3db339fb636f 100644 --- a/drivers/thermal/thermal_core.h +++ b/drivers/thermal/thermal_core.h @@ -77,4 +77,13 @@ static inline int thermal_gov_user_space_register(void) { return 0; } static inline void thermal_gov_user_space_unregister(void) {} #endif /* CONFIG_THERMAL_GOV_USER_SPACE */ +/* device tree support */ +#ifdef CONFIG_THERMAL_OF +int of_parse_thermal_zones(void); +void of_thermal_destroy_zones(void); +#else +static inline int of_parse_thermal_zones(void) { return 0; } +static inline void of_thermal_destroy_zones(void) { } +#endif + #endif /* __THERMAL_CORE_H__ */ diff --git a/drivers/thermal/thermal_hwmon.c b/drivers/thermal/thermal_hwmon.c new file mode 100644 index 000000000000..fdb07199d9c2 --- /dev/null +++ b/drivers/thermal/thermal_hwmon.c @@ -0,0 +1,269 @@ +/* + * thermal_hwmon.c - Generic Thermal Management hwmon support. + * + * Code based on Intel thermal_core.c. Copyrights of the original code: + * Copyright (C) 2008 Intel Corp + * Copyright (C) 2008 Zhang Rui <rui.zhang@intel.com> + * Copyright (C) 2008 Sujith Thomas <sujith.thomas@intel.com> + * + * Copyright (C) 2013 Texas Instruments + * Copyright (C) 2013 Eduardo Valentin <eduardo.valentin@ti.com> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; version 2 of the License. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. + * + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + */ +#include <linux/hwmon.h> +#include <linux/thermal.h> +#include <linux/slab.h> +#include <linux/err.h> +#include "thermal_hwmon.h" + +/* hwmon sys I/F */ +/* thermal zone devices with the same type share one hwmon device */ +struct thermal_hwmon_device { + char type[THERMAL_NAME_LENGTH]; + struct device *device; + int count; + struct list_head tz_list; + struct list_head node; +}; + +struct thermal_hwmon_attr { + struct device_attribute attr; + char name[16]; +}; + +/* one temperature input for each thermal zone */ +struct thermal_hwmon_temp { + struct list_head hwmon_node; + struct thermal_zone_device *tz; + struct thermal_hwmon_attr temp_input; /* hwmon sys attr */ + struct thermal_hwmon_attr temp_crit; /* hwmon sys attr */ +}; + +static LIST_HEAD(thermal_hwmon_list); + +static DEFINE_MUTEX(thermal_hwmon_list_lock); + +static ssize_t +name_show(struct device *dev, struct device_attribute *attr, char *buf) +{ + struct thermal_hwmon_device *hwmon = dev_get_drvdata(dev); + return sprintf(buf, "%s\n", hwmon->type); +} +static DEVICE_ATTR(name, 0444, name_show, NULL); + +static ssize_t +temp_input_show(struct device *dev, struct device_attribute *attr, char *buf) +{ + long temperature; + int ret; + struct thermal_hwmon_attr *hwmon_attr + = container_of(attr, struct thermal_hwmon_attr, attr); + struct thermal_hwmon_temp *temp + = container_of(hwmon_attr, struct thermal_hwmon_temp, + temp_input); + struct thermal_zone_device *tz = temp->tz; + + ret = thermal_zone_get_temp(tz, &temperature); + + if (ret) + return ret; + + return sprintf(buf, "%ld\n", temperature); +} + +static ssize_t +temp_crit_show(struct device *dev, struct device_attribute *attr, char *buf) +{ + struct thermal_hwmon_attr *hwmon_attr + = container_of(attr, struct thermal_hwmon_attr, attr); + struct thermal_hwmon_temp *temp + = container_of(hwmon_attr, struct thermal_hwmon_temp, + temp_crit); + struct thermal_zone_device *tz = temp->tz; + long temperature; + int ret; + + ret = tz->ops->get_trip_temp(tz, 0, &temperature); + if (ret) + return ret; + + return sprintf(buf, "%ld\n", temperature); +} + + +static struct thermal_hwmon_device * +thermal_hwmon_lookup_by_type(const struct thermal_zone_device *tz) +{ + struct thermal_hwmon_device *hwmon; + + mutex_lock(&thermal_hwmon_list_lock); + list_for_each_entry(hwmon, &thermal_hwmon_list, node) + if (!strcmp(hwmon->type, tz->type)) { + mutex_unlock(&thermal_hwmon_list_lock); + return hwmon; + } + mutex_unlock(&thermal_hwmon_list_lock); + + return NULL; +} + +/* Find the temperature input matching a given thermal zone */ +static struct thermal_hwmon_temp * +thermal_hwmon_lookup_temp(const struct thermal_hwmon_device *hwmon, + const struct thermal_zone_device *tz) +{ + struct thermal_hwmon_temp *temp; + + mutex_lock(&thermal_hwmon_list_lock); + list_for_each_entry(temp, &hwmon->tz_list, hwmon_node) + if (temp->tz == tz) { + mutex_unlock(&thermal_hwmon_list_lock); + return temp; + } + mutex_unlock(&thermal_hwmon_list_lock); + + return NULL; +} + +int thermal_add_hwmon_sysfs(struct thermal_zone_device *tz) +{ + struct thermal_hwmon_device *hwmon; + struct thermal_hwmon_temp *temp; + int new_hwmon_device = 1; + int result; + + hwmon = thermal_hwmon_lookup_by_type(tz); + if (hwmon) { + new_hwmon_device = 0; + goto register_sys_interface; + } + + hwmon = kzalloc(sizeof(*hwmon), GFP_KERNEL); + if (!hwmon) + return -ENOMEM; + + INIT_LIST_HEAD(&hwmon->tz_list); + strlcpy(hwmon->type, tz->type, THERMAL_NAME_LENGTH); + hwmon->device = hwmon_device_register(NULL); + if (IS_ERR(hwmon->device)) { + result = PTR_ERR(hwmon->device); + goto free_mem; + } + dev_set_drvdata(hwmon->device, hwmon); + result = device_create_file(hwmon->device, &dev_attr_name); + if (result) + goto free_mem; + + register_sys_interface: + temp = kzalloc(sizeof(*temp), GFP_KERNEL); + if (!temp) { + result = -ENOMEM; + goto unregister_name; + } + + temp->tz = tz; + hwmon->count++; + + snprintf(temp->temp_input.name, sizeof(temp->temp_input.name), + "temp%d_input", hwmon->count); + temp->temp_input.attr.attr.name = temp->temp_input.name; + temp->temp_input.attr.attr.mode = 0444; + temp->temp_input.attr.show = temp_input_show; + sysfs_attr_init(&temp->temp_input.attr.attr); + result = device_create_file(hwmon->device, &temp->temp_input.attr); + if (result) + goto free_temp_mem; + + if (tz->ops->get_crit_temp) { + unsigned long temperature; + if (!tz->ops->get_crit_temp(tz, &temperature)) { + snprintf(temp->temp_crit.name, + sizeof(temp->temp_crit.name), + "temp%d_crit", hwmon->count); + temp->temp_crit.attr.attr.name = temp->temp_crit.name; + temp->temp_crit.attr.attr.mode = 0444; + temp->temp_crit.attr.show = temp_crit_show; + sysfs_attr_init(&temp->temp_crit.attr.attr); + result = device_create_file(hwmon->device, + &temp->temp_crit.attr); + if (result) + goto unregister_input; + } + } + + mutex_lock(&thermal_hwmon_list_lock); + if (new_hwmon_device) + list_add_tail(&hwmon->node, &thermal_hwmon_list); + list_add_tail(&temp->hwmon_node, &hwmon->tz_list); + mutex_unlock(&thermal_hwmon_list_lock); + + return 0; + + unregister_input: + device_remove_file(hwmon->device, &temp->temp_input.attr); + free_temp_mem: + kfree(temp); + unregister_name: + if (new_hwmon_device) { + device_remove_file(hwmon->device, &dev_attr_name); + hwmon_device_unregister(hwmon->device); + } + free_mem: + if (new_hwmon_device) + kfree(hwmon); + + return result; +} + +void thermal_remove_hwmon_sysfs(struct thermal_zone_device *tz) +{ + struct thermal_hwmon_device *hwmon; + struct thermal_hwmon_temp *temp; + + hwmon = thermal_hwmon_lookup_by_type(tz); + if (unlikely(!hwmon)) { + /* Should never happen... */ + dev_dbg(&tz->device, "hwmon device lookup failed!\n"); + return; + } + + temp = thermal_hwmon_lookup_temp(hwmon, tz); + if (unlikely(!temp)) { + /* Should never happen... */ + dev_dbg(&tz->device, "temperature input lookup failed!\n"); + return; + } + + device_remove_file(hwmon->device, &temp->temp_input.attr); + if (tz->ops->get_crit_temp) + device_remove_file(hwmon->device, &temp->temp_crit.attr); + + mutex_lock(&thermal_hwmon_list_lock); + list_del(&temp->hwmon_node); + kfree(temp); + if (!list_empty(&hwmon->tz_list)) { + mutex_unlock(&thermal_hwmon_list_lock); + return; + } + list_del(&hwmon->node); + mutex_unlock(&thermal_hwmon_list_lock); + + device_remove_file(hwmon->device, &dev_attr_name); + hwmon_device_unregister(hwmon->device); + kfree(hwmon); +} diff --git a/drivers/thermal/thermal_hwmon.h b/drivers/thermal/thermal_hwmon.h new file mode 100644 index 000000000000..c798fdb2ae43 --- /dev/null +++ b/drivers/thermal/thermal_hwmon.h @@ -0,0 +1,49 @@ +/* + * thermal_hwmon.h - Generic Thermal Management hwmon support. + * + * Code based on Intel thermal_core.c. Copyrights of the original code: + * Copyright (C) 2008 Intel Corp + * Copyright (C) 2008 Zhang Rui <rui.zhang@intel.com> + * Copyright (C) 2008 Sujith Thomas <sujith.thomas@intel.com> + * + * Copyright (C) 2013 Texas Instruments + * Copyright (C) 2013 Eduardo Valentin <eduardo.valentin@ti.com> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; version 2 of the License. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. + * + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + */ +#ifndef __THERMAL_HWMON_H__ +#define __THERMAL_HWMON_H__ + +#include <linux/thermal.h> + +#ifdef CONFIG_THERMAL_HWMON +int thermal_add_hwmon_sysfs(struct thermal_zone_device *tz); +void thermal_remove_hwmon_sysfs(struct thermal_zone_device *tz); +#else +static int +thermal_add_hwmon_sysfs(struct thermal_zone_device *tz) +{ + return 0; +} + +static void +thermal_remove_hwmon_sysfs(struct thermal_zone_device *tz) +{ +} +#endif + +#endif /* __THERMAL_HWMON_H__ */ diff --git a/drivers/tty/serial/8250/8250_early.c b/drivers/tty/serial/8250/8250_early.c index 721904f8efa9..4858b8a99d3b 100644 --- a/drivers/tty/serial/8250/8250_early.c +++ b/drivers/tty/serial/8250/8250_early.c @@ -35,18 +35,8 @@ #include <linux/serial_8250.h> #include <asm/io.h> #include <asm/serial.h> -#ifdef CONFIG_FIX_EARLYCON_MEM -#include <asm/pgtable.h> -#include <asm/fixmap.h> -#endif -struct early_serial8250_device { - struct uart_port port; - char options[16]; /* e.g., 115200n8 */ - unsigned int baud; -}; - -static struct early_serial8250_device early_device; +static struct earlycon_device *early_device; unsigned int __weak __init serial8250_early_in(struct uart_port *port, int offset) { @@ -100,7 +90,7 @@ static void __init serial_putc(struct uart_port *port, int c) static void __init early_serial8250_write(struct console *console, const char *s, unsigned int count) { - struct uart_port *port = &early_device.port; + struct uart_port *port = &early_device->port; unsigned int ier; /* Save the IER and disable interrupts */ @@ -129,7 +119,7 @@ static unsigned int __init probe_baud(struct uart_port *port) return (port->uartclk / 16) / quot; } -static void __init init_port(struct early_serial8250_device *device) +static void __init init_port(struct earlycon_device *device) { struct uart_port *port = &device->port; unsigned int divisor; @@ -148,127 +138,45 @@ static void __init init_port(struct early_serial8250_device *device) serial8250_early_out(port, UART_LCR, c & ~UART_LCR_DLAB); } -static int __init parse_options(struct early_serial8250_device *device, - char *options) +static int __init early_serial8250_setup(struct earlycon_device *device, + const char *options) { - struct uart_port *port = &device->port; - int mmio, mmio32, length; - - if (!options) - return -ENODEV; - - port->uartclk = BASE_BAUD * 16; - - mmio = !strncmp(options, "mmio,", 5); - mmio32 = !strncmp(options, "mmio32,", 7); - if (mmio || mmio32) { - port->iotype = (mmio ? UPIO_MEM : UPIO_MEM32); - port->mapbase = simple_strtoul(options + (mmio ? 5 : 7), - &options, 0); - if (mmio32) - port->regshift = 2; -#ifdef CONFIG_FIX_EARLYCON_MEM - set_fixmap_nocache(FIX_EARLYCON_MEM_BASE, - port->mapbase & PAGE_MASK); - port->membase = - (void __iomem *)__fix_to_virt(FIX_EARLYCON_MEM_BASE); - port->membase += port->mapbase & ~PAGE_MASK; -#else - port->membase = ioremap_nocache(port->mapbase, 64); - if (!port->membase) { - printk(KERN_ERR "%s: Couldn't ioremap 0x%llx\n", - __func__, - (unsigned long long) port->mapbase); - return -ENOMEM; - } -#endif - } else if (!strncmp(options, "io,", 3)) { - port->iotype = UPIO_PORT; - port->iobase = simple_strtoul(options + 3, &options, 0); - mmio = 0; - } else - return -EINVAL; + if (!(device->port.membase || device->port.iobase)) + return 0; - options = strchr(options, ','); - if (options) { - options++; - device->baud = simple_strtoul(options, NULL, 0); - length = min(strcspn(options, " "), sizeof(device->options)); - strlcpy(device->options, options, length); - } else { - device->baud = probe_baud(port); + if (!device->baud) { + device->baud = probe_baud(&device->port); snprintf(device->options, sizeof(device->options), "%u", - device->baud); + device->baud); } - if (mmio || mmio32) - printk(KERN_INFO - "Early serial console at MMIO%s 0x%llx (options '%s')\n", - mmio32 ? "32" : "", - (unsigned long long)port->mapbase, - device->options); - else - printk(KERN_INFO - "Early serial console at I/O port 0x%lx (options '%s')\n", - port->iobase, - device->options); - - return 0; -} - -static struct console early_serial8250_console __initdata = { - .name = "uart", - .write = early_serial8250_write, - .flags = CON_PRINTBUFFER | CON_BOOT, - .index = -1, -}; - -static int __init early_serial8250_setup(char *options) -{ - struct early_serial8250_device *device = &early_device; - int err; - - if (device->port.membase || device->port.iobase) - return 0; - - err = parse_options(device, options); - if (err < 0) - return err; - init_port(device); + + early_device = device; + device->con->write = early_serial8250_write; return 0; } +EARLYCON_DECLARE(uart8250, early_serial8250_setup); +EARLYCON_DECLARE(uart, early_serial8250_setup); int __init setup_early_serial8250_console(char *cmdline) { - char *options; - int err; + char match[] = "uart8250"; - options = strstr(cmdline, "uart8250,"); - if (!options) { - options = strstr(cmdline, "uart,"); - if (!options) - return 0; - } - - options = strchr(cmdline, ',') + 1; - err = early_serial8250_setup(options); - if (err < 0) - return err; - - register_console(&early_serial8250_console); + if (cmdline && cmdline[4] == ',') + match[4] = '\0'; - return 0; + return setup_earlycon(cmdline, match, early_serial8250_setup); } int serial8250_find_port_for_earlycon(void) { - struct early_serial8250_device *device = &early_device; - struct uart_port *port = &device->port; + struct earlycon_device *device = early_device; + struct uart_port *port = device ? &device->port : NULL; int line; int ret; - if (!device->port.membase && !device->port.iobase) + if (!port || (!port->membase && !port->iobase)) return -ENODEV; line = serial8250_find_port(port); @@ -283,5 +191,3 @@ int serial8250_find_port_for_earlycon(void) return ret; } - -early_param("earlycon", setup_early_serial8250_console); diff --git a/drivers/tty/serial/8250/Kconfig b/drivers/tty/serial/8250/Kconfig index 80fe91e64a52..902c4bf15f0b 100644 --- a/drivers/tty/serial/8250/Kconfig +++ b/drivers/tty/serial/8250/Kconfig @@ -62,6 +62,7 @@ config SERIAL_8250_CONSOLE bool "Console on 8250/16550 and compatible serial port" depends on SERIAL_8250=y select SERIAL_CORE_CONSOLE + select SERIAL_EARLYCON ---help--- If you say Y here, it will be possible to use a serial port as the system console (the system console is the device which receives all diff --git a/drivers/tty/serial/Kconfig b/drivers/tty/serial/Kconfig index 7e7006fd404e..1ad9aadc0055 100644 --- a/drivers/tty/serial/Kconfig +++ b/drivers/tty/serial/Kconfig @@ -7,6 +7,13 @@ if TTY menu "Serial drivers" depends on HAS_IOMEM && GENERIC_HARDIRQS +config SERIAL_EARLYCON + bool + help + Support for early consoles with the earlycon parameter. This enables + the console before standard serial driver is probed. The console is + enabled when early_param is processed. + source "drivers/tty/serial/8250/Kconfig" comment "Non-8250 serial port support" diff --git a/drivers/tty/serial/Makefile b/drivers/tty/serial/Makefile index eedfec40e3dd..e51c680d6044 100644 --- a/drivers/tty/serial/Makefile +++ b/drivers/tty/serial/Makefile @@ -5,6 +5,8 @@ obj-$(CONFIG_SERIAL_CORE) += serial_core.o obj-$(CONFIG_SERIAL_21285) += 21285.o +obj-$(CONFIG_SERIAL_EARLYCON) += earlycon.o + # These Sparc drivers have to appear before others such as 8250 # which share ttySx minor node space. Otherwise console device # names change and other unplesantries. diff --git a/drivers/tty/serial/altera_uart.c b/drivers/tty/serial/altera_uart.c index 13471dd95793..d983370eb05b 100644 --- a/drivers/tty/serial/altera_uart.c +++ b/drivers/tty/serial/altera_uart.c @@ -231,7 +231,9 @@ static void altera_uart_rx_chars(struct altera_uart *pp) flag); } + spin_unlock(&port->lock); tty_flip_buffer_push(&port->state->port); + spin_lock(&port->lock); } static void altera_uart_tx_chars(struct altera_uart *pp) diff --git a/drivers/tty/serial/earlycon.c b/drivers/tty/serial/earlycon.c new file mode 100644 index 000000000000..4f27f788ac6f --- /dev/null +++ b/drivers/tty/serial/earlycon.c @@ -0,0 +1,152 @@ +/* + * Copyright (C) 2014 Linaro Ltd. + * Author: Rob Herring <robh@kernel.org> + * + * Based on 8250 earlycon: + * (c) Copyright 2004 Hewlett-Packard Development Company, L.P. + * Bjorn Helgaas <bjorn.helgaas@hp.com> + * + * This program is free software: you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ +#include <linux/console.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/io.h> +#include <linux/serial_core.h> + +#ifdef CONFIG_FIX_EARLYCON_MEM +#include <asm/fixmap.h> +#endif + +#include <asm/serial.h> + +static struct console early_con = { + .name = "uart", /* 8250 console switch requires this name */ + .flags = CON_PRINTBUFFER | CON_BOOT, + .index = -1, +}; + +static struct earlycon_device early_console_dev = { + .con = &early_con, +}; + +static void __iomem * __init earlycon_map(unsigned long paddr, size_t size) +{ + void __iomem *base; +#ifdef CONFIG_FIX_EARLYCON_MEM + set_fixmap_io(FIX_EARLYCON_MEM_BASE, paddr & PAGE_MASK); + base = (void __iomem *)__fix_to_virt(FIX_EARLYCON_MEM_BASE); + base += paddr & ~PAGE_MASK; +#else + base = ioremap(paddr, size); +#endif + if (!base) + pr_err("%s: Couldn't map 0x%llx\n", __func__, + (unsigned long long)paddr); + + return base; +} + +static int __init parse_options(struct earlycon_device *device, + char *options) +{ + struct uart_port *port = &device->port; + int mmio, mmio32, length, ret; + unsigned long addr; + + if (!options) + return -ENODEV; + + mmio = !strncmp(options, "mmio,", 5); + mmio32 = !strncmp(options, "mmio32,", 7); + if (mmio || mmio32) { + port->iotype = (mmio ? UPIO_MEM : UPIO_MEM32); + options += mmio ? 5 : 7; + ret = kstrtoul(options, 0, &addr); + if (ret) + return ret; + port->mapbase = addr; + if (mmio32) + port->regshift = 2; + } else if (!strncmp(options, "io,", 3)) { + port->iotype = UPIO_PORT; + options += 3; + ret = kstrtoul(options, 0, &addr); + if (ret) + return ret; + port->iobase = addr; + mmio = 0; + } else if (!strncmp(options, "0x", 2)) { + port->iotype = UPIO_MEM; + ret = kstrtoul(options, 0, &addr); + if (ret) + return ret; + port->mapbase = addr; + } else { + return -EINVAL; + } + + port->uartclk = BASE_BAUD * 16; + + options = strchr(options, ','); + if (options) { + options++; + ret = kstrtouint(options, 0, &device->baud); + if (ret) + return ret; + length = min(strcspn(options, " ") + 1, + (size_t)(sizeof(device->options))); + strlcpy(device->options, options, length); + } + + if (mmio || mmio32) + pr_info("Early serial console at MMIO%s 0x%llx (options '%s')\n", + mmio32 ? "32" : "", + (unsigned long long)port->mapbase, + device->options); + else + pr_info("Early serial console at I/O port 0x%lx (options '%s')\n", + port->iobase, + device->options); + + return 0; +} + +int __init setup_earlycon(char *buf, const char *match, + int (*setup)(struct earlycon_device *, const char *)) +{ + int err; + size_t len; + struct uart_port *port = &early_console_dev.port; + + if (!buf || !match || !setup) + return 0; + + len = strlen(match); + if (strncmp(buf, match, len)) + return 0; + if (buf[len] && (buf[len] != ',')) + return 0; + + buf += len + 1; + + err = parse_options(&early_console_dev, buf); + /* On parsing error, pass the options buf to the setup function */ + if (!err) + buf = NULL; + + if (port->mapbase) + port->membase = earlycon_map(port->mapbase, 64); + + early_console_dev.con->data = &early_console_dev; + err = setup(&early_console_dev, buf); + if (err < 0) + return err; + if (!early_console_dev.con->write) + return -ENODEV; + + register_console(early_console_dev.con); + return 0; +} diff --git a/drivers/tty/serial/samsung.c b/drivers/tty/serial/samsung.c index 0c8a9fa2be6c..aaa0df1f2fc6 100644 --- a/drivers/tty/serial/samsung.c +++ b/drivers/tty/serial/samsung.c @@ -249,6 +249,8 @@ s3c24xx_serial_rx_chars(int irq, void *dev_id) ufcon |= S3C2410_UFCON_RESETRX; wr_regl(port, S3C2410_UFCON, ufcon); rx_enabled(port) = 1; + spin_unlock_irqrestore(&port->lock, + flags); goto out; } continue; @@ -297,10 +299,11 @@ s3c24xx_serial_rx_chars(int irq, void *dev_id) ignore_char: continue; } + + spin_unlock_irqrestore(&port->lock, flags); tty_flip_buffer_push(&port->state->port); out: - spin_unlock_irqrestore(&port->lock, flags); return IRQ_HANDLED; } diff --git a/drivers/usb/Kconfig b/drivers/usb/Kconfig index 92e1dc94ecc8..73f62caa8609 100644 --- a/drivers/usb/Kconfig +++ b/drivers/usb/Kconfig @@ -2,59 +2,15 @@ # USB device configuration # -# many non-PCI SOC chips embed OHCI +# These are unused now, remove them once they are no longer selected config USB_ARCH_HAS_OHCI - boolean - # ARM: - default y if SA1111 - default y if ARCH_OMAP - default y if ARCH_S3C24XX - default y if PXA27x - default y if PXA3xx - default y if ARCH_EP93XX - default y if ARCH_AT91 - default y if MFD_TC6393XB - default y if ARCH_W90X900 - default y if ARCH_DAVINCI_DA8XX - default y if ARCH_CNS3XXX - default y if PLAT_SPEAR - default y if ARCH_EXYNOS - # PPC: - default y if STB03xxx - default y if PPC_MPC52xx - # MIPS: - default y if MIPS_ALCHEMY - default y if MACH_JZ4740 - # more: - default PCI - -# some non-PCI hcds implement EHCI + bool + config USB_ARCH_HAS_EHCI - boolean - default y if FSL_SOC - default y if PPC_MPC512x - default y if ARCH_IXP4XX - default y if ARCH_W90X900 - default y if ARCH_AT91 - default y if ARCH_MXC - default y if ARCH_MXS - default y if ARCH_OMAP3 - default y if ARCH_CNS3XXX - default y if ARCH_VT8500 - default y if PLAT_SPEAR - default y if PLAT_S5P - default y if ARCH_MSM - default y if MICROBLAZE - default y if SPARC_LEON - default y if ARCH_MMP - default y if MACH_LOONGSON1 - default y if PLAT_ORION - default PCI - -# some non-PCI HCDs implement xHCI + bool + config USB_ARCH_HAS_XHCI - boolean - default PCI + bool menuconfig USB_SUPPORT bool "USB support" @@ -71,19 +27,8 @@ config USB_COMMON default y depends on USB || USB_GADGET -# Host-side USB depends on having a host controller -# NOTE: dummy_hcd is always an option, but it's ignored here ... -# NOTE: SL-811 option should be board-specific ... config USB_ARCH_HAS_HCD - boolean - default y if USB_ARCH_HAS_OHCI - default y if USB_ARCH_HAS_EHCI - default y if USB_ARCH_HAS_XHCI - default y if PCMCIA && !M32R # sl811_cs - default y if ARM # SL-811 - default y if BLACKFIN # SL-811 - default y if SUPERH # r8a66597-hcd - default PCI + def_bool y # ARM SA1111 chips have a non-PCI based "OHCI-compatible" USB host interface. config USB diff --git a/drivers/usb/host/Kconfig b/drivers/usb/host/Kconfig index 344d5e2f87d7..8bc8e066c881 100644 --- a/drivers/usb/host/Kconfig +++ b/drivers/usb/host/Kconfig @@ -17,7 +17,6 @@ config USB_C67X00_HCD config USB_XHCI_HCD tristate "xHCI HCD (USB 3.0) support" - depends on USB_ARCH_HAS_XHCI ---help--- The eXtensible Host Controller Interface (xHCI) is standard for USB 3.0 "SuperSpeed" host controller hardware. @@ -43,7 +42,6 @@ endif # USB_XHCI_HCD config USB_EHCI_HCD tristate "EHCI HCD (USB 2.0) support" - depends on USB_ARCH_HAS_EHCI ---help--- The Enhanced Host Controller Interface (EHCI) is standard for USB 2.0 "high speed" (480 Mbit/sec, 60 Mbyte/sec) host controller hardware. @@ -177,6 +175,13 @@ config USB_EHCI_HCD_SPEAR Enables support for the on-chip EHCI controller on ST SPEAr chips. +config USB_EHCI_HCD_SYNOPSYS + tristate "Support for Synopsys Host-AHB USB 2.0 controller" + depends on USB_EHCI_HCD + ---help--- + Enable support for onchip USB controllers based on DesignWare USB 2.0 + Host-AHB Controller IP from Synopsys. + config USB_EHCI_HCD_AT91 tristate "Support for Atmel on-chip EHCI USB controller" depends on USB_EHCI_HCD && ARCH_AT91 @@ -347,7 +352,6 @@ config USB_ISP1362_HCD config USB_OHCI_HCD tristate "OHCI HCD support" - depends on USB_ARCH_HAS_OHCI select ISP1301_OMAP if MACH_OMAP_H2 || MACH_OMAP_H3 depends on USB_ISP1301 || !ARCH_LPC32XX ---help--- diff --git a/drivers/usb/host/Makefile b/drivers/usb/host/Makefile index 4fb73c156d72..70df7c8c6bcd 100644 --- a/drivers/usb/host/Makefile +++ b/drivers/usb/host/Makefile @@ -31,6 +31,7 @@ obj-$(CONFIG_USB_EHCI_HCD_OMAP) += ehci-omap.o obj-$(CONFIG_USB_EHCI_HCD_ORION) += ehci-orion.o obj-$(CONFIG_USB_EHCI_HCD_SPEAR) += ehci-spear.o obj-$(CONFIG_USB_EHCI_S5P) += ehci-s5p.o +obj-$(CONFIG_USB_EHCI_HCD_SYNOPSYS) += ehci-h20ahb.o obj-$(CONFIG_USB_EHCI_HCD_AT91) += ehci-atmel.o obj-$(CONFIG_USB_EHCI_MSM) += ehci-msm.o diff --git a/drivers/usb/host/ehci-h20ahb.c b/drivers/usb/host/ehci-h20ahb.c new file mode 100644 index 000000000000..7724bab1828b --- /dev/null +++ b/drivers/usb/host/ehci-h20ahb.c @@ -0,0 +1,341 @@ +/* + * Copyright (C) 2007-2013 Texas Instruments, Inc. + * Author: Vikram Pandita <vikram.pandita@ti.com> + * Author: Anand Gadiyar <gadiyar@ti.com> + * Author: Keshava Munegowda <keshava_mgowda@ti.com> + * Author: Roger Quadros <rogerq@ti.com> + * + * Copyright (C) 2009 Nokia Corporation + * Contact: Felipe Balbi <felipe.balbi@nokia.com> + * + * Based on ehci-omap.c - driver for USBHOST on OMAP3/4 processors + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file COPYING in the main directory of this archive + * for more details. + * + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/io.h> +#include <linux/platform_device.h> +#include <linux/slab.h> +#include <linux/usb/ulpi.h> +#include <linux/pm_runtime.h> +#include <linux/gpio.h> +#include <linux/clk.h> +#include <linux/usb.h> +#include <linux/usb/hcd.h> +#include <linux/of.h> +#include <linux/dma-mapping.h> + +#include "ehci.h" + +#define H20AHB_HS_USB_PORTS 1 + +/* EHCI Synopsys-specific Register Set */ +#define EHCI_INSNREG04 (0xA0) +#define EHCI_INSNREG04_DISABLE_UNSUSPEND (1 << 5) +#define EHCI_INSNREG05_ULPI (0xA4) +#define EHCI_INSNREG05_ULPI_CONTROL_SHIFT 31 +#define EHCI_INSNREG05_ULPI_PORTSEL_SHIFT 24 +#define EHCI_INSNREG05_ULPI_OPSEL_SHIFT 22 +#define EHCI_INSNREG05_ULPI_REGADD_SHIFT 16 +#define EHCI_INSNREG05_ULPI_EXTREGADD_SHIFT 8 +#define EHCI_INSNREG05_ULPI_WRDATA_SHIFT 0 + +#define DRIVER_DESC "H20AHB-EHCI Host Controller driver" + +static const char hcd_name[] = "ehci-h20ahb"; + +/*-------------------------------------------------------------------------*/ + +struct h20ahb_hcd { + struct usb_phy *phy[H20AHB_HS_USB_PORTS]; /* one PHY for each port */ + int nports; +}; + +static inline void ehci_write(void __iomem *base, u32 reg, u32 val) +{ + writel_relaxed(val, base + reg); +} + +static inline u32 ehci_read(void __iomem *base, u32 reg) +{ + return readl_relaxed(base + reg); +} + +/* configure so an HC device and id are always provided */ +/* always called with process context; sleeping is OK */ + +static struct hc_driver __read_mostly ehci_h20ahb_hc_driver; + +static const struct ehci_driver_overrides ehci_h20ahb_overrides __initdata = { + .extra_priv_size = sizeof(struct h20ahb_hcd), +}; + +static int ehci_h20ahb_phy_read(struct usb_phy *x, u32 reg) +{ + u32 val = (1 << EHCI_INSNREG05_ULPI_CONTROL_SHIFT) | + (1 << EHCI_INSNREG05_ULPI_PORTSEL_SHIFT) | + (3 << EHCI_INSNREG05_ULPI_OPSEL_SHIFT) | + (reg << EHCI_INSNREG05_ULPI_REGADD_SHIFT); + ehci_write(x->io_priv, 0, val); + while ((val = ehci_read(x->io_priv, 0)) & + (1 << EHCI_INSNREG05_ULPI_CONTROL_SHIFT)); + return val & 0xff; +} + +static int ehci_h20ahb_phy_write(struct usb_phy *x, u32 val, u32 reg) +{ + u32 v = (1 << EHCI_INSNREG05_ULPI_CONTROL_SHIFT) | + (1 << EHCI_INSNREG05_ULPI_PORTSEL_SHIFT) | + (2 << EHCI_INSNREG05_ULPI_OPSEL_SHIFT) | + (reg << EHCI_INSNREG05_ULPI_REGADD_SHIFT) | + (val & 0xff); + ehci_write(x->io_priv, 0, v); + while ((v = ehci_read(x->io_priv, 0)) & + (1 << EHCI_INSNREG05_ULPI_CONTROL_SHIFT)); + return 0; +} + +static struct usb_phy_io_ops ehci_h20ahb_phy_io_ops = { + .read = ehci_h20ahb_phy_read, + .write = ehci_h20ahb_phy_write, +}; + + +/** + * ehci_hcd_h20ahb_probe - initialize Synopsis-based HCDs + * + * Allocates basic resources for this USB host controller, and + * then invokes the start() method for the HCD associated with it + * through the hotplug entry's driver_data. + */ +static int ehci_hcd_h20ahb_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct resource *res; + struct usb_hcd *hcd; + void __iomem *regs; + int ret; + int irq; + int i; + struct h20ahb_hcd *h20ahb; + + if (usb_disabled()) + return -ENODEV; + + /* if (!dev->parent) { + dev_err(dev, "Missing parent device\n"); + return -ENODEV; + }*/ + + /* For DT boot, get platform data from parent. i.e. usbhshost */ + /*if (dev->of_node) { + pdata = dev_get_platdata(dev->parent); + dev->platform_data = pdata; + } + + if (!pdata) { + dev_err(dev, "Missing platform data\n"); + return -ENODEV; + }*/ + + irq = platform_get_irq(pdev, 0); + if (irq < 0) { + dev_err(dev, "EHCI irq failed\n"); + return -ENODEV; + } + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + regs = devm_ioremap_resource(dev, res); + if (IS_ERR(regs)) + return PTR_ERR(regs); + + /* + * Right now device-tree probed devices don't get dma_mask set. + * Since shared usb code relies on it, set it here for now. + * Once we have dma capability bindings this can go away. + */ + ret = dma_coerce_mask_and_coherent(dev, DMA_BIT_MASK(32)); + if (ret) + return ret; + + ret = -ENODEV; + hcd = usb_create_hcd(&ehci_h20ahb_hc_driver, dev, + dev_name(dev)); + if (!hcd) { + dev_err(dev, "Failed to create HCD\n"); + return -ENOMEM; + } + + hcd->rsrc_start = res->start; + hcd->rsrc_len = resource_size(res); + hcd->regs = regs; + hcd_to_ehci(hcd)->caps = regs; + + h20ahb = (struct h20ahb_hcd *)hcd_to_ehci(hcd)->priv; + h20ahb->nports = 1; + + platform_set_drvdata(pdev, hcd); + + /* get the PHY devices if needed */ + for (i = 0 ; i < h20ahb->nports ; i++) { + struct usb_phy *phy; + + /* get the PHY device */ +#if 0 + if (dev->of_node) + phy = devm_usb_get_phy_by_phandle(dev, "phys", i); + else + phy = devm_usb_get_phy_dev(dev, i); +#endif + phy = otg_ulpi_create(&ehci_h20ahb_phy_io_ops, 0); + if (IS_ERR(phy)) { + ret = PTR_ERR(phy); + dev_err(dev, "Can't get PHY device for port %d: %d\n", + i, ret); + goto err_phy; + } + phy->dev = dev; + usb_add_phy_dev(phy); + + h20ahb->phy[i] = phy; + phy->io_priv = hcd->regs + EHCI_INSNREG05_ULPI; + +#if 0 + usb_phy_init(h20ahb->phy[i]); + /* bring PHY out of suspend */ + usb_phy_set_suspend(h20ahb->phy[i], 0); +#endif + } + + /* make the first port's phy the one used by hcd as well */ + hcd->phy = h20ahb->phy[0]; + + pm_runtime_enable(dev); + pm_runtime_get_sync(dev); + + /* + * An undocumented "feature" in the H20AHB EHCI controller, + * causes suspended ports to be taken out of suspend when + * the USBCMD.Run/Stop bit is cleared (for example when + * we do ehci_bus_suspend). + * This breaks suspend-resume if the root-hub is allowed + * to suspend. Writing 1 to this undocumented register bit + * disables this feature and restores normal behavior. + */ + ehci_write(regs, EHCI_INSNREG04, + EHCI_INSNREG04_DISABLE_UNSUSPEND); + + ret = usb_add_hcd(hcd, irq, IRQF_SHARED); + if (ret) { + dev_err(dev, "failed to add hcd with err %d\n", ret); + goto err_pm_runtime; + } + device_wakeup_enable(hcd->self.controller); + + /* + * Bring PHYs out of reset for non PHY modes. + * Even though HSIC mode is a PHY-less mode, the reset + * line exists between the chips and can be modelled + * as a PHY device for reset control. + */ + for (i = 0; i < h20ahb->nports; i++) { + usb_phy_init(h20ahb->phy[i]); + /* bring PHY out of suspend */ + usb_phy_set_suspend(h20ahb->phy[i], 0); + } + + return 0; + +err_pm_runtime: + pm_runtime_put_sync(dev); + +err_phy: + for (i = 0; i < h20ahb->nports; i++) { + if (h20ahb->phy[i]) + usb_phy_shutdown(h20ahb->phy[i]); + } + + usb_put_hcd(hcd); + + return ret; +} + + +/** + * ehci_hcd_h20ahb_remove - shutdown processing for EHCI HCDs + * @pdev: USB Host Controller being removed + * + * Reverses the effect of usb_ehci_hcd_h20ahb_probe(), first invoking + * the HCD's stop() method. It is always called from a thread + * context, normally "rmmod", "apmd", or something similar. + */ +static int ehci_hcd_h20ahb_remove(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct usb_hcd *hcd = dev_get_drvdata(dev); + struct h20ahb_hcd *h20ahb = (struct h20ahb_hcd *)hcd_to_ehci(hcd)->priv; + int i; + + usb_remove_hcd(hcd); + + for (i = 0; i < h20ahb->nports; i++) { + if (h20ahb->phy[i]) + usb_phy_shutdown(h20ahb->phy[i]); + } + + usb_put_hcd(hcd); + pm_runtime_put_sync(dev); + pm_runtime_disable(dev); + + return 0; +} + +static const struct of_device_id h20ahb_ehci_dt_ids[] = { + { .compatible = "snps,ehci-h20ahb" }, + { } +}; + +MODULE_DEVICE_TABLE(of, h20ahb_ehci_dt_ids); + +static struct platform_driver ehci_hcd_h20ahb_driver = { + .probe = ehci_hcd_h20ahb_probe, + .remove = ehci_hcd_h20ahb_remove, + .shutdown = usb_hcd_platform_shutdown, + /*.suspend = ehci_hcd_h20ahb_suspend, */ + /*.resume = ehci_hcd_h20ahb_resume, */ + .driver = { + .name = hcd_name, + .of_match_table = h20ahb_ehci_dt_ids, + } +}; + +/*-------------------------------------------------------------------------*/ + +static int __init ehci_h20ahb_init(void) +{ + if (usb_disabled()) + return -ENODEV; + + pr_info("%s: " DRIVER_DESC "\n", hcd_name); + + ehci_init_driver(&ehci_h20ahb_hc_driver, &ehci_h20ahb_overrides); + return platform_driver_register(&ehci_hcd_h20ahb_driver); +} +module_init(ehci_h20ahb_init); + +static void __exit ehci_h20ahb_cleanup(void) +{ + platform_driver_unregister(&ehci_hcd_h20ahb_driver); +} +module_exit(ehci_h20ahb_cleanup); + +MODULE_ALIAS("platform:ehci-h20ahb"); +MODULE_AUTHOR("Liviu Dudau <Liviu.Dudau@arm.com>"); + +MODULE_DESCRIPTION(DRIVER_DESC); +MODULE_LICENSE("GPL"); diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c index 5a160063176d..f1c3ad69ca54 100644 --- a/drivers/usb/host/ehci-hcd.c +++ b/drivers/usb/host/ehci-hcd.c @@ -591,11 +591,16 @@ static int ehci_run (struct usb_hcd *hcd) */ hcc_params = ehci_readl(ehci, &ehci->caps->hcc_params); if (HCC_64BIT_ADDR(hcc_params)) { - ehci_writel(ehci, 0, &ehci->regs->segment); -#if 0 -// this is deeply broken on almost all architectures +#ifdef CONFIG_ARM64 + ehci_writel(ehci, ehci->periodic_dma >> 32, &ehci->regs->segment); + /* + * this is deeply broken on almost all architectures + * but arm64 can use it so enable it + */ if (!dma_set_mask(hcd->self.controller, DMA_BIT_MASK(64))) ehci_info(ehci, "enabled 64bit DMA\n"); +#else + ehci_writel(ehci, 0, &ehci->regs->segment); #endif } diff --git a/drivers/usb/host/ohci-hcd.c b/drivers/usb/host/ohci-hcd.c index 865946cde765..8c57e970bb98 100644 --- a/drivers/usb/host/ohci-hcd.c +++ b/drivers/usb/host/ohci-hcd.c @@ -1196,25 +1196,6 @@ MODULE_LICENSE ("GPL"); #define PLATFORM_DRIVER ohci_platform_driver #endif -#if !defined(PCI_DRIVER) && \ - !defined(PLATFORM_DRIVER) && \ - !defined(OMAP1_PLATFORM_DRIVER) && \ - !defined(OMAP3_PLATFORM_DRIVER) && \ - !defined(OF_PLATFORM_DRIVER) && \ - !defined(SA1111_DRIVER) && \ - !defined(PS3_SYSTEM_BUS_DRIVER) && \ - !defined(SM501_OHCI_DRIVER) && \ - !defined(TMIO_OHCI_DRIVER) && \ - !defined(S3C2410_PLATFORM_DRIVER) && \ - !defined(EXYNOS_PLATFORM_DRIVER) && \ - !defined(EP93XX_PLATFORM_DRIVER) && \ - !defined(AT91_PLATFORM_DRIVER) && \ - !defined(NXP_PLATFORM_DRIVER) && \ - !defined(DAVINCI_PLATFORM_DRIVER) && \ - !defined(SPEAR_PLATFORM_DRIVER) -#error "missing bus glue for ohci-hcd" -#endif - static int __init ohci_hcd_mod_init(void) { int retval = 0; diff --git a/drivers/usb/phy/Kconfig b/drivers/usb/phy/Kconfig index 2311b1e4e43c..be41733a5eac 100644 --- a/drivers/usb/phy/Kconfig +++ b/drivers/usb/phy/Kconfig @@ -198,7 +198,7 @@ config USB_RCAR_PHY config USB_ULPI bool "Generic ULPI Transceiver Driver" - depends on ARM + depends on ARM || ARM64 help Enable this to support ULPI connected USB OTG transceivers which are likely found on embedded boards. diff --git a/drivers/usb/phy/phy-ulpi.c b/drivers/usb/phy/phy-ulpi.c index 17ea3f271bd8..4e3877c329f2 100644 --- a/drivers/usb/phy/phy-ulpi.c +++ b/drivers/usb/phy/phy-ulpi.c @@ -48,6 +48,7 @@ static struct ulpi_info ulpi_ids[] = { ULPI_INFO(ULPI_ID(0x04cc, 0x1504), "NXP ISP1504"), ULPI_INFO(ULPI_ID(0x0424, 0x0006), "SMSC USB331x"), ULPI_INFO(ULPI_ID(0x0424, 0x0007), "SMSC USB3320"), + ULPI_INFO(ULPI_ID(0x0424, 0x0009), "SMSC USB334x"), ULPI_INFO(ULPI_ID(0x0451, 0x1507), "TI TUSB1210"), }; diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig index 2e937bdace6f..29a5121ce7fd 100644 --- a/drivers/video/Kconfig +++ b/drivers/video/Kconfig @@ -39,6 +39,11 @@ config VIDEOMODE_HELPERS config HDMI bool +config VEXPRESS_DVI_CONTROL + bool "Versatile Express DVI control" + depends on FB && VEXPRESS_CONFIG + default y + menuconfig FB tristate "Support for frame buffer devices" ---help--- @@ -312,7 +317,8 @@ config FB_PM2_FIFO_DISCONNECT config FB_ARMCLCD tristate "ARM PrimeCell PL110 support" - depends on FB && ARM && ARM_AMBA + depends on ARM || ARM64 || COMPILE_TEST + depends on FB && ARM_AMBA select FB_CFB_FILLRECT select FB_CFB_COPYAREA select FB_CFB_IMAGEBLIT @@ -326,6 +332,21 @@ config FB_ARMCLCD here and read <file:Documentation/kbuild/modules.txt>. The module will be called amba-clcd. +config FB_ARMHDLCD + tristate "ARM High Definition LCD support" + depends on FB && ARM + select FB_CFB_FILLRECT + select FB_CFB_COPYAREA + select FB_CFB_IMAGEBLIT + help + This framebuffer device driver is for the ARM High Definition + Colour LCD controller. + + If you want to compile this as a module (=code which can be + inserted into and removed from the running kernel), say M + here and read <file:Documentation/kbuild/modules.txt>. The module + will be called arm-hdlcd. + config FB_ACORN bool "Acorn VIDC support" depends on (FB = y) && ARM && ARCH_ACORN diff --git a/drivers/video/Makefile b/drivers/video/Makefile index e8bae8dd4804..33869eea4981 100644 --- a/drivers/video/Makefile +++ b/drivers/video/Makefile @@ -99,6 +99,7 @@ obj-$(CONFIG_FB_ATMEL) += atmel_lcdfb.o obj-$(CONFIG_FB_PVR2) += pvr2fb.o obj-$(CONFIG_FB_VOODOO1) += sstfb.o obj-$(CONFIG_FB_ARMCLCD) += amba-clcd.o +obj-$(CONFIG_FB_ARMHDLCD) += arm-hdlcd.o obj-$(CONFIG_FB_GOLDFISH) += goldfishfb.o obj-$(CONFIG_FB_68328) += 68328fb.o obj-$(CONFIG_FB_GBE) += gbefb.o @@ -177,3 +178,6 @@ obj-$(CONFIG_VIDEOMODE_HELPERS) += display_timing.o videomode.o ifeq ($(CONFIG_OF),y) obj-$(CONFIG_VIDEOMODE_HELPERS) += of_display_timing.o of_videomode.o endif + +# platform specific output drivers +obj-$(CONFIG_VEXPRESS_DVI_CONTROL) += vexpress-dvi.o diff --git a/drivers/video/amba-clcd.c b/drivers/video/amba-clcd.c index 0a2cce7285be..dba653f6badc 100644 --- a/drivers/video/amba-clcd.c +++ b/drivers/video/amba-clcd.c @@ -16,7 +16,10 @@ #include <linux/string.h> #include <linux/slab.h> #include <linux/delay.h> +#include <linux/dma-mapping.h> +#include <linux/memblock.h> #include <linux/mm.h> +#include <linux/of.h> #include <linux/fb.h> #include <linux/init.h> #include <linux/ioport.h> @@ -30,8 +33,20 @@ #define to_clcd(info) container_of(info, struct clcd_fb, fb) +#ifdef CONFIG_ARM +#define clcdfb_dma_alloc dma_alloc_writecombine +#define clcdfb_dma_free dma_free_writecombine +#define clcdfb_dma_mmap dma_mmap_writecombine +#else +#define clcdfb_dma_alloc dma_alloc_coherent +#define clcdfb_dma_free dma_free_coherent +#define clcdfb_dma_mmap dma_mmap_coherent +#endif + /* This is limited to 16 characters when displayed by X startup */ static const char *clcd_name = "CLCD FB"; +static char *def_mode; +module_param_named(mode, def_mode, charp, 0); /* * Unfortunately, the enable/disable functions may be called either from @@ -392,6 +407,44 @@ static int clcdfb_blank(int blank_mode, struct fb_info *info) return 0; } +int clcdfb_mmap_dma(struct clcd_fb *fb, struct vm_area_struct *vma) +{ + return clcdfb_dma_mmap(&fb->dev->dev, vma, + fb->fb.screen_base, + fb->fb.fix.smem_start, + fb->fb.fix.smem_len); +} + +int clcdfb_mmap_io(struct clcd_fb *fb, struct vm_area_struct *vma) +{ + unsigned long user_count, count, pfn, off; + + user_count = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT; + count = PAGE_ALIGN(fb->fb.fix.smem_len) >> PAGE_SHIFT; + pfn = fb->fb.fix.smem_start >> PAGE_SHIFT; + off = vma->vm_pgoff; + + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); + + if (off < count && user_count <= (count - off)) + return remap_pfn_range(vma, vma->vm_start, pfn + off, + user_count << PAGE_SHIFT, + vma->vm_page_prot); + + return -ENXIO; +} + +void clcdfb_remove_dma(struct clcd_fb *fb) +{ + clcdfb_dma_free(&fb->dev->dev, fb->fb.fix.smem_len, + fb->fb.screen_base, fb->fb.fix.smem_start); +} + +void clcdfb_remove_io(struct clcd_fb *fb) +{ + iounmap(fb->fb.screen_base); +} + static int clcdfb_mmap(struct fb_info *info, struct vm_area_struct *vma) { @@ -542,14 +595,247 @@ static int clcdfb_register(struct clcd_fb *fb) return ret; } +struct string_lookup { + const char *string; + const u32 val; +}; + +static struct string_lookup vmode_lookups[] = { + { "FB_VMODE_NONINTERLACED", FB_VMODE_NONINTERLACED}, + { "FB_VMODE_INTERLACED", FB_VMODE_INTERLACED}, + { "FB_VMODE_DOUBLE", FB_VMODE_DOUBLE}, + { "FB_VMODE_ODD_FLD_FIRST", FB_VMODE_ODD_FLD_FIRST}, + { NULL, 0 }, +}; + +static struct string_lookup tim2_lookups[] = { + { "TIM2_CLKSEL", TIM2_CLKSEL}, + { "TIM2_IVS", TIM2_IVS}, + { "TIM2_IHS", TIM2_IHS}, + { "TIM2_IPC", TIM2_IPC}, + { "TIM2_IOE", TIM2_IOE}, + { "TIM2_BCD", TIM2_BCD}, + { NULL, 0}, +}; +static struct string_lookup cntl_lookups[] = { + {"CNTL_LCDEN", CNTL_LCDEN}, + {"CNTL_LCDBPP1", CNTL_LCDBPP1}, + {"CNTL_LCDBPP2", CNTL_LCDBPP2}, + {"CNTL_LCDBPP4", CNTL_LCDBPP4}, + {"CNTL_LCDBPP8", CNTL_LCDBPP8}, + {"CNTL_LCDBPP16", CNTL_LCDBPP16}, + {"CNTL_LCDBPP16_565", CNTL_LCDBPP16_565}, + {"CNTL_LCDBPP16_444", CNTL_LCDBPP16_444}, + {"CNTL_LCDBPP24", CNTL_LCDBPP24}, + {"CNTL_LCDBW", CNTL_LCDBW}, + {"CNTL_LCDTFT", CNTL_LCDTFT}, + {"CNTL_LCDMONO8", CNTL_LCDMONO8}, + {"CNTL_LCDDUAL", CNTL_LCDDUAL}, + {"CNTL_BGR", CNTL_BGR}, + {"CNTL_BEBO", CNTL_BEBO}, + {"CNTL_BEPO", CNTL_BEPO}, + {"CNTL_LCDPWR", CNTL_LCDPWR}, + {"CNTL_LCDVCOMP(1)", CNTL_LCDVCOMP(1)}, + {"CNTL_LCDVCOMP(2)", CNTL_LCDVCOMP(2)}, + {"CNTL_LCDVCOMP(3)", CNTL_LCDVCOMP(3)}, + {"CNTL_LCDVCOMP(4)", CNTL_LCDVCOMP(4)}, + {"CNTL_LCDVCOMP(5)", CNTL_LCDVCOMP(5)}, + {"CNTL_LCDVCOMP(6)", CNTL_LCDVCOMP(6)}, + {"CNTL_LCDVCOMP(7)", CNTL_LCDVCOMP(7)}, + {"CNTL_LDMAFIFOTIME", CNTL_LDMAFIFOTIME}, + {"CNTL_WATERMARK", CNTL_WATERMARK}, + { NULL, 0}, +}; +static struct string_lookup caps_lookups[] = { + {"CLCD_CAP_RGB444", CLCD_CAP_RGB444}, + {"CLCD_CAP_RGB5551", CLCD_CAP_RGB5551}, + {"CLCD_CAP_RGB565", CLCD_CAP_RGB565}, + {"CLCD_CAP_RGB888", CLCD_CAP_RGB888}, + {"CLCD_CAP_BGR444", CLCD_CAP_BGR444}, + {"CLCD_CAP_BGR5551", CLCD_CAP_BGR5551}, + {"CLCD_CAP_BGR565", CLCD_CAP_BGR565}, + {"CLCD_CAP_BGR888", CLCD_CAP_BGR888}, + {"CLCD_CAP_444", CLCD_CAP_444}, + {"CLCD_CAP_5551", CLCD_CAP_5551}, + {"CLCD_CAP_565", CLCD_CAP_565}, + {"CLCD_CAP_888", CLCD_CAP_888}, + {"CLCD_CAP_RGB", CLCD_CAP_RGB}, + {"CLCD_CAP_BGR", CLCD_CAP_BGR}, + {"CLCD_CAP_ALL", CLCD_CAP_ALL}, + { NULL, 0}, +}; + +u32 parse_setting(struct string_lookup *lookup, const char *name) +{ + int i = 0; + while (lookup[i].string != NULL) { + if (strcmp(lookup[i].string, name) == 0) + return lookup[i].val; + ++i; + } + return -EINVAL; +} + +u32 get_string_lookup(struct device_node *node, const char *name, + struct string_lookup *lookup) +{ + const char *string; + int count, i, ret = 0; + + count = of_property_count_strings(node, name); + if (count >= 0) + for (i = 0; i < count; i++) + if (of_property_read_string_index(node, name, i, + &string) == 0) + ret |= parse_setting(lookup, string); + return ret; +} + +int get_val(struct device_node *node, const char *string) +{ + u32 ret = 0; + + if (of_property_read_u32(node, string, &ret)) + ret = -1; + return ret; +} + +struct clcd_panel *getPanel(struct device_node *node) +{ + static struct clcd_panel panel; + + panel.mode.refresh = get_val(node, "refresh"); + panel.mode.xres = get_val(node, "xres"); + panel.mode.yres = get_val(node, "yres"); + panel.mode.pixclock = get_val(node, "pixclock"); + panel.mode.left_margin = get_val(node, "left_margin"); + panel.mode.right_margin = get_val(node, "right_margin"); + panel.mode.upper_margin = get_val(node, "upper_margin"); + panel.mode.lower_margin = get_val(node, "lower_margin"); + panel.mode.hsync_len = get_val(node, "hsync_len"); + panel.mode.vsync_len = get_val(node, "vsync_len"); + panel.mode.sync = get_val(node, "sync"); + panel.bpp = get_val(node, "bpp"); + panel.width = (signed short) get_val(node, "width"); + panel.height = (signed short) get_val(node, "height"); + + panel.mode.vmode = get_string_lookup(node, "vmode", vmode_lookups); + panel.tim2 = get_string_lookup(node, "tim2", tim2_lookups); + panel.cntl = get_string_lookup(node, "cntl", cntl_lookups); + panel.caps = get_string_lookup(node, "caps", caps_lookups); + + return &panel; +} + +struct clcd_panel *clcdfb_get_panel(const char *name) +{ + struct device_node *node = NULL; + const char *mode; + struct clcd_panel *panel = NULL; + + do { + node = of_find_compatible_node(node, NULL, "panel"); + if (node) + if (of_property_read_string(node, "mode", &mode) == 0) + if (strcmp(mode, name) == 0) { + panel = getPanel(node); + panel->mode.name = name; + } + } while (node != NULL); + + return panel; +} + +#ifdef CONFIG_OF +static int clcdfb_dt_init(struct clcd_fb *fb) +{ + int err = 0; + struct device_node *node; + const char *mode; + dma_addr_t dma; + u32 use_dma; + const __be32 *prop; + int len, na, ns; + phys_addr_t fb_base, fb_size; + + node = fb->dev->dev.of_node; + if (!node) + return -ENODEV; + + na = of_n_addr_cells(node); + ns = of_n_size_cells(node); + + if (def_mode && strlen(def_mode) > 0) { + fb->panel = clcdfb_get_panel(def_mode); + if (!fb->panel) + printk(KERN_ERR "CLCD: invalid mode specified on the command line (%s)\n", def_mode); + } + + if (!fb->panel) { + if (WARN_ON(of_property_read_string(node, "mode", &mode))) + return -ENODEV; + fb->panel = clcdfb_get_panel(mode); + } + + if (!fb->panel) + return -EINVAL; + fb->fb.fix.smem_len = fb->panel->mode.xres * fb->panel->mode.yres * 2; + + fb->board->name = "Device Tree CLCD PL111"; + fb->board->caps = CLCD_CAP_5551 | CLCD_CAP_565; + fb->board->check = clcdfb_check; + fb->board->decode = clcdfb_decode; + + if (of_property_read_u32(node, "use_dma", &use_dma)) + use_dma = 0; + + if (use_dma) { + fb->fb.screen_base = clcdfb_dma_alloc(&fb->dev->dev, + fb->fb.fix.smem_len, + &dma, GFP_KERNEL); + if (!fb->fb.screen_base) { + pr_err("CLCD: unable to map framebuffer\n"); + return -ENOMEM; + } + + fb->fb.fix.smem_start = dma; + fb->board->mmap = clcdfb_mmap_dma; + fb->board->remove = clcdfb_remove_dma; + } else { + prop = of_get_property(node, "framebuffer", &len); + if (WARN_ON(!prop || len < (na + ns) * sizeof(*prop))) + return -EINVAL; + + fb_base = of_read_number(prop, na); + fb_size = of_read_number(prop + na, ns); + + fb->fb.fix.smem_start = fb_base; + fb->fb.screen_base = ioremap_wc(fb_base, fb_size); + fb->board->mmap = clcdfb_mmap_io; + fb->board->remove = clcdfb_remove_io; + } + + return err; +} +#endif /* CONFIG_OF */ + static int clcdfb_probe(struct amba_device *dev, const struct amba_id *id) { struct clcd_board *board = dev->dev.platform_data; struct clcd_fb *fb; int ret; - if (!board) - return -EINVAL; + if (!board) { +#ifdef CONFIG_OF + if (dev->dev.of_node) { + board = kzalloc(sizeof(struct clcd_board), GFP_KERNEL); + if (!board) + return -ENOMEM; + board->setup = clcdfb_dt_init; + } else +#endif + return -EINVAL; + } ret = amba_request_regions(dev, NULL); if (ret) { diff --git a/drivers/video/arm-hdlcd.c b/drivers/video/arm-hdlcd.c new file mode 100644 index 000000000000..cfd631e3dc52 --- /dev/null +++ b/drivers/video/arm-hdlcd.c @@ -0,0 +1,844 @@ +/* + * drivers/video/arm-hdlcd.c + * + * Copyright (C) 2011 ARM Limited + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file COPYING in the main directory of this archive + * for more details. + * + * ARM HDLCD Controller + */ + +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/errno.h> +#include <linux/string.h> +#include <linux/ctype.h> +#include <linux/mm.h> +#include <linux/delay.h> +#include <linux/of.h> +#include <linux/fb.h> +#include <linux/clk.h> +#include <linux/init.h> +#include <linux/interrupt.h> +#include <linux/ioport.h> +#include <linux/dma-mapping.h> +#include <linux/platform_device.h> +#include <linux/memblock.h> +#include <linux/arm-hdlcd.h> +#ifdef HDLCD_COUNT_BUFFERUNDERRUNS +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#endif + +#include "edid.h" + +#ifdef CONFIG_SERIAL_AMBA_PCU_UART +int get_edid(u8 *msgbuf); +#else +#endif + +#define to_hdlcd_device(info) container_of(info, struct hdlcd_device, fb) + +static struct of_device_id hdlcd_of_matches[] = { + { .compatible = "arm,hdlcd" }, + {}, +}; + +/* Framebuffer size. */ +static unsigned long framebuffer_size; + +#ifdef HDLCD_COUNT_BUFFERUNDERRUNS +static unsigned long buffer_underrun_events; +static DEFINE_SPINLOCK(hdlcd_underrun_lock); + +static void hdlcd_underrun_set(unsigned long val) +{ + spin_lock(&hdlcd_underrun_lock); + buffer_underrun_events = val; + spin_unlock(&hdlcd_underrun_lock); +} + +static unsigned long hdlcd_underrun_get(void) +{ + unsigned long val; + spin_lock(&hdlcd_underrun_lock); + val = buffer_underrun_events; + spin_unlock(&hdlcd_underrun_lock); + return val; +} + +#ifdef CONFIG_PROC_FS +static int hdlcd_underrun_show(struct seq_file *m, void *v) +{ + unsigned char underrun_string[32]; + snprintf(underrun_string, 32, "%lu\n", hdlcd_underrun_get()); + seq_puts(m, underrun_string); + return 0; +} + +static int proc_hdlcd_underrun_open(struct inode *inode, struct file *file) +{ + return single_open(file, hdlcd_underrun_show, NULL); +} + +static const struct file_operations proc_hdlcd_underrun_operations = { + .open = proc_hdlcd_underrun_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static int hdlcd_underrun_init(void) +{ + hdlcd_underrun_set(0); + proc_create("hdlcd_underrun", 0, NULL, &proc_hdlcd_underrun_operations); + return 0; +} +static void hdlcd_underrun_close(void) +{ + remove_proc_entry("hdlcd_underrun", NULL); +} +#else +static int hdlcd_underrun_init(void) { return 0; } +static void hdlcd_underrun_close(void) { } +#endif +#endif + +static char *fb_mode = "1680x1050-32@60\0\0\0\0\0"; + +static struct fb_var_screeninfo cached_var_screeninfo; + +static struct fb_videomode hdlcd_default_mode = { + .refresh = 60, + .xres = 1680, + .yres = 1050, + .pixclock = 8403, + .left_margin = 80, + .right_margin = 48, + .upper_margin = 21, + .lower_margin = 3, + .hsync_len = 32, + .vsync_len = 6, + .sync = FB_SYNC_HOR_HIGH_ACT | FB_SYNC_VERT_HIGH_ACT, + .vmode = FB_VMODE_NONINTERLACED +}; + +static inline void hdlcd_enable(struct hdlcd_device *hdlcd) +{ + dev_dbg(hdlcd->dev, "HDLCD: output enabled\n"); + writel(1, hdlcd->base + HDLCD_REG_COMMAND); +} + +static inline void hdlcd_disable(struct hdlcd_device *hdlcd) +{ + dev_dbg(hdlcd->dev, "HDLCD: output disabled\n"); + writel(0, hdlcd->base + HDLCD_REG_COMMAND); +} + +static int hdlcd_set_bitfields(struct hdlcd_device *hdlcd, + struct fb_var_screeninfo *var) +{ + int ret = 0; + + memset(&var->transp, 0, sizeof(var->transp)); + var->red.msb_right = 0; + var->green.msb_right = 0; + var->blue.msb_right = 0; + var->blue.offset = 0; + + switch (var->bits_per_pixel) { + case 8: + /* pseudocolor */ + var->red.length = 8; + var->green.length = 8; + var->blue.length = 8; + break; + case 16: + /* 565 format */ + var->red.length = 5; + var->green.length = 6; + var->blue.length = 5; + break; + case 32: + var->transp.length = 8; + case 24: + var->red.length = 8; + var->green.length = 8; + var->blue.length = 8; + break; + default: + ret = -EINVAL; + break; + } + + if (!ret) { + if(var->bits_per_pixel != 32) + { + var->green.offset = var->blue.length; + var->red.offset = var->green.offset + var->green.length; + } + else + { + /* Previously, the byte ordering for 32-bit color was + * (msb)<alpha><red><green><blue>(lsb) + * but this does not match what android expects and + * the colors are odd. Instead, use + * <alpha><blue><green><red> + * Since we tell fb what we are doing, console + * , X and directfb access should work fine. + */ + var->green.offset = var->red.length; + var->blue.offset = var->green.offset + var->green.length; + var->transp.offset = var->blue.offset + var->blue.length; + } + } + + return ret; +} + +static int hdlcd_check_var(struct fb_var_screeninfo *var, struct fb_info *info) +{ + struct hdlcd_device *hdlcd = to_hdlcd_device(info); + int bytes_per_pixel = var->bits_per_pixel / 8; + +#ifdef HDLCD_NO_VIRTUAL_SCREEN + var->yres_virtual = var->yres; +#else + var->yres_virtual = 2 * var->yres; +#endif + + if ((var->xres_virtual * bytes_per_pixel * var->yres_virtual) > hdlcd->fb.fix.smem_len) + return -ENOMEM; + + if (var->xres > HDLCD_MAX_XRES || var->yres > HDLCD_MAX_YRES) + return -EINVAL; + + /* make sure the bitfields are set appropriately */ + return hdlcd_set_bitfields(hdlcd, var); +} + +/* prototype */ +static int hdlcd_pan_display(struct fb_var_screeninfo *var, + struct fb_info *info); + +#define WRITE_HDLCD_REG(reg, value) writel((value), hdlcd->base + (reg)) +#define READ_HDLCD_REG(reg) readl(hdlcd->base + (reg)) + +static int hdlcd_set_par(struct fb_info *info) +{ + struct hdlcd_device *hdlcd = to_hdlcd_device(info); + int bytes_per_pixel = hdlcd->fb.var.bits_per_pixel / 8; + int polarities; + int old_yoffset; + + /* check for shortcuts */ + old_yoffset = cached_var_screeninfo.yoffset; + cached_var_screeninfo.yoffset = info->var.yoffset; + if (!memcmp(&info->var, &cached_var_screeninfo, + sizeof(struct fb_var_screeninfo))) { + if(old_yoffset != info->var.yoffset) { + /* we only changed yoffset, and we already + * already recorded it a couple lines up + */ + hdlcd_pan_display(&info->var, info); + } + /* or no change */ + return 0; + } + + hdlcd->fb.fix.line_length = hdlcd->fb.var.xres * bytes_per_pixel; + + if (hdlcd->fb.var.bits_per_pixel >= 16) + hdlcd->fb.fix.visual = FB_VISUAL_TRUECOLOR; + else + hdlcd->fb.fix.visual = FB_VISUAL_PSEUDOCOLOR; + + memcpy(&cached_var_screeninfo, &info->var, sizeof(struct fb_var_screeninfo)); + + polarities = HDLCD_POLARITY_DATAEN | +#ifndef CONFIG_ARCH_TUSCAN + HDLCD_POLARITY_PIXELCLK | +#endif + HDLCD_POLARITY_DATA; + polarities |= (hdlcd->fb.var.sync & FB_SYNC_HOR_HIGH_ACT) ? HDLCD_POLARITY_HSYNC : 0; + polarities |= (hdlcd->fb.var.sync & FB_SYNC_VERT_HIGH_ACT) ? HDLCD_POLARITY_VSYNC : 0; + + hdlcd_disable(hdlcd); + + WRITE_HDLCD_REG(HDLCD_REG_FB_LINE_LENGTH, hdlcd->fb.var.xres * bytes_per_pixel); + WRITE_HDLCD_REG(HDLCD_REG_FB_LINE_PITCH, hdlcd->fb.var.xres * bytes_per_pixel); + WRITE_HDLCD_REG(HDLCD_REG_FB_LINE_COUNT, hdlcd->fb.var.yres - 1); + WRITE_HDLCD_REG(HDLCD_REG_V_SYNC, hdlcd->fb.var.vsync_len - 1); + WRITE_HDLCD_REG(HDLCD_REG_V_BACK_PORCH, hdlcd->fb.var.upper_margin - 1); + WRITE_HDLCD_REG(HDLCD_REG_V_DATA, hdlcd->fb.var.yres - 1); + WRITE_HDLCD_REG(HDLCD_REG_V_FRONT_PORCH, hdlcd->fb.var.lower_margin - 1); + WRITE_HDLCD_REG(HDLCD_REG_H_SYNC, hdlcd->fb.var.hsync_len - 1); + WRITE_HDLCD_REG(HDLCD_REG_H_BACK_PORCH, hdlcd->fb.var.left_margin - 1); + WRITE_HDLCD_REG(HDLCD_REG_H_DATA, hdlcd->fb.var.xres - 1); + WRITE_HDLCD_REG(HDLCD_REG_H_FRONT_PORCH, hdlcd->fb.var.right_margin - 1); + WRITE_HDLCD_REG(HDLCD_REG_POLARITIES, polarities); + WRITE_HDLCD_REG(HDLCD_REG_PIXEL_FORMAT, (bytes_per_pixel - 1) << 3); +#ifdef HDLCD_RED_DEFAULT_COLOUR + WRITE_HDLCD_REG(HDLCD_REG_RED_SELECT, (0x00ff0000 | (hdlcd->fb.var.red.length & 0xf) << 8) \ + | hdlcd->fb.var.red.offset); +#else + WRITE_HDLCD_REG(HDLCD_REG_RED_SELECT, ((hdlcd->fb.var.red.length & 0xf) << 8) | hdlcd->fb.var.red.offset); +#endif + WRITE_HDLCD_REG(HDLCD_REG_GREEN_SELECT, ((hdlcd->fb.var.green.length & 0xf) << 8) | hdlcd->fb.var.green.offset); + WRITE_HDLCD_REG(HDLCD_REG_BLUE_SELECT, ((hdlcd->fb.var.blue.length & 0xf) << 8) | hdlcd->fb.var.blue.offset); + + clk_set_rate(hdlcd->clk, (1000000000 / hdlcd->fb.var.pixclock) * 1000); + clk_enable(hdlcd->clk); + + hdlcd_enable(hdlcd); + + return 0; +} + +static int hdlcd_setcolreg(unsigned int regno, unsigned int red, unsigned int green, + unsigned int blue, unsigned int transp, struct fb_info *info) +{ + if (regno < 16) { + u32 *pal = info->pseudo_palette; + + pal[regno] = ((red >> 8) << info->var.red.offset) | + ((green >> 8) << info->var.green.offset) | + ((blue >> 8) << info->var.blue.offset); + } + + return 0; +} + +static irqreturn_t hdlcd_irq(int irq, void *data) +{ + struct hdlcd_device *hdlcd = data; + unsigned long irq_mask, irq_status; + + irq_mask = READ_HDLCD_REG(HDLCD_REG_INT_MASK); + irq_status = READ_HDLCD_REG(HDLCD_REG_INT_STATUS); + + /* acknowledge interrupt(s) */ + WRITE_HDLCD_REG(HDLCD_REG_INT_CLEAR, irq_status); +#ifdef HDLCD_COUNT_BUFFERUNDERRUNS + if (irq_status & HDLCD_INTERRUPT_UNDERRUN) { + /* increment the count */ + hdlcd_underrun_set(hdlcd_underrun_get() + 1); + } +#endif + if (irq_status & HDLCD_INTERRUPT_VSYNC) { + /* disable future VSYNC interrupts */ + WRITE_HDLCD_REG(HDLCD_REG_INT_MASK, irq_mask & ~HDLCD_INTERRUPT_VSYNC); + + complete(&hdlcd->vsync_completion); + } + + return IRQ_HANDLED; +} + +static int hdlcd_wait_for_vsync(struct fb_info *info) +{ + struct hdlcd_device *hdlcd = to_hdlcd_device(info); + unsigned long irq_mask; + int err; + + /* enable VSYNC interrupt */ + irq_mask = READ_HDLCD_REG(HDLCD_REG_INT_MASK); + WRITE_HDLCD_REG(HDLCD_REG_INT_MASK, irq_mask | HDLCD_INTERRUPT_VSYNC); + + err = wait_for_completion_interruptible_timeout(&hdlcd->vsync_completion, + msecs_to_jiffies(100)); + + if (!err) + return -ETIMEDOUT; + + return 0; +} + +static int hdlcd_blank(int blank_mode, struct fb_info *info) +{ + struct hdlcd_device *hdlcd = to_hdlcd_device(info); + + switch (blank_mode) { + case FB_BLANK_POWERDOWN: + clk_disable(hdlcd->clk); + case FB_BLANK_NORMAL: + hdlcd_disable(hdlcd); + break; + case FB_BLANK_UNBLANK: + clk_enable(hdlcd->clk); + hdlcd_enable(hdlcd); + break; + case FB_BLANK_VSYNC_SUSPEND: + case FB_BLANK_HSYNC_SUSPEND: + default: + return 1; + } + + return 0; +} + +static void hdlcd_mmap_open(struct vm_area_struct *vma) +{ +} + +static void hdlcd_mmap_close(struct vm_area_struct *vma) +{ +} + +static struct vm_operations_struct hdlcd_mmap_ops = { + .open = hdlcd_mmap_open, + .close = hdlcd_mmap_close, +}; + +static int hdlcd_mmap(struct fb_info *info, struct vm_area_struct *vma) +{ + struct hdlcd_device *hdlcd = to_hdlcd_device(info); + unsigned long off; + unsigned long start; + unsigned long len = hdlcd->fb.fix.smem_len; + + if (vma->vm_end - vma->vm_start == 0) + return 0; + if (vma->vm_pgoff > (~0UL >> PAGE_SHIFT)) + return -EINVAL; + + off = vma->vm_pgoff << PAGE_SHIFT; + if ((off >= len) || (vma->vm_end - vma->vm_start + off) > len) + return -EINVAL; + + start = hdlcd->fb.fix.smem_start; + off += start; + + vma->vm_pgoff = off >> PAGE_SHIFT; + vma->vm_flags |= VM_IO; + vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot); + vma->vm_ops = &hdlcd_mmap_ops; + if (io_remap_pfn_range(vma, vma->vm_start, off >> PAGE_SHIFT, + vma->vm_end - vma->vm_start, + vma->vm_page_prot)) + return -EAGAIN; + + return 0; +} + +static int hdlcd_pan_display(struct fb_var_screeninfo *var, struct fb_info *info) +{ + struct hdlcd_device *hdlcd = to_hdlcd_device(info); + + hdlcd->fb.var.yoffset = var->yoffset; + WRITE_HDLCD_REG(HDLCD_REG_FB_BASE, hdlcd->fb.fix.smem_start + + (var->yoffset * hdlcd->fb.fix.line_length)); + + hdlcd_wait_for_vsync(info); + + return 0; +} + +static int hdlcd_ioctl(struct fb_info *info, unsigned int cmd, unsigned long arg) +{ + int err; + + switch (cmd) { + case FBIO_WAITFORVSYNC: + err = hdlcd_wait_for_vsync(info); + break; + default: + err = -ENOIOCTLCMD; + break; + } + + return err; +} + +static struct fb_ops hdlcd_ops = { + .owner = THIS_MODULE, + .fb_check_var = hdlcd_check_var, + .fb_set_par = hdlcd_set_par, + .fb_setcolreg = hdlcd_setcolreg, + .fb_blank = hdlcd_blank, + .fb_fillrect = cfb_fillrect, + .fb_copyarea = cfb_copyarea, + .fb_imageblit = cfb_imageblit, + .fb_mmap = hdlcd_mmap, + .fb_pan_display = hdlcd_pan_display, + .fb_ioctl = hdlcd_ioctl, + .fb_compat_ioctl = hdlcd_ioctl +}; + +static int hdlcd_setup(struct hdlcd_device *hdlcd) +{ + u32 version; + int err = -EFAULT; + + hdlcd->fb.device = hdlcd->dev; + + hdlcd->clk = clk_get(hdlcd->dev, NULL); + if (IS_ERR(hdlcd->clk)) { + dev_err(hdlcd->dev, "HDLCD: unable to find clock data\n"); + return PTR_ERR(hdlcd->clk); + } + + err = clk_prepare(hdlcd->clk); + if (err) + goto clk_prepare_err; + + hdlcd->base = ioremap_nocache(hdlcd->fb.fix.mmio_start, hdlcd->fb.fix.mmio_len); + if (!hdlcd->base) { + dev_err(hdlcd->dev, "HDLCD: unable to map registers\n"); + goto remap_err; + } + + hdlcd->fb.pseudo_palette = kmalloc(sizeof(u32) * 16, GFP_KERNEL); + if (!hdlcd->fb.pseudo_palette) { + dev_err(hdlcd->dev, "HDLCD: unable to allocate pseudo_palette memory\n"); + err = -ENOMEM; + goto kmalloc_err; + } + + version = readl(hdlcd->base + HDLCD_REG_VERSION); + if ((version & HDLCD_PRODUCT_MASK) != HDLCD_PRODUCT_ID) { + dev_err(hdlcd->dev, "HDLCD: unknown product id: 0x%x\n", version); + err = -EINVAL; + goto kmalloc_err; + } + dev_info(hdlcd->dev, "HDLCD: found ARM HDLCD version r%dp%d\n", + (version & HDLCD_VERSION_MAJOR_MASK) >> 8, + version & HDLCD_VERSION_MINOR_MASK); + + strcpy(hdlcd->fb.fix.id, "hdlcd"); + hdlcd->fb.fbops = &hdlcd_ops; + hdlcd->fb.flags = FBINFO_FLAG_DEFAULT/* | FBINFO_VIRTFB*/; + + hdlcd->fb.fix.type = FB_TYPE_PACKED_PIXELS; + hdlcd->fb.fix.type_aux = 0; + hdlcd->fb.fix.xpanstep = 0; + hdlcd->fb.fix.ypanstep = 1; + hdlcd->fb.fix.ywrapstep = 0; + hdlcd->fb.fix.accel = FB_ACCEL_NONE; + + hdlcd->fb.var.nonstd = 0; + hdlcd->fb.var.activate = FB_ACTIVATE_NOW; + hdlcd->fb.var.height = -1; + hdlcd->fb.var.width = -1; + hdlcd->fb.var.accel_flags = 0; + + init_completion(&hdlcd->vsync_completion); + + if (hdlcd->edid) { + /* build modedb from EDID */ + fb_edid_to_monspecs(hdlcd->edid, &hdlcd->fb.monspecs); + fb_videomode_to_modelist(hdlcd->fb.monspecs.modedb, + hdlcd->fb.monspecs.modedb_len, + &hdlcd->fb.modelist); + fb_find_mode(&hdlcd->fb.var, &hdlcd->fb, fb_mode, + hdlcd->fb.monspecs.modedb, + hdlcd->fb.monspecs.modedb_len, + &hdlcd_default_mode, 32); + } else { + hdlcd->fb.monspecs.hfmin = 0; + hdlcd->fb.monspecs.hfmax = 100000; + hdlcd->fb.monspecs.vfmin = 0; + hdlcd->fb.monspecs.vfmax = 400; + hdlcd->fb.monspecs.dclkmin = 1000000; + hdlcd->fb.monspecs.dclkmax = 100000000; + fb_find_mode(&hdlcd->fb.var, &hdlcd->fb, fb_mode, NULL, 0, &hdlcd_default_mode, 32); + } + + dev_info(hdlcd->dev, "using %dx%d-%d@%d mode\n", hdlcd->fb.var.xres, + hdlcd->fb.var.yres, hdlcd->fb.var.bits_per_pixel, + hdlcd->fb.mode ? hdlcd->fb.mode->refresh : 60); + hdlcd->fb.var.xres_virtual = hdlcd->fb.var.xres; +#ifdef HDLCD_NO_VIRTUAL_SCREEN + hdlcd->fb.var.yres_virtual = hdlcd->fb.var.yres; +#else + hdlcd->fb.var.yres_virtual = hdlcd->fb.var.yres * 2; +#endif + + /* initialise and set the palette */ + if (fb_alloc_cmap(&hdlcd->fb.cmap, NR_PALETTE, 0)) { + dev_err(hdlcd->dev, "failed to allocate cmap memory\n"); + err = -ENOMEM; + goto setup_err; + } + fb_set_cmap(&hdlcd->fb.cmap, &hdlcd->fb); + + /* Allow max number of outstanding requests with the largest beat burst */ + WRITE_HDLCD_REG(HDLCD_REG_BUS_OPTIONS, HDLCD_BUS_MAX_OUTSTAND | HDLCD_BUS_BURST_16); + /* Set the framebuffer base to start of allocated memory */ + WRITE_HDLCD_REG(HDLCD_REG_FB_BASE, hdlcd->fb.fix.smem_start); +#ifdef HDLCD_COUNT_BUFFERUNDERRUNS + /* turn on underrun interrupt for counting */ + WRITE_HDLCD_REG(HDLCD_REG_INT_MASK, HDLCD_INTERRUPT_UNDERRUN); +#else + /* Ensure interrupts are disabled */ + WRITE_HDLCD_REG(HDLCD_REG_INT_MASK, 0); +#endif + fb_set_var(&hdlcd->fb, &hdlcd->fb.var); + + if (!register_framebuffer(&hdlcd->fb)) { + return 0; + } + + dev_err(hdlcd->dev, "HDLCD: cannot register framebuffer\n"); + + fb_dealloc_cmap(&hdlcd->fb.cmap); +setup_err: + iounmap(hdlcd->base); +kmalloc_err: + kfree(hdlcd->fb.pseudo_palette); +remap_err: + clk_unprepare(hdlcd->clk); +clk_prepare_err: + clk_put(hdlcd->clk); + return err; +} + +static inline unsigned char atohex(u8 data) +{ + if (!isxdigit(data)) + return 0; + /* truncate the upper nibble and add 9 to non-digit values */ + return (data > 0x39) ? ((data & 0xf) + 9) : (data & 0xf); +} + +/* EDID data is passed from devicetree in a literal string that can contain spaces and + the hexadecimal dump of the data */ +static int parse_edid_data(struct hdlcd_device *hdlcd, const u8 *edid_data, int data_len) +{ + int i, j; + + if (!edid_data) + return -EINVAL; + + hdlcd->edid = kzalloc(EDID_LENGTH, GFP_KERNEL); + if (!hdlcd->edid) + return -ENOMEM; + + for (i = 0, j = 0; i < data_len; i++) { + if (isspace(edid_data[i])) + continue; + hdlcd->edid[j++] = atohex(edid_data[i]); + if (j >= EDID_LENGTH) + break; + } + + if (j < EDID_LENGTH) { + kfree(hdlcd->edid); + hdlcd->edid = NULL; + return -EINVAL; + } + + return 0; +} + +static int hdlcd_probe(struct platform_device *pdev) +{ + int err = 0, i; + struct hdlcd_device *hdlcd; + struct resource *mem; +#ifdef CONFIG_OF + struct device_node *of_node; +#endif + + memset(&cached_var_screeninfo, 0, sizeof(struct fb_var_screeninfo)); + + dev_dbg(&pdev->dev, "HDLCD: probing\n"); + + hdlcd = kzalloc(sizeof(*hdlcd), GFP_KERNEL); + if (!hdlcd) + return -ENOMEM; + +#ifdef CONFIG_OF + of_node = pdev->dev.of_node; + if (of_node) { + int len; + const u8 *edid; + const __be32 *prop = of_get_property(of_node, "mode", &len); + if (prop) + strncpy(fb_mode, (char *)prop, len); + prop = of_get_property(of_node, "framebuffer", &len); + if (prop) { + hdlcd->fb.fix.smem_start = of_read_ulong(prop, + of_n_addr_cells(of_node)); + prop += of_n_addr_cells(of_node); + framebuffer_size = of_read_ulong(prop, + of_n_size_cells(of_node)); + if (framebuffer_size > HDLCD_MAX_FRAMEBUFFER_SIZE) + framebuffer_size = HDLCD_MAX_FRAMEBUFFER_SIZE; + dev_dbg(&pdev->dev, "HDLCD: phys_addr = 0x%lx, size = 0x%lx\n", + hdlcd->fb.fix.smem_start, framebuffer_size); + } + edid = of_get_property(of_node, "edid", &len); + if (edid) { + err = parse_edid_data(hdlcd, edid, len); +#ifdef CONFIG_SERIAL_AMBA_PCU_UART + } else { + /* ask the firmware to fetch the EDID */ + dev_dbg(&pdev->dev, "HDLCD: Requesting EDID data\n"); + hdlcd->edid = kzalloc(EDID_LENGTH, GFP_KERNEL); + if (!hdlcd->edid) + return -ENOMEM; + err = get_edid(hdlcd->edid); +#endif /* CONFIG_SERIAL_AMBA_PCU_UART */ + } + if (err) + dev_info(&pdev->dev, "HDLCD: Failed to parse EDID data\n"); + } +#endif /* CONFIG_OF */ + + mem = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!mem) { + dev_err(&pdev->dev, "HDLCD: cannot get platform resources\n"); + err = -EINVAL; + goto resource_err; + } + + i = platform_get_irq(pdev, 0); + if (i < 0) { + dev_err(&pdev->dev, "HDLCD: no irq defined for vsync\n"); + err = -ENOENT; + goto resource_err; + } else { + err = request_irq(i, hdlcd_irq, 0, dev_name(&pdev->dev), hdlcd); + if (err) { + dev_err(&pdev->dev, "HDLCD: unable to request irq\n"); + goto resource_err; + } + hdlcd->irq = i; + } + + if (!request_mem_region(mem->start, resource_size(mem), dev_name(&pdev->dev))) { + err = -ENXIO; + goto request_err; + } + + if (!hdlcd->fb.fix.smem_start) { + dev_err(&pdev->dev, "platform did not allocate frame buffer memory\n"); + err = -ENOMEM; + goto memalloc_err; + } + hdlcd->fb.screen_base = ioremap_wc(hdlcd->fb.fix.smem_start, framebuffer_size); + if (!hdlcd->fb.screen_base) { + dev_err(&pdev->dev, "unable to ioremap framebuffer\n"); + err = -ENOMEM; + goto probe_err; + } + + hdlcd->fb.screen_size = framebuffer_size; + hdlcd->fb.fix.smem_len = framebuffer_size; + hdlcd->fb.fix.mmio_start = mem->start; + hdlcd->fb.fix.mmio_len = resource_size(mem); + + /* Clear the framebuffer */ + memset(hdlcd->fb.screen_base, 0, framebuffer_size); + + hdlcd->dev = &pdev->dev; + + dev_dbg(&pdev->dev, "HDLCD: framebuffer virt base %p, phys base 0x%lX\n", + hdlcd->fb.screen_base, (unsigned long)hdlcd->fb.fix.smem_start); + + err = hdlcd_setup(hdlcd); + + if (err) + goto probe_err; + + platform_set_drvdata(pdev, hdlcd); + return 0; + +probe_err: + iounmap(hdlcd->fb.screen_base); + memblock_free(hdlcd->fb.fix.smem_start, hdlcd->fb.fix.smem_start); + +memalloc_err: + release_mem_region(mem->start, resource_size(mem)); + +request_err: + free_irq(hdlcd->irq, hdlcd); + +resource_err: + kfree(hdlcd); + + return err; +} + +static int hdlcd_remove(struct platform_device *pdev) +{ + struct hdlcd_device *hdlcd = platform_get_drvdata(pdev); + + clk_disable(hdlcd->clk); + clk_unprepare(hdlcd->clk); + clk_put(hdlcd->clk); + + /* unmap memory */ + iounmap(hdlcd->fb.screen_base); + iounmap(hdlcd->base); + + /* deallocate fb memory */ + fb_dealloc_cmap(&hdlcd->fb.cmap); + kfree(hdlcd->fb.pseudo_palette); + memblock_free(hdlcd->fb.fix.smem_start, hdlcd->fb.fix.smem_start); + release_mem_region(hdlcd->fb.fix.mmio_start, hdlcd->fb.fix.mmio_len); + + free_irq(hdlcd->irq, NULL); + kfree(hdlcd); + + return 0; +} + +#ifdef CONFIG_PM +static int hdlcd_suspend(struct platform_device *pdev, pm_message_t state) +{ + /* not implemented yet */ + return 0; +} + +static int hdlcd_resume(struct platform_device *pdev) +{ + /* not implemented yet */ + return 0; +} +#else +#define hdlcd_suspend NULL +#define hdlcd_resume NULL +#endif + +static struct platform_driver hdlcd_driver = { + .probe = hdlcd_probe, + .remove = hdlcd_remove, + .suspend = hdlcd_suspend, + .resume = hdlcd_resume, + .driver = { + .name = "hdlcd", + .owner = THIS_MODULE, + .of_match_table = hdlcd_of_matches, + }, +}; + +static int __init hdlcd_init(void) +{ +#ifdef HDLCD_COUNT_BUFFERUNDERRUNS + int err = platform_driver_register(&hdlcd_driver); + if (!err) + hdlcd_underrun_init(); + return err; +#else + return platform_driver_register(&hdlcd_driver); +#endif +} + +void __exit hdlcd_exit(void) +{ +#ifdef HDLCD_COUNT_BUFFERUNDERRUNS + hdlcd_underrun_close(); +#endif + platform_driver_unregister(&hdlcd_driver); +} + +module_init(hdlcd_init); +module_exit(hdlcd_exit); + +MODULE_AUTHOR("Liviu Dudau"); +MODULE_DESCRIPTION("ARM HDLCD core driver"); +MODULE_LICENSE("GPL v2"); diff --git a/drivers/video/console/Kconfig b/drivers/video/console/Kconfig index bc922c47d046..d5fa5f3fe6d1 100644 --- a/drivers/video/console/Kconfig +++ b/drivers/video/console/Kconfig @@ -6,7 +6,7 @@ menu "Console display driver support" config VGA_CONSOLE bool "VGA text console" if EXPERT || !X86 - depends on !4xx && !8xx && !SPARC && !M68K && !PARISC && !FRV && !SUPERH && !BLACKFIN && !AVR32 && !MN10300 && (!ARM || ARCH_FOOTBRIDGE || ARCH_INTEGRATOR || ARCH_NETWINDER) + depends on !4xx && !8xx && !SPARC && !M68K && !PARISC && !FRV && !SUPERH && !BLACKFIN && !AVR32 && !MN10300 && (!ARM || ARCH_FOOTBRIDGE || ARCH_INTEGRATOR || ARCH_NETWINDER) && !ARM64 default y help Saying Y here will allow you to use Linux in text mode through a diff --git a/drivers/video/console/fbcon.c b/drivers/video/console/fbcon.c index a92783e480e6..0d8f98c79a6c 100644 --- a/drivers/video/console/fbcon.c +++ b/drivers/video/console/fbcon.c @@ -404,7 +404,7 @@ static void cursor_timer_handler(unsigned long dev_addr) struct fb_info *info = (struct fb_info *) dev_addr; struct fbcon_ops *ops = info->fbcon_par; - schedule_work(&info->queue); + queue_work(system_power_efficient_wq, &info->queue); mod_timer(&ops->cursor_timer, jiffies + HZ/5); } diff --git a/drivers/video/vexpress-dvi.c b/drivers/video/vexpress-dvi.c new file mode 100644 index 000000000000..f08753450ee4 --- /dev/null +++ b/drivers/video/vexpress-dvi.c @@ -0,0 +1,220 @@ +/* + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * Copyright (C) 2012 ARM Limited + */ + +#define pr_fmt(fmt) "vexpress-dvi: " fmt + +#include <linux/fb.h> +#include <linux/of.h> +#include <linux/of_device.h> +#include <linux/vexpress.h> + + +static struct vexpress_config_func *vexpress_dvimode_func; + +static struct { + u32 xres, yres, mode; +} vexpress_dvi_dvimodes[] = { + { 640, 480, 0 }, /* VGA */ + { 800, 600, 1 }, /* SVGA */ + { 1024, 768, 2 }, /* XGA */ + { 1280, 1024, 3 }, /* SXGA */ + { 1600, 1200, 4 }, /* UXGA */ + { 1920, 1080, 5 }, /* HD1080 */ +}; + +static void vexpress_dvi_mode_set(struct fb_info *info, u32 xres, u32 yres) +{ + int err = -ENOENT; + int i; + + if (!vexpress_dvimode_func) + return; + + for (i = 0; i < ARRAY_SIZE(vexpress_dvi_dvimodes); i++) { + if (vexpress_dvi_dvimodes[i].xres == xres && + vexpress_dvi_dvimodes[i].yres == yres) { + pr_debug("mode: %ux%u = %d\n", xres, yres, + vexpress_dvi_dvimodes[i].mode); + err = vexpress_config_write(vexpress_dvimode_func, 0, + vexpress_dvi_dvimodes[i].mode); + break; + } + } + + if (err) + pr_warn("Failed to set %ux%u mode! (%d)\n", xres, yres, err); +} + + +static struct vexpress_config_func *vexpress_muxfpga_func; +static int vexpress_dvi_fb = -1; + +static int vexpress_dvi_mux_set(struct fb_info *info) +{ + int err; + u32 site = vexpress_get_site_by_dev(info->device); + + if (!vexpress_muxfpga_func) + return -ENXIO; + + err = vexpress_config_write(vexpress_muxfpga_func, 0, site); + if (!err) { + pr_debug("Selected MUXFPGA input %d (fb%d)\n", site, + info->node); + vexpress_dvi_fb = info->node; + vexpress_dvi_mode_set(info, info->var.xres, + info->var.yres); + } else { + pr_warn("Failed to select MUXFPGA input %d (fb%d)! (%d)\n", + site, info->node, err); + } + + return err; +} + +static int vexpress_dvi_fb_select(int fb) +{ + int err; + struct fb_info *info; + + /* fb0 is the default */ + if (fb < 0) + fb = 0; + + info = registered_fb[fb]; + if (!info || !lock_fb_info(info)) + return -ENODEV; + + err = vexpress_dvi_mux_set(info); + + unlock_fb_info(info); + + return err; +} + +static ssize_t vexpress_dvi_fb_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + return sprintf(buf, "%d\n", vexpress_dvi_fb); +} + +static ssize_t vexpress_dvi_fb_store(struct device *dev, + struct device_attribute *attr, const char *buf, size_t count) +{ + long value; + int err = kstrtol(buf, 0, &value); + + if (!err) + err = vexpress_dvi_fb_select(value); + + return err ? err : count; +} + +DEVICE_ATTR(fb, S_IRUGO | S_IWUSR, vexpress_dvi_fb_show, + vexpress_dvi_fb_store); + + +static int vexpress_dvi_fb_event_notify(struct notifier_block *self, + unsigned long action, void *data) +{ + struct fb_event *event = data; + struct fb_info *info = event->info; + struct fb_videomode *mode = event->data; + + switch (action) { + case FB_EVENT_FB_REGISTERED: + if (vexpress_dvi_fb < 0) + vexpress_dvi_mux_set(info); + break; + case FB_EVENT_MODE_CHANGE: + case FB_EVENT_MODE_CHANGE_ALL: + if (info->node == vexpress_dvi_fb) + vexpress_dvi_mode_set(info, mode->xres, mode->yres); + break; + } + + return NOTIFY_OK; +} + +static struct notifier_block vexpress_dvi_fb_notifier = { + .notifier_call = vexpress_dvi_fb_event_notify, +}; +static bool vexpress_dvi_fb_notifier_registered; + + +enum vexpress_dvi_func { FUNC_MUXFPGA, FUNC_DVIMODE }; + +static struct of_device_id vexpress_dvi_of_match[] = { + { + .compatible = "arm,vexpress-muxfpga", + .data = (void *)FUNC_MUXFPGA, + }, { + .compatible = "arm,vexpress-dvimode", + .data = (void *)FUNC_DVIMODE, + }, + {} +}; + +static int vexpress_dvi_probe(struct platform_device *pdev) +{ + enum vexpress_dvi_func func; + const struct of_device_id *match = + of_match_device(vexpress_dvi_of_match, &pdev->dev); + + if (match) + func = (enum vexpress_dvi_func)match->data; + else + func = pdev->id_entry->driver_data; + + switch (func) { + case FUNC_MUXFPGA: + vexpress_muxfpga_func = + vexpress_config_func_get_by_dev(&pdev->dev); + device_create_file(&pdev->dev, &dev_attr_fb); + break; + case FUNC_DVIMODE: + vexpress_dvimode_func = + vexpress_config_func_get_by_dev(&pdev->dev); + break; + } + + if (!vexpress_dvi_fb_notifier_registered) { + fb_register_client(&vexpress_dvi_fb_notifier); + vexpress_dvi_fb_notifier_registered = true; + } + + vexpress_dvi_fb_select(vexpress_dvi_fb); + + return 0; +} + +static const struct platform_device_id vexpress_dvi_id_table[] = { + { .name = "vexpress-muxfpga", .driver_data = FUNC_MUXFPGA, }, + { .name = "vexpress-dvimode", .driver_data = FUNC_DVIMODE, }, + {} +}; + +static struct platform_driver vexpress_dvi_driver = { + .probe = vexpress_dvi_probe, + .driver = { + .name = "vexpress-dvi", + .of_match_table = vexpress_dvi_of_match, + }, + .id_table = vexpress_dvi_id_table, +}; + +static int __init vexpress_dvi_init(void) +{ + return platform_driver_register(&vexpress_dvi_driver); +} +device_initcall(vexpress_dvi_init); |