From 6c2e94033df5ca11149e52dd179b8dde3172e9bf Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 7 Nov 2008 12:33:49 +0100 Subject: x86: apic honour irq affinity which was set in early boot setup_ioapic_dest() is called after the non boot cpus have been brought up. It sets the irq affinity of all already configured interrupts to all cpus and ignores affinity settings which were done by the early bootup code. If the IRQ_NO_BALANCING or IRQ_AFFINITY_SET flags are set then use the affinity mask from the irq descriptor and not TARGET_CPUS. Signed-off-by: Thomas Gleixner Signed-off-by: Ingo Molnar --- arch/x86/kernel/io_apic.c | 26 +++++++++++++++++++++----- 1 file changed, 21 insertions(+), 5 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 7a3f2028e2e..988ee89467d 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -3761,7 +3761,9 @@ int acpi_get_override_irq(int bus_irq, int *trigger, int *polarity) void __init setup_ioapic_dest(void) { int pin, ioapic, irq, irq_entry; + struct irq_desc *desc; struct irq_cfg *cfg; + cpumask_t mask; if (skip_ioapic_setup == 1) return; @@ -3778,16 +3780,30 @@ void __init setup_ioapic_dest(void) * cpu is online. */ cfg = irq_cfg(irq); - if (!cfg->vector) + if (!cfg->vector) { setup_IO_APIC_irq(ioapic, pin, irq, irq_trigger(irq_entry), irq_polarity(irq_entry)); + continue; + + } + + /* + * Honour affinities which have been set in early boot + */ + desc = irq_to_desc(irq); + if (desc->status & + (IRQ_NO_BALANCING | IRQ_AFFINITY_SET)) + mask = desc->affinity; + else + mask = TARGET_CPUS; + #ifdef CONFIG_INTR_REMAP - else if (intr_remapping_enabled) - set_ir_ioapic_affinity_irq(irq, TARGET_CPUS); -#endif + if (intr_remapping_enabled) + set_ir_ioapic_affinity_irq(irq, mask); else - set_ioapic_affinity_irq(irq, TARGET_CPUS); +#endif + set_ioapic_affinity_irq(irq, mask); } } -- cgit v1.2.2 From d1f1e9c01006b4b050e090055c75278f80c2a5c5 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Sat, 15 Nov 2008 11:00:17 +0100 Subject: x86, bts: fix unlock problem in ds.c Fix a problem where ds_request() returned an error without releasing the ds lock. Reported-by: Stephane Eranian Signed-off-by: Markus Metzger Cc: Signed-off-by: Ingo Molnar --- arch/x86/kernel/ds.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 2b69994fd3a..ac1d5b0586b 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -384,8 +384,9 @@ static int ds_request(struct task_struct *task, void *base, size_t size, spin_lock(&ds_lock); + error = -EPERM; if (!check_tracer(task)) - return -EPERM; + goto out_unlock; error = -ENOMEM; context = ds_alloc_context(task); -- cgit v1.2.2 From d3c6aa1e69f705ac3ab64584101b1d38435b1353 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sun, 16 Nov 2008 00:49:31 -0800 Subject: x86: fix es7000 compiling Impact: fix es7000 build CC arch/x86/kernel/es7000_32.o arch/x86/kernel/es7000_32.c: In function find_unisys_acpi_oem_table: arch/x86/kernel/es7000_32.c:255: error: implicit declaration of function acpi_get_table_with_size arch/x86/kernel/es7000_32.c:261: error: implicit declaration of function early_acpi_os_unmap_memory arch/x86/kernel/es7000_32.c: In function unmap_unisys_acpi_oem_table: arch/x86/kernel/es7000_32.c:277: error: implicit declaration of function __acpi_unmap_table make[1]: *** [arch/x86/kernel/es7000_32.o] Error 1 we applied one patch out of order... | commit a73aaedd95703bd49f4c3f9df06fb7b7373ba905 | Author: Yinghai Lu | Date: Sun Sep 14 02:33:14 2008 -0700 | | x86: check dsdt before find oem table for es7000, v2 | | v2: use __acpi_unmap_table() that patch need: x86: use early_ioremap in __acpi_map_table x86: always explicitly map acpi memory acpi: remove final __acpi_map_table mapping before setting acpi_gbl_permanent_mmap acpi/x86: introduce __apci_map_table, v4 submitted to the ACPI tree but not upstream yet. fix it until those patches applied, need to revert this one Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/es7000_32.c | 9 +-------- 1 file changed, 1 insertion(+), 8 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/es7000_32.c b/arch/x86/kernel/es7000_32.c index f454c78fcef..0aa2c443d60 100644 --- a/arch/x86/kernel/es7000_32.c +++ b/arch/x86/kernel/es7000_32.c @@ -250,31 +250,24 @@ int __init find_unisys_acpi_oem_table(unsigned long *oem_addr) { struct acpi_table_header *header = NULL; int i = 0; - acpi_size tbl_size; - while (ACPI_SUCCESS(acpi_get_table_with_size("OEM1", i++, &header, &tbl_size))) { + while (ACPI_SUCCESS(acpi_get_table("OEM1", i++, &header))) { if (!memcmp((char *) &header->oem_id, "UNISYS", 6)) { struct oem_table *t = (struct oem_table *)header; oem_addrX = t->OEMTableAddr; oem_size = t->OEMTableSize; - early_acpi_os_unmap_memory(header, tbl_size); *oem_addr = (unsigned long)__acpi_map_table(oem_addrX, oem_size); return 0; } - early_acpi_os_unmap_memory(header, tbl_size); } return -1; } void __init unmap_unisys_acpi_oem_table(unsigned long oem_addr) { - if (!oem_addr) - return; - - __acpi_unmap_table((char *)oem_addr, oem_size); } #endif -- cgit v1.2.2 From 93ce99e849433ede4ce8b410b749dc0cad1100b2 Mon Sep 17 00:00:00 2001 From: Venki Pallipadi Date: Mon, 17 Nov 2008 14:43:58 -0800 Subject: x86: add rdtsc barrier to TSC sync check Impact: fix incorrectly marked unstable TSC clock Patch (commit 0d12cdd "sched: improve sched_clock() performance") has a regression on one of the test systems here. With the patch, I see: checking TSC synchronization [CPU#0 -> CPU#1]: Measured 28 cycles TSC warp between CPUs, turning off TSC clock. Marking TSC unstable due to check_tsc_sync_source failed Whereas, without the patch syncs pass fine on all CPUs: checking TSC synchronization [CPU#0 -> CPU#1]: passed. Due to this, TSC is marked unstable, when it is not actually unstable. This is because syncs in check_tsc_wrap() goes away due to this commit. As per the discussion on this thread, correct way to fix this is to add explicit syncs as below? Signed-off-by: Venkatesh Pallipadi Signed-off-by: Ingo Molnar --- arch/x86/kernel/tsc_sync.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/tsc_sync.c b/arch/x86/kernel/tsc_sync.c index 9ffb01c31c4..1c0dfbca87c 100644 --- a/arch/x86/kernel/tsc_sync.c +++ b/arch/x86/kernel/tsc_sync.c @@ -46,7 +46,9 @@ static __cpuinit void check_tsc_warp(void) cycles_t start, now, prev, end; int i; + rdtsc_barrier(); start = get_cycles(); + rdtsc_barrier(); /* * The measurement runs for 20 msecs: */ @@ -61,7 +63,9 @@ static __cpuinit void check_tsc_warp(void) */ __raw_spin_lock(&sync_lock); prev = last_tsc; + rdtsc_barrier(); now = get_cycles(); + rdtsc_barrier(); last_tsc = now; __raw_spin_unlock(&sync_lock); -- cgit v1.2.2 From 10db4ef7b9a65b86e4d047671a1886f4c101a859 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 18 Nov 2008 15:23:08 +0100 Subject: x86, PEBS/DS: fix code flow in ds_request() this compiler warning: arch/x86/kernel/ds.c: In function 'ds_request': arch/x86/kernel/ds.c:368: warning: 'context' may be used uninitialized in this function Shows that the code flow in ds_request() is buggy - it goes into the unlock+release-context path even when the context is not allocated yet. First allocate the context, then do the other checks. Also, take care with GFP allocations under the ds_lock spinlock. Cc: Signed-off-by: Ingo Molnar --- arch/x86/kernel/ds.c | 26 +++++++++++++++++++++----- 1 file changed, 21 insertions(+), 5 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index ac1d5b0586b..d1a121443bd 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -236,17 +236,33 @@ static inline struct ds_context *ds_alloc_context(struct task_struct *task) struct ds_context *context = *p_context; if (!context) { + spin_unlock(&ds_lock); + context = kzalloc(sizeof(*context), GFP_KERNEL); - if (!context) + if (!context) { + spin_lock(&ds_lock); return NULL; + } context->ds = kzalloc(ds_cfg.sizeof_ds, GFP_KERNEL); if (!context->ds) { kfree(context); + spin_lock(&ds_lock); return NULL; } + spin_lock(&ds_lock); + /* + * Check for race - another CPU could have allocated + * it meanwhile: + */ + if (*p_context) { + kfree(context->ds); + kfree(context); + return *p_context; + } + *p_context = context; context->this = p_context; @@ -384,15 +400,15 @@ static int ds_request(struct task_struct *task, void *base, size_t size, spin_lock(&ds_lock); - error = -EPERM; - if (!check_tracer(task)) - goto out_unlock; - error = -ENOMEM; context = ds_alloc_context(task); if (!context) goto out_unlock; + error = -EPERM; + if (!check_tracer(task)) + goto out_unlock; + error = -EALREADY; if (context->owner[qual] == current) goto out_unlock; -- cgit v1.2.2 From e5e1f606ecbf67e52ebe2df5d14f8b94ec6544d0 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Mon, 17 Nov 2008 15:07:17 +0100 Subject: AMD IOMMU: add parameter to disable device isolation Impact: add a new AMD IOMMU kernel command line parameter Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu_init.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 0cdcda35a05..838a2e1d5bb 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -1213,6 +1213,8 @@ static int __init parse_amd_iommu_options(char *str) for (; *str; ++str) { if (strncmp(str, "isolate", 7) == 0) amd_iommu_isolate = 1; + if (strncmp(str, "share", 5) == 0) + amd_iommu_isolate = 0; if (strncmp(str, "fullflush", 11) == 0) amd_iommu_unmap_flush = true; } -- cgit v1.2.2 From 3ce1f93c6d53c3f91c3846cf66b018276c8ac2e7 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Mon, 17 Nov 2008 15:09:20 +0100 Subject: AMD IOMMU: enable device isolation per default Impact: makes device isolation the default for AMD IOMMU Some device drivers showed double-free bugs of DMA memory while testing them with AMD IOMMU. If all devices share the same protection domain this can lead to data corruption and data loss. Prevent this by putting each device into its own protection domain per default. Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu_init.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 838a2e1d5bb..595edd2befc 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -121,7 +121,7 @@ u16 amd_iommu_last_bdf; /* largest PCI device id we have LIST_HEAD(amd_iommu_unity_map); /* a list of required unity mappings we find in ACPI */ unsigned amd_iommu_aperture_order = 26; /* size of aperture in power of 2 */ -int amd_iommu_isolate; /* if 1, device isolation is enabled */ +int amd_iommu_isolate = 1; /* if 1, device isolation is enabled */ bool amd_iommu_unmap_flush; /* if true, flush on every unmap */ LIST_HEAD(amd_iommu_list); /* list of all AMD IOMMUs in the -- cgit v1.2.2 From 695b5676c727d80921a2dc8737d5b3322222db85 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Mon, 17 Nov 2008 15:16:43 +0100 Subject: AMD IOMMU: fix fullflush comparison length Impact: fix comparison length for 'fullflush' Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu_init.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 595edd2befc..30ae2701b3d 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -1215,7 +1215,7 @@ static int __init parse_amd_iommu_options(char *str) amd_iommu_isolate = 1; if (strncmp(str, "share", 5) == 0) amd_iommu_isolate = 0; - if (strncmp(str, "fullflush", 11) == 0) + if (strncmp(str, "fullflush", 9) == 0) amd_iommu_unmap_flush = true; } -- cgit v1.2.2 From 8501c45cc32c311ae755a2d5ac8c4a5f04908d42 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Mon, 17 Nov 2008 19:11:46 +0100 Subject: AMD IOMMU: check for next_bit also in unmapped area Impact: fix possible use of stale IO/TLB entries Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 331b318304e..e4899e0e878 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -537,7 +537,7 @@ static void dma_ops_free_addresses(struct dma_ops_domain *dom, address >>= PAGE_SHIFT; iommu_area_free(dom->bitmap, address, pages); - if (address + pages >= dom->next_bit) + if (address >= dom->next_bit) dom->need_flush = true; } -- cgit v1.2.2 From 0af40a4b1050c050e62eb1dc30b82d5ab22bf221 Mon Sep 17 00:00:00 2001 From: Philipp Kohlbecher Date: Sun, 16 Nov 2008 12:11:01 +0100 Subject: x86: more general identifier for Phoenix BIOS Impact: widen the reach of the low-memory-protect DMI quirk Phoenix BIOSes variously identify their vendor as "Phoenix Technologies, LTD" or "Phoenix Technologies LTD" (without the comma.) This patch makes the identification string in the bad_bios_dmi_table more general (following a suggestion by Ingo Molnar), so that both versions are handled. Again, the patched file compiles cleanly and the patch has been tested successfully on my machine. Signed-off-by: Philipp Kohlbecher Cc: Signed-off-by: Ingo Molnar --- arch/x86/kernel/setup.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 0fa6790c1dd..9d5674f7b6c 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -764,7 +764,7 @@ static struct dmi_system_id __initdata bad_bios_dmi_table[] = { .callback = dmi_low_memory_corruption, .ident = "Phoenix BIOS", .matches = { - DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies, LTD"), + DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies"), }, }, #endif -- cgit v1.2.2 From 093bac154c142fa1fb31a3ac69ae1bc08930231b Mon Sep 17 00:00:00 2001 From: Steve Conklin Date: Fri, 14 Nov 2008 00:55:51 -0600 Subject: x86: quirk for reboot stalls on a Dell Optiplex 330 Dell Optiplex 330 appears to hang on reboot. This is resolved by adding a quirk to set bios reboot. Signed-off-by: Leann Ogasawara Signed-off-by: Steve Conklin Signed-off-by: Ingo Molnar --- arch/x86/kernel/reboot.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 724adfc63cb..cc5a2545dd4 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -169,6 +169,15 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = { DMI_MATCH(DMI_BOARD_NAME, "0KW626"), }, }, + { /* Handle problems with rebooting on Dell Optiplex 330 with 0KP561 */ + .callback = set_bios_reboot, + .ident = "Dell OptiPlex 330", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex 330"), + DMI_MATCH(DMI_BOARD_NAME, "0KP561"), + }, + }, { /* Handle problems with rebooting on Dell 2400's */ .callback = set_bios_reboot, .ident = "Dell PowerEdge 2400", -- cgit v1.2.2 From 9bc646f163b136684390081262fab0fd8f5343ca Mon Sep 17 00:00:00 2001 From: Rakib Mullick Date: Thu, 20 Nov 2008 19:08:45 +0600 Subject: x86: fix __cpuinit/__init tangle in init_thread_xstate() Impact: fix incorrect __init annotation This patch removes the following section mismatch warning. A patch set was send previously (http://lkml.org/lkml/2008/11/10/407). But introduce some other problem, reported by Rufus (http://lkml.org/lkml/2008/11/11/46). Then Ingo Molnar suggest that, it's best to remove __init from xsave_cntxt_init(void). Which is the second patch in this series. Now, this one removes the following warning. WARNING: arch/x86/kernel/built-in.o(.cpuinit.text+0x2237): Section mismatch in reference from the function cpu_init() to the function .init.text:init_thread_xstate() The function __cpuinit cpu_init() references a function __init init_thread_xstate(). If init_thread_xstate is only used by cpu_init then annotate init_thread_xstate with a matching annotation. Signed-off-by: Rakib Mullick Signed-off-by: Ingo Molnar --- arch/x86/kernel/i387.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c index 1f20608d4ca..b0f61f0dcd0 100644 --- a/arch/x86/kernel/i387.c +++ b/arch/x86/kernel/i387.c @@ -58,7 +58,7 @@ void __cpuinit mxcsr_feature_mask_init(void) stts(); } -void __init init_thread_xstate(void) +void __cpuinit init_thread_xstate(void) { if (!HAVE_HWFP) { xstate_size = sizeof(struct i387_soft_struct); -- cgit v1.2.2 From bfe085f62f98a49e1b864e4950389c7205174e4f Mon Sep 17 00:00:00 2001 From: Rakib Mullick Date: Thu, 20 Nov 2008 19:12:50 +0600 Subject: x86: fixing __cpuinit/__init tangle, xsave_cntxt_init() Annotate xsave_cntxt_init() as "can be called outside of __init". Signed-off-by: Rakib Mullick Signed-off-by: Ingo Molnar --- arch/x86/kernel/xsave.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/xsave.c b/arch/x86/kernel/xsave.c index b13acb75e82..15c3e699918 100644 --- a/arch/x86/kernel/xsave.c +++ b/arch/x86/kernel/xsave.c @@ -310,7 +310,7 @@ static void __init setup_xstate_init(void) /* * Enable and initialize the xsave feature. */ -void __init xsave_cntxt_init(void) +void __ref xsave_cntxt_init(void) { unsigned int eax, ebx, ecx, edx; -- cgit v1.2.2 From 0ca4b6b00113b064c080d26d803d0d7c80fb5dc8 Mon Sep 17 00:00:00 2001 From: Matthew Wilcox Date: Thu, 20 Nov 2008 14:09:33 -0700 Subject: x86: Fix interrupt leak due to migration When we migrate an interrupt from one CPU to another, we set the move_in_progress flag and clean up the vectors later once they're not being used. If you're unlucky and call destroy_irq() before the vectors become un-used, the move_in_progress flag is never cleared, which causes the interrupt to become unusable. This was discovered by Jesse Brandeburg for whom it manifested as an MSI-X device refusing to use MSI-X mode when the driver was unloaded and reloaded repeatedly. Signed-off-by: Matthew Wilcox Signed-off-by: Linus Torvalds --- arch/x86/kernel/io_apic.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 7a3f2028e2e..c9513e1ff28 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -1140,6 +1140,20 @@ static void __clear_irq_vector(int irq) cfg->vector = 0; cpus_clear(cfg->domain); + + if (likely(!cfg->move_in_progress)) + return; + cpus_and(mask, cfg->old_domain, cpu_online_map); + for_each_cpu_mask_nr(cpu, mask) { + for (vector = FIRST_EXTERNAL_VECTOR; vector < NR_VECTORS; + vector++) { + if (per_cpu(vector_irq, cpu)[vector] != irq) + continue; + per_cpu(vector_irq, cpu)[vector] = -1; + break; + } + } + cfg->move_in_progress = 0; } void __setup_vector_irq(int cpu) -- cgit v1.2.2 From a1967d64414dab500e86cbbddf8eae6ad2047903 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 21 Nov 2008 11:16:48 -0800 Subject: x86: revert irq number limitation Impact: fix MSIx not enough irq numbers available regression The manual revert of the sparse_irq patches missed to bring the number of possible irqs back to the .27 status. This resulted in a regression when two multichannel network cards were placed in a system with only one IO_APIC - causing the networking driver to not have the right IRQ and the device not coming up. Remove the dynamic allocation logic leftovers and simply return NR_IRQS in probe_nr_irqs() for now. Fixes: http://lkml.org/lkml/2008/11/19/354 Reported-by: Jesper Dangaard Brouer Signed-off-by: Thomas Gleixner Tested-by: Jesper Dangaard Brouer Acked-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/io_apic.c | 22 +--------------------- 1 file changed, 1 insertion(+), 21 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index c9513e1ff28..1fec0f9b150 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -3608,27 +3608,7 @@ int __init io_apic_get_redir_entries (int ioapic) int __init probe_nr_irqs(void) { - int idx; - int nr = 0; -#ifndef CONFIG_XEN - int nr_min = 32; -#else - int nr_min = NR_IRQS; -#endif - - for (idx = 0; idx < nr_ioapics; idx++) - nr += io_apic_get_redir_entries(idx) + 1; - - /* double it for hotplug and msi and nmi */ - nr <<= 1; - - /* something wrong ? */ - if (nr < nr_min) - nr = nr_min; - if (WARN_ON(nr > NR_IRQS)) - nr = NR_IRQS; - - return nr; + return NR_IRQS; } /* -------------------------------------------------------------------------- -- cgit v1.2.2 From eff79aee91dd07e944df65fa448c8baeee7709d8 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Tue, 25 Nov 2008 14:13:03 +0100 Subject: arch/x86/kernel/pci-calgary_64.c: change simple_strtol to simple_strtoul Impact: fix theoretical option string parsing overflow Since bridge is unsigned, it would seem better to use simple_strtoul that simple_strtol. A simplified version of the semantic patch that makes this change is as follows: (http://www.emn.fr/x-info/coccinelle/) // @r2@ long e; position p; @@ e = simple_strtol@p(...) @@ position p != r2.p; type T; T e; @@ e = - simple_strtol@p + simple_strtoul (...) // Signed-off-by: Julia Lawall Cc: muli@il.ibm.com Cc: jdmason@kudzu.us Cc: discuss@x86-64.org Signed-off-by: Ingo Molnar --- arch/x86/kernel/pci-calgary_64.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/pci-calgary_64.c b/arch/x86/kernel/pci-calgary_64.c index e1e731d78f3..d28bbdc35e4 100644 --- a/arch/x86/kernel/pci-calgary_64.c +++ b/arch/x86/kernel/pci-calgary_64.c @@ -1567,7 +1567,7 @@ static int __init calgary_parse_options(char *p) ++p; if (*p == '\0') break; - bridge = simple_strtol(p, &endp, 0); + bridge = simple_strtoul(p, &endp, 0); if (p == endp) break; -- cgit v1.2.2 From 292c669cd7087a090d6420e223eb1072f3e3c50b Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 25 Nov 2008 08:45:13 +0100 Subject: x86, bts: exclude ds.c from build when disabled Impact: cleanup Move the CONFIG guard from the .c file into the makefile. Reported-by: Andi Kleen Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/kernel/Makefile | 2 +- arch/x86/kernel/ds.c | 3 --- 2 files changed, 1 insertion(+), 4 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index e489ff9cb3e..b62a7667828 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -41,7 +41,7 @@ obj-$(CONFIG_X86_TRAMPOLINE) += trampoline.o obj-y += process.o obj-y += i387.o xsave.o obj-y += ptrace.o -obj-y += ds.o +obj-$(CONFIG_X86_DS) += ds.o obj-$(CONFIG_X86_32) += tls.o obj-$(CONFIG_IA32_EMULATION) += tls.o obj-y += step.o diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index d1a121443bd..4c8d57ec966 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -21,8 +21,6 @@ */ -#ifdef CONFIG_X86_DS - #include #include @@ -878,4 +876,3 @@ void ds_free(struct ds_context *context) while (leftovers--) ds_put_context(context); } -#endif /* CONFIG_X86_DS */ -- cgit v1.2.2 From c4858ffc8f2dc850cb1f609c679b1ac1ad36ef0c Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 25 Nov 2008 08:49:06 +0100 Subject: x86, pebs: fix PEBS record size configuration Impact: fix DS hw enablement on 64-bit x86 Fix the PEBS record size in the DS configuration. Reported-by: Stephane Eranian Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/kernel/ds.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 4c8d57ec966..04e38ef646a 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -816,13 +816,21 @@ static const struct ds_configuration ds_cfg_var = { .sizeof_ds = sizeof(long) * 12, .sizeof_field = sizeof(long), .sizeof_rec[ds_bts] = sizeof(long) * 3, +#ifdef __i386__ .sizeof_rec[ds_pebs] = sizeof(long) * 10 +#else + .sizeof_rec[ds_pebs] = sizeof(long) * 18 +#endif }; static const struct ds_configuration ds_cfg_64 = { .sizeof_ds = 8 * 12, .sizeof_field = 8, .sizeof_rec[ds_bts] = 8 * 3, +#ifdef __i386__ .sizeof_rec[ds_pebs] = 8 * 10 +#else + .sizeof_rec[ds_pebs] = 8 * 18 +#endif }; static inline void -- cgit v1.2.2 From de90add30e79261c3b5be68bb0f22d2ef98e8113 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 25 Nov 2008 08:52:56 +0100 Subject: x86, bts: fix wrmsr and spinlock over kmalloc Impact: fix sleeping-with-spinlock-held bugs/crashes - Turn a wrmsr to write the DS_AREA MSR into a wrmsrl. - Use irqsave variants of spinlocks. - Do not allocate memory while holding spinlocks. Reported-by: Stephane Eranian Reported-by: Ingo Molnar Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/kernel/ds.c | 77 +++++++++++++++++++++++++++------------------------- 1 file changed, 40 insertions(+), 37 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 04e38ef646a..a2d1176c38e 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -209,14 +209,15 @@ static DEFINE_PER_CPU(struct ds_context *, system_context); static inline struct ds_context *ds_get_context(struct task_struct *task) { struct ds_context *context; + unsigned long irq; - spin_lock(&ds_lock); + spin_lock_irqsave(&ds_lock, irq); context = (task ? task->thread.ds_ctx : this_system_context); if (context) context->count++; - spin_unlock(&ds_lock); + spin_unlock_irqrestore(&ds_lock, irq); return context; } @@ -224,55 +225,46 @@ static inline struct ds_context *ds_get_context(struct task_struct *task) /* * Same as ds_get_context, but allocates the context and it's DS * structure, if necessary; returns NULL; if out of memory. - * - * pre: requires ds_lock to be held */ static inline struct ds_context *ds_alloc_context(struct task_struct *task) { struct ds_context **p_context = (task ? &task->thread.ds_ctx : &this_system_context); struct ds_context *context = *p_context; + unsigned long irq; if (!context) { - spin_unlock(&ds_lock); - context = kzalloc(sizeof(*context), GFP_KERNEL); - - if (!context) { - spin_lock(&ds_lock); + if (!context) return NULL; - } context->ds = kzalloc(ds_cfg.sizeof_ds, GFP_KERNEL); if (!context->ds) { kfree(context); - spin_lock(&ds_lock); return NULL; } - spin_lock(&ds_lock); - /* - * Check for race - another CPU could have allocated - * it meanwhile: - */ + spin_lock_irqsave(&ds_lock, irq); + if (*p_context) { kfree(context->ds); kfree(context); - return *p_context; - } - - *p_context = context; - context->this = p_context; - context->task = task; + context = *p_context; + } else { + *p_context = context; - if (task) - set_tsk_thread_flag(task, TIF_DS_AREA_MSR); + context->this = p_context; + context->task = task; - if (!task || (task == current)) - wrmsr(MSR_IA32_DS_AREA, (unsigned long)context->ds, 0); + if (task) + set_tsk_thread_flag(task, TIF_DS_AREA_MSR); - get_tracer(task); + if (!task || (task == current)) + wrmsrl(MSR_IA32_DS_AREA, + (unsigned long)context->ds); + } + spin_unlock_irqrestore(&ds_lock, irq); } context->count++; @@ -286,10 +278,12 @@ static inline struct ds_context *ds_alloc_context(struct task_struct *task) */ static inline void ds_put_context(struct ds_context *context) { + unsigned long irq; + if (!context) return; - spin_lock(&ds_lock); + spin_lock_irqsave(&ds_lock, irq); if (--context->count) goto out; @@ -311,7 +305,7 @@ static inline void ds_put_context(struct ds_context *context) kfree(context->ds); kfree(context); out: - spin_unlock(&ds_lock); + spin_unlock_irqrestore(&ds_lock, irq); } @@ -382,6 +376,7 @@ static int ds_request(struct task_struct *task, void *base, size_t size, struct ds_context *context; unsigned long buffer, adj; const unsigned long alignment = (1 << 3); + unsigned long irq; int error = 0; if (!ds_cfg.sizeof_ds) @@ -396,26 +391,27 @@ static int ds_request(struct task_struct *task, void *base, size_t size, return -EOPNOTSUPP; - spin_lock(&ds_lock); - - error = -ENOMEM; context = ds_alloc_context(task); if (!context) - goto out_unlock; + return -ENOMEM; + + spin_lock_irqsave(&ds_lock, irq); error = -EPERM; if (!check_tracer(task)) goto out_unlock; + get_tracer(task); + error = -EALREADY; if (context->owner[qual] == current) - goto out_unlock; + goto out_put_tracer; error = -EPERM; if (context->owner[qual] != NULL) - goto out_unlock; + goto out_put_tracer; context->owner[qual] = current; - spin_unlock(&ds_lock); + spin_unlock_irqrestore(&ds_lock, irq); error = -ENOMEM; @@ -463,10 +459,17 @@ static int ds_request(struct task_struct *task, void *base, size_t size, out_release: context->owner[qual] = NULL; ds_put_context(context); + put_tracer(task); + return error; + + out_put_tracer: + spin_unlock_irqrestore(&ds_lock, irq); + ds_put_context(context); + put_tracer(task); return error; out_unlock: - spin_unlock(&ds_lock); + spin_unlock_irqrestore(&ds_lock, irq); ds_put_context(context); return error; } -- cgit v1.2.2 From a266d9f1253a38ec2d5655ebcd6846298b0554f4 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Fri, 21 Nov 2008 14:49:25 +0100 Subject: [CPUFREQ] powernow-k8: ignore out-of-range PstateStatus value A workaround for AMD CPU family 11h erratum 311 might cause that the P-state Status Register shows a "current P-state" which is larger than the "current P-state limit" in P-state Current Limit Register. For the wrong P-state value there is no ACPI _PSS object defined and powernow-k8/cpufreq can't determine the proper CPU frequency for that state. As a consequence this can cause a panic during boot (potentially with all recent kernel versions -- at least I have reproduced it with various 2.6.27 kernels and with the current .28 series), as an example: powernow-k8: Found 1 AMD Turion(tm)X2 Ultra DualCore Mobile ZM-82 processors (2 \ ) powernow-k8: 0 : pstate 0 (2200 MHz) powernow-k8: 1 : pstate 1 (1100 MHz) powernow-k8: 2 : pstate 2 (600 MHz) BUG: unable to handle kernel paging request at ffff88086e7528b8 IP: [] cpufreq_stats_update+0x4a/0x5f PGD 202063 PUD 0 Oops: 0002 [#1] SMP last sysfs file: CPU 1 Modules linked in: Pid: 1, comm: swapper Not tainted 2.6.28-rc3-dirty #16 RIP: 0010:[] [] cpufreq_stats_update+0x4a/0\ f Synaptics claims to have extended capabilities, but I'm not able to read them.<6\ 6 RAX: 0000000000000000 RBX: 0000000000000001 RCX: ffff88006e7528c0 RDX: 00000000ffffffff RSI: ffff88006e54af00 RDI: ffffffff808f056c RBP: 00000000fffee697 R08: 0000000000000003 R09: ffff88006e73f080 R10: 0000000000000001 R11: 00000000002191c0 R12: ffff88006fb83c10 R13: 00000000ffffffff R14: 0000000000000001 R15: 0000000000000000 FS: 0000000000000000(0000) GS:ffff88006fb50740(0000) knlGS:0000000000000000 Unable to initialize Synaptics hardware. CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: ffff88086e7528b8 CR3: 0000000000201000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process swapper (pid: 1, threadinfo ffff88006fb82000, task ffff88006fb816d0) Stack: ffff88006e74da50 0000000000000000 ffff88006e54af00 ffffffff804863c7 ffff88006e74da50 0000000000000000 00000000ffffffff 0000000000000000 ffff88006fb83c10 ffffffff8024b46c ffffffff808f0560 ffff88006fb83c10 Call Trace: [] ? cpufreq_stat_notifier_trans+0x51/0x83 [] ? notifier_call_chain+0x29/0x4c [] ? __srcu_notifier_call_chain+0x46/0x61 [] ? cpufreq_notify_transition+0x93/0xa9 [] ? powernowk8_target+0x1e8/0x5f3 [] ? cpufreq_governor_performance+0x1b/0x20 [] ? __cpufreq_governor+0x71/0xa8 [] ? __cpufreq_set_policy+0x101/0x13e [] ? cpufreq_add_dev+0x3f0/0x4cd [] ? handle_update+0x0/0x8 [] ? sysdev_driver_register+0xb6/0x10d [] ? powernowk8_init+0x0/0x7e [] ? cpufreq_register_driver+0x8f/0x140 [] ? _stext+0x56/0x14f [] ? proc_register+0x122/0x17d [] ? create_proc_entry+0x73/0x8a [] ? register_irq_proc+0x92/0xaa [] ? init_irq_proc+0x57/0x69 [] ? kernel_init+0x116/0x169 [] ? child_rip+0xa/0x11 [] ? kernel_init+0x0/0x169 [] ? child_rip+0x0/0x11 Code: 05 c5 83 36 00 48 c7 c2 48 5d 86 80 48 8b 04 d8 48 8b 40 08 48 8b 34 02 48\ RIP [] cpufreq_stats_update+0x4a/0x5f RSP CR2: ffff88086e7528b8 ---[ end trace 0678bac75e67a2f7 ]--- Kernel panic - not syncing: Attempted to kill init! In short, aftereffect of the wrong P-state is that cpufreq_stats_update() uses "-1" as index for some array in cpufreq_stats_update (unsigned int cpu) { ... if (stat->time_in_state) stat->time_in_state[stat->last_index] = cputime64_add(stat->time_in_state[stat->last_index], cputime_sub(cur_time, stat->last_time)); ... } Fortunately, the wrong P-state value is returned only if the core is in P-state 0. This fix solves the problem by detecting the out-of-range P-state, ignoring it, and using "0" instead. Cc: Mark Langsdorf Signed-off-by: Andreas Herrmann Signed-off-by: Dave Jones --- arch/x86/kernel/cpu/cpufreq/powernow-k8.c | 18 +++++++++++++++--- arch/x86/kernel/cpu/cpufreq/powernow-k8.h | 17 ++++++++++++++++- 2 files changed, 31 insertions(+), 4 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c index d3dcd58b87c..7f05f44b97e 100644 --- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c +++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c @@ -115,9 +115,20 @@ static int query_current_values_with_pending_wait(struct powernow_k8_data *data) u32 i = 0; if (cpu_family == CPU_HW_PSTATE) { - rdmsr(MSR_PSTATE_STATUS, lo, hi); - i = lo & HW_PSTATE_MASK; - data->currpstate = i; + if (data->currpstate == HW_PSTATE_INVALID) { + /* read (initial) hw pstate if not yet set */ + rdmsr(MSR_PSTATE_STATUS, lo, hi); + i = lo & HW_PSTATE_MASK; + + /* + * a workaround for family 11h erratum 311 might cause + * an "out-of-range Pstate if the core is in Pstate-0 + */ + if (i >= data->numps) + data->currpstate = HW_PSTATE_0; + else + data->currpstate = i; + } return 0; } do { @@ -1121,6 +1132,7 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol) } data->cpu = pol->cpu; + data->currpstate = HW_PSTATE_INVALID; if (powernow_k8_cpu_init_acpi(data)) { /* diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.h b/arch/x86/kernel/cpu/cpufreq/powernow-k8.h index ab48cfed4d9..65cfb5d7f77 100644 --- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.h +++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.h @@ -5,6 +5,19 @@ * http://www.gnu.org/licenses/gpl.html */ + +enum pstate { + HW_PSTATE_INVALID = 0xff, + HW_PSTATE_0 = 0, + HW_PSTATE_1 = 1, + HW_PSTATE_2 = 2, + HW_PSTATE_3 = 3, + HW_PSTATE_4 = 4, + HW_PSTATE_5 = 5, + HW_PSTATE_6 = 6, + HW_PSTATE_7 = 7, +}; + struct powernow_k8_data { unsigned int cpu; @@ -23,7 +36,9 @@ struct powernow_k8_data { u32 exttype; /* extended interface = 1 */ /* keep track of the current fid / vid or pstate */ - u32 currvid, currfid, currpstate; + u32 currvid; + u32 currfid; + enum pstate currpstate; /* the powernow_table includes all frequency and vid/fid pairings: * fid are the lower 8 bits of the index, vid are the upper 8 bits. -- cgit v1.2.2 From 2236d252e001ea57d53cec1954f680e503f3b8bc Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 22 Nov 2008 17:37:34 +0000 Subject: enable_IR_x2apic() needs to be __init calls __init, called only from __init Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- arch/x86/kernel/apic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 04a7f960bbc..16f94879b52 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -1315,7 +1315,7 @@ void enable_x2apic(void) } } -void enable_IR_x2apic(void) +void __init enable_IR_x2apic(void) { #ifdef CONFIG_INTR_REMAP int ret; -- cgit v1.2.2 From 23a14b9e9db49ed5f7683857557c26c874d4abb6 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 22 Nov 2008 17:37:44 +0000 Subject: kvm_setup_secondary_clock() is cpuinit Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- arch/x86/kernel/kvmclock.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c index 1c9cc431ea4..e169ae9b6a6 100644 --- a/arch/x86/kernel/kvmclock.c +++ b/arch/x86/kernel/kvmclock.c @@ -128,7 +128,7 @@ static int kvm_register_clock(char *txt) } #ifdef CONFIG_X86_LOCAL_APIC -static void __devinit kvm_setup_secondary_clock(void) +static void __cpuinit kvm_setup_secondary_clock(void) { /* * Now that the first cpu already had this clocksource initialized, -- cgit v1.2.2