From 0e4ccb1505a9e29c50170742ce26ac4655baab2d Mon Sep 17 00:00:00 2001 From: Konrad Rzeszutek Wilk Date: Wed, 6 Nov 2013 16:16:56 -0500 Subject: PCI: Add x86_msi.msi_mask_irq() and msix_mask_irq() Certain platforms do not allow writes in the MSI-X BARs to setup or tear down vector values. To combat against the generic code trying to write to that and either silently being ignored or crashing due to the pagetables being marked R/O this patch introduces a platform override. Note that we keep two separate, non-weak, functions default_mask_msi_irqs() and default_mask_msix_irqs() for the behavior of the arch_mask_msi_irqs() and arch_mask_msix_irqs(), as the default behavior is needed by x86 PCI code. For Xen, which does not allow the guest to write to MSI-X tables - as the hypervisor is solely responsible for setting the vector values - we implement two nops. This fixes a Xen guest crash when passing a PCI device with MSI-X to the guest. See the bugzilla for more details. [bhelgaas: add bugzilla info] Reference: https://bugzilla.kernel.org/show_bug.cgi?id=64581 Signed-off-by: Konrad Rzeszutek Wilk Signed-off-by: Bjorn Helgaas CC: Sucheta Chakraborty CC: Zhenzhong Duan --- arch/x86/include/asm/x86_init.h | 3 +++ arch/x86/kernel/x86_init.c | 10 ++++++++++ arch/x86/pci/xen.c | 13 ++++++++++++- drivers/pci/msi.c | 22 ++++++++++++++++------ include/linux/msi.h | 2 ++ 5 files changed, 43 insertions(+), 7 deletions(-) diff --git a/arch/x86/include/asm/x86_init.h b/arch/x86/include/asm/x86_init.h index 828a1565ba57..0f1be11e43d2 100644 --- a/arch/x86/include/asm/x86_init.h +++ b/arch/x86/include/asm/x86_init.h @@ -172,6 +172,7 @@ struct x86_platform_ops { struct pci_dev; struct msi_msg; +struct msi_desc; struct x86_msi_ops { int (*setup_msi_irqs)(struct pci_dev *dev, int nvec, int type); @@ -182,6 +183,8 @@ struct x86_msi_ops { void (*teardown_msi_irqs)(struct pci_dev *dev); void (*restore_msi_irqs)(struct pci_dev *dev, int irq); int (*setup_hpet_msi)(unsigned int irq, unsigned int id); + u32 (*msi_mask_irq)(struct msi_desc *desc, u32 mask, u32 flag); + u32 (*msix_mask_irq)(struct msi_desc *desc, u32 flag); }; struct IO_APIC_route_entry; diff --git a/arch/x86/kernel/x86_init.c b/arch/x86/kernel/x86_init.c index 8ce0072cd700..021783b1f46a 100644 --- a/arch/x86/kernel/x86_init.c +++ b/arch/x86/kernel/x86_init.c @@ -116,6 +116,8 @@ struct x86_msi_ops x86_msi = { .teardown_msi_irqs = default_teardown_msi_irqs, .restore_msi_irqs = default_restore_msi_irqs, .setup_hpet_msi = default_setup_hpet_msi, + .msi_mask_irq = default_msi_mask_irq, + .msix_mask_irq = default_msix_mask_irq, }; /* MSI arch specific hooks */ @@ -138,6 +140,14 @@ void arch_restore_msi_irqs(struct pci_dev *dev, int irq) { x86_msi.restore_msi_irqs(dev, irq); } +u32 arch_msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag) +{ + return x86_msi.msi_mask_irq(desc, mask, flag); +} +u32 arch_msix_mask_irq(struct msi_desc *desc, u32 flag) +{ + return x86_msi.msix_mask_irq(desc, flag); +} #endif struct x86_io_apic_ops x86_io_apic_ops = { diff --git a/arch/x86/pci/xen.c b/arch/x86/pci/xen.c index 48e8461057ba..5eee4959785d 100644 --- a/arch/x86/pci/xen.c +++ b/arch/x86/pci/xen.c @@ -382,7 +382,14 @@ static void xen_teardown_msi_irq(unsigned int irq) { xen_destroy_irq(irq); } - +static u32 xen_nop_msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag) +{ + return 0; +} +static u32 xen_nop_msix_mask_irq(struct msi_desc *desc, u32 flag) +{ + return 0; +} #endif int __init pci_xen_init(void) @@ -406,6 +413,8 @@ int __init pci_xen_init(void) x86_msi.setup_msi_irqs = xen_setup_msi_irqs; x86_msi.teardown_msi_irq = xen_teardown_msi_irq; x86_msi.teardown_msi_irqs = xen_teardown_msi_irqs; + x86_msi.msi_mask_irq = xen_nop_msi_mask_irq; + x86_msi.msix_mask_irq = xen_nop_msix_mask_irq; #endif return 0; } @@ -485,6 +494,8 @@ int __init pci_xen_initial_domain(void) x86_msi.setup_msi_irqs = xen_initdom_setup_msi_irqs; x86_msi.teardown_msi_irq = xen_teardown_msi_irq; x86_msi.restore_msi_irqs = xen_initdom_restore_msi_irqs; + x86_msi.msi_mask_irq = xen_nop_msi_mask_irq; + x86_msi.msix_mask_irq = xen_nop_msix_mask_irq; #endif xen_setup_acpi_sci(); __acpi_register_gsi = acpi_register_gsi_xen; diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c index 604265c40853..5e63645a7abe 100644 --- a/drivers/pci/msi.c +++ b/drivers/pci/msi.c @@ -185,7 +185,7 @@ static inline __attribute_const__ u32 msi_enabled_mask(u16 control) * reliably as devices without an INTx disable bit will then generate a * level IRQ which will never be cleared. */ -static u32 __msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag) +u32 default_msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag) { u32 mask_bits = desc->masked; @@ -199,9 +199,14 @@ static u32 __msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag) return mask_bits; } +__weak u32 arch_msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag) +{ + return default_msi_mask_irq(desc, mask, flag); +} + static void msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag) { - desc->masked = __msi_mask_irq(desc, mask, flag); + desc->masked = arch_msi_mask_irq(desc, mask, flag); } /* @@ -211,7 +216,7 @@ static void msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag) * file. This saves a few milliseconds when initialising devices with lots * of MSI-X interrupts. */ -static u32 __msix_mask_irq(struct msi_desc *desc, u32 flag) +u32 default_msix_mask_irq(struct msi_desc *desc, u32 flag) { u32 mask_bits = desc->masked; unsigned offset = desc->msi_attrib.entry_nr * PCI_MSIX_ENTRY_SIZE + @@ -224,9 +229,14 @@ static u32 __msix_mask_irq(struct msi_desc *desc, u32 flag) return mask_bits; } +__weak u32 arch_msix_mask_irq(struct msi_desc *desc, u32 flag) +{ + return default_msix_mask_irq(desc, flag); +} + static void msix_mask_irq(struct msi_desc *desc, u32 flag) { - desc->masked = __msix_mask_irq(desc, flag); + desc->masked = arch_msix_mask_irq(desc, flag); } static void msi_set_mask_bit(struct irq_data *data, u32 flag) @@ -902,7 +912,7 @@ void pci_msi_shutdown(struct pci_dev *dev) pci_read_config_word(dev, dev->msi_cap + PCI_MSI_FLAGS, &ctrl); mask = msi_capable_mask(ctrl); /* Keep cached state to be restored */ - __msi_mask_irq(desc, mask, ~mask); + arch_msi_mask_irq(desc, mask, ~mask); /* Restore dev->irq to its default pin-assertion irq */ dev->irq = desc->msi_attrib.default_irq; @@ -998,7 +1008,7 @@ void pci_msix_shutdown(struct pci_dev *dev) /* Return the device with MSI-X masked as initial states */ list_for_each_entry(entry, &dev->msi_list, list) { /* Keep cached states to be restored */ - __msix_mask_irq(entry, 1); + arch_msix_mask_irq(entry, 1); } msix_set_enable(dev, 0); diff --git a/include/linux/msi.h b/include/linux/msi.h index b17ead818aec..87cce50bd121 100644 --- a/include/linux/msi.h +++ b/include/linux/msi.h @@ -64,6 +64,8 @@ void arch_restore_msi_irqs(struct pci_dev *dev, int irq); void default_teardown_msi_irqs(struct pci_dev *dev); void default_restore_msi_irqs(struct pci_dev *dev, int irq); +u32 default_msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag); +u32 default_msix_mask_irq(struct msi_desc *desc, u32 flag); struct msi_chip { struct module *owner; -- cgit v1.2.3 From c6bde215acfd637708142ae671843b6f0eadbc6d Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 6 Nov 2013 10:11:48 -0700 Subject: PCI: Add pci_upstream_bridge() This adds a pci_upstream_bridge() interface to find the PCI-to-PCI bridge upstream from a device. This is typically just "dev->bus->self", but in the case of a VF on a virtual bus, we have to start from the corresponding PF. Returns NULL if there is no upstream PCI bridge, i.e., if the device is on a root bus. Signed-off-by: Bjorn Helgaas Acked-by: Yinghai Lu --- include/linux/pci.h | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/include/linux/pci.h b/include/linux/pci.h index d3a888ae4b2e..835ec7bf6c05 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -480,6 +480,15 @@ static inline bool pci_is_root_bus(struct pci_bus *pbus) return !(pbus->parent); } +static inline struct pci_dev *pci_upstream_bridge(struct pci_dev *dev) +{ + dev = pci_physfn(dev); + if (pci_is_root_bus(dev->bus)) + return NULL; + + return dev->bus->self; +} + #ifdef CONFIG_PCI_MSI static inline bool pci_dev_msi_enabled(struct pci_dev *pci_dev) { -- cgit v1.2.3 From 79272138bd5c35edb9be5ae7b473a2741751a706 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 6 Nov 2013 10:00:51 -0700 Subject: PCI: Enable upstream bridges even for VFs on virtual buses Previously we enabled the upstream PCI-to-PCI bridge only when "dev->bus->self != NULL". In the case of a VF on a virtual bus, where "bus->self == NULL", we didn't enable the upstream bridge. This fixes that by enabling the upstream bridge of the PF corresponding to the VF. Signed-off-by: Bjorn Helgaas Acked-by: Yinghai Lu --- drivers/pci/pci.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index ac40f90cfd5f..d3ed931c0770 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1148,12 +1148,12 @@ int pci_reenable_device(struct pci_dev *dev) static void pci_enable_bridge(struct pci_dev *dev) { + struct pci_dev *bridge; int retval; - if (!dev) - return; - - pci_enable_bridge(dev->bus->self); + bridge = pci_upstream_bridge(dev); + if (bridge) + pci_enable_bridge(bridge); if (pci_is_enabled(dev)) { if (!dev->is_busmaster) @@ -1170,6 +1170,7 @@ static void pci_enable_bridge(struct pci_dev *dev) static int pci_enable_device_flags(struct pci_dev *dev, unsigned long flags) { + struct pci_dev *bridge; int err; int i, bars = 0; @@ -1188,7 +1189,9 @@ static int pci_enable_device_flags(struct pci_dev *dev, unsigned long flags) if (atomic_inc_return(&dev->enable_cnt) > 1) return 0; /* already enabled */ - pci_enable_bridge(dev->bus->self); + bridge = pci_upstream_bridge(dev); + if (bridge) + pci_enable_bridge(bridge); /* only skip sriov related */ for (i = 0; i <= PCI_ROM_RESOURCE; i++) -- cgit v1.2.3