aboutsummaryrefslogtreecommitdiff
path: root/arch/x86
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86')
-rw-r--r--arch/x86/boot/string.c1
-rw-r--r--arch/x86/boot/string.h9
-rw-r--r--arch/x86/crypto/sha1_avx2_x86_64_asm.S67
-rw-r--r--arch/x86/crypto/sha1_ssse3_glue.c2
-rw-r--r--arch/x86/entry/entry_64.S2
-rw-r--r--arch/x86/events/intel/cstate.c2
-rw-r--r--arch/x86/events/intel/rapl.c59
-rw-r--r--arch/x86/events/intel/uncore.c1
-rw-r--r--arch/x86/include/asm/elf.h9
-rw-r--r--arch/x86/include/asm/io.h4
-rw-r--r--arch/x86/include/asm/mmu_context.h4
-rw-r--r--arch/x86/include/asm/xen/hypercall.h3
-rw-r--r--arch/x86/kernel/acpi/boot.c8
-rw-r--r--arch/x86/kernel/apic/io_apic.c2
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_amd.c3
-rw-r--r--arch/x86/kernel/kvm.c6
-rw-r--r--arch/x86/kernel/process_64.c236
-rw-r--r--arch/x86/kvm/cpuid.c2
-rw-r--r--arch/x86/pci/fixup.c32
-rw-r--r--arch/x86/platform/intel-mid/device_libs/Makefile2
-rw-r--r--arch/x86/platform/intel-mid/device_libs/platform_mrfld_spidev.c (renamed from arch/x86/platform/intel-mid/device_libs/platform_spidev.c)4
21 files changed, 274 insertions, 184 deletions
diff --git a/arch/x86/boot/string.c b/arch/x86/boot/string.c
index cc3bd583dce1..9e240fcba784 100644
--- a/arch/x86/boot/string.c
+++ b/arch/x86/boot/string.c
@@ -14,6 +14,7 @@
#include <linux/types.h>
#include "ctype.h"
+#include "string.h"
int memcmp(const void *s1, const void *s2, size_t len)
{
diff --git a/arch/x86/boot/string.h b/arch/x86/boot/string.h
index 725e820602b1..113588ddb43f 100644
--- a/arch/x86/boot/string.h
+++ b/arch/x86/boot/string.h
@@ -18,4 +18,13 @@ int memcmp(const void *s1, const void *s2, size_t len);
#define memset(d,c,l) __builtin_memset(d,c,l)
#define memcmp __builtin_memcmp
+extern int strcmp(const char *str1, const char *str2);
+extern int strncmp(const char *cs, const char *ct, size_t count);
+extern size_t strlen(const char *s);
+extern char *strstr(const char *s1, const char *s2);
+extern size_t strnlen(const char *s, size_t maxlen);
+extern unsigned int atou(const char *s);
+extern unsigned long long simple_strtoull(const char *cp, char **endp,
+ unsigned int base);
+
#endif /* BOOT_STRING_H */
diff --git a/arch/x86/crypto/sha1_avx2_x86_64_asm.S b/arch/x86/crypto/sha1_avx2_x86_64_asm.S
index 1cd792db15ef..1eab79c9ac48 100644
--- a/arch/x86/crypto/sha1_avx2_x86_64_asm.S
+++ b/arch/x86/crypto/sha1_avx2_x86_64_asm.S
@@ -117,11 +117,10 @@
.set T1, REG_T1
.endm
-#define K_BASE %r8
#define HASH_PTR %r9
+#define BLOCKS_CTR %r8
#define BUFFER_PTR %r10
#define BUFFER_PTR2 %r13
-#define BUFFER_END %r11
#define PRECALC_BUF %r14
#define WK_BUF %r15
@@ -205,14 +204,14 @@
* blended AVX2 and ALU instruction scheduling
* 1 vector iteration per 8 rounds
*/
- vmovdqu ((i * 2) + PRECALC_OFFSET)(BUFFER_PTR), W_TMP
+ vmovdqu (i * 2)(BUFFER_PTR), W_TMP
.elseif ((i & 7) == 1)
- vinsertf128 $1, (((i-1) * 2)+PRECALC_OFFSET)(BUFFER_PTR2),\
+ vinsertf128 $1, ((i-1) * 2)(BUFFER_PTR2),\
WY_TMP, WY_TMP
.elseif ((i & 7) == 2)
vpshufb YMM_SHUFB_BSWAP, WY_TMP, WY
.elseif ((i & 7) == 4)
- vpaddd K_XMM(K_BASE), WY, WY_TMP
+ vpaddd K_XMM + K_XMM_AR(%rip), WY, WY_TMP
.elseif ((i & 7) == 7)
vmovdqu WY_TMP, PRECALC_WK(i&~7)
@@ -255,7 +254,7 @@
vpxor WY, WY_TMP, WY_TMP
.elseif ((i & 7) == 7)
vpxor WY_TMP2, WY_TMP, WY
- vpaddd K_XMM(K_BASE), WY, WY_TMP
+ vpaddd K_XMM + K_XMM_AR(%rip), WY, WY_TMP
vmovdqu WY_TMP, PRECALC_WK(i&~7)
PRECALC_ROTATE_WY
@@ -291,7 +290,7 @@
vpsrld $30, WY, WY
vpor WY, WY_TMP, WY
.elseif ((i & 7) == 7)
- vpaddd K_XMM(K_BASE), WY, WY_TMP
+ vpaddd K_XMM + K_XMM_AR(%rip), WY, WY_TMP
vmovdqu WY_TMP, PRECALC_WK(i&~7)
PRECALC_ROTATE_WY
@@ -446,6 +445,16 @@
.endm
+/* Add constant only if (%2 > %3) condition met (uses RTA as temp)
+ * %1 + %2 >= %3 ? %4 : 0
+ */
+.macro ADD_IF_GE a, b, c, d
+ mov \a, RTA
+ add $\d, RTA
+ cmp $\c, \b
+ cmovge RTA, \a
+.endm
+
/*
* macro implements 80 rounds of SHA-1, for multiple blocks with s/w pipelining
*/
@@ -463,13 +472,16 @@
lea (2*4*80+32)(%rsp), WK_BUF
# Precalc WK for first 2 blocks
- PRECALC_OFFSET = 0
+ ADD_IF_GE BUFFER_PTR2, BLOCKS_CTR, 2, 64
.set i, 0
.rept 160
PRECALC i
.set i, i + 1
.endr
- PRECALC_OFFSET = 128
+
+ /* Go to next block if needed */
+ ADD_IF_GE BUFFER_PTR, BLOCKS_CTR, 3, 128
+ ADD_IF_GE BUFFER_PTR2, BLOCKS_CTR, 4, 128
xchg WK_BUF, PRECALC_BUF
.align 32
@@ -479,8 +491,8 @@ _loop:
* we use K_BASE value as a signal of a last block,
* it is set below by: cmovae BUFFER_PTR, K_BASE
*/
- cmp K_BASE, BUFFER_PTR
- jne _begin
+ test BLOCKS_CTR, BLOCKS_CTR
+ jnz _begin
.align 32
jmp _end
.align 32
@@ -512,10 +524,10 @@ _loop0:
.set j, j+2
.endr
- add $(2*64), BUFFER_PTR /* move to next odd-64-byte block */
- cmp BUFFER_END, BUFFER_PTR /* is current block the last one? */
- cmovae K_BASE, BUFFER_PTR /* signal the last iteration smartly */
-
+ /* Update Counter */
+ sub $1, BLOCKS_CTR
+ /* Move to the next block only if needed*/
+ ADD_IF_GE BUFFER_PTR, BLOCKS_CTR, 4, 128
/*
* rounds
* 60,62,64,66,68
@@ -532,8 +544,8 @@ _loop0:
UPDATE_HASH 12(HASH_PTR), D
UPDATE_HASH 16(HASH_PTR), E
- cmp K_BASE, BUFFER_PTR /* is current block the last one? */
- je _loop
+ test BLOCKS_CTR, BLOCKS_CTR
+ jz _loop
mov TB, B
@@ -575,10 +587,10 @@ _loop2:
.set j, j+2
.endr
- add $(2*64), BUFFER_PTR2 /* move to next even-64-byte block */
-
- cmp BUFFER_END, BUFFER_PTR2 /* is current block the last one */
- cmovae K_BASE, BUFFER_PTR /* signal the last iteration smartly */
+ /* update counter */
+ sub $1, BLOCKS_CTR
+ /* Move to the next block only if needed*/
+ ADD_IF_GE BUFFER_PTR2, BLOCKS_CTR, 4, 128
jmp _loop3
_loop3:
@@ -641,19 +653,12 @@ _loop3:
avx2_zeroupper
- lea K_XMM_AR(%rip), K_BASE
-
+ /* Setup initial values */
mov CTX, HASH_PTR
mov BUF, BUFFER_PTR
- lea 64(BUF), BUFFER_PTR2
-
- shl $6, CNT /* mul by 64 */
- add BUF, CNT
- add $64, CNT
- mov CNT, BUFFER_END
- cmp BUFFER_END, BUFFER_PTR2
- cmovae K_BASE, BUFFER_PTR2
+ mov BUF, BUFFER_PTR2
+ mov CNT, BLOCKS_CTR
xmm_mov BSWAP_SHUFB_CTL(%rip), YMM_SHUFB_BSWAP
diff --git a/arch/x86/crypto/sha1_ssse3_glue.c b/arch/x86/crypto/sha1_ssse3_glue.c
index f960a043cdeb..fc61739150e7 100644
--- a/arch/x86/crypto/sha1_ssse3_glue.c
+++ b/arch/x86/crypto/sha1_ssse3_glue.c
@@ -201,7 +201,7 @@ asmlinkage void sha1_transform_avx2(u32 *digest, const char *data,
static bool avx2_usable(void)
{
- if (false && avx_usable() && boot_cpu_has(X86_FEATURE_AVX2)
+ if (avx_usable() && boot_cpu_has(X86_FEATURE_AVX2)
&& boot_cpu_has(X86_FEATURE_BMI1)
&& boot_cpu_has(X86_FEATURE_BMI2))
return true;
diff --git a/arch/x86/entry/entry_64.S b/arch/x86/entry/entry_64.S
index ef766a358b37..e7b0e7ff4c58 100644
--- a/arch/x86/entry/entry_64.S
+++ b/arch/x86/entry/entry_64.S
@@ -1215,6 +1215,8 @@ ENTRY(nmi)
* other IST entries.
*/
+ ASM_CLAC
+
/* Use %rdx as our temp variable throughout */
pushq %rdx
diff --git a/arch/x86/events/intel/cstate.c b/arch/x86/events/intel/cstate.c
index fec8a461bdef..1076c9a77292 100644
--- a/arch/x86/events/intel/cstate.c
+++ b/arch/x86/events/intel/cstate.c
@@ -434,6 +434,7 @@ static struct pmu cstate_core_pmu = {
.stop = cstate_pmu_event_stop,
.read = cstate_pmu_event_update,
.capabilities = PERF_PMU_CAP_NO_INTERRUPT,
+ .module = THIS_MODULE,
};
static struct pmu cstate_pkg_pmu = {
@@ -447,6 +448,7 @@ static struct pmu cstate_pkg_pmu = {
.stop = cstate_pmu_event_stop,
.read = cstate_pmu_event_update,
.capabilities = PERF_PMU_CAP_NO_INTERRUPT,
+ .module = THIS_MODULE,
};
static const struct cstate_model nhm_cstates __initconst = {
diff --git a/arch/x86/events/intel/rapl.c b/arch/x86/events/intel/rapl.c
index 8b902b67342a..4c1b7ea18541 100644
--- a/arch/x86/events/intel/rapl.c
+++ b/arch/x86/events/intel/rapl.c
@@ -161,7 +161,13 @@ static u64 rapl_timer_ms;
static inline struct rapl_pmu *cpu_to_rapl_pmu(unsigned int cpu)
{
- return rapl_pmus->pmus[topology_logical_package_id(cpu)];
+ unsigned int pkgid = topology_logical_package_id(cpu);
+
+ /*
+ * The unsigned check also catches the '-1' return value for non
+ * existent mappings in the topology map.
+ */
+ return pkgid < rapl_pmus->maxpkg ? rapl_pmus->pmus[pkgid] : NULL;
}
static inline u64 rapl_read_counter(struct perf_event *event)
@@ -402,6 +408,8 @@ static int rapl_pmu_event_init(struct perf_event *event)
/* must be done before validate_group */
pmu = cpu_to_rapl_pmu(event->cpu);
+ if (!pmu)
+ return -EINVAL;
event->cpu = pmu->cpu;
event->pmu_private = pmu;
event->hw.event_base = msr;
@@ -585,6 +593,19 @@ static int rapl_cpu_online(unsigned int cpu)
struct rapl_pmu *pmu = cpu_to_rapl_pmu(cpu);
int target;
+ if (!pmu) {
+ pmu = kzalloc_node(sizeof(*pmu), GFP_KERNEL, cpu_to_node(cpu));
+ if (!pmu)
+ return -ENOMEM;
+
+ raw_spin_lock_init(&pmu->lock);
+ INIT_LIST_HEAD(&pmu->active_list);
+ pmu->pmu = &rapl_pmus->pmu;
+ pmu->timer_interval = ms_to_ktime(rapl_timer_ms);
+ rapl_hrtimer_init(pmu);
+
+ rapl_pmus->pmus[topology_logical_package_id(cpu)] = pmu;
+ }
/*
* Check if there is an online cpu in the package which collects rapl
* events already.
@@ -598,27 +619,6 @@ static int rapl_cpu_online(unsigned int cpu)
return 0;
}
-static int rapl_cpu_prepare(unsigned int cpu)
-{
- struct rapl_pmu *pmu = cpu_to_rapl_pmu(cpu);
-
- if (pmu)
- return 0;
-
- pmu = kzalloc_node(sizeof(*pmu), GFP_KERNEL, cpu_to_node(cpu));
- if (!pmu)
- return -ENOMEM;
-
- raw_spin_lock_init(&pmu->lock);
- INIT_LIST_HEAD(&pmu->active_list);
- pmu->pmu = &rapl_pmus->pmu;
- pmu->timer_interval = ms_to_ktime(rapl_timer_ms);
- pmu->cpu = -1;
- rapl_hrtimer_init(pmu);
- rapl_pmus->pmus[topology_logical_package_id(cpu)] = pmu;
- return 0;
-}
-
static int rapl_check_hw_unit(bool apply_quirk)
{
u64 msr_rapl_power_unit_bits;
@@ -697,6 +697,7 @@ static int __init init_rapl_pmus(void)
rapl_pmus->pmu.start = rapl_pmu_event_start;
rapl_pmus->pmu.stop = rapl_pmu_event_stop;
rapl_pmus->pmu.read = rapl_pmu_event_read;
+ rapl_pmus->pmu.module = THIS_MODULE;
return 0;
}
@@ -803,28 +804,21 @@ static int __init rapl_pmu_init(void)
* Install callbacks. Core will call them for each online cpu.
*/
- ret = cpuhp_setup_state(CPUHP_PERF_X86_RAPL_PREP, "PERF_X86_RAPL_PREP",
- rapl_cpu_prepare, NULL);
- if (ret)
- goto out;
-
ret = cpuhp_setup_state(CPUHP_AP_PERF_X86_RAPL_ONLINE,
"AP_PERF_X86_RAPL_ONLINE",
rapl_cpu_online, rapl_cpu_offline);
if (ret)
- goto out1;
+ goto out;
ret = perf_pmu_register(&rapl_pmus->pmu, "power", -1);
if (ret)
- goto out2;
+ goto out1;
rapl_advertise();
return 0;
-out2:
- cpuhp_remove_state(CPUHP_AP_PERF_X86_RAPL_ONLINE);
out1:
- cpuhp_remove_state(CPUHP_PERF_X86_RAPL_PREP);
+ cpuhp_remove_state(CPUHP_AP_PERF_X86_RAPL_ONLINE);
out:
pr_warn("Initialization failed (%d), disabled\n", ret);
cleanup_rapl_pmus();
@@ -835,7 +829,6 @@ module_init(rapl_pmu_init);
static void __exit intel_rapl_exit(void)
{
cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_RAPL_ONLINE);
- cpuhp_remove_state_nocalls(CPUHP_PERF_X86_RAPL_PREP);
perf_pmu_unregister(&rapl_pmus->pmu);
cleanup_rapl_pmus();
}
diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 19d646a783fd..aec6cc925af8 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -733,6 +733,7 @@ static int uncore_pmu_register(struct intel_uncore_pmu *pmu)
.start = uncore_pmu_event_start,
.stop = uncore_pmu_event_stop,
.read = uncore_pmu_event_read,
+ .module = THIS_MODULE,
};
} else {
pmu->pmu = *pmu->type->pmu;
diff --git a/arch/x86/include/asm/elf.h b/arch/x86/include/asm/elf.h
index c152db2ab687..7bcd138c3aa9 100644
--- a/arch/x86/include/asm/elf.h
+++ b/arch/x86/include/asm/elf.h
@@ -204,6 +204,7 @@ void set_personality_ia32(bool);
#define ELF_CORE_COPY_REGS(pr_reg, regs) \
do { \
+ unsigned long base; \
unsigned v; \
(pr_reg)[0] = (regs)->r15; \
(pr_reg)[1] = (regs)->r14; \
@@ -226,8 +227,8 @@ do { \
(pr_reg)[18] = (regs)->flags; \
(pr_reg)[19] = (regs)->sp; \
(pr_reg)[20] = (regs)->ss; \
- (pr_reg)[21] = current->thread.fsbase; \
- (pr_reg)[22] = current->thread.gsbase; \
+ rdmsrl(MSR_FS_BASE, base); (pr_reg)[21] = base; \
+ rdmsrl(MSR_KERNEL_GS_BASE, base); (pr_reg)[22] = base; \
asm("movl %%ds,%0" : "=r" (v)); (pr_reg)[23] = v; \
asm("movl %%es,%0" : "=r" (v)); (pr_reg)[24] = v; \
asm("movl %%fs,%0" : "=r" (v)); (pr_reg)[25] = v; \
@@ -247,11 +248,11 @@ extern int force_personality32;
/*
* This is the base location for PIE (ET_DYN with INTERP) loads. On
- * 64-bit, this is raised to 4GB to leave the entire 32-bit address
+ * 64-bit, this is above 4GB to leave the entire 32-bit address
* space open for things that want to use the area for 32-bit pointers.
*/
#define ELF_ET_DYN_BASE (mmap_is_ia32() ? 0x000400000UL : \
- 0x100000000UL)
+ (TASK_SIZE / 3 * 2))
/* This yields a mask that user programs can use to figure out what
instruction set this CPU supports. This could be done in user space,
diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
index d34bd370074b..6c5020163db0 100644
--- a/arch/x86/include/asm/io.h
+++ b/arch/x86/include/asm/io.h
@@ -304,13 +304,13 @@ static inline unsigned type in##bwl##_p(int port) \
static inline void outs##bwl(int port, const void *addr, unsigned long count) \
{ \
asm volatile("rep; outs" #bwl \
- : "+S"(addr), "+c"(count) : "d"(port)); \
+ : "+S"(addr), "+c"(count) : "d"(port) : "memory"); \
} \
\
static inline void ins##bwl(int port, void *addr, unsigned long count) \
{ \
asm volatile("rep; ins" #bwl \
- : "+D"(addr), "+c"(count) : "d"(port)); \
+ : "+D"(addr), "+c"(count) : "d"(port) : "memory"); \
}
BUILDIO(b, b, char)
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 8e0a9fe86de4..f9dd22469388 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -116,9 +116,7 @@ static inline int init_new_context(struct task_struct *tsk,
mm->context.execute_only_pkey = -1;
}
#endif
- init_new_context_ldt(tsk, mm);
-
- return 0;
+ return init_new_context_ldt(tsk, mm);
}
static inline void destroy_context(struct mm_struct *mm)
{
diff --git a/arch/x86/include/asm/xen/hypercall.h b/arch/x86/include/asm/xen/hypercall.h
index a12a047184ee..8b678af866f7 100644
--- a/arch/x86/include/asm/xen/hypercall.h
+++ b/arch/x86/include/asm/xen/hypercall.h
@@ -43,6 +43,7 @@
#include <asm/page.h>
#include <asm/pgtable.h>
+#include <asm/smap.h>
#include <xen/interface/xen.h>
#include <xen/interface/sched.h>
@@ -214,10 +215,12 @@ privcmd_call(unsigned call,
__HYPERCALL_DECLS;
__HYPERCALL_5ARG(a1, a2, a3, a4, a5);
+ stac();
asm volatile("call *%[call]"
: __HYPERCALL_5PARAM
: [call] "a" (&hypercall_page[call])
: __HYPERCALL_CLOBBER5);
+ clac();
return (long)__res;
}
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 931ced8ca345..d3e0d049a0c2 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -338,6 +338,14 @@ static void __init mp_override_legacy_irq(u8 bus_irq, u8 polarity, u8 trigger,
struct mpc_intsrc mp_irq;
/*
+ * Check bus_irq boundary.
+ */
+ if (bus_irq >= NR_IRQS_LEGACY) {
+ pr_warn("Invalid bus_irq %u for legacy override\n", bus_irq);
+ return;
+ }
+
+ /*
* Convert 'gsi' to 'ioapic.pin'.
*/
ioapic = mp_find_ioapic(gsi);
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 7249f1500bcb..cf89928dbd46 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -2116,7 +2116,7 @@ static inline void __init check_timer(void)
int idx;
idx = find_irq_entry(apic1, pin1, mp_INT);
if (idx != -1 && irq_trigger(idx))
- unmask_ioapic_irq(irq_get_chip_data(0));
+ unmask_ioapic_irq(irq_get_irq_data(0));
}
irq_domain_deactivate_irq(irq_data);
irq_domain_activate_irq(irq_data);
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index 3dfca7b302dc..a5b47c1361a0 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -955,6 +955,9 @@ static int threshold_create_bank(unsigned int cpu, unsigned int bank)
const char *name = get_name(bank, NULL);
int err = 0;
+ if (!dev)
+ return -ENODEV;
+
if (is_shared_bank(bank)) {
nb = node_to_amd_nb(amd_get_nb_id(cpu));
diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
index 9cf697ceedbf..55ffd9dc2258 100644
--- a/arch/x86/kernel/kvm.c
+++ b/arch/x86/kernel/kvm.c
@@ -152,6 +152,8 @@ void kvm_async_pf_task_wait(u32 token)
if (hlist_unhashed(&n.link))
break;
+ rcu_irq_exit();
+
if (!n.halted) {
local_irq_enable();
schedule();
@@ -160,11 +162,11 @@ void kvm_async_pf_task_wait(u32 token)
/*
* We cannot reschedule. So halt.
*/
- rcu_irq_exit();
native_safe_halt();
local_irq_disable();
- rcu_irq_enter();
}
+
+ rcu_irq_enter();
}
if (!n.halted)
finish_swait(&n.wq, &wait);
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index b3760b3c1ca0..0887d2ae3797 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -136,6 +136,123 @@ void release_thread(struct task_struct *dead_task)
}
}
+enum which_selector {
+ FS,
+ GS
+};
+
+/*
+ * Saves the FS or GS base for an outgoing thread if FSGSBASE extensions are
+ * not available. The goal is to be reasonably fast on non-FSGSBASE systems.
+ * It's forcibly inlined because it'll generate better code and this function
+ * is hot.
+ */
+static __always_inline void save_base_legacy(struct task_struct *prev_p,
+ unsigned short selector,
+ enum which_selector which)
+{
+ if (likely(selector == 0)) {
+ /*
+ * On Intel (without X86_BUG_NULL_SEG), the segment base could
+ * be the pre-existing saved base or it could be zero. On AMD
+ * (with X86_BUG_NULL_SEG), the segment base could be almost
+ * anything.
+ *
+ * This branch is very hot (it's hit twice on almost every
+ * context switch between 64-bit programs), and avoiding
+ * the RDMSR helps a lot, so we just assume that whatever
+ * value is already saved is correct. This matches historical
+ * Linux behavior, so it won't break existing applications.
+ *
+ * To avoid leaking state, on non-X86_BUG_NULL_SEG CPUs, if we
+ * report that the base is zero, it needs to actually be zero:
+ * see the corresponding logic in load_seg_legacy.
+ */
+ } else {
+ /*
+ * If the selector is 1, 2, or 3, then the base is zero on
+ * !X86_BUG_NULL_SEG CPUs and could be anything on
+ * X86_BUG_NULL_SEG CPUs. In the latter case, Linux
+ * has never attempted to preserve the base across context
+ * switches.
+ *
+ * If selector > 3, then it refers to a real segment, and
+ * saving the base isn't necessary.
+ */
+ if (which == FS)
+ prev_p->thread.fsbase = 0;
+ else
+ prev_p->thread.gsbase = 0;
+ }
+}
+
+static __always_inline void save_fsgs(struct task_struct *task)
+{
+ savesegment(fs, task->thread.fsindex);
+ savesegment(gs, task->thread.gsindex);
+ save_base_legacy(task, task->thread.fsindex, FS);
+ save_base_legacy(task, task->thread.gsindex, GS);
+}
+
+static __always_inline void loadseg(enum which_selector which,
+ unsigned short sel)
+{
+ if (which == FS)
+ loadsegment(fs, sel);
+ else
+ load_gs_index(sel);
+}
+
+static __always_inline void load_seg_legacy(unsigned short prev_index,
+ unsigned long prev_base,
+ unsigned short next_index,
+ unsigned long next_base,
+ enum which_selector which)
+{
+ if (likely(next_index <= 3)) {
+ /*
+ * The next task is using 64-bit TLS, is not using this
+ * segment at all, or is having fun with arcane CPU features.
+ */
+ if (next_base == 0) {
+ /*
+ * Nasty case: on AMD CPUs, we need to forcibly zero
+ * the base.
+ */
+ if (static_cpu_has_bug(X86_BUG_NULL_SEG)) {
+ loadseg(which, __USER_DS);
+ loadseg(which, next_index);
+ } else {
+ /*
+ * We could try to exhaustively detect cases
+ * under which we can skip the segment load,
+ * but there's really only one case that matters
+ * for performance: if both the previous and
+ * next states are fully zeroed, we can skip
+ * the load.
+ *
+ * (This assumes that prev_base == 0 has no
+ * false positives. This is the case on
+ * Intel-style CPUs.)
+ */
+ if (likely(prev_index | next_index | prev_base))
+ loadseg(which, next_index);
+ }
+ } else {
+ if (prev_index != next_index)
+ loadseg(which, next_index);
+ wrmsrl(which == FS ? MSR_FS_BASE : MSR_KERNEL_GS_BASE,
+ next_base);
+ }
+ } else {
+ /*
+ * The next task is using a real segment. Loading the selector
+ * is sufficient.
+ */
+ loadseg(which, next_index);
+ }
+}
+
int copy_thread_tls(unsigned long clone_flags, unsigned long sp,
unsigned long arg, struct task_struct *p, unsigned long tls)
{
@@ -216,10 +333,19 @@ start_thread_common(struct pt_regs *regs, unsigned long new_ip,
unsigned long new_sp,
unsigned int _cs, unsigned int _ss, unsigned int _ds)
{
+ WARN_ON_ONCE(regs != current_pt_regs());
+
+ if (static_cpu_has(X86_BUG_NULL_SEG)) {
+ /* Loading zero below won't clear the base. */
+ loadsegment(fs, __USER_DS);
+ load_gs_index(__USER_DS);
+ }
+
loadsegment(fs, 0);
loadsegment(es, _ds);
loadsegment(ds, _ds);
load_gs_index(0);
+
regs->ip = new_ip;
regs->sp = new_sp;
regs->cs = _cs;
@@ -264,7 +390,6 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
struct fpu *next_fpu = &next->fpu;
int cpu = smp_processor_id();
struct tss_struct *tss = &per_cpu(cpu_tss, cpu);
- unsigned prev_fsindex, prev_gsindex;
fpu_switch_t fpu_switch;
fpu_switch = switch_fpu_prepare(prev_fpu, next_fpu, cpu);
@@ -274,8 +399,7 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
*
* (e.g. xen_load_tls())
*/
- savesegment(fs, prev_fsindex);
- savesegment(gs, prev_gsindex);
+ save_fsgs(prev_p);
/*
* Load TLS before restoring any segments so that segment loads
@@ -314,108 +438,10 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
if (unlikely(next->ds | prev->ds))
loadsegment(ds, next->ds);
- /*
- * Switch FS and GS.
- *
- * These are even more complicated than DS and ES: they have
- * 64-bit bases are that controlled by arch_prctl. The bases
- * don't necessarily match the selectors, as user code can do
- * any number of things to cause them to be inconsistent.
- *
- * We don't promise to preserve the bases if the selectors are
- * nonzero. We also don't promise to preserve the base if the
- * selector is zero and the base doesn't match whatever was
- * most recently passed to ARCH_SET_FS/GS. (If/when the
- * FSGSBASE instructions are enabled, we'll need to offer
- * stronger guarantees.)
- *
- * As an invariant,
- * (fsbase != 0 && fsindex != 0) || (gsbase != 0 && gsindex != 0) is
- * impossible.
- */
- if (next->fsindex) {
- /* Loading a nonzero value into FS sets the index and base. */
- loadsegment(fs, next->fsindex);
- } else {
- if (next->fsbase) {
- /* Next index is zero but next base is nonzero. */
- if (prev_fsindex)
- loadsegment(fs, 0);
- wrmsrl(MSR_FS_BASE, next->fsbase);
- } else {
- /* Next base and index are both zero. */
- if (static_cpu_has_bug(X86_BUG_NULL_SEG)) {
- /*
- * We don't know the previous base and can't
- * find out without RDMSR. Forcibly clear it.
- */
- loadsegment(fs, __USER_DS);
- loadsegment(fs, 0);
- } else {
- /*
- * If the previous index is zero and ARCH_SET_FS
- * didn't change the base, then the base is
- * also zero and we don't need to do anything.
- */
- if (prev->fsbase || prev_fsindex)
- loadsegment(fs, 0);
- }
- }
- }
- /*
- * Save the old state and preserve the invariant.
- * NB: if prev_fsindex == 0, then we can't reliably learn the base
- * without RDMSR because Intel user code can zero it without telling
- * us and AMD user code can program any 32-bit value without telling
- * us.
- */
- if (prev_fsindex)
- prev->fsbase = 0;
- prev->fsindex = prev_fsindex;
-
- if (next->gsindex) {
- /* Loading a nonzero value into GS sets the index and base. */
- load_gs_index(next->gsindex);
- } else {
- if (next->gsbase) {
- /* Next index is zero but next base is nonzero. */
- if (prev_gsindex)
- load_gs_index(0);
- wrmsrl(MSR_KERNEL_GS_BASE, next->gsbase);
- } else {
- /* Next base and index are both zero. */
- if (static_cpu_has_bug(X86_BUG_NULL_SEG)) {
- /*
- * We don't know the previous base and can't
- * find out without RDMSR. Forcibly clear it.
- *
- * This contains a pointless SWAPGS pair.
- * Fixing it would involve an explicit check
- * for Xen or a new pvop.
- */
- load_gs_index(__USER_DS);
- load_gs_index(0);
- } else {
- /*
- * If the previous index is zero and ARCH_SET_GS
- * didn't change the base, then the base is
- * also zero and we don't need to do anything.
- */
- if (prev->gsbase || prev_gsindex)
- load_gs_index(0);
- }
- }
- }
- /*
- * Save the old state and preserve the invariant.
- * NB: if prev_gsindex == 0, then we can't reliably learn the base
- * without RDMSR because Intel user code can zero it without telling
- * us and AMD user code can program any 32-bit value without telling
- * us.
- */
- if (prev_gsindex)
- prev->gsbase = 0;
- prev->gsindex = prev_gsindex;
+ load_seg_legacy(prev->fsindex, prev->fsbase,
+ next->fsindex, next->fsbase, FS);
+ load_seg_legacy(prev->gsindex, prev->gsbase,
+ next->gsindex, next->gsbase, GS);
switch_fpu_finish(next_fpu, fpu_switch);
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index 649d8f2c1e40..91af75e37306 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -456,7 +456,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
entry->ecx &= kvm_cpuid_7_0_ecx_x86_features;
cpuid_mask(&entry->ecx, CPUID_7_ECX);
/* PKU is not yet implemented for shadow paging. */
- if (!tdp_enabled)
+ if (!tdp_enabled || !boot_cpu_has(X86_FEATURE_OSPKE))
entry->ecx &= ~F(PKU);
} else {
entry->ebx = 0;
diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c
index 6d52b94f4bb9..20fa7c84109d 100644
--- a/arch/x86/pci/fixup.c
+++ b/arch/x86/pci/fixup.c
@@ -571,3 +571,35 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2fc0, pci_invalid_bar);
DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x6f60, pci_invalid_bar);
DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x6fa0, pci_invalid_bar);
DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x6fc0, pci_invalid_bar);
+
+/*
+ * Apple MacBook Pro: Avoid [mem 0x7fa00000-0x7fbfffff]
+ *
+ * Using the [mem 0x7fa00000-0x7fbfffff] region, e.g., by assigning it to
+ * the 00:1c.0 Root Port, causes a conflict with [io 0x1804], which is used
+ * for soft poweroff and suspend-to-RAM.
+ *
+ * As far as we know, this is related to the address space, not to the Root
+ * Port itself. Attaching the quirk to the Root Port is a convenience, but
+ * it could probably also be a standalone DMI quirk.
+ *
+ * https://bugzilla.kernel.org/show_bug.cgi?id=103211
+ */
+static void quirk_apple_mbp_poweroff(struct pci_dev *pdev)
+{
+ struct device *dev = &pdev->dev;
+ struct resource *res;
+
+ if ((!dmi_match(DMI_PRODUCT_NAME, "MacBookPro11,4") &&
+ !dmi_match(DMI_PRODUCT_NAME, "MacBookPro11,5")) ||
+ pdev->bus->number != 0 || pdev->devfn != PCI_DEVFN(0x1c, 0))
+ return;
+
+ res = request_mem_region(0x7fa00000, 0x200000,
+ "MacBook Pro poweroff workaround");
+ if (res)
+ dev_info(dev, "claimed %s %pR\n", res->name, res);
+ else
+ dev_info(dev, "can't work around MacBook Pro poweroff issue\n");
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x8c10, quirk_apple_mbp_poweroff);
diff --git a/arch/x86/platform/intel-mid/device_libs/Makefile b/arch/x86/platform/intel-mid/device_libs/Makefile
index dd6cfa4ad3ac..75029d0cfa15 100644
--- a/arch/x86/platform/intel-mid/device_libs/Makefile
+++ b/arch/x86/platform/intel-mid/device_libs/Makefile
@@ -15,7 +15,7 @@ obj-$(subst m,y,$(CONFIG_INTEL_MID_POWER_BUTTON)) += platform_msic_power_btn.o
obj-$(subst m,y,$(CONFIG_GPIO_INTEL_PMIC)) += platform_pmic_gpio.o
obj-$(subst m,y,$(CONFIG_INTEL_MFLD_THERMAL)) += platform_msic_thermal.o
# SPI Devices
-obj-$(subst m,y,$(CONFIG_SPI_SPIDEV)) += platform_spidev.o
+obj-$(subst m,y,$(CONFIG_SPI_SPIDEV)) += platform_mrfld_spidev.o
# I2C Devices
obj-$(subst m,y,$(CONFIG_SENSORS_EMC1403)) += platform_emc1403.o
obj-$(subst m,y,$(CONFIG_SENSORS_LIS3LV02D)) += platform_lis331.o
diff --git a/arch/x86/platform/intel-mid/device_libs/platform_spidev.c b/arch/x86/platform/intel-mid/device_libs/platform_mrfld_spidev.c
index 30c601b399ee..27186ad654c9 100644
--- a/arch/x86/platform/intel-mid/device_libs/platform_spidev.c
+++ b/arch/x86/platform/intel-mid/device_libs/platform_mrfld_spidev.c
@@ -11,6 +11,7 @@
* of the License.
*/
+#include <linux/err.h>
#include <linux/init.h>
#include <linux/sfi.h>
#include <linux/spi/pxa2xx_spi.h>
@@ -34,6 +35,9 @@ static void __init *spidev_platform_data(void *info)
{
struct spi_board_info *spi_info = info;
+ if (intel_mid_identify_cpu() != INTEL_MID_CPU_CHIP_TANGIER)
+ return ERR_PTR(-ENODEV);
+
spi_info->mode = SPI_MODE_0;
spi_info->controller_data = &spidev_spi_chip;
generated by cgit v1.2.3 (git 2.25.1) at 2024-04-27 07:16:23 +0000