From 96ad262ff1969ad4b79774c761715a3d50f776c0 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Thu, 10 Sep 2015 14:36:21 +1000 Subject: [PATCH] powerpc/MSI: Fix race condition in tearing down MSI interrupts commit e297c939b745e420ef0b9dc989cb87bda617b399 upstream. This fixes a race which can result in the same virtual IRQ number being assigned to two different MSI interrupts. The most visible consequence of that is usually a warning and stack trace from the sysfs code about an attempt to create a duplicate entry in sysfs. The race happens when one CPU (say CPU 0) is disposing of an MSI while another CPU (say CPU 1) is setting up an MSI. CPU 0 calls (for example) pnv_teardown_msi_irqs(), which calls msi_bitmap_free_hwirqs() to indicate that the MSI (i.e. its hardware IRQ number) is no longer in use. Then, before CPU 0 gets to calling irq_dispose_mapping() to free up the virtal IRQ number, CPU 1 comes in and calls msi_bitmap_alloc_hwirqs() to allocate an MSI, and gets the same hardware IRQ number that CPU 0 just freed. CPU 1 then calls irq_create_mapping() to get a virtual IRQ number, which sees that there is currently a mapping for that hardware IRQ number and returns the corresponding virtual IRQ number (which is the same virtual IRQ number that CPU 0 was using). CPU 0 then calls irq_dispose_mapping() and frees that virtual IRQ number. Now, if another CPU comes along and calls irq_create_mapping(), it is likely to get the virtual IRQ number that was just freed, resulting in the same virtual IRQ number apparently being used for two different hardware interrupts. To fix this race, we just move the call to msi_bitmap_free_hwirqs() to after the call to irq_dispose_mapping(). Since virq_to_hw() doesn't work for the virtual IRQ number after irq_dispose_mapping() has been called, we need to call it before irq_dispose_mapping() and remember the result for the msi_bitmap_free_hwirqs() call. The pattern of calling msi_bitmap_free_hwirqs() before irq_dispose_mapping() appears in 5 places under arch/powerpc, and appears to have originated in commit 05af7bd2d75e ("[POWERPC] MPIC U3/U4 MSI backend") from 2007. Fixes: 05af7bd2d75e ("[POWERPC] MPIC U3/U4 MSI backend") Reported-by: Alexey Kardashevskiy Signed-off-by: Paul Mackerras Signed-off-by: Michael Ellerman [bwh: Backported to 3.2: - powernv uses a private functions instead of msi_bitmap_free_hwirqs() - Adjust filename, context] Signed-off-by: Ben Hutchings --- arch/powerpc/platforms/powernv/pci.c | 4 +++- arch/powerpc/sysdev/fsl_msi.c | 5 +++-- arch/powerpc/sysdev/mpic_pasemi_msi.c | 5 +++-- arch/powerpc/sysdev/mpic_u3msi.c | 5 +++-- arch/powerpc/sysdev/ppc4xx_msi.c | 5 +++-- 5 files changed, 15 insertions(+), 9 deletions(-) diff --git a/arch/powerpc/platforms/powernv/pci.c b/arch/powerpc/platforms/powernv/pci.c index 85bb66d7f933..15a05ca52e70 100644 --- a/arch/powerpc/platforms/powernv/pci.c +++ b/arch/powerpc/platforms/powernv/pci.c @@ -130,6 +130,7 @@ static void pnv_teardown_msi_irqs(struct pci_dev *pdev) struct pci_controller *hose = pci_bus_to_host(pdev->bus); struct pnv_phb *phb = hose->private_data; struct msi_desc *entry; + irq_hw_number_t hwirq; if (WARN_ON(!phb)) return; @@ -137,9 +138,10 @@ static void pnv_teardown_msi_irqs(struct pci_dev *pdev) list_for_each_entry(entry, &pdev->msi_list, list) { if (entry->irq == NO_IRQ) continue; + hwirq = virq_to_hw(entry->irq); irq_set_msi_desc(entry->irq, NULL); - pnv_put_msi(phb, virq_to_hw(entry->irq)); irq_dispose_mapping(entry->irq); + pnv_put_msi(phb, hwirq); } } #endif /* CONFIG_PCI_MSI */ diff --git a/arch/powerpc/sysdev/fsl_msi.c b/arch/powerpc/sysdev/fsl_msi.c index e5c344d336ea..8ebbdc6e82cd 100644 --- a/arch/powerpc/sysdev/fsl_msi.c +++ b/arch/powerpc/sysdev/fsl_msi.c @@ -106,15 +106,16 @@ static void fsl_teardown_msi_irqs(struct pci_dev *pdev) { struct msi_desc *entry; struct fsl_msi *msi_data; + irq_hw_number_t hwirq; list_for_each_entry(entry, &pdev->msi_list, list) { if (entry->irq == NO_IRQ) continue; + hwirq = virq_to_hw(entry->irq); msi_data = irq_get_chip_data(entry->irq); irq_set_msi_desc(entry->irq, NULL); - msi_bitmap_free_hwirqs(&msi_data->bitmap, - virq_to_hw(entry->irq), 1); irq_dispose_mapping(entry->irq); + msi_bitmap_free_hwirqs(&msi_data->bitmap, hwirq, 1); } return; diff --git a/arch/powerpc/sysdev/mpic_pasemi_msi.c b/arch/powerpc/sysdev/mpic_pasemi_msi.c index 38e62382070c..e8736165a7c7 100644 --- a/arch/powerpc/sysdev/mpic_pasemi_msi.c +++ b/arch/powerpc/sysdev/mpic_pasemi_msi.c @@ -74,6 +74,7 @@ static int pasemi_msi_check_device(struct pci_dev *pdev, int nvec, int type) static void pasemi_msi_teardown_msi_irqs(struct pci_dev *pdev) { struct msi_desc *entry; + irq_hw_number_t hwirq; pr_debug("pasemi_msi_teardown_msi_irqs, pdev %p\n", pdev); @@ -81,10 +82,10 @@ static void pasemi_msi_teardown_msi_irqs(struct pci_dev *pdev) if (entry->irq == NO_IRQ) continue; + hwirq = virq_to_hw(entry->irq); irq_set_msi_desc(entry->irq, NULL); - msi_bitmap_free_hwirqs(&msi_mpic->msi_bitmap, - virq_to_hw(entry->irq), ALLOC_CHUNK); irq_dispose_mapping(entry->irq); + msi_bitmap_free_hwirqs(&msi_mpic->msi_bitmap, hwirq, ALLOC_CHUNK); } return; diff --git a/arch/powerpc/sysdev/mpic_u3msi.c b/arch/powerpc/sysdev/mpic_u3msi.c index 9a7aa0ed9c1c..dfc3486bf802 100644 --- a/arch/powerpc/sysdev/mpic_u3msi.c +++ b/arch/powerpc/sysdev/mpic_u3msi.c @@ -124,15 +124,16 @@ static int u3msi_msi_check_device(struct pci_dev *pdev, int nvec, int type) static void u3msi_teardown_msi_irqs(struct pci_dev *pdev) { struct msi_desc *entry; + irq_hw_number_t hwirq; list_for_each_entry(entry, &pdev->msi_list, list) { if (entry->irq == NO_IRQ) continue; + hwirq = virq_to_hw(entry->irq); irq_set_msi_desc(entry->irq, NULL); - msi_bitmap_free_hwirqs(&msi_mpic->msi_bitmap, - virq_to_hw(entry->irq), 1); irq_dispose_mapping(entry->irq); + msi_bitmap_free_hwirqs(&msi_mpic->msi_bitmap, hwirq, 1); } return; diff --git a/arch/powerpc/sysdev/ppc4xx_msi.c b/arch/powerpc/sysdev/ppc4xx_msi.c index 1c2d7af17bbe..4aae9c8dec48 100644 --- a/arch/powerpc/sysdev/ppc4xx_msi.c +++ b/arch/powerpc/sysdev/ppc4xx_msi.c @@ -114,16 +114,17 @@ void ppc4xx_teardown_msi_irqs(struct pci_dev *dev) { struct msi_desc *entry; struct ppc4xx_msi *msi_data = &ppc4xx_msi; + irq_hw_number_t hwirq; dev_dbg(&dev->dev, "PCIE-MSI: tearing down msi irqs\n"); list_for_each_entry(entry, &dev->msi_list, list) { if (entry->irq == NO_IRQ) continue; + hwirq = virq_to_hw(entry->irq); irq_set_msi_desc(entry->irq, NULL); - msi_bitmap_free_hwirqs(&msi_data->bitmap, - virq_to_hw(entry->irq), 1); irq_dispose_mapping(entry->irq); + msi_bitmap_free_hwirqs(&msi_data->bitmap, hwirq, 1); } } -- 2.39.2