diff options
Diffstat (limited to 'arch/x86/kernel')
105 files changed, 6574 insertions, 5178 deletions
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index b62a7667828..d364df03c1d 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -12,6 +12,7 @@ CFLAGS_REMOVE_tsc.o = -pg CFLAGS_REMOVE_rtc.o = -pg CFLAGS_REMOVE_paravirt-spinlocks.o = -pg CFLAGS_REMOVE_ftrace.o = -pg +CFLAGS_REMOVE_early_printk.o = -pg endif # @@ -23,9 +24,9 @@ CFLAGS_vsyscall_64.o := $(PROFILING) -g0 $(nostackp) CFLAGS_hpet.o := $(nostackp) CFLAGS_tsc.o := $(nostackp) -obj-y := process_$(BITS).o signal_$(BITS).o entry_$(BITS).o +obj-y := process_$(BITS).o signal.o entry_$(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-$(CONFIG_X86_32) += probe_roms_32.o @@ -65,6 +66,7 @@ obj-$(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 @@ -105,6 +107,10 @@ microcode-$(CONFIG_MICROCODE_INTEL) += microcode_intel.o microcode-$(CONFIG_MICROCODE_AMD) += microcode_amd.o obj-$(CONFIG_MICROCODE) += microcode.o +obj-$(CONFIG_X86_CHECK_BIOS_CORRUPTION) += check.o + +obj-$(CONFIG_SWIOTLB) += pci-swiotlb_64.o # NB rename without _64 + ### # 64 bit specific files ifeq ($(CONFIG_X86_64),y) @@ -118,7 +124,6 @@ ifeq ($(CONFIG_X86_64),y) obj-$(CONFIG_GART_IOMMU) += pci-gart_64.o aperture_64.o obj-$(CONFIG_CALGARY_IOMMU) += pci-calgary_64.o tce_64.o obj-$(CONFIG_AMD_IOMMU) += amd_iommu_init.o amd_iommu.o - obj-$(CONFIG_SWIOTLB) += pci-swiotlb_64.o obj-$(CONFIG_PCI_MMCONFIG) += mmconf-fam10h_64.o endif diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 4c51a2f8fd3..29dc0c89d4a 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -538,9 +538,10 @@ static int __cpuinit _acpi_map_lsapic(acpi_handle handle, int *pcpu) struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL }; union acpi_object *obj; struct acpi_madt_local_apic *lapic; - cpumask_t tmp_map, new_map; + cpumask_var_t tmp_map, new_map; u8 physid; int cpu; + int retval = -ENOMEM; if (ACPI_FAILURE(acpi_evaluate_object(handle, "_MAT", NULL, &buffer))) return -EINVAL; @@ -569,23 +570,37 @@ static int __cpuinit _acpi_map_lsapic(acpi_handle handle, int *pcpu) buffer.length = ACPI_ALLOCATE_BUFFER; buffer.pointer = NULL; - tmp_map = cpu_present_map; + if (!alloc_cpumask_var(&tmp_map, GFP_KERNEL)) + goto out; + + if (!alloc_cpumask_var(&new_map, GFP_KERNEL)) + goto free_tmp_map; + + cpumask_copy(tmp_map, cpu_present_mask); acpi_register_lapic(physid, lapic->lapic_flags & ACPI_MADT_ENABLED); /* * If mp_register_lapic successfully generates a new logical cpu * number, then the following will get us exactly what was mapped */ - cpus_andnot(new_map, cpu_present_map, tmp_map); - if (cpus_empty(new_map)) { + cpumask_andnot(new_map, cpu_present_mask, tmp_map); + if (cpumask_empty(new_map)) { printk ("Unable to map lapic to logical cpu number\n"); - return -EINVAL; + retval = -EINVAL; + goto free_new_map; } - cpu = first_cpu(new_map); + cpu = cpumask_first(new_map); *pcpu = cpu; - return 0; + retval = 0; + +free_new_map: + free_cpumask_var(new_map); +free_tmp_map: + free_cpumask_var(tmp_map); +out: + return retval; } /* wrapper to silence section mismatch warning */ @@ -598,7 +613,7 @@ EXPORT_SYMBOL(acpi_map_lsapic); int acpi_unmap_lsapic(int cpu) { per_cpu(x86_cpu_to_apicid, cpu) = -1; - cpu_clear(cpu, cpu_present_map); + set_cpu_present(cpu, false); num_processors--; return (0); @@ -1360,6 +1375,17 @@ 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 a7b6dec6fc3..5113c080f0c 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -20,10 +20,15 @@ #include <linux/pci.h> #include <linux/gfp.h> #include <linux/bitops.h> +#include <linux/debugfs.h> #include <linux/scatterlist.h> #include <linux/iommu-helper.h> +#ifdef CONFIG_IOMMU_API +#include <linux/iommu.h> +#endif #include <asm/proto.h> #include <asm/iommu.h> +#include <asm/gart.h> #include <asm/amd_iommu_types.h> #include <asm/amd_iommu.h> @@ -37,6 +42,10 @@ static DEFINE_RWLOCK(amd_iommu_devtable_lock); static LIST_HEAD(iommu_pd_list); static DEFINE_SPINLOCK(iommu_pd_list_lock); +#ifdef CONFIG_IOMMU_API +static struct iommu_ops amd_iommu_ops; +#endif + /* * general struct to manage commands send to an IOMMU */ @@ -46,6 +55,68 @@ struct iommu_cmd { static int dma_ops_unity_map(struct dma_ops_domain *dma_dom, struct unity_map_entry *e); +static struct dma_ops_domain *find_protection_domain(u16 devid); + + +#ifdef CONFIG_AMD_IOMMU_STATS + +/* + * Initialization code for statistics collection + */ + +DECLARE_STATS_COUNTER(compl_wait); +DECLARE_STATS_COUNTER(cnt_map_single); +DECLARE_STATS_COUNTER(cnt_unmap_single); +DECLARE_STATS_COUNTER(cnt_map_sg); +DECLARE_STATS_COUNTER(cnt_unmap_sg); +DECLARE_STATS_COUNTER(cnt_alloc_coherent); +DECLARE_STATS_COUNTER(cnt_free_coherent); +DECLARE_STATS_COUNTER(cross_page); +DECLARE_STATS_COUNTER(domain_flush_single); +DECLARE_STATS_COUNTER(domain_flush_all); +DECLARE_STATS_COUNTER(alloced_io_mem); +DECLARE_STATS_COUNTER(total_map_requests); + +static struct dentry *stats_dir; +static struct dentry *de_isolate; +static struct dentry *de_fflush; + +static void amd_iommu_stats_add(struct __iommu_counter *cnt) +{ + if (stats_dir == NULL) + return; + + cnt->dent = debugfs_create_u64(cnt->name, 0444, stats_dir, + &cnt->value); +} + +static void amd_iommu_stats_init(void) +{ + stats_dir = debugfs_create_dir("amd-iommu", NULL); + if (stats_dir == NULL) + return; + + de_isolate = debugfs_create_bool("isolation", 0444, stats_dir, + (u32 *)&amd_iommu_isolate); + + de_fflush = debugfs_create_bool("fullflush", 0444, stats_dir, + (u32 *)&amd_iommu_unmap_flush); + + amd_iommu_stats_add(&compl_wait); + amd_iommu_stats_add(&cnt_map_single); + amd_iommu_stats_add(&cnt_unmap_single); + amd_iommu_stats_add(&cnt_map_sg); + amd_iommu_stats_add(&cnt_unmap_sg); + amd_iommu_stats_add(&cnt_alloc_coherent); + amd_iommu_stats_add(&cnt_free_coherent); + amd_iommu_stats_add(&cross_page); + amd_iommu_stats_add(&domain_flush_single); + amd_iommu_stats_add(&domain_flush_all); + amd_iommu_stats_add(&alloced_io_mem); + amd_iommu_stats_add(&total_map_requests); +} + +#endif /* returns !0 if the IOMMU is caching non-present entries in its TLB */ static int iommu_has_npcache(struct amd_iommu *iommu) @@ -188,13 +259,55 @@ 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; + iommu->need_sync = true; spin_unlock_irqrestore(&iommu->lock, flags); return ret; } /* + * This function waits until an IOMMU has completed a completion + * wait command + */ +static void __iommu_wait_for_completion(struct amd_iommu *iommu) +{ + int ready = 0; + unsigned status = 0; + unsigned long i = 0; + + INC_STATS_COUNTER(compl_wait); + + while (!ready && (i < EXIT_LOOP_COUNT)) { + ++i; + /* wait for the bit to become one */ + status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); + ready = status & MMIO_STATUS_COM_WAIT_INT_MASK; + } + + /* set bit back to zero */ + status &= ~MMIO_STATUS_COM_WAIT_INT_MASK; + writel(status, iommu->mmio_base + MMIO_STATUS_OFFSET); + + if (unlikely(i == EXIT_LOOP_COUNT)) + panic("AMD IOMMU: Completion wait loop failed\n"); +} + +/* + * This function queues a completion wait command into the command + * buffer of an IOMMU + */ +static int __iommu_completion_wait(struct amd_iommu *iommu) +{ + struct iommu_cmd cmd; + + memset(&cmd, 0, sizeof(cmd)); + cmd.data[0] = CMD_COMPL_WAIT_INT_MASK; + CMD_SET_TYPE(&cmd, CMD_COMPL_WAIT); + + return __iommu_queue_command(iommu, &cmd); +} + +/* * This function is called whenever we need to ensure that the IOMMU has * completed execution of all commands we sent. It sends a * COMPLETION_WAIT command and waits for it to finish. The IOMMU informs @@ -203,40 +316,23 @@ static int iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd) */ static int iommu_completion_wait(struct amd_iommu *iommu) { - int ret = 0, ready = 0; - unsigned status = 0; - struct iommu_cmd cmd; - unsigned long flags, i = 0; - - memset(&cmd, 0, sizeof(cmd)); - cmd.data[0] = CMD_COMPL_WAIT_INT_MASK; - CMD_SET_TYPE(&cmd, CMD_COMPL_WAIT); + int ret = 0; + unsigned long flags; spin_lock_irqsave(&iommu->lock, flags); if (!iommu->need_sync) goto out; - iommu->need_sync = 0; + ret = __iommu_completion_wait(iommu); - ret = __iommu_queue_command(iommu, &cmd); + iommu->need_sync = false; if (ret) goto out; - while (!ready && (i < EXIT_LOOP_COUNT)) { - ++i; - /* wait for the bit to become one */ - status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); - ready = status & MMIO_STATUS_COM_WAIT_INT_MASK; - } - - /* set bit back to zero */ - status &= ~MMIO_STATUS_COM_WAIT_INT_MASK; - writel(status, iommu->mmio_base + MMIO_STATUS_OFFSET); + __iommu_wait_for_completion(iommu); - if (unlikely((i == EXIT_LOOP_COUNT) && printk_ratelimit())) - printk(KERN_WARNING "AMD IOMMU: Completion wait loop failed\n"); out: spin_unlock_irqrestore(&iommu->lock, flags); @@ -262,6 +358,21 @@ static int iommu_queue_inv_dev_entry(struct amd_iommu *iommu, u16 devid) return ret; } +static void __iommu_build_inv_iommu_pages(struct iommu_cmd *cmd, u64 address, + u16 domid, int pde, int s) +{ + memset(cmd, 0, sizeof(*cmd)); + address &= PAGE_MASK; + CMD_SET_TYPE(cmd, CMD_INV_IOMMU_PAGES); + cmd->data[1] |= domid; + cmd->data[2] = lower_32_bits(address); + cmd->data[3] = upper_32_bits(address); + if (s) /* size bit - we flush more than one 4kb page */ + cmd->data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK; + if (pde) /* PDE bit - we wan't flush everything not only the PTEs */ + cmd->data[2] |= CMD_INV_IOMMU_PAGES_PDE_MASK; +} + /* * Generic command send function for invalidaing TLB entries */ @@ -271,16 +382,7 @@ static int iommu_queue_inv_iommu_pages(struct amd_iommu *iommu, struct iommu_cmd cmd; int ret; - memset(&cmd, 0, sizeof(cmd)); - address &= PAGE_MASK; - CMD_SET_TYPE(&cmd, CMD_INV_IOMMU_PAGES); - cmd.data[1] |= domid; - cmd.data[2] = lower_32_bits(address); - cmd.data[3] = upper_32_bits(address); - if (s) /* size bit - we flush more than one 4kb page */ - cmd.data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK; - if (pde) /* PDE bit - we wan't flush everything not only the PTEs */ - cmd.data[2] |= CMD_INV_IOMMU_PAGES_PDE_MASK; + __iommu_build_inv_iommu_pages(&cmd, address, domid, pde, s); ret = iommu_queue_command(iommu, &cmd); @@ -319,9 +421,35 @@ static void iommu_flush_tlb(struct amd_iommu *iommu, u16 domid) { u64 address = CMD_INV_IOMMU_ALL_PAGES_ADDRESS; + INC_STATS_COUNTER(domain_flush_single); + iommu_queue_inv_iommu_pages(iommu, address, domid, 0, 1); } +/* + * This function is used to flush the IO/TLB for a given protection domain + * on every IOMMU in the system + */ +static void iommu_flush_domain(u16 domid) +{ + unsigned long flags; + struct amd_iommu *iommu; + struct iommu_cmd cmd; + + INC_STATS_COUNTER(domain_flush_all); + + __iommu_build_inv_iommu_pages(&cmd, CMD_INV_IOMMU_ALL_PAGES_ADDRESS, + domid, 1, 1); + + list_for_each_entry(iommu, &amd_iommu_list, list) { + spin_lock_irqsave(&iommu->lock, flags); + __iommu_queue_command(iommu, &cmd); + __iommu_completion_wait(iommu); + __iommu_wait_for_completion(iommu); + spin_unlock_irqrestore(&iommu->lock, flags); + } +} + /**************************************************************************** * * The functions below are used the create the page table mappings for @@ -336,10 +464,10 @@ static void iommu_flush_tlb(struct amd_iommu *iommu, u16 domid) * supporting all features of AMD IOMMU page tables like level skipping * and full 64 bit address spaces. */ -static int iommu_map(struct protection_domain *dom, - unsigned long bus_addr, - unsigned long phys_addr, - int prot) +static int iommu_map_page(struct protection_domain *dom, + unsigned long bus_addr, + unsigned long phys_addr, + int prot) { u64 __pte, *pte, *page; @@ -386,6 +514,28 @@ static int iommu_map(struct protection_domain *dom, return 0; } +static void iommu_unmap_page(struct protection_domain *dom, + unsigned long bus_addr) +{ + u64 *pte; + + pte = &dom->pt_root[IOMMU_PTE_L2_INDEX(bus_addr)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return; + + pte = IOMMU_PTE_PAGE(*pte); + pte = &pte[IOMMU_PTE_L1_INDEX(bus_addr)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return; + + pte = IOMMU_PTE_PAGE(*pte); + pte = &pte[IOMMU_PTE_L1_INDEX(bus_addr)]; + + *pte = 0; +} + /* * This function checks if a specific unity mapping entry is needed for * this specific IOMMU. @@ -438,7 +588,7 @@ static int dma_ops_unity_map(struct dma_ops_domain *dma_dom, for (addr = e->address_start; addr < e->address_end; addr += PAGE_SIZE) { - ret = iommu_map(&dma_dom->domain, addr, addr, e->prot); + ret = iommu_map_page(&dma_dom->domain, addr, addr, e->prot); if (ret) return ret; /* @@ -569,6 +719,16 @@ static u16 domain_id_alloc(void) return id; } +static void domain_id_free(int id) +{ + unsigned long flags; + + write_lock_irqsave(&amd_iommu_devtable_lock, flags); + if (id > 0 && id < MAX_DOMAIN_ID) + __clear_bit(id, amd_iommu_pd_alloc_bitmap); + write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); +} + /* * Used to reserve address ranges in the aperture (e.g. for exclusion * ranges. @@ -585,12 +745,12 @@ static void dma_ops_reserve_addresses(struct dma_ops_domain *dom, iommu_area_reserve(dom->bitmap, start_page, pages); } -static void dma_ops_free_pagetable(struct dma_ops_domain *dma_dom) +static void free_pagetable(struct protection_domain *domain) { int i, j; u64 *p1, *p2, *p3; - p1 = dma_dom->domain.pt_root; + p1 = domain->pt_root; if (!p1) return; @@ -611,6 +771,8 @@ static void dma_ops_free_pagetable(struct dma_ops_domain *dma_dom) } free_page((unsigned long)p1); + + domain->pt_root = NULL; } /* @@ -622,7 +784,7 @@ static void dma_ops_domain_free(struct dma_ops_domain *dom) if (!dom) return; - dma_ops_free_pagetable(dom); + free_pagetable(&dom->domain); kfree(dom->pte_pages); @@ -661,6 +823,7 @@ static struct dma_ops_domain *dma_ops_domain_alloc(struct amd_iommu *iommu, goto free_dma_dom; dma_dom->domain.mode = PAGE_MODE_3_LEVEL; dma_dom->domain.pt_root = (void *)get_zeroed_page(GFP_KERNEL); + dma_dom->domain.flags = PD_DMA_OPS_MASK; dma_dom->domain.priv = dma_dom; if (!dma_dom->domain.pt_root) goto free_dma_dom; @@ -723,6 +886,15 @@ free_dma_dom: } /* + * little helper function to check whether a given protection domain is a + * dma_ops domain + */ +static bool dma_ops_domain(struct protection_domain *domain) +{ + return domain->flags & PD_DMA_OPS_MASK; +} + +/* * Find out the protection domain structure for a given PCI device. This * will give us the pointer to the page table root for example. */ @@ -742,14 +914,15 @@ static struct protection_domain *domain_for_device(u16 devid) * If a device is not yet associated with a domain, this function does * assigns it visible for the hardware */ -static void set_device_domain(struct amd_iommu *iommu, - struct protection_domain *domain, - u16 devid) +static void attach_device(struct amd_iommu *iommu, + struct protection_domain *domain, + u16 devid) { unsigned long flags; - u64 pte_root = virt_to_phys(domain->pt_root); + domain->dev_cnt += 1; + pte_root |= (domain->mode & DEV_ENTRY_MODE_MASK) << DEV_ENTRY_MODE_SHIFT; pte_root |= IOMMU_PTE_IR | IOMMU_PTE_IW | IOMMU_PTE_P | IOMMU_PTE_TV; @@ -765,6 +938,116 @@ static void set_device_domain(struct amd_iommu *iommu, iommu_queue_inv_dev_entry(iommu, devid); } +/* + * Removes a device from a protection domain (unlocked) + */ +static void __detach_device(struct protection_domain *domain, u16 devid) +{ + + /* lock domain */ + spin_lock(&domain->lock); + + /* remove domain from the lookup table */ + amd_iommu_pd_table[devid] = NULL; + + /* remove entry from the device table seen by the hardware */ + amd_iommu_dev_table[devid].data[0] = IOMMU_PTE_P | IOMMU_PTE_TV; + amd_iommu_dev_table[devid].data[1] = 0; + amd_iommu_dev_table[devid].data[2] = 0; + + /* decrease reference counter */ + domain->dev_cnt -= 1; + + /* ready */ + spin_unlock(&domain->lock); +} + +/* + * Removes a device from a protection domain (with devtable_lock held) + */ +static void detach_device(struct protection_domain *domain, u16 devid) +{ + unsigned long flags; + + /* lock device table */ + write_lock_irqsave(&amd_iommu_devtable_lock, flags); + __detach_device(domain, devid); + write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); +} + +static int device_change_notifier(struct notifier_block *nb, + unsigned long action, void *data) +{ + struct device *dev = data; + struct pci_dev *pdev = to_pci_dev(dev); + u16 devid = calc_devid(pdev->bus->number, pdev->devfn); + struct protection_domain *domain; + struct dma_ops_domain *dma_domain; + struct amd_iommu *iommu; + int order = amd_iommu_aperture_order; + unsigned long flags; + + if (devid > amd_iommu_last_bdf) + goto out; + + devid = amd_iommu_alias_table[devid]; + + iommu = amd_iommu_rlookup_table[devid]; + if (iommu == NULL) + goto out; + + domain = domain_for_device(devid); + + if (domain && !dma_ops_domain(domain)) + WARN_ONCE(1, "AMD IOMMU WARNING: device %s already bound " + "to a non-dma-ops domain\n", dev_name(dev)); + + switch (action) { + case BUS_NOTIFY_BOUND_DRIVER: + if (domain) + goto out; + dma_domain = find_protection_domain(devid); + if (!dma_domain) + dma_domain = iommu->default_dom; + attach_device(iommu, &dma_domain->domain, devid); + printk(KERN_INFO "AMD IOMMU: Using protection domain %d for " + "device %s\n", dma_domain->domain.id, dev_name(dev)); + break; + case BUS_NOTIFY_UNBIND_DRIVER: + if (!domain) + goto out; + detach_device(domain, devid); + break; + case BUS_NOTIFY_ADD_DEVICE: + /* allocate a protection domain if a device is added */ + dma_domain = find_protection_domain(devid); + if (dma_domain) + goto out; + dma_domain = dma_ops_domain_alloc(iommu, order); + if (!dma_domain) + goto out; + dma_domain->target_dev = devid; + + spin_lock_irqsave(&iommu_pd_list_lock, flags); + list_add_tail(&dma_domain->list, &iommu_pd_list); + spin_unlock_irqrestore(&iommu_pd_list_lock, flags); + + break; + default: + goto out; + } + + iommu_queue_inv_dev_entry(iommu, devid); + iommu_completion_wait(iommu); + +out: + return 0; +} + +struct notifier_block device_nb = { + .notifier_call = device_change_notifier, +}; + /***************************************************************************** * * The next functions belong to the dma_ops mapping/unmapping code. @@ -800,7 +1083,6 @@ static struct dma_ops_domain *find_protection_domain(u16 devid) list_for_each_entry(entry, &iommu_pd_list, list) { if (entry->target_dev == devid) { ret = entry; - list_del(&ret->list); break; } } @@ -851,14 +1133,13 @@ static int get_device_resources(struct device *dev, if (!dma_dom) dma_dom = (*iommu)->default_dom; *domain = &dma_dom->domain; - set_device_domain(*iommu, *domain, *bdf); + attach_device(*iommu, *domain, *bdf); printk(KERN_INFO "AMD IOMMU: Using protection domain %d for " - "device ", (*domain)->id); - print_devid(_bdf, 1); + "device %s\n", (*domain)->id, dev_name(dev)); } if (domain_for_device(_bdf) == NULL) - set_device_domain(*iommu, *domain, _bdf); + attach_device(*iommu, *domain, _bdf); return 1; } @@ -944,6 +1225,11 @@ static dma_addr_t __map_single(struct device *dev, pages = iommu_num_pages(paddr, size, PAGE_SIZE); paddr &= PAGE_MASK; + INC_STATS_COUNTER(total_map_requests); + + if (pages > 1) + INC_STATS_COUNTER(cross_page); + if (align) align_mask = (1UL << get_order(size)) - 1; @@ -960,6 +1246,8 @@ static dma_addr_t __map_single(struct device *dev, } address += offset; + ADD_STATS_COUNTER(alloced_io_mem, size); + if (unlikely(dma_dom->need_flush && !amd_iommu_unmap_flush)) { iommu_flush_tlb(iommu, dma_dom->domain.id); dma_dom->need_flush = false; @@ -996,6 +1284,8 @@ static void __unmap_single(struct amd_iommu *iommu, start += PAGE_SIZE; } + SUB_STATS_COUNTER(alloced_io_mem, size); + dma_ops_free_addresses(dma_dom, dma_addr, pages); if (amd_iommu_unmap_flush || dma_dom->need_flush) { @@ -1017,6 +1307,8 @@ static dma_addr_t map_single(struct device *dev, phys_addr_t paddr, dma_addr_t addr; u64 dma_mask; + INC_STATS_COUNTER(cnt_map_single); + if (!check_device(dev)) return bad_dma_address; @@ -1028,6 +1320,9 @@ static dma_addr_t map_single(struct device *dev, phys_addr_t paddr, /* device not handled by any AMD IOMMU */ return (dma_addr_t)paddr; + if (!dma_ops_domain(domain)) + return bad_dma_address; + spin_lock_irqsave(&domain->lock, flags); addr = __map_single(dev, iommu, domain->priv, paddr, size, dir, false, dma_mask); @@ -1053,11 +1348,16 @@ static void unmap_single(struct device *dev, dma_addr_t dma_addr, struct protection_domain *domain; u16 devid; + INC_STATS_COUNTER(cnt_unmap_single); + if (!check_device(dev) || !get_device_resources(dev, &iommu, &domain, &devid)) /* device not handled by any AMD IOMMU */ return; + if (!dma_ops_domain(domain)) + return; + spin_lock_irqsave(&domain->lock, flags); __unmap_single(iommu, domain->priv, dma_addr, size, dir); @@ -1102,6 +1402,8 @@ static int map_sg(struct device *dev, struct scatterlist *sglist, int mapped_elems = 0; u64 dma_mask; + INC_STATS_COUNTER(cnt_map_sg); + if (!check_device(dev)) return 0; @@ -1112,6 +1414,9 @@ static int map_sg(struct device *dev, struct scatterlist *sglist, if (!iommu || !domain) return map_sg_no_iommu(dev, sglist, nelems, dir); + if (!dma_ops_domain(domain)) + return 0; + spin_lock_irqsave(&domain->lock, flags); for_each_sg(sglist, s, nelems, i) { @@ -1161,10 +1466,15 @@ static void unmap_sg(struct device *dev, struct scatterlist *sglist, u16 devid; int i; + INC_STATS_COUNTER(cnt_unmap_sg); + if (!check_device(dev) || !get_device_resources(dev, &iommu, &domain, &devid)) return; + if (!dma_ops_domain(domain)) + return; + spin_lock_irqsave(&domain->lock, flags); for_each_sg(sglist, s, nelems, i) { @@ -1192,6 +1502,8 @@ static void *alloc_coherent(struct device *dev, size_t size, phys_addr_t paddr; u64 dma_mask = dev->coherent_dma_mask; + INC_STATS_COUNTER(cnt_alloc_coherent); + if (!check_device(dev)) return NULL; @@ -1210,6 +1522,9 @@ static void *alloc_coherent(struct device *dev, size_t size, return virt_addr; } + if (!dma_ops_domain(domain)) + goto out_free; + if (!dma_mask) dma_mask = *dev->dma_mask; @@ -1218,18 +1533,20 @@ static void *alloc_coherent(struct device *dev, size_t size, *dma_addr = __map_single(dev, iommu, domain->priv, paddr, size, DMA_BIDIRECTIONAL, true, dma_mask); - if (*dma_addr == bad_dma_address) { - free_pages((unsigned long)virt_addr, get_order(size)); - virt_addr = NULL; - goto out; - } + if (*dma_addr == bad_dma_address) + goto out_free; iommu_completion_wait(iommu); -out: spin_unlock_irqrestore(&domain->lock, flags); return virt_addr; + +out_free: + + free_pages((unsigned long)virt_addr, get_order(size)); + + return NULL; } /* @@ -1243,6 +1560,8 @@ static void free_coherent(struct device *dev, size_t size, struct protection_domain *domain; u16 devid; + INC_STATS_COUNTER(cnt_free_coherent); + if (!check_device(dev)) return; @@ -1251,6 +1570,9 @@ static void free_coherent(struct device *dev, size_t size, if (!iommu || !domain) goto free_mem; + if (!dma_ops_domain(domain)) + goto free_mem; + spin_lock_irqsave(&domain->lock, flags); __unmap_single(iommu, domain->priv, dma_addr, size, DMA_BIDIRECTIONAL); @@ -1294,7 +1616,7 @@ static int amd_iommu_dma_supported(struct device *dev, u64 mask) * we don't need to preallocate the protection domains anymore. * For now we have to. */ -void prealloc_protection_domains(void) +static void prealloc_protection_domains(void) { struct pci_dev *dev = NULL; struct dma_ops_domain *dma_dom; @@ -1303,7 +1625,7 @@ void prealloc_protection_domains(void) u16 devid; while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) { - devid = (dev->bus->number << 8) | dev->devfn; + devid = calc_devid(dev->bus->number, dev->devfn); if (devid > amd_iommu_last_bdf) continue; devid = amd_iommu_alias_table[devid]; @@ -1350,6 +1672,7 @@ int __init amd_iommu_init_dma_ops(void) iommu->default_dom = dma_ops_domain_alloc(iommu, order); if (iommu->default_dom == NULL) return -ENOMEM; + iommu->default_dom->domain.flags |= PD_DEFAULT_MASK; ret = iommu_init_unity_mappings(iommu); if (ret) goto free_domains; @@ -1373,6 +1696,12 @@ int __init amd_iommu_init_dma_ops(void) /* Make the driver finally visible to the drivers */ dma_ops = &amd_iommu_dma_ops; + register_iommu(&amd_iommu_ops); + + bus_register_notifier(&pci_bus_type, &device_nb); + + amd_iommu_stats_init(); + return 0; free_domains: @@ -1384,3 +1713,224 @@ free_domains: return ret; } + +/***************************************************************************** + * + * The following functions belong to the exported interface of AMD IOMMU + * + * This interface allows access to lower level functions of the IOMMU + * like protection domain handling and assignement of devices to domains + * which is not possible with the dma_ops interface. + * + *****************************************************************************/ + +static void cleanup_domain(struct protection_domain *domain) +{ + unsigned long flags; + u16 devid; + + write_lock_irqsave(&amd_iommu_devtable_lock, flags); + + for (devid = 0; devid <= amd_iommu_last_bdf; ++devid) + if (amd_iommu_pd_table[devid] == domain) + __detach_device(domain, devid); + + write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); +} + +static int amd_iommu_domain_init(struct iommu_domain *dom) +{ + struct protection_domain *domain; + + domain = kzalloc(sizeof(*domain), GFP_KERNEL); + if (!domain) + return -ENOMEM; + + spin_lock_init(&domain->lock); + domain->mode = PAGE_MODE_3_LEVEL; + domain->id = domain_id_alloc(); + if (!domain->id) + goto out_free; + domain->pt_root = (void *)get_zeroed_page(GFP_KERNEL); + if (!domain->pt_root) + goto out_free; + + dom->priv = domain; + + return 0; + +out_free: + kfree(domain); + + return -ENOMEM; +} + +static void amd_iommu_domain_destroy(struct iommu_domain *dom) +{ + struct protection_domain *domain = dom->priv; + + if (!domain) + return; + + if (domain->dev_cnt > 0) + cleanup_domain(domain); + + BUG_ON(domain->dev_cnt != 0); + + free_pagetable(domain); + + domain_id_free(domain->id); + + kfree(domain); + + dom->priv = NULL; +} + +static void amd_iommu_detach_device(struct iommu_domain *dom, + struct device *dev) +{ + struct protection_domain *domain = dom->priv; + struct amd_iommu *iommu; + struct pci_dev *pdev; + u16 devid; + + if (dev->bus != &pci_bus_type) + return; + + pdev = to_pci_dev(dev); + + devid = calc_devid(pdev->bus->number, pdev->devfn); + + if (devid > 0) + detach_device(domain, devid); + + iommu = amd_iommu_rlookup_table[devid]; + if (!iommu) + return; + + iommu_queue_inv_dev_entry(iommu, devid); + iommu_completion_wait(iommu); +} + +static int amd_iommu_attach_device(struct iommu_domain *dom, + struct device *dev) +{ + struct protection_domain *domain = dom->priv; + struct protection_domain *old_domain; + struct amd_iommu *iommu; + struct pci_dev *pdev; + u16 devid; + + if (dev->bus != &pci_bus_type) + return -EINVAL; + + pdev = to_pci_dev(dev); + + devid = calc_devid(pdev->bus->number, pdev->devfn); + + if (devid >= amd_iommu_last_bdf || + devid != amd_iommu_alias_table[devid]) + return -EINVAL; + + iommu = amd_iommu_rlookup_table[devid]; + if (!iommu) + return -EINVAL; + + old_domain = domain_for_device(devid); + if (old_domain) + return -EBUSY; + + attach_device(iommu, domain, devid); + + iommu_completion_wait(iommu); + + return 0; +} + +static int amd_iommu_map_range(struct iommu_domain *dom, + unsigned long iova, phys_addr_t paddr, + size_t size, int iommu_prot) +{ + struct protection_domain *domain = dom->priv; + unsigned long i, npages = iommu_num_pages(paddr, size, PAGE_SIZE); + int prot = 0; + int ret; + + if (iommu_prot & IOMMU_READ) + prot |= IOMMU_PROT_IR; + if (iommu_prot & IOMMU_WRITE) + prot |= IOMMU_PROT_IW; + + iova &= PAGE_MASK; + paddr &= PAGE_MASK; + + for (i = 0; i < npages; ++i) { + ret = iommu_map_page(domain, iova, paddr, prot); + if (ret) + return ret; + + iova += PAGE_SIZE; + paddr += PAGE_SIZE; + } + + return 0; +} + +static void amd_iommu_unmap_range(struct iommu_domain *dom, + unsigned long iova, size_t size) +{ + + struct protection_domain *domain = dom->priv; + unsigned long i, npages = iommu_num_pages(iova, size, PAGE_SIZE); + + iova &= PAGE_MASK; + + for (i = 0; i < npages; ++i) { + iommu_unmap_page(domain, iova); + iova += PAGE_SIZE; + } + + iommu_flush_domain(domain->id); +} + +static phys_addr_t amd_iommu_iova_to_phys(struct iommu_domain *dom, + unsigned long iova) +{ + struct protection_domain *domain = dom->priv; + unsigned long offset = iova & ~PAGE_MASK; + phys_addr_t paddr; + u64 *pte; + + pte = &domain->pt_root[IOMMU_PTE_L2_INDEX(iova)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return 0; + + pte = IOMMU_PTE_PAGE(*pte); + pte = &pte[IOMMU_PTE_L1_INDEX(iova)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return 0; + + pte = IOMMU_PTE_PAGE(*pte); + pte = &pte[IOMMU_PTE_L0_INDEX(iova)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return 0; + + paddr = *pte & IOMMU_PAGE_MASK; + paddr |= offset; + + return paddr; +} + +static struct iommu_ops amd_iommu_ops = { + .domain_init = amd_iommu_domain_init, + .domain_destroy = amd_iommu_domain_destroy, + .attach_dev = amd_iommu_attach_device, + .detach_dev = amd_iommu_detach_device, + .map = amd_iommu_map_range, + .unmap = amd_iommu_unmap_range, + .iova_to_phys = amd_iommu_iova_to_phys, +}; + diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 30ae2701b3d..42c33cebf00 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -28,6 +28,7 @@ #include <asm/amd_iommu_types.h> #include <asm/amd_iommu.h> #include <asm/iommu.h> +#include <asm/gart.h> /* * definitions for the ACPI scanning code @@ -121,7 +122,8 @@ 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 = 1; /* if 1, device isolation is enabled */ +bool amd_iommu_isolate = true; /* if true, 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 @@ -242,20 +244,16 @@ static void __init iommu_feature_disable(struct amd_iommu *iommu, u8 bit) } /* Function to enable the hardware */ -void __init iommu_enable(struct amd_iommu *iommu) +static void __init iommu_enable(struct amd_iommu *iommu) { - printk(KERN_INFO "AMD IOMMU: Enabling IOMMU " - "at %02x:%02x.%x cap 0x%hx\n", - iommu->dev->bus->number, - PCI_SLOT(iommu->dev->devfn), - PCI_FUNC(iommu->dev->devfn), - iommu->cap_ptr); + printk(KERN_INFO "AMD IOMMU: Enabling IOMMU at %s cap 0x%hx\n", + dev_name(&iommu->dev->dev), iommu->cap_ptr); iommu_feature_enable(iommu, CONTROL_IOMMU_EN); } /* Function to enable IOMMU event logging and event interrupts */ -void __init iommu_enable_event_logging(struct amd_iommu *iommu) +static void __init iommu_enable_event_logging(struct amd_iommu *iommu) { iommu_feature_enable(iommu, CONTROL_EVT_LOG_EN); iommu_feature_enable(iommu, CONTROL_EVT_INT_EN); @@ -427,6 +425,10 @@ static u8 * __init alloc_command_buffer(struct amd_iommu *iommu) memcpy_toio(iommu->mmio_base + MMIO_CMD_BUF_OFFSET, &entry, sizeof(entry)); + /* set head and tail to zero manually */ + writel(0x00, iommu->mmio_base + MMIO_CMD_HEAD_OFFSET); + writel(0x00, iommu->mmio_base + MMIO_CMD_TAIL_OFFSET); + iommu_feature_enable(iommu, CONTROL_CMDBUF_EN); return cmd_buf; @@ -1074,7 +1076,8 @@ int __init amd_iommu_init(void) goto free; /* IOMMU rlookup table - find the IOMMU for a specific device */ - amd_iommu_rlookup_table = (void *)__get_free_pages(GFP_KERNEL, + amd_iommu_rlookup_table = (void *)__get_free_pages( + GFP_KERNEL | __GFP_ZERO, get_order(rlookup_table_size)); if (amd_iommu_rlookup_table == NULL) goto free; @@ -1212,9 +1215,9 @@ static int __init parse_amd_iommu_options(char *str) { for (; *str; ++str) { if (strncmp(str, "isolate", 7) == 0) - amd_iommu_isolate = 1; + amd_iommu_isolate = true; if (strncmp(str, "share", 5) == 0) - amd_iommu_isolate = 0; + amd_iommu_isolate = false; if (strncmp(str, "fullflush", 9) == 0) amd_iommu_unmap_flush = true; } diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c index 9a32b37ee2e..676debfc170 100644 --- a/arch/x86/kernel/aperture_64.c +++ b/arch/x86/kernel/aperture_64.c @@ -1,8 +1,9 @@ /* * Firmware replacement code. * - * Work around broken BIOSes that don't set an aperture or only set the - * aperture in the AGP bridge. + * Work around broken BIOSes that don't set an aperture, only set the + * aperture in the AGP bridge, or set too small aperture. + * * If all fails map the aperture over some low memory. This is cheaper than * doing bounce buffering. The memory is lost. This is done at early boot * because only the bootmem allocator can allocate 32+MB. diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 16f94879b52..b13d3c4dbd4 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -30,6 +30,7 @@ #include <linux/module.h> #include <linux/dmi.h> #include <linux/dmar.h> +#include <linux/ftrace.h> #include <asm/atomic.h> #include <asm/smp.h> @@ -97,8 +98,8 @@ __setup("apicpmtimer", setup_apicpmtimer); #ifdef HAVE_X2APIC int x2apic; /* x2apic enabled before OS handover */ -int x2apic_preenabled; -int disable_x2apic; +static int x2apic_preenabled; +static int disable_x2apic; static __init int setup_nox2apic(char *str) { disable_x2apic = 1; @@ -118,8 +119,6 @@ EXPORT_SYMBOL_GPL(local_apic_timer_c2_ok); int first_system_vector = 0xfe; -char system_vectors[NR_VECTORS] = { [0 ... NR_VECTORS-1] = SYS_VECTOR_FREE}; - /* * Debug level, exported for io_apic.c */ @@ -141,7 +140,7 @@ 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 struct cpumask *mask); static void apic_pm_activate(void); /* @@ -227,7 +226,7 @@ void xapic_icr_write(u32 low, u32 id) apic_write(APIC_ICR, low); } -u64 xapic_icr_read(void) +static u64 xapic_icr_read(void) { u32 icr1, icr2; @@ -267,7 +266,7 @@ void x2apic_icr_write(u32 low, u32 id) wrmsrl(APIC_BASE_MSR + (APIC_ICR >> 4), ((__u64) id) << 32 | low); } -u64 x2apic_icr_read(void) +static u64 x2apic_icr_read(void) { unsigned long val; @@ -441,6 +440,7 @@ static void lapic_timer_setup(enum clock_event_mode mode, v = apic_read(APIC_LVTT); v |= (APIC_LVT_MASKED | LOCAL_TIMER_VECTOR); apic_write(APIC_LVTT, v); + apic_write(APIC_TMICT, 0xffffffff); break; case CLOCK_EVT_MODE_RESUME: /* Nothing to do here */ @@ -453,7 +453,7 @@ 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 struct cpumask *mask) { #ifdef CONFIG_SMP send_IPI_mask(mask, LOCAL_TIMER_VECTOR); @@ -469,7 +469,7 @@ 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); } @@ -559,13 +559,13 @@ static int __init calibrate_by_pmtimer(long deltapm, long *delta) } else { res = (((u64)deltapm) * mult) >> 22; do_div(res, 1000000); - printk(KERN_WARNING "APIC calibration not consistent " + pr_warning("APIC calibration not consistent " "with PM Timer: %ldms instead of 100ms\n", (long)res); /* Correct the lapic counter value */ res = (((u64)(*delta)) * pm_100ms); do_div(res, deltapm); - printk(KERN_INFO "APIC delta adjusted to PM-Timer: " + pr_info("APIC delta adjusted to PM-Timer: " "%lu (%ld)\n", (unsigned long)res, *delta); *delta = (long)res; } @@ -645,8 +645,7 @@ static int __init calibrate_APIC_clock(void) */ if (calibration_result < (1000000 / HZ)) { local_irq_enable(); - printk(KERN_WARNING - "APIC frequency too slow, disabling apic timer\n"); + pr_warning("APIC frequency too slow, disabling apic timer\n"); return -1; } @@ -672,13 +671,9 @@ static int __init calibrate_APIC_clock(void) while (lapic_cal_loops <= LAPIC_CAL_LOOPS) cpu_relax(); - local_irq_disable(); - /* Stop the lapic timer */ lapic_timer_setup(CLOCK_EVT_MODE_SHUTDOWN, levt); - local_irq_enable(); - /* Jiffies delta */ deltaj = lapic_cal_j2 - lapic_cal_j1; apic_printk(APIC_VERBOSE, "... jiffies delta = %lu\n", deltaj); @@ -692,8 +687,7 @@ static int __init calibrate_APIC_clock(void) local_irq_enable(); if (levt->features & CLOCK_EVT_FEAT_DUMMY) { - printk(KERN_WARNING - "APIC timer disabled due to verification failure.\n"); + pr_warning("APIC timer disabled due to verification failure.\n"); return -1; } @@ -714,7 +708,7 @@ void __init setup_boot_APIC_clock(void) * broadcast mechanism is used. On UP systems simply ignore it. */ if (disable_apic_timer) { - printk(KERN_INFO "Disabling APIC timer\n"); + pr_info("Disabling APIC timer\n"); /* No broadcast on UP ! */ if (num_possible_cpus() > 1) { lapic_clockevent.mult = 1; @@ -741,7 +735,7 @@ void __init setup_boot_APIC_clock(void) if (nmi_watchdog != NMI_IO_APIC) lapic_clockevent.features &= ~CLOCK_EVT_FEAT_DUMMY; else - printk(KERN_WARNING "APIC timer registered as dummy," + pr_warning("APIC timer registered as dummy," " due to nmi_watchdog=%d!\n", nmi_watchdog); /* Setup the lapic or request the broadcast */ @@ -773,8 +767,7 @@ static void local_apic_timer_interrupt(void) * spurious. */ if (!evt->event_handler) { - printk(KERN_WARNING - "Spurious LAPIC timer interrupt on cpu %d\n", cpu); + pr_warning("Spurious LAPIC timer interrupt on cpu %d\n", cpu); /* Switch it off */ lapic_timer_setup(CLOCK_EVT_MODE_SHUTDOWN, evt); return; @@ -783,11 +776,7 @@ static void local_apic_timer_interrupt(void) /* * the NMI deadlock-detector uses this. */ -#ifdef CONFIG_X86_64 - add_pda(apic_timer_irqs, 1); -#else - per_cpu(irq_stat, cpu).apic_timer_irqs++; -#endif + inc_irq_stat(apic_timer_irqs); evt->event_handler(evt); } @@ -800,7 +789,7 @@ static void local_apic_timer_interrupt(void) * [ if a single-CPU system runs an SMP kernel then we call the local * interrupt as well. Thus we cannot inline the local irq ... ] */ -void smp_apic_timer_interrupt(struct pt_regs *regs) +void __irq_entry smp_apic_timer_interrupt(struct pt_regs *regs) { struct pt_regs *old_regs = set_irq_regs(regs); @@ -814,9 +803,7 @@ void smp_apic_timer_interrupt(struct pt_regs *regs) * Besides, if we don't timer interrupts ignore the global * interrupt lock, which is the WrongThing (tm) to do. */ -#ifdef CONFIG_X86_64 exit_idle(); -#endif irq_enter(); local_apic_timer_interrupt(); irq_exit(); @@ -1093,7 +1080,7 @@ static void __cpuinit lapic_setup_esr(void) unsigned int oldvalue, value, maxlvt; if (!lapic_is_integrated()) { - printk(KERN_INFO "No ESR for 82489DX.\n"); + pr_info("No ESR for 82489DX.\n"); return; } @@ -1104,7 +1091,7 @@ static void __cpuinit lapic_setup_esr(void) * ESR disabled - we can't do anything useful with the * errors anyway - mbligh */ - printk(KERN_INFO "Leaving ESR disabled.\n"); + pr_info("Leaving ESR disabled.\n"); return; } @@ -1298,7 +1285,7 @@ void check_x2apic(void) rdmsr(MSR_IA32_APICBASE, msr, msr2); if (msr & X2APIC_ENABLE) { - printk("x2apic enabled by BIOS, switching to x2apic ops\n"); + pr_info("x2apic enabled by BIOS, switching to x2apic ops\n"); x2apic_preenabled = x2apic = 1; apic_ops = &x2apic_ops; } @@ -1310,7 +1297,7 @@ void enable_x2apic(void) rdmsr(MSR_IA32_APICBASE, msr, msr2); if (!(msr & X2APIC_ENABLE)) { - printk("Enabling x2apic\n"); + pr_info("Enabling x2apic\n"); wrmsr(MSR_IA32_APICBASE, msr | X2APIC_ENABLE, 0); } } @@ -1325,9 +1312,8 @@ void __init enable_IR_x2apic(void) return; if (!x2apic_preenabled && disable_x2apic) { - printk(KERN_INFO - "Skipped enabling x2apic and Interrupt-remapping " - "because of nox2apic\n"); + pr_info("Skipped enabling x2apic and Interrupt-remapping " + "because of nox2apic\n"); return; } @@ -1335,22 +1321,19 @@ void __init enable_IR_x2apic(void) panic("Bios already enabled x2apic, can't enforce nox2apic"); if (!x2apic_preenabled && skip_ioapic_setup) { - printk(KERN_INFO - "Skipped enabling x2apic and Interrupt-remapping " - "because of skipping io-apic setup\n"); + pr_info("Skipped enabling x2apic and Interrupt-remapping " + "because of skipping io-apic setup\n"); return; } ret = dmar_table_init(); if (ret) { - printk(KERN_INFO - "dmar_table_init() failed with %d:\n", ret); + pr_info("dmar_table_init() failed with %d:\n", ret); if (x2apic_preenabled) panic("x2apic enabled by bios. But IR enabling failed"); else - printk(KERN_INFO - "Not enabling x2apic,Intr-remapping\n"); + pr_info("Not enabling x2apic,Intr-remapping\n"); return; } @@ -1359,7 +1342,7 @@ void __init enable_IR_x2apic(void) ret = save_mask_IO_APIC_setup(); if (ret) { - printk(KERN_INFO "Saving IO-APIC state failed: %d\n", ret); + pr_info("Saving IO-APIC state failed: %d\n", ret); goto end; } @@ -1394,14 +1377,11 @@ end: if (!ret) { if (!x2apic_preenabled) - printk(KERN_INFO - "Enabled x2apic and interrupt-remapping\n"); + pr_info("Enabled x2apic and interrupt-remapping\n"); else - printk(KERN_INFO - "Enabled Interrupt-remapping\n"); + pr_info("Enabled Interrupt-remapping\n"); } else - printk(KERN_ERR - "Failed to enable Interrupt-remapping and x2apic\n"); + pr_err("Failed to enable Interrupt-remapping and x2apic\n"); #else if (!cpu_has_x2apic) return; @@ -1410,8 +1390,8 @@ end: panic("x2apic enabled prior OS handover," " enable CONFIG_INTR_REMAP"); - printk(KERN_INFO "Enable CONFIG_INTR_REMAP for enabling intr-remapping " - " and x2apic\n"); + pr_info("Enable CONFIG_INTR_REMAP for enabling intr-remapping " + " and x2apic\n"); #endif return; @@ -1428,7 +1408,7 @@ end: static int __init detect_init_APIC(void) { if (!cpu_has_apic) { - printk(KERN_INFO "No local APIC present\n"); + pr_info("No local APIC present\n"); return -1; } @@ -1469,8 +1449,8 @@ static int __init detect_init_APIC(void) * "lapic" specified. */ if (!force_enable_local_apic) { - printk(KERN_INFO "Local APIC disabled by BIOS -- " - "you can enable it with \"lapic\"\n"); + pr_info("Local APIC disabled by BIOS -- " + "you can enable it with \"lapic\"\n"); return -1; } /* @@ -1480,8 +1460,7 @@ static int __init detect_init_APIC(void) */ rdmsr(MSR_IA32_APICBASE, l, h); if (!(l & MSR_IA32_APICBASE_ENABLE)) { - printk(KERN_INFO - "Local APIC disabled by BIOS -- reenabling.\n"); + pr_info("Local APIC disabled by BIOS -- reenabling.\n"); l &= ~MSR_IA32_APICBASE_BASE; l |= MSR_IA32_APICBASE_ENABLE | APIC_DEFAULT_PHYS_BASE; wrmsr(MSR_IA32_APICBASE, l, h); @@ -1494,7 +1473,7 @@ static int __init detect_init_APIC(void) */ features = cpuid_edx(1); if (!(features & (1 << X86_FEATURE_APIC))) { - printk(KERN_WARNING "Could not enable APIC!\n"); + pr_warning("Could not enable APIC!\n"); return -1; } set_cpu_cap(&boot_cpu_data, X86_FEATURE_APIC); @@ -1505,14 +1484,14 @@ static int __init detect_init_APIC(void) if (l & MSR_IA32_APICBASE_ENABLE) mp_lapic_addr = l & MSR_IA32_APICBASE_BASE; - printk(KERN_INFO "Found and enabled local APIC!\n"); + pr_info("Found and enabled local APIC!\n"); apic_pm_activate(); return 0; no_apic: - printk(KERN_INFO "No local APIC present or hardware disabled\n"); + pr_info("No local APIC present or hardware disabled\n"); return -1; } #endif @@ -1588,12 +1567,12 @@ int __init APIC_init_uniprocessor(void) { #ifdef CONFIG_X86_64 if (disable_apic) { - printk(KERN_INFO "Apic disabled\n"); + pr_info("Apic disabled\n"); return -1; } if (!cpu_has_apic) { disable_apic = 1; - printk(KERN_INFO "Apic disabled by BIOS\n"); + pr_info("Apic disabled by BIOS\n"); return -1; } #else @@ -1605,8 +1584,8 @@ int __init APIC_init_uniprocessor(void) */ if (!cpu_has_apic && APIC_INTEGRATED(apic_version[boot_cpu_physical_apicid])) { - printk(KERN_ERR "BIOS bug, local APIC 0x%x not detected!...\n", - boot_cpu_physical_apicid); + pr_err("BIOS bug, local APIC 0x%x not detected!...\n", + boot_cpu_physical_apicid); clear_cpu_cap(&boot_cpu_data, X86_FEATURE_APIC); return -1; } @@ -1682,9 +1661,7 @@ void smp_spurious_interrupt(struct pt_regs *regs) { u32 v; -#ifdef CONFIG_X86_64 exit_idle(); -#endif irq_enter(); /* * Check if this really is a spurious interrupt and ACK it @@ -1695,14 +1672,11 @@ void smp_spurious_interrupt(struct pt_regs *regs) if (v & (1 << (SPURIOUS_APIC_VECTOR & 0x1f))) ack_APIC_irq(); -#ifdef CONFIG_X86_64 - add_pda(irq_spurious_count, 1); -#else + inc_irq_stat(irq_spurious_count); + /* see sw-dev-man vol 3, chapter 7.4.13.5 */ - printk(KERN_INFO "spurious APIC interrupt on CPU#%d, " - "should never happen.\n", smp_processor_id()); - __get_cpu_var(irq_stat).irq_spurious_count++; -#endif + pr_info("spurious APIC interrupt on CPU#%d, " + "should never happen.\n", smp_processor_id()); irq_exit(); } @@ -1713,9 +1687,7 @@ void smp_error_interrupt(struct pt_regs *regs) { u32 v, v1; -#ifdef CONFIG_X86_64 exit_idle(); -#endif irq_enter(); /* First tickle the hardware, only then report what went on. -- REW */ v = apic_read(APIC_ESR); @@ -1724,17 +1696,18 @@ void smp_error_interrupt(struct pt_regs *regs) ack_APIC_irq(); atomic_inc(&irq_err_count); - /* Here is what the APIC error bits mean: - 0: Send CS error - 1: Receive CS error - 2: Send accept error - 3: Receive accept error - 4: Reserved - 5: Send illegal vector - 6: Received illegal vector - 7: Illegal register address - */ - printk(KERN_DEBUG "APIC error on CPU%d: %02x(%02x)\n", + /* + * Here is what the APIC error bits mean: + * 0: Send CS error + * 1: Receive CS error + * 2: Send accept error + * 3: Receive accept error + * 4: Reserved + * 5: Send illegal vector + * 6: Received illegal vector + * 7: Illegal register address + */ + pr_debug("APIC error on CPU%d: %02x(%02x)\n", smp_processor_id(), v , v1); irq_exit(); } @@ -1832,28 +1805,32 @@ void disconnect_bsp_APIC(int virt_wire_setup) void __cpuinit generic_processor_info(int apicid, int version) { int cpu; - cpumask_t tmp_map; /* * Validate version */ if (version == 0x0) { - printk(KERN_WARNING "BIOS bug, APIC version is 0 for CPU#%d! " - "fixing up to 0x10. (tell your hw vendor)\n", + pr_warning("BIOS bug, APIC version is 0 for CPU#%d! " + "fixing up to 0x10. (tell your hw vendor)\n", version); version = 0x10; } apic_version[apicid] = version; - if (num_processors >= NR_CPUS) { - printk(KERN_WARNING "WARNING: NR_CPUS limit of %i reached." - " Processor ignored.\n", NR_CPUS); + if (num_processors >= nr_cpu_ids) { + int max = nr_cpu_ids; + int thiscpu = max + disabled_cpus; + + pr_warning( + "ACPI: NR_CPUS/possible_cpus limit of %i reached." + " Processor %d/0x%x ignored.\n", max, thiscpu, apicid); + + disabled_cpus++; return; } num_processors++; - cpus_complement(tmp_map, cpu_present_map); - cpu = first_cpu(tmp_map); + cpu = cpumask_next_zero(-1, cpu_present_mask); physid_set(apicid, phys_cpu_present_map); if (apicid == boot_cpu_physical_apicid) { @@ -1903,8 +1880,8 @@ 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 @@ -2106,7 +2083,7 @@ __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]; @@ -2209,7 +2186,7 @@ static int __init apic_set_verbosity(char *arg) else if (strcmp("verbose", arg) == 0) apic_verbosity = APIC_VERBOSE; else { - printk(KERN_WARNING "APIC Verbosity level %s not recognised" + pr_warning("APIC Verbosity level %s not recognised" " use apic=verbose or apic=debug\n", arg); return -EINVAL; } 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/asm-offsets_32.c b/arch/x86/kernel/asm-offsets_32.c index 6649d09ad88..ee4df08feee 100644 --- a/arch/x86/kernel/asm-offsets_32.c +++ b/arch/x86/kernel/asm-offsets_32.c @@ -11,7 +11,7 @@ #include <linux/suspend.h> #include <linux/kbuild.h> #include <asm/ucontext.h> -#include "sigframe.h" +#include <asm/sigframe.h> #include <asm/pgtable.h> #include <asm/fixmap.h> #include <asm/processor.h> diff --git a/arch/x86/kernel/asm-offsets_64.c b/arch/x86/kernel/asm-offsets_64.c index 7fcf63d22f8..1d41d3f1edb 100644 --- a/arch/x86/kernel/asm-offsets_64.c +++ b/arch/x86/kernel/asm-offsets_64.c @@ -20,6 +20,8 @@ #include <xen/interface/xen.h> +#include <asm/sigframe.h> + #define __NO_STUBS 1 #undef __SYSCALL #undef _ASM_X86_UNISTD_64_H @@ -87,7 +89,7 @@ int main(void) BLANK(); #undef ENTRY DEFINE(IA32_RT_SIGFRAME_sigcontext, - offsetof (struct rt_sigframe32, uc.uc_mcontext)); + offsetof (struct rt_sigframe_ia32, uc.uc_mcontext)); BLANK(); #endif DEFINE(pbe_address, offsetof(struct pbe, address)); diff --git a/arch/x86/kernel/bios_uv.c b/arch/x86/kernel/bios_uv.c index f0dfe6f17e7..f63882728d9 100644 --- a/arch/x86/kernel/bios_uv.c +++ b/arch/x86/kernel/bios_uv.c @@ -25,7 +25,7 @@ #include <asm/uv/bios.h> #include <asm/uv/uv_hub.h> -struct uv_systab uv_systab; +static struct uv_systab uv_systab; s64 uv_bios_call(enum uv_bios_cmd which, u64 a1, u64 a2, u64 a3, u64 a4, u64 a5) { @@ -69,10 +69,10 @@ s64 uv_bios_call_reentrant(enum uv_bios_cmd which, u64 a1, u64 a2, u64 a3, long sn_partition_id; EXPORT_SYMBOL_GPL(sn_partition_id); -long uv_coherency_id; -EXPORT_SYMBOL_GPL(uv_coherency_id); -long uv_region_size; -EXPORT_SYMBOL_GPL(uv_region_size); +long sn_coherency_id; +EXPORT_SYMBOL_GPL(sn_coherency_id); +long sn_region_size; +EXPORT_SYMBOL_GPL(sn_region_size); int uv_type; @@ -100,6 +100,56 @@ s64 uv_bios_get_sn_info(int fc, int *uvtype, long *partid, long *coher, return ret; } +int +uv_bios_mq_watchlist_alloc(int blade, unsigned long addr, unsigned int mq_size, + unsigned long *intr_mmr_offset) +{ + union uv_watchlist_u size_blade; + u64 watchlist; + s64 ret; + + size_blade.size = mq_size; + size_blade.blade = blade; + + /* + * bios returns watchlist number or negative error number. + */ + ret = (int)uv_bios_call_irqsave(UV_BIOS_WATCHLIST_ALLOC, addr, + size_blade.val, (u64)intr_mmr_offset, + (u64)&watchlist, 0); + if (ret < BIOS_STATUS_SUCCESS) + return ret; + + return watchlist; +} +EXPORT_SYMBOL_GPL(uv_bios_mq_watchlist_alloc); + +int +uv_bios_mq_watchlist_free(int blade, int watchlist_num) +{ + return (int)uv_bios_call_irqsave(UV_BIOS_WATCHLIST_FREE, + blade, watchlist_num, 0, 0, 0); +} +EXPORT_SYMBOL_GPL(uv_bios_mq_watchlist_free); + +s64 +uv_bios_change_memprotect(u64 paddr, u64 len, enum uv_memprotect perms) +{ + return uv_bios_call_irqsave(UV_BIOS_MEMPROTECT, paddr, len, + perms, 0, 0); +} +EXPORT_SYMBOL_GPL(uv_bios_change_memprotect); + +s64 +uv_bios_reserved_page_pa(u64 buf, u64 *cookie, u64 *addr, u64 *len) +{ + s64 ret; + + ret = uv_bios_call_irqsave(UV_BIOS_GET_PARTITION_ADDR, (u64)cookie, + (u64)addr, buf, (u64)len, 0); + return ret; +} +EXPORT_SYMBOL_GPL(uv_bios_reserved_page_pa); s64 uv_bios_freq_base(u64 clock_type, u64 *ticks_per_second) { diff --git a/arch/x86/kernel/check.c b/arch/x86/kernel/check.c new file mode 100644 index 00000000000..2ac0ab71412 --- /dev/null +++ b/arch/x86/kernel/check.c @@ -0,0 +1,161 @@ +#include <linux/module.h> +#include <linux/sched.h> +#include <linux/kthread.h> +#include <linux/workqueue.h> +#include <asm/e820.h> +#include <asm/proto.h> + +/* + * Some BIOSes seem to corrupt the low 64k of memory during events + * like suspend/resume and unplugging an HDMI cable. Reserve all + * remaining free memory in that area and fill it with a distinct + * pattern. + */ +#define MAX_SCAN_AREAS 8 + +static int __read_mostly memory_corruption_check = -1; + +static unsigned __read_mostly corruption_check_size = 64*1024; +static unsigned __read_mostly corruption_check_period = 60; /* seconds */ + +static struct e820entry scan_areas[MAX_SCAN_AREAS]; +static int num_scan_areas; + + +static __init int set_corruption_check(char *arg) +{ + char *end; + + memory_corruption_check = simple_strtol(arg, &end, 10); + + return (*end == 0) ? 0 : -EINVAL; +} +early_param("memory_corruption_check", set_corruption_check); + +static __init int set_corruption_check_period(char *arg) +{ + char *end; + + corruption_check_period = simple_strtoul(arg, &end, 10); + + return (*end == 0) ? 0 : -EINVAL; +} +early_param("memory_corruption_check_period", set_corruption_check_period); + +static __init int set_corruption_check_size(char *arg) +{ + char *end; + unsigned size; + + size = memparse(arg, &end); + + if (*end == '\0') + corruption_check_size = size; + + return (size == corruption_check_size) ? 0 : -EINVAL; +} +early_param("memory_corruption_check_size", set_corruption_check_size); + + +void __init setup_bios_corruption_check(void) +{ + u64 addr = PAGE_SIZE; /* assume first page is reserved anyway */ + + if (memory_corruption_check == -1) { + memory_corruption_check = +#ifdef CONFIG_X86_BOOTPARAM_MEMORY_CORRUPTION_CHECK + 1 +#else + 0 +#endif + ; + } + + if (corruption_check_size == 0) + memory_corruption_check = 0; + + if (!memory_corruption_check) + return; + + corruption_check_size = round_up(corruption_check_size, PAGE_SIZE); + + while (addr < corruption_check_size && num_scan_areas < MAX_SCAN_AREAS) { + u64 size; + addr = find_e820_area_size(addr, &size, PAGE_SIZE); + + if (addr == 0) + break; + + if ((addr + size) > corruption_check_size) + size = corruption_check_size - addr; + + if (size == 0) + break; + + e820_update_range(addr, size, E820_RAM, E820_RESERVED); + scan_areas[num_scan_areas].addr = addr; + scan_areas[num_scan_areas].size = size; + num_scan_areas++; + + /* Assume we've already mapped this early memory */ + memset(__va(addr), 0, size); + + addr += size; + } + + printk(KERN_INFO "Scanning %d areas for low memory corruption\n", + num_scan_areas); + update_e820(); +} + + +void check_for_bios_corruption(void) +{ + int i; + int corruption = 0; + + if (!memory_corruption_check) + return; + + for (i = 0; i < num_scan_areas; i++) { + unsigned long *addr = __va(scan_areas[i].addr); + unsigned long size = scan_areas[i].size; + + for (; size; addr++, size -= sizeof(unsigned long)) { + if (!*addr) + continue; + printk(KERN_ERR "Corrupted low memory at %p (%lx phys) = %08lx\n", + addr, __pa(addr), *addr); + corruption = 1; + *addr = 0; + } + } + + WARN_ONCE(corruption, KERN_ERR "Memory corruption detected in low memory\n"); +} + +static void check_corruption(struct work_struct *dummy); +static DECLARE_DELAYED_WORK(bios_check_work, check_corruption); + +static void check_corruption(struct work_struct *dummy) +{ + check_for_bios_corruption(); + schedule_delayed_work(&bios_check_work, + round_jiffies_relative(corruption_check_period*HZ)); +} + +static int start_periodic_check_for_corruption(void) +{ + if (!memory_corruption_check || corruption_check_period == 0) + return 0; + + printk(KERN_INFO "Scanning for low memory corruption every %d seconds\n", + corruption_check_period); + + /* First time we run the checks right away */ + schedule_delayed_work(&bios_check_work, 0); + return 0; +} + +module_init(start_periodic_check_for_corruption); + diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile index 82ec6075c05..82db7f45e2d 100644 --- a/arch/x86/kernel/cpu/Makefile +++ b/arch/x86/kernel/cpu/Makefile @@ -2,8 +2,14 @@ # Makefile for x86-compatible CPU details and quirks # +# Don't trace early stages of a secondary CPU boot +ifdef CONFIG_FUNCTION_TRACER +CFLAGS_REMOVE_common.o = -pg +endif + obj-y := intel_cacheinfo.o addon_cpuid_features.o obj-y += proc.o capflags.o powerflags.o common.o +obj-y += vmware.o hypervisor.o obj-$(CONFIG_X86_32) += bugs.o cmpxchg.o obj-$(CONFIG_X86_64) += bugs_64.o diff --git a/arch/x86/kernel/cpu/addon_cpuid_features.c b/arch/x86/kernel/cpu/addon_cpuid_features.c index ef8f831af82..2cf23634b6d 100644 --- a/arch/x86/kernel/cpu/addon_cpuid_features.c +++ b/arch/x86/kernel/cpu/addon_cpuid_features.c @@ -120,9 +120,17 @@ void __cpuinit detect_extended_topology(struct cpuinfo_x86 *c) c->cpu_core_id = phys_pkg_id(c->initial_apicid, ht_mask_width) & core_select_mask; c->phys_proc_id = phys_pkg_id(c->initial_apicid, core_plus_mask_width); + /* + * Reinit the apicid, now that we have extended initial_apicid. + */ + c->apicid = phys_pkg_id(c->initial_apicid, 0); #else c->cpu_core_id = phys_pkg_id(ht_mask_width) & core_select_mask; c->phys_proc_id = phys_pkg_id(core_plus_mask_width); + /* + * Reinit the apicid, now that we have extended initial_apicid. + */ + c->apicid = phys_pkg_id(0); #endif c->x86_max_cores = (core_level_siblings / smp_num_siblings); diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 8f1e31db2ad..7c878f6aa91 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -283,9 +283,14 @@ static void __cpuinit early_init_amd(struct cpuinfo_x86 *c) { early_init_amd_mc(c); - /* c->x86_power is 8000_0007 edx. Bit 8 is constant TSC */ - if (c->x86_power & (1<<8)) + /* + * c->x86_power is 8000_0007 edx. Bit 8 is TSC runs at constant rate + * with P/T states and does not stop in deep C-states + */ + if (c->x86_power & (1 << 8)) { set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); + set_cpu_cap(c, X86_FEATURE_NONSTOP_TSC); + } #ifdef CONFIG_X86_64 set_cpu_cap(c, X86_FEATURE_SYSCALL32); diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index b9c9ea0217a..3f95a40f718 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -36,6 +36,7 @@ #include <asm/proto.h> #include <asm/sections.h> #include <asm/setup.h> +#include <asm/hypervisor.h> #include "cpu.h" @@ -354,7 +355,7 @@ void __cpuinit detect_ht(struct cpuinfo_x86 *c) printk(KERN_INFO "CPU: Hyper-Threading is disabled\n"); } else if (smp_num_siblings > 1) { - if (smp_num_siblings > NR_CPUS) { + if (smp_num_siblings > nr_cpu_ids) { printk(KERN_WARNING "CPU: Unsupported number of siblings %d", smp_num_siblings); smp_num_siblings = 1; @@ -703,6 +704,7 @@ static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) detect_ht(c); #endif + init_hypervisor(c); /* * On SMP, boot_cpu_data holds the common feature set between * all CPUs; so make sure that we indicate which features are @@ -862,7 +864,7 @@ EXPORT_SYMBOL(_cpu_pda); struct desc_ptr idt_descr = { 256 * 16 - 1, (unsigned long) idt_table }; -char boot_cpu_stack[IRQSTACKSIZE] __page_aligned_bss; +static char boot_cpu_stack[IRQSTACKSIZE] __page_aligned_bss; void __cpuinit pda_init(int cpu) { @@ -903,8 +905,8 @@ void __cpuinit pda_init(int cpu) } } -char boot_exception_stacks[(N_EXCEPTION_STACKS - 1) * EXCEPTION_STKSZ + - DEBUG_STKSZ] __page_aligned_bss; +static char boot_exception_stacks[(N_EXCEPTION_STACKS - 1) * EXCEPTION_STKSZ + + DEBUG_STKSZ] __page_aligned_bss; extern asmlinkage void ignore_sysret(void); diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c index 8e48c5d4467..28102ad1a36 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; @@ -513,6 +517,17 @@ acpi_cpufreq_guess_freq(struct acpi_cpufreq_data *data, unsigned int cpu) } } +static void free_acpi_perf_data(void) +{ + unsigned int i; + + /* Freeing a NULL pointer is OK, and alloc_percpu zeroes. */ + for_each_possible_cpu(i) + free_cpumask_var(per_cpu_ptr(acpi_perf_data, i) + ->shared_cpu_map); + free_percpu(acpi_perf_data); +} + /* * acpi_cpufreq_early_init - initialize ACPI P-States library * @@ -523,6 +538,7 @@ acpi_cpufreq_guess_freq(struct acpi_cpufreq_data *data, unsigned int cpu) */ static int __init acpi_cpufreq_early_init(void) { + unsigned int i; dprintk("acpi_cpufreq_early_init\n"); acpi_perf_data = alloc_percpu(struct acpi_processor_performance); @@ -530,6 +546,16 @@ static int __init acpi_cpufreq_early_init(void) dprintk("Memory allocation error for acpi_perf_data.\n"); return -ENOMEM; } + for_each_possible_cpu(i) { + if (!alloc_cpumask_var_node( + &per_cpu_ptr(acpi_perf_data, i)->shared_cpu_map, + GFP_KERNEL, cpu_to_node(i))) { + + /* Freeing a NULL pointer is OK: alloc_percpu zeroes. */ + free_acpi_perf_data(); + return -ENOMEM; + } + } /* Do initialization in ACPI core */ acpi_processor_preregister_performance(acpi_perf_data); @@ -600,9 +626,9 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy) */ if (policy->shared_type == CPUFREQ_SHARED_TYPE_ALL || policy->shared_type == CPUFREQ_SHARED_TYPE_ANY) { - policy->cpus = perf->shared_cpu_map; + cpumask_copy(&policy->cpus, perf->shared_cpu_map); } - policy->related_cpus = perf->shared_cpu_map; + cpumask_copy(&policy->related_cpus, perf->shared_cpu_map); #ifdef CONFIG_SMP dmi_check_system(sw_any_bug_dmi_table); @@ -791,7 +817,7 @@ static int __init acpi_cpufreq_init(void) ret = cpufreq_register_driver(&acpi_cpufreq_driver); if (ret) - free_percpu(acpi_perf_data); + free_acpi_perf_data(); return ret; } diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k7.c b/arch/x86/kernel/cpu/cpufreq/powernow-k7.c index 7c7d56b4313..1b446d79a8f 100644 --- a/arch/x86/kernel/cpu/cpufreq/powernow-k7.c +++ b/arch/x86/kernel/cpu/cpufreq/powernow-k7.c @@ -310,6 +310,12 @@ static int powernow_acpi_init(void) goto err0; } + if (!alloc_cpumask_var(&acpi_processor_perf->shared_cpu_map, + GFP_KERNEL)) { + retval = -ENOMEM; + goto err05; + } + if (acpi_processor_register_performance(acpi_processor_perf, 0)) { retval = -EIO; goto err1; @@ -412,6 +418,8 @@ static int powernow_acpi_init(void) err2: acpi_processor_unregister_performance(acpi_processor_perf, 0); err1: + free_cpumask_var(acpi_processor_perf->shared_cpu_map); +err05: kfree(acpi_processor_perf); err0: printk(KERN_WARNING PFX "ACPI perflib can not be used in this platform\n"); @@ -652,6 +660,7 @@ static int powernow_cpu_exit (struct cpufreq_policy *policy) { #ifdef CONFIG_X86_POWERNOW_K7_ACPI if (acpi_processor_perf) { acpi_processor_unregister_performance(acpi_processor_perf, 0); + free_cpumask_var(acpi_processor_perf->shared_cpu_map); kfree(acpi_processor_perf); } #endif diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c index 7f05f44b97e..c3c9adbaa26 100644 --- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c +++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c @@ -766,7 +766,7 @@ static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) { struct cpufreq_frequency_table *powernow_table; - int ret_val; + int ret_val = -ENODEV; if (acpi_processor_register_performance(&data->acpi_data, data->cpu)) { dprintk("register performance failed: bad ACPI data\n"); @@ -815,6 +815,13 @@ static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) /* notify BIOS that we exist */ acpi_processor_notify_smm(THIS_MODULE); + if (!alloc_cpumask_var(&data->acpi_data.shared_cpu_map, GFP_KERNEL)) { + printk(KERN_ERR PFX + "unable to alloc powernow_k8_data cpumask\n"); + ret_val = -ENOMEM; + goto err_out_mem; + } + return 0; err_out_mem: @@ -826,7 +833,7 @@ err_out: /* data->acpi_data.state_count informs us at ->exit() whether ACPI was used */ data->acpi_data.state_count = 0; - return -ENODEV; + return ret_val; } static int fill_powernow_table_pstate(struct powernow_k8_data *data, struct cpufreq_frequency_table *powernow_table) @@ -929,6 +936,7 @@ static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data) { if (data->acpi_data.state_count) acpi_processor_unregister_performance(&data->acpi_data, data->cpu); + free_cpumask_var(data->acpi_data.shared_cpu_map); } #else @@ -1134,7 +1142,8 @@ 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)) { + rc = powernow_k8_cpu_init_acpi(data); + if (rc) { /* * Use the PSB BIOS structure. This is only availabe on * an UP version, and is deprecated by AMD. @@ -1152,20 +1161,17 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol) "ACPI maintainers and complain to your BIOS " "vendor.\n"); #endif - kfree(data); - return -ENODEV; + goto err_out; } if (pol->cpu != 0) { printk(KERN_ERR FW_BUG PFX "No ACPI _PSS objects for " "CPU other than CPU0. Complain to your BIOS " "vendor.\n"); - kfree(data); - return -ENODEV; + goto err_out; } rc = find_psb_table(data); if (rc) { - kfree(data); - return -ENODEV; + goto err_out; } } diff --git a/arch/x86/kernel/cpu/hypervisor.c b/arch/x86/kernel/cpu/hypervisor.c new file mode 100644 index 00000000000..fb5b86af0b0 --- /dev/null +++ b/arch/x86/kernel/cpu/hypervisor.c @@ -0,0 +1,58 @@ +/* + * Common hypervisor code + * + * Copyright (C) 2008, VMware, Inc. + * Author : Alok N Kataria <akataria@vmware.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. + * + */ + +#include <asm/processor.h> +#include <asm/vmware.h> +#include <asm/hypervisor.h> + +static inline void __cpuinit +detect_hypervisor_vendor(struct cpuinfo_x86 *c) +{ + if (vmware_platform()) { + c->x86_hyper_vendor = X86_HYPER_VENDOR_VMWARE; + } else { + c->x86_hyper_vendor = X86_HYPER_VENDOR_NONE; + } +} + +unsigned long get_hypervisor_tsc_freq(void) +{ + if (boot_cpu_data.x86_hyper_vendor == X86_HYPER_VENDOR_VMWARE) + return vmware_get_tsc_khz(); + return 0; +} + +static inline void __cpuinit +hypervisor_set_feature_bits(struct cpuinfo_x86 *c) +{ + if (boot_cpu_data.x86_hyper_vendor == X86_HYPER_VENDOR_VMWARE) { + vmware_set_feature_bits(c); + return; + } +} + +void __cpuinit init_hypervisor(struct cpuinfo_x86 *c) +{ + detect_hypervisor_vendor(c); + hypervisor_set_feature_bits(c); +} diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index cce0b6118d5..8ea6929e974 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -11,7 +11,6 @@ #include <asm/pgtable.h> #include <asm/msr.h> #include <asm/uaccess.h> -#include <asm/ptrace.h> #include <asm/ds.h> #include <asm/bugs.h> @@ -41,6 +40,16 @@ static void __cpuinit early_init_intel(struct cpuinfo_x86 *c) if (c->x86 == 15 && c->x86_cache_alignment == 64) c->x86_cache_alignment = 128; #endif + + /* + * c->x86_power is 8000_0007 edx. Bit 8 is TSC runs at constant rate + * with P/T states and does not stop in deep C-states + */ + if (c->x86_power & (1 << 8)) { + set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); + set_cpu_cap(c, X86_FEATURE_NONSTOP_TSC); + } + } #ifdef CONFIG_X86_32 @@ -242,6 +251,13 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) intel_workarounds(c); + /* + * Detect the extended topology information if available. This + * will reinitialise the initial_apicid which will be used + * in init_intel_cacheinfo() + */ + detect_extended_topology(c); + l2 = init_intel_cacheinfo(c); if (c->cpuid_level > 9) { unsigned eax = cpuid_eax(10); @@ -307,13 +323,8 @@ 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); - - if (cpu_has_bts) - ptrace_bts_init_intel(c); - #endif - detect_extended_topology(c); if (!cpu_has(c, X86_FEATURE_XTOPOLOGY)) { /* * let's use the legacy cpuid vector 0x1 and 0x4 for topology diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 3f46afbb1cf..48533d77be7 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 __cpuinit 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; -out: + per_cpu(cpuid4_info, cpu) = kzalloc( + sizeof(struct _cpuid4_info) * num_cache_leaves, GFP_KERNEL); + if (per_cpu(cpuid4_info, cpu) == NULL) + return -ENOMEM; + + 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'; } @@ -644,20 +641,17 @@ static inline ssize_t show_shared_cpu_list(struct _cpuid4_info *leaf, char *buf) return show_shared_cpu_map_func(leaf, 1, buf); } -static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) { - switch(this_leaf->eax.split.type) { - case CACHE_TYPE_DATA: +static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) +{ + switch (this_leaf->eax.split.type) { + case CACHE_TYPE_DATA: return sprintf(buf, "Data\n"); - break; - case CACHE_TYPE_INST: + case CACHE_TYPE_INST: return sprintf(buf, "Instruction\n"); - break; - case CACHE_TYPE_UNIFIED: + case CACHE_TYPE_UNIFIED: return sprintf(buf, "Unified\n"); - break; - default: + default: return sprintf(buf, "Unknown\n"); - break; } } diff --git a/arch/x86/kernel/cpu/mcheck/mce_64.c b/arch/x86/kernel/cpu/mcheck/mce_64.c index 4b031a4ac85..1c838032fd3 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_64.c @@ -510,12 +510,9 @@ static void __cpuinit mce_cpu_features(struct cpuinfo_x86 *c) */ void __cpuinit mcheck_init(struct cpuinfo_x86 *c) { - static cpumask_t mce_cpus = CPU_MASK_NONE; - mce_cpu_quirks(c); if (mce_dont_init || - cpu_test_and_set(smp_processor_id(), mce_cpus) || !mce_available(c)) return; diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c index 5eb390a4b2e..a5a5e053037 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); } } } @@ -237,7 +248,7 @@ asmlinkage void mce_threshold_interrupt(void) } } out: - add_pda(irq_threshold_count, 1); + inc_irq_stat(irq_threshold_count); irq_exit(); } @@ -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/cpu/mcheck/mce_intel_64.c b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c index c17eaf5dd6d..4b48f251fd3 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c @@ -26,7 +26,7 @@ asmlinkage void smp_thermal_interrupt(void) if (therm_throt_process(msr_val & 1)) mce_log_therm_throt_event(smp_processor_id(), msr_val); - add_pda(irq_thermal_count, 1); + inc_irq_stat(irq_thermal_count); irq_exit(); } diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index 4e8d77f01ee..b59ddcc88cd 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -14,14 +14,6 @@ #include <asm/pat.h> #include "mtrr.h" -struct mtrr_state { - struct mtrr_var_range var_ranges[MAX_VAR_RANGES]; - mtrr_type fixed_ranges[NUM_FIXED_RANGES]; - unsigned char enabled; - unsigned char have_fixed; - mtrr_type def_type; -}; - struct fixed_range_block { int base_msr; /* start address of an MTRR block */ int ranges; /* number of MTRRs in this block */ @@ -35,10 +27,12 @@ static struct fixed_range_block fixed_range_blocks[] = { }; static unsigned long smp_changes_mask; -static struct mtrr_state mtrr_state = {}; static int mtrr_state_set; u64 mtrr_tom2; +struct mtrr_state_type mtrr_state = {}; +EXPORT_SYMBOL_GPL(mtrr_state); + #undef MODULE_PARAM_PREFIX #define MODULE_PARAM_PREFIX "mtrr." diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c index c78c04821ea..d259e5d2e05 100644 --- a/arch/x86/kernel/cpu/mtrr/main.c +++ b/arch/x86/kernel/cpu/mtrr/main.c @@ -49,7 +49,7 @@ u32 num_var_ranges = 0; -unsigned int mtrr_usage_table[MAX_VAR_RANGES]; +unsigned int mtrr_usage_table[MTRR_MAX_VAR_RANGES]; static DEFINE_MUTEX(mtrr_mutex); u64 size_or_mask, size_and_mask; @@ -574,7 +574,7 @@ struct mtrr_value { unsigned long lsize; }; -static struct mtrr_value mtrr_state[MAX_VAR_RANGES]; +static struct mtrr_value mtrr_state[MTRR_MAX_VAR_RANGES]; static int mtrr_save(struct sys_device * sysdev, pm_message_t state) { @@ -803,6 +803,7 @@ x86_get_mtrr_mem_range(struct res_range *range, int nr_range, } static struct res_range __initdata range[RANGE_NUM]; +static int __initdata nr_range; #ifdef CONFIG_MTRR_SANITIZER @@ -823,16 +824,14 @@ static int enable_mtrr_cleanup __initdata = static int __init disable_mtrr_cleanup_setup(char *str) { - if (enable_mtrr_cleanup != -1) - enable_mtrr_cleanup = 0; + enable_mtrr_cleanup = 0; return 0; } early_param("disable_mtrr_cleanup", disable_mtrr_cleanup_setup); static int __init enable_mtrr_cleanup_setup(char *str) { - if (enable_mtrr_cleanup != -1) - enable_mtrr_cleanup = 1; + enable_mtrr_cleanup = 1; return 0; } early_param("enable_mtrr_cleanup", enable_mtrr_cleanup_setup); @@ -1206,39 +1205,43 @@ struct mtrr_cleanup_result { #define PSHIFT (PAGE_SHIFT - 10) static struct mtrr_cleanup_result __initdata result[NUM_RESULT]; -static struct res_range __initdata range_new[RANGE_NUM]; static unsigned long __initdata min_loss_pfn[RANGE_NUM]; -static int __init mtrr_cleanup(unsigned address_bits) +static void __init print_out_mtrr_range_state(void) { - unsigned long extra_remove_base, extra_remove_size; - unsigned long base, size, def, dummy; - mtrr_type type; - int nr_range, nr_range_new; - u64 chunk_size, gran_size; - unsigned long range_sums, range_sums_new; - int index_good; - int num_reg_good; int i; + char start_factor = 'K', size_factor = 'K'; + unsigned long start_base, size_base; + mtrr_type type; - /* extra one for all 0 */ - int num[MTRR_NUM_TYPES + 1]; + for (i = 0; i < num_var_ranges; i++) { - if (!is_cpu(INTEL) || enable_mtrr_cleanup < 1) - return 0; - rdmsr(MTRRdefType_MSR, def, dummy); - def &= 0xff; - if (def != MTRR_TYPE_UNCACHABLE) - return 0; + size_base = range_state[i].size_pfn << (PAGE_SHIFT - 10); + if (!size_base) + continue; - /* get it and store it aside */ - memset(range_state, 0, sizeof(range_state)); - for (i = 0; i < num_var_ranges; i++) { - mtrr_if->get(i, &base, &size, &type); - range_state[i].base_pfn = base; - range_state[i].size_pfn = size; - range_state[i].type = type; + size_base = to_size_factor(size_base, &size_factor), + start_base = range_state[i].base_pfn << (PAGE_SHIFT - 10); + start_base = to_size_factor(start_base, &start_factor), + type = range_state[i].type; + + printk(KERN_DEBUG "reg %d, base: %ld%cB, range: %ld%cB, type %s\n", + i, start_base, start_factor, + size_base, size_factor, + (type == MTRR_TYPE_UNCACHABLE) ? "UC" : + ((type == MTRR_TYPE_WRPROT) ? "WP" : + ((type == MTRR_TYPE_WRBACK) ? "WB" : "Other")) + ); } +} + +static int __init mtrr_need_cleanup(void) +{ + int i; + mtrr_type type; + unsigned long size; + /* extra one for all 0 */ + int num[MTRR_NUM_TYPES + 1]; /* check entries number */ memset(num, 0, sizeof(num)); @@ -1263,29 +1266,133 @@ static int __init mtrr_cleanup(unsigned address_bits) num_var_ranges - num[MTRR_NUM_TYPES]) return 0; - /* print original var MTRRs at first, for debugging: */ - printk(KERN_DEBUG "original variable MTRRs\n"); - for (i = 0; i < num_var_ranges; i++) { - char start_factor = 'K', size_factor = 'K'; - unsigned long start_base, size_base; + return 1; +} - size_base = range_state[i].size_pfn << (PAGE_SHIFT - 10); - if (!size_base) - continue; +static unsigned long __initdata range_sums; +static void __init mtrr_calc_range_state(u64 chunk_size, u64 gran_size, + unsigned long extra_remove_base, + unsigned long extra_remove_size, + int i) +{ + int num_reg; + static struct res_range range_new[RANGE_NUM]; + static int nr_range_new; + unsigned long range_sums_new; + + /* convert ranges to var ranges state */ + num_reg = x86_setup_var_mtrrs(range, nr_range, + chunk_size, gran_size); + + /* we got new setting in range_state, check it */ + memset(range_new, 0, sizeof(range_new)); + nr_range_new = x86_get_mtrr_mem_range(range_new, 0, + extra_remove_base, extra_remove_size); + range_sums_new = sum_ranges(range_new, nr_range_new); + + result[i].chunk_sizek = chunk_size >> 10; + result[i].gran_sizek = gran_size >> 10; + result[i].num_reg = num_reg; + if (range_sums < range_sums_new) { + result[i].lose_cover_sizek = + (range_sums_new - range_sums) << PSHIFT; + result[i].bad = 1; + } else + result[i].lose_cover_sizek = + (range_sums - range_sums_new) << PSHIFT; - size_base = to_size_factor(size_base, &size_factor), - start_base = range_state[i].base_pfn << (PAGE_SHIFT - 10); - start_base = to_size_factor(start_base, &start_factor), - type = range_state[i].type; + /* double check it */ + if (!result[i].bad && !result[i].lose_cover_sizek) { + if (nr_range_new != nr_range || + memcmp(range, range_new, sizeof(range))) + result[i].bad = 1; + } - printk(KERN_DEBUG "reg %d, base: %ld%cB, range: %ld%cB, type %s\n", - i, start_base, start_factor, - size_base, size_factor, - (type == MTRR_TYPE_UNCACHABLE) ? "UC" : - ((type == MTRR_TYPE_WRPROT) ? "WP" : - ((type == MTRR_TYPE_WRBACK) ? "WB" : "Other")) - ); + if (!result[i].bad && (range_sums - range_sums_new < + min_loss_pfn[num_reg])) { + min_loss_pfn[num_reg] = + range_sums - range_sums_new; } +} + +static void __init mtrr_print_out_one_result(int i) +{ + char gran_factor, chunk_factor, lose_factor; + unsigned long gran_base, chunk_base, lose_base; + + gran_base = to_size_factor(result[i].gran_sizek, &gran_factor), + chunk_base = to_size_factor(result[i].chunk_sizek, &chunk_factor), + lose_base = to_size_factor(result[i].lose_cover_sizek, &lose_factor), + printk(KERN_INFO "%sgran_size: %ld%c \tchunk_size: %ld%c \t", + result[i].bad ? "*BAD*" : " ", + gran_base, gran_factor, chunk_base, chunk_factor); + printk(KERN_CONT "num_reg: %d \tlose cover RAM: %s%ld%c\n", + result[i].num_reg, result[i].bad ? "-" : "", + lose_base, lose_factor); +} + +static int __init mtrr_search_optimal_index(void) +{ + int i; + int num_reg_good; + int index_good; + + if (nr_mtrr_spare_reg >= num_var_ranges) + nr_mtrr_spare_reg = num_var_ranges - 1; + num_reg_good = -1; + for (i = num_var_ranges - nr_mtrr_spare_reg; i > 0; i--) { + if (!min_loss_pfn[i]) + num_reg_good = i; + } + + index_good = -1; + if (num_reg_good != -1) { + for (i = 0; i < NUM_RESULT; i++) { + if (!result[i].bad && + result[i].num_reg == num_reg_good && + !result[i].lose_cover_sizek) { + index_good = i; + break; + } + } + } + + return index_good; +} + + +static int __init mtrr_cleanup(unsigned address_bits) +{ + unsigned long extra_remove_base, extra_remove_size; + unsigned long base, size, def, dummy; + mtrr_type type; + u64 chunk_size, gran_size; + int index_good; + int i; + + if (!is_cpu(INTEL) || enable_mtrr_cleanup < 1) + return 0; + rdmsr(MTRRdefType_MSR, def, dummy); + def &= 0xff; + if (def != MTRR_TYPE_UNCACHABLE) + return 0; + + /* get it and store it aside */ + memset(range_state, 0, sizeof(range_state)); + for (i = 0; i < num_var_ranges; i++) { + mtrr_if->get(i, &base, &size, &type); + range_state[i].base_pfn = base; + range_state[i].size_pfn = size; + range_state[i].type = type; + } + + /* check if we need handle it and can handle it */ + if (!mtrr_need_cleanup()) + return 0; + + /* print original var MTRRs at first, for debugging: */ + printk(KERN_DEBUG "original variable MTRRs\n"); + print_out_mtrr_range_state(); memset(range, 0, sizeof(range)); extra_remove_size = 0; @@ -1309,176 +1416,64 @@ static int __init mtrr_cleanup(unsigned address_bits) range_sums >> (20 - PAGE_SHIFT)); if (mtrr_chunk_size && mtrr_gran_size) { - int num_reg; - char gran_factor, chunk_factor, lose_factor; - unsigned long gran_base, chunk_base, lose_base; - - debug_print++; - /* convert ranges to var ranges state */ - num_reg = x86_setup_var_mtrrs(range, nr_range, mtrr_chunk_size, - mtrr_gran_size); + i = 0; + mtrr_calc_range_state(mtrr_chunk_size, mtrr_gran_size, + extra_remove_base, extra_remove_size, i); - /* we got new setting in range_state, check it */ - memset(range_new, 0, sizeof(range_new)); - nr_range_new = x86_get_mtrr_mem_range(range_new, 0, - extra_remove_base, - extra_remove_size); - range_sums_new = sum_ranges(range_new, nr_range_new); + mtrr_print_out_one_result(i); - i = 0; - result[i].chunk_sizek = mtrr_chunk_size >> 10; - result[i].gran_sizek = mtrr_gran_size >> 10; - result[i].num_reg = num_reg; - if (range_sums < range_sums_new) { - result[i].lose_cover_sizek = - (range_sums_new - range_sums) << PSHIFT; - result[i].bad = 1; - } else - result[i].lose_cover_sizek = - (range_sums - range_sums_new) << PSHIFT; - - gran_base = to_size_factor(result[i].gran_sizek, &gran_factor), - chunk_base = to_size_factor(result[i].chunk_sizek, &chunk_factor), - lose_base = to_size_factor(result[i].lose_cover_sizek, &lose_factor), - printk(KERN_INFO "%sgran_size: %ld%c \tchunk_size: %ld%c \t", - result[i].bad?"*BAD*":" ", - gran_base, gran_factor, chunk_base, chunk_factor); - printk(KERN_CONT "num_reg: %d \tlose cover RAM: %s%ld%c\n", - result[i].num_reg, result[i].bad?"-":"", - lose_base, lose_factor); if (!result[i].bad) { set_var_mtrr_all(address_bits); return 1; } printk(KERN_INFO "invalid mtrr_gran_size or mtrr_chunk_size, " "will find optimal one\n"); - debug_print--; - memset(result, 0, sizeof(result[0])); } i = 0; memset(min_loss_pfn, 0xff, sizeof(min_loss_pfn)); memset(result, 0, sizeof(result)); for (gran_size = (1ULL<<16); gran_size < (1ULL<<32); gran_size <<= 1) { - char gran_factor; - unsigned long gran_base; - - if (debug_print) - gran_base = to_size_factor(gran_size >> 10, &gran_factor); for (chunk_size = gran_size; chunk_size < (1ULL<<32); chunk_size <<= 1) { - int num_reg; - if (debug_print) { - char chunk_factor; - unsigned long chunk_base; - - chunk_base = to_size_factor(chunk_size>>10, &chunk_factor), - printk(KERN_INFO "\n"); - printk(KERN_INFO "gran_size: %ld%c chunk_size: %ld%c \n", - gran_base, gran_factor, chunk_base, chunk_factor); - } if (i >= NUM_RESULT) continue; - /* convert ranges to var ranges state */ - num_reg = x86_setup_var_mtrrs(range, nr_range, - chunk_size, gran_size); - - /* we got new setting in range_state, check it */ - memset(range_new, 0, sizeof(range_new)); - nr_range_new = x86_get_mtrr_mem_range(range_new, 0, - extra_remove_base, extra_remove_size); - range_sums_new = sum_ranges(range_new, nr_range_new); - - result[i].chunk_sizek = chunk_size >> 10; - result[i].gran_sizek = gran_size >> 10; - result[i].num_reg = num_reg; - if (range_sums < range_sums_new) { - result[i].lose_cover_sizek = - (range_sums_new - range_sums) << PSHIFT; - result[i].bad = 1; - } else - result[i].lose_cover_sizek = - (range_sums - range_sums_new) << PSHIFT; - - /* double check it */ - if (!result[i].bad && !result[i].lose_cover_sizek) { - if (nr_range_new != nr_range || - memcmp(range, range_new, sizeof(range))) - result[i].bad = 1; + mtrr_calc_range_state(chunk_size, gran_size, + extra_remove_base, extra_remove_size, i); + if (debug_print) { + mtrr_print_out_one_result(i); + printk(KERN_INFO "\n"); } - if (!result[i].bad && (range_sums - range_sums_new < - min_loss_pfn[num_reg])) { - min_loss_pfn[num_reg] = - range_sums - range_sums_new; - } i++; } } - /* print out all */ - for (i = 0; i < NUM_RESULT; i++) { - char gran_factor, chunk_factor, lose_factor; - unsigned long gran_base, chunk_base, lose_base; - - gran_base = to_size_factor(result[i].gran_sizek, &gran_factor), - chunk_base = to_size_factor(result[i].chunk_sizek, &chunk_factor), - lose_base = to_size_factor(result[i].lose_cover_sizek, &lose_factor), - printk(KERN_INFO "%sgran_size: %ld%c \tchunk_size: %ld%c \t", - result[i].bad?"*BAD*":" ", - gran_base, gran_factor, chunk_base, chunk_factor); - printk(KERN_CONT "num_reg: %d \tlose cover RAM: %s%ld%c\n", - result[i].num_reg, result[i].bad?"-":"", - lose_base, lose_factor); - } - /* try to find the optimal index */ - if (nr_mtrr_spare_reg >= num_var_ranges) - nr_mtrr_spare_reg = num_var_ranges - 1; - num_reg_good = -1; - for (i = num_var_ranges - nr_mtrr_spare_reg; i > 0; i--) { - if (!min_loss_pfn[i]) - num_reg_good = i; - } - - index_good = -1; - if (num_reg_good != -1) { - for (i = 0; i < NUM_RESULT; i++) { - if (!result[i].bad && - result[i].num_reg == num_reg_good && - !result[i].lose_cover_sizek) { - index_good = i; - break; - } - } - } + index_good = mtrr_search_optimal_index(); if (index_good != -1) { - char gran_factor, chunk_factor, lose_factor; - unsigned long gran_base, chunk_base, lose_base; - printk(KERN_INFO "Found optimal setting for mtrr clean up\n"); i = index_good; - gran_base = to_size_factor(result[i].gran_sizek, &gran_factor), - chunk_base = to_size_factor(result[i].chunk_sizek, &chunk_factor), - lose_base = to_size_factor(result[i].lose_cover_sizek, &lose_factor), - printk(KERN_INFO "gran_size: %ld%c \tchunk_size: %ld%c \t", - gran_base, gran_factor, chunk_base, chunk_factor); - printk(KERN_CONT "num_reg: %d \tlose RAM: %ld%c\n", - result[i].num_reg, lose_base, lose_factor); + mtrr_print_out_one_result(i); + /* convert ranges to var ranges state */ chunk_size = result[i].chunk_sizek; chunk_size <<= 10; gran_size = result[i].gran_sizek; gran_size <<= 10; - debug_print++; x86_setup_var_mtrrs(range, nr_range, chunk_size, gran_size); - debug_print--; set_var_mtrr_all(address_bits); + printk(KERN_DEBUG "New variable MTRRs\n"); + print_out_mtrr_range_state(); return 1; + } else { + /* print out all */ + for (i = 0; i < NUM_RESULT; i++) + mtrr_print_out_one_result(i); } printk(KERN_INFO "mtrr_cleanup: can not find optimal value\n"); @@ -1562,7 +1557,6 @@ int __init mtrr_trim_uncached_memory(unsigned long end_pfn) { unsigned long i, base, size, highest_pfn = 0, def, dummy; mtrr_type type; - int nr_range; u64 total_trim_size; /* extra one for all 0 */ diff --git a/arch/x86/kernel/cpu/mtrr/mtrr.h b/arch/x86/kernel/cpu/mtrr/mtrr.h index 2dc4ec656b2..ffd60409cc6 100644 --- a/arch/x86/kernel/cpu/mtrr/mtrr.h +++ b/arch/x86/kernel/cpu/mtrr/mtrr.h @@ -8,11 +8,6 @@ #define MTRRcap_MSR 0x0fe #define MTRRdefType_MSR 0x2ff -#define MTRRphysBase_MSR(reg) (0x200 + 2 * (reg)) -#define MTRRphysMask_MSR(reg) (0x200 + 2 * (reg) + 1) - -#define NUM_FIXED_RANGES 88 -#define MAX_VAR_RANGES 256 #define MTRRfix64K_00000_MSR 0x250 #define MTRRfix16K_80000_MSR 0x258 #define MTRRfix16K_A0000_MSR 0x259 @@ -29,11 +24,7 @@ #define MTRR_CHANGE_MASK_VARIABLE 0x02 #define MTRR_CHANGE_MASK_DEFTYPE 0x04 -/* In the Intel processor's MTRR interface, the MTRR type is always held in - an 8 bit field: */ -typedef u8 mtrr_type; - -extern unsigned int mtrr_usage_table[MAX_VAR_RANGES]; +extern unsigned int mtrr_usage_table[MTRR_MAX_VAR_RANGES]; struct mtrr_ops { u32 vendor; @@ -70,13 +61,6 @@ struct set_mtrr_context { u32 ccr3; }; -struct mtrr_var_range { - u32 base_lo; - u32 base_hi; - u32 mask_lo; - u32 mask_hi; -}; - void set_mtrr_done(struct set_mtrr_context *ctxt); void set_mtrr_cache_disable(struct set_mtrr_context *ctxt); void set_mtrr_prepare_save(struct set_mtrr_context *ctxt); diff --git a/arch/x86/kernel/cpu/vmware.c b/arch/x86/kernel/cpu/vmware.c new file mode 100644 index 00000000000..284c399e323 --- /dev/null +++ b/arch/x86/kernel/cpu/vmware.c @@ -0,0 +1,112 @@ +/* + * VMware Detection code. + * + * Copyright (C) 2008, VMware, Inc. + * Author : Alok N Kataria <akataria@vmware.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. + * + */ + +#include <linux/dmi.h> +#include <asm/div64.h> +#include <asm/vmware.h> + +#define CPUID_VMWARE_INFO_LEAF 0x40000000 +#define VMWARE_HYPERVISOR_MAGIC 0x564D5868 +#define VMWARE_HYPERVISOR_PORT 0x5658 + +#define VMWARE_PORT_CMD_GETVERSION 10 +#define VMWARE_PORT_CMD_GETHZ 45 + +#define VMWARE_PORT(cmd, eax, ebx, ecx, edx) \ + __asm__("inl (%%dx)" : \ + "=a"(eax), "=c"(ecx), "=d"(edx), "=b"(ebx) : \ + "0"(VMWARE_HYPERVISOR_MAGIC), \ + "1"(VMWARE_PORT_CMD_##cmd), \ + "2"(VMWARE_HYPERVISOR_PORT), "3"(UINT_MAX) : \ + "memory"); + +static inline int __vmware_platform(void) +{ + uint32_t eax, ebx, ecx, edx; + VMWARE_PORT(GETVERSION, eax, ebx, ecx, edx); + return eax != (uint32_t)-1 && ebx == VMWARE_HYPERVISOR_MAGIC; +} + +static unsigned long __vmware_get_tsc_khz(void) +{ + uint64_t tsc_hz; + uint32_t eax, ebx, ecx, edx; + + VMWARE_PORT(GETHZ, eax, ebx, ecx, edx); + + if (ebx == UINT_MAX) + return 0; + tsc_hz = eax | (((uint64_t)ebx) << 32); + do_div(tsc_hz, 1000); + BUG_ON(tsc_hz >> 32); + return tsc_hz; +} + +/* + * While checking the dmi string infomation, just checking the product + * serial key should be enough, as this will always have a VMware + * specific string when running under VMware hypervisor. + */ +int vmware_platform(void) +{ + if (cpu_has_hypervisor) { + unsigned int eax, ebx, ecx, edx; + char hyper_vendor_id[13]; + + cpuid(CPUID_VMWARE_INFO_LEAF, &eax, &ebx, &ecx, &edx); + memcpy(hyper_vendor_id + 0, &ebx, 4); + memcpy(hyper_vendor_id + 4, &ecx, 4); + memcpy(hyper_vendor_id + 8, &edx, 4); + hyper_vendor_id[12] = '\0'; + if (!strcmp(hyper_vendor_id, "VMwareVMware")) + return 1; + } else if (dmi_available && dmi_name_in_serial("VMware") && + __vmware_platform()) + return 1; + + return 0; +} + +unsigned long vmware_get_tsc_khz(void) +{ + BUG_ON(!vmware_platform()); + return __vmware_get_tsc_khz(); +} + +/* + * VMware hypervisor takes care of exporting a reliable TSC to the guest. + * Still, due to timing difference when running on virtual cpus, the TSC can + * be marked as unstable in some cases. For example, the TSC sync check at + * bootup can fail due to a marginal offset between vcpus' TSCs (though the + * TSCs do not drift from each other). Also, the ACPI PM timer clocksource + * is not suitable as a watchdog when running on a hypervisor because the + * kernel may miss a wrap of the counter if the vcpu is descheduled for a + * long time. To skip these checks at runtime we set these capability bits, + * so that the kernel could just trust the hypervisor with providing a + * reliable virtual TSC that is suitable for timekeeping. + */ +void __cpuinit vmware_set_feature_bits(struct cpuinfo_x86 *c) +{ + set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); + set_cpu_cap(c, X86_FEATURE_TSC_RELIABLE); +} diff --git a/arch/x86/kernel/cpuid.c b/arch/x86/kernel/cpuid.c index 72cefd1e649..2ac1f0c2beb 100644 --- a/arch/x86/kernel/cpuid.c +++ b/arch/x86/kernel/cpuid.c @@ -39,10 +39,10 @@ #include <linux/device.h> #include <linux/cpu.h> #include <linux/notifier.h> +#include <linux/uaccess.h> #include <asm/processor.h> #include <asm/msr.h> -#include <asm/uaccess.h> #include <asm/system.h> static struct class *cpuid_class; @@ -82,7 +82,7 @@ static loff_t cpuid_seek(struct file *file, loff_t offset, int orig) } static ssize_t cpuid_read(struct file *file, char __user *buf, - size_t count, loff_t * ppos) + size_t count, loff_t *ppos) { char __user *tmp = buf; struct cpuid_regs cmd; @@ -117,11 +117,11 @@ static int cpuid_open(struct inode *inode, struct file *file) unsigned int cpu; struct cpuinfo_x86 *c; int ret = 0; - + lock_kernel(); cpu = iminor(file->f_path.dentry->d_inode); - if (cpu >= NR_CPUS || !cpu_online(cpu)) { + if (cpu >= nr_cpu_ids || !cpu_online(cpu)) { ret = -ENXIO; /* No such CPU */ goto out; } diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index 26855381790..c689d19e35a 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -26,37 +26,21 @@ #include <linux/kdebug.h> #include <asm/smp.h> #include <asm/reboot.h> +#include <asm/virtext.h> #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 +49,28 @@ 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; -} + /* Disable VMX or SVM if needed. + * + * We need to disable virtualization on all CPUs. + * Having VMX or SVM enabled on any CPU may break rebooting + * after the kdump kernel has finished its task. + */ + cpu_emergency_vmxoff(); + cpu_emergency_svm_disable(); -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(); - - 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--; - } + nmi_shootdown_cpus(kdump_nmi_callback); - /* 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 +89,15 @@ 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(); + + /* Booting kdump kernel with VMX or SVM enabled won't work, + * because (among other limitations) we can't disable paging + * with the virt flags. + */ + cpu_emergency_vmxoff(); + cpu_emergency_svm_disable(); + 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 a2d1176c38e..da91701a234 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -6,14 +6,13 @@ * precise-event based sampling (PEBS). * * It manages: - * - per-thread and per-cpu allocation of BTS and PEBS - * - buffer memory allocation (optional) - * - buffer overflow handling + * - DS and BTS hardware configuration + * - buffer overflow handling (to be done) * - buffer access * - * It assumes: - * - get_task_struct on all parameter tasks - * - current is allowed to trace parameter tasks + * It does not do: + * - security checking (is the caller allowed to trace the task) + * - buffer allocation (memory accounting) * * * Copyright (C) 2007-2008 Intel Corporation. @@ -28,22 +27,69 @@ #include <linux/slab.h> #include <linux/sched.h> #include <linux/mm.h> +#include <linux/kernel.h> /* * The configuration for a particular DS hardware implementation. */ struct ds_configuration { - /* the size of the DS structure in bytes */ - unsigned char sizeof_ds; - /* the size of one pointer-typed field in the DS structure in bytes; - this covers the first 8 fields related to buffer management. */ + /* the name of the configuration */ + const char *name; + /* the size of one pointer-typed field in the DS structure and + in the BTS and PEBS buffers in bytes; + this covers the first 8 DS fields related to buffer management. */ unsigned char sizeof_field; /* the size of a BTS/PEBS record in bytes */ unsigned char sizeof_rec[2]; + /* a series of bit-masks to control various features indexed + * by enum ds_feature */ + unsigned long ctl[dsf_ctl_max]; }; -static struct ds_configuration ds_cfg; +static DEFINE_PER_CPU(struct ds_configuration, ds_cfg_array); +#define ds_cfg per_cpu(ds_cfg_array, smp_processor_id()) + +#define MAX_SIZEOF_DS (12 * 8) /* maximal size of a DS configuration */ +#define MAX_SIZEOF_BTS (3 * 8) /* maximal size of a BTS record */ +#define DS_ALIGNMENT (1 << 3) /* BTS and PEBS buffer alignment */ + +#define BTS_CONTROL \ + (ds_cfg.ctl[dsf_bts] | ds_cfg.ctl[dsf_bts_kernel] | ds_cfg.ctl[dsf_bts_user] |\ + ds_cfg.ctl[dsf_bts_overflow]) + + +/* + * 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; + /* the trace including the DS configuration */ + struct bts_trace trace; + /* buffer overflow notification function */ + bts_ovfl_callback_t ovfl; +}; + +struct pebs_tracer { + /* the common DS part */ + struct ds_tracer ds; + /* the trace including the DS configuration */ + struct pebs_trace trace; + /* buffer overflow notification function */ + pebs_ovfl_callback_t ovfl; +}; /* * Debug Store (DS) save area configuration (see Intel64 and IA32 @@ -109,32 +155,9 @@ static inline void ds_set(unsigned char *base, enum ds_qualifier qual, /* - * 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. + * Locking is done only for allocating BTS or PEBS resources. */ -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. - */ -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 DEFINE_SPINLOCK(ds_lock); /* @@ -150,27 +173,32 @@ static inline int ds_validate_access(struct ds_context *context, * >0 number of per-thread tracers * <0 number of per-cpu tracers * - * The below functions to get and put tracers and to check the - * allocation type require the ds_lock to be held by the caller. - * * Tracers essentially gives the number of ds contexts for a certain * type of allocation. */ -static long tracers; +static atomic_t tracers = ATOMIC_INIT(0); static inline void get_tracer(struct task_struct *task) { - tracers += (task ? 1 : -1); + if (task) + atomic_inc(&tracers); + else + atomic_dec(&tracers); } static inline void put_tracer(struct task_struct *task) { - tracers -= (task ? 1 : -1); + if (task) + atomic_dec(&tracers); + else + atomic_inc(&tracers); } static inline int check_tracer(struct task_struct *task) { - return (task ? (tracers >= 0) : (tracers <= 0)); + return task ? + (atomic_read(&tracers) >= 0) : + (atomic_read(&tracers) <= 0); } @@ -183,99 +211,70 @@ 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; - unsigned long irq; +struct ds_context { + /* pointer to the DS configuration; goes into MSR_IA32_DS_AREA */ + unsigned char ds[MAX_SIZEOF_DS]; + /* the owner of the BTS and PEBS configuration, respectively */ + struct bts_tracer *bts_master; + struct pebs_tracer *pebs_master; + /* use count */ + unsigned long count; + /* a pointer to the context location inside the thread_struct + * or the per_cpu context array */ + struct ds_context **this; + /* a pointer to the task owning this context, or NULL, if the + * context is owned by a cpu */ + struct task_struct *task; +}; - spin_lock_irqsave(&ds_lock, irq); +static DEFINE_PER_CPU(struct ds_context *, system_context_array); - context = (task ? task->thread.ds_ctx : this_system_context); - if (context) - context->count++; +#define system_context per_cpu(system_context_array, smp_processor_id()) - spin_unlock_irqrestore(&ds_lock, irq); - - return context; -} -/* - * Same as ds_get_context, but allocates the context and it's DS - * structure, if necessary; returns NULL; if out of memory. - */ -static inline struct ds_context *ds_alloc_context(struct task_struct *task) +static inline struct ds_context *ds_get_context(struct task_struct *task) { struct ds_context **p_context = - (task ? &task->thread.ds_ctx : &this_system_context); - struct ds_context *context = *p_context; + (task ? &task->thread.ds_ctx : &system_context); + struct ds_context *context = NULL; + struct ds_context *new_context = NULL; 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; - } + /* Chances are small that we already have a context. */ + new_context = kzalloc(sizeof(*new_context), GFP_KERNEL); + if (!new_context) + return NULL; - spin_lock_irqsave(&ds_lock, irq); + spin_lock_irqsave(&ds_lock, irq); - if (*p_context) { - kfree(context->ds); - kfree(context); + context = *p_context; + if (!context) { + context = new_context; - context = *p_context; - } else { - *p_context = context; + context->this = p_context; + context->task = task; + context->count = 0; - context->this = p_context; - context->task = task; + if (task) + set_tsk_thread_flag(task, TIF_DS_AREA_MSR); - if (task) - set_tsk_thread_flag(task, TIF_DS_AREA_MSR); + if (!task || (task == current)) + wrmsrl(MSR_IA32_DS_AREA, (unsigned long)context->ds); - if (!task || (task == current)) - wrmsrl(MSR_IA32_DS_AREA, - (unsigned long)context->ds); - } - spin_unlock_irqrestore(&ds_lock, irq); + *p_context = context; } context->count++; + spin_unlock_irqrestore(&ds_lock, irq); + + if (context != new_context) + kfree(new_context); + 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; @@ -285,8 +284,10 @@ static inline void ds_put_context(struct ds_context *context) spin_lock_irqsave(&ds_lock, irq); - if (--context->count) - goto out; + if (--context->count) { + spin_unlock_irqrestore(&ds_lock, irq); + return; + } *(context->this) = NULL; @@ -296,135 +297,263 @@ 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); + spin_unlock_irqrestore(&ds_lock, irq); - /* 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_irqrestore(&ds_lock, irq); } /* - * Handle a buffer overflow + * Call the tracer's callback on 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) +static void ds_overflow(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 */ + switch (qual) { + case ds_bts: + if (context->bts_master && + context->bts_master->ovfl) + context->bts_master->ovfl(context->bts_master); + break; + case ds_pebs: + if (context->pebs_master && + context->pebs_master->ovfl) + context->pebs_master->ovfl(context->pebs_master); + break; + } } /* - * Allocate a non-pageable buffer of the parameter size. - * Checks the memory and the locked memory rlimit. + * Write raw data into the BTS or PEBS buffer. * - * Returns the buffer, if successful; - * NULL, if out of memory or rlimit exceeded. + * The remainder of any partially written record is zeroed out. * - * size: the requested buffer size in bytes - * pages (out): if not NULL, contains the number of pages reserved + * context: the DS context + * qual: the buffer type + * record: the data to write + * size: the size of the data */ -static inline void *ds_allocate_buffer(size_t size, unsigned int *pages) +static int ds_write(struct ds_context *context, enum ds_qualifier qual, + const void *record, size_t size) { - unsigned long rlim, vm, pgsz; - void *buffer; + int bytes_written = 0; - pgsz = PAGE_ALIGN(size) >> PAGE_SHIFT; + if (!record) + return -EINVAL; - rlim = current->signal->rlim[RLIMIT_AS].rlim_cur >> PAGE_SHIFT; - vm = current->mm->total_vm + pgsz; - if (rlim < vm) - return NULL; + while (size) { + unsigned long base, index, end, write_end, int_th; + unsigned long write_size, adj_write_size; - rlim = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT; - vm = current->mm->locked_vm + pgsz; - if (rlim < vm) - return NULL; + /* + * write as much as possible without producing an + * overflow interrupt. + * + * interrupt_threshold must either be + * - bigger than absolute_maximum or + * - point to a record between buffer_base and absolute_maximum + * + * index points to a valid record. + */ + base = ds_get(context->ds, qual, ds_buffer_base); + index = ds_get(context->ds, qual, ds_index); + end = ds_get(context->ds, qual, ds_absolute_maximum); + int_th = ds_get(context->ds, qual, ds_interrupt_threshold); - buffer = kzalloc(size, GFP_KERNEL); - if (!buffer) - return NULL; + write_end = min(end, int_th); - current->mm->total_vm += pgsz; - current->mm->locked_vm += pgsz; + /* if we are already beyond the interrupt threshold, + * we fill the entire buffer */ + if (write_end <= index) + write_end = end; - if (pages) - *pages = pgsz; + if (write_end <= index) + break; + + write_size = min((unsigned long) size, write_end - index); + memcpy((void *)index, record, write_size); - return buffer; + record = (const char *)record + 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]; + + /* zero out trailing bytes */ + memset((char *)index + write_size, 0, + adj_write_size - write_size); + index += adj_write_size; + + if (index >= end) + index = base; + ds_set(context->ds, qual, ds_index, index); + + if (index >= int_th) + ds_overflow(context, qual); + } + + return bytes_written; } -static int ds_request(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl, enum ds_qualifier qual) + +/* + * Branch Trace Store (BTS) uses the following format. Different + * architectures vary in the size of those fields. + * - source linear address + * - destination linear address + * - flags + * + * Later architectures use 64bit pointers throughout, whereas earlier + * architectures use 32bit pointers in 32bit mode. + * + * We compute the base address for the first 8 fields based on: + * - the field size stored in the DS configuration + * - the relative field position + * + * In order to store additional information in the BTS buffer, we use + * a special source address to indicate that the record requires + * special interpretation. + * + * Netburst indicated via a bit in the flags field whether the branch + * was predicted; this is ignored. + * + * We use two levels of abstraction: + * - the raw data level defined here + * - an arch-independent level defined in ds.h + */ + +enum bts_field { + bts_from, + bts_to, + bts_flags, + + bts_qual = bts_from, + bts_jiffies = bts_to, + bts_pid = bts_flags, + + bts_qual_mask = (bts_qual_max - 1), + bts_escape = ((unsigned long)-1 & ~bts_qual_mask) +}; + +static inline unsigned long bts_get(const char *base, enum bts_field field) { - struct ds_context *context; - unsigned long buffer, adj; - const unsigned long alignment = (1 << 3); - unsigned long irq; - int error = 0; + base += (ds_cfg.sizeof_field * field); + return *(unsigned long *)base; +} + +static inline void bts_set(char *base, enum bts_field field, unsigned long val) +{ + base += (ds_cfg.sizeof_field * field);; + (*(unsigned long *)base) = val; +} - if (!ds_cfg.sizeof_ds) - return -EOPNOTSUPP; - /* we require some space to do alignment adjustments below */ - if (size < (alignment + ds_cfg.sizeof_rec[qual])) +/* + * The raw BTS data is architecture dependent. + * + * For higher-level users, we give an arch-independent view. + * - ds.h defines struct bts_struct + * - bts_read translates one raw bts record into a bts_struct + * - bts_write translates one bts_struct into the raw format and + * writes it into the top of the parameter tracer's buffer. + * + * return: bytes read/written on success; -Eerrno, otherwise + */ +static int bts_read(struct bts_tracer *tracer, const void *at, + struct bts_struct *out) +{ + if (!tracer) return -EINVAL; - /* buffer overflow notification is not yet implemented */ - if (ovfl) - return -EOPNOTSUPP; + if (at < tracer->trace.ds.begin) + return -EINVAL; + if (tracer->trace.ds.end < (at + tracer->trace.ds.size)) + return -EINVAL; - context = ds_alloc_context(task); - if (!context) - return -ENOMEM; + memset(out, 0, sizeof(*out)); + if ((bts_get(at, bts_qual) & ~bts_qual_mask) == bts_escape) { + out->qualifier = (bts_get(at, bts_qual) & bts_qual_mask); + out->variant.timestamp.jiffies = bts_get(at, bts_jiffies); + out->variant.timestamp.pid = bts_get(at, bts_pid); + } else { + out->qualifier = bts_branch; + out->variant.lbr.from = bts_get(at, bts_from); + out->variant.lbr.to = bts_get(at, bts_to); + + if (!out->variant.lbr.from && !out->variant.lbr.to) + out->qualifier = bts_invalid; + } - spin_lock_irqsave(&ds_lock, irq); + return ds_cfg.sizeof_rec[ds_bts]; +} - error = -EPERM; - if (!check_tracer(task)) - goto out_unlock; +static int bts_write(struct bts_tracer *tracer, const struct bts_struct *in) +{ + unsigned char raw[MAX_SIZEOF_BTS]; - get_tracer(task); + if (!tracer) + return -EINVAL; - error = -EALREADY; - if (context->owner[qual] == current) - goto out_put_tracer; - error = -EPERM; - if (context->owner[qual] != NULL) - goto out_put_tracer; - context->owner[qual] = current; + if (MAX_SIZEOF_BTS < ds_cfg.sizeof_rec[ds_bts]) + return -EOVERFLOW; - spin_unlock_irqrestore(&ds_lock, irq); + switch (in->qualifier) { + case bts_invalid: + bts_set(raw, bts_from, 0); + bts_set(raw, bts_to, 0); + bts_set(raw, bts_flags, 0); + break; + case bts_branch: + bts_set(raw, bts_from, in->variant.lbr.from); + bts_set(raw, bts_to, in->variant.lbr.to); + bts_set(raw, bts_flags, 0); + break; + case bts_task_arrives: + case bts_task_departs: + bts_set(raw, bts_qual, (bts_escape | in->qualifier)); + bts_set(raw, bts_jiffies, in->variant.timestamp.jiffies); + bts_set(raw, bts_pid, in->variant.timestamp.pid); + break; + default: + return -EINVAL; + } + return ds_write(tracer->ds.context, ds_bts, raw, + ds_cfg.sizeof_rec[ds_bts]); +} - error = -ENOMEM; - if (!base) { - base = ds_allocate_buffer(size, &context->pages[qual]); - if (!base) - goto out_release; - context->buffer[qual] = base; - } - error = 0; +static void ds_write_config(struct ds_context *context, + struct ds_trace *cfg, enum ds_qualifier qual) +{ + unsigned char *ds = context->ds; + + ds_set(ds, qual, ds_buffer_base, (unsigned long)cfg->begin); + ds_set(ds, qual, ds_index, (unsigned long)cfg->top); + ds_set(ds, qual, ds_absolute_maximum, (unsigned long)cfg->end); + ds_set(ds, qual, ds_interrupt_threshold, (unsigned long)cfg->ith); +} + +static void ds_read_config(struct ds_context *context, + struct ds_trace *cfg, enum ds_qualifier qual) +{ + unsigned char *ds = context->ds; - context->callback[qual] = ovfl; + cfg->begin = (void *)ds_get(ds, qual, ds_buffer_base); + cfg->top = (void *)ds_get(ds, qual, ds_index); + cfg->end = (void *)ds_get(ds, qual, ds_absolute_maximum); + cfg->ith = (void *)ds_get(ds, qual, ds_interrupt_threshold); +} + +static void ds_init_ds_trace(struct ds_trace *trace, enum ds_qualifier qual, + void *base, size_t size, size_t ith, + unsigned int flags) { + unsigned long buffer, adj; /* adjust the buffer address and size to meet alignment * constraints: @@ -436,410 +565,383 @@ static int ds_request(struct task_struct *task, void *base, size_t size, */ buffer = (unsigned long)base; - adj = ALIGN(buffer, alignment) - buffer; + adj = ALIGN(buffer, DS_ALIGNMENT) - buffer; buffer += adj; size -= adj; - size /= ds_cfg.sizeof_rec[qual]; - size *= ds_cfg.sizeof_rec[qual]; - - 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); + trace->n = size / ds_cfg.sizeof_rec[qual]; + trace->size = ds_cfg.sizeof_rec[qual]; - if (ovfl) { - /* todo: select a suitable interrupt threshold */ - } else - ds_set(context->ds, qual, - ds_interrupt_threshold, buffer + size + 1); + size = (trace->n * trace->size); - /* we keep the context until ds_release */ - return error; - - out_release: - context->owner[qual] = NULL; - ds_put_context(context); - put_tracer(task); - return error; - - out_put_tracer: - spin_unlock_irqrestore(&ds_lock, irq); - ds_put_context(context); - put_tracer(task); - return error; + trace->begin = (void *)buffer; + trace->top = trace->begin; + trace->end = (void *)(buffer + size); + /* The value for 'no threshold' is -1, which will set the + * threshold outside of the buffer, just like we want it. + */ + trace->ith = (void *)(buffer + size - ith); - out_unlock: - spin_unlock_irqrestore(&ds_lock, irq); - ds_put_context(context); - return error; + trace->flags = flags; } -int ds_request_bts(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl) -{ - return ds_request(task, base, size, ovfl, ds_bts); -} -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); -} - -static int ds_release(struct task_struct *task, enum ds_qualifier qual) +static int ds_request(struct ds_tracer *tracer, struct ds_trace *trace, + enum ds_qualifier qual, struct task_struct *task, + void *base, size_t size, size_t th, unsigned int flags) { struct ds_context *context; int error; - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) + error = -EINVAL; + if (!base) goto out; - kfree(context->buffer[qual]); - context->buffer[qual] = NULL; - - current->mm->total_vm -= context->pages[qual]; - current->mm->locked_vm -= context->pages[qual]; - context->pages[qual] = 0; - context->owner[qual] = NULL; - - /* - * we put the context twice: - * once for the ds_get_context - * once for the corresponding ds_request - */ - ds_put_context(context); - out: - ds_put_context(context); - return error; -} + /* we require some space to do alignment adjustments below */ + error = -EINVAL; + if (size < (DS_ALIGNMENT + ds_cfg.sizeof_rec[qual])) + goto out; -int ds_release_bts(struct task_struct *task) -{ - return ds_release(task, ds_bts); -} + if (th != (size_t)-1) { + th *= ds_cfg.sizeof_rec[qual]; -int ds_release_pebs(struct task_struct *task) -{ - return ds_release(task, ds_pebs); -} + error = -EINVAL; + if (size <= th) + goto out; + } -static int ds_get_index(struct task_struct *task, size_t *pos, - enum ds_qualifier qual) -{ - struct ds_context *context; - unsigned long base, index; - int error; + tracer->buffer = base; + tracer->size = size; + error = -ENOMEM; context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) + if (!context) goto out; + tracer->context = context; - base = ds_get(context->ds, qual, ds_buffer_base); - index = ds_get(context->ds, qual, ds_index); + ds_init_ds_trace(trace, qual, base, size, th, flags); - error = ((index - base) / ds_cfg.sizeof_rec[qual]); - if (pos) - *pos = error; + error = 0; out: - ds_put_context(context); return error; } -int ds_get_bts_index(struct task_struct *task, size_t *pos) -{ - return ds_get_index(task, pos, ds_bts); -} - -int ds_get_pebs_index(struct task_struct *task, size_t *pos) +struct bts_tracer *ds_request_bts(struct task_struct *task, + void *base, size_t size, + bts_ovfl_callback_t ovfl, size_t th, + unsigned int flags) { - return ds_get_index(task, pos, ds_pebs); -} - -static int ds_get_end(struct task_struct *task, size_t *pos, - enum ds_qualifier qual) -{ - struct ds_context *context; - unsigned long base, end; + struct bts_tracer *tracer; + unsigned long irq; int error; - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) + error = -EOPNOTSUPP; + if (!ds_cfg.ctl[dsf_bts]) goto out; - base = ds_get(context->ds, qual, ds_buffer_base); - end = ds_get(context->ds, qual, ds_absolute_maximum); + /* buffer overflow notification is not yet implemented */ + error = -EOPNOTSUPP; + if (ovfl) + goto out; - error = ((end - base) / ds_cfg.sizeof_rec[qual]); - if (pos) - *pos = error; - out: - ds_put_context(context); - return error; -} + error = -ENOMEM; + tracer = kzalloc(sizeof(*tracer), GFP_KERNEL); + if (!tracer) + goto out; + tracer->ovfl = ovfl; -int ds_get_bts_end(struct task_struct *task, size_t *pos) -{ - return ds_get_end(task, pos, ds_bts); -} + error = ds_request(&tracer->ds, &tracer->trace.ds, + ds_bts, task, base, size, th, flags); + if (error < 0) + goto out_tracer; -int ds_get_pebs_end(struct task_struct *task, size_t *pos) -{ - return ds_get_end(task, pos, ds_pebs); -} -static int ds_access(struct task_struct *task, size_t index, - const void **record, enum ds_qualifier qual) -{ - struct ds_context *context; - unsigned long base, idx; - int error; + spin_lock_irqsave(&ds_lock, irq); - if (!record) - return -EINVAL; + error = -EPERM; + if (!check_tracer(task)) + goto out_unlock; + get_tracer(task); - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; + error = -EPERM; + if (tracer->ds.context->bts_master) + goto out_put_tracer; + tracer->ds.context->bts_master = tracer; - base = ds_get(context->ds, qual, ds_buffer_base); - idx = base + (index * ds_cfg.sizeof_rec[qual]); + spin_unlock_irqrestore(&ds_lock, irq); - error = -EINVAL; - if (idx > ds_get(context->ds, qual, ds_absolute_maximum)) - goto out; - *record = (const void *)idx; - error = ds_cfg.sizeof_rec[qual]; - out: - ds_put_context(context); - return error; -} + tracer->trace.read = bts_read; + tracer->trace.write = bts_write; -int ds_access_bts(struct task_struct *task, size_t index, const void **record) -{ - return ds_access(task, index, record, ds_bts); -} + ds_write_config(tracer->ds.context, &tracer->trace.ds, ds_bts); + ds_resume_bts(tracer); -int ds_access_pebs(struct task_struct *task, size_t index, const void **record) -{ - return ds_access(task, index, record, ds_pebs); + return tracer; + + out_put_tracer: + put_tracer(task); + out_unlock: + spin_unlock_irqrestore(&ds_lock, irq); + ds_put_context(tracer->ds.context); + out_tracer: + kfree(tracer); + out: + return ERR_PTR(error); } -static int ds_write(struct task_struct *task, const void *record, size_t size, - enum ds_qualifier qual, int force) +struct pebs_tracer *ds_request_pebs(struct task_struct *task, + void *base, size_t size, + pebs_ovfl_callback_t ovfl, size_t th, + unsigned int flags) { - struct ds_context *context; + struct pebs_tracer *tracer; + unsigned long irq; int error; - if (!record) - return -EINVAL; + /* buffer overflow notification is not yet implemented */ + error = -EOPNOTSUPP; + if (ovfl) + goto out; - error = -EPERM; - context = ds_get_context(task); - if (!context) + error = -ENOMEM; + tracer = kzalloc(sizeof(*tracer), GFP_KERNEL); + if (!tracer) goto out; + tracer->ovfl = ovfl; - if (!force) { - error = ds_validate_access(context, qual); - if (error < 0) - goto out; - } + error = ds_request(&tracer->ds, &tracer->trace.ds, + ds_pebs, task, base, size, th, flags); + if (error < 0) + goto out_tracer; - error = 0; - while (size) { - unsigned long base, index, end, write_end, int_th; - unsigned long write_size, adj_write_size; + spin_lock_irqsave(&ds_lock, irq); - /* - * write as much as possible without producing an - * overflow interrupt. - * - * interrupt_threshold must either be - * - bigger than absolute_maximum or - * - point to a record between buffer_base and absolute_maximum - * - * index points to a valid record. - */ - base = ds_get(context->ds, qual, ds_buffer_base); - index = ds_get(context->ds, qual, ds_index); - end = ds_get(context->ds, qual, ds_absolute_maximum); - int_th = ds_get(context->ds, qual, ds_interrupt_threshold); + error = -EPERM; + if (!check_tracer(task)) + goto out_unlock; + get_tracer(task); - write_end = min(end, int_th); + error = -EPERM; + if (tracer->ds.context->pebs_master) + goto out_put_tracer; + tracer->ds.context->pebs_master = tracer; - /* if we are already beyond the interrupt threshold, - * we fill the entire buffer */ - if (write_end <= index) - write_end = end; + spin_unlock_irqrestore(&ds_lock, irq); - if (write_end <= index) - goto out; + ds_write_config(tracer->ds.context, &tracer->trace.ds, ds_bts); + ds_resume_pebs(tracer); - write_size = min((unsigned long) size, write_end - index); - memcpy((void *)index, record, write_size); + return tracer; - record = (const char *)record + write_size; - size -= write_size; - error += write_size; + out_put_tracer: + put_tracer(task); + out_unlock: + spin_unlock_irqrestore(&ds_lock, irq); + ds_put_context(tracer->ds.context); + out_tracer: + kfree(tracer); + out: + return ERR_PTR(error); +} - adj_write_size = write_size / ds_cfg.sizeof_rec[qual]; - adj_write_size *= ds_cfg.sizeof_rec[qual]; +void ds_release_bts(struct bts_tracer *tracer) +{ + if (!tracer) + return; - /* zero out trailing bytes */ - memset((char *)index + write_size, 0, - adj_write_size - write_size); - index += adj_write_size; + ds_suspend_bts(tracer); - if (index >= end) - index = base; - ds_set(context->ds, qual, ds_index, index); + WARN_ON_ONCE(tracer->ds.context->bts_master != tracer); + tracer->ds.context->bts_master = NULL; - if (index >= int_th) - ds_overflow(task, context, qual); - } + put_tracer(tracer->ds.context->task); + ds_put_context(tracer->ds.context); - out: - ds_put_context(context); - return error; + kfree(tracer); } -int ds_write_bts(struct task_struct *task, const void *record, size_t size) +void ds_suspend_bts(struct bts_tracer *tracer) { - return ds_write(task, record, size, ds_bts, /* force = */ 0); -} + struct task_struct *task; -int ds_write_pebs(struct task_struct *task, const void *record, size_t size) -{ - return ds_write(task, record, size, ds_pebs, /* force = */ 0); -} + if (!tracer) + return; -int ds_unchecked_write_bts(struct task_struct *task, - const void *record, size_t size) -{ - return ds_write(task, record, size, ds_bts, /* force = */ 1); -} + task = tracer->ds.context->task; -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); + if (!task || (task == current)) + update_debugctlmsr(get_debugctlmsr() & ~BTS_CONTROL); + + if (task) { + task->thread.debugctlmsr &= ~BTS_CONTROL; + + if (!task->thread.debugctlmsr) + clear_tsk_thread_flag(task, TIF_DEBUGCTLMSR); + } } -static int ds_reset_or_clear(struct task_struct *task, - enum ds_qualifier qual, int clear) +void ds_resume_bts(struct bts_tracer *tracer) { - struct ds_context *context; - unsigned long base, end; - int error; + struct task_struct *task; + unsigned long control; - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; + if (!tracer) + return; - base = ds_get(context->ds, qual, ds_buffer_base); - end = ds_get(context->ds, qual, ds_absolute_maximum); + task = tracer->ds.context->task; - if (clear) - memset((void *)base, 0, end - base); + control = ds_cfg.ctl[dsf_bts]; + if (!(tracer->trace.ds.flags & BTS_KERNEL)) + control |= ds_cfg.ctl[dsf_bts_kernel]; + if (!(tracer->trace.ds.flags & BTS_USER)) + control |= ds_cfg.ctl[dsf_bts_user]; - ds_set(context->ds, qual, ds_index, base); + if (task) { + task->thread.debugctlmsr |= control; + set_tsk_thread_flag(task, TIF_DEBUGCTLMSR); + } - error = 0; - out: - ds_put_context(context); - return error; + if (!task || (task == current)) + update_debugctlmsr(get_debugctlmsr() | control); } -int ds_reset_bts(struct task_struct *task) +void ds_release_pebs(struct pebs_tracer *tracer) { - return ds_reset_or_clear(task, ds_bts, /* clear = */ 0); + if (!tracer) + return; + + ds_suspend_pebs(tracer); + + WARN_ON_ONCE(tracer->ds.context->pebs_master != tracer); + tracer->ds.context->pebs_master = NULL; + + put_tracer(tracer->ds.context->task); + ds_put_context(tracer->ds.context); + + kfree(tracer); } -int ds_reset_pebs(struct task_struct *task) +void ds_suspend_pebs(struct pebs_tracer *tracer) { - return ds_reset_or_clear(task, ds_pebs, /* clear = */ 0); + } -int ds_clear_bts(struct task_struct *task) +void ds_resume_pebs(struct pebs_tracer *tracer) { - return ds_reset_or_clear(task, ds_bts, /* clear = */ 1); + } -int ds_clear_pebs(struct task_struct *task) +const struct bts_trace *ds_read_bts(struct bts_tracer *tracer) { - return ds_reset_or_clear(task, ds_pebs, /* clear = */ 1); + if (!tracer) + return NULL; + + ds_read_config(tracer->ds.context, &tracer->trace.ds, ds_bts); + return &tracer->trace; } -int ds_get_pebs_reset(struct task_struct *task, u64 *value) +const struct pebs_trace *ds_read_pebs(struct pebs_tracer *tracer) { - struct ds_context *context; - int error; + if (!tracer) + return NULL; + + ds_read_config(tracer->ds.context, &tracer->trace.ds, ds_pebs); + tracer->trace.reset_value = + *(u64 *)(tracer->ds.context->ds + (ds_cfg.sizeof_field * 8)); - if (!value) + return &tracer->trace; +} + +int ds_reset_bts(struct bts_tracer *tracer) +{ + if (!tracer) return -EINVAL; - context = ds_get_context(task); - error = ds_validate_access(context, ds_pebs); - if (error < 0) - goto out; + tracer->trace.ds.top = tracer->trace.ds.begin; - *value = *(u64 *)(context->ds + (ds_cfg.sizeof_field * 8)); + ds_set(tracer->ds.context->ds, ds_bts, ds_index, + (unsigned long)tracer->trace.ds.top); - error = 0; - out: - ds_put_context(context); - return error; + return 0; } -int ds_set_pebs_reset(struct task_struct *task, u64 value) +int ds_reset_pebs(struct pebs_tracer *tracer) { - struct ds_context *context; - int error; + if (!tracer) + return -EINVAL; - context = ds_get_context(task); - error = ds_validate_access(context, ds_pebs); - if (error < 0) - goto out; + tracer->trace.ds.top = tracer->trace.ds.begin; - *(u64 *)(context->ds + (ds_cfg.sizeof_field * 8)) = value; + ds_set(tracer->ds.context->ds, ds_bts, ds_index, + (unsigned long)tracer->trace.ds.top); - error = 0; - out: - ds_put_context(context); - return error; + return 0; +} + +int ds_set_pebs_reset(struct pebs_tracer *tracer, u64 value) +{ + if (!tracer) + return -EINVAL; + + *(u64 *)(tracer->ds.context->ds + (ds_cfg.sizeof_field * 8)) = value; + + 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, +static const struct ds_configuration ds_cfg_netburst = { + .name = "netburst", + .ctl[dsf_bts] = (1 << 2) | (1 << 3), + .ctl[dsf_bts_kernel] = (1 << 5), + .ctl[dsf_bts_user] = (1 << 6), + + .sizeof_field = sizeof(long), + .sizeof_rec[ds_bts] = sizeof(long) * 3, #ifdef __i386__ - .sizeof_rec[ds_pebs] = sizeof(long) * 10 + .sizeof_rec[ds_pebs] = sizeof(long) * 10, #else - .sizeof_rec[ds_pebs] = sizeof(long) * 18 + .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, +static const struct ds_configuration ds_cfg_pentium_m = { + .name = "pentium m", + .ctl[dsf_bts] = (1 << 6) | (1 << 7), + + .sizeof_field = sizeof(long), + .sizeof_rec[ds_bts] = sizeof(long) * 3, #ifdef __i386__ - .sizeof_rec[ds_pebs] = 8 * 10 + .sizeof_rec[ds_pebs] = sizeof(long) * 10, #else - .sizeof_rec[ds_pebs] = 8 * 18 + .sizeof_rec[ds_pebs] = sizeof(long) * 18, #endif }; +static const struct ds_configuration ds_cfg_core2 = { + .name = "core 2", + .ctl[dsf_bts] = (1 << 6) | (1 << 7), + .ctl[dsf_bts_kernel] = (1 << 9), + .ctl[dsf_bts_user] = (1 << 10), + + .sizeof_field = 8, + .sizeof_rec[ds_bts] = 8 * 3, + .sizeof_rec[ds_pebs] = 8 * 18, +}; -static inline void +static void ds_configure(const struct ds_configuration *cfg) { + memset(&ds_cfg, 0, sizeof(ds_cfg)); ds_cfg = *cfg; + + printk(KERN_INFO "[ds] using %s configuration\n", ds_cfg.name); + + if (!cpu_has_bts) { + ds_cfg.ctl[dsf_bts] = 0; + printk(KERN_INFO "[ds] bts not available\n"); + } + if (!cpu_has_pebs) + printk(KERN_INFO "[ds] pebs not available\n"); + + WARN_ON_ONCE(MAX_SIZEOF_DS < (12 * ds_cfg.sizeof_field)); } void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) @@ -847,16 +949,15 @@ 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); + ds_configure(&ds_cfg_pentium_m); break; - case 0xF: /* Core2 */ - case 0x1C: /* Atom */ - ds_configure(&ds_cfg_64); - break; - default: - /* sorry, don't know about them */ + default: /* Core2, Atom, ... */ + ds_configure(&ds_cfg_core2); break; } break; @@ -865,7 +966,7 @@ void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) case 0x0: case 0x1: case 0x2: /* Netburst */ - ds_configure(&ds_cfg_var); + ds_configure(&ds_cfg_netburst); break; default: /* sorry, don't know about them */ @@ -878,12 +979,52 @@ void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) } } -void ds_free(struct ds_context *context) +/* + * Change the DS configuration from tracing prev to tracing next. + */ +void ds_switch_to(struct task_struct *prev, struct task_struct *next) +{ + struct ds_context *prev_ctx = prev->thread.ds_ctx; + struct ds_context *next_ctx = next->thread.ds_ctx; + + if (prev_ctx) { + update_debugctlmsr(0); + + if (prev_ctx->bts_master && + (prev_ctx->bts_master->trace.ds.flags & BTS_TIMESTAMPS)) { + struct bts_struct ts = { + .qualifier = bts_task_departs, + .variant.timestamp.jiffies = jiffies_64, + .variant.timestamp.pid = prev->pid + }; + bts_write(prev_ctx->bts_master, &ts); + } + } + + if (next_ctx) { + if (next_ctx->bts_master && + (next_ctx->bts_master->trace.ds.flags & BTS_TIMESTAMPS)) { + struct bts_struct ts = { + .qualifier = bts_task_arrives, + .variant.timestamp.jiffies = jiffies_64, + .variant.timestamp.pid = next->pid + }; + bts_write(next_ctx->bts_master, &ts); + } + + wrmsrl(MSR_IA32_DS_AREA, (unsigned long)next_ctx->ds); + } + + update_debugctlmsr(next->thread.debugctlmsr); +} + +void ds_copy_thread(struct task_struct *tsk, struct task_struct *father) +{ + clear_tsk_thread_flag(tsk, TIF_DS_AREA_MSR); + tsk->thread.ds_ctx = NULL; +} + +void ds_exit_thread(struct task_struct *tsk) { - /* This is called when the task owning the parameter context - * is dying. There should not be any user of that context left - * to disturb us, anymore. */ - unsigned long leftovers = context->count; - while (leftovers--) - ds_put_context(context); + WARN_ON(tsk->thread.ds_ctx); } 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/e820.c b/arch/x86/kernel/e820.c index 7aafeb5263e..65a13943e09 100644 --- a/arch/x86/kernel/e820.c +++ b/arch/x86/kernel/e820.c @@ -677,22 +677,6 @@ struct early_res { }; static struct early_res early_res[MAX_EARLY_RES] __initdata = { { 0, PAGE_SIZE, "BIOS data page" }, /* BIOS data page */ -#if defined(CONFIG_X86_64) && defined(CONFIG_X86_TRAMPOLINE) - { TRAMPOLINE_BASE, TRAMPOLINE_BASE + 2 * PAGE_SIZE, "TRAMPOLINE" }, -#endif -#if defined(CONFIG_X86_32) && defined(CONFIG_SMP) - /* - * But first pinch a few for the stack/trampoline stuff - * FIXME: Don't need the extra page at 4K, but need to fix - * trampoline before removing it. (see the GDT stuff) - */ - { PAGE_SIZE, PAGE_SIZE + PAGE_SIZE, "EX TRAMPOLINE" }, - /* - * Has to be in very low memory so we can execute - * real-mode AP code. - */ - { TRAMPOLINE_BASE, TRAMPOLINE_BASE + PAGE_SIZE, "TRAMPOLINE" }, -#endif {} }; diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index 1b894b72c0f..744aa7fc49d 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -17,6 +17,7 @@ #include <asm/io_apic.h> #include <asm/apic.h> #include <asm/iommu.h> +#include <asm/gart.h> static void __init fix_hypertransport_config(int num, int slot, int func) { diff --git a/arch/x86/kernel/early_printk.c b/arch/x86/kernel/early_printk.c index 34ad997d383..504ad198e4a 100644 --- a/arch/x86/kernel/early_printk.c +++ b/arch/x86/kernel/early_printk.c @@ -875,49 +875,6 @@ static struct console early_dbgp_console = { }; #endif -/* Console interface to a host file on AMD's SimNow! */ - -static int simnow_fd; - -enum { - MAGIC1 = 0xBACCD00A, - MAGIC2 = 0xCA110000, - XOPEN = 5, - XWRITE = 4, -}; - -static noinline long simnow(long cmd, long a, long b, long c) -{ - long ret; - - asm volatile("cpuid" : - "=a" (ret) : - "b" (a), "c" (b), "d" (c), "0" (MAGIC1), "D" (cmd + MAGIC2)); - return ret; -} - -static void __init simnow_init(char *str) -{ - char *fn = "klog"; - - if (*str == '=') - fn = ++str; - /* error ignored */ - simnow_fd = simnow(XOPEN, (unsigned long)fn, O_WRONLY|O_APPEND|O_CREAT, 0644); -} - -static void simnow_write(struct console *con, const char *s, unsigned n) -{ - simnow(XWRITE, simnow_fd, (unsigned long)s, n); -} - -static struct console simnow_console = { - .name = "simnow", - .write = simnow_write, - .flags = CON_PRINTBUFFER, - .index = -1, -}; - /* Direct interface for emergencies */ static struct console *early_console = &early_vga_console; static int __initdata early_console_initialized; @@ -929,7 +886,7 @@ asmlinkage void early_printk(const char *fmt, ...) va_list ap; va_start(ap, fmt); - n = vscnprintf(buf, 512, fmt, ap); + n = vscnprintf(buf, sizeof(buf), fmt, ap); early_console->write(early_console, buf, n); va_end(ap); } @@ -960,10 +917,6 @@ static int __init setup_early_printk(char *buf) max_ypos = boot_params.screen_info.orig_video_lines; current_ypos = boot_params.screen_info.orig_y; early_console = &early_vga_console; - } else if (!strncmp(buf, "simnow", 6)) { - simnow_init(buf + 6); - early_console = &simnow_console; - keep_early = 1; #ifdef CONFIG_EARLY_PRINTK_DBGP } else if (!strncmp(buf, "dbgp", 4)) { if (early_dbgp_init(buf+4) < 0) diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 28b597ef9ca..d6f0490a739 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -619,28 +619,37 @@ END(syscall_badsys) 27:; /* - * Build the entry stubs and pointer table with - * some assembler magic. + * Build the entry stubs and pointer table with some assembler magic. + * We pack 7 stubs into a single 32-byte chunk, which will fit in a + * single cache line on all modern x86 implementations. */ -.section .rodata,"a" +.section .init.rodata,"a" ENTRY(interrupt) .text - + .p2align 5 + .p2align CONFIG_X86_L1_CACHE_SHIFT ENTRY(irq_entries_start) RING0_INT_FRAME -vector=0 -.rept NR_VECTORS - ALIGN - .if vector +vector=FIRST_EXTERNAL_VECTOR +.rept (NR_VECTORS-FIRST_EXTERNAL_VECTOR+6)/7 + .balign 32 + .rept 7 + .if vector < NR_VECTORS + .if vector <> FIRST_EXTERNAL_VECTOR CFI_ADJUST_CFA_OFFSET -4 - .endif -1: pushl $~(vector) + .endif +1: pushl $(~vector+0x80) /* Note: always in signed byte range */ CFI_ADJUST_CFA_OFFSET 4 - jmp common_interrupt - .previous + .if ((vector-FIRST_EXTERNAL_VECTOR)%7) <> 6 + jmp 2f + .endif + .previous .long 1b - .text + .text vector=vector+1 + .endif + .endr +2: jmp common_interrupt .endr END(irq_entries_start) @@ -652,8 +661,9 @@ END(interrupt) * the CPU automatically disables interrupts when executing an IRQ vector, * so IRQ-flags tracing has to follow that: */ - ALIGN + .p2align CONFIG_X86_L1_CACHE_SHIFT common_interrupt: + addl $-0x80,(%esp) /* Adjust vector into the [-256,-1] range */ SAVE_ALL TRACE_IRQS_OFF movl %esp,%eax @@ -678,65 +688,6 @@ ENDPROC(name) /* The include is where all of the SMP etc. interrupts come from */ #include "entry_arch.h" -KPROBE_ENTRY(page_fault) - RING0_EC_FRAME - pushl $do_page_fault - CFI_ADJUST_CFA_OFFSET 4 - ALIGN -error_code: - /* the function address is in %fs's slot on the stack */ - pushl %es - CFI_ADJUST_CFA_OFFSET 4 - /*CFI_REL_OFFSET es, 0*/ - pushl %ds - CFI_ADJUST_CFA_OFFSET 4 - /*CFI_REL_OFFSET ds, 0*/ - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET eax, 0 - pushl %ebp - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET ebp, 0 - pushl %edi - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET edi, 0 - pushl %esi - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET esi, 0 - pushl %edx - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET edx, 0 - pushl %ecx - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET ecx, 0 - pushl %ebx - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET ebx, 0 - cld - pushl %fs - CFI_ADJUST_CFA_OFFSET 4 - /*CFI_REL_OFFSET fs, 0*/ - movl $(__KERNEL_PERCPU), %ecx - movl %ecx, %fs - UNWIND_ESPFIX_STACK - popl %ecx - CFI_ADJUST_CFA_OFFSET -4 - /*CFI_REGISTER es, ecx*/ - movl PT_FS(%esp), %edi # get the function address - movl PT_ORIG_EAX(%esp), %edx # get the error code - movl $-1, PT_ORIG_EAX(%esp) # no syscall to restart - mov %ecx, PT_FS(%esp) - /*CFI_REL_OFFSET fs, ES*/ - movl $(__USER_DS), %ecx - movl %ecx, %ds - movl %ecx, %es - TRACE_IRQS_OFF - movl %esp,%eax # pt_regs pointer - call *%edi - jmp ret_from_exception - CFI_ENDPROC -KPROBE_END(page_fault) - ENTRY(coprocessor_error) RING0_INT_FRAME pushl $0 @@ -767,140 +718,6 @@ ENTRY(device_not_available) CFI_ENDPROC END(device_not_available) -/* - * Debug traps and NMI can happen at the one SYSENTER instruction - * that sets up the real kernel stack. Check here, since we can't - * allow the wrong stack to be used. - * - * "TSS_sysenter_sp0+12" is because the NMI/debug handler will have - * already pushed 3 words if it hits on the sysenter instruction: - * eflags, cs and eip. - * - * We just load the right stack, and push the three (known) values - * by hand onto the new stack - while updating the return eip past - * the instruction that would have done it for sysenter. - */ -#define FIX_STACK(offset, ok, label) \ - cmpw $__KERNEL_CS,4(%esp); \ - jne ok; \ -label: \ - movl TSS_sysenter_sp0+offset(%esp),%esp; \ - CFI_DEF_CFA esp, 0; \ - CFI_UNDEFINED eip; \ - pushfl; \ - CFI_ADJUST_CFA_OFFSET 4; \ - pushl $__KERNEL_CS; \ - CFI_ADJUST_CFA_OFFSET 4; \ - pushl $sysenter_past_esp; \ - CFI_ADJUST_CFA_OFFSET 4; \ - CFI_REL_OFFSET eip, 0 - -KPROBE_ENTRY(debug) - RING0_INT_FRAME - cmpl $ia32_sysenter_target,(%esp) - jne debug_stack_correct - FIX_STACK(12, debug_stack_correct, debug_esp_fix_insn) -debug_stack_correct: - pushl $-1 # mark this as an int - CFI_ADJUST_CFA_OFFSET 4 - SAVE_ALL - TRACE_IRQS_OFF - xorl %edx,%edx # error code 0 - movl %esp,%eax # pt_regs pointer - call do_debug - jmp ret_from_exception - CFI_ENDPROC -KPROBE_END(debug) - -/* - * NMI is doubly nasty. It can happen _while_ we're handling - * a debug fault, and the debug fault hasn't yet been able to - * clear up the stack. So we first check whether we got an - * NMI on the sysenter entry path, but after that we need to - * check whether we got an NMI on the debug path where the debug - * fault happened on the sysenter path. - */ -KPROBE_ENTRY(nmi) - RING0_INT_FRAME - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - movl %ss, %eax - cmpw $__ESPFIX_SS, %ax - popl %eax - CFI_ADJUST_CFA_OFFSET -4 - je nmi_espfix_stack - cmpl $ia32_sysenter_target,(%esp) - je nmi_stack_fixup - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - movl %esp,%eax - /* Do not access memory above the end of our stack page, - * it might not exist. - */ - andl $(THREAD_SIZE-1),%eax - cmpl $(THREAD_SIZE-20),%eax - popl %eax - CFI_ADJUST_CFA_OFFSET -4 - jae nmi_stack_correct - cmpl $ia32_sysenter_target,12(%esp) - je nmi_debug_stack_check -nmi_stack_correct: - /* We have a RING0_INT_FRAME here */ - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - SAVE_ALL - TRACE_IRQS_OFF - xorl %edx,%edx # zero error code - movl %esp,%eax # pt_regs pointer - call do_nmi - jmp restore_nocheck_notrace - CFI_ENDPROC - -nmi_stack_fixup: - RING0_INT_FRAME - FIX_STACK(12,nmi_stack_correct, 1) - jmp nmi_stack_correct - -nmi_debug_stack_check: - /* We have a RING0_INT_FRAME here */ - cmpw $__KERNEL_CS,16(%esp) - jne nmi_stack_correct - cmpl $debug,(%esp) - jb nmi_stack_correct - cmpl $debug_esp_fix_insn,(%esp) - ja nmi_stack_correct - FIX_STACK(24,nmi_stack_correct, 1) - jmp nmi_stack_correct - -nmi_espfix_stack: - /* We have a RING0_INT_FRAME here. - * - * create the pointer to lss back - */ - pushl %ss - CFI_ADJUST_CFA_OFFSET 4 - pushl %esp - CFI_ADJUST_CFA_OFFSET 4 - addw $4, (%esp) - /* copy the iret frame of 12 bytes */ - .rept 3 - pushl 16(%esp) - CFI_ADJUST_CFA_OFFSET 4 - .endr - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - SAVE_ALL - TRACE_IRQS_OFF - FIXUP_ESPFIX_STACK # %eax == %esp - xorl %edx,%edx # zero error code - call do_nmi - RESTORE_REGS - lss 12+4(%esp), %esp # back to espfix stack - CFI_ADJUST_CFA_OFFSET -24 - jmp irq_return - CFI_ENDPROC -KPROBE_END(nmi) - #ifdef CONFIG_PARAVIRT ENTRY(native_iret) iret @@ -916,19 +733,6 @@ ENTRY(native_irq_enable_sysexit) END(native_irq_enable_sysexit) #endif -KPROBE_ENTRY(int3) - RING0_INT_FRAME - pushl $-1 # mark this as an int - CFI_ADJUST_CFA_OFFSET 4 - SAVE_ALL - TRACE_IRQS_OFF - xorl %edx,%edx # zero error code - movl %esp,%eax # pt_regs pointer - call do_int3 - jmp ret_from_exception - CFI_ENDPROC -KPROBE_END(int3) - ENTRY(overflow) RING0_INT_FRAME pushl $0 @@ -993,14 +797,6 @@ ENTRY(stack_segment) CFI_ENDPROC END(stack_segment) -KPROBE_ENTRY(general_protection) - RING0_EC_FRAME - pushl $do_general_protection - CFI_ADJUST_CFA_OFFSET 4 - jmp error_code - CFI_ENDPROC -KPROBE_END(general_protection) - ENTRY(alignment_check) RING0_EC_FRAME pushl $do_alignment_check @@ -1051,6 +847,7 @@ ENTRY(kernel_thread_helper) push %eax CFI_ADJUST_CFA_OFFSET 4 call do_exit + ud2 # padding for call trace CFI_ENDPROC ENDPROC(kernel_thread_helper) @@ -1157,6 +954,9 @@ ENTRY(mcount) END(mcount) ENTRY(ftrace_caller) + cmpl $0, function_trace_stop + jne ftrace_stub + pushl %eax pushl %ecx pushl %edx @@ -1171,6 +971,11 @@ 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 +985,18 @@ 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 @@ -1200,13 +1015,268 @@ 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" syscall_table_size=(.-sys_call_table) + +/* + * Some functions should be protected against kprobes + */ + .pushsection .kprobes.text, "ax" + +ENTRY(page_fault) + RING0_EC_FRAME + pushl $do_page_fault + CFI_ADJUST_CFA_OFFSET 4 + ALIGN +error_code: + /* the function address is in %fs's slot on the stack */ + pushl %es + CFI_ADJUST_CFA_OFFSET 4 + /*CFI_REL_OFFSET es, 0*/ + pushl %ds + CFI_ADJUST_CFA_OFFSET 4 + /*CFI_REL_OFFSET ds, 0*/ + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET eax, 0 + pushl %ebp + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET ebp, 0 + pushl %edi + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET edi, 0 + pushl %esi + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET esi, 0 + pushl %edx + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET edx, 0 + pushl %ecx + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET ecx, 0 + pushl %ebx + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET ebx, 0 + cld + pushl %fs + CFI_ADJUST_CFA_OFFSET 4 + /*CFI_REL_OFFSET fs, 0*/ + movl $(__KERNEL_PERCPU), %ecx + movl %ecx, %fs + UNWIND_ESPFIX_STACK + popl %ecx + CFI_ADJUST_CFA_OFFSET -4 + /*CFI_REGISTER es, ecx*/ + movl PT_FS(%esp), %edi # get the function address + movl PT_ORIG_EAX(%esp), %edx # get the error code + movl $-1, PT_ORIG_EAX(%esp) # no syscall to restart + mov %ecx, PT_FS(%esp) + /*CFI_REL_OFFSET fs, ES*/ + movl $(__USER_DS), %ecx + movl %ecx, %ds + movl %ecx, %es + TRACE_IRQS_OFF + movl %esp,%eax # pt_regs pointer + call *%edi + jmp ret_from_exception + CFI_ENDPROC +END(page_fault) + +/* + * Debug traps and NMI can happen at the one SYSENTER instruction + * that sets up the real kernel stack. Check here, since we can't + * allow the wrong stack to be used. + * + * "TSS_sysenter_sp0+12" is because the NMI/debug handler will have + * already pushed 3 words if it hits on the sysenter instruction: + * eflags, cs and eip. + * + * We just load the right stack, and push the three (known) values + * by hand onto the new stack - while updating the return eip past + * the instruction that would have done it for sysenter. + */ +#define FIX_STACK(offset, ok, label) \ + cmpw $__KERNEL_CS,4(%esp); \ + jne ok; \ +label: \ + movl TSS_sysenter_sp0+offset(%esp),%esp; \ + CFI_DEF_CFA esp, 0; \ + CFI_UNDEFINED eip; \ + pushfl; \ + CFI_ADJUST_CFA_OFFSET 4; \ + pushl $__KERNEL_CS; \ + CFI_ADJUST_CFA_OFFSET 4; \ + pushl $sysenter_past_esp; \ + CFI_ADJUST_CFA_OFFSET 4; \ + CFI_REL_OFFSET eip, 0 + +ENTRY(debug) + RING0_INT_FRAME + cmpl $ia32_sysenter_target,(%esp) + jne debug_stack_correct + FIX_STACK(12, debug_stack_correct, debug_esp_fix_insn) +debug_stack_correct: + pushl $-1 # mark this as an int + CFI_ADJUST_CFA_OFFSET 4 + SAVE_ALL + TRACE_IRQS_OFF + xorl %edx,%edx # error code 0 + movl %esp,%eax # pt_regs pointer + call do_debug + jmp ret_from_exception + CFI_ENDPROC +END(debug) + +/* + * NMI is doubly nasty. It can happen _while_ we're handling + * a debug fault, and the debug fault hasn't yet been able to + * clear up the stack. So we first check whether we got an + * NMI on the sysenter entry path, but after that we need to + * check whether we got an NMI on the debug path where the debug + * fault happened on the sysenter path. + */ +ENTRY(nmi) + RING0_INT_FRAME + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + movl %ss, %eax + cmpw $__ESPFIX_SS, %ax + popl %eax + CFI_ADJUST_CFA_OFFSET -4 + je nmi_espfix_stack + cmpl $ia32_sysenter_target,(%esp) + je nmi_stack_fixup + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + movl %esp,%eax + /* Do not access memory above the end of our stack page, + * it might not exist. + */ + andl $(THREAD_SIZE-1),%eax + cmpl $(THREAD_SIZE-20),%eax + popl %eax + CFI_ADJUST_CFA_OFFSET -4 + jae nmi_stack_correct + cmpl $ia32_sysenter_target,12(%esp) + je nmi_debug_stack_check +nmi_stack_correct: + /* We have a RING0_INT_FRAME here */ + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + SAVE_ALL + TRACE_IRQS_OFF + xorl %edx,%edx # zero error code + movl %esp,%eax # pt_regs pointer + call do_nmi + jmp restore_nocheck_notrace + CFI_ENDPROC + +nmi_stack_fixup: + RING0_INT_FRAME + FIX_STACK(12,nmi_stack_correct, 1) + jmp nmi_stack_correct + +nmi_debug_stack_check: + /* We have a RING0_INT_FRAME here */ + cmpw $__KERNEL_CS,16(%esp) + jne nmi_stack_correct + cmpl $debug,(%esp) + jb nmi_stack_correct + cmpl $debug_esp_fix_insn,(%esp) + ja nmi_stack_correct + FIX_STACK(24,nmi_stack_correct, 1) + jmp nmi_stack_correct + +nmi_espfix_stack: + /* We have a RING0_INT_FRAME here. + * + * create the pointer to lss back + */ + pushl %ss + CFI_ADJUST_CFA_OFFSET 4 + pushl %esp + CFI_ADJUST_CFA_OFFSET 4 + addw $4, (%esp) + /* copy the iret frame of 12 bytes */ + .rept 3 + pushl 16(%esp) + CFI_ADJUST_CFA_OFFSET 4 + .endr + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + SAVE_ALL + TRACE_IRQS_OFF + FIXUP_ESPFIX_STACK # %eax == %esp + xorl %edx,%edx # zero error code + call do_nmi + RESTORE_REGS + lss 12+4(%esp), %esp # back to espfix stack + CFI_ADJUST_CFA_OFFSET -24 + jmp irq_return + CFI_ENDPROC +END(nmi) + +ENTRY(int3) + RING0_INT_FRAME + pushl $-1 # mark this as an int + CFI_ADJUST_CFA_OFFSET 4 + SAVE_ALL + TRACE_IRQS_OFF + xorl %edx,%edx # zero error code + movl %esp,%eax # pt_regs pointer + call do_int3 + jmp ret_from_exception + CFI_ENDPROC +END(int3) + +ENTRY(general_protection) + RING0_EC_FRAME + pushl $do_general_protection + CFI_ADJUST_CFA_OFFSET 4 + jmp error_code + CFI_ENDPROC +END(general_protection) + +/* + * End of kprobes section + */ + .popsection diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index b86f332c96a..e28c7a98779 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -11,15 +11,15 @@ * * NOTE: This code handles signal-recognition, which happens every time * after an interrupt and after each system call. - * - * Normal syscalls and interrupts don't save a full stack frame, this is + * + * Normal syscalls and interrupts don't save a full stack frame, this is * only done for syscall tracing, signals or fork/exec et.al. - * - * A note on terminology: - * - top of stack: Architecture defined interrupt frame from SS to RIP - * at the top of the kernel process stack. + * + * A note on terminology: + * - top of stack: Architecture defined interrupt frame from SS to RIP + * at the top of the kernel process stack. * - partial stack frame: partially saved registers upto R11. - * - full stack frame: Like partial stack frame, but all register saved. + * - full stack frame: Like partial stack frame, but all register saved. * * Some macro usage: * - CFI macros are used to generate dwarf2 unwind information for better @@ -60,7 +60,6 @@ #define __AUDIT_ARCH_LE 0x40000000 .code64 - #ifdef CONFIG_FUNCTION_TRACER #ifdef CONFIG_DYNAMIC_FTRACE ENTRY(mcount) @@ -68,16 +67,10 @@ ENTRY(mcount) END(mcount) ENTRY(ftrace_caller) + cmpl $0, function_trace_stop + jne ftrace_stub - /* taken from glibc */ - 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) + MCOUNT_SAVE_FRAME movq 0x38(%rsp), %rdi movq 8(%rbp), %rsi @@ -87,14 +80,13 @@ ENTRY(ftrace_caller) ftrace_call: call ftrace_stub - 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 + MCOUNT_RESTORE_FRAME + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +.globl ftrace_graph_call +ftrace_graph_call: + jmp ftrace_stub +#endif .globl ftrace_stub ftrace_stub: @@ -103,15 +95,63 @@ 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 trace: - /* taken from glibc */ - subq $0x38, %rsp + MCOUNT_SAVE_FRAME + + movq 0x38(%rsp), %rdi + movq 8(%rbp), %rsi + subq $MCOUNT_INSN_SIZE, %rdi + + call *ftrace_trace_function + + MCOUNT_RESTORE_FRAME + + 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 + + MCOUNT_SAVE_FRAME + + leaq 8(%rbp), %rdi + movq 0x38(%rsp), %rsi + subq $MCOUNT_INSN_SIZE, %rsi + + call prepare_ftrace_return + + MCOUNT_RESTORE_FRAME + + 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) @@ -119,13 +159,14 @@ trace: movq %rdi, 32(%rsp) movq %r8, 40(%rsp) movq %r9, 48(%rsp) + movq %r10, 56(%rsp) + movq %r11, 64(%rsp) - movq 0x38(%rsp), %rdi - movq 8(%rbp), %rsi - subq $MCOUNT_INSN_SIZE, %rdi - - call *ftrace_trace_function + 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 @@ -133,16 +174,14 @@ trace: movq 16(%rsp), %rdx movq 8(%rsp), %rcx movq (%rsp), %rax - addq $0x38, %rsp + addq $72, %rsp + retq +#endif - jmp ftrace_stub -END(mcount) -#endif /* CONFIG_DYNAMIC_FTRACE */ -#endif /* CONFIG_FUNCTION_TRACER */ #ifndef CONFIG_PREEMPT #define retint_kernel retint_restore_args -#endif +#endif #ifdef CONFIG_PARAVIRT ENTRY(native_usergs_sysret64) @@ -161,29 +200,29 @@ ENTRY(native_usergs_sysret64) .endm /* - * C code is not supposed to know about undefined top of stack. Every time - * a C function with an pt_regs argument is called from the SYSCALL based + * C code is not supposed to know about undefined top of stack. Every time + * a C function with an pt_regs argument is called from the SYSCALL based * fast path FIXUP_TOP_OF_STACK is needed. * RESTORE_TOP_OF_STACK syncs the syscall state after any possible ptregs * manipulation. - */ - - /* %rsp:at FRAMEEND */ - .macro FIXUP_TOP_OF_STACK tmp - movq %gs:pda_oldrsp,\tmp - movq \tmp,RSP(%rsp) - movq $__USER_DS,SS(%rsp) - movq $__USER_CS,CS(%rsp) - movq $-1,RCX(%rsp) - movq R11(%rsp),\tmp /* get eflags */ - movq \tmp,EFLAGS(%rsp) + */ + + /* %rsp:at FRAMEEND */ + .macro FIXUP_TOP_OF_STACK tmp offset=0 + movq %gs:pda_oldrsp,\tmp + movq \tmp,RSP+\offset(%rsp) + movq $__USER_DS,SS+\offset(%rsp) + movq $__USER_CS,CS+\offset(%rsp) + movq $-1,RCX+\offset(%rsp) + movq R11+\offset(%rsp),\tmp /* get eflags */ + movq \tmp,EFLAGS+\offset(%rsp) .endm - .macro RESTORE_TOP_OF_STACK tmp,offset=0 - movq RSP-\offset(%rsp),\tmp - movq \tmp,%gs:pda_oldrsp - movq EFLAGS-\offset(%rsp),\tmp - movq \tmp,R11-\offset(%rsp) + .macro RESTORE_TOP_OF_STACK tmp offset=0 + movq RSP+\offset(%rsp),\tmp + movq \tmp,%gs:pda_oldrsp + movq EFLAGS+\offset(%rsp),\tmp + movq \tmp,R11+\offset(%rsp) .endm .macro FAKE_STACK_FRAME child_rip @@ -195,7 +234,7 @@ ENTRY(native_usergs_sysret64) pushq %rax /* rsp */ CFI_ADJUST_CFA_OFFSET 8 CFI_REL_OFFSET rsp,0 - pushq $(1<<9) /* eflags - interrupts on */ + pushq $X86_EFLAGS_IF /* eflags - interrupts on */ CFI_ADJUST_CFA_OFFSET 8 /*CFI_REL_OFFSET rflags,0*/ pushq $__KERNEL_CS /* cs */ @@ -213,62 +252,184 @@ ENTRY(native_usergs_sysret64) CFI_ADJUST_CFA_OFFSET -(6*8) .endm - .macro CFI_DEFAULT_STACK start=1 +/* + * initial frame state for interrupts (and exceptions without error code) + */ + .macro EMPTY_FRAME start=1 offset=0 .if \start - CFI_STARTPROC simple + CFI_STARTPROC simple CFI_SIGNAL_FRAME - CFI_DEF_CFA rsp,SS+8 + CFI_DEF_CFA rsp,8+\offset .else - CFI_DEF_CFA_OFFSET SS+8 + CFI_DEF_CFA_OFFSET 8+\offset .endif - CFI_REL_OFFSET r15,R15 - CFI_REL_OFFSET r14,R14 - CFI_REL_OFFSET r13,R13 - CFI_REL_OFFSET r12,R12 - CFI_REL_OFFSET rbp,RBP - CFI_REL_OFFSET rbx,RBX - CFI_REL_OFFSET r11,R11 - CFI_REL_OFFSET r10,R10 - CFI_REL_OFFSET r9,R9 - CFI_REL_OFFSET r8,R8 - CFI_REL_OFFSET rax,RAX - CFI_REL_OFFSET rcx,RCX - CFI_REL_OFFSET rdx,RDX - CFI_REL_OFFSET rsi,RSI - CFI_REL_OFFSET rdi,RDI - CFI_REL_OFFSET rip,RIP - /*CFI_REL_OFFSET cs,CS*/ - /*CFI_REL_OFFSET rflags,EFLAGS*/ - CFI_REL_OFFSET rsp,RSP - /*CFI_REL_OFFSET ss,SS*/ .endm + +/* + * initial frame state for interrupts (and exceptions without error code) + */ + .macro INTR_FRAME start=1 offset=0 + EMPTY_FRAME \start, SS+8+\offset-RIP + /*CFI_REL_OFFSET ss, SS+\offset-RIP*/ + CFI_REL_OFFSET rsp, RSP+\offset-RIP + /*CFI_REL_OFFSET rflags, EFLAGS+\offset-RIP*/ + /*CFI_REL_OFFSET cs, CS+\offset-RIP*/ + CFI_REL_OFFSET rip, RIP+\offset-RIP + .endm + +/* + * initial frame state for exceptions with error code (and interrupts + * with vector already pushed) + */ + .macro XCPT_FRAME start=1 offset=0 + INTR_FRAME \start, RIP+\offset-ORIG_RAX + /*CFI_REL_OFFSET orig_rax, ORIG_RAX-ORIG_RAX*/ + .endm + /* - * A newly forked process directly context switches into this. - */ -/* rdi: prev */ + * frame that enables calling into C. + */ + .macro PARTIAL_FRAME start=1 offset=0 + XCPT_FRAME \start, ORIG_RAX+\offset-ARGOFFSET + CFI_REL_OFFSET rdi, RDI+\offset-ARGOFFSET + CFI_REL_OFFSET rsi, RSI+\offset-ARGOFFSET + CFI_REL_OFFSET rdx, RDX+\offset-ARGOFFSET + CFI_REL_OFFSET rcx, RCX+\offset-ARGOFFSET + CFI_REL_OFFSET rax, RAX+\offset-ARGOFFSET + CFI_REL_OFFSET r8, R8+\offset-ARGOFFSET + CFI_REL_OFFSET r9, R9+\offset-ARGOFFSET + CFI_REL_OFFSET r10, R10+\offset-ARGOFFSET + CFI_REL_OFFSET r11, R11+\offset-ARGOFFSET + .endm + +/* + * frame that enables passing a complete pt_regs to a C function. + */ + .macro DEFAULT_FRAME start=1 offset=0 + PARTIAL_FRAME \start, R11+\offset-R15 + CFI_REL_OFFSET rbx, RBX+\offset + CFI_REL_OFFSET rbp, RBP+\offset + CFI_REL_OFFSET r12, R12+\offset + CFI_REL_OFFSET r13, R13+\offset + CFI_REL_OFFSET r14, R14+\offset + CFI_REL_OFFSET r15, R15+\offset + .endm + +/* save partial stack frame */ +ENTRY(save_args) + XCPT_FRAME + cld + movq_cfi rdi, RDI+16-ARGOFFSET + movq_cfi rsi, RSI+16-ARGOFFSET + movq_cfi rdx, RDX+16-ARGOFFSET + movq_cfi rcx, RCX+16-ARGOFFSET + movq_cfi rax, RAX+16-ARGOFFSET + movq_cfi r8, R8+16-ARGOFFSET + movq_cfi r9, R9+16-ARGOFFSET + movq_cfi r10, R10+16-ARGOFFSET + movq_cfi r11, R11+16-ARGOFFSET + + leaq -ARGOFFSET+16(%rsp),%rdi /* arg1 for handler */ + movq_cfi rbp, 8 /* push %rbp */ + leaq 8(%rsp), %rbp /* mov %rsp, %ebp */ + testl $3, CS(%rdi) + je 1f + SWAPGS + /* + * irqcount is used to check if a CPU is already on an interrupt stack + * or not. While this is essentially redundant with preempt_count it is + * a little cheaper to use a separate counter in the PDA (short of + * moving irq_enter into assembly, which would be too much work) + */ +1: incl %gs:pda_irqcount + jne 2f + popq_cfi %rax /* move return address... */ + mov %gs:pda_irqstackptr,%rsp + EMPTY_FRAME 0 + pushq_cfi %rax /* ... to the new stack */ + /* + * We entered an interrupt context - irqs are off: + */ +2: TRACE_IRQS_OFF + ret + CFI_ENDPROC +END(save_args) + +ENTRY(save_rest) + PARTIAL_FRAME 1 REST_SKIP+8 + movq 5*8+16(%rsp), %r11 /* save return address */ + movq_cfi rbx, RBX+16 + movq_cfi rbp, RBP+16 + movq_cfi r12, R12+16 + movq_cfi r13, R13+16 + movq_cfi r14, R14+16 + movq_cfi r15, R15+16 + movq %r11, 8(%rsp) /* return address */ + FIXUP_TOP_OF_STACK %r11, 16 + ret + CFI_ENDPROC +END(save_rest) + +/* save complete stack frame */ +ENTRY(save_paranoid) + XCPT_FRAME 1 RDI+8 + cld + movq_cfi rdi, RDI+8 + movq_cfi rsi, RSI+8 + movq_cfi rdx, RDX+8 + movq_cfi rcx, RCX+8 + movq_cfi rax, RAX+8 + movq_cfi r8, R8+8 + movq_cfi r9, R9+8 + movq_cfi r10, R10+8 + movq_cfi r11, R11+8 + movq_cfi rbx, RBX+8 + movq_cfi rbp, RBP+8 + movq_cfi r12, R12+8 + movq_cfi r13, R13+8 + movq_cfi r14, R14+8 + movq_cfi r15, R15+8 + movl $1,%ebx + movl $MSR_GS_BASE,%ecx + rdmsr + testl %edx,%edx + js 1f /* negative -> in kernel */ + SWAPGS + xorl %ebx,%ebx +1: ret + CFI_ENDPROC +END(save_paranoid) + +/* + * A newly forked process directly context switches into this address. + * + * rdi: prev task we switched from + */ ENTRY(ret_from_fork) - CFI_DEFAULT_STACK + DEFAULT_FRAME + push kernel_eflags(%rip) CFI_ADJUST_CFA_OFFSET 8 - popf # reset kernel eflags + popf # reset kernel eflags CFI_ADJUST_CFA_OFFSET -8 - call schedule_tail + + call schedule_tail # rdi: 'prev' task parameter + GET_THREAD_INFO(%rcx) - testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT),TI_flags(%rcx) - jnz rff_trace -rff_action: + + CFI_REMEMBER_STATE RESTORE_REST - testl $3,CS-ARGOFFSET(%rsp) # from kernel_thread? + + testl $3, CS-ARGOFFSET(%rsp) # from kernel_thread? je int_ret_from_sys_call - testl $_TIF_IA32,TI_flags(%rcx) + + testl $_TIF_IA32, TI_flags(%rcx) # 32-bit compat task needs IRET jnz int_ret_from_sys_call - RESTORE_TOP_OF_STACK %rdi,ARGOFFSET - jmp ret_from_sys_call -rff_trace: - movq %rsp,%rdi - call syscall_trace_leave - GET_THREAD_INFO(%rcx) - jmp rff_action + + RESTORE_TOP_OF_STACK %rdi, -ARGOFFSET + jmp ret_from_sys_call # go to the SYSRET fastpath + + CFI_RESTORE_STATE CFI_ENDPROC END(ret_from_fork) @@ -278,20 +439,20 @@ END(ret_from_fork) * SYSCALL does not save anything on the stack and does not change the * stack pointer. */ - + /* - * Register setup: + * Register setup: * rax system call number * rdi arg0 - * rcx return address for syscall/sysret, C arg3 + * rcx return address for syscall/sysret, C arg3 * rsi arg1 - * rdx arg2 + * rdx arg2 * r10 arg3 (--> moved to rcx for C) * r8 arg4 * r9 arg5 * r11 eflags for syscall/sysret, temporary for C - * r12-r15,rbp,rbx saved by C code, not touched. - * + * r12-r15,rbp,rbx saved by C code, not touched. + * * Interrupts are off on entry. * Only called from user space. * @@ -301,7 +462,7 @@ END(ret_from_fork) * When user can change the frames always force IRET. That is because * it deals with uncanonical addresses better. SYSRET has trouble * with them due to bugs in both AMD and Intel CPUs. - */ + */ ENTRY(system_call) CFI_STARTPROC simple @@ -317,7 +478,7 @@ ENTRY(system_call) */ ENTRY(system_call_after_swapgs) - movq %rsp,%gs:pda_oldrsp + movq %rsp,%gs:pda_oldrsp movq %gs:pda_kernelstack,%rsp /* * No need to follow this irqs off/on section - it's straight @@ -325,7 +486,7 @@ ENTRY(system_call_after_swapgs) */ ENABLE_INTERRUPTS(CLBR_NONE) SAVE_ARGS 8,1 - movq %rax,ORIG_RAX-ARGOFFSET(%rsp) + movq %rax,ORIG_RAX-ARGOFFSET(%rsp) movq %rcx,RIP-ARGOFFSET(%rsp) CFI_REL_OFFSET rip,RIP-ARGOFFSET GET_THREAD_INFO(%rcx) @@ -339,19 +500,19 @@ system_call_fastpath: movq %rax,RAX-ARGOFFSET(%rsp) /* * Syscall return path ending with SYSRET (fast path) - * Has incomplete stack frame and undefined top of stack. - */ + * Has incomplete stack frame and undefined top of stack. + */ ret_from_sys_call: movl $_TIF_ALLWORK_MASK,%edi /* edi: flagmask */ -sysret_check: +sysret_check: LOCKDEP_SYS_EXIT GET_THREAD_INFO(%rcx) DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF movl TI_flags(%rcx),%edx andl %edi,%edx - jnz sysret_careful + jnz sysret_careful CFI_REMEMBER_STATE /* * sysretq will re-enable interrupts: @@ -366,7 +527,7 @@ sysret_check: CFI_RESTORE_STATE /* Handle reschedules */ - /* edx: work, edi: workmask */ + /* edx: work, edi: workmask */ sysret_careful: bt $TIF_NEED_RESCHED,%edx jnc sysret_signal @@ -379,7 +540,7 @@ sysret_careful: CFI_ADJUST_CFA_OFFSET -8 jmp sysret_check - /* Handle a signal */ + /* Handle a signal */ sysret_signal: TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) @@ -388,17 +549,20 @@ sysret_signal: jc sysret_audit #endif /* edx: work flags (arg3) */ - leaq do_notify_resume(%rip),%rax leaq -ARGOFFSET(%rsp),%rdi # &pt_regs -> arg1 xorl %esi,%esi # oldset -> arg2 - call ptregscall_common + SAVE_REST + FIXUP_TOP_OF_STACK %r11 + call do_notify_resume + RESTORE_TOP_OF_STACK %r11 + RESTORE_REST movl $_TIF_WORK_MASK,%edi /* Use IRET because user could have changed frame. This works because ptregscall_common has called FIXUP_TOP_OF_STACK. */ DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF jmp int_with_check - + badsys: movq $-ENOSYS,RAX-ARGOFFSET(%rsp) jmp ret_from_sys_call @@ -437,7 +601,7 @@ sysret_audit: #endif /* CONFIG_AUDITSYSCALL */ /* Do syscall tracing */ -tracesys: +tracesys: #ifdef CONFIG_AUDITSYSCALL testl $(_TIF_WORK_SYSCALL_ENTRY & ~_TIF_SYSCALL_AUDIT),TI_flags(%rcx) jz auditsys @@ -460,8 +624,8 @@ tracesys: call *sys_call_table(,%rax,8) movq %rax,RAX-ARGOFFSET(%rsp) /* Use IRET because user could have changed frame */ - -/* + +/* * Syscall return path ending with IRET. * Has correct top of stack, but partial stack frame. */ @@ -505,18 +669,18 @@ int_very_careful: TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) SAVE_REST - /* Check for syscall exit trace */ + /* Check for syscall exit trace */ testl $_TIF_WORK_SYSCALL_EXIT,%edx jz int_signal pushq %rdi CFI_ADJUST_CFA_OFFSET 8 - leaq 8(%rsp),%rdi # &ptregs -> arg1 + leaq 8(%rsp),%rdi # &ptregs -> arg1 call syscall_trace_leave popq %rdi CFI_ADJUST_CFA_OFFSET -8 andl $~(_TIF_WORK_SYSCALL_EXIT|_TIF_SYSCALL_EMU),%edi jmp int_restore_rest - + int_signal: testl $_TIF_DO_NOTIFY_MASK,%edx jz 1f @@ -531,22 +695,24 @@ int_restore_rest: jmp int_with_check CFI_ENDPROC END(system_call) - -/* + +/* * Certain special system calls that need to save a complete full stack frame. - */ - + */ .macro PTREGSCALL label,func,arg - .globl \label -\label: - leaq \func(%rip),%rax - leaq -ARGOFFSET+8(%rsp),\arg /* 8 for return address */ - jmp ptregscall_common +ENTRY(\label) + PARTIAL_FRAME 1 8 /* offset 8: return address */ + subq $REST_SKIP, %rsp + CFI_ADJUST_CFA_OFFSET REST_SKIP + call save_rest + DEFAULT_FRAME 0 8 /* offset 8: return address */ + leaq 8(%rsp), \arg /* pt_regs pointer */ + call \func + jmp ptregscall_common + CFI_ENDPROC END(\label) .endm - CFI_STARTPROC - PTREGSCALL stub_clone, sys_clone, %r8 PTREGSCALL stub_fork, sys_fork, %rdi PTREGSCALL stub_vfork, sys_vfork, %rdi @@ -554,25 +720,18 @@ END(\label) PTREGSCALL stub_iopl, sys_iopl, %rsi ENTRY(ptregscall_common) - popq %r11 - CFI_ADJUST_CFA_OFFSET -8 - CFI_REGISTER rip, r11 - SAVE_REST - movq %r11, %r15 - CFI_REGISTER rip, r15 - FIXUP_TOP_OF_STACK %r11 - call *%rax - RESTORE_TOP_OF_STACK %r11 - movq %r15, %r11 - CFI_REGISTER rip, r11 - RESTORE_REST - pushq %r11 - CFI_ADJUST_CFA_OFFSET 8 - CFI_REL_OFFSET rip, 0 - ret + DEFAULT_FRAME 1 8 /* offset 8: return address */ + RESTORE_TOP_OF_STACK %r11, 8 + movq_cfi_restore R15+8, r15 + movq_cfi_restore R14+8, r14 + movq_cfi_restore R13+8, r13 + movq_cfi_restore R12+8, r12 + movq_cfi_restore RBP+8, rbp + movq_cfi_restore RBX+8, rbx + ret $REST_SKIP /* pop extended registers */ CFI_ENDPROC END(ptregscall_common) - + ENTRY(stub_execve) CFI_STARTPROC popq %r11 @@ -588,11 +747,11 @@ ENTRY(stub_execve) jmp int_ret_from_sys_call CFI_ENDPROC END(stub_execve) - + /* * sigreturn is special because it needs to restore all registers on return. * This cannot be done with SYSRET, so use the IRET return path instead. - */ + */ ENTRY(stub_rt_sigreturn) CFI_STARTPROC addq $8, %rsp @@ -608,70 +767,70 @@ ENTRY(stub_rt_sigreturn) END(stub_rt_sigreturn) /* - * initial frame state for interrupts and exceptions + * Build the entry stubs and pointer table with some assembler magic. + * We pack 7 stubs into a single 32-byte chunk, which will fit in a + * single cache line on all modern x86 implementations. */ - .macro _frame ref - CFI_STARTPROC simple - CFI_SIGNAL_FRAME - CFI_DEF_CFA rsp,SS+8-\ref - /*CFI_REL_OFFSET ss,SS-\ref*/ - CFI_REL_OFFSET rsp,RSP-\ref - /*CFI_REL_OFFSET rflags,EFLAGS-\ref*/ - /*CFI_REL_OFFSET cs,CS-\ref*/ - CFI_REL_OFFSET rip,RIP-\ref - .endm + .section .init.rodata,"a" +ENTRY(interrupt) + .text + .p2align 5 + .p2align CONFIG_X86_L1_CACHE_SHIFT +ENTRY(irq_entries_start) + INTR_FRAME +vector=FIRST_EXTERNAL_VECTOR +.rept (NR_VECTORS-FIRST_EXTERNAL_VECTOR+6)/7 + .balign 32 + .rept 7 + .if vector < NR_VECTORS + .if vector <> FIRST_EXTERNAL_VECTOR + CFI_ADJUST_CFA_OFFSET -8 + .endif +1: pushq $(~vector+0x80) /* Note: always in signed byte range */ + CFI_ADJUST_CFA_OFFSET 8 + .if ((vector-FIRST_EXTERNAL_VECTOR)%7) <> 6 + jmp 2f + .endif + .previous + .quad 1b + .text +vector=vector+1 + .endif + .endr +2: jmp common_interrupt +.endr + CFI_ENDPROC +END(irq_entries_start) -/* initial frame state for interrupts (and exceptions without error code) */ -#define INTR_FRAME _frame RIP -/* initial frame state for exceptions with error code (and interrupts with - vector already pushed) */ -#define XCPT_FRAME _frame ORIG_RAX +.previous +END(interrupt) +.previous -/* +/* * Interrupt entry/exit. * * Interrupt entry points save only callee clobbered registers in fast path. - * - * Entry runs with interrupts off. - */ + * + * Entry runs with interrupts off. + */ -/* 0(%rsp): interrupt number */ +/* 0(%rsp): ~(interrupt number) */ .macro interrupt func - cld - SAVE_ARGS - leaq -ARGOFFSET(%rsp),%rdi # arg1 for handler - pushq %rbp - /* - * Save rbp twice: One is for marking the stack frame, as usual, and the - * other, to fill pt_regs properly. This is because bx comes right - * before the last saved register in that structure, and not bp. If the - * base pointer were in the place bx is today, this would not be needed. - */ - movq %rbp, -8(%rsp) - CFI_ADJUST_CFA_OFFSET 8 - CFI_REL_OFFSET rbp, 0 - movq %rsp,%rbp - CFI_DEF_CFA_REGISTER rbp - testl $3,CS(%rdi) - je 1f - SWAPGS - /* irqcount is used to check if a CPU is already on an interrupt - stack or not. While this is essentially redundant with preempt_count - it is a little cheaper to use a separate counter in the PDA - (short of moving irq_enter into assembly, which would be too - much work) */ -1: incl %gs:pda_irqcount - cmoveq %gs:pda_irqstackptr,%rsp - push %rbp # backlink for old unwinder - /* - * We entered an interrupt context - irqs are off: - */ - TRACE_IRQS_OFF + subq $10*8, %rsp + CFI_ADJUST_CFA_OFFSET 10*8 + call save_args + PARTIAL_FRAME 0 call \func .endm -ENTRY(common_interrupt) + /* + * The interrupt stubs push (~vector+0x80) onto the stack and + * then jump to common_interrupt. + */ + .p2align CONFIG_X86_L1_CACHE_SHIFT +common_interrupt: XCPT_FRAME + addq $-0x80,(%rsp) /* Adjust vector to [-256,-1] range */ interrupt do_IRQ /* 0(%rsp): oldrsp-ARGOFFSET */ ret_from_intr: @@ -685,12 +844,12 @@ exit_intr: GET_THREAD_INFO(%rcx) testl $3,CS-ARGOFFSET(%rsp) je retint_kernel - + /* Interrupt came from user space */ /* * Has a correct top of stack, but a partial stack frame * %rcx: thread info. Interrupts off. - */ + */ retint_with_reschedule: movl $_TIF_WORK_MASK,%edi retint_check: @@ -763,20 +922,20 @@ retint_careful: pushq %rdi CFI_ADJUST_CFA_OFFSET 8 call schedule - popq %rdi + popq %rdi CFI_ADJUST_CFA_OFFSET -8 GET_THREAD_INFO(%rcx) DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF jmp retint_check - + retint_signal: testl $_TIF_DO_NOTIFY_MASK,%edx jz retint_swapgs TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) SAVE_REST - movq $-1,ORIG_RAX(%rsp) + movq $-1,ORIG_RAX(%rsp) xorl %esi,%esi # oldset movq %rsp,%rdi # &pt_regs call do_notify_resume @@ -798,324 +957,211 @@ ENTRY(retint_kernel) jnc retint_restore_args call preempt_schedule_irq jmp exit_intr -#endif +#endif CFI_ENDPROC END(common_interrupt) - + /* * APIC interrupts. - */ - .macro apicinterrupt num,func + */ +.macro apicinterrupt num sym do_sym +ENTRY(\sym) INTR_FRAME pushq $~(\num) CFI_ADJUST_CFA_OFFSET 8 - interrupt \func + interrupt \do_sym jmp ret_from_intr CFI_ENDPROC - .endm - -ENTRY(thermal_interrupt) - apicinterrupt THERMAL_APIC_VECTOR,smp_thermal_interrupt -END(thermal_interrupt) - -ENTRY(threshold_interrupt) - apicinterrupt THRESHOLD_APIC_VECTOR,mce_threshold_interrupt -END(threshold_interrupt) - -#ifdef CONFIG_SMP -ENTRY(reschedule_interrupt) - apicinterrupt RESCHEDULE_VECTOR,smp_reschedule_interrupt -END(reschedule_interrupt) - - .macro INVALIDATE_ENTRY num -ENTRY(invalidate_interrupt\num) - apicinterrupt INVALIDATE_TLB_VECTOR_START+\num,smp_invalidate_interrupt -END(invalidate_interrupt\num) - .endm +END(\sym) +.endm - INVALIDATE_ENTRY 0 - INVALIDATE_ENTRY 1 - INVALIDATE_ENTRY 2 - INVALIDATE_ENTRY 3 - INVALIDATE_ENTRY 4 - INVALIDATE_ENTRY 5 - INVALIDATE_ENTRY 6 - INVALIDATE_ENTRY 7 - -ENTRY(call_function_interrupt) - apicinterrupt CALL_FUNCTION_VECTOR,smp_call_function_interrupt -END(call_function_interrupt) -ENTRY(call_function_single_interrupt) - apicinterrupt CALL_FUNCTION_SINGLE_VECTOR,smp_call_function_single_interrupt -END(call_function_single_interrupt) -ENTRY(irq_move_cleanup_interrupt) - apicinterrupt IRQ_MOVE_CLEANUP_VECTOR,smp_irq_move_cleanup_interrupt -END(irq_move_cleanup_interrupt) +#ifdef CONFIG_SMP +apicinterrupt IRQ_MOVE_CLEANUP_VECTOR \ + irq_move_cleanup_interrupt smp_irq_move_cleanup_interrupt #endif -ENTRY(apic_timer_interrupt) - apicinterrupt LOCAL_TIMER_VECTOR,smp_apic_timer_interrupt -END(apic_timer_interrupt) +apicinterrupt UV_BAU_MESSAGE \ + uv_bau_message_intr1 uv_bau_message_interrupt +apicinterrupt LOCAL_TIMER_VECTOR \ + apic_timer_interrupt smp_apic_timer_interrupt + +#ifdef CONFIG_SMP +apicinterrupt INVALIDATE_TLB_VECTOR_START+0 \ + invalidate_interrupt0 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+1 \ + invalidate_interrupt1 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+2 \ + invalidate_interrupt2 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+3 \ + invalidate_interrupt3 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+4 \ + invalidate_interrupt4 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+5 \ + invalidate_interrupt5 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+6 \ + invalidate_interrupt6 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+7 \ + invalidate_interrupt7 smp_invalidate_interrupt +#endif -ENTRY(uv_bau_message_intr1) - apicinterrupt 220,uv_bau_message_interrupt -END(uv_bau_message_intr1) +apicinterrupt THRESHOLD_APIC_VECTOR \ + threshold_interrupt mce_threshold_interrupt +apicinterrupt THERMAL_APIC_VECTOR \ + thermal_interrupt smp_thermal_interrupt + +#ifdef CONFIG_SMP +apicinterrupt CALL_FUNCTION_SINGLE_VECTOR \ + call_function_single_interrupt smp_call_function_single_interrupt +apicinterrupt CALL_FUNCTION_VECTOR \ + call_function_interrupt smp_call_function_interrupt +apicinterrupt RESCHEDULE_VECTOR \ + reschedule_interrupt smp_reschedule_interrupt +#endif -ENTRY(error_interrupt) - apicinterrupt ERROR_APIC_VECTOR,smp_error_interrupt -END(error_interrupt) +apicinterrupt ERROR_APIC_VECTOR \ + error_interrupt smp_error_interrupt +apicinterrupt SPURIOUS_APIC_VECTOR \ + spurious_interrupt smp_spurious_interrupt -ENTRY(spurious_interrupt) - apicinterrupt SPURIOUS_APIC_VECTOR,smp_spurious_interrupt -END(spurious_interrupt) - /* * Exception entry points. - */ - .macro zeroentry sym + */ +.macro zeroentry sym do_sym +ENTRY(\sym) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 /* push error code/oldrax */ - CFI_ADJUST_CFA_OFFSET 8 - pushq %rax /* push real oldrax to the rdi slot */ - CFI_ADJUST_CFA_OFFSET 8 - CFI_REL_OFFSET rax,0 - leaq \sym(%rip),%rax - jmp error_entry + pushq_cfi $-1 /* ORIG_RAX: no syscall to restart */ + subq $15*8,%rsp + CFI_ADJUST_CFA_OFFSET 15*8 + call error_entry + DEFAULT_FRAME 0 + movq %rsp,%rdi /* pt_regs pointer */ + xorl %esi,%esi /* no error code */ + call \do_sym + jmp error_exit /* %ebx: no swapgs flag */ CFI_ENDPROC - .endm +END(\sym) +.endm - .macro errorentry sym - XCPT_FRAME +.macro paranoidzeroentry sym do_sym +ENTRY(\sym) + INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq %rax + pushq $-1 /* ORIG_RAX: no syscall to restart */ CFI_ADJUST_CFA_OFFSET 8 - CFI_REL_OFFSET rax,0 - leaq \sym(%rip),%rax - jmp error_entry + subq $15*8, %rsp + call save_paranoid + TRACE_IRQS_OFF + movq %rsp,%rdi /* pt_regs pointer */ + xorl %esi,%esi /* no error code */ + call \do_sym + jmp paranoid_exit /* %ebx: no swapgs flag */ CFI_ENDPROC - .endm +END(\sym) +.endm - /* error code is on the stack already */ - /* handle NMI like exceptions that can happen everywhere */ - .macro paranoidentry sym, ist=0, irqtrace=1 - SAVE_ALL - cld - movl $1,%ebx - movl $MSR_GS_BASE,%ecx - rdmsr - testl %edx,%edx - js 1f - SWAPGS - xorl %ebx,%ebx -1: - .if \ist - movq %gs:pda_data_offset, %rbp - .endif - .if \irqtrace - TRACE_IRQS_OFF - .endif - movq %rsp,%rdi - movq ORIG_RAX(%rsp),%rsi - movq $-1,ORIG_RAX(%rsp) - .if \ist - subq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) - .endif - call \sym - .if \ist - addq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) - .endif - DISABLE_INTERRUPTS(CLBR_NONE) - .if \irqtrace +.macro paranoidzeroentry_ist sym do_sym ist +ENTRY(\sym) + INTR_FRAME + PARAVIRT_ADJUST_EXCEPTION_FRAME + pushq $-1 /* ORIG_RAX: no syscall to restart */ + CFI_ADJUST_CFA_OFFSET 8 + subq $15*8, %rsp + call save_paranoid TRACE_IRQS_OFF - .endif - .endm + movq %rsp,%rdi /* pt_regs pointer */ + xorl %esi,%esi /* no error code */ + movq %gs:pda_data_offset, %rbp + subq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) + call \do_sym + addq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) + jmp paranoid_exit /* %ebx: no swapgs flag */ + CFI_ENDPROC +END(\sym) +.endm - /* - * "Paranoid" exit path from exception stack. - * Paranoid because this is used by NMIs and cannot take - * any kernel state for granted. - * We don't do kernel preemption checks here, because only - * NMI should be common and it does not enable IRQs and - * cannot get reschedule ticks. - * - * "trace" is 0 for the NMI handler only, because irq-tracing - * is fundamentally NMI-unsafe. (we cannot change the soft and - * hard flags at once, atomically) - */ - .macro paranoidexit trace=1 - /* ebx: no swapgs flag */ -paranoid_exit\trace: - testl %ebx,%ebx /* swapgs needed? */ - jnz paranoid_restore\trace - testl $3,CS(%rsp) - jnz paranoid_userspace\trace -paranoid_swapgs\trace: - .if \trace - TRACE_IRQS_IRETQ 0 - .endif - SWAPGS_UNSAFE_STACK -paranoid_restore\trace: - RESTORE_ALL 8 - jmp irq_return -paranoid_userspace\trace: - GET_THREAD_INFO(%rcx) - movl TI_flags(%rcx),%ebx - andl $_TIF_WORK_MASK,%ebx - jz paranoid_swapgs\trace - movq %rsp,%rdi /* &pt_regs */ - call sync_regs - movq %rax,%rsp /* switch stack for scheduling */ - testl $_TIF_NEED_RESCHED,%ebx - jnz paranoid_schedule\trace - movl %ebx,%edx /* arg3: thread flags */ - .if \trace - TRACE_IRQS_ON - .endif - ENABLE_INTERRUPTS(CLBR_NONE) - xorl %esi,%esi /* arg2: oldset */ - movq %rsp,%rdi /* arg1: &pt_regs */ - call do_notify_resume - DISABLE_INTERRUPTS(CLBR_NONE) - .if \trace - TRACE_IRQS_OFF - .endif - jmp paranoid_userspace\trace -paranoid_schedule\trace: - .if \trace - TRACE_IRQS_ON - .endif - ENABLE_INTERRUPTS(CLBR_ANY) - call schedule - DISABLE_INTERRUPTS(CLBR_ANY) - .if \trace - TRACE_IRQS_OFF - .endif - jmp paranoid_userspace\trace +.macro errorentry sym do_sym +ENTRY(\sym) + XCPT_FRAME + PARAVIRT_ADJUST_EXCEPTION_FRAME + subq $15*8,%rsp + CFI_ADJUST_CFA_OFFSET 15*8 + call error_entry + DEFAULT_FRAME 0 + movq %rsp,%rdi /* pt_regs pointer */ + movq ORIG_RAX(%rsp),%rsi /* get error code */ + movq $-1,ORIG_RAX(%rsp) /* no syscall to restart */ + call \do_sym + jmp error_exit /* %ebx: no swapgs flag */ CFI_ENDPROC - .endm +END(\sym) +.endm -/* - * Exception entry point. This expects an error code/orig_rax on the stack - * and the exception handler in %rax. - */ -KPROBE_ENTRY(error_entry) - _frame RDI - CFI_REL_OFFSET rax,0 - /* rdi slot contains rax, oldrax contains error code */ - cld - subq $14*8,%rsp - CFI_ADJUST_CFA_OFFSET (14*8) - movq %rsi,13*8(%rsp) - CFI_REL_OFFSET rsi,RSI - movq 14*8(%rsp),%rsi /* load rax from rdi slot */ - CFI_REGISTER rax,rsi - movq %rdx,12*8(%rsp) - CFI_REL_OFFSET rdx,RDX - movq %rcx,11*8(%rsp) - CFI_REL_OFFSET rcx,RCX - movq %rsi,10*8(%rsp) /* store rax */ - CFI_REL_OFFSET rax,RAX - movq %r8, 9*8(%rsp) - CFI_REL_OFFSET r8,R8 - movq %r9, 8*8(%rsp) - CFI_REL_OFFSET r9,R9 - movq %r10,7*8(%rsp) - CFI_REL_OFFSET r10,R10 - movq %r11,6*8(%rsp) - CFI_REL_OFFSET r11,R11 - movq %rbx,5*8(%rsp) - CFI_REL_OFFSET rbx,RBX - movq %rbp,4*8(%rsp) - CFI_REL_OFFSET rbp,RBP - movq %r12,3*8(%rsp) - CFI_REL_OFFSET r12,R12 - movq %r13,2*8(%rsp) - CFI_REL_OFFSET r13,R13 - movq %r14,1*8(%rsp) - CFI_REL_OFFSET r14,R14 - movq %r15,(%rsp) - CFI_REL_OFFSET r15,R15 - xorl %ebx,%ebx - testl $3,CS(%rsp) - je error_kernelspace -error_swapgs: - SWAPGS -error_sti: - TRACE_IRQS_OFF - movq %rdi,RDI(%rsp) - CFI_REL_OFFSET rdi,RDI - movq %rsp,%rdi - movq ORIG_RAX(%rsp),%rsi /* get error code */ - movq $-1,ORIG_RAX(%rsp) - call *%rax - /* ebx: no swapgs flag (1: don't need swapgs, 0: need it) */ -error_exit: - movl %ebx,%eax - RESTORE_REST - DISABLE_INTERRUPTS(CLBR_NONE) + /* error code is on the stack already */ +.macro paranoiderrorentry sym do_sym +ENTRY(\sym) + XCPT_FRAME + PARAVIRT_ADJUST_EXCEPTION_FRAME + subq $15*8,%rsp + CFI_ADJUST_CFA_OFFSET 15*8 + call save_paranoid + DEFAULT_FRAME 0 TRACE_IRQS_OFF - GET_THREAD_INFO(%rcx) - testl %eax,%eax - jne retint_kernel - LOCKDEP_SYS_EXIT_IRQ - movl TI_flags(%rcx),%edx - movl $_TIF_WORK_MASK,%edi - andl %edi,%edx - jnz retint_careful - jmp retint_swapgs + movq %rsp,%rdi /* pt_regs pointer */ + movq ORIG_RAX(%rsp),%rsi /* get error code */ + movq $-1,ORIG_RAX(%rsp) /* no syscall to restart */ + call \do_sym + jmp paranoid_exit /* %ebx: no swapgs flag */ CFI_ENDPROC +END(\sym) +.endm -error_kernelspace: - incl %ebx - /* There are two places in the kernel that can potentially fault with - usergs. Handle them here. The exception handlers after - iret run with kernel gs again, so don't set the user space flag. - B stepping K8s sometimes report an truncated RIP for IRET - exceptions returning to compat mode. Check for these here too. */ - leaq irq_return(%rip),%rcx - cmpq %rcx,RIP(%rsp) - je error_swapgs - movl %ecx,%ecx /* zero extend */ - cmpq %rcx,RIP(%rsp) - je error_swapgs - cmpq $gs_change,RIP(%rsp) - je error_swapgs - jmp error_sti -KPROBE_END(error_entry) - - /* Reload gs selector with exception handling */ - /* edi: new selector */ +zeroentry divide_error do_divide_error +zeroentry overflow do_overflow +zeroentry bounds do_bounds +zeroentry invalid_op do_invalid_op +zeroentry device_not_available do_device_not_available +paranoiderrorentry double_fault do_double_fault +zeroentry coprocessor_segment_overrun do_coprocessor_segment_overrun +errorentry invalid_TSS do_invalid_TSS +errorentry segment_not_present do_segment_not_present +zeroentry spurious_interrupt_bug do_spurious_interrupt_bug +zeroentry coprocessor_error do_coprocessor_error +errorentry alignment_check do_alignment_check +zeroentry simd_coprocessor_error do_simd_coprocessor_error + + /* Reload gs selector with exception handling */ + /* edi: new selector */ ENTRY(native_load_gs_index) CFI_STARTPROC pushf CFI_ADJUST_CFA_OFFSET 8 DISABLE_INTERRUPTS(CLBR_ANY | ~(CLBR_RDI)) - SWAPGS -gs_change: - movl %edi,%gs + SWAPGS +gs_change: + movl %edi,%gs 2: mfence /* workaround */ SWAPGS - popf + popf CFI_ADJUST_CFA_OFFSET -8 - ret + ret CFI_ENDPROC -ENDPROC(native_load_gs_index) - - .section __ex_table,"a" - .align 8 - .quad gs_change,bad_gs - .previous - .section .fixup,"ax" +END(native_load_gs_index) + + .section __ex_table,"a" + .align 8 + .quad gs_change,bad_gs + .previous + .section .fixup,"ax" /* running with kernelgs */ -bad_gs: +bad_gs: SWAPGS /* switch back to user gs */ xorl %eax,%eax - movl %eax,%gs - jmp 2b - .previous - + movl %eax,%gs + jmp 2b + .previous + /* * Create a kernel thread. * @@ -1138,7 +1184,7 @@ ENTRY(kernel_thread) xorl %r8d,%r8d xorl %r9d,%r9d - + # clone now call do_fork movq %rax,RAX(%rsp) @@ -1149,15 +1195,15 @@ ENTRY(kernel_thread) * so internally to the x86_64 port you can rely on kernel_thread() * not to reschedule the child before returning, this avoids the need * of hacks for example to fork off the per-CPU idle tasks. - * [Hopefully no generic code relies on the reschedule -AK] + * [Hopefully no generic code relies on the reschedule -AK] */ RESTORE_ALL UNFAKE_STACK_FRAME ret CFI_ENDPROC -ENDPROC(kernel_thread) - -child_rip: +END(kernel_thread) + +ENTRY(child_rip) pushq $0 # fake return address CFI_STARTPROC /* @@ -1170,8 +1216,9 @@ child_rip: # exit mov %eax, %edi call do_exit + ud2 # padding for call trace CFI_ENDPROC -ENDPROC(child_rip) +END(child_rip) /* * execve(). This function needs to use IRET, not SYSRET, to set up all state properly. @@ -1191,10 +1238,10 @@ ENDPROC(child_rip) ENTRY(kernel_execve) CFI_STARTPROC FAKE_STACK_FRAME $0 - SAVE_ALL + SAVE_ALL movq %rsp,%rcx call sys_execve - movq %rax, RAX(%rsp) + movq %rax, RAX(%rsp) RESTORE_REST testq %rax,%rax je int_ret_from_sys_call @@ -1202,129 +1249,7 @@ ENTRY(kernel_execve) UNFAKE_STACK_FRAME ret CFI_ENDPROC -ENDPROC(kernel_execve) - -KPROBE_ENTRY(page_fault) - errorentry do_page_fault -KPROBE_END(page_fault) - -ENTRY(coprocessor_error) - zeroentry do_coprocessor_error -END(coprocessor_error) - -ENTRY(simd_coprocessor_error) - zeroentry do_simd_coprocessor_error -END(simd_coprocessor_error) - -ENTRY(device_not_available) - zeroentry do_device_not_available -END(device_not_available) - - /* runs on exception stack */ -KPROBE_ENTRY(debug) - INTR_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 - paranoidentry do_debug, DEBUG_STACK - paranoidexit -KPROBE_END(debug) - - /* runs on exception stack */ -KPROBE_ENTRY(nmi) - INTR_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $-1 - CFI_ADJUST_CFA_OFFSET 8 - paranoidentry do_nmi, 0, 0 -#ifdef CONFIG_TRACE_IRQFLAGS - paranoidexit 0 -#else - jmp paranoid_exit1 - CFI_ENDPROC -#endif -KPROBE_END(nmi) - -KPROBE_ENTRY(int3) - INTR_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 - paranoidentry do_int3, DEBUG_STACK - jmp paranoid_exit1 - CFI_ENDPROC -KPROBE_END(int3) - -ENTRY(overflow) - zeroentry do_overflow -END(overflow) - -ENTRY(bounds) - zeroentry do_bounds -END(bounds) - -ENTRY(invalid_op) - zeroentry do_invalid_op -END(invalid_op) - -ENTRY(coprocessor_segment_overrun) - zeroentry do_coprocessor_segment_overrun -END(coprocessor_segment_overrun) - - /* runs on exception stack */ -ENTRY(double_fault) - XCPT_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - paranoidentry do_double_fault - jmp paranoid_exit1 - CFI_ENDPROC -END(double_fault) - -ENTRY(invalid_TSS) - errorentry do_invalid_TSS -END(invalid_TSS) - -ENTRY(segment_not_present) - errorentry do_segment_not_present -END(segment_not_present) - - /* runs on exception stack */ -ENTRY(stack_segment) - XCPT_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - paranoidentry do_stack_segment - jmp paranoid_exit1 - CFI_ENDPROC -END(stack_segment) - -KPROBE_ENTRY(general_protection) - errorentry do_general_protection -KPROBE_END(general_protection) - -ENTRY(alignment_check) - errorentry do_alignment_check -END(alignment_check) - -ENTRY(divide_error) - zeroentry do_divide_error -END(divide_error) - -ENTRY(spurious_interrupt_bug) - zeroentry do_spurious_interrupt_bug -END(spurious_interrupt_bug) - -#ifdef CONFIG_X86_MCE - /* runs on exception stack */ -ENTRY(machine_check) - INTR_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 - paranoidentry do_machine_check - jmp paranoid_exit1 - CFI_ENDPROC -END(machine_check) -#endif +END(kernel_execve) /* Call softirq on interrupt stack. Interrupts are off. */ ENTRY(call_softirq) @@ -1344,40 +1269,33 @@ ENTRY(call_softirq) decl %gs:pda_irqcount ret CFI_ENDPROC -ENDPROC(call_softirq) - -KPROBE_ENTRY(ignore_sysret) - CFI_STARTPROC - mov $-ENOSYS,%eax - sysret - CFI_ENDPROC -ENDPROC(ignore_sysret) +END(call_softirq) #ifdef CONFIG_XEN -ENTRY(xen_hypervisor_callback) - zeroentry xen_do_hypervisor_callback -END(xen_hypervisor_callback) +zeroentry xen_hypervisor_callback xen_do_hypervisor_callback /* -# A note on the "critical region" in our callback handler. -# We want to avoid stacking callback handlers due to events occurring -# during handling of the last event. To do this, we keep events disabled -# until we've done all processing. HOWEVER, we must enable events before -# popping the stack frame (can't be done atomically) and so it would still -# be possible to get enough handler activations to overflow the stack. -# Although unlikely, bugs of that kind are hard to track down, so we'd -# like to avoid the possibility. -# So, on entry to the handler we detect whether we interrupted an -# existing activation in its critical region -- if so, we pop the current -# activation and restart the handler using the previous one. -*/ + * A note on the "critical region" in our callback handler. + * We want to avoid stacking callback handlers due to events occurring + * during handling of the last event. To do this, we keep events disabled + * until we've done all processing. HOWEVER, we must enable events before + * popping the stack frame (can't be done atomically) and so it would still + * be possible to get enough handler activations to overflow the stack. + * Although unlikely, bugs of that kind are hard to track down, so we'd + * like to avoid the possibility. + * So, on entry to the handler we detect whether we interrupted an + * existing activation in its critical region -- if so, we pop the current + * activation and restart the handler using the previous one. + */ ENTRY(xen_do_hypervisor_callback) # do_hypervisor_callback(struct *pt_regs) CFI_STARTPROC -/* Since we don't modify %rdi, evtchn_do_upall(struct *pt_regs) will - see the correct pointer to the pt_regs */ +/* + * Since we don't modify %rdi, evtchn_do_upall(struct *pt_regs) will + * see the correct pointer to the pt_regs + */ movq %rdi, %rsp # we don't return, adjust the stack frame CFI_ENDPROC - CFI_DEFAULT_STACK + DEFAULT_FRAME 11: incl %gs:pda_irqcount movq %rsp,%rbp CFI_DEF_CFA_REGISTER rbp @@ -1392,23 +1310,26 @@ ENTRY(xen_do_hypervisor_callback) # do_hypervisor_callback(struct *pt_regs) END(do_hypervisor_callback) /* -# Hypervisor uses this for application faults while it executes. -# We get here for two reasons: -# 1. Fault while reloading DS, ES, FS or GS -# 2. Fault while executing IRET -# Category 1 we do not need to fix up as Xen has already reloaded all segment -# registers that could be reloaded and zeroed the others. -# Category 2 we fix up by killing the current process. We cannot use the -# normal Linux return path in this case because if we use the IRET hypercall -# to pop the stack frame we end up in an infinite loop of failsafe callbacks. -# We distinguish between categories by comparing each saved segment register -# with its current contents: any discrepancy means we in category 1. -*/ + * Hypervisor uses this for application faults while it executes. + * We get here for two reasons: + * 1. Fault while reloading DS, ES, FS or GS + * 2. Fault while executing IRET + * Category 1 we do not need to fix up as Xen has already reloaded all segment + * registers that could be reloaded and zeroed the others. + * Category 2 we fix up by killing the current process. We cannot use the + * normal Linux return path in this case because if we use the IRET hypercall + * to pop the stack frame we end up in an infinite loop of failsafe callbacks. + * We distinguish between categories by comparing each saved segment register + * with its current contents: any discrepancy means we in category 1. + */ ENTRY(xen_failsafe_callback) - framesz = (RIP-0x30) /* workaround buggy gas */ - _frame framesz - CFI_REL_OFFSET rcx, 0 - CFI_REL_OFFSET r11, 8 + INTR_FRAME 1 (6*8) + /*CFI_REL_OFFSET gs,GS*/ + /*CFI_REL_OFFSET fs,FS*/ + /*CFI_REL_OFFSET es,ES*/ + /*CFI_REL_OFFSET ds,DS*/ + CFI_REL_OFFSET r11,8 + CFI_REL_OFFSET rcx,0 movw %ds,%cx cmpw %cx,0x10(%rsp) CFI_REMEMBER_STATE @@ -1429,12 +1350,9 @@ ENTRY(xen_failsafe_callback) CFI_RESTORE r11 addq $0x30,%rsp CFI_ADJUST_CFA_OFFSET -0x30 - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 - pushq %r11 - CFI_ADJUST_CFA_OFFSET 8 - pushq %rcx - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi $0 /* RIP */ + pushq_cfi %r11 + pushq_cfi %rcx jmp general_protection CFI_RESTORE_STATE 1: /* Segment mismatch => Category 1 (Bad segment). Retry the IRET. */ @@ -1444,11 +1362,223 @@ ENTRY(xen_failsafe_callback) CFI_RESTORE r11 addq $0x30,%rsp CFI_ADJUST_CFA_OFFSET -0x30 - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi $0 SAVE_ALL jmp error_exit CFI_ENDPROC END(xen_failsafe_callback) #endif /* CONFIG_XEN */ + +/* + * Some functions should be protected against kprobes + */ + .pushsection .kprobes.text, "ax" + +paranoidzeroentry_ist debug do_debug DEBUG_STACK +paranoidzeroentry_ist int3 do_int3 DEBUG_STACK +paranoiderrorentry stack_segment do_stack_segment +errorentry general_protection do_general_protection +errorentry page_fault do_page_fault +#ifdef CONFIG_X86_MCE +paranoidzeroentry machine_check do_machine_check +#endif + + /* + * "Paranoid" exit path from exception stack. + * Paranoid because this is used by NMIs and cannot take + * any kernel state for granted. + * We don't do kernel preemption checks here, because only + * NMI should be common and it does not enable IRQs and + * cannot get reschedule ticks. + * + * "trace" is 0 for the NMI handler only, because irq-tracing + * is fundamentally NMI-unsafe. (we cannot change the soft and + * hard flags at once, atomically) + */ + + /* ebx: no swapgs flag */ +ENTRY(paranoid_exit) + INTR_FRAME + DISABLE_INTERRUPTS(CLBR_NONE) + TRACE_IRQS_OFF + testl %ebx,%ebx /* swapgs needed? */ + jnz paranoid_restore + testl $3,CS(%rsp) + jnz paranoid_userspace +paranoid_swapgs: + TRACE_IRQS_IRETQ 0 + SWAPGS_UNSAFE_STACK +paranoid_restore: + RESTORE_ALL 8 + jmp irq_return +paranoid_userspace: + GET_THREAD_INFO(%rcx) + movl TI_flags(%rcx),%ebx + andl $_TIF_WORK_MASK,%ebx + jz paranoid_swapgs + movq %rsp,%rdi /* &pt_regs */ + call sync_regs + movq %rax,%rsp /* switch stack for scheduling */ + testl $_TIF_NEED_RESCHED,%ebx + jnz paranoid_schedule + movl %ebx,%edx /* arg3: thread flags */ + TRACE_IRQS_ON + ENABLE_INTERRUPTS(CLBR_NONE) + xorl %esi,%esi /* arg2: oldset */ + movq %rsp,%rdi /* arg1: &pt_regs */ + call do_notify_resume + DISABLE_INTERRUPTS(CLBR_NONE) + TRACE_IRQS_OFF + jmp paranoid_userspace +paranoid_schedule: + TRACE_IRQS_ON + ENABLE_INTERRUPTS(CLBR_ANY) + call schedule + DISABLE_INTERRUPTS(CLBR_ANY) + TRACE_IRQS_OFF + jmp paranoid_userspace + CFI_ENDPROC +END(paranoid_exit) + +/* + * Exception entry point. This expects an error code/orig_rax on the stack. + * returns in "no swapgs flag" in %ebx. + */ +ENTRY(error_entry) + XCPT_FRAME + CFI_ADJUST_CFA_OFFSET 15*8 + /* oldrax contains error code */ + cld + movq_cfi rdi, RDI+8 + movq_cfi rsi, RSI+8 + movq_cfi rdx, RDX+8 + movq_cfi rcx, RCX+8 + movq_cfi rax, RAX+8 + movq_cfi r8, R8+8 + movq_cfi r9, R9+8 + movq_cfi r10, R10+8 + movq_cfi r11, R11+8 + movq_cfi rbx, RBX+8 + movq_cfi rbp, RBP+8 + movq_cfi r12, R12+8 + movq_cfi r13, R13+8 + movq_cfi r14, R14+8 + movq_cfi r15, R15+8 + xorl %ebx,%ebx + testl $3,CS+8(%rsp) + je error_kernelspace +error_swapgs: + SWAPGS +error_sti: + TRACE_IRQS_OFF + ret + CFI_ENDPROC + +/* + * There are two places in the kernel that can potentially fault with + * usergs. Handle them here. The exception handlers after iret run with + * kernel gs again, so don't set the user space flag. B stepping K8s + * sometimes report an truncated RIP for IRET exceptions returning to + * compat mode. Check for these here too. + */ +error_kernelspace: + incl %ebx + leaq irq_return(%rip),%rcx + cmpq %rcx,RIP+8(%rsp) + je error_swapgs + movl %ecx,%ecx /* zero extend */ + cmpq %rcx,RIP+8(%rsp) + je error_swapgs + cmpq $gs_change,RIP+8(%rsp) + je error_swapgs + jmp error_sti +END(error_entry) + + +/* ebx: no swapgs flag (1: don't need swapgs, 0: need it) */ +ENTRY(error_exit) + DEFAULT_FRAME + movl %ebx,%eax + RESTORE_REST + DISABLE_INTERRUPTS(CLBR_NONE) + TRACE_IRQS_OFF + GET_THREAD_INFO(%rcx) + testl %eax,%eax + jne retint_kernel + LOCKDEP_SYS_EXIT_IRQ + movl TI_flags(%rcx),%edx + movl $_TIF_WORK_MASK,%edi + andl %edi,%edx + jnz retint_careful + jmp retint_swapgs + CFI_ENDPROC +END(error_exit) + + + /* runs on exception stack */ +ENTRY(nmi) + INTR_FRAME + PARAVIRT_ADJUST_EXCEPTION_FRAME + pushq_cfi $-1 + subq $15*8, %rsp + CFI_ADJUST_CFA_OFFSET 15*8 + call save_paranoid + DEFAULT_FRAME 0 + /* paranoidentry do_nmi, 0; without TRACE_IRQS_OFF */ + movq %rsp,%rdi + movq $-1,%rsi + call do_nmi +#ifdef CONFIG_TRACE_IRQFLAGS + /* paranoidexit; without TRACE_IRQS_OFF */ + /* ebx: no swapgs flag */ + DISABLE_INTERRUPTS(CLBR_NONE) + testl %ebx,%ebx /* swapgs needed? */ + jnz nmi_restore + testl $3,CS(%rsp) + jnz nmi_userspace +nmi_swapgs: + SWAPGS_UNSAFE_STACK +nmi_restore: + RESTORE_ALL 8 + jmp irq_return +nmi_userspace: + GET_THREAD_INFO(%rcx) + movl TI_flags(%rcx),%ebx + andl $_TIF_WORK_MASK,%ebx + jz nmi_swapgs + movq %rsp,%rdi /* &pt_regs */ + call sync_regs + movq %rax,%rsp /* switch stack for scheduling */ + testl $_TIF_NEED_RESCHED,%ebx + jnz nmi_schedule + movl %ebx,%edx /* arg3: thread flags */ + ENABLE_INTERRUPTS(CLBR_NONE) + xorl %esi,%esi /* arg2: oldset */ + movq %rsp,%rdi /* arg1: &pt_regs */ + call do_notify_resume + DISABLE_INTERRUPTS(CLBR_NONE) + jmp nmi_userspace +nmi_schedule: + ENABLE_INTERRUPTS(CLBR_ANY) + call schedule + DISABLE_INTERRUPTS(CLBR_ANY) + jmp nmi_userspace + CFI_ENDPROC +#else + jmp paranoid_exit + CFI_ENDPROC +#endif +END(nmi) + +ENTRY(ignore_sysret) + CFI_STARTPROC + mov $-ENOSYS,%eax + sysret + CFI_ENDPROC +END(ignore_sysret) + +/* + * End of kprobes section + */ + .popsection diff --git a/arch/x86/kernel/es7000_32.c b/arch/x86/kernel/es7000_32.c index 0aa2c443d60..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; } /* @@ -317,26 +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..34185488e4f 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 +259,31 @@ 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. + */ + for_each_cpu_and(cpu, cpumask, andmask) + if (cpumask_test_cpu(cpu, cpu_online_mask)) + break; + 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 +296,10 @@ 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..6ce497cc372 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) @@ -89,21 +109,38 @@ 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; /* - * We're using fixed IRQ delivery, can only return one phys APIC ID. + * We're using fixed IRQ delivery, can only return one logical 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 logical APIC ID. + * May as well be the first. + */ + for_each_cpu_and(cpu, cpumask, andmask) + if (cpumask_test_cpu(cpu, cpu_online_mask)) + break; + if (cpu < nr_cpu_ids) + return per_cpu(x86_cpu_to_logical_apicid, cpu); + return BAD_APICID; +} + static unsigned int get_apic_id(unsigned long x) { unsigned int id; @@ -150,8 +187,10 @@ 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..21bcc0e098b 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 +116,30 @@ 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. + */ + for_each_cpu_and(cpu, cpumask, andmask) + if (cpumask_test_cpu(cpu, cpu_online_mask)) + break; + 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; @@ -123,12 +161,12 @@ static unsigned int phys_pkg_id(int index_msb) return current_cpu_data.initial_apicid >> index_msb; } -void x2apic_send_IPI_self(int vector) +static void x2apic_send_IPI_self(int vector) { apic_write(APIC_SELF_IPI, vector); } -void init_x2apic_ldr(void) +static void init_x2apic_ldr(void) { return; } @@ -145,8 +183,10 @@ 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..b193e082f6c 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -10,6 +10,7 @@ #include <linux/kernel.h> #include <linux/threads.h> +#include <linux/cpu.h> #include <linux/cpumask.h> #include <linux/string.h> #include <linux/ctype.h> @@ -17,6 +18,9 @@ #include <linux/sched.h> #include <linux/module.h> #include <linux/hardirq.h> +#include <linux/timer.h> +#include <linux/proc_fs.h> +#include <asm/current.h> #include <asm/smp.h> #include <asm/ipi.h> #include <asm/genapic.h> @@ -75,16 +79,15 @@ 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 +126,37 @@ 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 +168,7 @@ 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 +176,30 @@ 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. + */ + for_each_cpu_and(cpu, cpumask, andmask) + if (cpumask_test_cpu(cpu, cpu_online_mask)) + break; + 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 +247,10 @@ 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, @@ -356,6 +387,103 @@ static __init void uv_rtc_init(void) } /* + * percpu heartbeat timer + */ +static void uv_heartbeat(unsigned long ignored) +{ + struct timer_list *timer = &uv_hub_info->scir.timer; + unsigned char bits = uv_hub_info->scir.state; + + /* flip heartbeat bit */ + bits ^= SCIR_CPU_HEARTBEAT; + + /* is this cpu idle? */ + if (idle_cpu(raw_smp_processor_id())) + bits &= ~SCIR_CPU_ACTIVITY; + else + bits |= SCIR_CPU_ACTIVITY; + + /* update system controller interface reg */ + uv_set_scir_bits(bits); + + /* enable next timer period */ + mod_timer(timer, jiffies + SCIR_CPU_HB_INTERVAL); +} + +static void __cpuinit uv_heartbeat_enable(int cpu) +{ + if (!uv_cpu_hub_info(cpu)->scir.enabled) { + struct timer_list *timer = &uv_cpu_hub_info(cpu)->scir.timer; + + uv_set_cpu_scir_bits(cpu, SCIR_CPU_HEARTBEAT|SCIR_CPU_ACTIVITY); + setup_timer(timer, uv_heartbeat, cpu); + timer->expires = jiffies + SCIR_CPU_HB_INTERVAL; + add_timer_on(timer, cpu); + uv_cpu_hub_info(cpu)->scir.enabled = 1; + } + + /* check boot cpu */ + if (!uv_cpu_hub_info(0)->scir.enabled) + uv_heartbeat_enable(0); +} + +#ifdef CONFIG_HOTPLUG_CPU +static void __cpuinit uv_heartbeat_disable(int cpu) +{ + if (uv_cpu_hub_info(cpu)->scir.enabled) { + uv_cpu_hub_info(cpu)->scir.enabled = 0; + del_timer(&uv_cpu_hub_info(cpu)->scir.timer); + } + uv_set_cpu_scir_bits(cpu, 0xff); +} + +/* + * cpu hotplug notifier + */ +static __cpuinit int uv_scir_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu) +{ + long cpu = (long)hcpu; + + switch (action) { + case CPU_ONLINE: + uv_heartbeat_enable(cpu); + break; + case CPU_DOWN_PREPARE: + uv_heartbeat_disable(cpu); + break; + default: + break; + } + return NOTIFY_OK; +} + +static __init void uv_scir_register_cpu_notifier(void) +{ + hotcpu_notifier(uv_scir_cpu_notify, 0); +} + +#else /* !CONFIG_HOTPLUG_CPU */ + +static __init void uv_scir_register_cpu_notifier(void) +{ +} + +static __init int uv_init_heartbeat(void) +{ + int cpu; + + if (is_uv_system()) + for_each_online_cpu(cpu) + uv_heartbeat_enable(cpu); + return 0; +} + +late_initcall(uv_init_heartbeat); + +#endif /* !CONFIG_HOTPLUG_CPU */ + +/* * Called on each cpu to initialize the per_cpu UV data area. * ZZZ hotplug not supported yet */ @@ -428,7 +556,7 @@ void __init uv_system_init(void) uv_bios_init(); uv_bios_get_sn_info(0, &uv_type, &sn_partition_id, - &uv_coherency_id, &uv_region_size); + &sn_coherency_id, &sn_region_size); uv_rtc_init(); for_each_present_cpu(cpu) { @@ -439,8 +567,7 @@ void __init uv_system_init(void) uv_blade_info[blade].nr_possible_cpus++; uv_cpu_hub_info(cpu)->lowmem_remap_base = lowmem_redir_base; - uv_cpu_hub_info(cpu)->lowmem_remap_top = - lowmem_redir_base + lowmem_redir_size; + uv_cpu_hub_info(cpu)->lowmem_remap_top = lowmem_redir_size; uv_cpu_hub_info(cpu)->m_val = m_val; uv_cpu_hub_info(cpu)->n_val = m_val; uv_cpu_hub_info(cpu)->numa_blade_id = blade; @@ -450,7 +577,8 @@ void __init uv_system_init(void) uv_cpu_hub_info(cpu)->gpa_mask = (1 << (m_val + n_val)) - 1; uv_cpu_hub_info(cpu)->gnode_upper = gnode_upper; uv_cpu_hub_info(cpu)->global_mmr_base = mmr_base; - uv_cpu_hub_info(cpu)->coherency_domain_number = uv_coherency_id; + uv_cpu_hub_info(cpu)->coherency_domain_number = sn_coherency_id; + uv_cpu_hub_info(cpu)->scir.offset = SCIR_LOCAL_MMR_BASE + lcpu; uv_node_to_blade[nid] = blade; uv_cpu_to_blade[cpu] = blade; max_pnode = max(pnode, max_pnode); @@ -467,4 +595,6 @@ void __init uv_system_init(void) map_mmioh_high(max_pnode); uv_cpu_init(); + uv_scir_register_cpu_notifier(); + proc_mkdir("sgi_uv", NULL); } diff --git a/arch/x86/kernel/head.c b/arch/x86/kernel/head.c index 1dcb0f13897..3e66bd364a9 100644 --- a/arch/x86/kernel/head.c +++ b/arch/x86/kernel/head.c @@ -35,7 +35,6 @@ void __init reserve_ebda_region(void) /* start of EBDA area */ ebda_addr = get_bios_ebda(); - printk(KERN_INFO "BIOS EBDA/lowmem at: %08x/%08x\n", ebda_addr, lowmem); /* Fixup: bios puts an EBDA in the top 64K segment */ /* of conventional memory, but does not adjust lowmem. */ diff --git a/arch/x86/kernel/head32.c b/arch/x86/kernel/head32.c index fa1d25dd83e..ac108d1fe18 100644 --- a/arch/x86/kernel/head32.c +++ b/arch/x86/kernel/head32.c @@ -12,9 +12,12 @@ #include <asm/sections.h> #include <asm/e820.h> #include <asm/bios_ebda.h> +#include <asm/trampoline.h> void __init i386_start_kernel(void) { + reserve_trampoline_memory(); + reserve_early(__pa_symbol(&_text), __pa_symbol(&_end), "TEXT DATA BSS"); #ifdef CONFIG_BLK_DEV_INITRD diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c index d16084f9064..b9a4d8c4b93 100644 --- a/arch/x86/kernel/head64.c +++ b/arch/x86/kernel/head64.c @@ -24,9 +24,10 @@ #include <asm/kdebug.h> #include <asm/e820.h> #include <asm/bios_ebda.h> +#include <asm/trampoline.h> /* boot cpu pda */ -static struct x8664_pda _boot_cpu_pda __read_mostly; +static struct x8664_pda _boot_cpu_pda; #ifdef CONFIG_SMP /* @@ -120,6 +121,8 @@ void __init x86_64_start_reservations(char *real_mode_data) { copy_bootdata(__va(real_mode_data)); + reserve_trampoline_memory(); + reserve_early(__pa_symbol(&_text), __pa_symbol(&_end), "TEXT DATA BSS"); #ifdef CONFIG_BLK_DEV_INITRD diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index 067d8de913f..cd759ad9069 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -33,7 +33,9 @@ * HPET address is set in acpi/boot.c, when an ACPI entry exists */ unsigned long hpet_address; -unsigned long hpet_num_timers; +#ifdef CONFIG_PCI_MSI +static unsigned long hpet_num_timers; +#endif static void __iomem *hpet_virt_address; struct hpet_dev { @@ -246,7 +248,7 @@ 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 +303,7 @@ 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; @@ -449,7 +451,7 @@ static int hpet_setup_irq(struct hpet_dev *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 +502,7 @@ 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); } @@ -811,7 +813,7 @@ int __init hpet_enable(void) out_nohpet: hpet_clear_mapping(); - boot_hpet_disable = 1; + hpet_address = 0; return 0; } @@ -834,10 +836,11 @@ static __init int hpet_late_init(void) hpet_address = force_hpet_address; hpet_enable(); - if (!hpet_virt_address) - return -ENODEV; } + if (!hpet_virt_address) + return -ENODEV; + hpet_reserve_platform_timers(hpet_readl(HPET_ID)); for_each_online_cpu(cpu) { 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/init_task.c b/arch/x86/kernel/init_task.c index a4f93b4120c..df3bf269bea 100644 --- a/arch/x86/kernel/init_task.c +++ b/arch/x86/kernel/init_task.c @@ -10,11 +10,9 @@ #include <asm/pgtable.h> #include <asm/desc.h> -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); -EXPORT_UNUSED_SYMBOL(init_mm); /* will be removed in 2.6.26 */ /* * Initial thread structure. diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 9043251210f..3639442aa7a 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -108,93 +108,276 @@ 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++) +int __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); + } + + return 0; +} + +#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) { + if (!alloc_cpumask_var_node(&cfg->domain, GFP_ATOMIC, node)) { + kfree(cfg); + cfg = NULL; + } else if (!alloc_cpumask_var_node(&cfg->old_domain, + GFP_ATOMIC, node)) { + 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) +int 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; -}; + return 0; +} -static struct irq_pin_list irq_2_pin_head[PIN_MAP_SIZE]; -static struct irq_pin_list *irq_2_pin_ptr; +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC -static void __init irq_2_pin_init(void) +static void +init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu) { - struct irq_pin_list *pin = irq_2_pin_head; - int i; + struct irq_pin_list *old_entry, *head, *tail, *entry; + + 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; + } - for (i = 1; i < PIN_MAP_SIZE; i++) - pin[i-1].next = &pin[i]; + 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 *entry, *next; + + if (old_cfg->irq_2_pin == cfg->irq_2_pin) + return; - irq_2_pin_ptr = &pin[0]; + entry = old_cfg->irq_2_pin; + + 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; @@ -237,11 +420,10 @@ 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 +505,32 @@ 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 +560,61 @@ 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); -static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t 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; + + 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 +623,18 @@ 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 +649,7 @@ 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 +658,10 @@ 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 +678,16 @@ 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; @@ -475,13 +700,13 @@ 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 -void io_apic_sync(struct irq_pin_list *entry) +static void io_apic_sync(struct irq_pin_list *entry) { /* * Synchronize the IO-APIC and the CPU by doing @@ -492,47 +717,64 @@ 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 +1051,7 @@ 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 +1276,8 @@ 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 @@ -1049,52 +1292,49 @@ 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; } if (unlikely(current_vector == vector)) continue; -#ifdef CONFIG_X86_64 - if (vector == IA32_SYSCALL_VECTOR) - goto next; -#else - if (vector == SYSCALL_VECTOR) + + if (test_bit(vector, used_vectors)) 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! */ @@ -1102,49 +1342,47 @@ 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; - cpus_and(mask, cfg->old_domain, cpu_online_map); - for_each_cpu_mask_nr(cpu, mask) { + 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) @@ -1162,10 +1400,12 @@ 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) { + cfg = desc->chip_data; + if (!cpumask_test_cpu(cpu, cfg->domain)) continue; vector = cfg->vector; per_cpu(vector_irq, cpu)[vector] = irq; @@ -1177,7 +1417,7 @@ 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; } } @@ -1215,11 +1455,8 @@ 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) @@ -1311,23 +1548,22 @@ 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 -> " @@ -1337,16 +1573,15 @@ 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); @@ -1356,6 +1591,9 @@ 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"); @@ -1387,9 +1625,15 @@ 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)); } } @@ -1448,6 +1692,7 @@ __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) @@ -1537,8 +1782,11 @@ __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; + + cfg = desc->chip_data; + entry = cfg->irq_2_pin; if (!entry) continue; printk(KERN_DEBUG "IRQ%d ", irq); @@ -2022,14 +2270,16 @@ 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; @@ -2043,7 +2293,7 @@ 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; @@ -2092,35 +2342,35 @@ 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); } @@ -2132,24 +2382,20 @@ 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 @@ -2161,14 +2407,15 @@ 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; } @@ -2189,7 +2436,7 @@ 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); } } @@ -2198,28 +2445,33 @@ 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 asmlinkage void smp_irq_move_cleanup_interrupt(void) { unsigned vector, me; + ack_APIC_irq(); -#ifdef CONFIG_X86_64 exit_idle(); -#endif irq_enter(); me = smp_processor_id(); @@ -2229,6 +2481,9 @@ 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; @@ -2238,7 +2493,7 @@ 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; @@ -2250,28 +2505,44 @@ 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 has not changed, but affinity did */ + 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) { @@ -2282,11 +2553,14 @@ 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(); } @@ -2295,18 +2569,21 @@ 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 @@ -2330,7 +2607,8 @@ 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 @@ -2369,17 +2647,18 @@ 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 @@ -2430,20 +2709,19 @@ 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) { + 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; - } } } } @@ -2468,7 +2746,7 @@ 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(); } @@ -2480,11 +2758,8 @@ 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"); @@ -2588,7 +2863,9 @@ 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; @@ -2603,7 +2880,7 @@ 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 @@ -2654,10 +2931,10 @@ 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(); @@ -2683,9 +2960,9 @@ 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"); @@ -2717,7 +2994,7 @@ 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); @@ -2902,22 +3179,26 @@ 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; - /* check if need to create one */ - if (!cfg_new) - cfg_new = irq_cfg_alloc(new); - if (__assign_irq_vector(new, TARGET_CPUS) == 0) + } + cfg_new = desc_new->chip_data; + + if (cfg_new->vector != 0) + continue; + if (__assign_irq_vector(new, cfg_new, TARGET_CPUS) == 0) irq = new; break; } @@ -2925,15 +3206,21 @@ 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; @@ -2944,14 +3231,22 @@ 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); } @@ -2964,16 +3259,13 @@ 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)) { @@ -3027,64 +3319,48 @@ 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 = desc->chip_data; - cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); - - 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_cfg *cfg; + struct irq_desc *desc = irq_to_desc(irq); + struct irq_cfg *cfg = desc->chip_data; 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); @@ -3098,16 +3374,10 @@ 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 */ @@ -3166,7 +3436,7 @@ 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; @@ -3175,7 +3445,7 @@ 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 @@ -3195,26 +3465,13 @@ 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; @@ -3228,7 +3485,7 @@ 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; @@ -3246,7 +3503,7 @@ 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 @@ -3254,10 +3511,11 @@ 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 @@ -3289,7 +3547,7 @@ 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++; @@ -3308,24 +3566,18 @@ 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); @@ -3335,9 +3587,8 @@ 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 = { @@ -3369,24 +3620,18 @@ 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); @@ -3396,9 +3641,8 @@ 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 = { @@ -3451,28 +3695,21 @@ 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)) + 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; target_ht_irq(irq, dest, cfg->vector); - desc = irq_to_desc(irq); - desc->affinity = mask; } + #endif static struct irq_chip ht_irq_chip = { @@ -3490,17 +3727,14 @@ 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); @@ -3536,7 +3770,7 @@ 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; @@ -3544,7 +3778,9 @@ 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; @@ -3553,8 +3789,6 @@ 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)); @@ -3565,7 +3799,7 @@ 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); @@ -3606,9 +3840,16 @@ 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) { - return NR_IRQS; + int idx; + int nr = 0; + + for (idx = 0; idx < nr_ioapics; idx++) + nr += io_apic_get_redir_entries(idx) + 1; + + if (nr > nr_irqs_gsi) + nr_irqs_gsi = nr; } /* -------------------------------------------------------------------------- @@ -3707,19 +3948,31 @@ 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; } @@ -3757,7 +4010,7 @@ void __init setup_ioapic_dest(void) int pin, ioapic, irq, irq_entry; struct irq_desc *desc; struct irq_cfg *cfg; - cpumask_t mask; + const struct cpumask *mask; if (skip_ioapic_setup == 1) return; @@ -3773,9 +4026,10 @@ 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); + desc = irq_to_desc(irq); + cfg = desc->chip_data; if (!cfg->vector) { - setup_IO_APIC_irq(ioapic, pin, irq, + setup_IO_APIC_irq(ioapic, pin, irq, desc, irq_trigger(irq_entry), irq_polarity(irq_entry)); continue; @@ -3785,19 +4039,18 @@ void __init setup_ioapic_dest(void) /* * Honour affinities which have been set in early boot */ - desc = irq_to_desc(irq); if (desc->status & (IRQ_NO_BALANCING | IRQ_AFFINITY_SET)) - mask = desc->affinity; + mask = &desc->affinity; else mask = TARGET_CPUS; #ifdef CONFIG_INTR_REMAP if (intr_remapping_enabled) - set_ir_ioapic_affinity_irq(irq, mask); + set_ir_ioapic_affinity_irq_desc(desc, mask); else #endif - set_ioapic_affinity_irq(irq, mask); + set_ioapic_affinity_irq_desc(desc, mask); } } @@ -3846,7 +4099,6 @@ 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..bce53e1352a 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -9,6 +9,7 @@ #include <asm/apic.h> #include <asm/io_apic.h> #include <asm/smp.h> +#include <asm/irq.h> atomic_t irq_err_count; @@ -118,6 +119,9 @@ 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); @@ -187,3 +191,5 @@ u64 arch_irq_stat(void) #endif return sum; } + +EXPORT_SYMBOL_GPL(vector_used_by_percpu_irq); 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..6383d50f82e 100644 --- a/arch/x86/kernel/irq_64.c +++ b/arch/x86/kernel/irq_64.c @@ -13,12 +13,12 @@ #include <linux/seq_file.h> #include <linux/module.h> #include <linux/delay.h> +#include <linux/ftrace.h> #include <asm/uaccess.h> #include <asm/io_apic.h> #include <asm/idle.h> #include <asm/smp.h> -#ifdef CONFIG_DEBUG_STACKOVERFLOW /* * Probabilistic stack overflow check: * @@ -28,26 +28,25 @@ */ static inline void stack_overflow_check(struct pt_regs *regs) { +#ifdef CONFIG_DEBUG_STACKOVERFLOW u64 curbase = (u64)task_stack_page(current); - static unsigned long warned = -60*HZ; - - if (regs->sp >= curbase && regs->sp <= curbase + THREAD_SIZE && - regs->sp < curbase + sizeof(struct thread_info) + 128 && - time_after(jiffies, warned + 60*HZ)) { - printk("do_IRQ: %s near stack overflow (cur:%Lx,sp:%lx)\n", - current->comm, curbase, regs->sp); - show_stack(NULL,NULL); - warned = jiffies; - } -} + + WARN_ONCE(regs->sp >= curbase && + regs->sp <= curbase + THREAD_SIZE && + regs->sp < curbase + sizeof(struct thread_info) + + sizeof(struct pt_regs) + 128, + + "do_IRQ: %s near stack overflow (cur:%Lx,sp:%lx)\n", + current->comm, curbase, regs->sp); #endif +} /* * do_IRQ handles all normal device IRQ's (the special * SMP cross-CPU interrupts have their own specific * handlers). */ -asmlinkage unsigned int do_IRQ(struct pt_regs *regs) +asmlinkage unsigned int __irq_entry do_IRQ(struct pt_regs *regs) { struct pt_regs *old_regs = set_irq_regs(regs); struct irq_desc *desc; @@ -60,9 +59,7 @@ asmlinkage unsigned int do_IRQ(struct pt_regs *regs) irq_enter(); irq = __get_cpu_var(vector_irq)[vector]; -#ifdef CONFIG_DEBUG_STACKOVERFLOW stack_overflow_check(regs); -#endif desc = irq_to_desc(irq); if (likely(desc)) @@ -83,40 +80,43 @@ 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..84723295f88 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; @@ -111,6 +110,18 @@ DEFINE_PER_CPU(vector_irq_t, vector_irq) = { [IRQ15_VECTOR + 1 ... NR_VECTORS - 1] = -1 }; +int vector_used_by_percpu_irq(unsigned int vector) +{ + int cpu; + + for_each_online_cpu(cpu) { + if (per_cpu(vector_irq, cpu)[vector] != -1) + return 1; + } + + return 0; +} + /* Overridden in paravirt.c */ void init_IRQ(void) __attribute__((weak, alias("native_init_IRQ"))); @@ -129,7 +140,7 @@ void __init native_init_IRQ(void) for (i = FIRST_EXTERNAL_VECTOR; i < NR_VECTORS; i++) { /* SYSCALL_VECTOR was reserved in trap_init. */ if (i != SYSCALL_VECTOR) - set_intr_gate(i, interrupt[i]); + set_intr_gate(i, interrupt[i-FIRST_EXTERNAL_VECTOR]); } @@ -147,10 +158,12 @@ void __init native_init_IRQ(void) alloc_intr_gate(CALL_FUNCTION_VECTOR, call_function_interrupt); /* IPI for single call function */ - set_intr_gate(CALL_FUNCTION_SINGLE_VECTOR, call_function_single_interrupt); + alloc_intr_gate(CALL_FUNCTION_SINGLE_VECTOR, + call_function_single_interrupt); /* Low priority IPI to cleanup after moving an irq */ set_intr_gate(IRQ_MOVE_CLEANUP_VECTOR, irq_move_cleanup_interrupt); + set_bit(IRQ_MOVE_CLEANUP_VECTOR, used_vectors); #endif #ifdef CONFIG_X86_LOCAL_APIC diff --git a/arch/x86/kernel/irqinit_64.c b/arch/x86/kernel/irqinit_64.c index ff023539128..31ebfe38e96 100644 --- a/arch/x86/kernel/irqinit_64.c +++ b/arch/x86/kernel/irqinit_64.c @@ -24,41 +24,6 @@ #include <asm/i8259.h> /* - * Common place to define all x86 IRQ vectors - * - * This builds up the IRQ handler stubs using some ugly macros in irq.h - * - * These macros create the low-level assembly IRQ routines that save - * register context and call do_IRQ(). do_IRQ() then does all the - * operations that are needed to keep the AT (or SMP IOAPIC) - * interrupt-controller happy. - */ - -#define IRQ_NAME2(nr) nr##_interrupt(void) -#define IRQ_NAME(nr) IRQ_NAME2(IRQ##nr) - -/* - * SMP has a few special interrupts for IPI messages - */ - -#define BUILD_IRQ(nr) \ - asmlinkage void IRQ_NAME(nr); \ - asm("\n.text\n.p2align\n" \ - "IRQ" #nr "_interrupt:\n\t" \ - "push $~(" #nr ") ; " \ - "jmp common_interrupt\n" \ - ".previous"); - -#define BI(x,y) \ - BUILD_IRQ(x##y) - -#define BUILD_16_IRQS(x) \ - BI(x,0) BI(x,1) BI(x,2) BI(x,3) \ - BI(x,4) BI(x,5) BI(x,6) BI(x,7) \ - BI(x,8) BI(x,9) BI(x,a) BI(x,b) \ - BI(x,c) BI(x,d) BI(x,e) BI(x,f) - -/* * ISA PIC or low IO-APIC triggered (INTA-cycle or APIC) interrupts: * (these are usually mapped to vectors 0x30-0x3f) */ @@ -73,37 +38,6 @@ * * (these are usually mapped into the 0x30-0xff vector range) */ - BUILD_16_IRQS(0x2) BUILD_16_IRQS(0x3) -BUILD_16_IRQS(0x4) BUILD_16_IRQS(0x5) BUILD_16_IRQS(0x6) BUILD_16_IRQS(0x7) -BUILD_16_IRQS(0x8) BUILD_16_IRQS(0x9) BUILD_16_IRQS(0xa) BUILD_16_IRQS(0xb) -BUILD_16_IRQS(0xc) BUILD_16_IRQS(0xd) BUILD_16_IRQS(0xe) BUILD_16_IRQS(0xf) - -#undef BUILD_16_IRQS -#undef BI - - -#define IRQ(x,y) \ - IRQ##x##y##_interrupt - -#define IRQLIST_16(x) \ - IRQ(x,0), IRQ(x,1), IRQ(x,2), IRQ(x,3), \ - IRQ(x,4), IRQ(x,5), IRQ(x,6), IRQ(x,7), \ - IRQ(x,8), IRQ(x,9), IRQ(x,a), IRQ(x,b), \ - IRQ(x,c), IRQ(x,d), IRQ(x,e), IRQ(x,f) - -/* for the irq vectors */ -static void (*__initdata interrupt[NR_VECTORS - FIRST_EXTERNAL_VECTOR])(void) = { - IRQLIST_16(0x2), IRQLIST_16(0x3), - IRQLIST_16(0x4), IRQLIST_16(0x5), IRQLIST_16(0x6), IRQLIST_16(0x7), - IRQLIST_16(0x8), IRQLIST_16(0x9), IRQLIST_16(0xa), IRQLIST_16(0xb), - IRQLIST_16(0xc), IRQLIST_16(0xd), IRQLIST_16(0xe), IRQLIST_16(0xf) -}; - -#undef IRQ -#undef IRQLIST_16 - - - /* * IRQ2 is cascade interrupt to second interrupt controller @@ -135,6 +69,18 @@ DEFINE_PER_CPU(vector_irq_t, vector_irq) = { [IRQ15_VECTOR + 1 ... NR_VECTORS - 1] = -1 }; +int vector_used_by_percpu_irq(unsigned int vector) +{ + int cpu; + + for_each_online_cpu(cpu) { + if (per_cpu(vector_irq, cpu)[vector] != -1) + return 1; + } + + return 0; +} + void __init init_ISA_irqs(void) { int i; @@ -142,8 +88,7 @@ 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; @@ -188,6 +133,7 @@ static void __init smp_intr_init(void) /* Low priority IPI to cleanup after moving an irq */ set_intr_gate(IRQ_MOVE_CLEANUP_VECTOR, irq_move_cleanup_interrupt); + set_bit(IRQ_MOVE_CLEANUP_VECTOR, used_vectors); #endif } diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c index e169ae9b6a6..652fce6d2cc 100644 --- a/arch/x86/kernel/kvmclock.c +++ b/arch/x86/kernel/kvmclock.c @@ -89,17 +89,17 @@ static cycle_t kvm_clock_read(void) */ static unsigned long kvm_get_tsc_khz(void) { - return preset_lpj; + struct pvclock_vcpu_time_info *src; + src = &per_cpu(hv_clock, 0); + return pvclock_tsc_khz(src); } static void kvm_get_preset_lpj(void) { - struct pvclock_vcpu_time_info *src; unsigned long khz; u64 lpj; - src = &per_cpu(hv_clock, 0); - khz = pvclock_tsc_khz(src); + khz = kvm_get_tsc_khz(); lpj = ((u64)khz * 1000); do_div(lpj, HZ); @@ -194,5 +194,7 @@ void __init kvmclock_init(void) #endif kvm_get_preset_lpj(); clocksource_register(&kvm_clock); + pv_info.paravirt_enabled = 1; + pv_info.name = "KVM"; } } diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c index eee32b43fee..71f1d99a635 100644 --- a/arch/x86/kernel/ldt.c +++ b/arch/x86/kernel/ldt.c @@ -12,8 +12,8 @@ #include <linux/mm.h> #include <linux/smp.h> #include <linux/vmalloc.h> +#include <linux/uaccess.h> -#include <asm/uaccess.h> #include <asm/system.h> #include <asm/ldt.h> #include <asm/desc.h> @@ -93,7 +93,7 @@ static inline int copy_ldt(mm_context_t *new, mm_context_t *old) if (err < 0) return err; - for(i = 0; i < old->size; i++) + for (i = 0; i < old->size; i++) write_ldt_entry(new->ldt, i, old->ldt + i * LDT_ENTRY_SIZE); return 0; } 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/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index 5f8e5d75a25..c25fdb38229 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -10,7 +10,7 @@ * This driver allows to upgrade microcode on AMD * family 0x10 and 0x11 processors. * - * Licensed unter the terms of the GNU General Public + * Licensed under the terms of the GNU General Public * License version 2. See file COPYING for details. */ @@ -32,9 +32,9 @@ #include <linux/platform_device.h> #include <linux/pci.h> #include <linux/pci_ids.h> +#include <linux/uaccess.h> #include <asm/msr.h> -#include <asm/uaccess.h> #include <asm/processor.h> #include <asm/microcode.h> @@ -47,43 +47,38 @@ MODULE_LICENSE("GPL v2"); #define UCODE_UCODE_TYPE 0x00000001 struct equiv_cpu_entry { - unsigned int installed_cpu; - unsigned int fixed_errata_mask; - unsigned int fixed_errata_compare; - unsigned int equiv_cpu; -}; + u32 installed_cpu; + u32 fixed_errata_mask; + u32 fixed_errata_compare; + u16 equiv_cpu; + u16 res; +} __attribute__((packed)); struct microcode_header_amd { - unsigned int data_code; - unsigned int patch_id; - unsigned char mc_patch_data_id[2]; - unsigned char mc_patch_data_len; - unsigned char init_flag; - unsigned int mc_patch_data_checksum; - unsigned int nb_dev_id; - unsigned int sb_dev_id; - unsigned char processor_rev_id[2]; - unsigned char nb_rev_id; - unsigned char sb_rev_id; - unsigned char bios_api_rev; - unsigned char reserved1[3]; - unsigned int match_reg[8]; -}; + u32 data_code; + u32 patch_id; + u16 mc_patch_data_id; + u8 mc_patch_data_len; + u8 init_flag; + u32 mc_patch_data_checksum; + u32 nb_dev_id; + u32 sb_dev_id; + u16 processor_rev_id; + u8 nb_rev_id; + u8 sb_rev_id; + u8 bios_api_rev; + u8 reserved1[3]; + u32 match_reg[8]; +} __attribute__((packed)); struct microcode_amd { struct microcode_header_amd hdr; unsigned int mpb[0]; }; -#define UCODE_MAX_SIZE (2048) -#define DEFAULT_UCODE_DATASIZE (896) -#define MC_HEADER_SIZE (sizeof(struct microcode_header_amd)) -#define DEFAULT_UCODE_TOTALSIZE (DEFAULT_UCODE_DATASIZE + MC_HEADER_SIZE) -#define DWSIZE (sizeof(u32)) -/* For now we support a fixed ucode total size only */ -#define get_totalsize(mc) \ - ((((struct microcode_amd *)mc)->hdr.mc_patch_data_len * 28) \ - + MC_HEADER_SIZE) +#define UCODE_MAX_SIZE 2048 +#define UCODE_CONTAINER_SECTION_HDR 8 +#define UCODE_CONTAINER_HEADER_SIZE 12 /* serialize access to the physical write */ static DEFINE_SPINLOCK(microcode_update_lock); @@ -93,31 +88,24 @@ static struct equiv_cpu_entry *equiv_cpu_table; static int collect_cpu_info_amd(int cpu, struct cpu_signature *csig) { struct cpuinfo_x86 *c = &cpu_data(cpu); + u32 dummy; memset(csig, 0, sizeof(*csig)); - if (c->x86_vendor != X86_VENDOR_AMD || c->x86 < 0x10) { - printk(KERN_ERR "microcode: CPU%d not a capable AMD processor\n", - cpu); + printk(KERN_WARNING "microcode: CPU%d: AMD CPU family 0x%x not " + "supported\n", cpu, c->x86); return -1; } - - asm volatile("movl %1, %%ecx; rdmsr" - : "=a" (csig->rev) - : "i" (0x0000008B) : "ecx"); - - printk(KERN_INFO "microcode: collect_cpu_info_amd : patch_id=0x%x\n", - csig->rev); - + rdmsr(MSR_AMD64_PATCH_LEVEL, csig->rev, dummy); + printk(KERN_INFO "microcode: CPU%d: patch_level=0x%x\n", cpu, csig->rev); return 0; } static int get_matching_microcode(int cpu, void *mc, int rev) { struct microcode_header_amd *mc_header = mc; - struct pci_dev *nb_pci_dev, *sb_pci_dev; unsigned int current_cpu_id; - unsigned int equiv_cpu_id = 0x00; + u16 equiv_cpu_id = 0; unsigned int i = 0; BUG_ON(equiv_cpu_table == NULL); @@ -132,57 +120,25 @@ static int get_matching_microcode(int cpu, void *mc, int rev) } if (!equiv_cpu_id) { - printk(KERN_ERR "microcode: CPU%d cpu_id " - "not found in equivalent cpu table \n", cpu); + printk(KERN_WARNING "microcode: CPU%d: cpu revision " + "not listed in equivalent cpu table\n", cpu); return 0; } - if ((mc_header->processor_rev_id[0]) != (equiv_cpu_id & 0xff)) { - printk(KERN_ERR - "microcode: CPU%d patch does not match " - "(patch is %x, cpu extended is %x) \n", - cpu, mc_header->processor_rev_id[0], - (equiv_cpu_id & 0xff)); + if (mc_header->processor_rev_id != equiv_cpu_id) { + printk(KERN_ERR "microcode: CPU%d: patch mismatch " + "(processor_rev_id: %x, equiv_cpu_id: %x)\n", + cpu, mc_header->processor_rev_id, equiv_cpu_id); return 0; } - if ((mc_header->processor_rev_id[1]) != ((equiv_cpu_id >> 16) & 0xff)) { - printk(KERN_ERR "microcode: CPU%d patch does not match " - "(patch is %x, cpu base id is %x) \n", - cpu, mc_header->processor_rev_id[1], - ((equiv_cpu_id >> 16) & 0xff)); - + /* ucode might be chipset specific -- currently we don't support this */ + if (mc_header->nb_dev_id || mc_header->sb_dev_id) { + printk(KERN_ERR "microcode: CPU%d: loading of chipset " + "specific code not yet supported\n", cpu); return 0; } - /* ucode may be northbridge specific */ - if (mc_header->nb_dev_id) { - nb_pci_dev = pci_get_device(PCI_VENDOR_ID_AMD, - (mc_header->nb_dev_id & 0xff), - NULL); - if ((!nb_pci_dev) || - (mc_header->nb_rev_id != nb_pci_dev->revision)) { - printk(KERN_ERR "microcode: CPU%d NB mismatch \n", cpu); - pci_dev_put(nb_pci_dev); - return 0; - } - pci_dev_put(nb_pci_dev); - } - - /* ucode may be southbridge specific */ - if (mc_header->sb_dev_id) { - sb_pci_dev = pci_get_device(PCI_VENDOR_ID_AMD, - (mc_header->sb_dev_id & 0xff), - NULL); - if ((!sb_pci_dev) || - (mc_header->sb_rev_id != sb_pci_dev->revision)) { - printk(KERN_ERR "microcode: CPU%d SB mismatch \n", cpu); - pci_dev_put(sb_pci_dev); - return 0; - } - pci_dev_put(sb_pci_dev); - } - if (mc_header->patch_id <= rev) return 0; @@ -192,12 +148,10 @@ static int get_matching_microcode(int cpu, void *mc, int rev) static void apply_microcode_amd(int cpu) { unsigned long flags; - unsigned int eax, edx; - unsigned int rev; + u32 rev, dummy; int cpu_num = raw_smp_processor_id(); struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; struct microcode_amd *mc_amd = uci->mc; - unsigned long addr; /* We should bind the task to the CPU */ BUG_ON(cpu_num != cpu); @@ -206,42 +160,34 @@ static void apply_microcode_amd(int cpu) return; spin_lock_irqsave(µcode_update_lock, flags); - - addr = (unsigned long)&mc_amd->hdr.data_code; - edx = (unsigned int)(((unsigned long)upper_32_bits(addr))); - eax = (unsigned int)(((unsigned long)lower_32_bits(addr))); - - asm volatile("movl %0, %%ecx; wrmsr" : - : "i" (0xc0010020), "a" (eax), "d" (edx) : "ecx"); - + wrmsrl(MSR_AMD64_PATCH_LOADER, (u64)(long)&mc_amd->hdr.data_code); /* get patch id after patching */ - asm volatile("movl %1, %%ecx; rdmsr" - : "=a" (rev) - : "i" (0x0000008B) : "ecx"); - + rdmsr(MSR_AMD64_PATCH_LEVEL, rev, dummy); spin_unlock_irqrestore(µcode_update_lock, flags); /* check current patch id and patch's id for match */ if (rev != mc_amd->hdr.patch_id) { - printk(KERN_ERR "microcode: CPU%d update from revision " - "0x%x to 0x%x failed\n", cpu_num, - mc_amd->hdr.patch_id, rev); + printk(KERN_ERR "microcode: CPU%d: update failed " + "(for patch_level=0x%x)\n", cpu, mc_amd->hdr.patch_id); return; } - printk(KERN_INFO "microcode: CPU%d updated from revision " - "0x%x to 0x%x \n", - cpu_num, uci->cpu_sig.rev, mc_amd->hdr.patch_id); + printk(KERN_INFO "microcode: CPU%d: updated (new patch_level=0x%x)\n", + cpu, rev); uci->cpu_sig.rev = rev; } -static void * get_next_ucode(u8 *buf, unsigned int size, - int (*get_ucode_data)(void *, const void *, size_t), - unsigned int *mc_size) +static int get_ucode_data(void *to, const u8 *from, size_t n) +{ + memcpy(to, from, n); + return 0; +} + +static void *get_next_ucode(const u8 *buf, unsigned int size, + unsigned int *mc_size) { unsigned int total_size; -#define UCODE_CONTAINER_SECTION_HDR 8 u8 section_hdr[UCODE_CONTAINER_SECTION_HDR]; void *mc; @@ -249,39 +195,37 @@ static void * get_next_ucode(u8 *buf, unsigned int size, return NULL; if (section_hdr[0] != UCODE_UCODE_TYPE) { - printk(KERN_ERR "microcode: error! " - "Wrong microcode payload type field\n"); + printk(KERN_ERR "microcode: error: invalid type field in " + "container file section header\n"); return NULL; } total_size = (unsigned long) (section_hdr[4] + (section_hdr[5] << 8)); - printk(KERN_INFO "microcode: size %u, total_size %u\n", - size, total_size); + printk(KERN_DEBUG "microcode: size %u, total_size %u\n", + size, total_size); if (total_size > size || total_size > UCODE_MAX_SIZE) { - printk(KERN_ERR "microcode: error! Bad data in microcode data file\n"); + printk(KERN_ERR "microcode: error: size mismatch\n"); return NULL; } mc = vmalloc(UCODE_MAX_SIZE); if (mc) { memset(mc, 0, UCODE_MAX_SIZE); - if (get_ucode_data(mc, buf + UCODE_CONTAINER_SECTION_HDR, total_size)) { + if (get_ucode_data(mc, buf + UCODE_CONTAINER_SECTION_HDR, + total_size)) { vfree(mc); mc = NULL; } else *mc_size = total_size + UCODE_CONTAINER_SECTION_HDR; } -#undef UCODE_CONTAINER_SECTION_HDR return mc; } -static int install_equiv_cpu_table(u8 *buf, - int (*get_ucode_data)(void *, const void *, size_t)) +static int install_equiv_cpu_table(const u8 *buf) { -#define UCODE_CONTAINER_HEADER_SIZE 12 u8 *container_hdr[UCODE_CONTAINER_HEADER_SIZE]; unsigned int *buf_pos = (unsigned int *)container_hdr; unsigned long size; @@ -292,14 +236,15 @@ static int install_equiv_cpu_table(u8 *buf, size = buf_pos[2]; if (buf_pos[1] != UCODE_EQUIV_CPU_TABLE_TYPE || !size) { - printk(KERN_ERR "microcode: error! " - "Wrong microcode equivalnet cpu table\n"); + printk(KERN_ERR "microcode: error: invalid type field in " + "container file section header\n"); return 0; } equiv_cpu_table = (struct equiv_cpu_entry *) vmalloc(size); if (!equiv_cpu_table) { - printk(KERN_ERR "microcode: error, can't allocate memory for equiv CPU table\n"); + printk(KERN_ERR "microcode: failed to allocate " + "equivalent CPU table\n"); return 0; } @@ -310,7 +255,6 @@ static int install_equiv_cpu_table(u8 *buf, } return size + UCODE_CONTAINER_HEADER_SIZE; /* add header length */ -#undef UCODE_CONTAINER_HEADER_SIZE } static void free_equiv_cpu_table(void) @@ -321,18 +265,20 @@ static void free_equiv_cpu_table(void) } } -static int generic_load_microcode(int cpu, void *data, size_t size, - int (*get_ucode_data)(void *, const void *, size_t)) +static int generic_load_microcode(int cpu, const u8 *data, size_t size) { struct ucode_cpu_info *uci = ucode_cpu_info + cpu; - u8 *ucode_ptr = data, *new_mc = NULL, *mc; + const u8 *ucode_ptr = data; + void *new_mc = NULL; + void *mc; int new_rev = uci->cpu_sig.rev; unsigned int leftover; unsigned long offset; - offset = install_equiv_cpu_table(ucode_ptr, get_ucode_data); + offset = install_equiv_cpu_table(ucode_ptr); if (!offset) { - printk(KERN_ERR "microcode: installing equivalent cpu table failed\n"); + printk(KERN_ERR "microcode: failed to create " + "equivalent cpu table\n"); return -EINVAL; } @@ -343,7 +289,7 @@ static int generic_load_microcode(int cpu, void *data, size_t size, unsigned int uninitialized_var(mc_size); struct microcode_header_amd *mc_header; - mc = get_next_ucode(ucode_ptr, leftover, get_ucode_data, &mc_size); + mc = get_next_ucode(ucode_ptr, leftover, &mc_size); if (!mc) break; @@ -353,7 +299,7 @@ static int generic_load_microcode(int cpu, void *data, size_t size, vfree(new_mc); new_rev = mc_header->patch_id; new_mc = mc; - } else + } else vfree(mc); ucode_ptr += mc_size; @@ -365,9 +311,9 @@ static int generic_load_microcode(int cpu, void *data, size_t size, if (uci->mc) vfree(uci->mc); uci->mc = new_mc; - pr_debug("microcode: CPU%d found a matching microcode update with" - " version 0x%x (current=0x%x)\n", - cpu, new_rev, uci->cpu_sig.rev); + pr_debug("microcode: CPU%d found a matching microcode " + "update with version 0x%x (current=0x%x)\n", + cpu, new_rev, uci->cpu_sig.rev); } else vfree(new_mc); } @@ -377,12 +323,6 @@ static int generic_load_microcode(int cpu, void *data, size_t size, return (int)leftover; } -static int get_ucode_fw(void *to, const void *from, size_t n) -{ - memcpy(to, from, n); - return 0; -} - static int request_microcode_fw(int cpu, struct device *device) { const char *fw_name = "amd-ucode/microcode_amd.bin"; @@ -394,12 +334,11 @@ static int request_microcode_fw(int cpu, struct device *device) ret = request_firmware(&firmware, fw_name, device); if (ret) { - printk(KERN_ERR "microcode: ucode data file %s load failed\n", fw_name); + printk(KERN_ERR "microcode: failed to load file %s\n", fw_name); return ret; } - ret = generic_load_microcode(cpu, (void*)firmware->data, firmware->size, - &get_ucode_fw); + ret = generic_load_microcode(cpu, firmware->data, firmware->size); release_firmware(firmware); @@ -408,8 +347,8 @@ static int request_microcode_fw(int cpu, struct device *device) static int request_microcode_user(int cpu, const void __user *buf, size_t size) { - printk(KERN_WARNING "microcode: AMD microcode update via /dev/cpu/microcode" - "is not supported\n"); + printk(KERN_INFO "microcode: AMD microcode update via " + "/dev/cpu/microcode not supported\n"); return -1; } @@ -433,3 +372,4 @@ struct microcode_ops * __init init_amd_microcode(void) { return µcode_amd_ops; } + diff --git a/arch/x86/kernel/microcode_core.c b/arch/x86/kernel/microcode_core.c index 82fb2809ce3..c9b721ba968 100644 --- a/arch/x86/kernel/microcode_core.c +++ b/arch/x86/kernel/microcode_core.c @@ -99,7 +99,7 @@ MODULE_LICENSE("GPL"); #define MICROCODE_VERSION "2.00" -struct microcode_ops *microcode_ops; +static struct microcode_ops *microcode_ops; /* no concurrent ->write()s are allowed on /dev/cpu/microcode */ static DEFINE_MUTEX(microcode_mutex); @@ -203,7 +203,7 @@ MODULE_ALIAS_MISCDEV(MICROCODE_MINOR); #endif /* fake device for request_firmware */ -struct platform_device *microcode_pdev; +static struct platform_device *microcode_pdev; static ssize_t reload_store(struct sys_device *dev, struct sysdev_attribute *attr, @@ -272,13 +272,18 @@ static struct attribute_group mc_attr_group = { .name = "microcode", }; -static void microcode_fini_cpu(int cpu) +static void __microcode_fini_cpu(int cpu) { struct ucode_cpu_info *uci = ucode_cpu_info + cpu; - mutex_lock(µcode_mutex); microcode_ops->microcode_fini_cpu(cpu); uci->valid = 0; +} + +static void microcode_fini_cpu(int cpu) +{ + mutex_lock(µcode_mutex); + __microcode_fini_cpu(cpu); mutex_unlock(µcode_mutex); } @@ -306,12 +311,16 @@ static int microcode_resume_cpu(int cpu) * to this cpu (a bit of paranoia): */ if (microcode_ops->collect_cpu_info(cpu, &nsig)) { - microcode_fini_cpu(cpu); + __microcode_fini_cpu(cpu); + printk(KERN_ERR "failed to collect_cpu_info for resuming cpu #%d\n", + cpu); return -1; } - if (memcmp(&nsig, &uci->cpu_sig, sizeof(nsig))) { - microcode_fini_cpu(cpu); + if ((nsig.sig != uci->cpu_sig.sig) || (nsig.pf != uci->cpu_sig.pf)) { + __microcode_fini_cpu(cpu); + printk(KERN_ERR "cached ucode doesn't match the resuming cpu #%d\n", + cpu); /* Should we look for a new ucode here? */ return 1; } @@ -319,7 +328,7 @@ static int microcode_resume_cpu(int cpu) return 0; } -void microcode_update_cpu(int cpu) +static void microcode_update_cpu(int cpu) { struct ucode_cpu_info *uci = ucode_cpu_info + cpu; int err = 0; diff --git a/arch/x86/kernel/microcode_intel.c b/arch/x86/kernel/microcode_intel.c index 622dc4a2178..b7f4c929e61 100644 --- a/arch/x86/kernel/microcode_intel.c +++ b/arch/x86/kernel/microcode_intel.c @@ -155,6 +155,7 @@ static DEFINE_SPINLOCK(microcode_update_lock); static int collect_cpu_info(int cpu_num, struct cpu_signature *csig) { struct cpuinfo_x86 *c = &cpu_data(cpu_num); + unsigned long flags; unsigned int val[2]; memset(csig, 0, sizeof(*csig)); @@ -174,11 +175,16 @@ static int collect_cpu_info(int cpu_num, struct cpu_signature *csig) csig->pf = 1 << ((val[1] >> 18) & 7); } + /* serialize access to the physical write to MSR 0x79 */ + spin_lock_irqsave(µcode_update_lock, flags); + wrmsr(MSR_IA32_UCODE_REV, 0, 0); /* see notes above for revision 1.07. Apparent chip bug */ sync_core(); /* get the current revision from MSR 0x8B */ rdmsr(MSR_IA32_UCODE_REV, val[0], csig->rev); + spin_unlock_irqrestore(µcode_update_lock, flags); + pr_debug("microcode: collect_cpu_info : sig=0x%x, pf=0x%x, rev=0x%x\n", csig->sig, csig->pf, csig->rev); @@ -465,7 +471,7 @@ static void microcode_fini_cpu(int cpu) uci->mc = NULL; } -struct microcode_ops microcode_intel_ops = { +static struct microcode_ops microcode_intel_ops = { .request_microcode_user = request_microcode_user, .request_microcode_fw = request_microcode_fw, .collect_cpu_info = collect_cpu_info, diff --git a/arch/x86/kernel/mmconf-fam10h_64.c b/arch/x86/kernel/mmconf-fam10h_64.c index efc2f361fe8..666e43df51f 100644 --- a/arch/x86/kernel/mmconf-fam10h_64.c +++ b/arch/x86/kernel/mmconf-fam10h_64.c @@ -13,8 +13,7 @@ #include <asm/msr.h> #include <asm/acpi.h> #include <asm/mmconfig.h> - -#include "../pci/pci.h" +#include <asm/pci_x86.h> struct pci_hostbridge_probe { u32 bus; diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c index 0f4c1fd5a1f..c5c5b8df1db 100644 --- a/arch/x86/kernel/mpparse.c +++ b/arch/x86/kernel/mpparse.c @@ -16,14 +16,14 @@ #include <linux/bitops.h> #include <linux/acpi.h> #include <linux/module.h> +#include <linux/smp.h> +#include <linux/acpi.h> -#include <asm/smp.h> #include <asm/mtrr.h> #include <asm/mpspec.h> #include <asm/pgalloc.h> #include <asm/io_apic.h> #include <asm/proto.h> -#include <asm/acpi.h> #include <asm/bios_ebda.h> #include <asm/e820.h> #include <asm/trampoline.h> @@ -95,8 +95,8 @@ static void __init MP_bus_info(struct mpc_config_bus *m) #endif if (strncmp(str, BUSTYPE_ISA, sizeof(BUSTYPE_ISA) - 1) == 0) { - set_bit(m->mpc_busid, mp_bus_not_pci); -#if defined(CONFIG_EISA) || defined (CONFIG_MCA) + set_bit(m->mpc_busid, mp_bus_not_pci); +#if defined(CONFIG_EISA) || defined(CONFIG_MCA) mp_bus_id_to_type[m->mpc_busid] = MP_BUS_ISA; #endif } else if (strncmp(str, BUSTYPE_PCI, sizeof(BUSTYPE_PCI) - 1) == 0) { @@ -104,7 +104,7 @@ static void __init MP_bus_info(struct mpc_config_bus *m) x86_quirks->mpc_oem_pci_bus(m); clear_bit(m->mpc_busid, mp_bus_not_pci); -#if defined(CONFIG_EISA) || defined (CONFIG_MCA) +#if defined(CONFIG_EISA) || defined(CONFIG_MCA) mp_bus_id_to_type[m->mpc_busid] = MP_BUS_PCI; } else if (strncmp(str, BUSTYPE_EISA, sizeof(BUSTYPE_EISA) - 1) == 0) { mp_bus_id_to_type[m->mpc_busid] = MP_BUS_EISA; @@ -586,26 +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 (!mpf) - return; + 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/msr.c b/arch/x86/kernel/msr.c index 82a7c7ed6d4..726266695b2 100644 --- a/arch/x86/kernel/msr.c +++ b/arch/x86/kernel/msr.c @@ -136,7 +136,7 @@ static int msr_open(struct inode *inode, struct file *file) lock_kernel(); cpu = iminor(file->f_path.dentry->d_inode); - if (cpu >= NR_CPUS || !cpu_online(cpu)) { + if (cpu >= nr_cpu_ids || !cpu_online(cpu)) { ret = -ENXIO; /* No such CPU */ goto out; } diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c index 2c97f07f1c2..45a09ccdc21 100644 --- a/arch/x86/kernel/nmi.c +++ b/arch/x86/kernel/nmi.c @@ -26,11 +26,10 @@ #include <linux/kernel_stat.h> #include <linux/kdebug.h> #include <linux/smp.h> +#include <linux/nmi.h> #include <asm/i8259.h> #include <asm/io_apic.h> -#include <asm/smp.h> -#include <asm/nmi.h> #include <asm/proto.h> #include <asm/timer.h> @@ -131,6 +130,11 @@ static void report_broken_nmi(int cpu, int *prev_nmi_count) atomic_dec(&nmi_active); } +static void __acpi_nmi_disable(void *__unused) +{ + apic_write(APIC_LVT0, APIC_DM_NMI | APIC_LVT_MASKED); +} + int __init check_nmi_watchdog(void) { unsigned int *prev_nmi_count; @@ -179,8 +183,12 @@ int __init check_nmi_watchdog(void) kfree(prev_nmi_count); return 0; error: - if (nmi_watchdog == NMI_IO_APIC && !timer_through_8259) - disable_8259A_irq(0); + if (nmi_watchdog == NMI_IO_APIC) { + if (!timer_through_8259) + disable_8259A_irq(0); + on_each_cpu(__acpi_nmi_disable, NULL, 1); + } + #ifdef CONFIG_X86_32 timer_ack = 0; #endif @@ -199,12 +207,17 @@ static int __init setup_nmi_watchdog(char *str) ++str; } - get_option(&str, &nmi); - - if (nmi >= NMI_INVALID) - return 0; + if (!strncmp(str, "lapic", 5)) + nmi_watchdog = NMI_LOCAL_APIC; + else if (!strncmp(str, "ioapic", 6)) + nmi_watchdog = NMI_IO_APIC; + else { + get_option(&str, &nmi); + if (nmi >= NMI_INVALID) + return 0; + nmi_watchdog = nmi; + } - nmi_watchdog = nmi; return 1; } __setup("nmi_watchdog=", setup_nmi_watchdog); @@ -285,11 +298,6 @@ void acpi_nmi_enable(void) on_each_cpu(__acpi_nmi_enable, NULL, 1); } -static void __acpi_nmi_disable(void *__unused) -{ - apic_write(APIC_LVT0, APIC_DM_NMI | APIC_LVT_MASKED); -} - /* * Disable timer based NMIs on all CPUs: */ @@ -340,6 +348,8 @@ void stop_apic_nmi_watchdog(void *unused) return; if (nmi_watchdog == NMI_LOCAL_APIC) lapic_watchdog_stop(); + else + __acpi_nmi_disable(NULL); __get_cpu_var(wd_enabled) = 0; atomic_dec(&nmi_active); } @@ -465,6 +475,24 @@ nmi_watchdog_tick(struct pt_regs *regs, unsigned reason) #ifdef CONFIG_SYSCTL +static void enable_ioapic_nmi_watchdog_single(void *unused) +{ + __get_cpu_var(wd_enabled) = 1; + atomic_inc(&nmi_active); + __acpi_nmi_enable(NULL); +} + +static void enable_ioapic_nmi_watchdog(void) +{ + on_each_cpu(enable_ioapic_nmi_watchdog_single, NULL, 1); + touch_nmi_watchdog(); +} + +static void disable_ioapic_nmi_watchdog(void) +{ + on_each_cpu(stop_apic_nmi_watchdog, NULL, 1); +} + static int __init setup_unknown_nmi_panic(char *str) { unknown_nmi_panic = 1; @@ -507,6 +535,11 @@ int proc_nmi_enabled(struct ctl_table *table, int write, struct file *file, enable_lapic_nmi_watchdog(); else disable_lapic_nmi_watchdog(); + } else if (nmi_watchdog == NMI_IO_APIC) { + if (nmi_watchdog_enabled) + enable_ioapic_nmi_watchdog(); + else + disable_ioapic_nmi_watchdog(); } else { printk(KERN_WARNING "NMI watchdog doesn't know what hardware to touch\n"); 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/pci-dma.c b/arch/x86/kernel/pci-dma.c index 19262482021..19a1044a0cd 100644 --- a/arch/x86/kernel/pci-dma.c +++ b/arch/x86/kernel/pci-dma.c @@ -6,6 +6,7 @@ #include <asm/proto.h> #include <asm/dma.h> #include <asm/iommu.h> +#include <asm/gart.h> #include <asm/calgary.h> #include <asm/amd_iommu.h> @@ -30,11 +31,6 @@ int no_iommu __read_mostly; /* Set this to 1 if there is a HW IOMMU in the system */ int iommu_detected __read_mostly = 0; -/* This tells the BIO block layer to assume merging. Default to off - because we cannot guarantee merging later. */ -int iommu_bio_merge __read_mostly = 0; -EXPORT_SYMBOL(iommu_bio_merge); - dma_addr_t bad_dma_address __read_mostly = 0; EXPORT_SYMBOL(bad_dma_address); @@ -105,11 +101,15 @@ static void __init dma32_free_bootmem(void) dma32_bootmem_ptr = NULL; dma32_bootmem_size = 0; } +#endif void __init pci_iommu_alloc(void) { +#ifdef CONFIG_X86_64 /* free the range so iommu could get some range less than 4G */ dma32_free_bootmem(); +#endif + /* * The order of these functions is important for * fall-back/fail-over reasons @@ -125,15 +125,6 @@ void __init pci_iommu_alloc(void) pci_swiotlb_init(); } -unsigned long iommu_nr_pages(unsigned long addr, unsigned long len) -{ - unsigned long size = roundup((addr & ~PAGE_MASK) + len, PAGE_SIZE); - - return size >> PAGE_SHIFT; -} -EXPORT_SYMBOL(iommu_nr_pages); -#endif - void *dma_generic_alloc_coherent(struct device *dev, size_t size, dma_addr_t *dma_addr, gfp_t flag) { @@ -188,7 +179,6 @@ static __init int iommu_setup(char *p) } if (!strncmp(p, "biomerge", 8)) { - iommu_bio_merge = 4096; iommu_merge = 1; force_iommu = 1; } @@ -300,8 +290,8 @@ fs_initcall(pci_iommu_init); static __devinit void via_no_dac(struct pci_dev *dev) { if ((dev->class >> 8) == PCI_CLASS_BRIDGE_PCI && forbid_dac == 0) { - printk(KERN_INFO "PCI: VIA PCI bridge detected." - "Disabling DAC.\n"); + printk(KERN_INFO + "PCI: VIA PCI bridge detected. Disabling DAC.\n"); forbid_dac = 1; } } diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c index ba7ad83e20a..00c2bcd4146 100644 --- a/arch/x86/kernel/pci-gart_64.c +++ b/arch/x86/kernel/pci-gart_64.c @@ -52,7 +52,7 @@ static u32 *iommu_gatt_base; /* Remapping table */ * to trigger bugs with some popular PCI cards, in particular 3ware (but * has been also also seen with Qlogic at least). */ -int iommu_fullflush = 1; +static int iommu_fullflush = 1; /* Allocation bitmap for the remapping area: */ static DEFINE_SPINLOCK(iommu_bitmap_lock); @@ -745,10 +745,8 @@ void __init gart_iommu_init(void) unsigned long scratch; long i; - if (cache_k8_northbridges() < 0 || num_k8_northbridges == 0) { - printk(KERN_INFO "PCI-GART: No AMD GART found.\n"); + if (cache_k8_northbridges() < 0 || num_k8_northbridges == 0) return; - } #ifndef CONFIG_AGP_AMD64 no_agp = 1; diff --git a/arch/x86/kernel/pci-swiotlb_64.c b/arch/x86/kernel/pci-swiotlb_64.c index 3c539d111ab..242c3440687 100644 --- a/arch/x86/kernel/pci-swiotlb_64.c +++ b/arch/x86/kernel/pci-swiotlb_64.c @@ -3,6 +3,8 @@ #include <linux/pci.h> #include <linux/cache.h> #include <linux/module.h> +#include <linux/swiotlb.h> +#include <linux/bootmem.h> #include <linux/dma-mapping.h> #include <asm/iommu.h> @@ -11,6 +13,31 @@ int swiotlb __read_mostly; +void *swiotlb_alloc_boot(size_t size, unsigned long nslabs) +{ + return alloc_bootmem_low_pages(size); +} + +void *swiotlb_alloc(unsigned order, unsigned long nslabs) +{ + return (void *)__get_free_pages(GFP_DMA | __GFP_NOWARN, order); +} + +dma_addr_t swiotlb_phys_to_bus(phys_addr_t paddr) +{ + return paddr; +} + +phys_addr_t swiotlb_bus_to_phys(dma_addr_t baddr) +{ + return baddr; +} + +int __weak swiotlb_arch_range_needs_mapping(void *ptr, size_t size) +{ + return 0; +} + static dma_addr_t swiotlb_map_single_phys(struct device *hwdev, phys_addr_t paddr, size_t size, int direction) @@ -50,8 +77,10 @@ struct dma_mapping_ops swiotlb_dma_ops = { void __init pci_swiotlb_init(void) { /* don't initialize swiotlb if iommu=off (no_iommu=1) */ +#ifdef CONFIG_X86_64 if (!iommu_detected && !no_iommu && max_pfn > MAX_DMA32_PFN) swiotlb = 1; +#endif if (swiotlb_force) swiotlb = 1; if (swiotlb) { diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index c622772744d..e68bb9e3086 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -1,13 +1,16 @@ #include <linux/errno.h> #include <linux/kernel.h> #include <linux/mm.h> +#include <asm/idle.h> #include <linux/smp.h> #include <linux/slab.h> #include <linux/sched.h> #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 +103,9 @@ 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 +118,7 @@ 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 +129,21 @@ 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 +176,31 @@ 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 +212,13 @@ 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); } /* @@ -270,7 +303,7 @@ static void c1e_idle(void) rdmsr(MSR_K8_INT_PENDING_MSG, lo, hi); if (lo & K8_INTP_C1E_ACTIVE_MASK) { c1e_detected = 1; - if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC)) + if (!boot_cpu_has(X86_FEATURE_NONSTOP_TSC)) mark_tsc_unstable("TSC halt in AMD C1E"); printk(KERN_INFO "System has AMD C1E enabled\n"); set_cpu_cap(&boot_cpu_data, X86_FEATURE_AMDC1E); diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 0a1302fe6d4..3ba155d2488 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> @@ -59,6 +60,7 @@ #include <asm/idle.h> #include <asm/syscalls.h> #include <asm/smp.h> +#include <asm/ds.h> asmlinkage void ret_from_fork(void) __asm__("ret_from_fork"); @@ -250,14 +252,8 @@ void exit_thread(void) tss->x86_tss.io_bitmap_base = INVALID_IO_BITMAP_OFFSET; put_cpu(); } -#ifdef CONFIG_X86_DS - /* Free any DS contexts that have not been properly released. */ - if (unlikely(current->thread.ds_ctx)) { - /* we clear debugctl to make sure DS is not used. */ - update_debugctlmsr(0); - ds_free(current->thread.ds_ctx); - } -#endif /* CONFIG_X86_DS */ + + ds_exit_thread(current); } void flush_thread(void) @@ -339,6 +335,12 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long sp, kfree(p->thread.io_bitmap_ptr); p->thread.io_bitmap_max = 0; } + + ds_copy_thread(p, current); + + clear_tsk_thread_flag(p, TIF_DEBUGCTLMSR); + p->thread.debugctlmsr = 0; + return err; } @@ -419,48 +421,19 @@ int set_tsc_mode(unsigned int val) return 0; } -#ifdef CONFIG_X86_DS -static int update_debugctl(struct thread_struct *prev, - struct thread_struct *next, unsigned long debugctl) -{ - unsigned long ds_prev = 0; - unsigned long ds_next = 0; - - if (prev->ds_ctx) - ds_prev = (unsigned long)prev->ds_ctx->ds; - if (next->ds_ctx) - ds_next = (unsigned long)next->ds_ctx->ds; - - if (ds_next != ds_prev) { - /* we clear debugctl to make sure DS - * is not in use when we change it */ - debugctl = 0; - update_debugctlmsr(0); - wrmsr(MSR_IA32_DS_AREA, ds_next, 0); - } - return debugctl; -} -#else -static int update_debugctl(struct thread_struct *prev, - struct thread_struct *next, unsigned long debugctl) -{ - return debugctl; -} -#endif /* CONFIG_X86_DS */ - static noinline void __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p, struct tss_struct *tss) { struct thread_struct *prev, *next; - unsigned long debugctl; prev = &prev_p->thread; next = &next_p->thread; - debugctl = update_debugctl(prev, next, prev->debugctlmsr); - - if (next->debugctlmsr != debugctl) + if (test_tsk_thread_flag(next_p, TIF_DS_AREA_MSR) || + test_tsk_thread_flag(prev_p, TIF_DS_AREA_MSR)) + ds_switch_to(prev_p, next_p); + else if (next->debugctlmsr != prev->debugctlmsr) update_debugctlmsr(next->debugctlmsr); if (test_tsk_thread_flag(next_p, TIF_DEBUG)) { @@ -482,15 +455,6 @@ __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p, hard_enable_TSC(); } -#ifdef CONFIG_X86_PTRACE_BTS - if (test_tsk_thread_flag(prev_p, TIF_BTS_TRACE_TS)) - ptrace_bts_take_timestamp(prev_p, BTS_TASK_DEPARTS); - - if (test_tsk_thread_flag(next_p, TIF_BTS_TRACE_TS)) - ptrace_bts_take_timestamp(next_p, BTS_TASK_ARRIVES); -#endif /* CONFIG_X86_PTRACE_BTS */ - - if (!test_tsk_thread_flag(next_p, TIF_IO_BITMAP)) { /* * Disable the bitmap via an invalid offset. We still cache @@ -548,7 +512,8 @@ __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..416fb9282f4 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> @@ -52,6 +53,7 @@ #include <asm/ia32.h> #include <asm/idle.h> #include <asm/syscalls.h> +#include <asm/ds.h> asmlinkage extern void ret_from_fork(void); @@ -235,14 +237,8 @@ void exit_thread(void) t->io_bitmap_max = 0; put_cpu(); } -#ifdef CONFIG_X86_DS - /* Free any DS contexts that have not been properly released. */ - if (unlikely(t->ds_ctx)) { - /* we clear debugctl to make sure DS is not used. */ - update_debugctlmsr(0); - ds_free(t->ds_ctx); - } -#endif /* CONFIG_X86_DS */ + + ds_exit_thread(current); } void flush_thread(void) @@ -372,6 +368,12 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long sp, if (err) goto out; } + + ds_copy_thread(p, me); + + clear_tsk_thread_flag(p, TIF_DEBUGCTLMSR); + p->thread.debugctlmsr = 0; + err = 0; out: if (err && p->thread.io_bitmap_ptr) { @@ -470,35 +472,14 @@ static inline void __switch_to_xtra(struct task_struct *prev_p, struct tss_struct *tss) { struct thread_struct *prev, *next; - unsigned long debugctl; prev = &prev_p->thread, next = &next_p->thread; - debugctl = prev->debugctlmsr; - -#ifdef CONFIG_X86_DS - { - unsigned long ds_prev = 0, ds_next = 0; - - if (prev->ds_ctx) - ds_prev = (unsigned long)prev->ds_ctx->ds; - if (next->ds_ctx) - ds_next = (unsigned long)next->ds_ctx->ds; - - if (ds_next != ds_prev) { - /* - * We clear debugctl to make sure DS - * is not in use when we change it: - */ - debugctl = 0; - update_debugctlmsr(0); - wrmsrl(MSR_IA32_DS_AREA, ds_next); - } - } -#endif /* CONFIG_X86_DS */ - - if (next->debugctlmsr != debugctl) + if (test_tsk_thread_flag(next_p, TIF_DS_AREA_MSR) || + test_tsk_thread_flag(prev_p, TIF_DS_AREA_MSR)) + ds_switch_to(prev_p, next_p); + else if (next->debugctlmsr != prev->debugctlmsr) update_debugctlmsr(next->debugctlmsr); if (test_tsk_thread_flag(next_p, TIF_DEBUG)) { @@ -533,14 +514,6 @@ static inline void __switch_to_xtra(struct task_struct *prev_p, */ memset(tss->io_bitmap, 0xff, prev->io_bitmap_max); } - -#ifdef CONFIG_X86_PTRACE_BTS - if (test_tsk_thread_flag(prev_p, TIF_BTS_TRACE_TS)) - ptrace_bts_take_timestamp(prev_p, BTS_TASK_DEPARTS); - - if (test_tsk_thread_flag(next_p, TIF_BTS_TRACE_TS)) - ptrace_bts_take_timestamp(next_p, BTS_TASK_ARRIVES); -#endif /* CONFIG_X86_PTRACE_BTS */ } /* @@ -551,8 +524,9 @@ 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..0a5df5f82fb 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -581,158 +581,91 @@ static int ioperm_get(struct task_struct *target, } #ifdef CONFIG_X86_PTRACE_BTS -/* - * The configuration for a particular BTS hardware implementation. - */ -struct bts_configuration { - /* the size of a BTS record in bytes; at most BTS_MAX_RECORD_SIZE */ - unsigned char sizeof_bts; - /* the size of a field in the BTS record in bytes */ - unsigned char sizeof_field; - /* a bitmask to enable/disable BTS in DEBUGCTL MSR */ - unsigned long debugctl_mask; -}; -static struct bts_configuration bts_cfg; - -#define BTS_MAX_RECORD_SIZE (8 * 3) - - -/* - * Branch Trace Store (BTS) uses the following format. Different - * architectures vary in the size of those fields. - * - source linear address - * - destination linear address - * - flags - * - * Later architectures use 64bit pointers throughout, whereas earlier - * architectures use 32bit pointers in 32bit mode. - * - * We compute the base address for the first 8 fields based on: - * - the field size stored in the DS configuration - * - the relative field position - * - * In order to store additional information in the BTS buffer, we use - * a special source address to indicate that the record requires - * special interpretation. - * - * Netburst indicated via a bit in the flags field whether the branch - * was predicted; this is ignored. - */ - -enum bts_field { - bts_from = 0, - bts_to, - bts_flags, - - bts_escape = (unsigned long)-1, - bts_qual = bts_to, - bts_jiffies = bts_flags -}; - -static inline unsigned long bts_get(const char *base, enum bts_field field) -{ - base += (bts_cfg.sizeof_field * field); - return *(unsigned long *)base; -} - -static inline void bts_set(char *base, enum bts_field field, unsigned long val) -{ - base += (bts_cfg.sizeof_field * field);; - (*(unsigned long *)base) = val; -} - -/* - * Translate a BTS record from the raw format into the bts_struct format - * - * out (out): bts_struct interpretation - * raw: raw BTS record - */ -static void ptrace_bts_translate_record(struct bts_struct *out, const void *raw) -{ - memset(out, 0, sizeof(*out)); - if (bts_get(raw, bts_from) == bts_escape) { - out->qualifier = bts_get(raw, bts_qual); - out->variant.jiffies = bts_get(raw, bts_jiffies); - } else { - out->qualifier = BTS_BRANCH; - out->variant.lbr.from_ip = bts_get(raw, bts_from); - out->variant.lbr.to_ip = bts_get(raw, bts_to); - } -} - static int ptrace_bts_read_record(struct task_struct *child, size_t index, struct bts_struct __user *out) { - struct bts_struct ret; - const void *bts_record; - size_t bts_index, bts_end; + const struct bts_trace *trace; + struct bts_struct bts; + const unsigned char *at; int error; - error = ds_get_bts_end(child, &bts_end); - if (error < 0) - return error; - - if (bts_end <= index) - return -EINVAL; + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; - error = ds_get_bts_index(child, &bts_index); - if (error < 0) - return error; + at = trace->ds.top - ((index + 1) * trace->ds.size); + if ((void *)at < trace->ds.begin) + at += (trace->ds.n * trace->ds.size); - /* translate the ptrace bts index into the ds bts index */ - bts_index += bts_end - (index + 1); - if (bts_end <= bts_index) - bts_index -= bts_end; + if (!trace->read) + return -EOPNOTSUPP; - error = ds_access_bts(child, bts_index, &bts_record); + error = trace->read(child->bts, at, &bts); if (error < 0) return error; - ptrace_bts_translate_record(&ret, bts_record); - - if (copy_to_user(out, &ret, sizeof(ret))) + if (copy_to_user(out, &bts, sizeof(bts))) return -EFAULT; - return sizeof(ret); + return sizeof(bts); } static int ptrace_bts_drain(struct task_struct *child, long size, struct bts_struct __user *out) { - struct bts_struct ret; - const unsigned char *raw; - size_t end, i; - int error; + const struct bts_trace *trace; + const unsigned char *at; + int error, drained = 0; - error = ds_get_bts_index(child, &end); - if (error < 0) - return error; + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; - if (size < (end * sizeof(struct bts_struct))) + if (!trace->read) + return -EOPNOTSUPP; + + if (size < (trace->ds.top - trace->ds.begin)) return -EIO; - error = ds_access_bts(child, 0, (const void **)&raw); - if (error < 0) - return error; + for (at = trace->ds.begin; (void *)at < trace->ds.top; + out++, drained++, at += trace->ds.size) { + struct bts_struct bts; + int error; - for (i = 0; i < end; i++, out++, raw += bts_cfg.sizeof_bts) { - ptrace_bts_translate_record(&ret, raw); + error = trace->read(child->bts, at, &bts); + if (error < 0) + return error; - if (copy_to_user(out, &ret, sizeof(ret))) + if (copy_to_user(out, &bts, sizeof(bts))) return -EFAULT; } - error = ds_clear_bts(child); + memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size); + + error = ds_reset_bts(child->bts); if (error < 0) return error; - return end; + return drained; } -static void ptrace_bts_ovfl(struct task_struct *child) +static int ptrace_bts_allocate_buffer(struct task_struct *child, size_t size) { - send_sig(child->thread.bts_ovfl_signal, child, 0); + child->bts_buffer = alloc_locked_buffer(size); + if (!child->bts_buffer) + return -ENOMEM; + + child->bts_size = size; + + return 0; +} + +static void ptrace_bts_free_buffer(struct task_struct *child) +{ + free_locked_buffer(child->bts_buffer, child->bts_size); + child->bts_buffer = NULL; + child->bts_size = 0; } static int ptrace_bts_config(struct task_struct *child, @@ -740,114 +673,86 @@ static int ptrace_bts_config(struct task_struct *child, const struct ptrace_bts_config __user *ucfg) { struct ptrace_bts_config cfg; - int error = 0; - - error = -EOPNOTSUPP; - if (!bts_cfg.sizeof_bts) - goto errout; + unsigned int flags = 0; - error = -EIO; if (cfg_size < sizeof(cfg)) - goto errout; + return -EIO; - error = -EFAULT; if (copy_from_user(&cfg, ucfg, sizeof(cfg))) - goto errout; + return -EFAULT; - error = -EINVAL; - if ((cfg.flags & PTRACE_BTS_O_SIGNAL) && - !(cfg.flags & PTRACE_BTS_O_ALLOC)) - goto errout; + if (child->bts) { + ds_release_bts(child->bts); + child->bts = NULL; + } - if (cfg.flags & PTRACE_BTS_O_ALLOC) { - ds_ovfl_callback_t ovfl = NULL; - unsigned int sig = 0; + if (cfg.flags & PTRACE_BTS_O_SIGNAL) { + if (!cfg.signal) + return -EINVAL; - /* we ignore the error in case we were not tracing child */ - (void)ds_release_bts(child); + return -EOPNOTSUPP; - if (cfg.flags & PTRACE_BTS_O_SIGNAL) { - if (!cfg.signal) - goto errout; + child->thread.bts_ovfl_signal = cfg.signal; + } - sig = cfg.signal; - ovfl = ptrace_bts_ovfl; - } + if ((cfg.flags & PTRACE_BTS_O_ALLOC) && + (cfg.size != child->bts_size)) { + int error; - error = ds_request_bts(child, /* base = */ NULL, cfg.size, ovfl); - if (error < 0) - goto errout; + ptrace_bts_free_buffer(child); - child->thread.bts_ovfl_signal = sig; + error = ptrace_bts_allocate_buffer(child, cfg.size); + if (error < 0) + return error; } - error = -EINVAL; - if (!child->thread.ds_ctx && cfg.flags) - goto errout; - if (cfg.flags & PTRACE_BTS_O_TRACE) - child->thread.debugctlmsr |= bts_cfg.debugctl_mask; - else - child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; + flags |= BTS_USER; if (cfg.flags & PTRACE_BTS_O_SCHED) - set_tsk_thread_flag(child, TIF_BTS_TRACE_TS); - else - clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS); + flags |= BTS_TIMESTAMPS; - error = sizeof(cfg); + child->bts = ds_request_bts(child, child->bts_buffer, child->bts_size, + /* ovfl = */ NULL, /* th = */ (size_t)-1, + flags); + if (IS_ERR(child->bts)) { + int error = PTR_ERR(child->bts); -out: - if (child->thread.debugctlmsr) - set_tsk_thread_flag(child, TIF_DEBUGCTLMSR); - else - clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR); + ptrace_bts_free_buffer(child); + child->bts = NULL; - return error; + return error; + } -errout: - child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; - clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS); - goto out; + return sizeof(cfg); } static int ptrace_bts_status(struct task_struct *child, long cfg_size, struct ptrace_bts_config __user *ucfg) { + const struct bts_trace *trace; struct ptrace_bts_config cfg; - size_t end; - const void *base, *max; - int error; if (cfg_size < sizeof(cfg)) return -EIO; - error = ds_get_bts_end(child, &end); - if (error < 0) - return error; - - error = ds_access_bts(child, /* index = */ 0, &base); - if (error < 0) - return error; - - error = ds_access_bts(child, /* index = */ end, &max); - if (error < 0) - return error; + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; memset(&cfg, 0, sizeof(cfg)); - cfg.size = (max - base); + cfg.size = trace->ds.end - trace->ds.begin; cfg.signal = child->thread.bts_ovfl_signal; cfg.bts_size = sizeof(struct bts_struct); if (cfg.signal) cfg.flags |= PTRACE_BTS_O_SIGNAL; - if (test_tsk_thread_flag(child, TIF_DEBUGCTLMSR) && - child->thread.debugctlmsr & bts_cfg.debugctl_mask) + if (trace->ds.flags & BTS_USER) cfg.flags |= PTRACE_BTS_O_TRACE; - if (test_tsk_thread_flag(child, TIF_BTS_TRACE_TS)) + if (trace->ds.flags & BTS_TIMESTAMPS) cfg.flags |= PTRACE_BTS_O_SCHED; if (copy_to_user(ucfg, &cfg, sizeof(cfg))) @@ -856,110 +761,77 @@ static int ptrace_bts_status(struct task_struct *child, return sizeof(cfg); } -static int ptrace_bts_write_record(struct task_struct *child, - const struct bts_struct *in) +static int ptrace_bts_clear(struct task_struct *child) { - unsigned char bts_record[BTS_MAX_RECORD_SIZE]; + const struct bts_trace *trace; - BUG_ON(BTS_MAX_RECORD_SIZE < bts_cfg.sizeof_bts); + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; - memset(bts_record, 0, bts_cfg.sizeof_bts); - switch (in->qualifier) { - case BTS_INVALID: - break; + memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size); - case BTS_BRANCH: - bts_set(bts_record, bts_from, in->variant.lbr.from_ip); - bts_set(bts_record, bts_to, in->variant.lbr.to_ip); - break; + return ds_reset_bts(child->bts); +} - case BTS_TASK_ARRIVES: - case BTS_TASK_DEPARTS: - bts_set(bts_record, bts_from, bts_escape); - bts_set(bts_record, bts_qual, in->qualifier); - bts_set(bts_record, bts_jiffies, in->variant.jiffies); - break; +static int ptrace_bts_size(struct task_struct *child) +{ + const struct bts_trace *trace; - default: - return -EINVAL; - } + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; - /* 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 (trace->ds.top - trace->ds.begin) / trace->ds.size; } -void ptrace_bts_take_timestamp(struct task_struct *tsk, - enum bts_qualifier qualifier) +static void ptrace_bts_fork(struct task_struct *tsk) { - struct bts_struct rec = { - .qualifier = qualifier, - .variant.jiffies = jiffies_64 - }; - - ptrace_bts_write_record(tsk, &rec); + tsk->bts = NULL; + tsk->bts_buffer = NULL; + tsk->bts_size = 0; + tsk->thread.bts_ovfl_signal = 0; } -static const struct bts_configuration bts_cfg_netburst = { - .sizeof_bts = sizeof(long) * 3, - .sizeof_field = sizeof(long), - .debugctl_mask = (1<<2)|(1<<3)|(1<<5) -}; +static void ptrace_bts_untrace(struct task_struct *child) +{ + if (unlikely(child->bts)) { + ds_release_bts(child->bts); + child->bts = NULL; + + /* We cannot update total_vm and locked_vm since + child's mm is already gone. But we can reclaim the + memory. */ + kfree(child->bts_buffer); + child->bts_buffer = NULL; + child->bts_size = 0; + } +} -static const struct bts_configuration bts_cfg_pentium_m = { - .sizeof_bts = sizeof(long) * 3, - .sizeof_field = sizeof(long), - .debugctl_mask = (1<<6)|(1<<7) -}; +static void ptrace_bts_detach(struct task_struct *child) +{ + if (unlikely(child->bts)) { + ds_release_bts(child->bts); + child->bts = NULL; -static const struct bts_configuration bts_cfg_core2 = { - .sizeof_bts = 8 * 3, - .sizeof_field = 8, - .debugctl_mask = (1<<6)|(1<<7)|(1<<9) -}; + ptrace_bts_free_buffer(child); + } +} +#else +static inline void ptrace_bts_fork(struct task_struct *tsk) {} +static inline void ptrace_bts_detach(struct task_struct *child) {} +static inline void ptrace_bts_untrace(struct task_struct *child) {} +#endif /* CONFIG_X86_PTRACE_BTS */ -static inline void bts_configure(const struct bts_configuration *cfg) +void x86_ptrace_fork(struct task_struct *child, unsigned long clone_flags) { - bts_cfg = *cfg; + ptrace_bts_fork(child); } -void __cpuinit ptrace_bts_init_intel(struct cpuinfo_x86 *c) +void x86_ptrace_untrace(struct task_struct *child) { - switch (c->x86) { - case 0x6: - switch (c->x86_model) { - case 0xD: - case 0xE: /* Pentium M */ - bts_configure(&bts_cfg_pentium_m); - break; - case 0xF: /* Core2 */ - case 0x1C: /* Atom */ - bts_configure(&bts_cfg_core2); - break; - default: - /* sorry, don't know about them */ - break; - } - break; - case 0xF: - switch (c->x86_model) { - case 0x0: - case 0x1: - case 0x2: /* Netburst */ - bts_configure(&bts_cfg_netburst); - break; - default: - /* sorry, don't know about them */ - break; - } - break; - default: - /* sorry, don't know about them */ - break; - } + ptrace_bts_untrace(child); } -#endif /* CONFIG_X86_PTRACE_BTS */ /* * Called by kernel/ptrace.c when detaching.. @@ -972,15 +844,7 @@ void ptrace_disable(struct task_struct *child) #ifdef TIF_SYSCALL_EMU clear_tsk_thread_flag(child, TIF_SYSCALL_EMU); #endif -#ifdef CONFIG_X86_PTRACE_BTS - (void)ds_release_bts(child); - - 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); -#endif /* CONFIG_X86_PTRACE_BTS */ + ptrace_bts_detach(child); } #if defined CONFIG_X86_32 || defined CONFIG_IA32_EMULATION @@ -1112,7 +976,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) break; case PTRACE_BTS_SIZE: - ret = ds_get_bts_index(child, /* pos = */ NULL); + ret = ptrace_bts_size(child); break; case PTRACE_BTS_GET: @@ -1121,7 +985,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) break; case PTRACE_BTS_CLEAR: - ret = ds_clear_bts(child); + ret = ptrace_bts_clear(child); break; case PTRACE_BTS_DRAIN: @@ -1384,6 +1248,14 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request, case PTRACE_GET_THREAD_AREA: case PTRACE_SET_THREAD_AREA: +#ifdef CONFIG_X86_PTRACE_BTS + case PTRACE_BTS_CONFIG: + case PTRACE_BTS_STATUS: + case PTRACE_BTS_SIZE: + case PTRACE_BTS_GET: + case PTRACE_BTS_CLEAR: + case PTRACE_BTS_DRAIN: +#endif /* CONFIG_X86_PTRACE_BTS */ return arch_ptrace(child, request, addr, data); default: diff --git a/arch/x86/kernel/quirks.c b/arch/x86/kernel/quirks.c index 67465ed8931..309949e9e1c 100644 --- a/arch/x86/kernel/quirks.c +++ b/arch/x86/kernel/quirks.c @@ -168,6 +168,8 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH7_31, ich_force_enable_hpet); DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH8_1, ich_force_enable_hpet); +DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH8_4, + ich_force_enable_hpet); DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH9_7, ich_force_enable_hpet); diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index cc5a2545dd4..2b46eb41643 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -12,6 +12,8 @@ #include <asm/proto.h> #include <asm/reboot_fixups.h> #include <asm/reboot.h> +#include <asm/pci_x86.h> +#include <asm/virtext.h> #ifdef CONFIG_X86_32 # include <linux/dmi.h> @@ -21,6 +23,8 @@ # include <asm/iommu.h> #endif +#include <mach_ipi.h> + /* * Power off function, if any */ @@ -36,7 +40,16 @@ 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 if we need to go through the 'emergency' path. + * When machine_emergency_restart() is called, we may be on + * an inconsistent state and won't be able to do a clean cleanup + */ +static int reboot_emergency; + +/* 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 +58,7 @@ 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 +93,7 @@ static int __init reboot_setup(char *str) case 'k': case 't': case 'e': + case 'p': reboot_type = *str; break; @@ -360,6 +375,48 @@ static inline void kb_wait(void) } } +static void vmxoff_nmi(int cpu, struct die_args *args) +{ + cpu_emergency_vmxoff(); +} + +/* Use NMIs as IPIs to tell all CPUs to disable virtualization + */ +static void emergency_vmx_disable_all(void) +{ + /* Just make sure we won't change CPUs while doing this */ + local_irq_disable(); + + /* We need to disable VMX on all CPUs before rebooting, otherwise + * we risk hanging up the machine, because the CPU ignore INIT + * signals when VMX is enabled. + * + * We can't take any locks and we may be on an inconsistent + * state, so we use NMIs as IPIs to tell the other CPUs to disable + * VMX and halt. + * + * For safety, we will avoid running the nmi_shootdown_cpus() + * stuff unnecessarily, but we don't have a way to check + * if other CPUs have VMX enabled. So we will call it only if the + * CPU we are running on has VMX enabled. + * + * We will miss cases where VMX is not enabled on all CPUs. This + * shouldn't do much harm because KVM always enable VMX on all + * CPUs anyway. But we can miss it on the small window where KVM + * is still enabling VMX. + */ + if (cpu_has_vmx() && cpu_vmx_enabled()) { + /* Disable VMX on this CPU. + */ + cpu_vmxoff(); + + /* Halt and disable VMX on the other CPUs */ + nmi_shootdown_cpus(vmxoff_nmi); + + } +} + + void __attribute__((weak)) mach_reboot_fixups(void) { } @@ -368,6 +425,9 @@ static void native_machine_emergency_restart(void) { int i; + if (reboot_emergency) + emergency_vmx_disable_all(); + /* Tell the BIOS if we want cold or warm reboot */ *((unsigned short *)__va(0x472)) = reboot_mode; @@ -404,12 +464,27 @@ 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; } @@ -426,7 +501,7 @@ void native_machine_shutdown(void) #ifdef CONFIG_X86_32 /* See if there has been given a command line override */ - if ((reboot_cpu != -1) && (reboot_cpu < NR_CPUS) && + if ((reboot_cpu != -1) && (reboot_cpu < nr_cpu_ids) && cpu_online(reboot_cpu)) reboot_cpu_id = reboot_cpu; #endif @@ -436,7 +511,7 @@ void native_machine_shutdown(void) reboot_cpu_id = smp_processor_id(); /* Make certain I only run on the appropriate processor */ - set_cpus_allowed_ptr(current, &cpumask_of_cpu(reboot_cpu_id)); + set_cpus_allowed_ptr(current, cpumask_of(reboot_cpu_id)); /* O.K Now that I'm on the appropriate processor, * stop all of the others. @@ -459,17 +534,28 @@ void native_machine_shutdown(void) #endif } +static void __machine_emergency_restart(int emergency) +{ + reboot_emergency = emergency; + machine_ops.emergency_restart(); +} + static void native_machine_restart(char *__unused) { printk("machine restart\n"); if (!reboot_force) machine_shutdown(); - machine_emergency_restart(); + __machine_emergency_restart(0); } 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) @@ -504,7 +590,7 @@ void machine_shutdown(void) void machine_emergency_restart(void) { - machine_ops.emergency_restart(); + __machine_emergency_restart(1); } void machine_restart(char *cmd) @@ -523,3 +609,92 @@ 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 9d5674f7b6c..ae0d8042cf6 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -93,11 +93,13 @@ #include <asm/desc.h> #include <asm/dma.h> #include <asm/iommu.h> +#include <asm/gart.h> #include <asm/mmu_context.h> #include <asm/proto.h> #include <mach_apic.h> #include <asm/paravirt.h> +#include <asm/hypervisor.h> #include <asm/percpu.h> #include <asm/topology.h> @@ -448,6 +450,7 @@ static void __init reserve_early_setup_data(void) * @size: Size of the crashkernel memory to reserve. * Returns the base address on success, and -1ULL on failure. */ +static unsigned long long __init find_and_reserve_crashkernel(unsigned long long size) { const unsigned long long alignment = 16<<20; /* 16M */ @@ -583,161 +586,24 @@ static int __init setup_elfcorehdr(char *arg) early_param("elfcorehdr", setup_elfcorehdr); #endif -static struct x86_quirks default_x86_quirks __initdata; - -struct x86_quirks *x86_quirks __initdata = &default_x86_quirks; - -/* - * Some BIOSes seem to corrupt the low 64k of memory during events - * like suspend/resume and unplugging an HDMI cable. Reserve all - * remaining free memory in that area and fill it with a distinct - * pattern. - */ -#ifdef CONFIG_X86_CHECK_BIOS_CORRUPTION -#define MAX_SCAN_AREAS 8 - -static int __read_mostly memory_corruption_check = -1; - -static unsigned __read_mostly corruption_check_size = 64*1024; -static unsigned __read_mostly corruption_check_period = 60; /* seconds */ - -static struct e820entry scan_areas[MAX_SCAN_AREAS]; -static int num_scan_areas; - - -static int set_corruption_check(char *arg) +static int __init default_update_genapic(void) { - char *end; - - memory_corruption_check = simple_strtol(arg, &end, 10); - - return (*end == 0) ? 0 : -EINVAL; -} -early_param("memory_corruption_check", set_corruption_check); - -static int set_corruption_check_period(char *arg) -{ - char *end; - - corruption_check_period = simple_strtoul(arg, &end, 10); - - return (*end == 0) ? 0 : -EINVAL; -} -early_param("memory_corruption_check_period", set_corruption_check_period); - -static int set_corruption_check_size(char *arg) -{ - char *end; - unsigned size; - - size = memparse(arg, &end); - - if (*end == '\0') - corruption_check_size = size; - - return (size == corruption_check_size) ? 0 : -EINVAL; -} -early_param("memory_corruption_check_size", set_corruption_check_size); - - -static void __init setup_bios_corruption_check(void) -{ - u64 addr = PAGE_SIZE; /* assume first page is reserved anyway */ - - if (memory_corruption_check == -1) { - memory_corruption_check = -#ifdef CONFIG_X86_BOOTPARAM_MEMORY_CORRUPTION_CHECK - 1 -#else - 0 +#ifdef CONFIG_X86_SMP +# if defined(CONFIG_X86_GENERICARCH) || defined(CONFIG_X86_64) + genapic->wakeup_cpu = wakeup_secondary_cpu_via_init; +# endif #endif - ; - } - - if (corruption_check_size == 0) - memory_corruption_check = 0; - - if (!memory_corruption_check) - return; - - corruption_check_size = round_up(corruption_check_size, PAGE_SIZE); - - while(addr < corruption_check_size && num_scan_areas < MAX_SCAN_AREAS) { - u64 size; - addr = find_e820_area_size(addr, &size, PAGE_SIZE); - - if (addr == 0) - break; - if ((addr + size) > corruption_check_size) - size = corruption_check_size - addr; - - if (size == 0) - break; - - e820_update_range(addr, size, E820_RAM, E820_RESERVED); - scan_areas[num_scan_areas].addr = addr; - scan_areas[num_scan_areas].size = size; - num_scan_areas++; - - /* Assume we've already mapped this early memory */ - memset(__va(addr), 0, size); - - addr += size; - } - - printk(KERN_INFO "Scanning %d areas for low memory corruption\n", - num_scan_areas); - update_e820(); -} - -static struct timer_list periodic_check_timer; - -void check_for_bios_corruption(void) -{ - int i; - int corruption = 0; - - if (!memory_corruption_check) - return; - - for(i = 0; i < num_scan_areas; i++) { - unsigned long *addr = __va(scan_areas[i].addr); - unsigned long size = scan_areas[i].size; - - for(; size; addr++, size -= sizeof(unsigned long)) { - if (!*addr) - continue; - printk(KERN_ERR "Corrupted low memory at %p (%lx phys) = %08lx\n", - addr, __pa(addr), *addr); - corruption = 1; - *addr = 0; - } - } - - WARN(corruption, KERN_ERR "Memory corruption detected in low memory\n"); -} - -static void periodic_check_for_corruption(unsigned long data) -{ - check_for_bios_corruption(); - mod_timer(&periodic_check_timer, round_jiffies(jiffies + corruption_check_period*HZ)); + return 0; } -void start_periodic_check_for_corruption(void) -{ - if (!memory_corruption_check || corruption_check_period == 0) - return; - - printk(KERN_INFO "Scanning for low memory corruption every %d seconds\n", - corruption_check_period); +static struct x86_quirks default_x86_quirks __initdata = { + .update_genapic = default_update_genapic, +}; - init_timer(&periodic_check_timer); - periodic_check_timer.function = &periodic_check_for_corruption; - periodic_check_for_corruption(0); -} -#endif +struct x86_quirks *x86_quirks __initdata = &default_x86_quirks; +#ifdef CONFIG_X86_RESERVE_LOW_64K static int __init dmi_low_memory_corruption(const struct dmi_system_id *d) { printk(KERN_NOTICE @@ -749,6 +615,7 @@ static int __init dmi_low_memory_corruption(const struct dmi_system_id *d) return 0; } +#endif /* List of systems that have known low memory corruption BIOS problems */ static struct dmi_system_id __initdata bad_bios_dmi_table[] = { @@ -794,6 +661,9 @@ 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 +750,8 @@ 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(); @@ -909,6 +774,12 @@ void __init setup_arch(char **cmdline_p) dmi_check_system(bad_bios_dmi_table); + /* + * VMware detection requires dmi to be available, so this + * needs to be done after dmi_scan_machine, for the BP. + */ + init_hypervisor(&boot_cpu_data); + #ifdef CONFIG_X86_32 probe_roms(); #endif @@ -1082,7 +953,7 @@ 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..a4b619c3310 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -152,8 +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 "PERCPU: Allocating %zd bytes of per cpu data\n", - size); + + pr_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); + + pr_info("PERCPU: Allocating %zd bytes of per cpu data\n", size); for_each_possible_cpu(cpu) { #ifndef CONFIG_NEED_MULTIPLE_NODES @@ -164,28 +167,21 @@ void __init setup_per_cpu_areas(void) if (!node_online(node) || !NODE_DATA(node)) { ptr = __alloc_bootmem(size, align, __pa(MAX_DMA_ADDRESS)); - printk(KERN_INFO - "cpu %d has no node %d or node-local memory\n", + pr_info("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", - cpu, __pa(ptr)); - } - else { + pr_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)); + pr_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 +278,7 @@ 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); } @@ -334,25 +330,25 @@ static const cpumask_t cpu_mask_none; /* * Returns a pointer to the bitmask of CPUs on Node 'node'. */ -const cpumask_t *_node_to_cpumask_ptr(int node) +const cpumask_t *cpumask_of_node(int node) { if (node_to_cpumask_map == NULL) { printk(KERN_WARNING - "_node_to_cpumask_ptr(%d): no node_to_cpumask_map!\n", + "cpumask_of_node(%d): no node_to_cpumask_map!\n", node); dump_stack(); return (const cpumask_t *)&cpu_online_map; } if (node >= nr_node_ids) { printk(KERN_WARNING - "_node_to_cpumask_ptr(%d): node > nr_node_ids(%d)\n", + "cpumask_of_node(%d): node > nr_node_ids(%d)\n", node, nr_node_ids); dump_stack(); return &cpu_mask_none; } return &node_to_cpumask_map[node]; } -EXPORT_SYMBOL(_node_to_cpumask_ptr); +EXPORT_SYMBOL(cpumask_of_node); /* * Returns a bitmask of CPUs on Node 'node'. diff --git a/arch/x86/kernel/sigframe.h b/arch/x86/kernel/sigframe.h deleted file mode 100644 index cc673aa55ce..00000000000 --- a/arch/x86/kernel/sigframe.h +++ /dev/null @@ -1,42 +0,0 @@ -#ifdef CONFIG_X86_32 -struct sigframe { - char __user *pretcode; - int sig; - struct sigcontext sc; - /* - * fpstate is unused. fpstate is moved/allocated after - * retcode[] below. This movement allows to have the FP state and the - * future state extensions (xsave) stay together. - * And at the same time retaining the unused fpstate, prevents changing - * the offset of extramask[] in the sigframe and thus prevent any - * legacy application accessing/modifying it. - */ - struct _fpstate fpstate_unused; - unsigned long extramask[_NSIG_WORDS-1]; - char retcode[8]; - /* fp state follows here */ -}; - -struct rt_sigframe { - char __user *pretcode; - int sig; - struct siginfo __user *pinfo; - void __user *puc; - struct siginfo info; - struct ucontext uc; - char retcode[8]; - /* fp state follows here */ -}; -#else -struct rt_sigframe { - char __user *pretcode; - struct ucontext uc; - struct siginfo info; - /* fp state follows here */ -}; - -int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs); -int ia32_setup_frame(int sig, struct k_sigaction *ka, - sigset_t *set, struct pt_regs *regs); -#endif diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal.c index d6dd057d0f2..89bb7668041 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal.c @@ -1,36 +1,41 @@ /* * Copyright (C) 1991, 1992 Linus Torvalds + * Copyright (C) 2000, 2001, 2002 Andi Kleen SuSE Labs * * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson * 2000-06-20 Pentium III FXSR, SSE support by Gareth Hughes + * 2000-2002 x86-64 support by Andi Kleen */ -#include <linux/list.h> -#include <linux/personality.h> -#include <linux/binfmts.h> -#include <linux/suspend.h> +#include <linux/sched.h> +#include <linux/mm.h> +#include <linux/smp.h> #include <linux/kernel.h> -#include <linux/ptrace.h> #include <linux/signal.h> -#include <linux/stddef.h> -#include <linux/unistd.h> #include <linux/errno.h> -#include <linux/sched.h> #include <linux/wait.h> +#include <linux/ptrace.h> #include <linux/tracehook.h> -#include <linux/elf.h> -#include <linux/smp.h> -#include <linux/mm.h> +#include <linux/unistd.h> +#include <linux/stddef.h> +#include <linux/personality.h> +#include <linux/uaccess.h> #include <asm/processor.h> #include <asm/ucontext.h> -#include <asm/uaccess.h> #include <asm/i387.h> #include <asm/vdso.h> + +#ifdef CONFIG_X86_64 +#include <asm/proto.h> +#include <asm/ia32_unistd.h> +#include <asm/mce.h> +#endif /* CONFIG_X86_64 */ + #include <asm/syscall.h> #include <asm/syscalls.h> -#include "sigframe.h" +#include <asm/sigframe.h> #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) @@ -45,74 +50,6 @@ # define FIX_EFLAGS __FIX_EFLAGS #endif -/* - * Atomically swap in the new signal mask, and wait for a signal. - */ -asmlinkage int -sys_sigsuspend(int history0, int history1, old_sigset_t mask) -{ - mask &= _BLOCKABLE; - spin_lock_irq(¤t->sighand->siglock); - current->saved_sigmask = current->blocked; - siginitset(¤t->blocked, mask); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - current->state = TASK_INTERRUPTIBLE; - schedule(); - set_restore_sigmask(); - - return -ERESTARTNOHAND; -} - -asmlinkage int -sys_sigaction(int sig, const struct old_sigaction __user *act, - struct old_sigaction __user *oact) -{ - struct k_sigaction new_ka, old_ka; - int ret; - - if (act) { - old_sigset_t mask; - - if (!access_ok(VERIFY_READ, act, sizeof(*act)) || - __get_user(new_ka.sa.sa_handler, &act->sa_handler) || - __get_user(new_ka.sa.sa_restorer, &act->sa_restorer)) - return -EFAULT; - - __get_user(new_ka.sa.sa_flags, &act->sa_flags); - __get_user(mask, &act->sa_mask); - siginitset(&new_ka.sa.sa_mask, mask); - } - - ret = do_sigaction(sig, act ? &new_ka : NULL, oact ? &old_ka : NULL); - - if (!ret && oact) { - if (!access_ok(VERIFY_WRITE, oact, sizeof(*oact)) || - __put_user(old_ka.sa.sa_handler, &oact->sa_handler) || - __put_user(old_ka.sa.sa_restorer, &oact->sa_restorer)) - return -EFAULT; - - __put_user(old_ka.sa.sa_flags, &oact->sa_flags); - __put_user(old_ka.sa.sa_mask.sig[0], &oact->sa_mask); - } - - return ret; -} - -asmlinkage int sys_sigaltstack(unsigned long bx) -{ - /* - * This is needed to make gcc realize it doesn't own the - * "struct pt_regs" - */ - struct pt_regs *regs = (struct pt_regs *)&bx; - const stack_t __user *uss = (const stack_t __user *)bx; - stack_t __user *uoss = (stack_t __user *)regs->cx; - - return do_sigaltstack(uss, uoss, regs->sp); -} - #define COPY(x) { \ err |= __get_user(regs->x, &sc->x); \ } @@ -123,7 +60,7 @@ asmlinkage int sys_sigaltstack(unsigned long bx) regs->seg = tmp; \ } -#define COPY_SEG_STRICT(seg) { \ +#define COPY_SEG_CPL3(seg) { \ unsigned short tmp; \ err |= __get_user(tmp, &sc->seg); \ regs->seg = tmp | 3; \ @@ -135,9 +72,6 @@ asmlinkage int sys_sigaltstack(unsigned long bx) loadsegment(seg, tmp); \ } -/* - * Do a signal return; undo the signal stack. - */ static int restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, unsigned long *pax) @@ -149,14 +83,36 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, /* Always make any pending restarted system calls return -EINTR */ current_thread_info()->restart_block.fn = do_no_restart_syscall; +#ifdef CONFIG_X86_32 GET_SEG(gs); COPY_SEG(fs); COPY_SEG(es); COPY_SEG(ds); +#endif /* CONFIG_X86_32 */ + COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); COPY(dx); COPY(cx); COPY(ip); - COPY_SEG_STRICT(cs); - COPY_SEG_STRICT(ss); + +#ifdef CONFIG_X86_64 + COPY(r8); + COPY(r9); + COPY(r10); + COPY(r11); + COPY(r12); + COPY(r13); + COPY(r14); + COPY(r15); +#endif /* CONFIG_X86_64 */ + +#ifdef CONFIG_X86_32 + COPY_SEG_CPL3(cs); + COPY_SEG_CPL3(ss); +#else /* !CONFIG_X86_32 */ + /* Kernel saves and restores only the CS segment register on signals, + * which is the bare minimum needed to allow mixed 32/64-bit code. + * App's signal handler can save/restore other segments if needed. */ + COPY_SEG_CPL3(cs); +#endif /* CONFIG_X86_32 */ err |= __get_user(tmpflags, &sc->flags); regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); @@ -169,102 +125,24 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, return err; } -asmlinkage unsigned long sys_sigreturn(unsigned long __unused) -{ - struct sigframe __user *frame; - struct pt_regs *regs; - unsigned long ax; - sigset_t set; - - regs = (struct pt_regs *) &__unused; - frame = (struct sigframe __user *)(regs->sp - 8); - - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1 - && __copy_from_user(&set.sig[1], &frame->extramask, - sizeof(frame->extramask)))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->sc, &ax)) - goto badframe; - return ax; - -badframe: - if (show_unhandled_signals && printk_ratelimit()) { - printk("%s%s[%d] bad frame in sigreturn frame:" - "%p ip:%lx sp:%lx oeax:%lx", - task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, - current->comm, task_pid_nr(current), frame, regs->ip, - regs->sp, regs->orig_ax); - print_vma_addr(" in ", regs->ip); - printk(KERN_CONT "\n"); - } - - force_sig(SIGSEGV, current); - - return 0; -} - -static long do_rt_sigreturn(struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - unsigned long ax; - sigset_t set; - - frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) - goto badframe; - - if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) - goto badframe; - - return ax; - -badframe: - signal_fault(regs, frame, "rt_sigreturn"); - return 0; -} - -asmlinkage int sys_rt_sigreturn(unsigned long __unused) -{ - struct pt_regs *regs = (struct pt_regs *)&__unused; - - return do_rt_sigreturn(regs); -} - -/* - * Set up a signal frame. - */ static int setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, struct pt_regs *regs, unsigned long mask) { - int tmp, err = 0; + int err = 0; - err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); - savesegment(gs, tmp); - err |= __put_user(tmp, (unsigned int __user *)&sc->gs); +#ifdef CONFIG_X86_32 + { + unsigned int tmp; + savesegment(gs, tmp); + err |= __put_user(tmp, (unsigned int __user *)&sc->gs); + } + err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); err |= __put_user(regs->es, (unsigned int __user *)&sc->es); err |= __put_user(regs->ds, (unsigned int __user *)&sc->ds); +#endif /* CONFIG_X86_32 */ + err |= __put_user(regs->di, &sc->di); err |= __put_user(regs->si, &sc->si); err |= __put_user(regs->bp, &sc->bp); @@ -273,19 +151,33 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, err |= __put_user(regs->dx, &sc->dx); err |= __put_user(regs->cx, &sc->cx); err |= __put_user(regs->ax, &sc->ax); +#ifdef CONFIG_X86_64 + err |= __put_user(regs->r8, &sc->r8); + err |= __put_user(regs->r9, &sc->r9); + err |= __put_user(regs->r10, &sc->r10); + err |= __put_user(regs->r11, &sc->r11); + err |= __put_user(regs->r12, &sc->r12); + err |= __put_user(regs->r13, &sc->r13); + err |= __put_user(regs->r14, &sc->r14); + err |= __put_user(regs->r15, &sc->r15); +#endif /* CONFIG_X86_64 */ + err |= __put_user(current->thread.trap_no, &sc->trapno); err |= __put_user(current->thread.error_code, &sc->err); err |= __put_user(regs->ip, &sc->ip); +#ifdef CONFIG_X86_32 err |= __put_user(regs->cs, (unsigned int __user *)&sc->cs); err |= __put_user(regs->flags, &sc->flags); err |= __put_user(regs->sp, &sc->sp_at_signal); err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); +#else /* !CONFIG_X86_32 */ + err |= __put_user(regs->flags, &sc->flags); + err |= __put_user(regs->cs, &sc->cs); + err |= __put_user(0, &sc->gs); + err |= __put_user(0, &sc->fs); +#endif /* CONFIG_X86_32 */ - tmp = save_i387_xstate(fpstate); - if (tmp < 0) - err = 1; - else - err |= __put_user(tmp ? fpstate : NULL, &sc->fpstate); + err |= __put_user(fpstate, &sc->fpstate); /* non-iBCS2 extensions.. */ err |= __put_user(mask, &sc->oldmask); @@ -295,6 +187,32 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, } /* + * Set up a signal frame. + */ +#ifdef CONFIG_X86_32 +static const struct { + u16 poplmovl; + u32 val; + u16 int80; +} __attribute__((packed)) retcode = { + 0xb858, /* popl %eax; movl $..., %eax */ + __NR_sigreturn, + 0x80cd, /* int $0x80 */ +}; + +static const struct { + u8 movl; + u32 val; + u16 int80; + u8 pad; +} __attribute__((packed)) rt_retcode = { + 0xb8, /* movl $..., %eax */ + __NR_rt_sigreturn, + 0x80cd, /* int $0x80 */ + 0 +}; + +/* * Determine which stack to use.. */ static inline void __user * @@ -328,6 +246,8 @@ get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size, if (used_math()) { sp = sp - sig_xstate_size; *fpstate = (struct _fpstate *) sp; + if (save_i387_xstate(*fpstate) < 0) + return (void __user *)-1L; } sp -= frame_size; @@ -383,9 +303,7 @@ __setup_frame(int sig, struct k_sigaction *ka, sigset_t *set, * reasons and because gdb uses it as a signature to notice * signal handler stack frames. */ - err |= __put_user(0xb858, (short __user *)(frame->retcode+0)); - err |= __put_user(__NR_sigreturn, (int __user *)(frame->retcode+2)); - err |= __put_user(0x80cd, (short __user *)(frame->retcode+6)); + err |= __put_user(*((u64 *)&retcode), (u64 *)frame->retcode); if (err) return -EFAULT; @@ -454,9 +372,7 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, * reasons and because gdb uses it as a signature to notice * signal handler stack frames. */ - err |= __put_user(0xb8, (char __user *)(frame->retcode+0)); - err |= __put_user(__NR_rt_sigreturn, (int __user *)(frame->retcode+1)); - err |= __put_user(0x80cd, (short __user *)(frame->retcode+5)); + err |= __put_user(*((u64 *)&rt_retcode), (u64 *)frame->retcode); if (err) return -EFAULT; @@ -475,23 +391,293 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, return 0; } +#else /* !CONFIG_X86_32 */ +/* + * Determine which stack to use.. + */ +static void __user * +get_stack(struct k_sigaction *ka, unsigned long sp, unsigned long size) +{ + /* Default to using normal stack - redzone*/ + sp -= 128; + + /* This is the X/Open sanctioned signal stack switching. */ + if (ka->sa.sa_flags & SA_ONSTACK) { + if (sas_ss_flags(sp) == 0) + sp = current->sas_ss_sp + current->sas_ss_size; + } + + return (void __user *)round_down(sp - size, 64); +} + +static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + void __user *fp = NULL; + int err = 0; + struct task_struct *me = current; + + if (used_math()) { + fp = get_stack(ka, regs->sp, sig_xstate_size); + frame = (void __user *)round_down( + (unsigned long)fp - sizeof(struct rt_sigframe), 16) - 8; + + if (save_i387_xstate(fp) < 0) + return -EFAULT; + } else + frame = get_stack(ka, regs->sp, sizeof(struct rt_sigframe)) - 8; + + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + return -EFAULT; + + if (ka->sa.sa_flags & SA_SIGINFO) { + if (copy_siginfo_to_user(&frame->info, info)) + return -EFAULT; + } + + /* Create the ucontext. */ + if (cpu_has_xsave) + err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); + else + err |= __put_user(0, &frame->uc.uc_flags); + err |= __put_user(0, &frame->uc.uc_link); + err |= __put_user(me->sas_ss_sp, &frame->uc.uc_stack.ss_sp); + err |= __put_user(sas_ss_flags(regs->sp), + &frame->uc.uc_stack.ss_flags); + err |= __put_user(me->sas_ss_size, &frame->uc.uc_stack.ss_size); + err |= setup_sigcontext(&frame->uc.uc_mcontext, fp, regs, set->sig[0]); + err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); + + /* Set up to return from userspace. If provided, use a stub + already in userspace. */ + /* x86-64 should always use SA_RESTORER. */ + if (ka->sa.sa_flags & SA_RESTORER) { + err |= __put_user(ka->sa.sa_restorer, &frame->pretcode); + } else { + /* could use a vstub here */ + return -EFAULT; + } + + if (err) + return -EFAULT; + + /* Set up registers for signal handler */ + regs->di = sig; + /* In case the signal handler was declared without prototypes */ + regs->ax = 0; + + /* This also works for non SA_SIGINFO handlers because they expect the + next argument after the signal number on the stack. */ + regs->si = (unsigned long)&frame->info; + regs->dx = (unsigned long)&frame->uc; + regs->ip = (unsigned long) ka->sa.sa_handler; + + regs->sp = (unsigned long)frame; + + /* Set up the CS register to run signal handlers in 64-bit mode, + even if the handler happens to be interrupting 32-bit code. */ + regs->cs = __USER_CS; + + return 0; +} +#endif /* CONFIG_X86_32 */ + +#ifdef CONFIG_X86_32 +/* + * Atomically swap in the new signal mask, and wait for a signal. + */ +asmlinkage int +sys_sigsuspend(int history0, int history1, old_sigset_t mask) +{ + mask &= _BLOCKABLE; + spin_lock_irq(¤t->sighand->siglock); + current->saved_sigmask = current->blocked; + siginitset(¤t->blocked, mask); + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + current->state = TASK_INTERRUPTIBLE; + schedule(); + set_restore_sigmask(); + + return -ERESTARTNOHAND; +} + +asmlinkage int +sys_sigaction(int sig, const struct old_sigaction __user *act, + struct old_sigaction __user *oact) +{ + struct k_sigaction new_ka, old_ka; + int ret; + + if (act) { + old_sigset_t mask; + + if (!access_ok(VERIFY_READ, act, sizeof(*act)) || + __get_user(new_ka.sa.sa_handler, &act->sa_handler) || + __get_user(new_ka.sa.sa_restorer, &act->sa_restorer)) + return -EFAULT; + + __get_user(new_ka.sa.sa_flags, &act->sa_flags); + __get_user(mask, &act->sa_mask); + siginitset(&new_ka.sa.sa_mask, mask); + } + + ret = do_sigaction(sig, act ? &new_ka : NULL, oact ? &old_ka : NULL); + + if (!ret && oact) { + if (!access_ok(VERIFY_WRITE, oact, sizeof(*oact)) || + __put_user(old_ka.sa.sa_handler, &oact->sa_handler) || + __put_user(old_ka.sa.sa_restorer, &oact->sa_restorer)) + return -EFAULT; + + __put_user(old_ka.sa.sa_flags, &oact->sa_flags); + __put_user(old_ka.sa.sa_mask.sig[0], &oact->sa_mask); + } + + return ret; +} +#endif /* CONFIG_X86_32 */ + +#ifdef CONFIG_X86_32 +asmlinkage int sys_sigaltstack(unsigned long bx) +{ + /* + * This is needed to make gcc realize it doesn't own the + * "struct pt_regs" + */ + struct pt_regs *regs = (struct pt_regs *)&bx; + const stack_t __user *uss = (const stack_t __user *)bx; + stack_t __user *uoss = (stack_t __user *)regs->cx; + + return do_sigaltstack(uss, uoss, regs->sp); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long +sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, + struct pt_regs *regs) +{ + return do_sigaltstack(uss, uoss, regs->sp); +} +#endif /* CONFIG_X86_32 */ + +/* + * Do a signal return; undo the signal stack. + */ +#ifdef CONFIG_X86_32 +asmlinkage unsigned long sys_sigreturn(unsigned long __unused) +{ + struct sigframe __user *frame; + struct pt_regs *regs; + unsigned long ax; + sigset_t set; + + regs = (struct pt_regs *) &__unused; + frame = (struct sigframe __user *)(regs->sp - 8); + + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1 + && __copy_from_user(&set.sig[1], &frame->extramask, + sizeof(frame->extramask)))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->sc, &ax)) + goto badframe; + return ax; + +badframe: + signal_fault(regs, frame, "sigreturn"); + + return 0; +} +#endif /* CONFIG_X86_32 */ + +static long do_rt_sigreturn(struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + unsigned long ax; + sigset_t set; + + frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) + goto badframe; + + if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) + goto badframe; + + return ax; + +badframe: + signal_fault(regs, frame, "rt_sigreturn"); + return 0; +} + +#ifdef CONFIG_X86_32 +asmlinkage int sys_rt_sigreturn(struct pt_regs regs) +{ + return do_rt_sigreturn(®s); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) +{ + return do_rt_sigreturn(regs); +} +#endif /* CONFIG_X86_32 */ /* * OK, we're invoking a handler: */ static int signr_convert(int sig) { +#ifdef CONFIG_X86_32 struct thread_info *info = current_thread_info(); if (info->exec_domain && info->exec_domain->signal_invmap && sig < 32) return info->exec_domain->signal_invmap[sig]; +#endif /* CONFIG_X86_32 */ return sig; } +#ifdef CONFIG_X86_32 + #define is_ia32 1 #define ia32_setup_frame __setup_frame #define ia32_setup_rt_frame __setup_rt_frame +#else /* !CONFIG_X86_32 */ + +#ifdef CONFIG_IA32_EMULATION +#define is_ia32 test_thread_flag(TIF_IA32) +#else /* !CONFIG_IA32_EMULATION */ +#define is_ia32 0 +#endif /* CONFIG_IA32_EMULATION */ + +int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs *regs); +int ia32_setup_frame(int sig, struct k_sigaction *ka, + sigset_t *set, struct pt_regs *regs); + +#endif /* CONFIG_X86_32 */ + static int setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, sigset_t *set, struct pt_regs *regs) @@ -592,7 +778,13 @@ handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, return 0; } +#ifdef CONFIG_X86_32 #define NR_restart_syscall __NR_restart_syscall +#else /* !CONFIG_X86_32 */ +#define NR_restart_syscall \ + test_thread_flag(TIF_IA32) ? __NR_ia32_restart_syscall : __NR_restart_syscall +#endif /* CONFIG_X86_32 */ + /* * Note that 'init' is a special process: it doesn't get signals it doesn't * want to handle. Thus you cannot kill init even with a SIGKILL even by @@ -704,8 +896,9 @@ void signal_fault(struct pt_regs *regs, void __user *frame, char *where) struct task_struct *me = current; if (show_unhandled_signals && printk_ratelimit()) { - printk(KERN_INFO + printk("%s" "%s[%d] bad frame in %s frame:%p ip:%lx sp:%lx orax:%lx", + task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, me->comm, me->pid, where, frame, regs->ip, regs->sp, regs->orig_ax); print_vma_addr(" in ", regs->ip); diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c deleted file mode 100644 index a5c9627f4db..00000000000 --- a/arch/x86/kernel/signal_64.c +++ /dev/null @@ -1,516 +0,0 @@ -/* - * Copyright (C) 1991, 1992 Linus Torvalds - * Copyright (C) 2000, 2001, 2002 Andi Kleen SuSE Labs - * - * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson - * 2000-06-20 Pentium III FXSR, SSE support by Gareth Hughes - * 2000-2002 x86-64 support by Andi Kleen - */ - -#include <linux/sched.h> -#include <linux/mm.h> -#include <linux/smp.h> -#include <linux/kernel.h> -#include <linux/signal.h> -#include <linux/errno.h> -#include <linux/wait.h> -#include <linux/ptrace.h> -#include <linux/tracehook.h> -#include <linux/unistd.h> -#include <linux/stddef.h> -#include <linux/personality.h> -#include <linux/compiler.h> -#include <linux/uaccess.h> - -#include <asm/processor.h> -#include <asm/ucontext.h> -#include <asm/i387.h> -#include <asm/proto.h> -#include <asm/ia32_unistd.h> -#include <asm/mce.h> -#include <asm/syscall.h> -#include <asm/syscalls.h> -#include "sigframe.h" - -#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) - -#define __FIX_EFLAGS (X86_EFLAGS_AC | X86_EFLAGS_OF | \ - X86_EFLAGS_DF | X86_EFLAGS_TF | X86_EFLAGS_SF | \ - X86_EFLAGS_ZF | X86_EFLAGS_AF | X86_EFLAGS_PF | \ - X86_EFLAGS_CF) - -#ifdef CONFIG_X86_32 -# define FIX_EFLAGS (__FIX_EFLAGS | X86_EFLAGS_RF) -#else -# define FIX_EFLAGS __FIX_EFLAGS -#endif - -asmlinkage long -sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, - struct pt_regs *regs) -{ - return do_sigaltstack(uss, uoss, regs->sp); -} - -#define COPY(x) { \ - err |= __get_user(regs->x, &sc->x); \ -} - -#define COPY_SEG_STRICT(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - regs->seg = tmp | 3; \ -} - -/* - * Do a signal return; undo the signal stack. - */ -static int -restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, - unsigned long *pax) -{ - void __user *buf; - unsigned int tmpflags; - unsigned int err = 0; - - /* Always make any pending restarted system calls return -EINTR */ - current_thread_info()->restart_block.fn = do_no_restart_syscall; - - COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); - COPY(dx); COPY(cx); COPY(ip); - COPY(r8); - COPY(r9); - COPY(r10); - COPY(r11); - COPY(r12); - COPY(r13); - COPY(r14); - COPY(r15); - - /* Kernel saves and restores only the CS segment register on signals, - * which is the bare minimum needed to allow mixed 32/64-bit code. - * App's signal handler can save/restore other segments if needed. */ - COPY_SEG_STRICT(cs); - - err |= __get_user(tmpflags, &sc->flags); - regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); - regs->orig_ax = -1; /* disable syscall checks */ - - err |= __get_user(buf, &sc->fpstate); - err |= restore_i387_xstate(buf); - - err |= __get_user(*pax, &sc->ax); - return err; -} - -static long do_rt_sigreturn(struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - unsigned long ax; - sigset_t set; - - frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) - goto badframe; - - if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) - goto badframe; - - return ax; - -badframe: - signal_fault(regs, frame, "rt_sigreturn"); - return 0; -} - -asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) -{ - return do_rt_sigreturn(regs); -} - -/* - * Set up a signal frame. - */ - -static inline int -setup_sigcontext(struct sigcontext __user *sc, struct pt_regs *regs, - unsigned long mask, struct task_struct *me) -{ - int err = 0; - - err |= __put_user(regs->cs, &sc->cs); - err |= __put_user(0, &sc->gs); - err |= __put_user(0, &sc->fs); - - err |= __put_user(regs->di, &sc->di); - err |= __put_user(regs->si, &sc->si); - err |= __put_user(regs->bp, &sc->bp); - err |= __put_user(regs->sp, &sc->sp); - err |= __put_user(regs->bx, &sc->bx); - err |= __put_user(regs->dx, &sc->dx); - err |= __put_user(regs->cx, &sc->cx); - err |= __put_user(regs->ax, &sc->ax); - err |= __put_user(regs->r8, &sc->r8); - err |= __put_user(regs->r9, &sc->r9); - err |= __put_user(regs->r10, &sc->r10); - err |= __put_user(regs->r11, &sc->r11); - err |= __put_user(regs->r12, &sc->r12); - err |= __put_user(regs->r13, &sc->r13); - err |= __put_user(regs->r14, &sc->r14); - err |= __put_user(regs->r15, &sc->r15); - err |= __put_user(me->thread.trap_no, &sc->trapno); - err |= __put_user(me->thread.error_code, &sc->err); - err |= __put_user(regs->ip, &sc->ip); - err |= __put_user(regs->flags, &sc->flags); - err |= __put_user(mask, &sc->oldmask); - err |= __put_user(me->thread.cr2, &sc->cr2); - - return err; -} - -/* - * Determine which stack to use.. - */ - -static void __user * -get_stack(struct k_sigaction *ka, struct pt_regs *regs, unsigned long size) -{ - unsigned long sp; - - /* Default to using normal stack - redzone*/ - sp = regs->sp - 128; - - /* This is the X/Open sanctioned signal stack switching. */ - if (ka->sa.sa_flags & SA_ONSTACK) { - if (sas_ss_flags(sp) == 0) - sp = current->sas_ss_sp + current->sas_ss_size; - } - - return (void __user *)round_down(sp - size, 64); -} - -static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - void __user *fp = NULL; - int err = 0; - struct task_struct *me = current; - - if (used_math()) { - fp = get_stack(ka, regs, sig_xstate_size); - frame = (void __user *)round_down( - (unsigned long)fp - sizeof(struct rt_sigframe), 16) - 8; - - if (save_i387_xstate(fp) < 0) - return -EFAULT; - } else - frame = get_stack(ka, regs, sizeof(struct rt_sigframe)) - 8; - - if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) - return -EFAULT; - - if (ka->sa.sa_flags & SA_SIGINFO) { - if (copy_siginfo_to_user(&frame->info, info)) - return -EFAULT; - } - - /* Create the ucontext. */ - if (cpu_has_xsave) - err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); - else - err |= __put_user(0, &frame->uc.uc_flags); - err |= __put_user(0, &frame->uc.uc_link); - err |= __put_user(me->sas_ss_sp, &frame->uc.uc_stack.ss_sp); - err |= __put_user(sas_ss_flags(regs->sp), - &frame->uc.uc_stack.ss_flags); - err |= __put_user(me->sas_ss_size, &frame->uc.uc_stack.ss_size); - err |= setup_sigcontext(&frame->uc.uc_mcontext, regs, set->sig[0], me); - err |= __put_user(fp, &frame->uc.uc_mcontext.fpstate); - if (sizeof(*set) == 16) { - __put_user(set->sig[0], &frame->uc.uc_sigmask.sig[0]); - __put_user(set->sig[1], &frame->uc.uc_sigmask.sig[1]); - } else - err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); - - /* Set up to return from userspace. If provided, use a stub - already in userspace. */ - /* x86-64 should always use SA_RESTORER. */ - if (ka->sa.sa_flags & SA_RESTORER) { - err |= __put_user(ka->sa.sa_restorer, &frame->pretcode); - } else { - /* could use a vstub here */ - return -EFAULT; - } - - if (err) - return -EFAULT; - - /* Set up registers for signal handler */ - regs->di = sig; - /* In case the signal handler was declared without prototypes */ - regs->ax = 0; - - /* This also works for non SA_SIGINFO handlers because they expect the - next argument after the signal number on the stack. */ - regs->si = (unsigned long)&frame->info; - regs->dx = (unsigned long)&frame->uc; - regs->ip = (unsigned long) ka->sa.sa_handler; - - regs->sp = (unsigned long)frame; - - /* Set up the CS register to run signal handlers in 64-bit mode, - even if the handler happens to be interrupting 32-bit code. */ - regs->cs = __USER_CS; - - return 0; -} - -/* - * OK, we're invoking a handler - */ -static int signr_convert(int sig) -{ - return sig; -} - -#ifdef CONFIG_IA32_EMULATION -#define is_ia32 test_thread_flag(TIF_IA32) -#else -#define is_ia32 0 -#endif - -static int -setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs) -{ - int usig = signr_convert(sig); - int ret; - - /* Set up the stack frame */ - if (is_ia32) { - if (ka->sa.sa_flags & SA_SIGINFO) - ret = ia32_setup_rt_frame(usig, ka, info, set, regs); - else - ret = ia32_setup_frame(usig, ka, set, regs); - } else - ret = __setup_rt_frame(sig, ka, info, set, regs); - - if (ret) { - force_sigsegv(sig, current); - return -EFAULT; - } - - return ret; -} - -static int -handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, - sigset_t *oldset, struct pt_regs *regs) -{ - int ret; - - /* Are we from a system call? */ - if (syscall_get_nr(current, regs) >= 0) { - /* If so, check system call restarting.. */ - switch (syscall_get_error(current, regs)) { - case -ERESTART_RESTARTBLOCK: - case -ERESTARTNOHAND: - regs->ax = -EINTR; - break; - - case -ERESTARTSYS: - if (!(ka->sa.sa_flags & SA_RESTART)) { - regs->ax = -EINTR; - break; - } - /* fallthrough */ - case -ERESTARTNOINTR: - regs->ax = regs->orig_ax; - regs->ip -= 2; - break; - } - } - - /* - * If TF is set due to a debugger (TIF_FORCED_TF), clear the TF - * flag so that register information in the sigcontext is correct. - */ - if (unlikely(regs->flags & X86_EFLAGS_TF) && - likely(test_and_clear_thread_flag(TIF_FORCED_TF))) - regs->flags &= ~X86_EFLAGS_TF; - - ret = setup_rt_frame(sig, ka, info, oldset, regs); - - if (ret) - return ret; - -#ifdef CONFIG_X86_64 - /* - * This has nothing to do with segment registers, - * despite the name. This magic affects uaccess.h - * macros' behavior. Reset it to the normal setting. - */ - set_fs(USER_DS); -#endif - - /* - * Clear the direction flag as per the ABI for function entry. - */ - regs->flags &= ~X86_EFLAGS_DF; - - /* - * Clear TF when entering the signal handler, but - * notify any tracer that was single-stepping it. - * The tracer may want to single-step inside the - * handler too. - */ - regs->flags &= ~X86_EFLAGS_TF; - - spin_lock_irq(¤t->sighand->siglock); - sigorsets(¤t->blocked, ¤t->blocked, &ka->sa.sa_mask); - if (!(ka->sa.sa_flags & SA_NODEFER)) - sigaddset(¤t->blocked, sig); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - tracehook_signal_handler(sig, info, ka, regs, - test_thread_flag(TIF_SINGLESTEP)); - - return 0; -} - -#define NR_restart_syscall \ - test_thread_flag(TIF_IA32) ? __NR_ia32_restart_syscall : __NR_restart_syscall -/* - * Note that 'init' is a special process: it doesn't get signals it doesn't - * want to handle. Thus you cannot kill init even with a SIGKILL even by - * mistake. - */ -static void do_signal(struct pt_regs *regs) -{ - struct k_sigaction ka; - siginfo_t info; - int signr; - sigset_t *oldset; - - /* - * We want the common case to go fast, which is why we may in certain - * cases get here from kernel mode. Just return without doing anything - * if so. - * X86_32: vm86 regs switched out by assembly code before reaching - * here, so testing against kernel CS suffices. - */ - if (!user_mode(regs)) - return; - - if (current_thread_info()->status & TS_RESTORE_SIGMASK) - oldset = ¤t->saved_sigmask; - else - oldset = ¤t->blocked; - - signr = get_signal_to_deliver(&info, &ka, regs, NULL); - if (signr > 0) { - /* - * Re-enable any watchpoints before delivering the - * signal to user space. The processor register will - * have been cleared if the watchpoint triggered - * inside the kernel. - */ - if (current->thread.debugreg7) - set_debugreg(current->thread.debugreg7, 7); - - /* Whee! Actually deliver the signal. */ - if (handle_signal(signr, &info, &ka, oldset, regs) == 0) { - /* - * A signal was successfully delivered; the saved - * sigmask will have been stored in the signal frame, - * and will be restored by sigreturn, so we can simply - * clear the TS_RESTORE_SIGMASK flag. - */ - current_thread_info()->status &= ~TS_RESTORE_SIGMASK; - } - return; - } - - /* Did we come from a system call? */ - if (syscall_get_nr(current, regs) >= 0) { - /* Restart the system call - no handlers present */ - switch (syscall_get_error(current, regs)) { - case -ERESTARTNOHAND: - case -ERESTARTSYS: - case -ERESTARTNOINTR: - regs->ax = regs->orig_ax; - regs->ip -= 2; - break; - - case -ERESTART_RESTARTBLOCK: - regs->ax = NR_restart_syscall; - regs->ip -= 2; - break; - } - } - - /* - * If there's no signal to deliver, we just put the saved sigmask - * back. - */ - if (current_thread_info()->status & TS_RESTORE_SIGMASK) { - current_thread_info()->status &= ~TS_RESTORE_SIGMASK; - sigprocmask(SIG_SETMASK, ¤t->saved_sigmask, NULL); - } -} - -/* - * notification of userspace execution resumption - * - triggered by the TIF_WORK_MASK flags - */ -void -do_notify_resume(struct pt_regs *regs, void *unused, __u32 thread_info_flags) -{ -#if defined(CONFIG_X86_64) && defined(CONFIG_X86_MCE) - /* notify userspace of pending MCEs */ - if (thread_info_flags & _TIF_MCE_NOTIFY) - mce_notify_user(); -#endif /* CONFIG_X86_64 && CONFIG_X86_MCE */ - - /* deal with pending signal delivery */ - if (thread_info_flags & _TIF_SIGPENDING) - do_signal(regs); - - if (thread_info_flags & _TIF_NOTIFY_RESUME) { - clear_thread_flag(TIF_NOTIFY_RESUME); - tracehook_notify_resume(regs); - } - -#ifdef CONFIG_X86_32 - clear_thread_flag(TIF_IRET); -#endif /* CONFIG_X86_32 */ -} - -void signal_fault(struct pt_regs *regs, void __user *frame, char *where) -{ - struct task_struct *me = current; - - if (show_unhandled_signals && printk_ratelimit()) { - printk(KERN_INFO - "%s[%d] bad frame in %s frame:%p ip:%lx sp:%lx orax:%lx", - me->comm, me->pid, where, frame, - regs->ip, regs->sp, regs->orig_ax); - print_vma_addr(" in ", regs->ip); - printk(KERN_CONT "\n"); - } - - force_sig(SIGSEGV, me); -} diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c index 18f9b19f5f8..beea2649a24 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. */ @@ -178,11 +165,7 @@ static void native_smp_send_stop(void) void smp_reschedule_interrupt(struct pt_regs *regs) { ack_APIC_irq(); -#ifdef CONFIG_X86_32 - __get_cpu_var(irq_stat).irq_resched_count++; -#else - add_pda(irq_resched_count, 1); -#endif + inc_irq_stat(irq_resched_count); } void smp_call_function_interrupt(struct pt_regs *regs) @@ -190,11 +173,7 @@ void smp_call_function_interrupt(struct pt_regs *regs) ack_APIC_irq(); irq_enter(); generic_smp_call_function_interrupt(); -#ifdef CONFIG_X86_32 - __get_cpu_var(irq_stat).irq_call_count++; -#else - add_pda(irq_call_count, 1); -#endif + inc_irq_stat(irq_call_count); irq_exit(); } @@ -203,11 +182,7 @@ void smp_call_function_single_interrupt(struct pt_regs *regs) ack_APIC_irq(); irq_enter(); generic_smp_call_function_single_interrupt(); -#ifdef CONFIG_X86_32 - __get_cpu_var(irq_stat).irq_call_count++; -#else - add_pda(irq_call_count, 1); -#endif + inc_irq_stat(irq_call_count); irq_exit(); } diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 7b109339731..6bd4d9b7387 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); @@ -287,16 +282,14 @@ static int __cpuinitdata unsafe_smp; /* * Activate a secondary processor. */ -static void __cpuinit start_secondary(void *unused) +notrace static void __cpuinit start_secondary(void *unused) { /* * Don't put *anything* before cpu_init(), SMP booting is too * 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(); @@ -503,7 +496,7 @@ void __cpuinit set_cpu_sibling_map(int cpu) } /* maps the cpu to the sched domain representing multi-core */ -cpumask_t cpu_coregroup_map(int cpu) +const struct cpumask *cpu_coregroup_mask(int cpu) { struct cpuinfo_x86 *c = &cpu_data(cpu); /* @@ -511,9 +504,14 @@ cpumask_t cpu_coregroup_map(int cpu) * And for power savings, we return cpu_core_map */ if (sched_mc_power_savings || sched_smt_power_savings) - return per_cpu(cpu_core_map, cpu); + return &per_cpu(cpu_core_map, cpu); else - return c->llc_shared_map; + return &c->llc_shared_map; +} + +cpumask_t cpu_coregroup_map(int cpu) +{ + return *cpu_coregroup_mask(cpu); } static void impress_friends(void) @@ -536,7 +534,7 @@ 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 +573,13 @@ 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 +596,7 @@ 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 +611,9 @@ 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 +732,6 @@ 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; @@ -1086,8 +1080,10 @@ static int __init smp_sanity_check(unsigned max_cpus) #endif if (!physid_isset(hard_smp_processor_id(), phys_cpu_present_map)) { - printk(KERN_WARNING "weird, boot CPU (#%d) not listed" - "by the BIOS.\n", hard_smp_processor_id()); + printk(KERN_WARNING + "weird, boot CPU (#%d) not listed by the BIOS.\n", + hard_smp_processor_id()); + physid_set(hard_smp_processor_id(), phys_cpu_present_map); } @@ -1158,7 +1154,7 @@ static void __init smp_cpu_index_default(void) for_each_possible_cpu(i) { c = &cpu_data(i); /* mark all to hotplug */ - c->cpu_index = NR_CPUS; + c->cpu_index = nr_cpu_ids; } } @@ -1263,6 +1259,15 @@ void __init native_smp_cpus_done(unsigned int max_cpus) check_nmi_watchdog(); } +static int __initdata setup_possible_cpus = -1; +static int __init _setup_possible_cpus(char *str) +{ + get_option(&str, &setup_possible_cpus); + return 0; +} +early_param("possible_cpus", _setup_possible_cpus); + + /* * cpu_possible_map should be static, it cannot change as cpu's * are onlined, or offlined. The reason is per-cpu data-structures @@ -1275,7 +1280,7 @@ void __init native_smp_cpus_done(unsigned int max_cpus) * * Three ways to find out the number of additional hotplug CPUs: * - If the BIOS specified disabled CPUs in ACPI/mptables use that. - * - The user can overwrite it with additional_cpus=NUM + * - The user can overwrite it with possible_cpus=NUM * - Otherwise don't reserve additional CPUs. * We do this because additional CPUs waste a lot of memory. * -AK @@ -1288,9 +1293,19 @@ __init void prefill_possible_map(void) if (!num_processors) num_processors = 1; - possible = num_processors + disabled_cpus; - if (possible > NR_CPUS) - possible = NR_CPUS; + if (setup_possible_cpus == -1) + possible = num_processors + disabled_cpus; + else + possible = setup_possible_cpus; + + total_cpus = max_t(int, possible, num_processors + disabled_cpus); + + if (possible > CONFIG_NR_CPUS) { + printk(KERN_WARNING + "%d Processors exceeds NR_CPUS limit of %d\n", + possible, CONFIG_NR_CPUS); + possible = CONFIG_NR_CPUS; + } printk(KERN_INFO "SMP: Allowing %d CPUs, %d hotplug CPUs\n", possible, max_t(int, possible - num_processors, 0)); @@ -1355,7 +1370,7 @@ 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/time_32.c b/arch/x86/kernel/time_32.c index 77b400f06ea..65309e4cb1c 100644 --- a/arch/x86/kernel/time_32.c +++ b/arch/x86/kernel/time_32.c @@ -75,7 +75,7 @@ EXPORT_SYMBOL(profile_pc); irqreturn_t timer_interrupt(int irq, void *dev_id) { /* Keep nmi watchdog up to date */ - per_cpu(irq_stat, smp_processor_id()).irq0_irqs++; + inc_irq_stat(irq0_irqs); #ifdef CONFIG_X86_IO_APIC if (timer_ack) { diff --git a/arch/x86/kernel/time_64.c b/arch/x86/kernel/time_64.c index cb19d650c21..891e7a7c433 100644 --- a/arch/x86/kernel/time_64.c +++ b/arch/x86/kernel/time_64.c @@ -49,9 +49,9 @@ unsigned long profile_pc(struct pt_regs *regs) } EXPORT_SYMBOL(profile_pc); -irqreturn_t timer_interrupt(int irq, void *dev_id) +static irqreturn_t timer_interrupt(int irq, void *dev_id) { - add_pda(irq0_irqs, 1); + inc_irq_stat(irq0_irqs); global_clock_event->event_handler(global_clock_event); @@ -80,6 +80,8 @@ unsigned long __init calibrate_cpu(void) break; no_ctr_free = (i == 4); if (no_ctr_free) { + WARN(1, KERN_WARNING "Warning: AMD perfctrs busy ... " + "cpu_khz value may be incorrect.\n"); i = 3; rdmsrl(MSR_K7_EVNTSEL3, evntsel3); wrmsrl(MSR_K7_EVNTSEL3, 0); diff --git a/arch/x86/kernel/tlb_32.c b/arch/x86/kernel/tlb_32.c index f4049f3513b..ce505464224 100644 --- a/arch/x86/kernel/tlb_32.c +++ b/arch/x86/kernel/tlb_32.c @@ -34,9 +34,8 @@ static DEFINE_SPINLOCK(tlbstate_lock); */ void leave_mm(int cpu) { - if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK) - BUG(); - cpu_clear(cpu, per_cpu(cpu_tlbstate, cpu).active_mm->cpu_vm_mask); + BUG_ON(x86_read_percpu(cpu_tlbstate.state) == TLBSTATE_OK); + cpu_clear(cpu, x86_read_percpu(cpu_tlbstate.active_mm)->cpu_vm_mask); load_cr3(swapper_pg_dir); } EXPORT_SYMBOL_GPL(leave_mm); @@ -104,8 +103,8 @@ void smp_invalidate_interrupt(struct pt_regs *regs) * BUG(); */ - if (flush_mm == per_cpu(cpu_tlbstate, cpu).active_mm) { - if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK) { + if (flush_mm == x86_read_percpu(cpu_tlbstate.active_mm)) { + if (x86_read_percpu(cpu_tlbstate.state) == TLBSTATE_OK) { if (flush_va == TLB_FLUSH_ALL) local_flush_tlb(); else @@ -119,7 +118,7 @@ void smp_invalidate_interrupt(struct pt_regs *regs) smp_mb__after_clear_bit(); out: put_cpu_no_resched(); - __get_cpu_var(irq_stat).irq_tlb_count++; + inc_irq_stat(irq_tlb_count); } void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm, @@ -164,7 +163,7 @@ 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 */ @@ -238,7 +237,7 @@ static void do_flush_tlb_all(void *info) unsigned long cpu = smp_processor_id(); __flush_tlb_all(); - if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_LAZY) + if (x86_read_percpu(cpu_tlbstate.state) == TLBSTATE_LAZY) leave_mm(cpu); } diff --git a/arch/x86/kernel/tlb_64.c b/arch/x86/kernel/tlb_64.c index 8f919ca6949..f8be6f1d2e4 100644 --- a/arch/x86/kernel/tlb_64.c +++ b/arch/x86/kernel/tlb_64.c @@ -154,7 +154,7 @@ asmlinkage void smp_invalidate_interrupt(struct pt_regs *regs) out: ack_APIC_irq(); cpu_clear(cpu, f->flush_cpumask); - add_pda(irq_tlb_count, 1); + inc_irq_stat(irq_tlb_count); } void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm, @@ -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/tlb_uv.c b/arch/x86/kernel/tlb_uv.c index 04431f34fd1..f885023167e 100644 --- a/arch/x86/kernel/tlb_uv.c +++ b/arch/x86/kernel/tlb_uv.c @@ -566,14 +566,10 @@ static int __init uv_ptc_init(void) if (!is_uv_system()) return 0; - if (!proc_mkdir("sgi_uv", NULL)) - return -EINVAL; - proc_uv_ptc = create_proc_entry(UV_PTC_BASENAME, 0444, NULL); if (!proc_uv_ptc) { printk(KERN_ERR "unable to create %s proc entry\n", UV_PTC_BASENAME); - remove_proc_entry("sgi_uv", NULL); return -EINVAL; } proc_uv_ptc->proc_fops = &proc_uv_ptc_operations; @@ -586,7 +582,6 @@ static int __init uv_ptc_init(void) static struct bau_control * __init uv_table_bases_init(int blade, int node) { int i; - int *ip; struct bau_msg_status *msp; struct bau_control *bau_tabp; @@ -603,13 +598,6 @@ static struct bau_control * __init uv_table_bases_init(int blade, int node) bau_cpubits_clear(&msp->seen_by, (int) uv_blade_nr_possible_cpus(blade)); - bau_tabp->watching = - kmalloc_node(sizeof(int) * DEST_NUM_RESOURCES, GFP_KERNEL, node); - BUG_ON(!bau_tabp->watching); - - for (i = 0, ip = bau_tabp->watching; i < DEST_Q_SIZE; i++, ip++) - *ip = 0; - uv_bau_table_bases[blade] = bau_tabp; return bau_tabp; @@ -632,7 +620,6 @@ uv_table_bases_finish(int blade, int node, int cur_cpu, bcp->bau_msg_head = bau_tablesp->va_queue_first; bcp->va_queue_first = bau_tablesp->va_queue_first; bcp->va_queue_last = bau_tablesp->va_queue_last; - bcp->watching = bau_tablesp->watching; bcp->msg_statuses = bau_tablesp->msg_statuses; bcp->descriptor_base = adp; } diff --git a/arch/x86/kernel/trampoline.c b/arch/x86/kernel/trampoline.c index 1106fac6024..808031a5ba1 100644 --- a/arch/x86/kernel/trampoline.c +++ b/arch/x86/kernel/trampoline.c @@ -1,10 +1,26 @@ #include <linux/io.h> #include <asm/trampoline.h> +#include <asm/e820.h> /* ready for x86_64 and x86 */ unsigned char *trampoline_base = __va(TRAMPOLINE_BASE); +void __init reserve_trampoline_memory(void) +{ +#ifdef CONFIG_X86_32 + /* + * But first pinch a few for the stack/trampoline stuff + * FIXME: Don't need the extra page at 4K, but need to fix + * trampoline before removing it. (see the GDT stuff) + */ + reserve_early(PAGE_SIZE, PAGE_SIZE + PAGE_SIZE, "EX TRAMPOLINE"); +#endif + /* Has to be in very low memory so we can execute real-mode AP code. */ + reserve_early(TRAMPOLINE_BASE, TRAMPOLINE_BASE + TRAMPOLINE_SIZE, + "TRAMPOLINE"); +} + /* * Currently trivial. Write the real->protected mode * bootstrap into the page concerned. The caller @@ -12,7 +28,6 @@ unsigned char *trampoline_base = __va(TRAMPOLINE_BASE); */ unsigned long setup_trampoline(void) { - memcpy(trampoline_base, trampoline_data, - trampoline_end - trampoline_data); + memcpy(trampoline_base, trampoline_data, TRAMPOLINE_SIZE); return virt_to_phys(trampoline_base); } diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 04d242ab016..ce6650eb64e 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -72,9 +72,6 @@ #include "cpu/mcheck/mce.h" -DECLARE_BITMAP(used_vectors, NR_VECTORS); -EXPORT_SYMBOL_GPL(used_vectors); - asmlinkage int system_call(void); /* Do we ignore FPU interrupts ? */ @@ -89,6 +86,9 @@ gate_desc idt_table[256] __attribute__((__section__(".data.idt"))) = { { { { 0, 0 } } }, }; #endif +DECLARE_BITMAP(used_vectors, NR_VECTORS); +EXPORT_SYMBOL_GPL(used_vectors); + static int ignore_nmis; static inline void conditional_sti(struct pt_regs *regs) @@ -292,8 +292,10 @@ dotraplinkage void do_double_fault(struct pt_regs *regs, long error_code) tsk->thread.error_code = error_code; tsk->thread.trap_no = 8; - /* This is always a kernel trap and never fixable (and thus must - never return). */ + /* + * This is always a kernel trap and never fixable (and thus must + * never return). + */ for (;;) die(str, regs, error_code); } @@ -481,11 +483,7 @@ do_nmi(struct pt_regs *regs, long error_code) { nmi_enter(); -#ifdef CONFIG_X86_32 - { int cpu; cpu = smp_processor_id(); ++nmi_count(cpu); } -#else - add_pda(__nmi_count, 1); -#endif + inc_irq_stat(__nmi_count); if (!ignore_nmis) default_do_nmi(regs); @@ -524,9 +522,11 @@ dotraplinkage void __kprobes do_int3(struct pt_regs *regs, long error_code) } #ifdef CONFIG_X86_64 -/* Help handler running on IST stack to switch back to user stack - for scheduling or signal handling. The actual stack switch is done in - entry.S */ +/* + * Help handler running on IST stack to switch back to user stack + * for scheduling or signal handling. The actual stack switch is done in + * entry.S + */ asmlinkage __kprobes struct pt_regs *sync_regs(struct pt_regs *eregs) { struct pt_regs *regs = eregs; @@ -536,8 +536,10 @@ asmlinkage __kprobes struct pt_regs *sync_regs(struct pt_regs *eregs) /* Exception from user space */ else if (user_mode(eregs)) regs = task_pt_regs(current); - /* Exception from kernel and interrupts are enabled. Move to - kernel process stack. */ + /* + * Exception from kernel and interrupts are enabled. Move to + * kernel process stack. + */ else if (eregs->flags & X86_EFLAGS_IF) regs = (struct pt_regs *)(eregs->sp -= sizeof(struct pt_regs)); if (eregs != regs) @@ -664,7 +666,7 @@ void math_error(void __user *ip) { struct task_struct *task; siginfo_t info; - unsigned short cwd, swd; + unsigned short cwd, swd, err; /* * Save the info for the exception handler and clear the error. @@ -675,7 +677,6 @@ void math_error(void __user *ip) task->thread.error_code = 0; info.si_signo = SIGFPE; info.si_errno = 0; - info.si_code = __SI_FAULT; info.si_addr = ip; /* * (~cwd & swd) will mask out exceptions that are not set to unmasked @@ -689,34 +690,30 @@ void math_error(void __user *ip) */ cwd = get_fpu_cwd(task); swd = get_fpu_swd(task); - switch (swd & ~cwd & 0x3f) { - case 0x000: /* No unmasked exception */ -#ifdef CONFIG_X86_32 - return; -#endif - default: /* Multiple exceptions */ - break; - case 0x001: /* Invalid Op */ + + err = swd & ~cwd; + + if (err & 0x001) { /* Invalid op */ /* * swd & 0x240 == 0x040: Stack Underflow * swd & 0x240 == 0x240: Stack Overflow * User must clear the SF bit (0x40) if set */ info.si_code = FPE_FLTINV; - break; - case 0x002: /* Denormalize */ - case 0x010: /* Underflow */ - info.si_code = FPE_FLTUND; - break; - case 0x004: /* Zero Divide */ + } else if (err & 0x004) { /* Divide by Zero */ info.si_code = FPE_FLTDIV; - break; - case 0x008: /* Overflow */ + } else if (err & 0x008) { /* Overflow */ info.si_code = FPE_FLTOVF; - break; - case 0x020: /* Precision */ + } else if (err & 0x012) { /* Denormal, Underflow */ + info.si_code = FPE_FLTUND; + } else if (err & 0x020) { /* Precision */ info.si_code = FPE_FLTRES; - break; + } else { + /* + * If we're using IRQ 13, or supposedly even some trap 16 + * implementations, it's possible we get a spurious trap... + */ + return; /* Spurious trap, no error */ } force_sig_info(SIGFPE, &info, task); } @@ -949,9 +946,7 @@ dotraplinkage void do_iret_error(struct pt_regs *regs, long error_code) void __init trap_init(void) { -#ifdef CONFIG_X86_32 int i; -#endif #ifdef CONFIG_EISA void __iomem *p = early_ioremap(0x0FFFD9, 4); @@ -1008,11 +1003,15 @@ void __init trap_init(void) } set_system_trap_gate(SYSCALL_VECTOR, &system_call); +#endif /* Reserve all the builtin and the syscall vector: */ for (i = 0; i < FIRST_EXTERNAL_VECTOR; i++) set_bit(i, used_vectors); +#ifdef CONFIG_X86_64 + set_bit(IA32_SYSCALL_VECTOR, used_vectors); +#else set_bit(SYSCALL_VECTOR, used_vectors); #endif /* diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index 424093b157d..599e5816863 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -15,6 +15,7 @@ #include <asm/vgtod.h> #include <asm/time.h> #include <asm/delay.h> +#include <asm/hypervisor.h> unsigned int cpu_khz; /* TSC clocks / usec, not used here */ EXPORT_SYMBOL(cpu_khz); @@ -31,6 +32,7 @@ static int tsc_unstable; erroneous rdtsc usage on !cpu_has_tsc processors */ static int tsc_disabled = -1; +static int tsc_clocksource_reliable; /* * Scheduler clock - returns current time in nanosec units. */ @@ -98,6 +100,15 @@ int __init notsc_setup(char *str) __setup("notsc", notsc_setup); +static int __init tsc_setup(char *str) +{ + if (!strcmp(str, "reliable")) + tsc_clocksource_reliable = 1; + return 1; +} + +__setup("tsc=", tsc_setup); + #define MAX_RETRIES 5 #define SMI_TRESHOLD 50000 @@ -352,9 +363,15 @@ unsigned long native_calibrate_tsc(void) { u64 tsc1, tsc2, delta, ref1, ref2; unsigned long tsc_pit_min = ULONG_MAX, tsc_ref_min = ULONG_MAX; - unsigned long flags, latch, ms, fast_calibrate; + unsigned long flags, latch, ms, fast_calibrate, tsc_khz; int hpet = is_hpet_enabled(), i, loopmin; + tsc_khz = get_hypervisor_tsc_freq(); + if (tsc_khz) { + printk(KERN_INFO "TSC: Frequency read from the hypervisor\n"); + return tsc_khz; + } + local_irq_save(flags); fast_calibrate = quick_pit_calibrate(); local_irq_restore(flags); @@ -731,24 +748,21 @@ static struct dmi_system_id __initdata bad_tsc_dmi_table[] = { {} }; -/* - * Geode_LX - the OLPC CPU has a possibly a very reliable TSC - */ +static void __init check_system_tsc_reliable(void) +{ #ifdef CONFIG_MGEODE_LX -/* RTSC counts during suspend */ + /* RTSC counts during suspend */ #define RTSC_SUSP 0x100 - -static void __init check_geode_tsc_reliable(void) -{ unsigned long res_low, res_high; rdmsr_safe(MSR_GEODE_BUSCONT_CONF0, &res_low, &res_high); + /* Geode_LX - the OLPC CPU has a possibly a very reliable TSC */ if (res_low & RTSC_SUSP) - clocksource_tsc.flags &= ~CLOCK_SOURCE_MUST_VERIFY; -} -#else -static inline void check_geode_tsc_reliable(void) { } + tsc_clocksource_reliable = 1; #endif + if (boot_cpu_has(X86_FEATURE_TSC_RELIABLE)) + tsc_clocksource_reliable = 1; +} /* * Make an educated guess if the TSC is trustworthy and synchronized @@ -783,6 +797,8 @@ static void __init init_tsc_clocksource(void) { clocksource_tsc.mult = clocksource_khz2mult(tsc_khz, clocksource_tsc.shift); + if (tsc_clocksource_reliable) + clocksource_tsc.flags &= ~CLOCK_SOURCE_MUST_VERIFY; /* lower the rating if we already know its unstable: */ if (check_tsc_unstable()) { clocksource_tsc.rating = 0; @@ -843,7 +859,7 @@ void __init tsc_init(void) if (unsynchronized_tsc()) mark_tsc_unstable("TSCs unsynchronized"); - check_geode_tsc_reliable(); + check_system_tsc_reliable(); init_tsc_clocksource(); } diff --git a/arch/x86/kernel/tsc_sync.c b/arch/x86/kernel/tsc_sync.c index 1c0dfbca87c..bf36328f6ef 100644 --- a/arch/x86/kernel/tsc_sync.c +++ b/arch/x86/kernel/tsc_sync.c @@ -112,6 +112,12 @@ void __cpuinit check_tsc_sync_source(int cpu) if (unsynchronized_tsc()) return; + if (boot_cpu_has(X86_FEATURE_TSC_RELIABLE)) { + printk(KERN_INFO + "Skipping synchronization checks as TSC is reliable.\n"); + return; + } + printk(KERN_INFO "checking TSC synchronization [CPU#%d -> CPU#%d]:", smp_processor_id(), cpu); @@ -165,7 +171,7 @@ void __cpuinit check_tsc_sync_target(void) { int cpus = 2; - if (unsynchronized_tsc()) + if (unsynchronized_tsc() || boot_cpu_has(X86_FEATURE_TSC_RELIABLE)) return; /* diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c index 8b6c393ab9f..23206ba1687 100644 --- a/arch/x86/kernel/vmi_32.c +++ b/arch/x86/kernel/vmi_32.c @@ -266,109 +266,6 @@ static void vmi_nop(void) { } -#ifdef CONFIG_DEBUG_PAGE_TYPE - -#ifdef CONFIG_X86_PAE -#define MAX_BOOT_PTS (2048+4+1) -#else -#define MAX_BOOT_PTS (1024+1) -#endif - -/* - * During boot, mem_map is not yet available in paging_init, so stash - * all the boot page allocations here. - */ -static struct { - u32 pfn; - int type; -} boot_page_allocations[MAX_BOOT_PTS]; -static int num_boot_page_allocations; -static int boot_allocations_applied; - -void vmi_apply_boot_page_allocations(void) -{ - int i; - BUG_ON(!mem_map); - for (i = 0; i < num_boot_page_allocations; i++) { - struct page *page = pfn_to_page(boot_page_allocations[i].pfn); - page->type = boot_page_allocations[i].type; - page->type = boot_page_allocations[i].type & - ~(VMI_PAGE_ZEROED | VMI_PAGE_CLONE); - } - boot_allocations_applied = 1; -} - -static void record_page_type(u32 pfn, int type) -{ - BUG_ON(num_boot_page_allocations >= MAX_BOOT_PTS); - boot_page_allocations[num_boot_page_allocations].pfn = pfn; - boot_page_allocations[num_boot_page_allocations].type = type; - num_boot_page_allocations++; -} - -static void check_zeroed_page(u32 pfn, int type, struct page *page) -{ - u32 *ptr; - int i; - int limit = PAGE_SIZE / sizeof(int); - - if (page_address(page)) - ptr = (u32 *)page_address(page); - else - ptr = (u32 *)__va(pfn << PAGE_SHIFT); - /* - * When cloning the root in non-PAE mode, only the userspace - * pdes need to be zeroed. - */ - if (type & VMI_PAGE_CLONE) - limit = KERNEL_PGD_BOUNDARY; - for (i = 0; i < limit; i++) - BUG_ON(ptr[i]); -} - -/* - * We stash the page type into struct page so we can verify the page - * types are used properly. - */ -static void vmi_set_page_type(u32 pfn, int type) -{ - /* PAE can have multiple roots per page - don't track */ - if (PTRS_PER_PMD > 1 && (type & VMI_PAGE_PDP)) - return; - - if (boot_allocations_applied) { - struct page *page = pfn_to_page(pfn); - if (type != VMI_PAGE_NORMAL) - BUG_ON(page->type); - else - BUG_ON(page->type == VMI_PAGE_NORMAL); - page->type = type & ~(VMI_PAGE_ZEROED | VMI_PAGE_CLONE); - if (type & VMI_PAGE_ZEROED) - check_zeroed_page(pfn, type, page); - } else { - record_page_type(pfn, type); - } -} - -static void vmi_check_page_type(u32 pfn, int type) -{ - /* PAE can have multiple roots per page - skip checks */ - if (PTRS_PER_PMD > 1 && (type & VMI_PAGE_PDP)) - return; - - type &= ~(VMI_PAGE_ZEROED | VMI_PAGE_CLONE); - if (boot_allocations_applied) { - struct page *page = pfn_to_page(pfn); - BUG_ON((page->type ^ type) & VMI_PAGE_PAE); - BUG_ON(type == VMI_PAGE_NORMAL && page->type); - BUG_ON((type & page->type) == 0); - } -} -#else -#define vmi_set_page_type(p,t) do { } while (0) -#define vmi_check_page_type(p,t) do { } while (0) -#endif - #ifdef CONFIG_HIGHPTE static void *vmi_kmap_atomic_pte(struct page *page, enum km_type type) { @@ -395,7 +292,6 @@ static void *vmi_kmap_atomic_pte(struct page *page, enum km_type type) static void vmi_allocate_pte(struct mm_struct *mm, unsigned long pfn) { - vmi_set_page_type(pfn, VMI_PAGE_L1); vmi_ops.allocate_page(pfn, VMI_PAGE_L1, 0, 0, 0); } @@ -406,27 +302,22 @@ static void vmi_allocate_pmd(struct mm_struct *mm, unsigned long pfn) * It is called only for swapper_pg_dir, which already has * data on it. */ - vmi_set_page_type(pfn, VMI_PAGE_L2); vmi_ops.allocate_page(pfn, VMI_PAGE_L2, 0, 0, 0); } static void vmi_allocate_pmd_clone(unsigned long pfn, unsigned long clonepfn, unsigned long start, unsigned long count) { - vmi_set_page_type(pfn, VMI_PAGE_L2 | VMI_PAGE_CLONE); - vmi_check_page_type(clonepfn, VMI_PAGE_L2); vmi_ops.allocate_page(pfn, VMI_PAGE_L2 | VMI_PAGE_CLONE, clonepfn, start, count); } static void vmi_release_pte(unsigned long pfn) { vmi_ops.release_page(pfn, VMI_PAGE_L1); - vmi_set_page_type(pfn, VMI_PAGE_NORMAL); } static void vmi_release_pmd(unsigned long pfn) { vmi_ops.release_page(pfn, VMI_PAGE_L2); - vmi_set_page_type(pfn, VMI_PAGE_NORMAL); } /* @@ -450,26 +341,22 @@ static void vmi_release_pmd(unsigned long pfn) static void vmi_update_pte(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.update_pte(ptep, vmi_flags_addr(mm, addr, VMI_PAGE_PT, 0)); } static void vmi_update_pte_defer(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.update_pte(ptep, vmi_flags_addr_defer(mm, addr, VMI_PAGE_PT, 0)); } static void vmi_set_pte(pte_t *ptep, pte_t pte) { /* XXX because of set_pmd_pte, this can be called on PT or PD layers */ - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE | VMI_PAGE_PD); vmi_ops.set_pte(pte, ptep, VMI_PAGE_PT); } static void vmi_set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte) { - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.set_pte(pte, ptep, vmi_flags_addr(mm, addr, VMI_PAGE_PT, 0)); } @@ -477,10 +364,8 @@ static void vmi_set_pmd(pmd_t *pmdp, pmd_t pmdval) { #ifdef CONFIG_X86_PAE const pte_t pte = { .pte = pmdval.pmd }; - vmi_check_page_type(__pa(pmdp) >> PAGE_SHIFT, VMI_PAGE_PMD); #else const pte_t pte = { pmdval.pud.pgd.pgd }; - vmi_check_page_type(__pa(pmdp) >> PAGE_SHIFT, VMI_PAGE_PGD); #endif vmi_ops.set_pte(pte, (pte_t *)pmdp, VMI_PAGE_PD); } @@ -502,7 +387,6 @@ static void vmi_set_pte_atomic(pte_t *ptep, pte_t pteval) static void vmi_set_pte_present(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte) { - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.set_pte(pte, ptep, vmi_flags_addr_defer(mm, addr, VMI_PAGE_PT, 1)); } @@ -510,21 +394,18 @@ static void vmi_set_pud(pud_t *pudp, pud_t pudval) { /* Um, eww */ const pte_t pte = { .pte = pudval.pgd.pgd }; - vmi_check_page_type(__pa(pudp) >> PAGE_SHIFT, VMI_PAGE_PGD); vmi_ops.set_pte(pte, (pte_t *)pudp, VMI_PAGE_PDP); } static void vmi_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { const pte_t pte = { .pte = 0 }; - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.set_pte(pte, ptep, vmi_flags_addr(mm, addr, VMI_PAGE_PT, 0)); } static void vmi_pmd_clear(pmd_t *pmd) { const pte_t pte = { .pte = 0 }; - vmi_check_page_type(__pa(pmd) >> PAGE_SHIFT, VMI_PAGE_PMD); vmi_ops.set_pte(pte, (pte_t *)pmd, VMI_PAGE_PD); } #endif @@ -960,8 +841,6 @@ static inline int __init activate_vmi(void) void __init vmi_init(void) { - unsigned long flags; - if (!vmi_rom) probe_vmi_rom(); else @@ -973,13 +852,21 @@ 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/vmlinux_32.lds.S b/arch/x86/kernel/vmlinux_32.lds.S index a9b8560adbc..82c67559dde 100644 --- a/arch/x86/kernel/vmlinux_32.lds.S +++ b/arch/x86/kernel/vmlinux_32.lds.S @@ -44,6 +44,7 @@ SECTIONS SCHED_TEXT LOCK_TEXT KPROBES_TEXT + IRQENTRY_TEXT *(.fixup) *(.gnu.warning) _etext = .; /* End of text section */ diff --git a/arch/x86/kernel/vmlinux_64.lds.S b/arch/x86/kernel/vmlinux_64.lds.S index 46e05447405..1a614c0e6be 100644 --- a/arch/x86/kernel/vmlinux_64.lds.S +++ b/arch/x86/kernel/vmlinux_64.lds.S @@ -35,6 +35,7 @@ SECTIONS SCHED_TEXT LOCK_TEXT KPROBES_TEXT + IRQENTRY_TEXT *(.fixup) *(.gnu.warning) _etext = .; /* End of text section */ diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c index 0b8b6690a86..44153afc906 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> @@ -128,7 +131,16 @@ static __always_inline void do_vgettimeofday(struct timeval * tv) gettimeofday(tv,NULL); return; } + + /* + * Surround the RDTSC by barriers, to make sure it's not + * speculated to outside the seqlock critical section and + * does not cause time warps: + */ + rdtsc_barrier(); now = vread(); + rdtsc_barrier(); + base = __vsyscall_gtod_data.clock.cycle_last; mask = __vsyscall_gtod_data.clock.mask; mult = __vsyscall_gtod_data.clock.mult; diff --git a/arch/x86/kernel/xsave.c b/arch/x86/kernel/xsave.c index 15c3e699918..2b54fe002e9 100644 --- a/arch/x86/kernel/xsave.c +++ b/arch/x86/kernel/xsave.c @@ -159,7 +159,7 @@ int save_i387_xstate(void __user *buf) * Restore the extended state if present. Otherwise, restore the FP/SSE * state. */ -int restore_user_xstate(void __user *buf) +static int restore_user_xstate(void __user *buf) { struct _fpx_sw_bytes fx_sw_user; u64 mask; |