diff options
Diffstat (limited to 'arch/x86/kernel')
64 files changed, 2897 insertions, 1979 deletions
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index e489ff9cb3e..1cad9318d21 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -25,7 +25,7 @@ CFLAGS_tsc.o := $(nostackp) obj-y := process_$(BITS).o signal_$(BITS).o entry_$(BITS).o obj-y += traps.o irq.o irq_$(BITS).o dumpstack_$(BITS).o -obj-y += time_$(BITS).o ioport.o ldt.o +obj-y += time_$(BITS).o ioport.o ldt.o dumpstack.o obj-y += setup.o i8259.o irqinit_$(BITS).o setup_percpu.o obj-$(CONFIG_X86_VISWS) += visws_quirks.o obj-$(CONFIG_X86_32) += probe_roms_32.o @@ -41,7 +41,7 @@ obj-$(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-$(CONFIG_IA32_EMULATION) += tls.o obj-y += step.o @@ -65,6 +65,7 @@ obj-$(CONFIG_X86_LOCAL_APIC) += apic.o nmi.o obj-$(CONFIG_X86_IO_APIC) += io_apic.o obj-$(CONFIG_X86_REBOOTFIXUPS) += reboot_fixups_32.o obj-$(CONFIG_DYNAMIC_FTRACE) += ftrace.o +obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += ftrace.o obj-$(CONFIG_KEXEC) += machine_kexec_$(BITS).o obj-$(CONFIG_KEXEC) += relocate_kernel_$(BITS).o crash.o obj-$(CONFIG_CRASH_DUMP) += crash_dump_$(BITS).o diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 8c1f76abae9..65d0b72777e 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -1343,7 +1343,6 @@ static void __init acpi_process_madt(void) error = acpi_parse_madt_ioapic_entries(); if (!error) { acpi_irq_model = ACPI_IRQ_MODEL_IOAPIC; - acpi_irq_balance_set(NULL); acpi_ioapic = 1; smp_found_config = 1; @@ -1361,6 +1360,17 @@ static void __init acpi_process_madt(void) disable_acpi(); } } + + /* + * ACPI supports both logical (e.g. Hyper-Threading) and physical + * processors, where MPS only supports physical. + */ + if (acpi_lapic && acpi_ioapic) + printk(KERN_INFO "Using ACPI (MADT) for SMP configuration " + "information\n"); + else if (acpi_lapic) + printk(KERN_INFO "Using ACPI for processor (LAPIC) " + "configuration information\n"); #endif return; } diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 331b318304e..a7b6dec6fc3 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -187,6 +187,8 @@ static int iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd) spin_lock_irqsave(&iommu->lock, flags); ret = __iommu_queue_command(iommu, cmd); + if (!ret) + iommu->need_sync = 1; spin_unlock_irqrestore(&iommu->lock, flags); return ret; @@ -210,10 +212,13 @@ static int iommu_completion_wait(struct amd_iommu *iommu) cmd.data[0] = CMD_COMPL_WAIT_INT_MASK; CMD_SET_TYPE(&cmd, CMD_COMPL_WAIT); - iommu->need_sync = 0; - spin_lock_irqsave(&iommu->lock, flags); + if (!iommu->need_sync) + goto out; + + iommu->need_sync = 0; + ret = __iommu_queue_command(iommu, &cmd); if (ret) @@ -254,8 +259,6 @@ static int iommu_queue_inv_dev_entry(struct amd_iommu *iommu, u16 devid) ret = iommu_queue_command(iommu, &cmd); - iommu->need_sync = 1; - return ret; } @@ -281,8 +284,6 @@ static int iommu_queue_inv_iommu_pages(struct amd_iommu *iommu, ret = iommu_queue_command(iommu, &cmd); - iommu->need_sync = 1; - return ret; } @@ -343,7 +344,7 @@ static int iommu_map(struct protection_domain *dom, u64 __pte, *pte, *page; bus_addr = PAGE_ALIGN(bus_addr); - phys_addr = PAGE_ALIGN(bus_addr); + phys_addr = PAGE_ALIGN(phys_addr); /* only support 512GB address spaces for now */ if (bus_addr > IOMMU_MAP_SIZE_L3 || !(prot & IOMMU_PROT_MASK)) @@ -537,7 +538,7 @@ 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; } @@ -599,7 +600,7 @@ static void dma_ops_free_pagetable(struct dma_ops_domain *dma_dom) continue; p2 = IOMMU_PTE_PAGE(p1[i]); - for (j = 0; j < 512; ++i) { + for (j = 0; j < 512; ++j) { if (!IOMMU_PTE_PRESENT(p2[j])) continue; p3 = IOMMU_PTE_PAGE(p2[j]); @@ -762,8 +763,6 @@ static void set_device_domain(struct amd_iommu *iommu, write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); iommu_queue_inv_dev_entry(iommu, devid); - - iommu->need_sync = 1; } /***************************************************************************** @@ -858,6 +857,9 @@ static int get_device_resources(struct device *dev, print_devid(_bdf, 1); } + if (domain_for_device(_bdf) == NULL) + set_device_domain(*iommu, *domain, _bdf); + return 1; } @@ -908,7 +910,7 @@ static void dma_ops_domain_unmap(struct amd_iommu *iommu, if (address >= dom->aperture_size) return; - WARN_ON(address & 0xfffULL || address > dom->aperture_size); + WARN_ON(address & ~PAGE_MASK || address >= dom->aperture_size); pte = dom->pte_pages[IOMMU_PTE_L1_INDEX(address)]; pte += IOMMU_PTE_L0_INDEX(address); @@ -920,8 +922,8 @@ static void dma_ops_domain_unmap(struct amd_iommu *iommu, /* * This function contains common code for mapping of a physically - * contiguous memory region into DMA address space. It is uses by all - * mapping functions provided by this IOMMU driver. + * contiguous memory region into DMA address space. It is used by all + * mapping functions provided with this IOMMU driver. * Must be called with the domain lock held. */ static dma_addr_t __map_single(struct device *dev, @@ -981,7 +983,8 @@ static void __unmap_single(struct amd_iommu *iommu, dma_addr_t i, start; unsigned int pages; - if ((dma_addr == 0) || (dma_addr + size > dma_dom->aperture_size)) + if ((dma_addr == bad_dma_address) || + (dma_addr + size > dma_dom->aperture_size)) return; pages = iommu_num_pages(dma_addr, size, PAGE_SIZE); @@ -1031,8 +1034,7 @@ static dma_addr_t map_single(struct device *dev, phys_addr_t paddr, if (addr == bad_dma_address) goto out; - if (unlikely(iommu->need_sync)) - iommu_completion_wait(iommu); + iommu_completion_wait(iommu); out: spin_unlock_irqrestore(&domain->lock, flags); @@ -1060,8 +1062,7 @@ static void unmap_single(struct device *dev, dma_addr_t dma_addr, __unmap_single(iommu, domain->priv, dma_addr, size, dir); - if (unlikely(iommu->need_sync)) - iommu_completion_wait(iommu); + iommu_completion_wait(iommu); spin_unlock_irqrestore(&domain->lock, flags); } @@ -1127,8 +1128,7 @@ static int map_sg(struct device *dev, struct scatterlist *sglist, goto unmap; } - if (unlikely(iommu->need_sync)) - iommu_completion_wait(iommu); + iommu_completion_wait(iommu); out: spin_unlock_irqrestore(&domain->lock, flags); @@ -1173,8 +1173,7 @@ static void unmap_sg(struct device *dev, struct scatterlist *sglist, s->dma_address = s->dma_length = 0; } - if (unlikely(iommu->need_sync)) - iommu_completion_wait(iommu); + iommu_completion_wait(iommu); spin_unlock_irqrestore(&domain->lock, flags); } @@ -1225,8 +1224,7 @@ static void *alloc_coherent(struct device *dev, size_t size, goto out; } - if (unlikely(iommu->need_sync)) - iommu_completion_wait(iommu); + iommu_completion_wait(iommu); out: spin_unlock_irqrestore(&domain->lock, flags); @@ -1257,8 +1255,7 @@ static void free_coherent(struct device *dev, size_t size, __unmap_single(iommu, domain->priv, dma_addr, size, DMA_BIDIRECTIONAL); - if (unlikely(iommu->need_sync)) - iommu_completion_wait(iommu); + iommu_completion_wait(iommu); spin_unlock_irqrestore(&domain->lock, flags); diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 0cdcda35a05..30ae2701b3d 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 @@ -1213,7 +1213,9 @@ 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, "fullflush", 11) == 0) + if (strncmp(str, "share", 5) == 0) + amd_iommu_isolate = 0; + if (strncmp(str, "fullflush", 9) == 0) amd_iommu_unmap_flush = true; } diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 1d410ee4b06..93cf2d13f33 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -141,7 +141,7 @@ static int lapic_next_event(unsigned long delta, struct clock_event_device *evt); static void lapic_timer_setup(enum clock_event_mode mode, struct clock_event_device *evt); -static void lapic_timer_broadcast(cpumask_t mask); +static void lapic_timer_broadcast(const cpumask_t *mask); static void apic_pm_activate(void); /* @@ -454,7 +454,7 @@ static void lapic_timer_setup(enum clock_event_mode mode, /* * Local APIC timer broadcast function */ -static void lapic_timer_broadcast(cpumask_t mask) +static void lapic_timer_broadcast(const cpumask_t *mask) { #ifdef CONFIG_SMP send_IPI_mask(mask, LOCAL_TIMER_VECTOR); @@ -470,7 +470,7 @@ static void __cpuinit setup_APIC_timer(void) struct clock_event_device *levt = &__get_cpu_var(lapic_events); memcpy(levt, &lapic_clockevent, sizeof(*levt)); - levt->cpumask = cpumask_of_cpu(smp_processor_id()); + levt->cpumask = cpumask_of(smp_processor_id()); clockevents_register_device(levt); } @@ -1309,7 +1309,7 @@ void enable_x2apic(void) } } -void enable_IR_x2apic(void) +void __init enable_IR_x2apic(void) { #ifdef CONFIG_INTR_REMAP int ret; @@ -1890,8 +1890,8 @@ void __cpuinit generic_processor_info(int apicid, int version) } #endif - cpu_set(cpu, cpu_possible_map); - cpu_set(cpu, cpu_present_map); + set_cpu_possible(cpu, true); + set_cpu_present(cpu, true); } #ifdef CONFIG_X86_64 @@ -2093,7 +2093,7 @@ __cpuinit int apic_is_clustered_box(void) bios_cpu_apicid = early_per_cpu_ptr(x86_bios_cpu_apicid); bitmap_zero(clustermap, NUM_APIC_CLUSTERS); - for (i = 0; i < NR_CPUS; i++) { + for (i = 0; i < nr_cpu_ids; i++) { /* are we being called early in kernel startup? */ if (bios_cpu_apicid) { id = bios_cpu_apicid[i]; diff --git a/arch/x86/kernel/apm_32.c b/arch/x86/kernel/apm_32.c index 5145a6e72bb..3a26525a3f3 100644 --- a/arch/x86/kernel/apm_32.c +++ b/arch/x86/kernel/apm_32.c @@ -391,11 +391,7 @@ static int power_off; #else static int power_off = 1; #endif -#ifdef CONFIG_APM_REAL_MODE_POWER_OFF -static int realmode_power_off = 1; -#else static int realmode_power_off; -#endif #ifdef CONFIG_APM_ALLOW_INTS static int allow_ints = 1; #else diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c index 8e48c5d4467..88ea02dcb62 100644 --- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c @@ -33,6 +33,7 @@ #include <linux/cpufreq.h> #include <linux/compiler.h> #include <linux/dmi.h> +#include <linux/ftrace.h> #include <linux/acpi.h> #include <acpi/processor.h> @@ -391,6 +392,7 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy, unsigned int next_perf_state = 0; /* Index into perf table */ unsigned int i; int result = 0; + struct power_trace it; dprintk("acpi_cpufreq_target %d (%d)\n", target_freq, policy->cpu); @@ -427,6 +429,8 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy, } } + trace_power_mark(&it, POWER_PSTATE, next_perf_state); + switch (data->cpu_feature) { case SYSTEM_INTEL_MSR_CAPABLE: cmd.type = SYSTEM_INTEL_MSR_CAPABLE; 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. diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index cce0b6118d5..816f27f289b 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -307,12 +307,11 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) set_cpu_cap(c, X86_FEATURE_P4); if (c->x86 == 6) set_cpu_cap(c, X86_FEATURE_P3); +#endif if (cpu_has_bts) ptrace_bts_init_intel(c); -#endif - detect_extended_topology(c); if (!cpu_has(c, X86_FEATURE_XTOPOLOGY)) { /* diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 3f46afbb1cf..fb7f946cb65 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -534,31 +534,16 @@ static void __cpuinit free_cache_attributes(unsigned int cpu) per_cpu(cpuid4_info, cpu) = NULL; } -static int __cpuinit detect_cache_attributes(unsigned int cpu) +static void get_cpu_leaves(void *_retval) { - struct _cpuid4_info *this_leaf; - unsigned long j; - int retval; - cpumask_t oldmask; - - if (num_cache_leaves == 0) - return -ENOENT; - - per_cpu(cpuid4_info, cpu) = kzalloc( - sizeof(struct _cpuid4_info) * num_cache_leaves, GFP_KERNEL); - if (per_cpu(cpuid4_info, cpu) == NULL) - return -ENOMEM; - - oldmask = current->cpus_allowed; - retval = set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu)); - if (retval) - goto out; + int j, *retval = _retval, cpu = smp_processor_id(); /* Do cpuid and store the results */ for (j = 0; j < num_cache_leaves; j++) { + struct _cpuid4_info *this_leaf; this_leaf = CPUID4_INFO_IDX(cpu, j); - retval = cpuid4_cache_lookup(j, this_leaf); - if (unlikely(retval < 0)) { + *retval = cpuid4_cache_lookup(j, this_leaf); + if (unlikely(*retval < 0)) { int i; for (i = 0; i < j; i++) @@ -567,9 +552,21 @@ static int __cpuinit detect_cache_attributes(unsigned int cpu) } cache_shared_cpu_map_setup(cpu, j); } - set_cpus_allowed_ptr(current, &oldmask); +} + +static int __cpuinit detect_cache_attributes(unsigned int cpu) +{ + int retval; + + if (num_cache_leaves == 0) + return -ENOENT; + + per_cpu(cpuid4_info, cpu) = kzalloc( + sizeof(struct _cpuid4_info) * num_cache_leaves, GFP_KERNEL); + if (per_cpu(cpuid4_info, cpu) == NULL) + return -ENOMEM; -out: + smp_call_function_single(cpu, get_cpu_leaves, &retval, true); if (retval) { kfree(per_cpu(cpuid4_info, cpu)); per_cpu(cpuid4_info, cpu) = NULL; @@ -626,8 +623,8 @@ static ssize_t show_shared_cpu_map_func(struct _cpuid4_info *this_leaf, cpumask_t *mask = &this_leaf->shared_cpu_map; n = type? - cpulist_scnprintf(buf, len-2, *mask): - cpumask_scnprintf(buf, len-2, *mask); + cpulist_scnprintf(buf, len-2, mask) : + cpumask_scnprintf(buf, len-2, mask); buf[n++] = '\n'; buf[n] = '\0'; } diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c index 5eb390a4b2e..a1de80f368f 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c @@ -83,34 +83,41 @@ static DEFINE_PER_CPU(unsigned char, bank_map); /* see which banks are on */ * CPU Initialization */ +struct thresh_restart { + struct threshold_block *b; + int reset; + u16 old_limit; +}; + /* must be called with correct cpu affinity */ -static void threshold_restart_bank(struct threshold_block *b, - int reset, u16 old_limit) +static long threshold_restart_bank(void *_tr) { + struct thresh_restart *tr = _tr; u32 mci_misc_hi, mci_misc_lo; - rdmsr(b->address, mci_misc_lo, mci_misc_hi); + rdmsr(tr->b->address, mci_misc_lo, mci_misc_hi); - if (b->threshold_limit < (mci_misc_hi & THRESHOLD_MAX)) - reset = 1; /* limit cannot be lower than err count */ + if (tr->b->threshold_limit < (mci_misc_hi & THRESHOLD_MAX)) + tr->reset = 1; /* limit cannot be lower than err count */ - if (reset) { /* reset err count and overflow bit */ + if (tr->reset) { /* reset err count and overflow bit */ mci_misc_hi = (mci_misc_hi & ~(MASK_ERR_COUNT_HI | MASK_OVERFLOW_HI)) | - (THRESHOLD_MAX - b->threshold_limit); - } else if (old_limit) { /* change limit w/o reset */ + (THRESHOLD_MAX - tr->b->threshold_limit); + } else if (tr->old_limit) { /* change limit w/o reset */ int new_count = (mci_misc_hi & THRESHOLD_MAX) + - (old_limit - b->threshold_limit); + (tr->old_limit - tr->b->threshold_limit); mci_misc_hi = (mci_misc_hi & ~MASK_ERR_COUNT_HI) | (new_count & THRESHOLD_MAX); } - b->interrupt_enable ? + tr->b->interrupt_enable ? (mci_misc_hi = (mci_misc_hi & ~MASK_INT_TYPE_HI) | INT_TYPE_APIC) : (mci_misc_hi &= ~MASK_INT_TYPE_HI); mci_misc_hi |= MASK_COUNT_EN_HI; - wrmsr(b->address, mci_misc_lo, mci_misc_hi); + wrmsr(tr->b->address, mci_misc_lo, mci_misc_hi); + return 0; } /* cpu init entry point, called from mce.c with preempt off */ @@ -120,6 +127,7 @@ void __cpuinit mce_amd_feature_init(struct cpuinfo_x86 *c) unsigned int cpu = smp_processor_id(); u8 lvt_off; u32 low = 0, high = 0, address = 0; + struct thresh_restart tr; for (bank = 0; bank < NR_BANKS; ++bank) { for (block = 0; block < NR_BLOCKS; ++block) { @@ -162,7 +170,10 @@ void __cpuinit mce_amd_feature_init(struct cpuinfo_x86 *c) wrmsr(address, low, high); threshold_defaults.address = address; - threshold_restart_bank(&threshold_defaults, 0, 0); + tr.b = &threshold_defaults; + tr.reset = 0; + tr.old_limit = 0; + threshold_restart_bank(&tr); } } } @@ -251,20 +262,6 @@ struct threshold_attr { ssize_t(*store) (struct threshold_block *, const char *, size_t count); }; -static void affinity_set(unsigned int cpu, cpumask_t *oldmask, - cpumask_t *newmask) -{ - *oldmask = current->cpus_allowed; - cpus_clear(*newmask); - cpu_set(cpu, *newmask); - set_cpus_allowed_ptr(current, newmask); -} - -static void affinity_restore(const cpumask_t *oldmask) -{ - set_cpus_allowed_ptr(current, oldmask); -} - #define SHOW_FIELDS(name) \ static ssize_t show_ ## name(struct threshold_block * b, char *buf) \ { \ @@ -277,15 +274,16 @@ static ssize_t store_interrupt_enable(struct threshold_block *b, const char *buf, size_t count) { char *end; - cpumask_t oldmask, newmask; + struct thresh_restart tr; unsigned long new = simple_strtoul(buf, &end, 0); if (end == buf) return -EINVAL; b->interrupt_enable = !!new; - affinity_set(b->cpu, &oldmask, &newmask); - threshold_restart_bank(b, 0, 0); - affinity_restore(&oldmask); + tr.b = b; + tr.reset = 0; + tr.old_limit = 0; + work_on_cpu(b->cpu, threshold_restart_bank, &tr); return end - buf; } @@ -294,8 +292,7 @@ static ssize_t store_threshold_limit(struct threshold_block *b, const char *buf, size_t count) { char *end; - cpumask_t oldmask, newmask; - u16 old; + struct thresh_restart tr; unsigned long new = simple_strtoul(buf, &end, 0); if (end == buf) return -EINVAL; @@ -303,34 +300,36 @@ static ssize_t store_threshold_limit(struct threshold_block *b, new = THRESHOLD_MAX; if (new < 1) new = 1; - old = b->threshold_limit; + tr.old_limit = b->threshold_limit; b->threshold_limit = new; + tr.b = b; + tr.reset = 0; - affinity_set(b->cpu, &oldmask, &newmask); - threshold_restart_bank(b, 0, old); - affinity_restore(&oldmask); + work_on_cpu(b->cpu, threshold_restart_bank, &tr); return end - buf; } -static ssize_t show_error_count(struct threshold_block *b, char *buf) +static long local_error_count(void *_b) { - u32 high, low; - cpumask_t oldmask, newmask; - affinity_set(b->cpu, &oldmask, &newmask); + struct threshold_block *b = _b; + u32 low, high; + rdmsr(b->address, low, high); - affinity_restore(&oldmask); - return sprintf(buf, "%x\n", - (high & 0xFFF) - (THRESHOLD_MAX - b->threshold_limit)); + return (high & 0xFFF) - (THRESHOLD_MAX - b->threshold_limit); +} + +static ssize_t show_error_count(struct threshold_block *b, char *buf) +{ + return sprintf(buf, "%lx\n", work_on_cpu(b->cpu, local_error_count, b)); } static ssize_t store_error_count(struct threshold_block *b, const char *buf, size_t count) { - cpumask_t oldmask, newmask; - affinity_set(b->cpu, &oldmask, &newmask); - threshold_restart_bank(b, 1, 0); - affinity_restore(&oldmask); + struct thresh_restart tr = { .b = b, .reset = 1, .old_limit = 0 }; + + work_on_cpu(b->cpu, threshold_restart_bank, &tr); return 1; } @@ -463,12 +462,19 @@ out_free: return err; } +static long local_allocate_threshold_blocks(void *_bank) +{ + unsigned int *bank = _bank; + + return allocate_threshold_blocks(smp_processor_id(), *bank, 0, + MSR_IA32_MC0_MISC + *bank * 4); +} + /* symlinks sibling shared banks to first core. first core owns dir/files. */ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) { int i, err = 0; struct threshold_bank *b = NULL; - cpumask_t oldmask, newmask; char name[32]; sprintf(name, "threshold_bank%i", bank); @@ -519,11 +525,7 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) per_cpu(threshold_banks, cpu)[bank] = b; - affinity_set(cpu, &oldmask, &newmask); - err = allocate_threshold_blocks(cpu, bank, 0, - MSR_IA32_MC0_MISC + bank * 4); - affinity_restore(&oldmask); - + err = work_on_cpu(cpu, local_allocate_threshold_blocks, &bank); if (err) goto out_free; diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index 26855381790..d84a852e4cd 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -29,34 +29,17 @@ #include <mach_ipi.h> -/* This keeps a track of which one is crashing cpu. */ -static int crashing_cpu; #if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) -static atomic_t waiting_for_crash_ipi; -static int crash_nmi_callback(struct notifier_block *self, - unsigned long val, void *data) +static void kdump_nmi_callback(int cpu, struct die_args *args) { struct pt_regs *regs; #ifdef CONFIG_X86_32 struct pt_regs fixed_regs; #endif - int cpu; - if (val != DIE_NMI_IPI) - return NOTIFY_OK; - - regs = ((struct die_args *)data)->regs; - cpu = raw_smp_processor_id(); - - /* Don't do anything if this handler is invoked on crashing cpu. - * Otherwise, system will completely hang. Crashing cpu can get - * an NMI if system was initially booted with nmi_watchdog parameter. - */ - if (cpu == crashing_cpu) - return NOTIFY_STOP; - local_irq_disable(); + regs = args->regs; #ifdef CONFIG_X86_32 if (!user_mode_vm(regs)) { @@ -65,54 +48,19 @@ static int crash_nmi_callback(struct notifier_block *self, } #endif crash_save_cpu(regs, cpu); - disable_local_APIC(); - atomic_dec(&waiting_for_crash_ipi); - /* Assume hlt works */ - halt(); - for (;;) - cpu_relax(); - - return 1; -} -static void smp_send_nmi_allbutself(void) -{ - cpumask_t mask = cpu_online_map; - cpu_clear(safe_smp_processor_id(), mask); - if (!cpus_empty(mask)) - send_IPI_mask(mask, NMI_VECTOR); + disable_local_APIC(); } -static struct notifier_block crash_nmi_nb = { - .notifier_call = crash_nmi_callback, -}; - -static void nmi_shootdown_cpus(void) +static void kdump_nmi_shootdown_cpus(void) { - unsigned long msecs; - - atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1); - /* Would it be better to replace the trap vector here? */ - if (register_die_notifier(&crash_nmi_nb)) - return; /* return what? */ - /* Ensure the new callback function is set before sending - * out the NMI - */ - wmb(); + nmi_shootdown_cpus(kdump_nmi_callback); - smp_send_nmi_allbutself(); - - msecs = 1000; /* Wait at most a second for the other cpus to stop */ - while ((atomic_read(&waiting_for_crash_ipi) > 0) && msecs) { - mdelay(1); - msecs--; - } - - /* Leave the nmi callback set */ disable_local_APIC(); } + #else -static void nmi_shootdown_cpus(void) +static void kdump_nmi_shootdown_cpus(void) { /* There are no cpus to shootdown */ } @@ -131,9 +79,7 @@ void native_machine_crash_shutdown(struct pt_regs *regs) /* The kernel is broken so disable interrupts */ local_irq_disable(); - /* Make a note of crashing cpu. Will be used in NMI callback.*/ - crashing_cpu = safe_smp_processor_id(); - nmi_shootdown_cpus(); + kdump_nmi_shootdown_cpus(); lapic_shutdown(); #if defined(CONFIG_X86_IO_APIC) disable_IO_APIC(); diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 2b69994fd3a..19a8c2c0389 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -7,13 +7,12 @@ * * It manages: * - per-thread and per-cpu allocation of BTS and PEBS - * - buffer memory allocation (optional) - * - buffer overflow handling + * - buffer overflow handling (to be done) * - buffer access * * It assumes: - * - get_task_struct on all parameter tasks - * - current is allowed to trace parameter tasks + * - get_task_struct on all traced tasks + * - current is allowed to trace tasks * * * Copyright (C) 2007-2008 Intel Corporation. @@ -21,8 +20,6 @@ */ -#ifdef CONFIG_X86_DS - #include <asm/ds.h> #include <linux/errno.h> @@ -30,6 +27,7 @@ #include <linux/slab.h> #include <linux/sched.h> #include <linux/mm.h> +#include <linux/kernel.h> /* @@ -46,6 +44,33 @@ struct ds_configuration { }; static struct ds_configuration ds_cfg; +/* + * A BTS or PEBS tracer. + * + * This holds the configuration of the tracer and serves as a handle + * to identify tracers. + */ +struct ds_tracer { + /* the DS context (partially) owned by this tracer */ + struct ds_context *context; + /* the buffer provided on ds_request() and its size in bytes */ + void *buffer; + size_t size; +}; + +struct bts_tracer { + /* the common DS part */ + struct ds_tracer ds; + /* buffer overflow notification function */ + bts_ovfl_callback_t ovfl; +}; + +struct pebs_tracer { + /* the common DS part */ + struct ds_tracer ds; + /* buffer overflow notification function */ + pebs_ovfl_callback_t ovfl; +}; /* * Debug Store (DS) save area configuration (see Intel64 and IA32 @@ -109,34 +134,13 @@ static inline void ds_set(unsigned char *base, enum ds_qualifier qual, (*(unsigned long *)base) = value; } +#define DS_ALIGNMENT (1 << 3) /* BTS and PEBS buffer alignment */ -/* - * Locking is done only for allocating BTS or PEBS resources and for - * guarding context and buffer memory allocation. - * - * Most functions require the current task to own the ds context part - * they are going to access. All the locking is done when validating - * access to the context. - */ -static spinlock_t ds_lock = __SPIN_LOCK_UNLOCKED(ds_lock); /* - * Validate that the current task is allowed to access the BTS/PEBS - * buffer of the parameter task. - * - * Returns 0, if access is granted; -Eerrno, otherwise. + * Locking is done only for allocating BTS or PEBS resources. */ -static inline int ds_validate_access(struct ds_context *context, - enum ds_qualifier qual) -{ - if (!context) - return -EPERM; - - if (context->owner[qual] == current) - return 0; - - return -EPERM; -} +static spinlock_t ds_lock = __SPIN_LOCK_UNLOCKED(ds_lock); /* @@ -185,80 +189,43 @@ static inline int check_tracer(struct task_struct *task) * * Contexts are use-counted. They are allocated on first access and * deallocated when the last user puts the context. - * - * We distinguish between an allocating and a non-allocating get of a - * context: - * - the allocating get is used for requesting BTS/PEBS resources. It - * requires the caller to hold the global ds_lock. - * - the non-allocating get is used for all other cases. A - * non-existing context indicates an error. It acquires and releases - * the ds_lock itself for obtaining the context. - * - * A context and its DS configuration are allocated and deallocated - * together. A context always has a DS configuration of the - * appropriate size. */ static DEFINE_PER_CPU(struct ds_context *, system_context); #define this_system_context per_cpu(system_context, smp_processor_id()) -/* - * Returns the pointer to the parameter task's context or to the - * system-wide context, if task is NULL. - * - * Increases the use count of the returned context, if not NULL. - */ static inline struct ds_context *ds_get_context(struct task_struct *task) { - struct ds_context *context; - - spin_lock(&ds_lock); - - context = (task ? task->thread.ds_ctx : this_system_context); - if (context) - context->count++; - - spin_unlock(&ds_lock); - - return context; -} - -/* - * 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) { context = kzalloc(sizeof(*context), GFP_KERNEL); - if (!context) return NULL; - context->ds = kzalloc(ds_cfg.sizeof_ds, GFP_KERNEL); - if (!context->ds) { - kfree(context); - return NULL; - } + spin_lock_irqsave(&ds_lock, irq); - *p_context = context; + if (*p_context) { + kfree(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++; @@ -266,16 +233,14 @@ static inline struct ds_context *ds_alloc_context(struct task_struct *task) return context; } -/* - * Decreases the use count of the parameter context, if not NULL. - * Deallocates the context, if the use count reaches zero. - */ 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; @@ -288,351 +253,351 @@ static inline void ds_put_context(struct ds_context *context) if (!context->task || (context->task == current)) wrmsrl(MSR_IA32_DS_AREA, 0); - put_tracer(context->task); - - /* free any leftover buffers from tracers that did not - * deallocate them properly. */ - kfree(context->buffer[ds_bts]); - kfree(context->buffer[ds_pebs]); - kfree(context->ds); kfree(context); out: - spin_unlock(&ds_lock); + spin_unlock_irqrestore(&ds_lock, irq); } /* * Handle a buffer overflow * - * task: the task whose buffers are overflowing; - * NULL for a buffer overflow on the current cpu * context: the ds context * qual: the buffer type */ -static void ds_overflow(struct task_struct *task, struct ds_context *context, - enum ds_qualifier qual) -{ - if (!context) - return; - - if (context->callback[qual]) - (*context->callback[qual])(task); - - /* todo: do some more overflow handling */ +static void ds_overflow(struct ds_context *context, enum ds_qualifier qual) +{ + switch (qual) { + case ds_bts: { + struct bts_tracer *tracer = + container_of(context->owner[qual], + struct bts_tracer, ds); + if (tracer->ovfl) + tracer->ovfl(tracer); + } + break; + case ds_pebs: { + struct pebs_tracer *tracer = + container_of(context->owner[qual], + struct pebs_tracer, ds); + if (tracer->ovfl) + tracer->ovfl(tracer); + } + break; + } } -/* - * Allocate a non-pageable buffer of the parameter size. - * Checks the memory and the locked memory rlimit. - * - * Returns the buffer, if successful; - * NULL, if out of memory or rlimit exceeded. - * - * size: the requested buffer size in bytes - * pages (out): if not NULL, contains the number of pages reserved - */ -static inline void *ds_allocate_buffer(size_t size, unsigned int *pages) +static void ds_install_ds_config(struct ds_context *context, + enum ds_qualifier qual, + void *base, size_t size, size_t ith) { - unsigned long rlim, vm, pgsz; - void *buffer; - - pgsz = PAGE_ALIGN(size) >> PAGE_SHIFT; - - rlim = current->signal->rlim[RLIMIT_AS].rlim_cur >> PAGE_SHIFT; - vm = current->mm->total_vm + pgsz; - if (rlim < vm) - return NULL; + unsigned long buffer, adj; - rlim = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT; - vm = current->mm->locked_vm + pgsz; - if (rlim < vm) - return NULL; + /* adjust the buffer address and size to meet alignment + * constraints: + * - buffer is double-word aligned + * - size is multiple of record size + * + * We checked the size at the very beginning; we have enough + * space to do the adjustment. + */ + buffer = (unsigned long)base; - buffer = kzalloc(size, GFP_KERNEL); - if (!buffer) - return NULL; + adj = ALIGN(buffer, DS_ALIGNMENT) - buffer; + buffer += adj; + size -= adj; - current->mm->total_vm += pgsz; - current->mm->locked_vm += pgsz; + size /= ds_cfg.sizeof_rec[qual]; + size *= ds_cfg.sizeof_rec[qual]; - if (pages) - *pages = pgsz; + ds_set(context->ds, qual, ds_buffer_base, buffer); + ds_set(context->ds, qual, ds_index, buffer); + ds_set(context->ds, qual, ds_absolute_maximum, buffer + size); - return buffer; + /* The value for 'no threshold' is -1, which will set the + * threshold outside of the buffer, just like we want it. + */ + ds_set(context->ds, qual, + ds_interrupt_threshold, buffer + size - ith); } -static int ds_request(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl, enum ds_qualifier qual) +static int ds_request(struct ds_tracer *tracer, enum ds_qualifier qual, + struct task_struct *task, + void *base, size_t size, size_t th) { struct ds_context *context; - unsigned long buffer, adj; - const unsigned long alignment = (1 << 3); - int error = 0; + unsigned long irq; + int error; + error = -EOPNOTSUPP; if (!ds_cfg.sizeof_ds) - return -EOPNOTSUPP; + goto out; - /* we require some space to do alignment adjustments below */ - if (size < (alignment + ds_cfg.sizeof_rec[qual])) - return -EINVAL; + error = -EINVAL; + if (!base) + goto out; - /* buffer overflow notification is not yet implemented */ - if (ovfl) - return -EOPNOTSUPP; + /* we require some space to do alignment adjustments below */ + error = -EINVAL; + if (size < (DS_ALIGNMENT + ds_cfg.sizeof_rec[qual])) + goto out; + if (th != (size_t)-1) { + th *= ds_cfg.sizeof_rec[qual]; - spin_lock(&ds_lock); + error = -EINVAL; + if (size <= th) + goto out; + } - if (!check_tracer(task)) - return -EPERM; + tracer->buffer = base; + tracer->size = size; error = -ENOMEM; - context = ds_alloc_context(task); + context = ds_get_context(task); if (!context) - goto out_unlock; - - error = -EALREADY; - if (context->owner[qual] == current) - goto out_unlock; - error = -EPERM; - if (context->owner[qual] != NULL) - goto out_unlock; - context->owner[qual] = current; - - spin_unlock(&ds_lock); + goto out; + tracer->context = context; - error = -ENOMEM; - if (!base) { - base = ds_allocate_buffer(size, &context->pages[qual]); - if (!base) - goto out_release; - - context->buffer[qual] = base; - } - error = 0; + spin_lock_irqsave(&ds_lock, irq); - context->callback[qual] = ovfl; - - /* adjust the buffer address and size to meet alignment - * constraints: - * - buffer is double-word aligned - * - size is multiple of record size - * - * We checked the size at the very beginning; we have enough - * space to do the adjustment. - */ - buffer = (unsigned long)base; + error = -EPERM; + if (!check_tracer(task)) + goto out_unlock; + get_tracer(task); - adj = ALIGN(buffer, alignment) - buffer; - buffer += adj; - size -= adj; + error = -EPERM; + if (context->owner[qual]) + goto out_put_tracer; + context->owner[qual] = tracer; - size /= ds_cfg.sizeof_rec[qual]; - size *= ds_cfg.sizeof_rec[qual]; + spin_unlock_irqrestore(&ds_lock, irq); - ds_set(context->ds, qual, ds_buffer_base, buffer); - ds_set(context->ds, qual, ds_index, buffer); - ds_set(context->ds, qual, ds_absolute_maximum, buffer + size); - if (ovfl) { - /* todo: select a suitable interrupt threshold */ - } else - ds_set(context->ds, qual, - ds_interrupt_threshold, buffer + size + 1); + ds_install_ds_config(context, qual, base, size, th); - /* we keep the context until ds_release */ - return error; - - out_release: - context->owner[qual] = NULL; - ds_put_context(context); - return error; + return 0; + out_put_tracer: + put_tracer(task); out_unlock: - spin_unlock(&ds_lock); + spin_unlock_irqrestore(&ds_lock, irq); ds_put_context(context); + tracer->context = NULL; + out: return error; } -int ds_request_bts(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl) +struct bts_tracer *ds_request_bts(struct task_struct *task, + void *base, size_t size, + bts_ovfl_callback_t ovfl, size_t th) { - return ds_request(task, base, size, ovfl, ds_bts); -} + struct bts_tracer *tracer; + int error; -int ds_request_pebs(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl) -{ - return ds_request(task, base, size, ovfl, ds_pebs); + /* buffer overflow notification is not yet implemented */ + error = -EOPNOTSUPP; + if (ovfl) + goto out; + + error = -ENOMEM; + tracer = kzalloc(sizeof(*tracer), GFP_KERNEL); + if (!tracer) + goto out; + tracer->ovfl = ovfl; + + error = ds_request(&tracer->ds, ds_bts, task, base, size, th); + if (error < 0) + goto out_tracer; + + return tracer; + + out_tracer: + kfree(tracer); + out: + return ERR_PTR(error); } -static int ds_release(struct task_struct *task, enum ds_qualifier qual) +struct pebs_tracer *ds_request_pebs(struct task_struct *task, + void *base, size_t size, + pebs_ovfl_callback_t ovfl, size_t th) { - struct ds_context *context; + struct pebs_tracer *tracer; int error; - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) + /* buffer overflow notification is not yet implemented */ + error = -EOPNOTSUPP; + if (ovfl) + goto out; + + error = -ENOMEM; + tracer = kzalloc(sizeof(*tracer), GFP_KERNEL); + if (!tracer) goto out; + tracer->ovfl = ovfl; - kfree(context->buffer[qual]); - context->buffer[qual] = NULL; + error = ds_request(&tracer->ds, ds_pebs, task, base, size, th); + if (error < 0) + goto out_tracer; - current->mm->total_vm -= context->pages[qual]; - current->mm->locked_vm -= context->pages[qual]; - context->pages[qual] = 0; - context->owner[qual] = NULL; + return tracer; - /* - * we put the context twice: - * once for the ds_get_context - * once for the corresponding ds_request - */ - ds_put_context(context); + out_tracer: + kfree(tracer); out: - ds_put_context(context); - return error; + return ERR_PTR(error); } -int ds_release_bts(struct task_struct *task) +static void ds_release(struct ds_tracer *tracer, enum ds_qualifier qual) { - return ds_release(task, ds_bts); + BUG_ON(tracer->context->owner[qual] != tracer); + tracer->context->owner[qual] = NULL; + + put_tracer(tracer->context->task); + ds_put_context(tracer->context); } -int ds_release_pebs(struct task_struct *task) +int ds_release_bts(struct bts_tracer *tracer) { - return ds_release(task, ds_pebs); + if (!tracer) + return -EINVAL; + + ds_release(&tracer->ds, ds_bts); + kfree(tracer); + + return 0; } -static int ds_get_index(struct task_struct *task, size_t *pos, - enum ds_qualifier qual) +int ds_release_pebs(struct pebs_tracer *tracer) { - struct ds_context *context; - unsigned long base, index; - int error; + if (!tracer) + return -EINVAL; - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; + ds_release(&tracer->ds, ds_pebs); + kfree(tracer); + + return 0; +} + +static size_t ds_get_index(struct ds_context *context, enum ds_qualifier qual) +{ + unsigned long base, index; base = ds_get(context->ds, qual, ds_buffer_base); index = ds_get(context->ds, qual, ds_index); - error = ((index - base) / ds_cfg.sizeof_rec[qual]); - if (pos) - *pos = error; - out: - ds_put_context(context); - return error; + return (index - base) / ds_cfg.sizeof_rec[qual]; } -int ds_get_bts_index(struct task_struct *task, size_t *pos) +int ds_get_bts_index(struct bts_tracer *tracer, size_t *pos) { - return ds_get_index(task, pos, ds_bts); + if (!tracer) + return -EINVAL; + + if (!pos) + return -EINVAL; + + *pos = ds_get_index(tracer->ds.context, ds_bts); + + return 0; } -int ds_get_pebs_index(struct task_struct *task, size_t *pos) +int ds_get_pebs_index(struct pebs_tracer *tracer, size_t *pos) { - return ds_get_index(task, pos, ds_pebs); + if (!tracer) + return -EINVAL; + + if (!pos) + return -EINVAL; + + *pos = ds_get_index(tracer->ds.context, ds_pebs); + + return 0; } -static int ds_get_end(struct task_struct *task, size_t *pos, - enum ds_qualifier qual) +static size_t ds_get_end(struct ds_context *context, enum ds_qualifier qual) { - struct ds_context *context; - unsigned long base, end; - int error; - - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; + unsigned long base, max; base = ds_get(context->ds, qual, ds_buffer_base); - end = ds_get(context->ds, qual, ds_absolute_maximum); + max = ds_get(context->ds, qual, ds_absolute_maximum); - error = ((end - base) / ds_cfg.sizeof_rec[qual]); - if (pos) - *pos = error; - out: - ds_put_context(context); - return error; + return (max - base) / ds_cfg.sizeof_rec[qual]; } -int ds_get_bts_end(struct task_struct *task, size_t *pos) +int ds_get_bts_end(struct bts_tracer *tracer, size_t *pos) { - return ds_get_end(task, pos, ds_bts); + if (!tracer) + return -EINVAL; + + if (!pos) + return -EINVAL; + + *pos = ds_get_end(tracer->ds.context, ds_bts); + + return 0; } -int ds_get_pebs_end(struct task_struct *task, size_t *pos) +int ds_get_pebs_end(struct pebs_tracer *tracer, size_t *pos) { - return ds_get_end(task, pos, ds_pebs); + if (!tracer) + return -EINVAL; + + if (!pos) + return -EINVAL; + + *pos = ds_get_end(tracer->ds.context, ds_pebs); + + return 0; } -static int ds_access(struct task_struct *task, size_t index, - const void **record, enum ds_qualifier qual) +static int ds_access(struct ds_context *context, enum ds_qualifier qual, + size_t index, const void **record) { - struct ds_context *context; unsigned long base, idx; - int error; if (!record) return -EINVAL; - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; - base = ds_get(context->ds, qual, ds_buffer_base); idx = base + (index * ds_cfg.sizeof_rec[qual]); - error = -EINVAL; if (idx > ds_get(context->ds, qual, ds_absolute_maximum)) - goto out; + return -EINVAL; *record = (const void *)idx; - error = ds_cfg.sizeof_rec[qual]; - out: - ds_put_context(context); - return error; + + return ds_cfg.sizeof_rec[qual]; } -int ds_access_bts(struct task_struct *task, size_t index, const void **record) +int ds_access_bts(struct bts_tracer *tracer, size_t index, + const void **record) { - return ds_access(task, index, record, ds_bts); + if (!tracer) + return -EINVAL; + + return ds_access(tracer->ds.context, ds_bts, index, record); } -int ds_access_pebs(struct task_struct *task, size_t index, const void **record) +int ds_access_pebs(struct pebs_tracer *tracer, size_t index, + const void **record) { - return ds_access(task, index, record, ds_pebs); + if (!tracer) + return -EINVAL; + + return ds_access(tracer->ds.context, ds_pebs, index, record); } -static int ds_write(struct task_struct *task, const void *record, size_t size, - enum ds_qualifier qual, int force) +static int ds_write(struct ds_context *context, enum ds_qualifier qual, + const void *record, size_t size) { - struct ds_context *context; - int error; + int bytes_written = 0; if (!record) return -EINVAL; - error = -EPERM; - context = ds_get_context(task); - if (!context) - goto out; - - if (!force) { - error = ds_validate_access(context, qual); - if (error < 0) - goto out; - } - - error = 0; while (size) { unsigned long base, index, end, write_end, int_th; unsigned long write_size, adj_write_size; @@ -660,14 +625,14 @@ static int ds_write(struct task_struct *task, const void *record, size_t size, write_end = end; if (write_end <= index) - goto out; + break; write_size = min((unsigned long) size, write_end - index); memcpy((void *)index, record, write_size); record = (const char *)record + write_size; - size -= write_size; - error += write_size; + size -= write_size; + bytes_written += write_size; adj_write_size = write_size / ds_cfg.sizeof_rec[qual]; adj_write_size *= ds_cfg.sizeof_rec[qual]; @@ -682,47 +647,32 @@ static int ds_write(struct task_struct *task, const void *record, size_t size, ds_set(context->ds, qual, ds_index, index); if (index >= int_th) - ds_overflow(task, context, qual); + ds_overflow(context, qual); } - out: - ds_put_context(context); - return error; + return bytes_written; } -int ds_write_bts(struct task_struct *task, const void *record, size_t size) +int ds_write_bts(struct bts_tracer *tracer, const void *record, size_t size) { - return ds_write(task, record, size, ds_bts, /* force = */ 0); -} + if (!tracer) + return -EINVAL; -int ds_write_pebs(struct task_struct *task, const void *record, size_t size) -{ - return ds_write(task, record, size, ds_pebs, /* force = */ 0); + return ds_write(tracer->ds.context, ds_bts, record, size); } -int ds_unchecked_write_bts(struct task_struct *task, - const void *record, size_t size) +int ds_write_pebs(struct pebs_tracer *tracer, const void *record, size_t size) { - return ds_write(task, record, size, ds_bts, /* force = */ 1); -} + if (!tracer) + return -EINVAL; -int ds_unchecked_write_pebs(struct task_struct *task, - const void *record, size_t size) -{ - return ds_write(task, record, size, ds_pebs, /* force = */ 1); + return ds_write(tracer->ds.context, ds_pebs, record, size); } -static int ds_reset_or_clear(struct task_struct *task, - enum ds_qualifier qual, int clear) +static void ds_reset_or_clear(struct ds_context *context, + enum ds_qualifier qual, int clear) { - struct ds_context *context; unsigned long base, end; - int error; - - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; base = ds_get(context->ds, qual, ds_buffer_base); end = ds_get(context->ds, qual, ds_absolute_maximum); @@ -731,89 +681,100 @@ static int ds_reset_or_clear(struct task_struct *task, memset((void *)base, 0, end - base); ds_set(context->ds, qual, ds_index, base); - - error = 0; - out: - ds_put_context(context); - return error; } -int ds_reset_bts(struct task_struct *task) +int ds_reset_bts(struct bts_tracer *tracer) { - return ds_reset_or_clear(task, ds_bts, /* clear = */ 0); + if (!tracer) + return -EINVAL; + + ds_reset_or_clear(tracer->ds.context, ds_bts, /* clear = */ 0); + + return 0; } -int ds_reset_pebs(struct task_struct *task) +int ds_reset_pebs(struct pebs_tracer *tracer) { - return ds_reset_or_clear(task, ds_pebs, /* clear = */ 0); + if (!tracer) + return -EINVAL; + + ds_reset_or_clear(tracer->ds.context, ds_pebs, /* clear = */ 0); + + return 0; } -int ds_clear_bts(struct task_struct *task) +int ds_clear_bts(struct bts_tracer *tracer) { - return ds_reset_or_clear(task, ds_bts, /* clear = */ 1); + if (!tracer) + return -EINVAL; + + ds_reset_or_clear(tracer->ds.context, ds_bts, /* clear = */ 1); + + return 0; } -int ds_clear_pebs(struct task_struct *task) +int ds_clear_pebs(struct pebs_tracer *tracer) { - return ds_reset_or_clear(task, ds_pebs, /* clear = */ 1); + if (!tracer) + return -EINVAL; + + ds_reset_or_clear(tracer->ds.context, ds_pebs, /* clear = */ 1); + + return 0; } -int ds_get_pebs_reset(struct task_struct *task, u64 *value) +int ds_get_pebs_reset(struct pebs_tracer *tracer, u64 *value) { - struct ds_context *context; - int error; + if (!tracer) + return -EINVAL; if (!value) return -EINVAL; - context = ds_get_context(task); - error = ds_validate_access(context, ds_pebs); - if (error < 0) - goto out; + *value = *(u64 *)(tracer->ds.context->ds + (ds_cfg.sizeof_field * 8)); - *value = *(u64 *)(context->ds + (ds_cfg.sizeof_field * 8)); - - error = 0; - out: - ds_put_context(context); - return error; + return 0; } -int ds_set_pebs_reset(struct task_struct *task, u64 value) +int ds_set_pebs_reset(struct pebs_tracer *tracer, u64 value) { - struct ds_context *context; - int error; - - context = ds_get_context(task); - error = ds_validate_access(context, ds_pebs); - if (error < 0) - goto out; + if (!tracer) + return -EINVAL; - *(u64 *)(context->ds + (ds_cfg.sizeof_field * 8)) = value; + *(u64 *)(tracer->ds.context->ds + (ds_cfg.sizeof_field * 8)) = value; - error = 0; - out: - ds_put_context(context); - return error; + return 0; } 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 ds_configure(const struct ds_configuration *cfg) { ds_cfg = *cfg; + + printk(KERN_INFO "DS available\n"); + + BUG_ON(MAX_SIZEOF_DS < ds_cfg.sizeof_ds); } void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) @@ -821,17 +782,16 @@ void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) switch (c->x86) { case 0x6: switch (c->x86_model) { + case 0 ... 0xC: + /* sorry, don't know about them */ + break; case 0xD: case 0xE: /* Pentium M */ ds_configure(&ds_cfg_var); break; - case 0xF: /* Core2 */ - case 0x1C: /* Atom */ + default: /* Core2, Atom, ... */ ds_configure(&ds_cfg_64); break; - default: - /* sorry, don't know about them */ - break; } break; case 0xF: @@ -858,7 +818,8 @@ void ds_free(struct ds_context *context) * is dying. There should not be any user of that context left * to disturb us, anymore. */ unsigned long leftovers = context->count; - while (leftovers--) + while (leftovers--) { + put_tracer(context->task); ds_put_context(context); + } } -#endif /* CONFIG_X86_DS */ diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c new file mode 100644 index 00000000000..6b1f6f6f866 --- /dev/null +++ b/arch/x86/kernel/dumpstack.c @@ -0,0 +1,351 @@ +/* + * Copyright (C) 1991, 1992 Linus Torvalds + * Copyright (C) 2000, 2001, 2002 Andi Kleen, SuSE Labs + */ +#include <linux/kallsyms.h> +#include <linux/kprobes.h> +#include <linux/uaccess.h> +#include <linux/utsname.h> +#include <linux/hardirq.h> +#include <linux/kdebug.h> +#include <linux/module.h> +#include <linux/ptrace.h> +#include <linux/kexec.h> +#include <linux/bug.h> +#include <linux/nmi.h> +#include <linux/sysfs.h> + +#include <asm/stacktrace.h> + +#include "dumpstack.h" + +int panic_on_unrecovered_nmi; +unsigned int code_bytes = 64; +int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE; +static int die_counter; + +void printk_address(unsigned long address, int reliable) +{ + printk(" [<%p>] %s%pS\n", (void *) address, + reliable ? "" : "? ", (void *) address); +} + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +static void +print_ftrace_graph_addr(unsigned long addr, void *data, + const struct stacktrace_ops *ops, + struct thread_info *tinfo, int *graph) +{ + struct task_struct *task = tinfo->task; + unsigned long ret_addr; + int index = task->curr_ret_stack; + + if (addr != (unsigned long)return_to_handler) + return; + + if (!task->ret_stack || index < *graph) + return; + + index -= *graph; + ret_addr = task->ret_stack[index].ret; + + ops->address(data, ret_addr, 1); + + (*graph)++; +} +#else +static inline void +print_ftrace_graph_addr(unsigned long addr, void *data, + const struct stacktrace_ops *ops, + struct thread_info *tinfo, int *graph) +{ } +#endif + +/* + * x86-64 can have up to three kernel stacks: + * process stack + * interrupt stack + * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack + */ + +static inline int valid_stack_ptr(struct thread_info *tinfo, + void *p, unsigned int size, void *end) +{ + void *t = tinfo; + if (end) { + if (p < end && p >= (end-THREAD_SIZE)) + return 1; + else + return 0; + } + return p > t && p < t + THREAD_SIZE - size; +} + +unsigned long +print_context_stack(struct thread_info *tinfo, + unsigned long *stack, unsigned long bp, + const struct stacktrace_ops *ops, void *data, + unsigned long *end, int *graph) +{ + struct stack_frame *frame = (struct stack_frame *)bp; + + while (valid_stack_ptr(tinfo, stack, sizeof(*stack), end)) { + unsigned long addr; + + addr = *stack; + if (__kernel_text_address(addr)) { + if ((unsigned long) stack == bp + sizeof(long)) { + ops->address(data, addr, 1); + frame = frame->next_frame; + bp = (unsigned long) frame; + } else { + ops->address(data, addr, bp == 0); + } + print_ftrace_graph_addr(addr, data, ops, tinfo, graph); + } + stack++; + } + return bp; +} + + +static void +print_trace_warning_symbol(void *data, char *msg, unsigned long symbol) +{ + printk(data); + print_symbol(msg, symbol); + printk("\n"); +} + +static void print_trace_warning(void *data, char *msg) +{ + printk("%s%s\n", (char *)data, msg); +} + +static int print_trace_stack(void *data, char *name) +{ + printk("%s <%s> ", (char *)data, name); + return 0; +} + +/* + * Print one address/symbol entries per line. + */ +static void print_trace_address(void *data, unsigned long addr, int reliable) +{ + touch_nmi_watchdog(); + printk(data); + printk_address(addr, reliable); +} + +static const struct stacktrace_ops print_trace_ops = { + .warning = print_trace_warning, + .warning_symbol = print_trace_warning_symbol, + .stack = print_trace_stack, + .address = print_trace_address, +}; + +void +show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, + unsigned long *stack, unsigned long bp, char *log_lvl) +{ + printk("%sCall Trace:\n", log_lvl); + dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl); +} + +void show_trace(struct task_struct *task, struct pt_regs *regs, + unsigned long *stack, unsigned long bp) +{ + show_trace_log_lvl(task, regs, stack, bp, ""); +} + +void show_stack(struct task_struct *task, unsigned long *sp) +{ + show_stack_log_lvl(task, NULL, sp, 0, ""); +} + +/* + * The architecture-independent dump_stack generator + */ +void dump_stack(void) +{ + unsigned long bp = 0; + unsigned long stack; + +#ifdef CONFIG_FRAME_POINTER + if (!bp) + get_bp(bp); +#endif + + printk("Pid: %d, comm: %.20s %s %s %.*s\n", + current->pid, current->comm, print_tainted(), + init_utsname()->release, + (int)strcspn(init_utsname()->version, " "), + init_utsname()->version); + show_trace(NULL, NULL, &stack, bp); +} +EXPORT_SYMBOL(dump_stack); + +static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED; +static int die_owner = -1; +static unsigned int die_nest_count; + +unsigned __kprobes long oops_begin(void) +{ + int cpu; + unsigned long flags; + + oops_enter(); + + /* racy, but better than risking deadlock. */ + raw_local_irq_save(flags); + cpu = smp_processor_id(); + if (!__raw_spin_trylock(&die_lock)) { + if (cpu == die_owner) + /* nested oops. should stop eventually */; + else + __raw_spin_lock(&die_lock); + } + die_nest_count++; + die_owner = cpu; + console_verbose(); + bust_spinlocks(1); + return flags; +} + +void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) +{ + if (regs && kexec_should_crash(current)) + crash_kexec(regs); + + bust_spinlocks(0); + die_owner = -1; + add_taint(TAINT_DIE); + die_nest_count--; + if (!die_nest_count) + /* Nest count reaches zero, release the lock. */ + __raw_spin_unlock(&die_lock); + raw_local_irq_restore(flags); + oops_exit(); + + if (!signr) + return; + if (in_interrupt()) + panic("Fatal exception in interrupt"); + if (panic_on_oops) + panic("Fatal exception"); + do_exit(signr); +} + +int __kprobes __die(const char *str, struct pt_regs *regs, long err) +{ +#ifdef CONFIG_X86_32 + unsigned short ss; + unsigned long sp; +#endif + printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter); +#ifdef CONFIG_PREEMPT + printk("PREEMPT "); +#endif +#ifdef CONFIG_SMP + printk("SMP "); +#endif +#ifdef CONFIG_DEBUG_PAGEALLOC + printk("DEBUG_PAGEALLOC"); +#endif + printk("\n"); + sysfs_printk_last_file(); + if (notify_die(DIE_OOPS, str, regs, err, + current->thread.trap_no, SIGSEGV) == NOTIFY_STOP) + return 1; + + show_registers(regs); +#ifdef CONFIG_X86_32 + sp = (unsigned long) (®s->sp); + savesegment(ss, ss); + if (user_mode(regs)) { + sp = regs->sp; + ss = regs->ss & 0xffff; + } + printk(KERN_EMERG "EIP: [<%08lx>] ", regs->ip); + print_symbol("%s", regs->ip); + printk(" SS:ESP %04x:%08lx\n", ss, sp); +#else + /* Executive summary in case the oops scrolled away */ + printk(KERN_ALERT "RIP "); + printk_address(regs->ip, 1); + printk(" RSP <%016lx>\n", regs->sp); +#endif + return 0; +} + +/* + * This is gone through when something in the kernel has done something bad + * and is about to be terminated: + */ +void die(const char *str, struct pt_regs *regs, long err) +{ + unsigned long flags = oops_begin(); + int sig = SIGSEGV; + + if (!user_mode_vm(regs)) + report_bug(regs->ip, regs); + + if (__die(str, regs, err)) + sig = 0; + oops_end(flags, regs, sig); +} + +void notrace __kprobes +die_nmi(char *str, struct pt_regs *regs, int do_panic) +{ + unsigned long flags; + + if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP) + return; + + /* + * We are in trouble anyway, lets at least try + * to get a message out. + */ + flags = oops_begin(); + printk(KERN_EMERG "%s", str); + printk(" on CPU%d, ip %08lx, registers:\n", + smp_processor_id(), regs->ip); + show_registers(regs); + oops_end(flags, regs, 0); + if (do_panic || panic_on_oops) + panic("Non maskable interrupt"); + nmi_exit(); + local_irq_enable(); + do_exit(SIGBUS); +} + +static int __init oops_setup(char *s) +{ + if (!s) + return -EINVAL; + if (!strcmp(s, "panic")) + panic_on_oops = 1; + return 0; +} +early_param("oops", oops_setup); + +static int __init kstack_setup(char *s) +{ + if (!s) + return -EINVAL; + kstack_depth_to_print = simple_strtoul(s, NULL, 0); + return 0; +} +early_param("kstack", kstack_setup); + +static int __init code_bytes_setup(char *s) +{ + code_bytes = simple_strtoul(s, NULL, 0); + if (code_bytes > 8192) + code_bytes = 8192; + + return 1; +} +__setup("code_bytes=", code_bytes_setup); diff --git a/arch/x86/kernel/dumpstack.h b/arch/x86/kernel/dumpstack.h new file mode 100644 index 00000000000..da87590b869 --- /dev/null +++ b/arch/x86/kernel/dumpstack.h @@ -0,0 +1,39 @@ +/* + * Copyright (C) 1991, 1992 Linus Torvalds + * Copyright (C) 2000, 2001, 2002 Andi Kleen, SuSE Labs + */ + +#ifndef DUMPSTACK_H +#define DUMPSTACK_H + +#ifdef CONFIG_X86_32 +#define STACKSLOTS_PER_LINE 8 +#define get_bp(bp) asm("movl %%ebp, %0" : "=r" (bp) :) +#else +#define STACKSLOTS_PER_LINE 4 +#define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :) +#endif + +extern unsigned long +print_context_stack(struct thread_info *tinfo, + unsigned long *stack, unsigned long bp, + const struct stacktrace_ops *ops, void *data, + unsigned long *end, int *graph); + +extern void +show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, + unsigned long *stack, unsigned long bp, char *log_lvl); + +extern void +show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, + unsigned long *sp, unsigned long bp, char *log_lvl); + +extern unsigned int code_bytes; +extern int kstack_depth_to_print; + +/* The form of the top of the frame on the stack */ +struct stack_frame { + struct stack_frame *next_frame; + unsigned long return_address; +}; +#endif diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index b3614752197..d593cd1f58d 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -17,69 +17,14 @@ #include <asm/stacktrace.h> -#define STACKSLOTS_PER_LINE 8 -#define get_bp(bp) asm("movl %%ebp, %0" : "=r" (bp) :) - -int panic_on_unrecovered_nmi; -int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE; -static unsigned int code_bytes = 64; -static int die_counter; - -void printk_address(unsigned long address, int reliable) -{ - printk(" [<%p>] %s%pS\n", (void *) address, - reliable ? "" : "? ", (void *) address); -} - -static inline int valid_stack_ptr(struct thread_info *tinfo, - void *p, unsigned int size, void *end) -{ - void *t = tinfo; - if (end) { - if (p < end && p >= (end-THREAD_SIZE)) - return 1; - else - return 0; - } - return p > t && p < t + THREAD_SIZE - size; -} - -/* The form of the top of the frame on the stack */ -struct stack_frame { - struct stack_frame *next_frame; - unsigned long return_address; -}; - -static inline unsigned long -print_context_stack(struct thread_info *tinfo, - unsigned long *stack, unsigned long bp, - const struct stacktrace_ops *ops, void *data, - unsigned long *end) -{ - struct stack_frame *frame = (struct stack_frame *)bp; - - while (valid_stack_ptr(tinfo, stack, sizeof(*stack), end)) { - unsigned long addr; - - addr = *stack; - if (__kernel_text_address(addr)) { - if ((unsigned long) stack == bp + sizeof(long)) { - ops->address(data, addr, 1); - frame = frame->next_frame; - bp = (unsigned long) frame; - } else { - ops->address(data, addr, bp == 0); - } - } - stack++; - } - return bp; -} +#include "dumpstack.h" void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *stack, unsigned long bp, const struct stacktrace_ops *ops, void *data) { + int graph = 0; + if (!task) task = current; @@ -107,7 +52,8 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, context = (struct thread_info *) ((unsigned long)stack & (~(THREAD_SIZE - 1))); - bp = print_context_stack(context, stack, bp, ops, data, NULL); + bp = print_context_stack(context, stack, bp, ops, + data, NULL, &graph); stack = (unsigned long *)context->previous_esp; if (!stack) @@ -119,57 +65,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, } EXPORT_SYMBOL(dump_trace); -static void -print_trace_warning_symbol(void *data, char *msg, unsigned long symbol) -{ - printk(data); - print_symbol(msg, symbol); - printk("\n"); -} - -static void print_trace_warning(void *data, char *msg) -{ - printk("%s%s\n", (char *)data, msg); -} - -static int print_trace_stack(void *data, char *name) -{ - printk("%s <%s> ", (char *)data, name); - return 0; -} - -/* - * Print one address/symbol entries per line. - */ -static void print_trace_address(void *data, unsigned long addr, int reliable) -{ - touch_nmi_watchdog(); - printk(data); - printk_address(addr, reliable); -} - -static const struct stacktrace_ops print_trace_ops = { - .warning = print_trace_warning, - .warning_symbol = print_trace_warning_symbol, - .stack = print_trace_stack, - .address = print_trace_address, -}; - -static void -show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp, char *log_lvl) -{ - printk("%sCall Trace:\n", log_lvl); - dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl); -} - -void show_trace(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp) -{ - show_trace_log_lvl(task, regs, stack, bp, ""); -} - -static void +void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, unsigned long *sp, unsigned long bp, char *log_lvl) { @@ -196,33 +92,6 @@ show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, show_trace_log_lvl(task, regs, sp, bp, log_lvl); } -void show_stack(struct task_struct *task, unsigned long *sp) -{ - show_stack_log_lvl(task, NULL, sp, 0, ""); -} - -/* - * The architecture-independent dump_stack generator - */ -void dump_stack(void) -{ - unsigned long bp = 0; - unsigned long stack; - -#ifdef CONFIG_FRAME_POINTER - if (!bp) - get_bp(bp); -#endif - - printk("Pid: %d, comm: %.20s %s %s %.*s\n", - current->pid, current->comm, print_tainted(), - init_utsname()->release, - (int)strcspn(init_utsname()->version, " "), - init_utsname()->version); - show_trace(NULL, NULL, &stack, bp); -} - -EXPORT_SYMBOL(dump_stack); void show_registers(struct pt_regs *regs) { @@ -283,167 +152,3 @@ int is_valid_bugaddr(unsigned long ip) return ud2 == 0x0b0f; } -static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED; -static int die_owner = -1; -static unsigned int die_nest_count; - -unsigned __kprobes long oops_begin(void) -{ - unsigned long flags; - - oops_enter(); - - if (die_owner != raw_smp_processor_id()) { - console_verbose(); - raw_local_irq_save(flags); - __raw_spin_lock(&die_lock); - die_owner = smp_processor_id(); - die_nest_count = 0; - bust_spinlocks(1); - } else { - raw_local_irq_save(flags); - } - die_nest_count++; - return flags; -} - -void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) -{ - bust_spinlocks(0); - die_owner = -1; - add_taint(TAINT_DIE); - __raw_spin_unlock(&die_lock); - raw_local_irq_restore(flags); - - if (!regs) - return; - - if (kexec_should_crash(current)) - crash_kexec(regs); - if (in_interrupt()) - panic("Fatal exception in interrupt"); - if (panic_on_oops) - panic("Fatal exception"); - oops_exit(); - do_exit(signr); -} - -int __kprobes __die(const char *str, struct pt_regs *regs, long err) -{ - unsigned short ss; - unsigned long sp; - - printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter); -#ifdef CONFIG_PREEMPT - printk("PREEMPT "); -#endif -#ifdef CONFIG_SMP - printk("SMP "); -#endif -#ifdef CONFIG_DEBUG_PAGEALLOC - printk("DEBUG_PAGEALLOC"); -#endif - printk("\n"); - sysfs_printk_last_file(); - if (notify_die(DIE_OOPS, str, regs, err, - current->thread.trap_no, SIGSEGV) == NOTIFY_STOP) - return 1; - - show_registers(regs); - /* Executive summary in case the oops scrolled away */ - sp = (unsigned long) (®s->sp); - savesegment(ss, ss); - if (user_mode(regs)) { - sp = regs->sp; - ss = regs->ss & 0xffff; - } - printk(KERN_EMERG "EIP: [<%08lx>] ", regs->ip); - print_symbol("%s", regs->ip); - printk(" SS:ESP %04x:%08lx\n", ss, sp); - return 0; -} - -/* - * This is gone through when something in the kernel has done something bad - * and is about to be terminated: - */ -void die(const char *str, struct pt_regs *regs, long err) -{ - unsigned long flags = oops_begin(); - - if (die_nest_count < 3) { - report_bug(regs->ip, regs); - - if (__die(str, regs, err)) - regs = NULL; - } else { - printk(KERN_EMERG "Recursive die() failure, output suppressed\n"); - } - - oops_end(flags, regs, SIGSEGV); -} - -static DEFINE_SPINLOCK(nmi_print_lock); - -void notrace __kprobes -die_nmi(char *str, struct pt_regs *regs, int do_panic) -{ - if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP) - return; - - spin_lock(&nmi_print_lock); - /* - * We are in trouble anyway, lets at least try - * to get a message out: - */ - bust_spinlocks(1); - printk(KERN_EMERG "%s", str); - printk(" on CPU%d, ip %08lx, registers:\n", - smp_processor_id(), regs->ip); - show_registers(regs); - if (do_panic) - panic("Non maskable interrupt"); - console_silent(); - spin_unlock(&nmi_print_lock); - - /* - * If we are in kernel we are probably nested up pretty bad - * and might aswell get out now while we still can: - */ - if (!user_mode_vm(regs)) { - current->thread.trap_no = 2; - crash_kexec(regs); - } - - bust_spinlocks(0); - do_exit(SIGSEGV); -} - -static int __init oops_setup(char *s) -{ - if (!s) - return -EINVAL; - if (!strcmp(s, "panic")) - panic_on_oops = 1; - return 0; -} -early_param("oops", oops_setup); - -static int __init kstack_setup(char *s) -{ - if (!s) - return -EINVAL; - kstack_depth_to_print = simple_strtoul(s, NULL, 0); - return 0; -} -early_param("kstack", kstack_setup); - -static int __init code_bytes_setup(char *s) -{ - code_bytes = simple_strtoul(s, NULL, 0); - if (code_bytes > 8192) - code_bytes = 8192; - - return 1; -} -__setup("code_bytes=", code_bytes_setup); diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 96a5db7da8a..c302d070704 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -17,19 +17,7 @@ #include <asm/stacktrace.h> -#define STACKSLOTS_PER_LINE 4 -#define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :) - -int panic_on_unrecovered_nmi; -int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE; -static unsigned int code_bytes = 64; -static int die_counter; - -void printk_address(unsigned long address, int reliable) -{ - printk(" [<%p>] %s%pS\n", (void *) address, - reliable ? "" : "? ", (void *) address); -} +#include "dumpstack.h" static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, unsigned *usedp, char **idp) @@ -113,51 +101,6 @@ static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack */ -static inline int valid_stack_ptr(struct thread_info *tinfo, - void *p, unsigned int size, void *end) -{ - void *t = tinfo; - if (end) { - if (p < end && p >= (end-THREAD_SIZE)) - return 1; - else - return 0; - } - return p > t && p < t + THREAD_SIZE - size; -} - -/* The form of the top of the frame on the stack */ -struct stack_frame { - struct stack_frame *next_frame; - unsigned long return_address; -}; - -static inline unsigned long -print_context_stack(struct thread_info *tinfo, - unsigned long *stack, unsigned long bp, - const struct stacktrace_ops *ops, void *data, - unsigned long *end) -{ - struct stack_frame *frame = (struct stack_frame *)bp; - - while (valid_stack_ptr(tinfo, stack, sizeof(*stack), end)) { - unsigned long addr; - - addr = *stack; - if (__kernel_text_address(addr)) { - if ((unsigned long) stack == bp + sizeof(long)) { - ops->address(data, addr, 1); - frame = frame->next_frame; - bp = (unsigned long) frame; - } else { - ops->address(data, addr, bp == 0); - } - } - stack++; - } - return bp; -} - void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *stack, unsigned long bp, const struct stacktrace_ops *ops, void *data) @@ -166,6 +109,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *irqstack_end = (unsigned long *)cpu_pda(cpu)->irqstackptr; unsigned used = 0; struct thread_info *tinfo; + int graph = 0; if (!task) task = current; @@ -206,7 +150,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, break; bp = print_context_stack(tinfo, stack, bp, ops, - data, estack_end); + data, estack_end, &graph); ops->stack(data, "<EOE>"); /* * We link to the next stack via the @@ -225,7 +169,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, if (ops->stack(data, "IRQ") < 0) break; bp = print_context_stack(tinfo, stack, bp, - ops, data, irqstack_end); + ops, data, irqstack_end, &graph); /* * We link to the next stack (which would be * the process stack normally) the last @@ -243,62 +187,12 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, /* * This handles the process stack: */ - bp = print_context_stack(tinfo, stack, bp, ops, data, NULL); + bp = print_context_stack(tinfo, stack, bp, ops, data, NULL, &graph); put_cpu(); } EXPORT_SYMBOL(dump_trace); -static void -print_trace_warning_symbol(void *data, char *msg, unsigned long symbol) -{ - printk(data); - print_symbol(msg, symbol); - printk("\n"); -} - -static void print_trace_warning(void *data, char *msg) -{ - printk("%s%s\n", (char *)data, msg); -} - -static int print_trace_stack(void *data, char *name) -{ - printk("%s <%s> ", (char *)data, name); - return 0; -} - -/* - * Print one address/symbol entries per line. - */ -static void print_trace_address(void *data, unsigned long addr, int reliable) -{ - touch_nmi_watchdog(); - printk(data); - printk_address(addr, reliable); -} - -static const struct stacktrace_ops print_trace_ops = { - .warning = print_trace_warning, - .warning_symbol = print_trace_warning_symbol, - .stack = print_trace_stack, - .address = print_trace_address, -}; - -static void -show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp, char *log_lvl) -{ - printk("%sCall Trace:\n", log_lvl); - dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl); -} - -void show_trace(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp) -{ - show_trace_log_lvl(task, regs, stack, bp, ""); -} - -static void +void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, unsigned long *sp, unsigned long bp, char *log_lvl) { @@ -342,33 +236,6 @@ show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, show_trace_log_lvl(task, regs, sp, bp, log_lvl); } -void show_stack(struct task_struct *task, unsigned long *sp) -{ - show_stack_log_lvl(task, NULL, sp, 0, ""); -} - -/* - * The architecture-independent dump_stack generator - */ -void dump_stack(void) -{ - unsigned long bp = 0; - unsigned long stack; - -#ifdef CONFIG_FRAME_POINTER - if (!bp) - get_bp(bp); -#endif - - printk("Pid: %d, comm: %.20s %s %s %.*s\n", - current->pid, current->comm, print_tainted(), - init_utsname()->release, - (int)strcspn(init_utsname()->version, " "), - init_utsname()->version); - show_trace(NULL, NULL, &stack, bp); -} -EXPORT_SYMBOL(dump_stack); - void show_registers(struct pt_regs *regs) { int i; @@ -429,147 +296,3 @@ int is_valid_bugaddr(unsigned long ip) return ud2 == 0x0b0f; } -static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED; -static int die_owner = -1; -static unsigned int die_nest_count; - -unsigned __kprobes long oops_begin(void) -{ - int cpu; - unsigned long flags; - - oops_enter(); - - /* racy, but better than risking deadlock. */ - raw_local_irq_save(flags); - cpu = smp_processor_id(); - if (!__raw_spin_trylock(&die_lock)) { - if (cpu == die_owner) - /* nested oops. should stop eventually */; - else - __raw_spin_lock(&die_lock); - } - die_nest_count++; - die_owner = cpu; - console_verbose(); - bust_spinlocks(1); - return flags; -} - -void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) -{ - die_owner = -1; - bust_spinlocks(0); - die_nest_count--; - if (!die_nest_count) - /* Nest count reaches zero, release the lock. */ - __raw_spin_unlock(&die_lock); - raw_local_irq_restore(flags); - if (!regs) { - oops_exit(); - return; - } - if (in_interrupt()) - panic("Fatal exception in interrupt"); - if (panic_on_oops) - panic("Fatal exception"); - oops_exit(); - do_exit(signr); -} - -int __kprobes __die(const char *str, struct pt_regs *regs, long err) -{ - printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter); -#ifdef CONFIG_PREEMPT - printk("PREEMPT "); -#endif -#ifdef CONFIG_SMP - printk("SMP "); -#endif -#ifdef CONFIG_DEBUG_PAGEALLOC - printk("DEBUG_PAGEALLOC"); -#endif - printk("\n"); - sysfs_printk_last_file(); - if (notify_die(DIE_OOPS, str, regs, err, - current->thread.trap_no, SIGSEGV) == NOTIFY_STOP) - return 1; - - show_registers(regs); - add_taint(TAINT_DIE); - /* Executive summary in case the oops scrolled away */ - printk(KERN_ALERT "RIP "); - printk_address(regs->ip, 1); - printk(" RSP <%016lx>\n", regs->sp); - if (kexec_should_crash(current)) - crash_kexec(regs); - return 0; -} - -void die(const char *str, struct pt_regs *regs, long err) -{ - unsigned long flags = oops_begin(); - - if (!user_mode(regs)) - report_bug(regs->ip, regs); - - if (__die(str, regs, err)) - regs = NULL; - oops_end(flags, regs, SIGSEGV); -} - -notrace __kprobes void -die_nmi(char *str, struct pt_regs *regs, int do_panic) -{ - unsigned long flags; - - if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP) - return; - - flags = oops_begin(); - /* - * We are in trouble anyway, lets at least try - * to get a message out. - */ - printk(KERN_EMERG "%s", str); - printk(" on CPU%d, ip %08lx, registers:\n", - smp_processor_id(), regs->ip); - show_registers(regs); - if (kexec_should_crash(current)) - crash_kexec(regs); - if (do_panic || panic_on_oops) - panic("Non maskable interrupt"); - oops_end(flags, NULL, SIGBUS); - nmi_exit(); - local_irq_enable(); - do_exit(SIGBUS); -} - -static int __init oops_setup(char *s) -{ - if (!s) - return -EINVAL; - if (!strcmp(s, "panic")) - panic_on_oops = 1; - return 0; -} -early_param("oops", oops_setup); - -static int __init kstack_setup(char *s) -{ - if (!s) - return -EINVAL; - kstack_depth_to_print = simple_strtoul(s, NULL, 0); - return 0; -} -early_param("kstack", kstack_setup); - -static int __init code_bytes_setup(char *s) -{ - code_bytes = simple_strtoul(s, NULL, 0); - if (code_bytes > 8192) - code_bytes = 8192; - - return 1; -} -__setup("code_bytes=", code_bytes_setup); diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index 3ce029ffaa5..1b894b72c0f 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -188,20 +188,6 @@ static void __init ati_bugs_contd(int num, int slot, int func) } #endif -#ifdef CONFIG_DMAR -static void __init intel_g33_dmar(int num, int slot, int func) -{ - struct acpi_table_header *dmar_tbl; - acpi_status status; - - status = acpi_get_table(ACPI_SIG_DMAR, 0, &dmar_tbl); - if (ACPI_SUCCESS(status)) { - printk(KERN_INFO "BIOS BUG: DMAR advertised on Intel G31/G33 chipset -- ignoring\n"); - dmar_disabled = 1; - } -} -#endif - #define QFLAG_APPLY_ONCE 0x1 #define QFLAG_APPLIED 0x2 #define QFLAG_DONE (QFLAG_APPLY_ONCE|QFLAG_APPLIED) @@ -225,10 +211,6 @@ static struct chipset early_qrk[] __initdata = { PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs }, { PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS, PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs_contd }, -#ifdef CONFIG_DMAR - { PCI_VENDOR_ID_INTEL, 0x29c0, - PCI_CLASS_BRIDGE_HOST, PCI_ANY_ID, 0, intel_g33_dmar }, -#endif {} }; diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 28b597ef9ca..43ceb3f454b 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1157,6 +1157,9 @@ ENTRY(mcount) END(mcount) ENTRY(ftrace_caller) + cmpl $0, function_trace_stop + jne ftrace_stub + pushl %eax pushl %ecx pushl %edx @@ -1171,6 +1174,11 @@ ftrace_call: popl %edx popl %ecx popl %eax +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +.globl ftrace_graph_call +ftrace_graph_call: + jmp ftrace_stub +#endif .globl ftrace_stub ftrace_stub: @@ -1180,8 +1188,18 @@ END(ftrace_caller) #else /* ! CONFIG_DYNAMIC_FTRACE */ ENTRY(mcount) + cmpl $0, function_trace_stop + jne ftrace_stub + cmpl $ftrace_stub, ftrace_trace_function jnz trace +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + cmpl $ftrace_stub, ftrace_graph_return + jnz ftrace_graph_caller + + cmpl $ftrace_graph_entry_stub, ftrace_graph_entry + jnz ftrace_graph_caller +#endif .globl ftrace_stub ftrace_stub: ret @@ -1200,12 +1218,43 @@ trace: popl %edx popl %ecx popl %eax - jmp ftrace_stub END(mcount) #endif /* CONFIG_DYNAMIC_FTRACE */ #endif /* CONFIG_FUNCTION_TRACER */ +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +ENTRY(ftrace_graph_caller) + cmpl $0, function_trace_stop + jne ftrace_stub + + pushl %eax + pushl %ecx + pushl %edx + movl 0xc(%esp), %edx + lea 0x4(%ebp), %eax + subl $MCOUNT_INSN_SIZE, %edx + call prepare_ftrace_return + popl %edx + popl %ecx + popl %eax + ret +END(ftrace_graph_caller) + +.globl return_to_handler +return_to_handler: + pushl $0 + pushl %eax + pushl %ecx + pushl %edx + call ftrace_return_to_handler + movl %eax, 0xc(%esp) + popl %edx + popl %ecx + popl %eax + ret +#endif + .section .rodata,"a" #include "syscall_table_32.S" diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index b86f332c96a..54e0bbdccb9 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -68,6 +68,8 @@ ENTRY(mcount) END(mcount) ENTRY(ftrace_caller) + cmpl $0, function_trace_stop + jne ftrace_stub /* taken from glibc */ subq $0x38, %rsp @@ -96,6 +98,12 @@ ftrace_call: movq (%rsp), %rax addq $0x38, %rsp +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +.globl ftrace_graph_call +ftrace_graph_call: + jmp ftrace_stub +#endif + .globl ftrace_stub ftrace_stub: retq @@ -103,8 +111,20 @@ END(ftrace_caller) #else /* ! CONFIG_DYNAMIC_FTRACE */ ENTRY(mcount) + cmpl $0, function_trace_stop + jne ftrace_stub + cmpq $ftrace_stub, ftrace_trace_function jnz trace + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + cmpq $ftrace_stub, ftrace_graph_return + jnz ftrace_graph_caller + + cmpq $ftrace_graph_entry_stub, ftrace_graph_entry + jnz ftrace_graph_caller +#endif + .globl ftrace_stub ftrace_stub: retq @@ -140,6 +160,69 @@ END(mcount) #endif /* CONFIG_DYNAMIC_FTRACE */ #endif /* CONFIG_FUNCTION_TRACER */ +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +ENTRY(ftrace_graph_caller) + cmpl $0, function_trace_stop + jne ftrace_stub + + subq $0x38, %rsp + movq %rax, (%rsp) + movq %rcx, 8(%rsp) + movq %rdx, 16(%rsp) + movq %rsi, 24(%rsp) + movq %rdi, 32(%rsp) + movq %r8, 40(%rsp) + movq %r9, 48(%rsp) + + leaq 8(%rbp), %rdi + movq 0x38(%rsp), %rsi + subq $MCOUNT_INSN_SIZE, %rsi + + call prepare_ftrace_return + + movq 48(%rsp), %r9 + movq 40(%rsp), %r8 + movq 32(%rsp), %rdi + movq 24(%rsp), %rsi + movq 16(%rsp), %rdx + movq 8(%rsp), %rcx + movq (%rsp), %rax + addq $0x38, %rsp + retq +END(ftrace_graph_caller) + + +.globl return_to_handler +return_to_handler: + subq $80, %rsp + + movq %rax, (%rsp) + movq %rcx, 8(%rsp) + movq %rdx, 16(%rsp) + movq %rsi, 24(%rsp) + movq %rdi, 32(%rsp) + movq %r8, 40(%rsp) + movq %r9, 48(%rsp) + movq %r10, 56(%rsp) + movq %r11, 64(%rsp) + + call ftrace_return_to_handler + + movq %rax, 72(%rsp) + movq 64(%rsp), %r11 + movq 56(%rsp), %r10 + movq 48(%rsp), %r9 + movq 40(%rsp), %r8 + movq 32(%rsp), %rdi + movq 24(%rsp), %rsi + movq 16(%rsp), %rdx + movq 8(%rsp), %rcx + movq (%rsp), %rax + addq $72, %rsp + retq +#endif + + #ifndef CONFIG_PREEMPT #define retint_kernel retint_restore_args #endif diff --git a/arch/x86/kernel/es7000_32.c b/arch/x86/kernel/es7000_32.c index f454c78fcef..53699c931ad 100644 --- a/arch/x86/kernel/es7000_32.c +++ b/arch/x86/kernel/es7000_32.c @@ -38,8 +38,11 @@ #include <asm/io.h> #include <asm/nmi.h> #include <asm/smp.h> +#include <asm/atomic.h> #include <asm/apicdef.h> #include <mach_mpparse.h> +#include <asm/genapic.h> +#include <asm/setup.h> /* * ES7000 chipsets @@ -161,6 +164,43 @@ es7000_rename_gsi(int ioapic, int gsi) return gsi; } +static int wakeup_secondary_cpu_via_mip(int cpu, unsigned long eip) +{ + unsigned long vect = 0, psaival = 0; + + if (psai == NULL) + return -1; + + vect = ((unsigned long)__pa(eip)/0x1000) << 16; + psaival = (0x1000000 | vect | cpu); + + while (*psai & 0x1000000) + ; + + *psai = psaival; + + return 0; +} + +static void noop_wait_for_deassert(atomic_t *deassert_not_used) +{ +} + +static int __init es7000_update_genapic(void) +{ + genapic->wakeup_cpu = wakeup_secondary_cpu_via_mip; + + /* MPENTIUMIII */ + if (boot_cpu_data.x86 == 6 && + (boot_cpu_data.x86_model >= 7 || boot_cpu_data.x86_model <= 11)) { + es7000_update_genapic_to_cluster(); + genapic->wait_for_init_deassert = noop_wait_for_deassert; + genapic->wakeup_cpu = wakeup_secondary_cpu_via_mip; + } + + return 0; +} + void __init setup_unisys(void) { @@ -176,6 +216,8 @@ setup_unisys(void) else es7000_plat = ES7000_CLASSIC; ioapic_renumber_irq = es7000_rename_gsi; + + x86_quirks->update_genapic = es7000_update_genapic; } /* @@ -250,31 +292,24 @@ 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 @@ -324,26 +359,6 @@ es7000_mip_write(struct mip_reg *mip_reg) return status; } -int -es7000_start_cpu(int cpu, unsigned long eip) -{ - unsigned long vect = 0, psaival = 0; - - if (psai == NULL) - return -1; - - vect = ((unsigned long)__pa(eip)/0x1000) << 16; - psaival = (0x1000000 | vect | cpu); - - while (*psai & 0x1000000) - ; - - *psai = psaival; - - return 0; - -} - void __init es7000_sw_apic(void) { diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 50ea0ac8c9b..1b43086b097 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -14,14 +14,17 @@ #include <linux/uaccess.h> #include <linux/ftrace.h> #include <linux/percpu.h> +#include <linux/sched.h> #include <linux/init.h> #include <linux/list.h> #include <asm/ftrace.h> +#include <linux/ftrace.h> #include <asm/nops.h> +#include <asm/nmi.h> -static unsigned char ftrace_nop[MCOUNT_INSN_SIZE]; +#ifdef CONFIG_DYNAMIC_FTRACE union ftrace_code_union { char code[MCOUNT_INSN_SIZE]; @@ -31,18 +34,12 @@ union ftrace_code_union { } __attribute__((packed)); }; - static int ftrace_calc_offset(long ip, long addr) { return (int)(addr - ip); } -unsigned char *ftrace_nop_replace(void) -{ - return ftrace_nop; -} - -unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) +static unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) { static union ftrace_code_union calc; @@ -56,7 +53,142 @@ unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) return calc.code; } -int +/* + * Modifying code must take extra care. On an SMP machine, if + * the code being modified is also being executed on another CPU + * that CPU will have undefined results and possibly take a GPF. + * We use kstop_machine to stop other CPUS from exectuing code. + * But this does not stop NMIs from happening. We still need + * to protect against that. We separate out the modification of + * the code to take care of this. + * + * Two buffers are added: An IP buffer and a "code" buffer. + * + * 1) Put the instruction pointer into the IP buffer + * and the new code into the "code" buffer. + * 2) Set a flag that says we are modifying code + * 3) Wait for any running NMIs to finish. + * 4) Write the code + * 5) clear the flag. + * 6) Wait for any running NMIs to finish. + * + * If an NMI is executed, the first thing it does is to call + * "ftrace_nmi_enter". This will check if the flag is set to write + * and if it is, it will write what is in the IP and "code" buffers. + * + * The trick is, it does not matter if everyone is writing the same + * content to the code location. Also, if a CPU is executing code + * it is OK to write to that code location if the contents being written + * are the same as what exists. + */ + +static atomic_t in_nmi = ATOMIC_INIT(0); +static int mod_code_status; /* holds return value of text write */ +static int mod_code_write; /* set when NMI should do the write */ +static void *mod_code_ip; /* holds the IP to write to */ +static void *mod_code_newcode; /* holds the text to write to the IP */ + +static unsigned nmi_wait_count; +static atomic_t nmi_update_count = ATOMIC_INIT(0); + +int ftrace_arch_read_dyn_info(char *buf, int size) +{ + int r; + + r = snprintf(buf, size, "%u %u", + nmi_wait_count, + atomic_read(&nmi_update_count)); + return r; +} + +static void ftrace_mod_code(void) +{ + /* + * Yes, more than one CPU process can be writing to mod_code_status. + * (and the code itself) + * But if one were to fail, then they all should, and if one were + * to succeed, then they all should. + */ + mod_code_status = probe_kernel_write(mod_code_ip, mod_code_newcode, + MCOUNT_INSN_SIZE); +} + +void ftrace_nmi_enter(void) +{ + atomic_inc(&in_nmi); + /* Must have in_nmi seen before reading write flag */ + smp_mb(); + if (mod_code_write) { + ftrace_mod_code(); + atomic_inc(&nmi_update_count); + } +} + +void ftrace_nmi_exit(void) +{ + /* Finish all executions before clearing in_nmi */ + smp_wmb(); + atomic_dec(&in_nmi); +} + +static void wait_for_nmi(void) +{ + int waited = 0; + + while (atomic_read(&in_nmi)) { + waited = 1; + cpu_relax(); + } + + if (waited) + nmi_wait_count++; +} + +static int +do_ftrace_mod_code(unsigned long ip, void *new_code) +{ + mod_code_ip = (void *)ip; + mod_code_newcode = new_code; + + /* The buffers need to be visible before we let NMIs write them */ + smp_wmb(); + + mod_code_write = 1; + + /* Make sure write bit is visible before we wait on NMIs */ + smp_mb(); + + wait_for_nmi(); + + /* Make sure all running NMIs have finished before we write the code */ + smp_mb(); + + ftrace_mod_code(); + + /* Make sure the write happens before clearing the bit */ + smp_wmb(); + + mod_code_write = 0; + + /* make sure NMIs see the cleared bit */ + smp_mb(); + + wait_for_nmi(); + + return mod_code_status; +} + + + + +static unsigned char ftrace_nop[MCOUNT_INSN_SIZE]; + +static unsigned char *ftrace_nop_replace(void) +{ + return ftrace_nop; +} + +static int ftrace_modify_code(unsigned long ip, unsigned char *old_code, unsigned char *new_code) { @@ -81,7 +213,7 @@ ftrace_modify_code(unsigned long ip, unsigned char *old_code, return -EINVAL; /* replace the text with the new text */ - if (probe_kernel_write((void *)ip, new_code, MCOUNT_INSN_SIZE)) + if (do_ftrace_mod_code(ip, new_code)) return -EPERM; sync_core(); @@ -89,6 +221,29 @@ ftrace_modify_code(unsigned long ip, unsigned char *old_code, return 0; } +int ftrace_make_nop(struct module *mod, + struct dyn_ftrace *rec, unsigned long addr) +{ + unsigned char *new, *old; + unsigned long ip = rec->ip; + + old = ftrace_call_replace(ip, addr); + new = ftrace_nop_replace(); + + return ftrace_modify_code(rec->ip, old, new); +} + +int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr) +{ + unsigned char *new, *old; + unsigned long ip = rec->ip; + + old = ftrace_nop_replace(); + new = ftrace_call_replace(ip, addr); + + return ftrace_modify_code(rec->ip, old, new); +} + int ftrace_update_ftrace_func(ftrace_func_t func) { unsigned long ip = (unsigned long)(&ftrace_call); @@ -165,3 +320,218 @@ int __init ftrace_dyn_arch_init(void *data) return 0; } +#endif + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + +#ifdef CONFIG_DYNAMIC_FTRACE +extern void ftrace_graph_call(void); + +static int ftrace_mod_jmp(unsigned long ip, + int old_offset, int new_offset) +{ + unsigned char code[MCOUNT_INSN_SIZE]; + + if (probe_kernel_read(code, (void *)ip, MCOUNT_INSN_SIZE)) + return -EFAULT; + + if (code[0] != 0xe9 || old_offset != *(int *)(&code[1])) + return -EINVAL; + + *(int *)(&code[1]) = new_offset; + + if (do_ftrace_mod_code(ip, &code)) + return -EPERM; + + return 0; +} + +int ftrace_enable_ftrace_graph_caller(void) +{ + unsigned long ip = (unsigned long)(&ftrace_graph_call); + int old_offset, new_offset; + + old_offset = (unsigned long)(&ftrace_stub) - (ip + MCOUNT_INSN_SIZE); + new_offset = (unsigned long)(&ftrace_graph_caller) - (ip + MCOUNT_INSN_SIZE); + + return ftrace_mod_jmp(ip, old_offset, new_offset); +} + +int ftrace_disable_ftrace_graph_caller(void) +{ + unsigned long ip = (unsigned long)(&ftrace_graph_call); + int old_offset, new_offset; + + old_offset = (unsigned long)(&ftrace_graph_caller) - (ip + MCOUNT_INSN_SIZE); + new_offset = (unsigned long)(&ftrace_stub) - (ip + MCOUNT_INSN_SIZE); + + return ftrace_mod_jmp(ip, old_offset, new_offset); +} + +#else /* CONFIG_DYNAMIC_FTRACE */ + +/* + * These functions are picked from those used on + * this page for dynamic ftrace. They have been + * simplified to ignore all traces in NMI context. + */ +static atomic_t in_nmi; + +void ftrace_nmi_enter(void) +{ + atomic_inc(&in_nmi); +} + +void ftrace_nmi_exit(void) +{ + atomic_dec(&in_nmi); +} + +#endif /* !CONFIG_DYNAMIC_FTRACE */ + +/* Add a function return address to the trace stack on thread info.*/ +static int push_return_trace(unsigned long ret, unsigned long long time, + unsigned long func, int *depth) +{ + int index; + + if (!current->ret_stack) + return -EBUSY; + + /* The return trace stack is full */ + if (current->curr_ret_stack == FTRACE_RETFUNC_DEPTH - 1) { + atomic_inc(¤t->trace_overrun); + return -EBUSY; + } + + index = ++current->curr_ret_stack; + barrier(); + current->ret_stack[index].ret = ret; + current->ret_stack[index].func = func; + current->ret_stack[index].calltime = time; + *depth = index; + + return 0; +} + +/* Retrieve a function return address to the trace stack on thread info.*/ +static void pop_return_trace(struct ftrace_graph_ret *trace, unsigned long *ret) +{ + int index; + + index = current->curr_ret_stack; + + if (unlikely(index < 0)) { + ftrace_graph_stop(); + WARN_ON(1); + /* Might as well panic, otherwise we have no where to go */ + *ret = (unsigned long)panic; + return; + } + + *ret = current->ret_stack[index].ret; + trace->func = current->ret_stack[index].func; + trace->calltime = current->ret_stack[index].calltime; + trace->overrun = atomic_read(¤t->trace_overrun); + trace->depth = index; + barrier(); + current->curr_ret_stack--; + +} + +/* + * Send the trace to the ring-buffer. + * @return the original return address. + */ +unsigned long ftrace_return_to_handler(void) +{ + struct ftrace_graph_ret trace; + unsigned long ret; + + pop_return_trace(&trace, &ret); + trace.rettime = cpu_clock(raw_smp_processor_id()); + ftrace_graph_return(&trace); + + if (unlikely(!ret)) { + ftrace_graph_stop(); + WARN_ON(1); + /* Might as well panic. What else to do? */ + ret = (unsigned long)panic; + } + + return ret; +} + +/* + * Hook the return address and push it in the stack of return addrs + * in current thread info. + */ +void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) +{ + unsigned long old; + unsigned long long calltime; + int faulted; + struct ftrace_graph_ent trace; + unsigned long return_hooker = (unsigned long) + &return_to_handler; + + /* Nmi's are currently unsupported */ + if (unlikely(atomic_read(&in_nmi))) + return; + + if (unlikely(atomic_read(¤t->tracing_graph_pause))) + return; + + /* + * Protect against fault, even if it shouldn't + * happen. This tool is too much intrusive to + * ignore such a protection. + */ + asm volatile( + "1: " _ASM_MOV " (%[parent_old]), %[old]\n" + "2: " _ASM_MOV " %[return_hooker], (%[parent_replaced])\n" + " movl $0, %[faulted]\n" + + ".section .fixup, \"ax\"\n" + "3: movl $1, %[faulted]\n" + ".previous\n" + + _ASM_EXTABLE(1b, 3b) + _ASM_EXTABLE(2b, 3b) + + : [parent_replaced] "=r" (parent), [old] "=r" (old), + [faulted] "=r" (faulted) + : [parent_old] "0" (parent), [return_hooker] "r" (return_hooker) + : "memory" + ); + + if (unlikely(faulted)) { + ftrace_graph_stop(); + WARN_ON(1); + return; + } + + if (unlikely(!__kernel_text_address(old))) { + ftrace_graph_stop(); + *parent = old; + WARN_ON(1); + return; + } + + calltime = cpu_clock(raw_smp_processor_id()); + + if (push_return_trace(old, calltime, + self_addr, &trace.depth) == -EBUSY) { + *parent = old; + return; + } + + trace.func = self_addr; + + /* Only trace if the calling function expects to */ + if (!ftrace_graph_entry(&trace)) { + current->curr_ret_stack--; + *parent = old; + } +} +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ diff --git a/arch/x86/kernel/genapic_64.c b/arch/x86/kernel/genapic_64.c index 6c9bfc9e1e9..2bced78b0b8 100644 --- a/arch/x86/kernel/genapic_64.c +++ b/arch/x86/kernel/genapic_64.c @@ -21,6 +21,7 @@ #include <asm/smp.h> #include <asm/ipi.h> #include <asm/genapic.h> +#include <asm/setup.h> extern struct genapic apic_flat; extern struct genapic apic_physflat; @@ -53,6 +54,9 @@ void __init setup_apic_routing(void) genapic = &apic_physflat; printk(KERN_INFO "Setting APIC routing to %s\n", genapic->name); } + + if (x86_quirks->update_genapic) + x86_quirks->update_genapic(); } /* Same for both flat and physical. */ diff --git a/arch/x86/kernel/genapic_flat_64.c b/arch/x86/kernel/genapic_flat_64.c index c0262791bda..7fa5f49c2dd 100644 --- a/arch/x86/kernel/genapic_flat_64.c +++ b/arch/x86/kernel/genapic_flat_64.c @@ -30,12 +30,12 @@ static int flat_acpi_madt_oem_check(char *oem_id, char *oem_table_id) return 1; } -static cpumask_t flat_target_cpus(void) +static const struct cpumask *flat_target_cpus(void) { - return cpu_online_map; + return cpu_online_mask; } -static cpumask_t flat_vector_allocation_domain(int cpu) +static void flat_vector_allocation_domain(int cpu, struct cpumask *retmask) { /* Careful. Some cpus do not strictly honor the set of cpus * specified in the interrupt destination when using lowest @@ -45,8 +45,8 @@ static cpumask_t flat_vector_allocation_domain(int cpu) * deliver interrupts to the wrong hyperthread when only one * hyperthread was specified in the interrupt desitination. */ - cpumask_t domain = { { [0] = APIC_ALL_CPUS, } }; - return domain; + cpumask_clear(retmask); + cpumask_bits(retmask)[0] = APIC_ALL_CPUS; } /* @@ -69,9 +69,8 @@ static void flat_init_apic_ldr(void) apic_write(APIC_LDR, val); } -static void flat_send_IPI_mask(cpumask_t cpumask, int vector) +static inline void _flat_send_IPI_mask(unsigned long mask, int vector) { - unsigned long mask = cpus_addr(cpumask)[0]; unsigned long flags; local_irq_save(flags); @@ -79,20 +78,41 @@ static void flat_send_IPI_mask(cpumask_t cpumask, int vector) local_irq_restore(flags); } +static void flat_send_IPI_mask(const struct cpumask *cpumask, int vector) +{ + unsigned long mask = cpumask_bits(cpumask)[0]; + + _flat_send_IPI_mask(mask, vector); +} + +static void flat_send_IPI_mask_allbutself(const struct cpumask *cpumask, + int vector) +{ + unsigned long mask = cpumask_bits(cpumask)[0]; + int cpu = smp_processor_id(); + + if (cpu < BITS_PER_LONG) + clear_bit(cpu, &mask); + _flat_send_IPI_mask(mask, vector); +} + static void flat_send_IPI_allbutself(int vector) { + int cpu = smp_processor_id(); #ifdef CONFIG_HOTPLUG_CPU int hotplug = 1; #else int hotplug = 0; #endif if (hotplug || vector == NMI_VECTOR) { - cpumask_t allbutme = cpu_online_map; + if (!cpumask_equal(cpu_online_mask, cpumask_of(cpu))) { + unsigned long mask = cpumask_bits(cpu_online_mask)[0]; - cpu_clear(smp_processor_id(), allbutme); + if (cpu < BITS_PER_LONG) + clear_bit(cpu, &mask); - if (!cpus_empty(allbutme)) - flat_send_IPI_mask(allbutme, vector); + _flat_send_IPI_mask(mask, vector); + } } else if (num_online_cpus() > 1) { __send_IPI_shortcut(APIC_DEST_ALLBUT, vector,APIC_DEST_LOGICAL); } @@ -101,7 +121,7 @@ static void flat_send_IPI_allbutself(int vector) static void flat_send_IPI_all(int vector) { if (vector == NMI_VECTOR) - flat_send_IPI_mask(cpu_online_map, vector); + flat_send_IPI_mask(cpu_online_mask, vector); else __send_IPI_shortcut(APIC_DEST_ALLINC, vector, APIC_DEST_LOGICAL); } @@ -135,9 +155,18 @@ static int flat_apic_id_registered(void) return physid_isset(read_xapic_id(), phys_cpu_present_map); } -static unsigned int flat_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int flat_cpu_mask_to_apicid(const struct cpumask *cpumask) +{ + return cpumask_bits(cpumask)[0] & APIC_ALL_CPUS; +} + +static unsigned int flat_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { - return cpus_addr(cpumask)[0] & APIC_ALL_CPUS; + unsigned long mask1 = cpumask_bits(cpumask)[0] & APIC_ALL_CPUS; + unsigned long mask2 = cpumask_bits(andmask)[0] & APIC_ALL_CPUS; + + return mask1 & mask2; } static unsigned int phys_pkg_id(int index_msb) @@ -157,8 +186,10 @@ struct genapic apic_flat = { .send_IPI_all = flat_send_IPI_all, .send_IPI_allbutself = flat_send_IPI_allbutself, .send_IPI_mask = flat_send_IPI_mask, + .send_IPI_mask_allbutself = flat_send_IPI_mask_allbutself, .send_IPI_self = apic_send_IPI_self, .cpu_mask_to_apicid = flat_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = flat_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, @@ -188,35 +219,39 @@ static int physflat_acpi_madt_oem_check(char *oem_id, char *oem_table_id) return 0; } -static cpumask_t physflat_target_cpus(void) +static const struct cpumask *physflat_target_cpus(void) { - return cpu_online_map; + return cpu_online_mask; } -static cpumask_t physflat_vector_allocation_domain(int cpu) +static void physflat_vector_allocation_domain(int cpu, struct cpumask *retmask) { - return cpumask_of_cpu(cpu); + cpumask_clear(retmask); + cpumask_set_cpu(cpu, retmask); } -static void physflat_send_IPI_mask(cpumask_t cpumask, int vector) +static void physflat_send_IPI_mask(const struct cpumask *cpumask, int vector) { send_IPI_mask_sequence(cpumask, vector); } -static void physflat_send_IPI_allbutself(int vector) +static void physflat_send_IPI_mask_allbutself(const struct cpumask *cpumask, + int vector) { - cpumask_t allbutme = cpu_online_map; + send_IPI_mask_allbutself(cpumask, vector); +} - cpu_clear(smp_processor_id(), allbutme); - physflat_send_IPI_mask(allbutme, vector); +static void physflat_send_IPI_allbutself(int vector) +{ + send_IPI_mask_allbutself(cpu_online_mask, vector); } static void physflat_send_IPI_all(int vector) { - physflat_send_IPI_mask(cpu_online_map, vector); + physflat_send_IPI_mask(cpu_online_mask, vector); } -static unsigned int physflat_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int physflat_cpu_mask_to_apicid(const struct cpumask *cpumask) { int cpu; @@ -224,13 +259,29 @@ static unsigned int physflat_cpu_mask_to_apicid(cpumask_t cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(cpumask); + cpu = cpumask_first(cpumask); if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); else return BAD_APICID; } +static unsigned int +physflat_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) +{ + int cpu; + + /* + * We're using fixed IRQ delivery, can only return one phys APIC ID. + * May as well be the first. + */ + cpu = cpumask_any_and(cpumask, andmask); + if (cpu < nr_cpu_ids) + return per_cpu(x86_cpu_to_apicid, cpu); + return BAD_APICID; +} + struct genapic apic_physflat = { .name = "physical flat", .acpi_madt_oem_check = physflat_acpi_madt_oem_check, @@ -243,8 +294,10 @@ struct genapic apic_physflat = { .send_IPI_all = physflat_send_IPI_all, .send_IPI_allbutself = physflat_send_IPI_allbutself, .send_IPI_mask = physflat_send_IPI_mask, + .send_IPI_mask_allbutself = physflat_send_IPI_mask_allbutself, .send_IPI_self = apic_send_IPI_self, .cpu_mask_to_apicid = physflat_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = physflat_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, diff --git a/arch/x86/kernel/genx2apic_cluster.c b/arch/x86/kernel/genx2apic_cluster.c index f6a2c8eb48a..4716a0c9f93 100644 --- a/arch/x86/kernel/genx2apic_cluster.c +++ b/arch/x86/kernel/genx2apic_cluster.c @@ -22,19 +22,18 @@ static int x2apic_acpi_madt_oem_check(char *oem_id, char *oem_table_id) /* Start with all IRQs pointing to boot CPU. IRQ balancing will shift them. */ -static cpumask_t x2apic_target_cpus(void) +static const struct cpumask *x2apic_target_cpus(void) { - return cpumask_of_cpu(0); + return cpumask_of(0); } /* * for now each logical cpu is in its own vector allocation domain. */ -static cpumask_t x2apic_vector_allocation_domain(int cpu) +static void x2apic_vector_allocation_domain(int cpu, struct cpumask *retmask) { - cpumask_t domain = CPU_MASK_NONE; - cpu_set(cpu, domain); - return domain; + cpumask_clear(retmask); + cpumask_set_cpu(cpu, retmask); } static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, @@ -56,32 +55,53 @@ static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, * at once. We have 16 cpu's in a cluster. This will minimize IPI register * writes. */ -static void x2apic_send_IPI_mask(cpumask_t mask, int vector) +static void x2apic_send_IPI_mask(const struct cpumask *mask, int vector) { unsigned long flags; unsigned long query_cpu; local_irq_save(flags); - for_each_cpu_mask(query_cpu, mask) { - __x2apic_send_IPI_dest(per_cpu(x86_cpu_to_logical_apicid, query_cpu), - vector, APIC_DEST_LOGICAL); - } + for_each_cpu(query_cpu, mask) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_logical_apicid, query_cpu), + vector, APIC_DEST_LOGICAL); local_irq_restore(flags); } -static void x2apic_send_IPI_allbutself(int vector) +static void x2apic_send_IPI_mask_allbutself(const struct cpumask *mask, + int vector) { - cpumask_t mask = cpu_online_map; + unsigned long flags; + unsigned long query_cpu; + unsigned long this_cpu = smp_processor_id(); - cpu_clear(smp_processor_id(), mask); + local_irq_save(flags); + for_each_cpu(query_cpu, mask) + if (query_cpu != this_cpu) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_logical_apicid, query_cpu), + vector, APIC_DEST_LOGICAL); + local_irq_restore(flags); +} + +static void x2apic_send_IPI_allbutself(int vector) +{ + unsigned long flags; + unsigned long query_cpu; + unsigned long this_cpu = smp_processor_id(); - if (!cpus_empty(mask)) - x2apic_send_IPI_mask(mask, vector); + local_irq_save(flags); + for_each_online_cpu(query_cpu) + if (query_cpu != this_cpu) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_logical_apicid, query_cpu), + vector, APIC_DEST_LOGICAL); + local_irq_restore(flags); } static void x2apic_send_IPI_all(int vector) { - x2apic_send_IPI_mask(cpu_online_map, vector); + x2apic_send_IPI_mask(cpu_online_mask, vector); } static int x2apic_apic_id_registered(void) @@ -89,7 +109,7 @@ static int x2apic_apic_id_registered(void) return 1; } -static unsigned int x2apic_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int x2apic_cpu_mask_to_apicid(const struct cpumask *cpumask) { int cpu; @@ -97,13 +117,28 @@ static unsigned int x2apic_cpu_mask_to_apicid(cpumask_t cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(cpumask); - if ((unsigned)cpu < NR_CPUS) + cpu = cpumask_first(cpumask); + if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_logical_apicid, cpu); else return BAD_APICID; } +static unsigned int x2apic_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) +{ + int cpu; + + /* + * We're using fixed IRQ delivery, can only return one phys APIC ID. + * May as well be the first. + */ + cpu = cpumask_any_and(cpumask, andmask); + if (cpu < nr_cpu_ids) + return per_cpu(x86_cpu_to_apicid, cpu); + return BAD_APICID; +} + static unsigned int get_apic_id(unsigned long x) { unsigned int id; @@ -150,8 +185,10 @@ struct genapic apic_x2apic_cluster = { .send_IPI_all = x2apic_send_IPI_all, .send_IPI_allbutself = x2apic_send_IPI_allbutself, .send_IPI_mask = x2apic_send_IPI_mask, + .send_IPI_mask_allbutself = x2apic_send_IPI_mask_allbutself, .send_IPI_self = x2apic_send_IPI_self, .cpu_mask_to_apicid = x2apic_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = x2apic_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, diff --git a/arch/x86/kernel/genx2apic_phys.c b/arch/x86/kernel/genx2apic_phys.c index d042211768b..b255507884f 100644 --- a/arch/x86/kernel/genx2apic_phys.c +++ b/arch/x86/kernel/genx2apic_phys.c @@ -29,16 +29,15 @@ static int x2apic_acpi_madt_oem_check(char *oem_id, char *oem_table_id) /* Start with all IRQs pointing to boot CPU. IRQ balancing will shift them. */ -static cpumask_t x2apic_target_cpus(void) +static const struct cpumask *x2apic_target_cpus(void) { - return cpumask_of_cpu(0); + return cpumask_of(0); } -static cpumask_t x2apic_vector_allocation_domain(int cpu) +static void x2apic_vector_allocation_domain(int cpu, struct cpumask *retmask) { - cpumask_t domain = CPU_MASK_NONE; - cpu_set(cpu, domain); - return domain; + cpumask_clear(retmask); + cpumask_set_cpu(cpu, retmask); } static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, @@ -54,32 +53,54 @@ static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, x2apic_icr_write(cfg, apicid); } -static void x2apic_send_IPI_mask(cpumask_t mask, int vector) +static void x2apic_send_IPI_mask(const struct cpumask *mask, int vector) { unsigned long flags; unsigned long query_cpu; local_irq_save(flags); - for_each_cpu_mask(query_cpu, mask) { + for_each_cpu(query_cpu, mask) { __x2apic_send_IPI_dest(per_cpu(x86_cpu_to_apicid, query_cpu), vector, APIC_DEST_PHYSICAL); } local_irq_restore(flags); } -static void x2apic_send_IPI_allbutself(int vector) +static void x2apic_send_IPI_mask_allbutself(const struct cpumask *mask, + int vector) { - cpumask_t mask = cpu_online_map; + unsigned long flags; + unsigned long query_cpu; + unsigned long this_cpu = smp_processor_id(); + + local_irq_save(flags); + for_each_cpu(query_cpu, mask) { + if (query_cpu != this_cpu) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_apicid, query_cpu), + vector, APIC_DEST_PHYSICAL); + } + local_irq_restore(flags); +} - cpu_clear(smp_processor_id(), mask); +static void x2apic_send_IPI_allbutself(int vector) +{ + unsigned long flags; + unsigned long query_cpu; + unsigned long this_cpu = smp_processor_id(); - if (!cpus_empty(mask)) - x2apic_send_IPI_mask(mask, vector); + local_irq_save(flags); + for_each_online_cpu(query_cpu) + if (query_cpu != this_cpu) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_apicid, query_cpu), + vector, APIC_DEST_PHYSICAL); + local_irq_restore(flags); } static void x2apic_send_IPI_all(int vector) { - x2apic_send_IPI_mask(cpu_online_map, vector); + x2apic_send_IPI_mask(cpu_online_mask, vector); } static int x2apic_apic_id_registered(void) @@ -87,7 +108,7 @@ static int x2apic_apic_id_registered(void) return 1; } -static unsigned int x2apic_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int x2apic_cpu_mask_to_apicid(const struct cpumask *cpumask) { int cpu; @@ -95,13 +116,28 @@ static unsigned int x2apic_cpu_mask_to_apicid(cpumask_t cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(cpumask); - if ((unsigned)cpu < NR_CPUS) + cpu = cpumask_first(cpumask); + if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); else return BAD_APICID; } +static unsigned int x2apic_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) +{ + int cpu; + + /* + * We're using fixed IRQ delivery, can only return one phys APIC ID. + * May as well be the first. + */ + cpu = cpumask_any_and(cpumask, andmask); + if (cpu < nr_cpu_ids) + return per_cpu(x86_cpu_to_apicid, cpu); + return BAD_APICID; +} + static unsigned int get_apic_id(unsigned long x) { unsigned int id; @@ -145,8 +181,10 @@ struct genapic apic_x2apic_phys = { .send_IPI_all = x2apic_send_IPI_all, .send_IPI_allbutself = x2apic_send_IPI_allbutself, .send_IPI_mask = x2apic_send_IPI_mask, + .send_IPI_mask_allbutself = x2apic_send_IPI_mask_allbutself, .send_IPI_self = x2apic_send_IPI_self, .cpu_mask_to_apicid = x2apic_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = x2apic_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index 2c7dbdb9827..3984682cd84 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -75,16 +75,15 @@ EXPORT_SYMBOL(sn_rtc_cycles_per_second); /* Start with all IRQs pointing to boot CPU. IRQ balancing will shift them. */ -static cpumask_t uv_target_cpus(void) +static const struct cpumask *uv_target_cpus(void) { - return cpumask_of_cpu(0); + return cpumask_of(0); } -static cpumask_t uv_vector_allocation_domain(int cpu) +static void uv_vector_allocation_domain(int cpu, struct cpumask *retmask) { - cpumask_t domain = CPU_MASK_NONE; - cpu_set(cpu, domain); - return domain; + cpumask_clear(retmask); + cpumask_set_cpu(cpu, retmask); } int uv_wakeup_secondary(int phys_apicid, unsigned int start_rip) @@ -123,28 +122,37 @@ static void uv_send_IPI_one(int cpu, int vector) uv_write_global_mmr64(pnode, UVH_IPI_INT, val); } -static void uv_send_IPI_mask(cpumask_t mask, int vector) +static void uv_send_IPI_mask(const struct cpumask *mask, int vector) { unsigned int cpu; - for_each_possible_cpu(cpu) - if (cpu_isset(cpu, mask)) + for_each_cpu(cpu, mask) + uv_send_IPI_one(cpu, vector); +} + +static void uv_send_IPI_mask_allbutself(const struct cpumask *mask, int vector) +{ + unsigned int cpu; + unsigned int this_cpu = smp_processor_id(); + + for_each_cpu(cpu, mask) + if (cpu != this_cpu) uv_send_IPI_one(cpu, vector); } static void uv_send_IPI_allbutself(int vector) { - cpumask_t mask = cpu_online_map; - - cpu_clear(smp_processor_id(), mask); + unsigned int cpu; + unsigned int this_cpu = smp_processor_id(); - if (!cpus_empty(mask)) - uv_send_IPI_mask(mask, vector); + for_each_online_cpu(cpu) + if (cpu != this_cpu) + uv_send_IPI_one(cpu, vector); } static void uv_send_IPI_all(int vector) { - uv_send_IPI_mask(cpu_online_map, vector); + uv_send_IPI_mask(cpu_online_mask, vector); } static int uv_apic_id_registered(void) @@ -156,7 +164,7 @@ static void uv_init_apic_ldr(void) { } -static unsigned int uv_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int uv_cpu_mask_to_apicid(const struct cpumask *cpumask) { int cpu; @@ -164,13 +172,28 @@ static unsigned int uv_cpu_mask_to_apicid(cpumask_t cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(cpumask); + cpu = cpumask_first(cpumask); if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); else return BAD_APICID; } +static unsigned int uv_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) +{ + int cpu; + + /* + * We're using fixed IRQ delivery, can only return one phys APIC ID. + * May as well be the first. + */ + cpu = cpumask_any_and(cpumask, andmask); + if (cpu < nr_cpu_ids) + return per_cpu(x86_cpu_to_apicid, cpu); + return BAD_APICID; +} + static unsigned int get_apic_id(unsigned long x) { unsigned int id; @@ -218,8 +241,10 @@ struct genapic apic_x2apic_uv_x = { .send_IPI_all = uv_send_IPI_all, .send_IPI_allbutself = uv_send_IPI_allbutself, .send_IPI_mask = uv_send_IPI_mask, + .send_IPI_mask_allbutself = uv_send_IPI_mask_allbutself, .send_IPI_self = uv_send_IPI_self, .cpu_mask_to_apicid = uv_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = uv_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index 77017e834cf..e76d7e27297 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -246,7 +246,7 @@ static void hpet_legacy_clockevent_register(void) * Start hpet with the boot cpu mask and make it * global after the IO_APIC has been initialized. */ - hpet_clockevent.cpumask = cpumask_of_cpu(smp_processor_id()); + hpet_clockevent.cpumask = cpumask_of(smp_processor_id()); clockevents_register_device(&hpet_clockevent); global_clock_event = &hpet_clockevent; printk(KERN_DEBUG "hpet clockevent registered\n"); @@ -301,7 +301,7 @@ static void hpet_set_mode(enum clock_event_mode mode, struct hpet_dev *hdev = EVT_TO_HPET_DEV(evt); hpet_setup_msi_irq(hdev->irq); disable_irq(hdev->irq); - irq_set_affinity(hdev->irq, cpumask_of_cpu(hdev->cpu)); + irq_set_affinity(hdev->irq, cpumask_of(hdev->cpu)); enable_irq(hdev->irq); } break; @@ -322,7 +322,7 @@ static int hpet_next_event(unsigned long delta, * what we wrote hit the chip before we compare it to the * counter. */ - WARN_ON((u32)hpet_readl(HPET_T0_CMP) != cnt); + WARN_ON_ONCE((u32)hpet_readl(HPET_Tn_CMP(timer)) != cnt); return (s32)((u32)hpet_readl(HPET_COUNTER) - cnt) >= 0 ? -ETIME : 0; } @@ -445,11 +445,11 @@ static int hpet_setup_irq(struct hpet_dev *dev) { if (request_irq(dev->irq, hpet_interrupt_handler, - IRQF_SHARED|IRQF_NOBALANCING, dev->name, dev)) + IRQF_DISABLED|IRQF_NOBALANCING, dev->name, dev)) return -1; disable_irq(dev->irq); - irq_set_affinity(dev->irq, cpumask_of_cpu(dev->cpu)); + irq_set_affinity(dev->irq, cpumask_of(dev->cpu)); enable_irq(dev->irq); printk(KERN_DEBUG "hpet: %s irq %d for MSI\n", @@ -500,7 +500,7 @@ static void init_one_hpet_msi_clockevent(struct hpet_dev *hdev, int cpu) /* 5 usec minimum reprogramming delta. */ evt->min_delta_ns = 5000; - evt->cpumask = cpumask_of_cpu(hdev->cpu); + evt->cpumask = cpumask_of(hdev->cpu); clockevents_register_device(evt); } 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); diff --git a/arch/x86/kernel/i8253.c b/arch/x86/kernel/i8253.c index c1b5e3ece1f..10f92fb532f 100644 --- a/arch/x86/kernel/i8253.c +++ b/arch/x86/kernel/i8253.c @@ -114,7 +114,7 @@ void __init setup_pit_timer(void) * Start pit with the boot cpu mask and make it global after the * IO_APIC has been initialized. */ - pit_clockevent.cpumask = cpumask_of_cpu(smp_processor_id()); + pit_clockevent.cpumask = cpumask_of(smp_processor_id()); pit_clockevent.mult = div_sc(CLOCK_TICK_RATE, NSEC_PER_SEC, pit_clockevent.shift); pit_clockevent.max_delta_ns = diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 7a3f2028e2e..58938cc4b7d 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -108,94 +108,273 @@ static int __init parse_noapic(char *str) early_param("noapic", parse_noapic); struct irq_pin_list; + +/* + * This is performance-critical, we want to do it O(1) + * + * the indexing order of this array favors 1:1 mappings + * between pins and IRQs. + */ + +struct irq_pin_list { + int apic, pin; + struct irq_pin_list *next; +}; + +static struct irq_pin_list *get_one_free_irq_2_pin(int cpu) +{ + struct irq_pin_list *pin; + int node; + + node = cpu_to_node(cpu); + + pin = kzalloc_node(sizeof(*pin), GFP_ATOMIC, node); + printk(KERN_DEBUG " alloc irq_2_pin on cpu %d node %d\n", cpu, node); + + return pin; +} + struct irq_cfg { - unsigned int irq; struct irq_pin_list *irq_2_pin; - cpumask_t domain; - cpumask_t old_domain; + cpumask_var_t domain; + cpumask_var_t old_domain; unsigned move_cleanup_count; u8 vector; u8 move_in_progress : 1; +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC + u8 move_desc_pending : 1; +#endif }; /* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */ +#ifdef CONFIG_SPARSE_IRQ +static struct irq_cfg irq_cfgx[] = { +#else static struct irq_cfg irq_cfgx[NR_IRQS] = { - [0] = { .irq = 0, .domain = CPU_MASK_ALL, .vector = IRQ0_VECTOR, }, - [1] = { .irq = 1, .domain = CPU_MASK_ALL, .vector = IRQ1_VECTOR, }, - [2] = { .irq = 2, .domain = CPU_MASK_ALL, .vector = IRQ2_VECTOR, }, - [3] = { .irq = 3, .domain = CPU_MASK_ALL, .vector = IRQ3_VECTOR, }, - [4] = { .irq = 4, .domain = CPU_MASK_ALL, .vector = IRQ4_VECTOR, }, - [5] = { .irq = 5, .domain = CPU_MASK_ALL, .vector = IRQ5_VECTOR, }, - [6] = { .irq = 6, .domain = CPU_MASK_ALL, .vector = IRQ6_VECTOR, }, - [7] = { .irq = 7, .domain = CPU_MASK_ALL, .vector = IRQ7_VECTOR, }, - [8] = { .irq = 8, .domain = CPU_MASK_ALL, .vector = IRQ8_VECTOR, }, - [9] = { .irq = 9, .domain = CPU_MASK_ALL, .vector = IRQ9_VECTOR, }, - [10] = { .irq = 10, .domain = CPU_MASK_ALL, .vector = IRQ10_VECTOR, }, - [11] = { .irq = 11, .domain = CPU_MASK_ALL, .vector = IRQ11_VECTOR, }, - [12] = { .irq = 12, .domain = CPU_MASK_ALL, .vector = IRQ12_VECTOR, }, - [13] = { .irq = 13, .domain = CPU_MASK_ALL, .vector = IRQ13_VECTOR, }, - [14] = { .irq = 14, .domain = CPU_MASK_ALL, .vector = IRQ14_VECTOR, }, - [15] = { .irq = 15, .domain = CPU_MASK_ALL, .vector = IRQ15_VECTOR, }, +#endif + [0] = { .vector = IRQ0_VECTOR, }, + [1] = { .vector = IRQ1_VECTOR, }, + [2] = { .vector = IRQ2_VECTOR, }, + [3] = { .vector = IRQ3_VECTOR, }, + [4] = { .vector = IRQ4_VECTOR, }, + [5] = { .vector = IRQ5_VECTOR, }, + [6] = { .vector = IRQ6_VECTOR, }, + [7] = { .vector = IRQ7_VECTOR, }, + [8] = { .vector = IRQ8_VECTOR, }, + [9] = { .vector = IRQ9_VECTOR, }, + [10] = { .vector = IRQ10_VECTOR, }, + [11] = { .vector = IRQ11_VECTOR, }, + [12] = { .vector = IRQ12_VECTOR, }, + [13] = { .vector = IRQ13_VECTOR, }, + [14] = { .vector = IRQ14_VECTOR, }, + [15] = { .vector = IRQ15_VECTOR, }, }; -#define for_each_irq_cfg(irq, cfg) \ - for (irq = 0, cfg = irq_cfgx; irq < nr_irqs; irq++, cfg++) +void __init arch_early_irq_init(void) +{ + struct irq_cfg *cfg; + struct irq_desc *desc; + int count; + int i; + cfg = irq_cfgx; + count = ARRAY_SIZE(irq_cfgx); + + for (i = 0; i < count; i++) { + desc = irq_to_desc(i); + desc->chip_data = &cfg[i]; + alloc_bootmem_cpumask_var(&cfg[i].domain); + alloc_bootmem_cpumask_var(&cfg[i].old_domain); + if (i < NR_IRQS_LEGACY) + cpumask_setall(cfg[i].domain); + } +} + +#ifdef CONFIG_SPARSE_IRQ static struct irq_cfg *irq_cfg(unsigned int irq) { - return irq < nr_irqs ? irq_cfgx + irq : NULL; + struct irq_cfg *cfg = NULL; + struct irq_desc *desc; + + desc = irq_to_desc(irq); + if (desc) + cfg = desc->chip_data; + + return cfg; } -static struct irq_cfg *irq_cfg_alloc(unsigned int irq) +static struct irq_cfg *get_one_free_irq_cfg(int cpu) { - return irq_cfg(irq); + struct irq_cfg *cfg; + int node; + + node = cpu_to_node(cpu); + + cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node); + if (cfg) { + /* FIXME: needs alloc_cpumask_var_node() */ + if (!alloc_cpumask_var(&cfg->domain, GFP_ATOMIC)) { + kfree(cfg); + cfg = NULL; + } else if (!alloc_cpumask_var(&cfg->old_domain, GFP_ATOMIC)) { + free_cpumask_var(cfg->domain); + kfree(cfg); + cfg = NULL; + } else { + cpumask_clear(cfg->domain); + cpumask_clear(cfg->old_domain); + } + } + printk(KERN_DEBUG " alloc irq_cfg on cpu %d node %d\n", cpu, node); + + return cfg; } -/* - * Rough estimation of how many shared IRQs there are, can be changed - * anytime. - */ -#define MAX_PLUS_SHARED_IRQS NR_IRQS -#define PIN_MAP_SIZE (MAX_PLUS_SHARED_IRQS + NR_IRQS) +void arch_init_chip_data(struct irq_desc *desc, int cpu) +{ + struct irq_cfg *cfg; -/* - * This is performance-critical, we want to do it O(1) - * - * the indexing order of this array favors 1:1 mappings - * between pins and IRQs. - */ + cfg = desc->chip_data; + if (!cfg) { + desc->chip_data = get_one_free_irq_cfg(cpu); + if (!desc->chip_data) { + printk(KERN_ERR "can not alloc irq_cfg\n"); + BUG_ON(1); + } + } +} -struct irq_pin_list { - int apic, pin; - struct irq_pin_list *next; -}; +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC -static struct irq_pin_list irq_2_pin_head[PIN_MAP_SIZE]; -static struct irq_pin_list *irq_2_pin_ptr; +static void +init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu) +{ + struct irq_pin_list *old_entry, *head, *tail, *entry; -static void __init irq_2_pin_init(void) + cfg->irq_2_pin = NULL; + old_entry = old_cfg->irq_2_pin; + if (!old_entry) + return; + + entry = get_one_free_irq_2_pin(cpu); + if (!entry) + return; + + entry->apic = old_entry->apic; + entry->pin = old_entry->pin; + head = entry; + tail = entry; + old_entry = old_entry->next; + while (old_entry) { + entry = get_one_free_irq_2_pin(cpu); + if (!entry) { + entry = head; + while (entry) { + head = entry->next; + kfree(entry); + entry = head; + } + /* still use the old one */ + return; + } + entry->apic = old_entry->apic; + entry->pin = old_entry->pin; + tail->next = entry; + tail = entry; + old_entry = old_entry->next; + } + + tail->next = NULL; + cfg->irq_2_pin = head; +} + +static void free_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg) { - struct irq_pin_list *pin = irq_2_pin_head; - int i; + struct irq_pin_list *entry, *next; + + if (old_cfg->irq_2_pin == cfg->irq_2_pin) + return; - for (i = 1; i < PIN_MAP_SIZE; i++) - pin[i-1].next = &pin[i]; + entry = old_cfg->irq_2_pin; - irq_2_pin_ptr = &pin[0]; + while (entry) { + next = entry->next; + kfree(entry); + entry = next; + } + old_cfg->irq_2_pin = NULL; } -static struct irq_pin_list *get_one_free_irq_2_pin(void) +void arch_init_copy_chip_data(struct irq_desc *old_desc, + struct irq_desc *desc, int cpu) { - struct irq_pin_list *pin = irq_2_pin_ptr; + struct irq_cfg *cfg; + struct irq_cfg *old_cfg; - if (!pin) - panic("can not get more irq_2_pin\n"); + cfg = get_one_free_irq_cfg(cpu); - irq_2_pin_ptr = pin->next; - pin->next = NULL; - return pin; + if (!cfg) + return; + + desc->chip_data = cfg; + + old_cfg = old_desc->chip_data; + + memcpy(cfg, old_cfg, sizeof(struct irq_cfg)); + + init_copy_irq_2_pin(old_cfg, cfg, cpu); +} + +static void free_irq_cfg(struct irq_cfg *old_cfg) +{ + kfree(old_cfg); } +void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) +{ + struct irq_cfg *old_cfg, *cfg; + + old_cfg = old_desc->chip_data; + cfg = desc->chip_data; + + if (old_cfg == cfg) + return; + + if (old_cfg) { + free_irq_2_pin(old_cfg, cfg); + free_irq_cfg(old_cfg); + old_desc->chip_data = NULL; + } +} + +static void +set_extra_move_desc(struct irq_desc *desc, const struct cpumask *mask) +{ + struct irq_cfg *cfg = desc->chip_data; + + if (!cfg->move_in_progress) { + /* it means that domain is not changed */ + if (!cpumask_intersects(&desc->affinity, mask)) + cfg->move_desc_pending = 1; + } +} +#endif + +#else +static struct irq_cfg *irq_cfg(unsigned int irq) +{ + return irq < nr_irqs ? irq_cfgx + irq : NULL; +} + +#endif + +#ifndef CONFIG_NUMA_MIGRATE_IRQ_DESC +static inline void +set_extra_move_desc(struct irq_desc *desc, const struct cpumask *mask) +{ +} +#endif + struct io_apic { unsigned int index; unsigned int unused[3]; @@ -237,11 +416,10 @@ static inline void io_apic_modify(unsigned int apic, unsigned int reg, unsigned writel(value, &io_apic->data); } -static bool io_apic_level_ack_pending(unsigned int irq) +static bool io_apic_level_ack_pending(struct irq_cfg *cfg) { struct irq_pin_list *entry; unsigned long flags; - struct irq_cfg *cfg = irq_cfg(irq); spin_lock_irqsave(&ioapic_lock, flags); entry = cfg->irq_2_pin; @@ -323,13 +501,32 @@ static void ioapic_mask_entry(int apic, int pin) } #ifdef CONFIG_SMP -static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, u8 vector) +static void send_cleanup_vector(struct irq_cfg *cfg) +{ + cpumask_var_t cleanup_mask; + + if (unlikely(!alloc_cpumask_var(&cleanup_mask, GFP_ATOMIC))) { + unsigned int i; + cfg->move_cleanup_count = 0; + for_each_cpu_and(i, cfg->old_domain, cpu_online_mask) + cfg->move_cleanup_count++; + for_each_cpu_and(i, cfg->old_domain, cpu_online_mask) + send_IPI_mask(cpumask_of(i), IRQ_MOVE_CLEANUP_VECTOR); + } else { + cpumask_and(cleanup_mask, cfg->old_domain, cpu_online_mask); + cfg->move_cleanup_count = cpumask_weight(cleanup_mask); + send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); + free_cpumask_var(cleanup_mask); + } + cfg->move_in_progress = 0; +} + +static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq_cfg *cfg) { int apic, pin; - struct irq_cfg *cfg; struct irq_pin_list *entry; + u8 vector = cfg->vector; - cfg = irq_cfg(irq); entry = cfg->irq_2_pin; for (;;) { unsigned int reg; @@ -359,36 +556,61 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, u8 vector) } } -static int assign_irq_vector(int irq, cpumask_t mask); +static int +assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask); + +/* + * Either sets desc->affinity to a valid value, and returns cpu_mask_to_apicid + * of that, or returns BAD_APICID and leaves desc->affinity untouched. + */ +static unsigned int +set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask) +{ + struct irq_cfg *cfg; + unsigned int irq; + + if (!cpumask_intersects(mask, cpu_online_mask)) + return BAD_APICID; -static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) + irq = desc->irq; + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) + return BAD_APICID; + + cpumask_and(&desc->affinity, cfg->domain, mask); + set_extra_move_desc(desc, mask); + return cpu_mask_to_apicid_and(&desc->affinity, cpu_online_mask); +} + +static void +set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) { struct irq_cfg *cfg; unsigned long flags; unsigned int dest; - cpumask_t tmp; - struct irq_desc *desc; + unsigned int irq; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) - return; + irq = desc->irq; + cfg = desc->chip_data; - cfg = irq_cfg(irq); - if (assign_irq_vector(irq, mask)) - return; + spin_lock_irqsave(&ioapic_lock, flags); + dest = set_desc_affinity(desc, mask); + if (dest != BAD_APICID) { + /* Only the high 8 bits are valid. */ + dest = SET_APIC_LOGICAL_ID(dest); + __target_IO_APIC_irq(irq, dest, cfg); + } + spin_unlock_irqrestore(&ioapic_lock, flags); +} - cpus_and(tmp, cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); - /* - * Only the high 8 bits are valid. - */ - dest = SET_APIC_LOGICAL_ID(dest); +static void +set_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) +{ + struct irq_desc *desc; desc = irq_to_desc(irq); - spin_lock_irqsave(&ioapic_lock, flags); - __target_IO_APIC_irq(irq, dest, cfg->vector); - desc->affinity = mask; - spin_unlock_irqrestore(&ioapic_lock, flags); + + set_ioapic_affinity_irq_desc(desc, mask); } #endif /* CONFIG_SMP */ @@ -397,16 +619,18 @@ static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) * shared ISA-space IRQs, so we have to support them. We are super * fast in the common case, and fast for shared ISA-space IRQs. */ -static void add_pin_to_irq(unsigned int irq, int apic, int pin) +static void add_pin_to_irq_cpu(struct irq_cfg *cfg, int cpu, int apic, int pin) { - struct irq_cfg *cfg; struct irq_pin_list *entry; - /* first time to refer irq_cfg, so with new */ - cfg = irq_cfg_alloc(irq); entry = cfg->irq_2_pin; if (!entry) { - entry = get_one_free_irq_2_pin(); + entry = get_one_free_irq_2_pin(cpu); + if (!entry) { + printk(KERN_ERR "can not alloc irq_2_pin to add %d - %d\n", + apic, pin); + return; + } cfg->irq_2_pin = entry; entry->apic = apic; entry->pin = pin; @@ -421,7 +645,7 @@ static void add_pin_to_irq(unsigned int irq, int apic, int pin) entry = entry->next; } - entry->next = get_one_free_irq_2_pin(); + entry->next = get_one_free_irq_2_pin(cpu); entry = entry->next; entry->apic = apic; entry->pin = pin; @@ -430,11 +654,10 @@ static void add_pin_to_irq(unsigned int irq, int apic, int pin) /* * Reroute an IRQ to a different pin. */ -static void __init replace_pin_at_irq(unsigned int irq, +static void __init replace_pin_at_irq_cpu(struct irq_cfg *cfg, int cpu, int oldapic, int oldpin, int newapic, int newpin) { - struct irq_cfg *cfg = irq_cfg(irq); struct irq_pin_list *entry = cfg->irq_2_pin; int replaced = 0; @@ -451,18 +674,16 @@ static void __init replace_pin_at_irq(unsigned int irq, /* why? call replace before add? */ if (!replaced) - add_pin_to_irq(irq, newapic, newpin); + add_pin_to_irq_cpu(cfg, cpu, newapic, newpin); } -static inline void io_apic_modify_irq(unsigned int irq, +static inline void io_apic_modify_irq(struct irq_cfg *cfg, int mask_and, int mask_or, void (*final)(struct irq_pin_list *entry)) { int pin; - struct irq_cfg *cfg; struct irq_pin_list *entry; - cfg = irq_cfg(irq); for (entry = cfg->irq_2_pin; entry != NULL; entry = entry->next) { unsigned int reg; pin = entry->pin; @@ -475,9 +696,9 @@ static inline void io_apic_modify_irq(unsigned int irq, } } -static void __unmask_IO_APIC_irq(unsigned int irq) +static void __unmask_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~IO_APIC_REDIR_MASKED, 0, NULL); + io_apic_modify_irq(cfg, ~IO_APIC_REDIR_MASKED, 0, NULL); } #ifdef CONFIG_X86_64 @@ -492,47 +713,64 @@ void io_apic_sync(struct irq_pin_list *entry) readl(&io_apic->data); } -static void __mask_IO_APIC_irq(unsigned int irq) +static void __mask_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~0, IO_APIC_REDIR_MASKED, &io_apic_sync); + io_apic_modify_irq(cfg, ~0, IO_APIC_REDIR_MASKED, &io_apic_sync); } #else /* CONFIG_X86_32 */ -static void __mask_IO_APIC_irq(unsigned int irq) +static void __mask_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~0, IO_APIC_REDIR_MASKED, NULL); + io_apic_modify_irq(cfg, ~0, IO_APIC_REDIR_MASKED, NULL); } -static void __mask_and_edge_IO_APIC_irq(unsigned int irq) +static void __mask_and_edge_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~IO_APIC_REDIR_LEVEL_TRIGGER, + io_apic_modify_irq(cfg, ~IO_APIC_REDIR_LEVEL_TRIGGER, IO_APIC_REDIR_MASKED, NULL); } -static void __unmask_and_level_IO_APIC_irq(unsigned int irq) +static void __unmask_and_level_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~IO_APIC_REDIR_MASKED, + io_apic_modify_irq(cfg, ~IO_APIC_REDIR_MASKED, IO_APIC_REDIR_LEVEL_TRIGGER, NULL); } #endif /* CONFIG_X86_32 */ -static void mask_IO_APIC_irq (unsigned int irq) +static void mask_IO_APIC_irq_desc(struct irq_desc *desc) { + struct irq_cfg *cfg = desc->chip_data; unsigned long flags; + BUG_ON(!cfg); + spin_lock_irqsave(&ioapic_lock, flags); - __mask_IO_APIC_irq(irq); + __mask_IO_APIC_irq(cfg); spin_unlock_irqrestore(&ioapic_lock, flags); } -static void unmask_IO_APIC_irq (unsigned int irq) +static void unmask_IO_APIC_irq_desc(struct irq_desc *desc) { + struct irq_cfg *cfg = desc->chip_data; unsigned long flags; spin_lock_irqsave(&ioapic_lock, flags); - __unmask_IO_APIC_irq(irq); + __unmask_IO_APIC_irq(cfg); spin_unlock_irqrestore(&ioapic_lock, flags); } +static void mask_IO_APIC_irq(unsigned int irq) +{ + struct irq_desc *desc = irq_to_desc(irq); + + mask_IO_APIC_irq_desc(desc); +} +static void unmask_IO_APIC_irq(unsigned int irq) +{ + struct irq_desc *desc = irq_to_desc(irq); + + unmask_IO_APIC_irq_desc(desc); +} + static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin) { struct IO_APIC_route_entry entry; @@ -809,7 +1047,7 @@ EXPORT_SYMBOL(IO_APIC_get_PCI_irq_vector); */ static int EISA_ELCR(unsigned int irq) { - if (irq < 16) { + if (irq < NR_IRQS_LEGACY) { unsigned int port = 0x4d0 + (irq >> 3); return (inb(port) >> (irq & 7)) & 1; } @@ -1034,7 +1272,8 @@ void unlock_vector_lock(void) spin_unlock(&vector_lock); } -static int __assign_irq_vector(int irq, cpumask_t mask) +static int +__assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask) { /* * NOTE! The local APIC isn't very good at handling @@ -1049,39 +1288,39 @@ static int __assign_irq_vector(int irq, cpumask_t mask) */ static int current_vector = FIRST_DEVICE_VECTOR, current_offset = 0; unsigned int old_vector; - int cpu; - struct irq_cfg *cfg; - - cfg = irq_cfg(irq); - - /* Only try and allocate irqs on cpus that are present */ - cpus_and(mask, mask, cpu_online_map); + int cpu, err; + cpumask_var_t tmp_mask; if ((cfg->move_in_progress) || cfg->move_cleanup_count) return -EBUSY; + if (!alloc_cpumask_var(&tmp_mask, GFP_ATOMIC)) + return -ENOMEM; + old_vector = cfg->vector; if (old_vector) { - cpumask_t tmp; - cpus_and(tmp, cfg->domain, mask); - if (!cpus_empty(tmp)) + cpumask_and(tmp_mask, mask, cpu_online_mask); + cpumask_and(tmp_mask, cfg->domain, tmp_mask); + if (!cpumask_empty(tmp_mask)) { + free_cpumask_var(tmp_mask); return 0; + } } - for_each_cpu_mask_nr(cpu, mask) { - cpumask_t domain, new_mask; + /* Only try and allocate irqs on cpus that are present */ + err = -ENOSPC; + for_each_cpu_and(cpu, mask, cpu_online_mask) { int new_cpu; int vector, offset; - domain = vector_allocation_domain(cpu); - cpus_and(new_mask, domain, cpu_online_map); + vector_allocation_domain(cpu, tmp_mask); vector = current_vector; offset = current_offset; next: vector += 8; if (vector >= first_system_vector) { - /* If we run out of vectors on large boxen, must share them. */ + /* If out of vectors on large boxen, must share them. */ offset = (offset + 1) % 8; vector = FIRST_DEVICE_VECTOR + offset; } @@ -1094,7 +1333,7 @@ next: if (vector == SYSCALL_VECTOR) goto next; #endif - for_each_cpu_mask_nr(new_cpu, new_mask) + for_each_cpu_and(new_cpu, tmp_mask, cpu_online_mask) if (per_cpu(vector_irq, new_cpu)[vector] != -1) goto next; /* Found one! */ @@ -1102,44 +1341,56 @@ next: current_offset = offset; if (old_vector) { cfg->move_in_progress = 1; - cfg->old_domain = cfg->domain; + cpumask_copy(cfg->old_domain, cfg->domain); } - for_each_cpu_mask_nr(new_cpu, new_mask) + for_each_cpu_and(new_cpu, tmp_mask, cpu_online_mask) per_cpu(vector_irq, new_cpu)[vector] = irq; cfg->vector = vector; - cfg->domain = domain; - return 0; + cpumask_copy(cfg->domain, tmp_mask); + err = 0; + break; } - return -ENOSPC; + free_cpumask_var(tmp_mask); + return err; } -static int assign_irq_vector(int irq, cpumask_t mask) +static int +assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask) { int err; unsigned long flags; spin_lock_irqsave(&vector_lock, flags); - err = __assign_irq_vector(irq, mask); + err = __assign_irq_vector(irq, cfg, mask); spin_unlock_irqrestore(&vector_lock, flags); return err; } -static void __clear_irq_vector(int irq) +static void __clear_irq_vector(int irq, struct irq_cfg *cfg) { - struct irq_cfg *cfg; - cpumask_t mask; int cpu, vector; - cfg = irq_cfg(irq); BUG_ON(!cfg->vector); vector = cfg->vector; - cpus_and(mask, cfg->domain, cpu_online_map); - for_each_cpu_mask_nr(cpu, mask) + for_each_cpu_and(cpu, cfg->domain, cpu_online_mask) per_cpu(vector_irq, cpu)[vector] = -1; cfg->vector = 0; - cpus_clear(cfg->domain); + cpumask_clear(cfg->domain); + + if (likely(!cfg->move_in_progress)) + return; + for_each_cpu_and(cpu, cfg->old_domain, cpu_online_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) @@ -1148,10 +1399,14 @@ void __setup_vector_irq(int cpu) /* This function must be called with vector_lock held */ int irq, vector; struct irq_cfg *cfg; + struct irq_desc *desc; /* Mark the inuse vectors */ - for_each_irq_cfg(irq, cfg) { - if (!cpu_isset(cpu, cfg->domain)) + for_each_irq_desc(irq, desc) { + if (!desc) + continue; + cfg = desc->chip_data; + if (!cpumask_test_cpu(cpu, cfg->domain)) continue; vector = cfg->vector; per_cpu(vector_irq, cpu)[vector] = irq; @@ -1163,7 +1418,7 @@ void __setup_vector_irq(int cpu) continue; cfg = irq_cfg(irq); - if (!cpu_isset(cpu, cfg->domain)) + if (!cpumask_test_cpu(cpu, cfg->domain)) per_cpu(vector_irq, cpu)[vector] = -1; } } @@ -1201,11 +1456,8 @@ static inline int IO_APIC_irq_trigger(int irq) } #endif -static void ioapic_register_intr(int irq, unsigned long trigger) +static void ioapic_register_intr(int irq, struct irq_desc *desc, unsigned long trigger) { - struct irq_desc *desc; - - desc = irq_to_desc(irq); if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) || trigger == IOAPIC_LEVEL) @@ -1297,23 +1549,22 @@ static int setup_ioapic_entry(int apic, int irq, return 0; } -static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, +static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, struct irq_desc *desc, int trigger, int polarity) { struct irq_cfg *cfg; struct IO_APIC_route_entry entry; - cpumask_t mask; + unsigned int dest; if (!IO_APIC_IRQ(irq)) return; - cfg = irq_cfg(irq); + cfg = desc->chip_data; - mask = TARGET_CPUS; - if (assign_irq_vector(irq, mask)) + if (assign_irq_vector(irq, cfg, TARGET_CPUS)) return; - cpus_and(mask, cfg->domain, mask); + dest = cpu_mask_to_apicid_and(cfg->domain, TARGET_CPUS); apic_printk(APIC_VERBOSE,KERN_DEBUG "IOAPIC[%d]: Set routing entry (%d-%d -> 0x%x -> " @@ -1323,16 +1574,15 @@ static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, if (setup_ioapic_entry(mp_ioapics[apic].mp_apicid, irq, &entry, - cpu_mask_to_apicid(mask), trigger, polarity, - cfg->vector)) { + dest, trigger, polarity, cfg->vector)) { printk("Failed to setup ioapic entry for ioapic %d, pin %d\n", mp_ioapics[apic].mp_apicid, pin); - __clear_irq_vector(irq); + __clear_irq_vector(irq, cfg); return; } - ioapic_register_intr(irq, trigger); - if (irq < 16) + ioapic_register_intr(irq, desc, trigger); + if (irq < NR_IRQS_LEGACY) disable_8259A_irq(irq); ioapic_write_entry(apic, pin, entry); @@ -1342,6 +1592,9 @@ static void __init setup_IO_APIC_irqs(void) { int apic, pin, idx, irq; int notcon = 0; + struct irq_desc *desc; + struct irq_cfg *cfg; + int cpu = boot_cpu_id; apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n"); @@ -1373,9 +1626,15 @@ static void __init setup_IO_APIC_irqs(void) if (multi_timer_check(apic, irq)) continue; #endif - add_pin_to_irq(irq, apic, pin); + desc = irq_to_desc_alloc_cpu(irq, cpu); + if (!desc) { + printk(KERN_INFO "can not get irq_desc for %d\n", irq); + continue; + } + cfg = desc->chip_data; + add_pin_to_irq_cpu(cfg, cpu, apic, pin); - setup_IO_APIC_irq(apic, pin, irq, + setup_IO_APIC_irq(apic, pin, irq, desc, irq_trigger(idx), irq_polarity(idx)); } } @@ -1434,6 +1693,7 @@ __apicdebuginit(void) print_IO_APIC(void) union IO_APIC_reg_03 reg_03; unsigned long flags; struct irq_cfg *cfg; + struct irq_desc *desc; unsigned int irq; if (apic_verbosity == APIC_QUIET) @@ -1523,8 +1783,13 @@ __apicdebuginit(void) print_IO_APIC(void) } } printk(KERN_DEBUG "IRQ to pin mappings:\n"); - for_each_irq_cfg(irq, cfg) { - struct irq_pin_list *entry = cfg->irq_2_pin; + for_each_irq_desc(irq, desc) { + struct irq_pin_list *entry; + + if (!desc) + continue; + cfg = desc->chip_data; + entry = cfg->irq_2_pin; if (!entry) continue; printk(KERN_DEBUG "IRQ%d ", irq); @@ -2008,14 +2273,16 @@ static unsigned int startup_ioapic_irq(unsigned int irq) { int was_pending = 0; unsigned long flags; + struct irq_cfg *cfg; spin_lock_irqsave(&ioapic_lock, flags); - if (irq < 16) { + if (irq < NR_IRQS_LEGACY) { disable_8259A_irq(irq); if (i8259A_irq_pending(irq)) was_pending = 1; } - __unmask_IO_APIC_irq(irq); + cfg = irq_cfg(irq); + __unmask_IO_APIC_irq(cfg); spin_unlock_irqrestore(&ioapic_lock, flags); return was_pending; @@ -2029,7 +2296,7 @@ static int ioapic_retrigger_irq(unsigned int irq) unsigned long flags; spin_lock_irqsave(&vector_lock, flags); - send_IPI_mask(cpumask_of_cpu(first_cpu(cfg->domain)), cfg->vector); + send_IPI_mask(cpumask_of(cpumask_first(cfg->domain)), cfg->vector); spin_unlock_irqrestore(&vector_lock, flags); return 1; @@ -2078,35 +2345,35 @@ static DECLARE_DELAYED_WORK(ir_migration_work, ir_irq_migration); * as simple as edge triggered migration and we can do the irq migration * with a simple atomic update to IO-APIC RTE. */ -static void migrate_ioapic_irq(int irq, cpumask_t mask) +static void +migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) { struct irq_cfg *cfg; - struct irq_desc *desc; - cpumask_t tmp, cleanup_mask; struct irte irte; int modify_ioapic_rte; unsigned int dest; unsigned long flags; + unsigned int irq; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + if (!cpumask_intersects(mask, cpu_online_mask)) return; + irq = desc->irq; if (get_irte(irq, &irte)) return; - if (assign_irq_vector(irq, mask)) + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) return; - cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + set_extra_move_desc(desc, mask); + + dest = cpu_mask_to_apicid_and(cfg->domain, mask); - desc = irq_to_desc(irq); modify_ioapic_rte = desc->status & IRQ_LEVEL; if (modify_ioapic_rte) { spin_lock_irqsave(&ioapic_lock, flags); - __target_IO_APIC_irq(irq, dest, cfg->vector); + __target_IO_APIC_irq(irq, dest, cfg); spin_unlock_irqrestore(&ioapic_lock, flags); } @@ -2118,24 +2385,20 @@ static void migrate_ioapic_irq(int irq, cpumask_t mask) */ modify_irte(irq, &irte); - if (cfg->move_in_progress) { - cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); - cfg->move_cleanup_count = cpus_weight(cleanup_mask); - send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); - cfg->move_in_progress = 0; - } + if (cfg->move_in_progress) + send_cleanup_vector(cfg); - desc->affinity = mask; + cpumask_copy(&desc->affinity, mask); } -static int migrate_irq_remapped_level(int irq) +static int migrate_irq_remapped_level_desc(struct irq_desc *desc) { int ret = -1; - struct irq_desc *desc = irq_to_desc(irq); + struct irq_cfg *cfg = desc->chip_data; - mask_IO_APIC_irq(irq); + mask_IO_APIC_irq_desc(desc); - if (io_apic_level_ack_pending(irq)) { + if (io_apic_level_ack_pending(cfg)) { /* * Interrupt in progress. Migrating irq now will change the * vector information in the IO-APIC RTE and that will confuse @@ -2147,14 +2410,15 @@ static int migrate_irq_remapped_level(int irq) } /* everthing is clear. we have right of way */ - migrate_ioapic_irq(irq, desc->pending_mask); + migrate_ioapic_irq_desc(desc, &desc->pending_mask); ret = 0; desc->status &= ~IRQ_MOVE_PENDING; - cpus_clear(desc->pending_mask); + cpumask_clear(&desc->pending_mask); unmask: - unmask_IO_APIC_irq(irq); + unmask_IO_APIC_irq_desc(desc); + return ret; } @@ -2164,6 +2428,9 @@ static void ir_irq_migration(struct work_struct *work) struct irq_desc *desc; for_each_irq_desc(irq, desc) { + if (!desc) + continue; + if (desc->status & IRQ_MOVE_PENDING) { unsigned long flags; @@ -2175,7 +2442,7 @@ static void ir_irq_migration(struct work_struct *work) continue; } - desc->chip->set_affinity(irq, desc->pending_mask); + desc->chip->set_affinity(irq, &desc->pending_mask); spin_unlock_irqrestore(&desc->lock, flags); } } @@ -2184,18 +2451,24 @@ static void ir_irq_migration(struct work_struct *work) /* * Migrates the IRQ destination in the process context. */ -static void set_ir_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) +static void set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, + const struct cpumask *mask) { - struct irq_desc *desc = irq_to_desc(irq); - if (desc->status & IRQ_LEVEL) { desc->status |= IRQ_MOVE_PENDING; - desc->pending_mask = mask; - migrate_irq_remapped_level(irq); + cpumask_copy(&desc->pending_mask, mask); + migrate_irq_remapped_level_desc(desc); return; } - migrate_ioapic_irq(irq, mask); + migrate_ioapic_irq_desc(desc, mask); +} +static void set_ir_ioapic_affinity_irq(unsigned int irq, + const struct cpumask *mask) +{ + struct irq_desc *desc = irq_to_desc(irq); + + set_ir_ioapic_affinity_irq_desc(desc, mask); } #endif @@ -2215,6 +2488,9 @@ asmlinkage void smp_irq_move_cleanup_interrupt(void) struct irq_cfg *cfg; irq = __get_cpu_var(vector_irq)[vector]; + if (irq == -1) + continue; + desc = irq_to_desc(irq); if (!desc) continue; @@ -2224,7 +2500,7 @@ asmlinkage void smp_irq_move_cleanup_interrupt(void) if (!cfg->move_cleanup_count) goto unlock; - if ((vector == cfg->vector) && cpu_isset(me, cfg->domain)) + if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain)) goto unlock; __get_cpu_var(vector_irq)[vector] = -1; @@ -2236,28 +2512,44 @@ unlock: irq_exit(); } -static void irq_complete_move(unsigned int irq) +static void irq_complete_move(struct irq_desc **descp) { - struct irq_cfg *cfg = irq_cfg(irq); + struct irq_desc *desc = *descp; + struct irq_cfg *cfg = desc->chip_data; unsigned vector, me; - if (likely(!cfg->move_in_progress)) + if (likely(!cfg->move_in_progress)) { +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC + if (likely(!cfg->move_desc_pending)) + return; + + /* domain is not change, but affinity is changed */ + me = smp_processor_id(); + if (cpu_isset(me, desc->affinity)) { + *descp = desc = move_irq_desc(desc, me); + /* get the new one */ + cfg = desc->chip_data; + cfg->move_desc_pending = 0; + } +#endif return; + } vector = ~get_irq_regs()->orig_ax; me = smp_processor_id(); - if ((vector == cfg->vector) && cpu_isset(me, cfg->domain)) { - cpumask_t cleanup_mask; +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC + *descp = desc = move_irq_desc(desc, me); + /* get the new one */ + cfg = desc->chip_data; +#endif - cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); - cfg->move_cleanup_count = cpus_weight(cleanup_mask); - send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); - cfg->move_in_progress = 0; - } + if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain)) + send_cleanup_vector(cfg); } #else -static inline void irq_complete_move(unsigned int irq) {} +static inline void irq_complete_move(struct irq_desc **descp) {} #endif + #ifdef CONFIG_INTR_REMAP static void ack_x2apic_level(unsigned int irq) { @@ -2268,11 +2560,14 @@ static void ack_x2apic_edge(unsigned int irq) { ack_x2APIC_irq(); } + #endif static void ack_apic_edge(unsigned int irq) { - irq_complete_move(irq); + struct irq_desc *desc = irq_to_desc(irq); + + irq_complete_move(&desc); move_native_irq(irq); ack_APIC_irq(); } @@ -2281,18 +2576,21 @@ atomic_t irq_mis_count; static void ack_apic_level(unsigned int irq) { + struct irq_desc *desc = irq_to_desc(irq); + #ifdef CONFIG_X86_32 unsigned long v; int i; #endif + struct irq_cfg *cfg; int do_unmask_irq = 0; - irq_complete_move(irq); + irq_complete_move(&desc); #ifdef CONFIG_GENERIC_PENDING_IRQ /* If we are moving the irq we need to mask it */ - if (unlikely(irq_to_desc(irq)->status & IRQ_MOVE_PENDING)) { + if (unlikely(desc->status & IRQ_MOVE_PENDING)) { do_unmask_irq = 1; - mask_IO_APIC_irq(irq); + mask_IO_APIC_irq_desc(desc); } #endif @@ -2316,7 +2614,8 @@ static void ack_apic_level(unsigned int irq) * operation to prevent an edge-triggered interrupt escaping meanwhile. * The idea is from Manfred Spraul. --macro */ - i = irq_cfg(irq)->vector; + cfg = desc->chip_data; + i = cfg->vector; v = apic_read(APIC_TMR + ((i & ~0x1f) >> 1)); #endif @@ -2355,17 +2654,18 @@ static void ack_apic_level(unsigned int irq) * accurate and is causing problems then it is a hardware bug * and you can go talk to the chipset vendor about it. */ - if (!io_apic_level_ack_pending(irq)) + cfg = desc->chip_data; + if (!io_apic_level_ack_pending(cfg)) move_masked_irq(irq); - unmask_IO_APIC_irq(irq); + unmask_IO_APIC_irq_desc(desc); } #ifdef CONFIG_X86_32 if (!(v & (1 << (i & 0x1f)))) { atomic_inc(&irq_mis_count); spin_lock(&ioapic_lock); - __mask_and_edge_IO_APIC_irq(irq); - __unmask_and_level_IO_APIC_irq(irq); + __mask_and_edge_IO_APIC_irq(cfg); + __unmask_and_level_IO_APIC_irq(cfg); spin_unlock(&ioapic_lock); } #endif @@ -2416,20 +2716,22 @@ static inline void init_IO_APIC_traps(void) * Also, we've got to be careful not to trash gate * 0x80, because int 0x80 is hm, kind of importantish. ;) */ - for_each_irq_cfg(irq, cfg) { - if (IO_APIC_IRQ(irq) && !cfg->vector) { + for_each_irq_desc(irq, desc) { + if (!desc) + continue; + + cfg = desc->chip_data; + if (IO_APIC_IRQ(irq) && cfg && !cfg->vector) { /* * Hmm.. We don't have an entry for this, * so default to an old-fashioned 8259 * interrupt if we can.. */ - if (irq < 16) + if (irq < NR_IRQS_LEGACY) make_8259A_irq(irq); - else { - desc = irq_to_desc(irq); + else /* Strange. Oh, well.. */ desc->chip = &no_irq_chip; - } } } } @@ -2454,7 +2756,7 @@ static void unmask_lapic_irq(unsigned int irq) apic_write(APIC_LVT0, v & ~APIC_LVT_MASKED); } -static void ack_lapic_irq (unsigned int irq) +static void ack_lapic_irq(unsigned int irq) { ack_APIC_irq(); } @@ -2466,11 +2768,8 @@ static struct irq_chip lapic_chip __read_mostly = { .ack = ack_lapic_irq, }; -static void lapic_register_intr(int irq) +static void lapic_register_intr(int irq, struct irq_desc *desc) { - struct irq_desc *desc; - - desc = irq_to_desc(irq); desc->status &= ~IRQ_LEVEL; set_irq_chip_and_handler_name(irq, &lapic_chip, handle_edge_irq, "edge"); @@ -2574,7 +2873,9 @@ int timer_through_8259 __initdata; */ static inline void __init check_timer(void) { - struct irq_cfg *cfg = irq_cfg(0); + struct irq_desc *desc = irq_to_desc(0); + struct irq_cfg *cfg = desc->chip_data; + int cpu = boot_cpu_id; int apic1, pin1, apic2, pin2; unsigned long flags; unsigned int ver; @@ -2589,7 +2890,7 @@ static inline void __init check_timer(void) * get/set the timer IRQ vector: */ disable_8259A_irq(0); - assign_irq_vector(0, TARGET_CPUS); + assign_irq_vector(0, cfg, TARGET_CPUS); /* * As IRQ0 is to be enabled in the 8259A, the virtual @@ -2640,10 +2941,10 @@ static inline void __init check_timer(void) * Ok, does IRQ0 through the IOAPIC work? */ if (no_pin1) { - add_pin_to_irq(0, apic1, pin1); + add_pin_to_irq_cpu(cfg, cpu, apic1, pin1); setup_timer_IRQ0_pin(apic1, pin1, cfg->vector); } - unmask_IO_APIC_irq(0); + unmask_IO_APIC_irq_desc(desc); if (timer_irq_works()) { if (nmi_watchdog == NMI_IO_APIC) { setup_nmi(); @@ -2669,9 +2970,9 @@ static inline void __init check_timer(void) /* * legacy devices should be connected to IO APIC #0 */ - replace_pin_at_irq(0, apic1, pin1, apic2, pin2); + replace_pin_at_irq_cpu(cfg, cpu, apic1, pin1, apic2, pin2); setup_timer_IRQ0_pin(apic2, pin2, cfg->vector); - unmask_IO_APIC_irq(0); + unmask_IO_APIC_irq_desc(desc); enable_8259A_irq(0); if (timer_irq_works()) { apic_printk(APIC_QUIET, KERN_INFO "....... works.\n"); @@ -2703,7 +3004,7 @@ static inline void __init check_timer(void) apic_printk(APIC_QUIET, KERN_INFO "...trying to set up timer as Virtual Wire IRQ...\n"); - lapic_register_intr(0); + lapic_register_intr(0, desc); apic_write(APIC_LVT0, APIC_DM_FIXED | cfg->vector); /* Fixed mode */ enable_8259A_irq(0); @@ -2888,22 +3189,26 @@ unsigned int create_irq_nr(unsigned int irq_want) unsigned int irq; unsigned int new; unsigned long flags; - struct irq_cfg *cfg_new; - - irq_want = nr_irqs - 1; + struct irq_cfg *cfg_new = NULL; + int cpu = boot_cpu_id; + struct irq_desc *desc_new = NULL; irq = 0; spin_lock_irqsave(&vector_lock, flags); - for (new = irq_want; new > 0; new--) { + for (new = irq_want; new < NR_IRQS; new++) { if (platform_legacy_irq(new)) continue; - cfg_new = irq_cfg(new); - if (cfg_new && cfg_new->vector != 0) + + desc_new = irq_to_desc_alloc_cpu(new, cpu); + if (!desc_new) { + printk(KERN_INFO "can not get irq_desc for %d\n", new); + continue; + } + cfg_new = desc_new->chip_data; + + if (cfg_new->vector != 0) continue; - /* check if need to create one */ - if (!cfg_new) - cfg_new = irq_cfg_alloc(new); - if (__assign_irq_vector(new, TARGET_CPUS) == 0) + if (__assign_irq_vector(new, cfg_new, TARGET_CPUS) == 0) irq = new; break; } @@ -2911,15 +3216,21 @@ unsigned int create_irq_nr(unsigned int irq_want) if (irq > 0) { dynamic_irq_init(irq); + /* restore it, in case dynamic_irq_init clear it */ + if (desc_new) + desc_new->chip_data = cfg_new; } return irq; } +static int nr_irqs_gsi = NR_IRQS_LEGACY; int create_irq(void) { + unsigned int irq_want; int irq; - irq = create_irq_nr(nr_irqs - 1); + irq_want = nr_irqs_gsi; + irq = create_irq_nr(irq_want); if (irq == 0) irq = -1; @@ -2930,14 +3241,22 @@ int create_irq(void) void destroy_irq(unsigned int irq) { unsigned long flags; + struct irq_cfg *cfg; + struct irq_desc *desc; + /* store it, in case dynamic_irq_cleanup clear it */ + desc = irq_to_desc(irq); + cfg = desc->chip_data; dynamic_irq_cleanup(irq); + /* connect back irq_cfg */ + if (desc) + desc->chip_data = cfg; #ifdef CONFIG_INTR_REMAP free_irte(irq); #endif spin_lock_irqsave(&vector_lock, flags); - __clear_irq_vector(irq); + __clear_irq_vector(irq, cfg); spin_unlock_irqrestore(&vector_lock, flags); } @@ -2950,16 +3269,13 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms struct irq_cfg *cfg; int err; unsigned dest; - cpumask_t tmp; - tmp = TARGET_CPUS; - err = assign_irq_vector(irq, tmp); + cfg = irq_cfg(irq); + err = assign_irq_vector(irq, cfg, TARGET_CPUS); if (err) return err; - cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, tmp); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid_and(cfg->domain, TARGET_CPUS); #ifdef CONFIG_INTR_REMAP if (irq_remapped(irq)) { @@ -3013,64 +3329,48 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms } #ifdef CONFIG_SMP -static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) +static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; struct msi_msg msg; unsigned int dest; - cpumask_t tmp; - struct irq_desc *desc; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; - if (assign_irq_vector(irq, mask)) - return; - - cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + cfg = desc->chip_data; - read_msi_msg(irq, &msg); + read_msi_msg_desc(desc, &msg); msg.data &= ~MSI_DATA_VECTOR_MASK; msg.data |= MSI_DATA_VECTOR(cfg->vector); msg.address_lo &= ~MSI_ADDR_DEST_ID_MASK; msg.address_lo |= MSI_ADDR_DEST_ID(dest); - write_msi_msg(irq, &msg); - desc = irq_to_desc(irq); - desc->affinity = mask; + write_msi_msg_desc(desc, &msg); } - #ifdef CONFIG_INTR_REMAP /* * Migrate the MSI irq to another cpumask. This migration is * done in the process context using interrupt-remapping hardware. */ -static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask) +static void +ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; unsigned int dest; - cpumask_t tmp, cleanup_mask; struct irte irte; - struct irq_desc *desc; - - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) - return; if (get_irte(irq, &irte)) return; - if (assign_irq_vector(irq, mask)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; - cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); - irte.vector = cfg->vector; irte.dest_id = IRTE_DEST(dest); @@ -3084,16 +3384,10 @@ static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask) * at the new destination. So, time to cleanup the previous * vector allocation. */ - if (cfg->move_in_progress) { - cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); - cfg->move_cleanup_count = cpus_weight(cleanup_mask); - send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); - cfg->move_in_progress = 0; - } - - desc = irq_to_desc(irq); - desc->affinity = mask; + if (cfg->move_in_progress) + send_cleanup_vector(cfg); } + #endif #endif /* CONFIG_SMP */ @@ -3152,7 +3446,7 @@ static int msi_alloc_irte(struct pci_dev *dev, int irq, int nvec) } #endif -static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc, int irq) +static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc, int irq) { int ret; struct msi_msg msg; @@ -3161,7 +3455,7 @@ static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc, int irq) if (ret < 0) return ret; - set_irq_msi(irq, desc); + set_irq_msi(irq, msidesc); write_msi_msg(irq, &msg); #ifdef CONFIG_INTR_REMAP @@ -3181,26 +3475,13 @@ static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc, int irq) return 0; } -static unsigned int build_irq_for_pci_dev(struct pci_dev *dev) -{ - unsigned int irq; - - irq = dev->bus->number; - irq <<= 8; - irq |= dev->devfn; - irq <<= 12; - - return irq; -} - -int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc) +int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc) { unsigned int irq; int ret; unsigned int irq_want; - irq_want = build_irq_for_pci_dev(dev) + 0x100; - + irq_want = nr_irqs_gsi; irq = create_irq_nr(irq_want); if (irq == 0) return -1; @@ -3214,7 +3495,7 @@ int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc) goto error; no_ir: #endif - ret = setup_msi_irq(dev, desc, irq); + ret = setup_msi_irq(dev, msidesc, irq); if (ret < 0) { destroy_irq(irq); return ret; @@ -3232,7 +3513,7 @@ int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) { unsigned int irq; int ret, sub_handle; - struct msi_desc *desc; + struct msi_desc *msidesc; unsigned int irq_want; #ifdef CONFIG_INTR_REMAP @@ -3240,10 +3521,11 @@ int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) int index = 0; #endif - irq_want = build_irq_for_pci_dev(dev) + 0x100; + irq_want = nr_irqs_gsi; sub_handle = 0; - list_for_each_entry(desc, &dev->msi_list, list) { - irq = create_irq_nr(irq_want--); + list_for_each_entry(msidesc, &dev->msi_list, list) { + irq = create_irq_nr(irq_want); + irq_want++; if (irq == 0) return -1; #ifdef CONFIG_INTR_REMAP @@ -3275,7 +3557,7 @@ int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) } no_ir: #endif - ret = setup_msi_irq(dev, desc, irq); + ret = setup_msi_irq(dev, msidesc, irq); if (ret < 0) goto error; sub_handle++; @@ -3294,24 +3576,18 @@ void arch_teardown_msi_irq(unsigned int irq) #ifdef CONFIG_DMAR #ifdef CONFIG_SMP -static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) +static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; struct msi_msg msg; unsigned int dest; - cpumask_t tmp; - struct irq_desc *desc; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; - if (assign_irq_vector(irq, mask)) - return; - - cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + cfg = desc->chip_data; dmar_msi_read(irq, &msg); @@ -3321,9 +3597,8 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); dmar_msi_write(irq, &msg); - desc = irq_to_desc(irq); - desc->affinity = mask; } + #endif /* CONFIG_SMP */ struct irq_chip dmar_msi_type = { @@ -3355,24 +3630,18 @@ int arch_setup_dmar_msi(unsigned int irq) #ifdef CONFIG_HPET_TIMER #ifdef CONFIG_SMP -static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask) +static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; - struct irq_desc *desc; struct msi_msg msg; unsigned int dest; - cpumask_t tmp; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; - if (assign_irq_vector(irq, mask)) - return; - - cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + cfg = desc->chip_data; hpet_msi_read(irq, &msg); @@ -3382,9 +3651,8 @@ static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); hpet_msi_write(irq, &msg); - desc = irq_to_desc(irq); - desc->affinity = mask; } + #endif /* CONFIG_SMP */ struct irq_chip hpet_msi_type = { @@ -3437,28 +3705,21 @@ static void target_ht_irq(unsigned int irq, unsigned int dest, u8 vector) write_ht_irq_msg(irq, &msg); } -static void set_ht_irq_affinity(unsigned int irq, cpumask_t mask) +static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; unsigned int dest; - cpumask_t tmp; - struct irq_desc *desc; - - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) - return; - if (assign_irq_vector(irq, mask)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; - cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + cfg = desc->chip_data; target_ht_irq(irq, dest, cfg->vector); - desc = irq_to_desc(irq); - desc->affinity = mask; } + #endif static struct irq_chip ht_irq_chip = { @@ -3476,17 +3737,14 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) { struct irq_cfg *cfg; int err; - cpumask_t tmp; - tmp = TARGET_CPUS; - err = assign_irq_vector(irq, tmp); + cfg = irq_cfg(irq); + err = assign_irq_vector(irq, cfg, TARGET_CPUS); if (!err) { struct ht_irq_msg msg; unsigned dest; - cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, tmp); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid_and(cfg->domain, TARGET_CPUS); msg.address_hi = HT_IRQ_HIGH_DEST_ID(dest); @@ -3522,7 +3780,7 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, unsigned long mmr_offset) { - const cpumask_t *eligible_cpu = get_cpu_mask(cpu); + const struct cpumask *eligible_cpu = cpumask_of(cpu); struct irq_cfg *cfg; int mmr_pnode; unsigned long mmr_value; @@ -3530,7 +3788,9 @@ int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, unsigned long flags; int err; - err = assign_irq_vector(irq, *eligible_cpu); + cfg = irq_cfg(irq); + + err = assign_irq_vector(irq, cfg, eligible_cpu); if (err != 0) return err; @@ -3539,8 +3799,6 @@ int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, irq_name); spin_unlock_irqrestore(&vector_lock, flags); - cfg = irq_cfg(irq); - mmr_value = 0; entry = (struct uv_IO_APIC_route_entry *)&mmr_value; BUG_ON(sizeof(struct uv_IO_APIC_route_entry) != sizeof(unsigned long)); @@ -3551,7 +3809,7 @@ int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, entry->polarity = 0; entry->trigger = 0; entry->mask = 0; - entry->dest = cpu_mask_to_apicid(*eligible_cpu); + entry->dest = cpu_mask_to_apicid(eligible_cpu); mmr_pnode = uv_blade_to_pnode(mmr_blade); uv_write_global_mmr64(mmr_pnode, mmr_offset, mmr_value); @@ -3592,29 +3850,16 @@ int __init io_apic_get_redir_entries (int ioapic) return reg_01.bits.entries; } -int __init probe_nr_irqs(void) +void __init probe_nr_irqs_gsi(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; + if (nr > nr_irqs_gsi) + nr_irqs_gsi = nr; } /* -------------------------------------------------------------------------- @@ -3713,19 +3958,31 @@ int __init io_apic_get_version(int ioapic) int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int polarity) { + struct irq_desc *desc; + struct irq_cfg *cfg; + int cpu = boot_cpu_id; + if (!IO_APIC_IRQ(irq)) { apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n", ioapic); return -EINVAL; } + desc = irq_to_desc_alloc_cpu(irq, cpu); + if (!desc) { + printk(KERN_INFO "can not get irq_desc %d\n", irq); + return 0; + } + /* * IRQs < 16 are already in the irq_2_pin[] map */ - if (irq >= 16) - add_pin_to_irq(irq, ioapic, pin); + if (irq >= NR_IRQS_LEGACY) { + cfg = desc->chip_data; + add_pin_to_irq_cpu(cfg, cpu, ioapic, pin); + } - setup_IO_APIC_irq(ioapic, pin, irq, triggering, polarity); + setup_IO_APIC_irq(ioapic, pin, irq, desc, triggering, polarity); return 0; } @@ -3761,7 +4018,9 @@ 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; + const struct cpumask *mask; if (skip_ioapic_setup == 1) return; @@ -3777,17 +4036,31 @@ void __init setup_ioapic_dest(void) * when you have too many devices, because at that time only boot * cpu is online. */ - cfg = irq_cfg(irq); - if (!cfg->vector) - setup_IO_APIC_irq(ioapic, pin, irq, + desc = irq_to_desc(irq); + cfg = desc->chip_data; + if (!cfg->vector) { + setup_IO_APIC_irq(ioapic, pin, irq, desc, irq_trigger(irq_entry), irq_polarity(irq_entry)); + continue; + + } + + /* + * Honour affinities which have been set in early boot + */ + 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_desc(desc, mask); else - set_ioapic_affinity_irq(irq, TARGET_CPUS); +#endif + set_ioapic_affinity_irq_desc(desc, mask); } } @@ -3836,7 +4109,6 @@ void __init ioapic_init_mappings(void) struct resource *ioapic_res; int i; - irq_2_pin_init(); ioapic_res = ioapic_setup_resources(); for (i = 0; i < nr_ioapics; i++) { if (smp_found_config) { diff --git a/arch/x86/kernel/ipi.c b/arch/x86/kernel/ipi.c index f1c688e46f3..285bbf8831f 100644 --- a/arch/x86/kernel/ipi.c +++ b/arch/x86/kernel/ipi.c @@ -116,18 +116,18 @@ static inline void __send_IPI_dest_field(unsigned long mask, int vector) /* * This is only used on smaller machines. */ -void send_IPI_mask_bitmask(cpumask_t cpumask, int vector) +void send_IPI_mask_bitmask(const struct cpumask *cpumask, int vector) { - unsigned long mask = cpus_addr(cpumask)[0]; + unsigned long mask = cpumask_bits(cpumask)[0]; unsigned long flags; local_irq_save(flags); - WARN_ON(mask & ~cpus_addr(cpu_online_map)[0]); + WARN_ON(mask & ~cpumask_bits(cpu_online_mask)[0]); __send_IPI_dest_field(mask, vector); local_irq_restore(flags); } -void send_IPI_mask_sequence(cpumask_t mask, int vector) +void send_IPI_mask_sequence(const struct cpumask *mask, int vector) { unsigned long flags; unsigned int query_cpu; @@ -139,12 +139,24 @@ void send_IPI_mask_sequence(cpumask_t mask, int vector) */ local_irq_save(flags); - for_each_possible_cpu(query_cpu) { - if (cpu_isset(query_cpu, mask)) { + for_each_cpu(query_cpu, mask) + __send_IPI_dest_field(cpu_to_logical_apicid(query_cpu), vector); + local_irq_restore(flags); +} + +void send_IPI_mask_allbutself(const struct cpumask *mask, int vector) +{ + unsigned long flags; + unsigned int query_cpu; + unsigned int this_cpu = smp_processor_id(); + + /* See Hack comment above */ + + local_irq_save(flags); + for_each_cpu(query_cpu, mask) + if (query_cpu != this_cpu) __send_IPI_dest_field(cpu_to_logical_apicid(query_cpu), vector); - } - } local_irq_restore(flags); } diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index d1d4dc52f64..3f1d9d18df6 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -118,6 +118,9 @@ int show_interrupts(struct seq_file *p, void *v) } desc = irq_to_desc(i); + if (!desc) + return 0; + spin_lock_irqsave(&desc->lock, flags); #ifndef CONFIG_SMP any_count = kstat_irqs(i); diff --git a/arch/x86/kernel/irq_32.c b/arch/x86/kernel/irq_32.c index a51382672de..9dc5588f336 100644 --- a/arch/x86/kernel/irq_32.c +++ b/arch/x86/kernel/irq_32.c @@ -233,25 +233,28 @@ unsigned int do_IRQ(struct pt_regs *regs) #ifdef CONFIG_HOTPLUG_CPU #include <mach_apic.h> -void fixup_irqs(cpumask_t map) +/* A cpu has been removed from cpu_online_mask. Reset irq affinities. */ +void fixup_irqs(void) { unsigned int irq; static int warned; struct irq_desc *desc; for_each_irq_desc(irq, desc) { - cpumask_t mask; + const struct cpumask *affinity; + if (!desc) + continue; if (irq == 2) continue; - cpus_and(mask, desc->affinity, map); - if (any_online_cpu(mask) == NR_CPUS) { + affinity = &desc->affinity; + if (cpumask_any_and(affinity, cpu_online_mask) >= nr_cpu_ids) { printk("Breaking affinity for irq %i\n", irq); - mask = map; + affinity = cpu_all_mask; } if (desc->chip->set_affinity) - desc->chip->set_affinity(irq, mask); + desc->chip->set_affinity(irq, affinity); else if (desc->action && !(warned++)) printk("Cannot set affinity for irq %i\n", irq); } diff --git a/arch/x86/kernel/irq_64.c b/arch/x86/kernel/irq_64.c index 60eb84eb77a..fca2991443f 100644 --- a/arch/x86/kernel/irq_64.c +++ b/arch/x86/kernel/irq_64.c @@ -83,40 +83,43 @@ asmlinkage unsigned int do_IRQ(struct pt_regs *regs) } #ifdef CONFIG_HOTPLUG_CPU -void fixup_irqs(cpumask_t map) +/* A cpu has been removed from cpu_online_mask. Reset irq affinities. */ +void fixup_irqs(void) { unsigned int irq; static int warned; struct irq_desc *desc; for_each_irq_desc(irq, desc) { - cpumask_t mask; int break_affinity = 0; int set_affinity = 1; + const struct cpumask *affinity; + if (!desc) + continue; if (irq == 2) continue; /* interrupt's are disabled at this point */ spin_lock(&desc->lock); + affinity = &desc->affinity; if (!irq_has_action(irq) || - cpus_equal(desc->affinity, map)) { + cpumask_equal(affinity, cpu_online_mask)) { spin_unlock(&desc->lock); continue; } - cpus_and(mask, desc->affinity, map); - if (cpus_empty(mask)) { + if (cpumask_any_and(affinity, cpu_online_mask) >= nr_cpu_ids) { break_affinity = 1; - mask = map; + affinity = cpu_all_mask; } if (desc->chip->mask) desc->chip->mask(irq); if (desc->chip->set_affinity) - desc->chip->set_affinity(irq, mask); + desc->chip->set_affinity(irq, affinity); else if (!(warned++)) set_affinity = 0; diff --git a/arch/x86/kernel/irqinit_32.c b/arch/x86/kernel/irqinit_32.c index 845aa9803e8..6a92f47c52e 100644 --- a/arch/x86/kernel/irqinit_32.c +++ b/arch/x86/kernel/irqinit_32.c @@ -68,8 +68,7 @@ void __init init_ISA_irqs (void) /* * 16 old-style INTA-cycle interrupts: */ - for (i = 0; i < 16; i++) { - /* first time call this irq_desc */ + for (i = 0; i < NR_IRQS_LEGACY; i++) { struct irq_desc *desc = irq_to_desc(i); desc->status = IRQ_DISABLED; diff --git a/arch/x86/kernel/irqinit_64.c b/arch/x86/kernel/irqinit_64.c index ff023539128..40c1e62ec78 100644 --- a/arch/x86/kernel/irqinit_64.c +++ b/arch/x86/kernel/irqinit_64.c @@ -142,8 +142,7 @@ void __init init_ISA_irqs(void) init_bsp_APIC(); init_8259A(0); - for (i = 0; i < 16; i++) { - /* first time call this irq_desc */ + for (i = 0; i < NR_IRQS_LEGACY; i++) { struct irq_desc *desc = irq_to_desc(i); desc->status = IRQ_DISABLED; diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c index 774ac499156..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 kvm_setup_secondary_clock(void) +static void __cpuinit kvm_setup_secondary_clock(void) { /* * Now that the first cpu already had this clocksource initialized, diff --git a/arch/x86/kernel/machine_kexec_32.c b/arch/x86/kernel/machine_kexec_32.c index 7a385746509..37f420018a4 100644 --- a/arch/x86/kernel/machine_kexec_32.c +++ b/arch/x86/kernel/machine_kexec_32.c @@ -13,6 +13,7 @@ #include <linux/numa.h> #include <linux/ftrace.h> #include <linux/suspend.h> +#include <linux/gfp.h> #include <asm/pgtable.h> #include <asm/pgalloc.h> @@ -25,15 +26,6 @@ #include <asm/system.h> #include <asm/cacheflush.h> -#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE))) -static u32 kexec_pgd[1024] PAGE_ALIGNED; -#ifdef CONFIG_X86_PAE -static u32 kexec_pmd0[1024] PAGE_ALIGNED; -static u32 kexec_pmd1[1024] PAGE_ALIGNED; -#endif -static u32 kexec_pte0[1024] PAGE_ALIGNED; -static u32 kexec_pte1[1024] PAGE_ALIGNED; - static void set_idt(void *newidt, __u16 limit) { struct desc_ptr curidt; @@ -76,6 +68,76 @@ static void load_segments(void) #undef __STR } +static void machine_kexec_free_page_tables(struct kimage *image) +{ + free_page((unsigned long)image->arch.pgd); +#ifdef CONFIG_X86_PAE + free_page((unsigned long)image->arch.pmd0); + free_page((unsigned long)image->arch.pmd1); +#endif + free_page((unsigned long)image->arch.pte0); + free_page((unsigned long)image->arch.pte1); +} + +static int machine_kexec_alloc_page_tables(struct kimage *image) +{ + image->arch.pgd = (pgd_t *)get_zeroed_page(GFP_KERNEL); +#ifdef CONFIG_X86_PAE + image->arch.pmd0 = (pmd_t *)get_zeroed_page(GFP_KERNEL); + image->arch.pmd1 = (pmd_t *)get_zeroed_page(GFP_KERNEL); +#endif + image->arch.pte0 = (pte_t *)get_zeroed_page(GFP_KERNEL); + image->arch.pte1 = (pte_t *)get_zeroed_page(GFP_KERNEL); + if (!image->arch.pgd || +#ifdef CONFIG_X86_PAE + !image->arch.pmd0 || !image->arch.pmd1 || +#endif + !image->arch.pte0 || !image->arch.pte1) { + machine_kexec_free_page_tables(image); + return -ENOMEM; + } + return 0; +} + +static void machine_kexec_page_table_set_one( + pgd_t *pgd, pmd_t *pmd, pte_t *pte, + unsigned long vaddr, unsigned long paddr) +{ + pud_t *pud; + + pgd += pgd_index(vaddr); +#ifdef CONFIG_X86_PAE + if (!(pgd_val(*pgd) & _PAGE_PRESENT)) + set_pgd(pgd, __pgd(__pa(pmd) | _PAGE_PRESENT)); +#endif + pud = pud_offset(pgd, vaddr); + pmd = pmd_offset(pud, vaddr); + if (!(pmd_val(*pmd) & _PAGE_PRESENT)) + set_pmd(pmd, __pmd(__pa(pte) | _PAGE_TABLE)); + pte = pte_offset_kernel(pmd, vaddr); + set_pte(pte, pfn_pte(paddr >> PAGE_SHIFT, PAGE_KERNEL_EXEC)); +} + +static void machine_kexec_prepare_page_tables(struct kimage *image) +{ + void *control_page; + pmd_t *pmd = 0; + + control_page = page_address(image->control_code_page); +#ifdef CONFIG_X86_PAE + pmd = image->arch.pmd0; +#endif + machine_kexec_page_table_set_one( + image->arch.pgd, pmd, image->arch.pte0, + (unsigned long)control_page, __pa(control_page)); +#ifdef CONFIG_X86_PAE + pmd = image->arch.pmd1; +#endif + machine_kexec_page_table_set_one( + image->arch.pgd, pmd, image->arch.pte1, + __pa(control_page), __pa(control_page)); +} + /* * A architecture hook called to validate the * proposed image and prepare the control pages @@ -87,12 +149,20 @@ static void load_segments(void) * reboot code buffer to allow us to avoid allocations * later. * - * Make control page executable. + * - Make control page executable. + * - Allocate page tables + * - Setup page tables */ int machine_kexec_prepare(struct kimage *image) { + int error; + if (nx_enabled) set_pages_x(image->control_code_page, 1); + error = machine_kexec_alloc_page_tables(image); + if (error) + return error; + machine_kexec_prepare_page_tables(image); return 0; } @@ -104,6 +174,7 @@ void machine_kexec_cleanup(struct kimage *image) { if (nx_enabled) set_pages_nx(image->control_code_page, 1); + machine_kexec_free_page_tables(image); } /* @@ -150,18 +221,7 @@ void machine_kexec(struct kimage *image) relocate_kernel_ptr = control_page; page_list[PA_CONTROL_PAGE] = __pa(control_page); page_list[VA_CONTROL_PAGE] = (unsigned long)control_page; - page_list[PA_PGD] = __pa(kexec_pgd); - page_list[VA_PGD] = (unsigned long)kexec_pgd; -#ifdef CONFIG_X86_PAE - page_list[PA_PMD_0] = __pa(kexec_pmd0); - page_list[VA_PMD_0] = (unsigned long)kexec_pmd0; - page_list[PA_PMD_1] = __pa(kexec_pmd1); - page_list[VA_PMD_1] = (unsigned long)kexec_pmd1; -#endif - page_list[PA_PTE_0] = __pa(kexec_pte0); - page_list[VA_PTE_0] = (unsigned long)kexec_pte0; - page_list[PA_PTE_1] = __pa(kexec_pte1); - page_list[VA_PTE_1] = (unsigned long)kexec_pte1; + page_list[PA_PGD] = __pa(image->arch.pgd); if (image->type == KEXEC_TYPE_DEFAULT) page_list[PA_SWAP_PAGE] = (page_to_pfn(image->swap_page) diff --git a/arch/x86/kernel/mfgpt_32.c b/arch/x86/kernel/mfgpt_32.c index 3b599518c32..c12314c9e86 100644 --- a/arch/x86/kernel/mfgpt_32.c +++ b/arch/x86/kernel/mfgpt_32.c @@ -287,7 +287,7 @@ static struct clock_event_device mfgpt_clockevent = { .set_mode = mfgpt_set_mode, .set_next_event = mfgpt_next_event, .rating = 250, - .cpumask = CPU_MASK_ALL, + .cpumask = cpu_all_mask, .shift = 32 }; diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c index f98f4e1dba0..45e3b69808b 100644 --- a/arch/x86/kernel/mpparse.c +++ b/arch/x86/kernel/mpparse.c @@ -586,23 +586,23 @@ static void __init __get_smp_config(unsigned int early) { struct intel_mp_floating *mpf = mpf_found; - if (x86_quirks->mach_get_smp_config) { - if (x86_quirks->mach_get_smp_config(early)) - return; - } + if (!mpf) + return; + if (acpi_lapic && early) return; + /* - * ACPI supports both logical (e.g. Hyper-Threading) and physical - * processors, where MPS only supports physical. + * MPS doesn't support hyperthreading, aka only have + * thread 0 apic id in MPS table */ - if (acpi_lapic && acpi_ioapic) { - printk(KERN_INFO "Using ACPI (MADT) for SMP configuration " - "information\n"); + if (acpi_lapic && acpi_ioapic) return; - } else if (acpi_lapic) - printk(KERN_INFO "Using ACPI for processor (LAPIC) " - "configuration information\n"); + + if (x86_quirks->mach_get_smp_config) { + if (x86_quirks->mach_get_smp_config(early)) + return; + } printk(KERN_INFO "Intel MultiProcessor Specification v1.%d\n", mpf->mpf_specification); diff --git a/arch/x86/kernel/numaq_32.c b/arch/x86/kernel/numaq_32.c index 4caff39078e..0deea37a53c 100644 --- a/arch/x86/kernel/numaq_32.c +++ b/arch/x86/kernel/numaq_32.c @@ -31,7 +31,7 @@ #include <asm/numaq.h> #include <asm/topology.h> #include <asm/processor.h> -#include <asm/mpspec.h> +#include <asm/genapic.h> #include <asm/e820.h> #include <asm/setup.h> @@ -235,6 +235,13 @@ static int __init numaq_setup_ioapic_ids(void) return 1; } +static int __init numaq_update_genapic(void) +{ + genapic->wakeup_cpu = wakeup_secondary_cpu_via_nmi; + + return 0; +} + static struct x86_quirks numaq_x86_quirks __initdata = { .arch_pre_time_init = numaq_pre_time_init, .arch_time_init = NULL, @@ -250,6 +257,7 @@ static struct x86_quirks numaq_x86_quirks __initdata = { .mpc_oem_pci_bus = mpc_oem_pci_bus, .smp_read_mpc_oem = smp_read_mpc_oem, .setup_ioapic_ids = numaq_setup_ioapic_ids, + .update_genapic = numaq_update_genapic, }; void numaq_mps_oem_check(struct mp_config_table *mpc, char *oem, diff --git a/arch/x86/kernel/paravirt-spinlocks.c b/arch/x86/kernel/paravirt-spinlocks.c index 0e9f1982b1d..95777b0faa7 100644 --- a/arch/x86/kernel/paravirt-spinlocks.c +++ b/arch/x86/kernel/paravirt-spinlocks.c @@ -7,7 +7,8 @@ #include <asm/paravirt.h> -static void default_spin_lock_flags(struct raw_spinlock *lock, unsigned long flags) +static inline void +default_spin_lock_flags(raw_spinlock_t *lock, unsigned long flags) { __raw_spin_lock(lock); } 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; diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c index a42b02b4df6..ba7ad83e20a 100644 --- a/arch/x86/kernel/pci-gart_64.c +++ b/arch/x86/kernel/pci-gart_64.c @@ -123,6 +123,8 @@ static void free_iommu(unsigned long offset, int size) spin_lock_irqsave(&iommu_bitmap_lock, flags); iommu_area_free(iommu_gart_bitmap, offset, size); + if (offset >= next_bit) + next_bit = offset + size; spin_unlock_irqrestore(&iommu_bitmap_lock, flags); } diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index c622772744d..95d811a9594 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -7,7 +7,9 @@ #include <linux/module.h> #include <linux/pm.h> #include <linux/clockchips.h> +#include <linux/ftrace.h> #include <asm/system.h> +#include <asm/apic.h> unsigned long idle_halt; EXPORT_SYMBOL(idle_halt); @@ -100,6 +102,9 @@ static inline int hlt_use_halt(void) void default_idle(void) { if (hlt_use_halt()) { + struct power_trace it; + + trace_power_start(&it, POWER_CSTATE, 1); current_thread_info()->status &= ~TS_POLLING; /* * TS_POLLING-cleared state must be visible before we @@ -112,6 +117,7 @@ void default_idle(void) else local_irq_enable(); current_thread_info()->status |= TS_POLLING; + trace_power_end(&it); } else { local_irq_enable(); /* loop is done by the caller */ @@ -122,6 +128,21 @@ void default_idle(void) EXPORT_SYMBOL(default_idle); #endif +void stop_this_cpu(void *dummy) +{ + local_irq_disable(); + /* + * Remove this CPU: + */ + cpu_clear(smp_processor_id(), cpu_online_map); + disable_local_APIC(); + + for (;;) { + if (hlt_works(smp_processor_id())) + halt(); + } +} + static void do_nothing(void *unused) { } @@ -154,24 +175,31 @@ EXPORT_SYMBOL_GPL(cpu_idle_wait); */ void mwait_idle_with_hints(unsigned long ax, unsigned long cx) { + struct power_trace it; + + trace_power_start(&it, POWER_CSTATE, (ax>>4)+1); if (!need_resched()) { __monitor((void *)¤t_thread_info()->flags, 0, 0); smp_mb(); if (!need_resched()) __mwait(ax, cx); } + trace_power_end(&it); } /* Default MONITOR/MWAIT with no hints, used for default C1 state */ static void mwait_idle(void) { + struct power_trace it; if (!need_resched()) { + trace_power_start(&it, POWER_CSTATE, 1); __monitor((void *)¤t_thread_info()->flags, 0, 0); smp_mb(); if (!need_resched()) __sti_mwait(0, 0); else local_irq_enable(); + trace_power_end(&it); } else local_irq_enable(); } @@ -183,9 +211,13 @@ static void mwait_idle(void) */ static void poll_idle(void) { + struct power_trace it; + + trace_power_start(&it, POWER_CSTATE, 0); local_irq_enable(); while (!need_resched()) cpu_relax(); + trace_power_end(&it); } /* diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 0a1302fe6d4..24c2276aa45 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -38,6 +38,7 @@ #include <linux/percpu.h> #include <linux/prctl.h> #include <linux/dmi.h> +#include <linux/ftrace.h> #include <asm/uaccess.h> #include <asm/pgtable.h> @@ -548,7 +549,8 @@ __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p, * the task-switch, and shows up in ret_from_fork in entry.S, * for example. */ -struct task_struct * __switch_to(struct task_struct *prev_p, struct task_struct *next_p) +__notrace_funcgraph struct task_struct * +__switch_to(struct task_struct *prev_p, struct task_struct *next_p) { struct thread_struct *prev = &prev_p->thread, *next = &next_p->thread; diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index c958120fb1b..fbb321d53d3 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -39,6 +39,7 @@ #include <linux/prctl.h> #include <linux/uaccess.h> #include <linux/io.h> +#include <linux/ftrace.h> #include <asm/pgtable.h> #include <asm/system.h> @@ -551,8 +552,9 @@ static inline void __switch_to_xtra(struct task_struct *prev_p, * - could test fs/gs bitsliced * * Kprobes not supported here. Set the probe on schedule instead. + * Function graph tracer not supported too. */ -struct task_struct * +__notrace_funcgraph struct task_struct * __switch_to(struct task_struct *prev_p, struct task_struct *next_p) { struct thread_struct *prev = &prev_p->thread; diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 0a6d8c12e10..2c8ec1ba75e 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -668,14 +668,14 @@ static int ptrace_bts_read_record(struct task_struct *child, size_t index, size_t bts_index, bts_end; int error; - error = ds_get_bts_end(child, &bts_end); + error = ds_get_bts_end(child->bts, &bts_end); if (error < 0) return error; if (bts_end <= index) return -EINVAL; - error = ds_get_bts_index(child, &bts_index); + error = ds_get_bts_index(child->bts, &bts_index); if (error < 0) return error; @@ -684,7 +684,7 @@ static int ptrace_bts_read_record(struct task_struct *child, size_t index, if (bts_end <= bts_index) bts_index -= bts_end; - error = ds_access_bts(child, bts_index, &bts_record); + error = ds_access_bts(child->bts, bts_index, &bts_record); if (error < 0) return error; @@ -705,14 +705,14 @@ static int ptrace_bts_drain(struct task_struct *child, size_t end, i; int error; - error = ds_get_bts_index(child, &end); + error = ds_get_bts_index(child->bts, &end); if (error < 0) return error; if (size < (end * sizeof(struct bts_struct))) return -EIO; - error = ds_access_bts(child, 0, (const void **)&raw); + error = ds_access_bts(child->bts, 0, (const void **)&raw); if (error < 0) return error; @@ -723,18 +723,13 @@ static int ptrace_bts_drain(struct task_struct *child, return -EFAULT; } - error = ds_clear_bts(child); + error = ds_clear_bts(child->bts); if (error < 0) return error; return end; } -static void ptrace_bts_ovfl(struct task_struct *child) -{ - send_sig(child->thread.bts_ovfl_signal, child, 0); -} - static int ptrace_bts_config(struct task_struct *child, long cfg_size, const struct ptrace_bts_config __user *ucfg) @@ -760,23 +755,45 @@ static int ptrace_bts_config(struct task_struct *child, goto errout; if (cfg.flags & PTRACE_BTS_O_ALLOC) { - ds_ovfl_callback_t ovfl = NULL; + bts_ovfl_callback_t ovfl = NULL; unsigned int sig = 0; - /* we ignore the error in case we were not tracing child */ - (void)ds_release_bts(child); + error = -EINVAL; + if (cfg.size < (10 * bts_cfg.sizeof_bts)) + goto errout; if (cfg.flags & PTRACE_BTS_O_SIGNAL) { if (!cfg.signal) goto errout; + error = -EOPNOTSUPP; + goto errout; + sig = cfg.signal; - ovfl = ptrace_bts_ovfl; } - error = ds_request_bts(child, /* base = */ NULL, cfg.size, ovfl); - if (error < 0) + if (child->bts) { + (void)ds_release_bts(child->bts); + kfree(child->bts_buffer); + + child->bts = NULL; + child->bts_buffer = NULL; + } + + error = -ENOMEM; + child->bts_buffer = kzalloc(cfg.size, GFP_KERNEL); + if (!child->bts_buffer) + goto errout; + + child->bts = ds_request_bts(child, child->bts_buffer, cfg.size, + ovfl, /* th = */ (size_t)-1); + if (IS_ERR(child->bts)) { + error = PTR_ERR(child->bts); + kfree(child->bts_buffer); + child->bts = NULL; + child->bts_buffer = NULL; goto errout; + } child->thread.bts_ovfl_signal = sig; } @@ -823,15 +840,15 @@ static int ptrace_bts_status(struct task_struct *child, if (cfg_size < sizeof(cfg)) return -EIO; - error = ds_get_bts_end(child, &end); + error = ds_get_bts_end(child->bts, &end); if (error < 0) return error; - error = ds_access_bts(child, /* index = */ 0, &base); + error = ds_access_bts(child->bts, /* index = */ 0, &base); if (error < 0) return error; - error = ds_access_bts(child, /* index = */ end, &max); + error = ds_access_bts(child->bts, /* index = */ end, &max); if (error < 0) return error; @@ -884,10 +901,7 @@ static int ptrace_bts_write_record(struct task_struct *child, return -EINVAL; } - /* The writing task will be the switched-to task on a context - * switch. It needs to write into the switched-from task's BTS - * buffer. */ - return ds_unchecked_write_bts(child, bts_record, bts_cfg.sizeof_bts); + return ds_write_bts(child->bts, bts_record, bts_cfg.sizeof_bts); } void ptrace_bts_take_timestamp(struct task_struct *tsk, @@ -929,17 +943,16 @@ void __cpuinit ptrace_bts_init_intel(struct cpuinfo_x86 *c) switch (c->x86) { case 0x6: switch (c->x86_model) { + case 0 ... 0xC: + /* sorry, don't know about them */ + break; case 0xD: case 0xE: /* Pentium M */ bts_configure(&bts_cfg_pentium_m); break; - case 0xF: /* Core2 */ - case 0x1C: /* Atom */ + default: /* Core2, Atom, ... */ bts_configure(&bts_cfg_core2); break; - default: - /* sorry, don't know about them */ - break; } break; case 0xF: @@ -973,13 +986,17 @@ void ptrace_disable(struct task_struct *child) clear_tsk_thread_flag(child, TIF_SYSCALL_EMU); #endif #ifdef CONFIG_X86_PTRACE_BTS - (void)ds_release_bts(child); + if (child->bts) { + (void)ds_release_bts(child->bts); + kfree(child->bts_buffer); + child->bts_buffer = NULL; - child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; - if (!child->thread.debugctlmsr) - clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR); + child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; + if (!child->thread.debugctlmsr) + clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR); - clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS); + clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS); + } #endif /* CONFIG_X86_PTRACE_BTS */ } @@ -1111,9 +1128,16 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) (child, data, (struct ptrace_bts_config __user *)addr); break; - case PTRACE_BTS_SIZE: - ret = ds_get_bts_index(child, /* pos = */ NULL); + case PTRACE_BTS_SIZE: { + size_t size; + + ret = ds_get_bts_index(child->bts, &size); + if (ret == 0) { + BUG_ON(size != (int) size); + ret = (int) size; + } break; + } case PTRACE_BTS_GET: ret = ptrace_bts_read_record @@ -1121,7 +1145,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) break; case PTRACE_BTS_CLEAR: - ret = ds_clear_bts(child); + ret = ds_clear_bts(child->bts); break; case PTRACE_BTS_DRAIN: diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 724adfc63cb..ba7b9a0e606 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -21,6 +21,9 @@ # include <asm/iommu.h> #endif +#include <mach_ipi.h> + + /* * Power off function, if any */ @@ -36,7 +39,10 @@ int reboot_force; static int reboot_cpu = -1; #endif -/* reboot=b[ios] | s[mp] | t[riple] | k[bd] | e[fi] [, [w]arm | [c]old] +/* This is set by the PCI code if either type 1 or type 2 PCI is detected */ +bool port_cf9_safe = false; + +/* reboot=b[ios] | s[mp] | t[riple] | k[bd] | e[fi] [, [w]arm | [c]old] | p[ci] warm Don't set the cold reboot flag cold Set the cold reboot flag bios Reboot by jumping through the BIOS (only for X86_32) @@ -45,6 +51,7 @@ static int reboot_cpu = -1; kbd Use the keyboard controller. cold reset (default) acpi Use the RESET_REG in the FADT efi Use efi reset_system runtime service + pci Use the so-called "PCI reset register", CF9 force Avoid anything that could hang. */ static int __init reboot_setup(char *str) @@ -79,6 +86,7 @@ static int __init reboot_setup(char *str) case 'k': case 't': case 'e': + case 'p': reboot_type = *str; break; @@ -169,6 +177,15 @@ 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", @@ -395,12 +412,27 @@ static void native_machine_emergency_restart(void) reboot_type = BOOT_KBD; break; - case BOOT_EFI: if (efi_enabled) - efi.reset_system(reboot_mode ? EFI_RESET_WARM : EFI_RESET_COLD, + efi.reset_system(reboot_mode ? + EFI_RESET_WARM : + EFI_RESET_COLD, EFI_SUCCESS, 0, NULL); + reboot_type = BOOT_KBD; + break; + case BOOT_CF9: + port_cf9_safe = true; + /* fall through */ + + case BOOT_CF9_COND: + if (port_cf9_safe) { + u8 cf9 = inb(0xcf9) & ~6; + outb(cf9|2, 0xcf9); /* Request hard reset */ + udelay(50); + outb(cf9|6, 0xcf9); /* Actually do the reset */ + udelay(50); + } reboot_type = BOOT_KBD; break; } @@ -461,6 +493,11 @@ static void native_machine_restart(char *__unused) static void native_machine_halt(void) { + /* stop other cpus and apics */ + machine_shutdown(); + + /* stop this cpu */ + stop_this_cpu(NULL); } static void native_machine_power_off(void) @@ -514,3 +551,92 @@ void machine_crash_shutdown(struct pt_regs *regs) machine_ops.crash_shutdown(regs); } #endif + + +#if defined(CONFIG_SMP) + +/* This keeps a track of which one is crashing cpu. */ +static int crashing_cpu; +static nmi_shootdown_cb shootdown_callback; + +static atomic_t waiting_for_crash_ipi; + +static int crash_nmi_callback(struct notifier_block *self, + unsigned long val, void *data) +{ + int cpu; + + if (val != DIE_NMI_IPI) + return NOTIFY_OK; + + cpu = raw_smp_processor_id(); + + /* Don't do anything if this handler is invoked on crashing cpu. + * Otherwise, system will completely hang. Crashing cpu can get + * an NMI if system was initially booted with nmi_watchdog parameter. + */ + if (cpu == crashing_cpu) + return NOTIFY_STOP; + local_irq_disable(); + + shootdown_callback(cpu, (struct die_args *)data); + + atomic_dec(&waiting_for_crash_ipi); + /* Assume hlt works */ + halt(); + for (;;) + cpu_relax(); + + return 1; +} + +static void smp_send_nmi_allbutself(void) +{ + send_IPI_allbutself(NMI_VECTOR); +} + +static struct notifier_block crash_nmi_nb = { + .notifier_call = crash_nmi_callback, +}; + +/* Halt all other CPUs, calling the specified function on each of them + * + * This function can be used to halt all other CPUs on crash + * or emergency reboot time. The function passed as parameter + * will be called inside a NMI handler on all CPUs. + */ +void nmi_shootdown_cpus(nmi_shootdown_cb callback) +{ + unsigned long msecs; + local_irq_disable(); + + /* Make a note of crashing cpu. Will be used in NMI callback.*/ + crashing_cpu = safe_smp_processor_id(); + + shootdown_callback = callback; + + atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1); + /* Would it be better to replace the trap vector here? */ + if (register_die_notifier(&crash_nmi_nb)) + return; /* return what? */ + /* Ensure the new callback function is set before sending + * out the NMI + */ + wmb(); + + smp_send_nmi_allbutself(); + + msecs = 1000; /* Wait at most a second for the other cpus to stop */ + while ((atomic_read(&waiting_for_crash_ipi) > 0) && msecs) { + mdelay(1); + msecs--; + } + + /* Leave the nmi callback set */ +} +#else /* !CONFIG_SMP */ +void nmi_shootdown_cpus(nmi_shootdown_cb callback) +{ + /* No other CPUs to shoot down */ +} +#endif diff --git a/arch/x86/kernel/relocate_kernel_32.S b/arch/x86/kernel/relocate_kernel_32.S index 6f50664b2ba..a160f311972 100644 --- a/arch/x86/kernel/relocate_kernel_32.S +++ b/arch/x86/kernel/relocate_kernel_32.S @@ -10,15 +10,12 @@ #include <asm/page.h> #include <asm/kexec.h> #include <asm/processor-flags.h> -#include <asm/pgtable.h> /* * Must be relocatable PIC code callable as a C function */ #define PTR(x) (x << 2) -#define PAGE_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY) -#define PAE_PGD_ATTR (_PAGE_PRESENT) /* control_page + KEXEC_CONTROL_CODE_MAX_SIZE * ~ control_page + PAGE_SIZE are used as data storage and stack for @@ -39,7 +36,6 @@ #define CP_PA_BACKUP_PAGES_MAP DATA(0x1c) .text - .align PAGE_SIZE .globl relocate_kernel relocate_kernel: /* Save the CPU context, used for jumping back */ @@ -60,117 +56,6 @@ relocate_kernel: movl %cr4, %eax movl %eax, CR4(%edi) -#ifdef CONFIG_X86_PAE - /* map the control page at its virtual address */ - - movl PTR(VA_PGD)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0xc0000000, %eax - shrl $27, %eax - addl %edi, %eax - - movl PTR(PA_PMD_0)(%ebp), %edx - orl $PAE_PGD_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PMD_0)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0x3fe00000, %eax - shrl $18, %eax - addl %edi, %eax - - movl PTR(PA_PTE_0)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PTE_0)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0x001ff000, %eax - shrl $9, %eax - addl %edi, %eax - - movl PTR(PA_CONTROL_PAGE)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - /* identity map the control page at its physical address */ - - movl PTR(VA_PGD)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0xc0000000, %eax - shrl $27, %eax - addl %edi, %eax - - movl PTR(PA_PMD_1)(%ebp), %edx - orl $PAE_PGD_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PMD_1)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0x3fe00000, %eax - shrl $18, %eax - addl %edi, %eax - - movl PTR(PA_PTE_1)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PTE_1)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0x001ff000, %eax - shrl $9, %eax - addl %edi, %eax - - movl PTR(PA_CONTROL_PAGE)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) -#else - /* map the control page at its virtual address */ - - movl PTR(VA_PGD)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0xffc00000, %eax - shrl $20, %eax - addl %edi, %eax - - movl PTR(PA_PTE_0)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PTE_0)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0x003ff000, %eax - shrl $10, %eax - addl %edi, %eax - - movl PTR(PA_CONTROL_PAGE)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - /* identity map the control page at its physical address */ - - movl PTR(VA_PGD)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0xffc00000, %eax - shrl $20, %eax - addl %edi, %eax - - movl PTR(PA_PTE_1)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PTE_1)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0x003ff000, %eax - shrl $10, %eax - addl %edi, %eax - - movl PTR(PA_CONTROL_PAGE)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) -#endif - -relocate_new_kernel: /* read the arguments and say goodbye to the stack */ movl 20+4(%esp), %ebx /* page_list */ movl 20+8(%esp), %ebp /* list of pages */ diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 0fa6790c1dd..32fe42f26e7 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -583,7 +583,20 @@ static int __init setup_elfcorehdr(char *arg) early_param("elfcorehdr", setup_elfcorehdr); #endif -static struct x86_quirks default_x86_quirks __initdata; +static int __init default_update_genapic(void) +{ +#ifdef CONFIG_X86_SMP +# if defined(CONFIG_X86_GENERICARCH) || defined(CONFIG_X86_64) + genapic->wakeup_cpu = wakeup_secondary_cpu_via_init; +# endif +#endif + + return 0; +} + +static struct x86_quirks default_x86_quirks __initdata = { + .update_genapic = default_update_genapic, +}; struct x86_quirks *x86_quirks __initdata = &default_x86_quirks; @@ -764,7 +777,7 @@ 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 @@ -794,6 +807,9 @@ void __init setup_arch(char **cmdline_p) printk(KERN_INFO "Command line: %s\n", boot_command_line); #endif + /* VMI may relocate the fixmap; do this before touching ioremap area */ + vmi_init(); + early_cpu_init(); early_ioremap_init(); @@ -880,13 +896,8 @@ void __init setup_arch(char **cmdline_p) check_efer(); #endif -#if defined(CONFIG_VMI) && defined(CONFIG_X86_32) - /* - * Must be before kernel pagetables are setup - * or fixmap area is touched. - */ - vmi_init(); -#endif + /* Must be before kernel pagetables are setup */ + vmi_activate(); /* after early param, so could get panic from serial */ reserve_early_setup_data(); @@ -1082,7 +1093,7 @@ void __init setup_arch(char **cmdline_p) ioapic_init_mappings(); /* need to wait for io_apic is mapped */ - nr_irqs = probe_nr_irqs(); + probe_nr_irqs_gsi(); kvm_guest_init(); diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index ae0c0d3bb77..0b63b08e753 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -152,6 +152,11 @@ void __init setup_per_cpu_areas(void) old_size = PERCPU_ENOUGH_ROOM; align = max_t(unsigned long, PAGE_SIZE, align); size = roundup(old_size, align); + + printk(KERN_INFO + "NR_CPUS:%d nr_cpumask_bits:%d nr_cpu_ids:%d nr_node_ids:%d\n", + NR_CPUS, nr_cpumask_bits, nr_cpu_ids, nr_node_ids); + printk(KERN_INFO "PERCPU: Allocating %zd bytes of per cpu data\n", size); @@ -168,24 +173,24 @@ void __init setup_per_cpu_areas(void) "cpu %d has no node %d or node-local memory\n", cpu, node); if (ptr) - printk(KERN_DEBUG "per cpu data for cpu%d at %016lx\n", + printk(KERN_DEBUG + "per cpu data for cpu%d at %016lx\n", cpu, __pa(ptr)); } else { ptr = __alloc_bootmem_node(NODE_DATA(node), size, align, __pa(MAX_DMA_ADDRESS)); if (ptr) - printk(KERN_DEBUG "per cpu data for cpu%d on node%d at %016lx\n", - cpu, node, __pa(ptr)); + printk(KERN_DEBUG + "per cpu data for cpu%d on node%d " + "at %016lx\n", + cpu, node, __pa(ptr)); } #endif per_cpu_offset(cpu) = ptr - __per_cpu_start; memcpy(ptr, __per_cpu_start, __per_cpu_end - __per_cpu_start); } - printk(KERN_DEBUG "NR_CPUS: %d, nr_cpu_ids: %d, nr_node_ids %d\n", - NR_CPUS, nr_cpu_ids, nr_node_ids); - /* Setup percpu data maps */ setup_per_cpu_maps(); @@ -282,7 +287,7 @@ static void __cpuinit numa_set_cpumask(int cpu, int enable) else cpu_clear(cpu, *mask); - cpulist_scnprintf(buf, sizeof(buf), *mask); + cpulist_scnprintf(buf, sizeof(buf), mask); printk(KERN_DEBUG "%s cpu %d node %d: mask now %s\n", enable? "numa_add_cpu":"numa_remove_cpu", cpu, node, buf); } diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c index 18f9b19f5f8..49ed667b06f 100644 --- a/arch/x86/kernel/smp.c +++ b/arch/x86/kernel/smp.c @@ -118,41 +118,28 @@ static void native_smp_send_reschedule(int cpu) WARN_ON(1); return; } - send_IPI_mask(cpumask_of_cpu(cpu), RESCHEDULE_VECTOR); + send_IPI_mask(cpumask_of(cpu), RESCHEDULE_VECTOR); } void native_send_call_func_single_ipi(int cpu) { - send_IPI_mask(cpumask_of_cpu(cpu), CALL_FUNCTION_SINGLE_VECTOR); + send_IPI_mask(cpumask_of(cpu), CALL_FUNCTION_SINGLE_VECTOR); } -void native_send_call_func_ipi(cpumask_t mask) +void native_send_call_func_ipi(const struct cpumask *mask) { cpumask_t allbutself; allbutself = cpu_online_map; cpu_clear(smp_processor_id(), allbutself); - if (cpus_equal(mask, allbutself) && + if (cpus_equal(*mask, allbutself) && cpus_equal(cpu_online_map, cpu_callout_map)) send_IPI_allbutself(CALL_FUNCTION_VECTOR); else send_IPI_mask(mask, CALL_FUNCTION_VECTOR); } -static void stop_this_cpu(void *dummy) -{ - local_irq_disable(); - /* - * Remove this CPU: - */ - cpu_clear(smp_processor_id(), cpu_online_map); - disable_local_APIC(); - if (hlt_works(smp_processor_id())) - for (;;) halt(); - for (;;); -} - /* * this function calls the 'stop' function on all other CPUs in the system. */ diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 7b109339731..be946678804 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -62,6 +62,7 @@ #include <asm/mtrr.h> #include <asm/vmi.h> #include <asm/genapic.h> +#include <asm/setup.h> #include <linux/mc146818rtc.h> #include <mach_apic.h> @@ -101,14 +102,8 @@ EXPORT_SYMBOL(smp_num_siblings); /* Last level cache ID of each logical CPU */ DEFINE_PER_CPU(u16, cpu_llc_id) = BAD_APICID; -/* bitmap of online cpus */ -cpumask_t cpu_online_map __read_mostly; -EXPORT_SYMBOL(cpu_online_map); - cpumask_t cpu_callin_map; cpumask_t cpu_callout_map; -cpumask_t cpu_possible_map; -EXPORT_SYMBOL(cpu_possible_map); /* representing HT siblings of each logical CPU */ DEFINE_PER_CPU(cpumask_t, cpu_sibling_map); @@ -294,9 +289,7 @@ static void __cpuinit start_secondary(void *unused) * fragile that we want to limit the things done here to the * most necessary things. */ -#ifdef CONFIG_VMI vmi_bringup(); -#endif cpu_init(); preempt_disable(); smp_callin(); @@ -536,7 +529,7 @@ static void impress_friends(void) pr_debug("Before bogocount - setting activated=1.\n"); } -static inline void __inquire_remote_apic(int apicid) +void __inquire_remote_apic(int apicid) { unsigned i, regs[] = { APIC_ID >> 4, APIC_LVR >> 4, APIC_SPIV >> 4 }; char *names[] = { "ID", "VERSION", "SPIV" }; @@ -575,14 +568,13 @@ static inline void __inquire_remote_apic(int apicid) } } -#ifdef WAKE_SECONDARY_VIA_NMI /* * Poke the other CPU in the eye via NMI to wake it up. Remember that the normal * INIT, INIT, STARTUP sequence will reset the chip hard for us, and this * won't ... remember to clear down the APIC, etc later. */ -static int __devinit -wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip) +int __devinit +wakeup_secondary_cpu_via_nmi(int logical_apicid, unsigned long start_eip) { unsigned long send_status, accept_status = 0; int maxlvt; @@ -599,7 +591,7 @@ wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip) * Give the other CPU some time to accept the IPI. */ udelay(200); - if (APIC_INTEGRATED(apic_version[phys_apicid])) { + if (APIC_INTEGRATED(apic_version[boot_cpu_physical_apicid])) { maxlvt = lapic_get_maxlvt(); if (maxlvt > 3) /* Due to the Pentium erratum 3AP. */ apic_write(APIC_ESR, 0); @@ -614,11 +606,9 @@ wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip) return (send_status | accept_status); } -#endif /* WAKE_SECONDARY_VIA_NMI */ -#ifdef WAKE_SECONDARY_VIA_INIT -static int __devinit -wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip) +int __devinit +wakeup_secondary_cpu_via_init(int phys_apicid, unsigned long start_eip) { unsigned long send_status, accept_status = 0; int maxlvt, num_starts, j; @@ -737,7 +727,6 @@ wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip) return (send_status | accept_status); } -#endif /* WAKE_SECONDARY_VIA_INIT */ struct create_idle { struct work_struct work; @@ -1355,7 +1344,7 @@ void cpu_disable_common(void) lock_vector_lock(); remove_cpu_from_maps(cpu); unlock_vector_lock(); - fixup_irqs(cpu_online_map); + fixup_irqs(); } int native_cpu_disable(void) diff --git a/arch/x86/kernel/stacktrace.c b/arch/x86/kernel/stacktrace.c index a03e7f6d90c..10786af9554 100644 --- a/arch/x86/kernel/stacktrace.c +++ b/arch/x86/kernel/stacktrace.c @@ -6,6 +6,7 @@ #include <linux/sched.h> #include <linux/stacktrace.h> #include <linux/module.h> +#include <linux/uaccess.h> #include <asm/stacktrace.h> static void save_stack_warning(void *data, char *msg) @@ -83,3 +84,66 @@ void save_stack_trace_tsk(struct task_struct *tsk, struct stack_trace *trace) trace->entries[trace->nr_entries++] = ULONG_MAX; } EXPORT_SYMBOL_GPL(save_stack_trace_tsk); + +/* Userspace stacktrace - based on kernel/trace/trace_sysprof.c */ + +struct stack_frame { + const void __user *next_fp; + unsigned long ret_addr; +}; + +static int copy_stack_frame(const void __user *fp, struct stack_frame *frame) +{ + int ret; + + if (!access_ok(VERIFY_READ, fp, sizeof(*frame))) + return 0; + + ret = 1; + pagefault_disable(); + if (__copy_from_user_inatomic(frame, fp, sizeof(*frame))) + ret = 0; + pagefault_enable(); + + return ret; +} + +static inline void __save_stack_trace_user(struct stack_trace *trace) +{ + const struct pt_regs *regs = task_pt_regs(current); + const void __user *fp = (const void __user *)regs->bp; + + if (trace->nr_entries < trace->max_entries) + trace->entries[trace->nr_entries++] = regs->ip; + + while (trace->nr_entries < trace->max_entries) { + struct stack_frame frame; + + frame.next_fp = NULL; + frame.ret_addr = 0; + if (!copy_stack_frame(fp, &frame)) + break; + if ((unsigned long)fp < regs->sp) + break; + if (frame.ret_addr) { + trace->entries[trace->nr_entries++] = + frame.ret_addr; + } + if (fp == frame.next_fp) + break; + fp = frame.next_fp; + } +} + +void save_stack_trace_user(struct stack_trace *trace) +{ + /* + * Trace user stack if we are not a kernel thread + */ + if (current->mm) { + __save_stack_trace_user(trace); + } + if (trace->nr_entries < trace->max_entries) + trace->entries[trace->nr_entries++] = ULONG_MAX; +} + diff --git a/arch/x86/kernel/tlb_32.c b/arch/x86/kernel/tlb_32.c index f4049f3513b..174ea90d1cb 100644 --- a/arch/x86/kernel/tlb_32.c +++ b/arch/x86/kernel/tlb_32.c @@ -164,7 +164,7 @@ void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm, * We have to send the IPI only to * CPUs affected. */ - send_IPI_mask(cpumask, INVALIDATE_TLB_VECTOR); + send_IPI_mask(&cpumask, INVALIDATE_TLB_VECTOR); while (!cpus_empty(flush_cpumask)) /* nothing. lockup detection does not belong here */ diff --git a/arch/x86/kernel/tlb_64.c b/arch/x86/kernel/tlb_64.c index 8f919ca6949..de6f1bda0c5 100644 --- a/arch/x86/kernel/tlb_64.c +++ b/arch/x86/kernel/tlb_64.c @@ -191,7 +191,7 @@ void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm, * We have to send the IPI only to * CPUs affected. */ - send_IPI_mask(cpumask, INVALIDATE_TLB_VECTOR_START + sender); + send_IPI_mask(&cpumask, INVALIDATE_TLB_VECTOR_START + sender); while (!cpus_empty(f->flush_cpumask)) cpu_relax(); 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); diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c index 8b6c393ab9f..22fd6577156 100644 --- a/arch/x86/kernel/vmi_32.c +++ b/arch/x86/kernel/vmi_32.c @@ -960,8 +960,6 @@ static inline int __init activate_vmi(void) void __init vmi_init(void) { - unsigned long flags; - if (!vmi_rom) probe_vmi_rom(); else @@ -973,13 +971,21 @@ void __init vmi_init(void) reserve_top_address(-vmi_rom->virtual_top); - local_irq_save(flags); - activate_vmi(); - #ifdef CONFIG_X86_IO_APIC /* This is virtual hardware; timer routing is wired correctly */ no_timer_check = 1; #endif +} + +void vmi_activate(void) +{ + unsigned long flags; + + if (!vmi_rom) + return; + + local_irq_save(flags); + activate_vmi(); local_irq_restore(flags & X86_EFLAGS_IF); } diff --git a/arch/x86/kernel/vmiclock_32.c b/arch/x86/kernel/vmiclock_32.c index 254ee07f863..c4c1f9e0940 100644 --- a/arch/x86/kernel/vmiclock_32.c +++ b/arch/x86/kernel/vmiclock_32.c @@ -226,7 +226,7 @@ static void __devinit vmi_time_init_clockevent(void) /* Upper bound is clockevent's use of ulong for cycle deltas. */ evt->max_delta_ns = clockevent_delta2ns(ULONG_MAX, evt); evt->min_delta_ns = clockevent_delta2ns(1, evt); - evt->cpumask = cpumask_of_cpu(cpu); + evt->cpumask = cpumask_of(cpu); printk(KERN_WARNING "vmi: registering clock event %s. mult=%lu shift=%u\n", evt->name, evt->mult, evt->shift); diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c index 0b8b6690a86..6f3d3d4cd97 100644 --- a/arch/x86/kernel/vsyscall_64.c +++ b/arch/x86/kernel/vsyscall_64.c @@ -17,6 +17,9 @@ * want per guest time just set the kernel.vsyscall64 sysctl to 0. */ +/* Disable profiling for userspace code: */ +#define DISABLE_BRANCH_PROFILING + #include <linux/time.h> #include <linux/init.h> #include <linux/kernel.h> 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; |