From fd35fbcdd1b2579a6e00a1545f7124e4005d0474 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Fri, 22 Oct 2010 15:33:38 -0700 Subject: x86-64, asm: Use fxsaveq/fxrestorq in more places Checkin d7acb92fea932ad2e7846480aeacddc2c03c8485 made use of fxsaveq in fpu_fxsave() if the assembler supports it; this adds fxsaveq/fxrstorq to fxrstor_checking() and fxsave_user() as well. Reported-by: Linus Torvalds LKML-Reference: Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/i387.h | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index 4aa2bb3b242a..ef328901c802 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -93,6 +93,17 @@ static inline int fxrstor_checking(struct i387_fxsave_struct *fx) int err; /* See comment in fxsave() below. */ +#ifdef CONFIG_AS_FXSAVEQ + asm volatile("1: fxrstorq %[fx]\n\t" + "2:\n" + ".section .fixup,\"ax\"\n" + "3: movl $-1,%[err]\n" + " jmp 2b\n" + ".previous\n" + _ASM_EXTABLE(1b, 3b) + : [err] "=r" (err) + : [fx] "m" (*fx), "0" (0)); +#else asm volatile("1: rex64/fxrstor (%[fx])\n\t" "2:\n" ".section .fixup,\"ax\"\n" @@ -102,6 +113,7 @@ static inline int fxrstor_checking(struct i387_fxsave_struct *fx) _ASM_EXTABLE(1b, 3b) : [err] "=r" (err) : [fx] "R" (fx), "m" (*fx), "0" (0)); +#endif return err; } @@ -119,6 +131,17 @@ static inline int fxsave_user(struct i387_fxsave_struct __user *fx) return -EFAULT; /* See comment in fxsave() below. */ +#ifdef CONFIG_AS_FXSAVEQ + asm volatile("1: fxsaveq %[fx]\n\t" + "2:\n" + ".section .fixup,\"ax\"\n" + "3: movl $-1,%[err]\n" + " jmp 2b\n" + ".previous\n" + _ASM_EXTABLE(1b, 3b) + : [err] "=r" (err), [fx] "=m" (*fx) + : "0" (0)); +#else asm volatile("1: rex64/fxsave (%[fx])\n\t" "2:\n" ".section .fixup,\"ax\"\n" @@ -128,6 +151,7 @@ static inline int fxsave_user(struct i387_fxsave_struct __user *fx) _ASM_EXTABLE(1b, 3b) : [err] "=r" (err), "=m" (*fx) : [fx] "R" (fx), "0" (0)); +#endif if (unlikely(err) && __clear_user(fx, sizeof(struct i387_fxsave_struct))) err = -EFAULT; -- cgit v1.2.3 From c7657ac0c3e4d4ab569296911164b7a2b0ff871a Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Mon, 1 Nov 2010 23:36:53 +0100 Subject: x86, microcode, AMD: Cleanup code a bit get_ucode_data is a memcpy() wrapper which always returns 0. Move it into the header and make it an inline. Remove all code checking its return value and turn it into a void. There should be no functionality change resulting from this patch. Signed-off-by: Borislav Petkov --- arch/x86/include/asm/microcode.h | 6 ++++++ arch/x86/kernel/microcode_amd.c | 25 +++++-------------------- 2 files changed, 11 insertions(+), 20 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/microcode.h b/arch/x86/include/asm/microcode.h index ef51b501e22a..24215072d0e1 100644 --- a/arch/x86/include/asm/microcode.h +++ b/arch/x86/include/asm/microcode.h @@ -48,6 +48,12 @@ static inline struct microcode_ops * __init init_intel_microcode(void) #ifdef CONFIG_MICROCODE_AMD extern struct microcode_ops * __init init_amd_microcode(void); + +static inline void get_ucode_data(void *to, const u8 *from, size_t n) +{ + memcpy(to, from, n); +} + #else static inline struct microcode_ops * __init init_amd_microcode(void) { diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index 383d4f8ec9e1..15831336bda8 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -155,12 +155,6 @@ static int apply_microcode_amd(int cpu) return 0; } -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) { @@ -168,8 +162,7 @@ get_next_ucode(const u8 *buf, unsigned int size, unsigned int *mc_size) u8 section_hdr[UCODE_CONTAINER_SECTION_HDR]; void *mc; - if (get_ucode_data(section_hdr, buf, UCODE_CONTAINER_SECTION_HDR)) - return NULL; + get_ucode_data(section_hdr, buf, UCODE_CONTAINER_SECTION_HDR); if (section_hdr[0] != UCODE_UCODE_TYPE) { pr_err("error: invalid type field in container file section header\n"); @@ -187,12 +180,8 @@ get_next_ucode(const u8 *buf, unsigned int size, unsigned int *mc_size) if (!mc) return NULL; - 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; - } + get_ucode_data(mc, buf + UCODE_CONTAINER_SECTION_HDR, total_size); + *mc_size = total_size + UCODE_CONTAINER_SECTION_HDR; return mc; } @@ -203,8 +192,7 @@ static int install_equiv_cpu_table(const u8 *buf) unsigned int *buf_pos = (unsigned int *)container_hdr; unsigned long size; - if (get_ucode_data(&container_hdr, buf, UCODE_CONTAINER_HEADER_SIZE)) - return 0; + get_ucode_data(&container_hdr, buf, UCODE_CONTAINER_HEADER_SIZE); size = buf_pos[2]; @@ -220,10 +208,7 @@ static int install_equiv_cpu_table(const u8 *buf) } buf += UCODE_CONTAINER_HEADER_SIZE; - if (get_ucode_data(equiv_cpu_table, buf, size)) { - vfree(equiv_cpu_table); - return 0; - } + get_ucode_data(equiv_cpu_table, buf, size); return size + UCODE_CONTAINER_HEADER_SIZE; /* add header length */ } -- cgit v1.2.3 From b5908548537ccd3ada258ca5348df7ffc93e5a06 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 10 Nov 2010 22:29:49 -0500 Subject: tracing: Force arch_local_irq_* notrace for paravirt When running ktest.pl randconfig tests, I would sometimes trigger a lockdep annotation bug (possible reason: unannotated irqs-on). This triggering happened right after function tracer self test was executed. After doing a config bisect I found that this was caused with having function tracer, paravirt guest, prove locking, and rcu torture all enabled. The rcu torture just enhanced the likelyhood of triggering the bug. Prove locking was needed, since it was the thing that was bugging. Function tracer would trace and disable interrupts in all sorts of funny places. paravirt guest would turn arch_local_irq_* into functions that would be traced. Besides the fact that tracing arch_local_irq_* is just a bad idea, this is what is happening. The bug happened simply in the local_irq_restore() code: if (raw_irqs_disabled_flags(flags)) { \ raw_local_irq_restore(flags); \ trace_hardirqs_off(); \ } else { \ trace_hardirqs_on(); \ raw_local_irq_restore(flags); \ } \ The raw_local_irq_restore() was defined as arch_local_irq_restore(). Now imagine, we are about to enable interrupts. We go into the else case and call trace_hardirqs_on() which tells lockdep that we are enabling interrupts, so it sets the current->hardirqs_enabled = 1. Then we call raw_local_irq_restore() which calls arch_local_irq_restore() which gets traced! Now in the function tracer we disable interrupts with local_irq_save(). This is fine, but flags is stored that we have interrupts disabled. When the function tracer calls local_irq_restore() it does it, but this time with flags set as disabled, so we go into the if () path. This keeps interrupts disabled and calls trace_hardirqs_off() which sets current->hardirqs_enabled = 0. When the tracer is finished and proceeds with the original code, we enable interrupts but leave current->hardirqs_enabled as 0. Which now breaks lockdeps internal processing. Cc: Thomas Gleixner Signed-off-by: Steven Rostedt --- arch/x86/include/asm/paravirt.h | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h index 18e3b8a8709f..ef9975812c77 100644 --- a/arch/x86/include/asm/paravirt.h +++ b/arch/x86/include/asm/paravirt.h @@ -824,27 +824,27 @@ static __always_inline void arch_spin_unlock(struct arch_spinlock *lock) #define __PV_IS_CALLEE_SAVE(func) \ ((struct paravirt_callee_save) { func }) -static inline unsigned long arch_local_save_flags(void) +static inline notrace unsigned long arch_local_save_flags(void) { return PVOP_CALLEE0(unsigned long, pv_irq_ops.save_fl); } -static inline void arch_local_irq_restore(unsigned long f) +static inline notrace void arch_local_irq_restore(unsigned long f) { PVOP_VCALLEE1(pv_irq_ops.restore_fl, f); } -static inline void arch_local_irq_disable(void) +static inline notrace void arch_local_irq_disable(void) { PVOP_VCALLEE0(pv_irq_ops.irq_disable); } -static inline void arch_local_irq_enable(void) +static inline notrace void arch_local_irq_enable(void) { PVOP_VCALLEE0(pv_irq_ops.irq_enable); } -static inline unsigned long arch_local_irq_save(void) +static inline notrace unsigned long arch_local_irq_save(void) { unsigned long f; -- cgit v1.2.3 From 7e77506a5918d82cafa2ffa783ab57c23f9e9817 Mon Sep 17 00:00:00 2001 From: Ian Campbell Date: Thu, 30 Sep 2010 12:37:26 +0100 Subject: xen: implement XENMEM_machphys_mapping This hypercall allows Xen to specify a non-default location for the machine to physical mapping. This capability is used when running a 32 bit domain 0 on a 64 bit hypervisor to shrink the hypervisor hole to exactly the size required. [ Impact: add Xen hypercall definitions ] Signed-off-by: Ian Campbell Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Stefano Stabellini --- arch/x86/include/asm/xen/interface.h | 6 +++--- arch/x86/include/asm/xen/interface_32.h | 5 +++++ arch/x86/include/asm/xen/interface_64.h | 13 +------------ arch/x86/include/asm/xen/page.h | 7 ++++--- arch/x86/xen/enlighten.c | 7 +++++++ arch/x86/xen/mmu.c | 14 ++++++++++++++ include/xen/interface/memory.h | 13 +++++++++++++ 7 files changed, 47 insertions(+), 18 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/xen/interface.h b/arch/x86/include/asm/xen/interface.h index e8506c1f0c55..1c10c88ee4e1 100644 --- a/arch/x86/include/asm/xen/interface.h +++ b/arch/x86/include/asm/xen/interface.h @@ -61,9 +61,9 @@ DEFINE_GUEST_HANDLE(void); #define HYPERVISOR_VIRT_START mk_unsigned_long(__HYPERVISOR_VIRT_START) #endif -#ifndef machine_to_phys_mapping -#define machine_to_phys_mapping ((unsigned long *)HYPERVISOR_VIRT_START) -#endif +#define MACH2PHYS_VIRT_START mk_unsigned_long(__MACH2PHYS_VIRT_START) +#define MACH2PHYS_VIRT_END mk_unsigned_long(__MACH2PHYS_VIRT_END) +#define MACH2PHYS_NR_ENTRIES ((MACH2PHYS_VIRT_END-MACH2PHYS_VIRT_START)>>__MACH2PHYS_SHIFT) /* Maximum number of virtual CPUs in multi-processor guests. */ #define MAX_VIRT_CPUS 32 diff --git a/arch/x86/include/asm/xen/interface_32.h b/arch/x86/include/asm/xen/interface_32.h index 42a7e004ae5c..8413688b2571 100644 --- a/arch/x86/include/asm/xen/interface_32.h +++ b/arch/x86/include/asm/xen/interface_32.h @@ -32,6 +32,11 @@ /* And the trap vector is... */ #define TRAP_INSTR "int $0x82" +#define __MACH2PHYS_VIRT_START 0xF5800000 +#define __MACH2PHYS_VIRT_END 0xF6800000 + +#define __MACH2PHYS_SHIFT 2 + /* * Virtual addresses beyond this are not modifiable by guest OSes. The * machine->physical mapping table starts at this address, read-only. diff --git a/arch/x86/include/asm/xen/interface_64.h b/arch/x86/include/asm/xen/interface_64.h index 100d2662b97c..839a4811cf98 100644 --- a/arch/x86/include/asm/xen/interface_64.h +++ b/arch/x86/include/asm/xen/interface_64.h @@ -39,18 +39,7 @@ #define __HYPERVISOR_VIRT_END 0xFFFF880000000000 #define __MACH2PHYS_VIRT_START 0xFFFF800000000000 #define __MACH2PHYS_VIRT_END 0xFFFF804000000000 - -#ifndef HYPERVISOR_VIRT_START -#define HYPERVISOR_VIRT_START mk_unsigned_long(__HYPERVISOR_VIRT_START) -#define HYPERVISOR_VIRT_END mk_unsigned_long(__HYPERVISOR_VIRT_END) -#endif - -#define MACH2PHYS_VIRT_START mk_unsigned_long(__MACH2PHYS_VIRT_START) -#define MACH2PHYS_VIRT_END mk_unsigned_long(__MACH2PHYS_VIRT_END) -#define MACH2PHYS_NR_ENTRIES ((MACH2PHYS_VIRT_END-MACH2PHYS_VIRT_START)>>3) -#ifndef machine_to_phys_mapping -#define machine_to_phys_mapping ((unsigned long *)HYPERVISOR_VIRT_START) -#endif +#define __MACH2PHYS_SHIFT 3 /* * int HYPERVISOR_set_segment_base(unsigned int which, unsigned long base) diff --git a/arch/x86/include/asm/xen/page.h b/arch/x86/include/asm/xen/page.h index dd8c1414b3d5..8760cc60a21c 100644 --- a/arch/x86/include/asm/xen/page.h +++ b/arch/x86/include/asm/xen/page.h @@ -5,6 +5,7 @@ #include #include #include +#include #include #include @@ -35,6 +36,8 @@ typedef struct xpaddr { #define MAX_DOMAIN_PAGES \ ((unsigned long)((u64)CONFIG_XEN_MAX_DOMAIN_MEMORY * 1024 * 1024 * 1024 / PAGE_SIZE)) +extern unsigned long *machine_to_phys_mapping; +extern unsigned int machine_to_phys_order; extern unsigned long get_phys_to_machine(unsigned long pfn); extern bool set_phys_to_machine(unsigned long pfn, unsigned long mfn); @@ -69,10 +72,8 @@ static inline unsigned long mfn_to_pfn(unsigned long mfn) if (xen_feature(XENFEAT_auto_translated_physmap)) return mfn; -#if 0 if (unlikely((mfn >> machine_to_phys_order) != 0)) - return max_mapnr; -#endif + return ~0; pfn = 0; /* diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 235c0f4d3861..bd3554934613 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -75,6 +75,11 @@ DEFINE_PER_CPU(struct vcpu_info, xen_vcpu_info); enum xen_domain_type xen_domain_type = XEN_NATIVE; EXPORT_SYMBOL_GPL(xen_domain_type); +unsigned long *machine_to_phys_mapping = (void *)MACH2PHYS_VIRT_START; +EXPORT_SYMBOL(machine_to_phys_mapping); +unsigned int machine_to_phys_order; +EXPORT_SYMBOL(machine_to_phys_order); + struct start_info *xen_start_info; EXPORT_SYMBOL_GPL(xen_start_info); @@ -1097,6 +1102,8 @@ asmlinkage void __init xen_start_kernel(void) xen_domain_type = XEN_PV_DOMAIN; + xen_setup_machphys_mapping(); + /* Install Xen paravirt ops */ pv_info = xen_info; pv_init_ops = xen_init_ops; diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index 21ed8d7f75a5..bd2713a82571 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -2034,6 +2034,20 @@ static __init void xen_map_identity_early(pmd_t *pmd, unsigned long max_pfn) set_page_prot(pmd, PAGE_KERNEL_RO); } +void __init xen_setup_machphys_mapping(void) +{ + struct xen_machphys_mapping mapping; + unsigned long machine_to_phys_nr_ents; + + if (HYPERVISOR_memory_op(XENMEM_machphys_mapping, &mapping) == 0) { + machine_to_phys_mapping = (unsigned long *)mapping.v_start; + machine_to_phys_nr_ents = mapping.max_mfn + 1; + } else { + machine_to_phys_nr_ents = MACH2PHYS_NR_ENTRIES; + } + machine_to_phys_order = fls(machine_to_phys_nr_ents - 1); +} + #ifdef CONFIG_X86_64 static void convert_pfn_mfn(void *v) { diff --git a/include/xen/interface/memory.h b/include/xen/interface/memory.h index d7a6c13bde69..eac3ce153719 100644 --- a/include/xen/interface/memory.h +++ b/include/xen/interface/memory.h @@ -140,6 +140,19 @@ struct xen_machphys_mfn_list { }; DEFINE_GUEST_HANDLE_STRUCT(xen_machphys_mfn_list); +/* + * Returns the location in virtual address space of the machine_to_phys + * mapping table. Architectures which do not have a m2p table, or which do not + * map it by default into guest address space, do not implement this command. + * arg == addr of xen_machphys_mapping_t. + */ +#define XENMEM_machphys_mapping 12 +struct xen_machphys_mapping { + unsigned long v_start, v_end; /* Start and end virtual addresses. */ + unsigned long max_mfn; /* Maximum MFN that can be looked up. */ +}; +DEFINE_GUEST_HANDLE_STRUCT(xen_machphys_mapping_t); + /* * Sets the GPFN at which a particular page appears in the specified guest's * pseudophysical address space. -- cgit v1.2.3 From 5f2b0ba4d94b3ac23cbc4b7f675d98eb677a760a Mon Sep 17 00:00:00 2001 From: Don Zickus Date: Fri, 12 Nov 2010 11:22:23 -0500 Subject: x86, nmi_watchdog: Remove the old nmi_watchdog Now that we have a new nmi_watchdog that is more generic and sits on top of the perf subsystem, we really do not need the old nmi_watchdog any more. In addition, the old nmi_watchdog doesn't really work if you are using the default clocksource, hpet. The old nmi_watchdog code relied on local apic interrupts to determine if the cpu is still alive. With hpet as the clocksource, these interrupts don't increment any more and the old nmi_watchdog triggers false postives. This piece removes the old nmi_watchdog code and stubs out any variables and functions calls. The stubs are the same ones used by the new nmi_watchdog code, so it should be well tested. Signed-off-by: Don Zickus Cc: fweisbec@gmail.com Cc: gorcunov@openvz.org LKML-Reference: <1289578944-28564-2-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/nmi.h | 4 - arch/x86/kernel/apic/Makefile | 5 +- arch/x86/kernel/apic/hw_nmi.c | 6 +- arch/x86/kernel/apic/nmi.c | 567 ------------------------------------------ arch/x86/kernel/traps.c | 9 - include/linux/nmi.h | 6 +- kernel/sysctl.c | 16 -- 7 files changed, 5 insertions(+), 608 deletions(-) delete mode 100644 arch/x86/kernel/apic/nmi.c (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/nmi.h b/arch/x86/include/asm/nmi.h index 932f0f86b4b7..33292ec848ca 100644 --- a/arch/x86/include/asm/nmi.h +++ b/arch/x86/include/asm/nmi.h @@ -17,9 +17,6 @@ int do_nmi_callback(struct pt_regs *regs, int cpu); extern void die_nmi(char *str, struct pt_regs *regs, int do_panic); extern int check_nmi_watchdog(void); -#if !defined(CONFIG_LOCKUP_DETECTOR) -extern int nmi_watchdog_enabled; -#endif extern int avail_to_resrv_perfctr_nmi_bit(unsigned int); extern int reserve_perfctr_nmi(unsigned int); extern void release_perfctr_nmi(unsigned int); @@ -30,7 +27,6 @@ extern void setup_apic_nmi_watchdog(void *); extern void stop_apic_nmi_watchdog(void *); extern void disable_timer_nmi_watchdog(void); extern void enable_timer_nmi_watchdog(void); -extern int nmi_watchdog_tick(struct pt_regs *regs, unsigned reason); extern void cpu_nmi_set_wd_enabled(void); extern atomic_t nmi_active; diff --git a/arch/x86/kernel/apic/Makefile b/arch/x86/kernel/apic/Makefile index 910f20b457c4..3966b564ea47 100644 --- a/arch/x86/kernel/apic/Makefile +++ b/arch/x86/kernel/apic/Makefile @@ -3,10 +3,7 @@ # obj-$(CONFIG_X86_LOCAL_APIC) += apic.o apic_noop.o probe_$(BITS).o ipi.o -ifneq ($(CONFIG_HARDLOCKUP_DETECTOR),y) -obj-$(CONFIG_X86_LOCAL_APIC) += nmi.o -endif -obj-$(CONFIG_HARDLOCKUP_DETECTOR) += hw_nmi.o +obj-y += hw_nmi.o obj-$(CONFIG_X86_IO_APIC) += io_apic.o obj-$(CONFIG_SMP) += ipi.o diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c index cefd6942f0e9..b68b17460016 100644 --- a/arch/x86/kernel/apic/hw_nmi.c +++ b/arch/x86/kernel/apic/hw_nmi.c @@ -20,12 +20,14 @@ /* For reliability, we're prepared to waste bits here. */ static DECLARE_BITMAP(backtrace_mask, NR_CPUS) __read_mostly; +#ifdef CONFIG_HARDLOCKUP_DETECTOR u64 hw_nmi_get_sample_period(void) { return (u64)(cpu_khz) * 1000 * 60; } +#endif -#ifdef ARCH_HAS_NMI_WATCHDOG +#ifdef arch_trigger_all_cpu_backtrace void arch_trigger_all_cpu_backtrace(void) { int i; @@ -95,8 +97,6 @@ early_initcall(register_trigger_all_cpu_backtrace); #if defined(CONFIG_X86_LOCAL_APIC) unsigned int nmi_watchdog = NMI_NONE; EXPORT_SYMBOL(nmi_watchdog); -void acpi_nmi_enable(void) { return; } -void acpi_nmi_disable(void) { return; } #endif atomic_t nmi_active = ATOMIC_INIT(0); /* oprofile uses this */ EXPORT_SYMBOL(nmi_active); diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c deleted file mode 100644 index c90041ccb742..000000000000 --- a/arch/x86/kernel/apic/nmi.c +++ /dev/null @@ -1,567 +0,0 @@ -/* - * NMI watchdog support on APIC systems - * - * Started by Ingo Molnar - * - * Fixes: - * Mikael Pettersson : AMD K7 support for local APIC NMI watchdog. - * Mikael Pettersson : Power Management for local APIC NMI watchdog. - * Mikael Pettersson : Pentium 4 support for local APIC NMI watchdog. - * Pavel Machek and - * Mikael Pettersson : PM converted to driver model. Disable/enable API. - */ - -#include - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include - -#include - -#include - -int unknown_nmi_panic; -int nmi_watchdog_enabled; - -/* For reliability, we're prepared to waste bits here. */ -static DECLARE_BITMAP(backtrace_mask, NR_CPUS) __read_mostly; - -/* nmi_active: - * >0: the lapic NMI watchdog is active, but can be disabled - * <0: the lapic NMI watchdog has not been set up, and cannot - * be enabled - * 0: the lapic NMI watchdog is disabled, but can be enabled - */ -atomic_t nmi_active = ATOMIC_INIT(0); /* oprofile uses this */ -EXPORT_SYMBOL(nmi_active); - -unsigned int nmi_watchdog = NMI_NONE; -EXPORT_SYMBOL(nmi_watchdog); - -static int panic_on_timeout; - -static unsigned int nmi_hz = HZ; -static DEFINE_PER_CPU(short, wd_enabled); -static int endflag __initdata; - -static inline unsigned int get_nmi_count(int cpu) -{ - return per_cpu(irq_stat, cpu).__nmi_count; -} - -static inline int mce_in_progress(void) -{ -#if defined(CONFIG_X86_MCE) - return atomic_read(&mce_entry) > 0; -#endif - return 0; -} - -/* - * Take the local apic timer and PIT/HPET into account. We don't - * know which one is active, when we have highres/dyntick on - */ -static inline unsigned int get_timer_irqs(int cpu) -{ - return per_cpu(irq_stat, cpu).apic_timer_irqs + - per_cpu(irq_stat, cpu).irq0_irqs; -} - -#ifdef CONFIG_SMP -/* - * The performance counters used by NMI_LOCAL_APIC don't trigger when - * the CPU is idle. To make sure the NMI watchdog really ticks on all - * CPUs during the test make them busy. - */ -static __init void nmi_cpu_busy(void *data) -{ - local_irq_enable_in_hardirq(); - /* - * Intentionally don't use cpu_relax here. This is - * to make sure that the performance counter really ticks, - * even if there is a simulator or similar that catches the - * pause instruction. On a real HT machine this is fine because - * all other CPUs are busy with "useless" delay loops and don't - * care if they get somewhat less cycles. - */ - while (endflag == 0) - mb(); -} -#endif - -static void report_broken_nmi(int cpu, unsigned int *prev_nmi_count) -{ - printk(KERN_CONT "\n"); - - printk(KERN_WARNING - "WARNING: CPU#%d: NMI appears to be stuck (%d->%d)!\n", - cpu, prev_nmi_count[cpu], get_nmi_count(cpu)); - - printk(KERN_WARNING - "Please report this to bugzilla.kernel.org,\n"); - printk(KERN_WARNING - "and attach the output of the 'dmesg' command.\n"); - - per_cpu(wd_enabled, cpu) = 0; - 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; - int cpu; - - if (!nmi_watchdog_active() || !atomic_read(&nmi_active)) - return 0; - - prev_nmi_count = kmalloc(nr_cpu_ids * sizeof(int), GFP_KERNEL); - if (!prev_nmi_count) - goto error; - - printk(KERN_INFO "Testing NMI watchdog ... "); - -#ifdef CONFIG_SMP - if (nmi_watchdog == NMI_LOCAL_APIC) - smp_call_function(nmi_cpu_busy, (void *)&endflag, 0); -#endif - - for_each_possible_cpu(cpu) - prev_nmi_count[cpu] = get_nmi_count(cpu); - local_irq_enable(); - mdelay((20 * 1000) / nmi_hz); /* wait 20 ticks */ - - for_each_online_cpu(cpu) { - if (!per_cpu(wd_enabled, cpu)) - continue; - if (get_nmi_count(cpu) - prev_nmi_count[cpu] <= 5) - report_broken_nmi(cpu, prev_nmi_count); - } - endflag = 1; - if (!atomic_read(&nmi_active)) { - kfree(prev_nmi_count); - atomic_set(&nmi_active, -1); - goto error; - } - printk("OK.\n"); - - /* - * now that we know it works we can reduce NMI frequency to - * something more reasonable; makes a difference in some configs - */ - if (nmi_watchdog == NMI_LOCAL_APIC) - nmi_hz = lapic_adjust_nmi_hz(1); - - kfree(prev_nmi_count); - return 0; -error: - if (nmi_watchdog == NMI_IO_APIC) { - if (!timer_through_8259) - legacy_pic->mask(0); - on_each_cpu(__acpi_nmi_disable, NULL, 1); - } - -#ifdef CONFIG_X86_32 - timer_ack = 0; -#endif - return -1; -} - -static int __init setup_nmi_watchdog(char *str) -{ - unsigned int nmi; - - if (!strncmp(str, "panic", 5)) { - panic_on_timeout = 1; - str = strchr(str, ','); - if (!str) - return 1; - ++str; - } - - 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; - } - - return 1; -} -__setup("nmi_watchdog=", setup_nmi_watchdog); - -/* - * Suspend/resume support - */ -#ifdef CONFIG_PM - -static int nmi_pm_active; /* nmi_active before suspend */ - -static int lapic_nmi_suspend(struct sys_device *dev, pm_message_t state) -{ - /* only CPU0 goes here, other CPUs should be offline */ - nmi_pm_active = atomic_read(&nmi_active); - stop_apic_nmi_watchdog(NULL); - BUG_ON(atomic_read(&nmi_active) != 0); - return 0; -} - -static int lapic_nmi_resume(struct sys_device *dev) -{ - /* only CPU0 goes here, other CPUs should be offline */ - if (nmi_pm_active > 0) { - setup_apic_nmi_watchdog(NULL); - touch_nmi_watchdog(); - } - return 0; -} - -static struct sysdev_class nmi_sysclass = { - .name = "lapic_nmi", - .resume = lapic_nmi_resume, - .suspend = lapic_nmi_suspend, -}; - -static struct sys_device device_lapic_nmi = { - .id = 0, - .cls = &nmi_sysclass, -}; - -static int __init init_lapic_nmi_sysfs(void) -{ - int error; - - /* - * should really be a BUG_ON but b/c this is an - * init call, it just doesn't work. -dcz - */ - if (nmi_watchdog != NMI_LOCAL_APIC) - return 0; - - if (atomic_read(&nmi_active) < 0) - return 0; - - error = sysdev_class_register(&nmi_sysclass); - if (!error) - error = sysdev_register(&device_lapic_nmi); - return error; -} - -/* must come after the local APIC's device_initcall() */ -late_initcall(init_lapic_nmi_sysfs); - -#endif /* CONFIG_PM */ - -static void __acpi_nmi_enable(void *__unused) -{ - apic_write(APIC_LVT0, APIC_DM_NMI); -} - -/* - * Enable timer based NMIs on all CPUs: - */ -void acpi_nmi_enable(void) -{ - if (atomic_read(&nmi_active) && nmi_watchdog == NMI_IO_APIC) - on_each_cpu(__acpi_nmi_enable, NULL, 1); -} - -/* - * Disable timer based NMIs on all CPUs: - */ -void acpi_nmi_disable(void) -{ - if (atomic_read(&nmi_active) && nmi_watchdog == NMI_IO_APIC) - on_each_cpu(__acpi_nmi_disable, NULL, 1); -} - -/* - * This function is called as soon the LAPIC NMI watchdog driver has everything - * in place and it's ready to check if the NMIs belong to the NMI watchdog - */ -void cpu_nmi_set_wd_enabled(void) -{ - __get_cpu_var(wd_enabled) = 1; -} - -void setup_apic_nmi_watchdog(void *unused) -{ - if (__get_cpu_var(wd_enabled)) - return; - - /* cheap hack to support suspend/resume */ - /* if cpu0 is not active neither should the other cpus */ - if (smp_processor_id() != 0 && atomic_read(&nmi_active) <= 0) - return; - - switch (nmi_watchdog) { - case NMI_LOCAL_APIC: - if (lapic_watchdog_init(nmi_hz) < 0) { - __get_cpu_var(wd_enabled) = 0; - return; - } - /* FALL THROUGH */ - case NMI_IO_APIC: - __get_cpu_var(wd_enabled) = 1; - atomic_inc(&nmi_active); - } -} - -void stop_apic_nmi_watchdog(void *unused) -{ - /* only support LOCAL and IO APICs for now */ - if (!nmi_watchdog_active()) - return; - if (__get_cpu_var(wd_enabled) == 0) - 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); -} - -/* - * the best way to detect whether a CPU has a 'hard lockup' problem - * is to check it's local APIC timer IRQ counts. If they are not - * changing then that CPU has some problem. - * - * as these watchdog NMI IRQs are generated on every CPU, we only - * have to check the current processor. - * - * since NMIs don't listen to _any_ locks, we have to be extremely - * careful not to rely on unsafe variables. The printk might lock - * up though, so we have to break up any console locks first ... - * [when there will be more tty-related locks, break them up here too!] - */ - -static DEFINE_PER_CPU(unsigned, last_irq_sum); -static DEFINE_PER_CPU(long, alert_counter); -static DEFINE_PER_CPU(int, nmi_touch); - -void touch_nmi_watchdog(void) -{ - if (nmi_watchdog_active()) { - unsigned cpu; - - /* - * Tell other CPUs to reset their alert counters. We cannot - * do it ourselves because the alert count increase is not - * atomic. - */ - for_each_present_cpu(cpu) { - if (per_cpu(nmi_touch, cpu) != 1) - per_cpu(nmi_touch, cpu) = 1; - } - } - - /* - * Tickle the softlockup detector too: - */ - touch_softlockup_watchdog(); -} -EXPORT_SYMBOL(touch_nmi_watchdog); - -notrace __kprobes int -nmi_watchdog_tick(struct pt_regs *regs, unsigned reason) -{ - /* - * Since current_thread_info()-> is always on the stack, and we - * always switch the stack NMI-atomically, it's safe to use - * smp_processor_id(). - */ - unsigned int sum; - int touched = 0; - int cpu = smp_processor_id(); - int rc = 0; - - sum = get_timer_irqs(cpu); - - if (__get_cpu_var(nmi_touch)) { - __get_cpu_var(nmi_touch) = 0; - touched = 1; - } - - /* We can be called before check_nmi_watchdog, hence NULL check. */ - if (cpumask_test_cpu(cpu, to_cpumask(backtrace_mask))) { - static DEFINE_RAW_SPINLOCK(lock); /* Serialise the printks */ - - raw_spin_lock(&lock); - printk(KERN_WARNING "NMI backtrace for cpu %d\n", cpu); - show_regs(regs); - dump_stack(); - raw_spin_unlock(&lock); - cpumask_clear_cpu(cpu, to_cpumask(backtrace_mask)); - - rc = 1; - } - - /* Could check oops_in_progress here too, but it's safer not to */ - if (mce_in_progress()) - touched = 1; - - /* if the none of the timers isn't firing, this cpu isn't doing much */ - if (!touched && __get_cpu_var(last_irq_sum) == sum) { - /* - * Ayiee, looks like this CPU is stuck ... - * wait a few IRQs (5 seconds) before doing the oops ... - */ - __this_cpu_inc(alert_counter); - if (__this_cpu_read(alert_counter) == 5 * nmi_hz) - /* - * die_nmi will return ONLY if NOTIFY_STOP happens.. - */ - die_nmi("BUG: NMI Watchdog detected LOCKUP", - regs, panic_on_timeout); - } else { - __get_cpu_var(last_irq_sum) = sum; - __this_cpu_write(alert_counter, 0); - } - - /* see if the nmi watchdog went off */ - if (!__get_cpu_var(wd_enabled)) - return rc; - switch (nmi_watchdog) { - case NMI_LOCAL_APIC: - rc |= lapic_wd_event(nmi_hz); - break; - case NMI_IO_APIC: - /* - * don't know how to accurately check for this. - * just assume it was a watchdog timer interrupt - * This matches the old behaviour. - */ - rc = 1; - break; - } - return rc; -} - -#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; - return 1; -} -__setup("unknown_nmi_panic", setup_unknown_nmi_panic); - -static int unknown_nmi_panic_callback(struct pt_regs *regs, int cpu) -{ - unsigned char reason = get_nmi_reason(); - char buf[64]; - - sprintf(buf, "NMI received for unknown reason %02x\n", reason); - die_nmi(buf, regs, 1); /* Always panic here */ - return 0; -} - -/* - * proc handler for /proc/sys/kernel/nmi - */ -int proc_nmi_enabled(struct ctl_table *table, int write, - void __user *buffer, size_t *length, loff_t *ppos) -{ - int old_state; - - nmi_watchdog_enabled = (atomic_read(&nmi_active) > 0) ? 1 : 0; - old_state = nmi_watchdog_enabled; - proc_dointvec(table, write, buffer, length, ppos); - if (!!old_state == !!nmi_watchdog_enabled) - return 0; - - if (atomic_read(&nmi_active) < 0 || !nmi_watchdog_active()) { - printk(KERN_WARNING - "NMI watchdog is permanently disabled\n"); - return -EIO; - } - - if (nmi_watchdog == NMI_LOCAL_APIC) { - if (nmi_watchdog_enabled) - 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"); - return -EIO; - } - return 0; -} - -#endif /* CONFIG_SYSCTL */ - -int do_nmi_callback(struct pt_regs *regs, int cpu) -{ -#ifdef CONFIG_SYSCTL - if (unknown_nmi_panic) - return unknown_nmi_panic_callback(regs, cpu); -#endif - return 0; -} - -void arch_trigger_all_cpu_backtrace(void) -{ - int i; - - cpumask_copy(to_cpumask(backtrace_mask), cpu_online_mask); - - printk(KERN_INFO "sending NMI to all CPUs:\n"); - apic->send_IPI_all(NMI_VECTOR); - - /* Wait for up to 10 seconds for all CPUs to do the backtrace */ - for (i = 0; i < 10 * 1000; i++) { - if (cpumask_empty(to_cpumask(backtrace_mask))) - break; - mdelay(1); - } -} diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index cb838ca42c96..db30d9cb9dd6 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -398,15 +398,6 @@ static notrace __kprobes void default_do_nmi(struct pt_regs *regs) == NOTIFY_STOP) return; -#ifndef CONFIG_LOCKUP_DETECTOR - /* - * Ok, so this is none of the documented NMI sources, - * so it must be the NMI watchdog. - */ - if (nmi_watchdog_tick(regs, reason)) - return; - if (!do_nmi_callback(regs, cpu)) -#endif /* !CONFIG_LOCKUP_DETECTOR */ unknown_nmi_error(reason, regs); #else unknown_nmi_error(reason, regs); diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 06aab5eee134..0cb3e5c246d0 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -16,10 +16,7 @@ */ #ifdef ARCH_HAS_NMI_WATCHDOG #include -extern void touch_nmi_watchdog(void); -extern void acpi_nmi_disable(void); -extern void acpi_nmi_enable(void); -#else +#endif #ifndef CONFIG_HARDLOCKUP_DETECTOR static inline void touch_nmi_watchdog(void) { @@ -30,7 +27,6 @@ extern void touch_nmi_watchdog(void); #endif static inline void acpi_nmi_disable(void) { } static inline void acpi_nmi_enable(void) { } -#endif /* * Create trigger_all_cpu_backtrace() out of the arch-provided diff --git a/kernel/sysctl.c b/kernel/sysctl.c index b65bf634035e..ce33e2a2afea 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -746,22 +746,6 @@ static struct ctl_table kern_table[] = { .extra2 = &one, }, #endif -#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86) && !defined(CONFIG_LOCKUP_DETECTOR) - { - .procname = "unknown_nmi_panic", - .data = &unknown_nmi_panic, - .maxlen = sizeof (int), - .mode = 0644, - .proc_handler = proc_dointvec, - }, - { - .procname = "nmi_watchdog", - .data = &nmi_watchdog_enabled, - .maxlen = sizeof (int), - .mode = 0644, - .proc_handler = proc_nmi_enabled, - }, -#endif #if defined(CONFIG_X86) { .procname = "panic_on_unrecovered_nmi", -- cgit v1.2.3 From 072b198a4ad48bd722ec6d203d65422a4698eae7 Mon Sep 17 00:00:00 2001 From: Don Zickus Date: Fri, 12 Nov 2010 11:22:24 -0500 Subject: x86, nmi_watchdog: Remove all stub function calls from old nmi_watchdog Now that the bulk of the old nmi_watchdog is gone, remove all the stub variables and hooks associated with it. This touches lots of files mainly because of how the io_apic nmi_watchdog was implemented. Now that the io_apic nmi_watchdog is forever gone, remove all its fingers. Most of this code was not being exercised by virtue of nmi_watchdog != NMI_IO_APIC, so there shouldn't be anything to risky here. Signed-off-by: Don Zickus Cc: fweisbec@gmail.com Cc: gorcunov@openvz.org LKML-Reference: <1289578944-28564-3-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/nmi.h | 47 --- arch/x86/include/asm/smpboot_hooks.h | 1 - arch/x86/include/asm/timer.h | 6 - arch/x86/kernel/apic/apic.c | 15 +- arch/x86/kernel/apic/hw_nmi.c | 10 - arch/x86/kernel/apic/io_apic.c | 46 --- arch/x86/kernel/cpu/perf_event.c | 9 - arch/x86/kernel/cpu/perfctr-watchdog.c | 642 --------------------------------- arch/x86/kernel/smpboot.c | 11 - arch/x86/kernel/time.c | 18 - arch/x86/kernel/traps.c | 2 - arch/x86/oprofile/nmi_timer_int.c | 3 - drivers/acpi/acpica/nsinit.c | 2 - drivers/watchdog/hpwdt.c | 7 +- include/linux/nmi.h | 2 - 15 files changed, 2 insertions(+), 819 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/nmi.h b/arch/x86/include/asm/nmi.h index 33292ec848ca..3545838cddeb 100644 --- a/arch/x86/include/asm/nmi.h +++ b/arch/x86/include/asm/nmi.h @@ -7,35 +7,13 @@ #ifdef ARCH_HAS_NMI_WATCHDOG -/** - * do_nmi_callback - * - * Check to see if a callback exists and execute it. Return 1 - * if the handler exists and was handled successfully. - */ -int do_nmi_callback(struct pt_regs *regs, int cpu); - extern void die_nmi(char *str, struct pt_regs *regs, int do_panic); -extern int check_nmi_watchdog(void); extern int avail_to_resrv_perfctr_nmi_bit(unsigned int); extern int reserve_perfctr_nmi(unsigned int); extern void release_perfctr_nmi(unsigned int); extern int reserve_evntsel_nmi(unsigned int); extern void release_evntsel_nmi(unsigned int); -extern void setup_apic_nmi_watchdog(void *); -extern void stop_apic_nmi_watchdog(void *); -extern void disable_timer_nmi_watchdog(void); -extern void enable_timer_nmi_watchdog(void); -extern void cpu_nmi_set_wd_enabled(void); - -extern atomic_t nmi_active; -extern unsigned int nmi_watchdog; -#define NMI_NONE 0 -#define NMI_IO_APIC 1 -#define NMI_LOCAL_APIC 2 -#define NMI_INVALID 3 - struct ctl_table; extern int proc_nmi_enabled(struct ctl_table *, int , void __user *, size_t *, loff_t *); @@ -43,33 +21,8 @@ extern int unknown_nmi_panic; void arch_trigger_all_cpu_backtrace(void); #define arch_trigger_all_cpu_backtrace arch_trigger_all_cpu_backtrace - -static inline void localise_nmi_watchdog(void) -{ - if (nmi_watchdog == NMI_IO_APIC) - nmi_watchdog = NMI_LOCAL_APIC; -} - -/* check if nmi_watchdog is active (ie was specified at boot) */ -static inline int nmi_watchdog_active(void) -{ - /* - * actually it should be: - * return (nmi_watchdog == NMI_LOCAL_APIC || - * nmi_watchdog == NMI_IO_APIC) - * but since they are power of two we could use a - * cheaper way --cvg - */ - return nmi_watchdog & (NMI_LOCAL_APIC | NMI_IO_APIC); -} #endif -void lapic_watchdog_stop(void); -int lapic_watchdog_init(unsigned nmi_hz); -int lapic_wd_event(unsigned nmi_hz); -unsigned lapic_adjust_nmi_hz(unsigned hz); -void disable_lapic_nmi_watchdog(void); -void enable_lapic_nmi_watchdog(void); void stop_nmi(void); void restart_nmi(void); diff --git a/arch/x86/include/asm/smpboot_hooks.h b/arch/x86/include/asm/smpboot_hooks.h index 1def60114906..6c22bf353f26 100644 --- a/arch/x86/include/asm/smpboot_hooks.h +++ b/arch/x86/include/asm/smpboot_hooks.h @@ -48,7 +48,6 @@ static inline void __init smpboot_setup_io_apic(void) setup_IO_APIC(); else { nr_ioapics = 0; - localise_nmi_watchdog(); } #endif } diff --git a/arch/x86/include/asm/timer.h b/arch/x86/include/asm/timer.h index 5469630b27f5..fa7b9176b76c 100644 --- a/arch/x86/include/asm/timer.h +++ b/arch/x86/include/asm/timer.h @@ -10,12 +10,6 @@ unsigned long long native_sched_clock(void); extern int recalibrate_cpu_khz(void); -#if defined(CONFIG_X86_32) && defined(CONFIG_X86_IO_APIC) -extern int timer_ack; -#else -# define timer_ack (0) -#endif - extern int no_timer_check; /* Accelerators for sched_clock() diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 3f838d537392..e9e2a93783f9 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -31,7 +31,6 @@ #include #include #include -#include #include #include @@ -799,11 +798,7 @@ void __init setup_boot_APIC_clock(void) * PIT/HPET going. Otherwise register lapic as a dummy * device. */ - if (nmi_watchdog != NMI_IO_APIC) - lapic_clockevent.features &= ~CLOCK_EVT_FEAT_DUMMY; - else - pr_warning("APIC timer registered as dummy," - " due to nmi_watchdog=%d!\n", nmi_watchdog); + lapic_clockevent.features &= ~CLOCK_EVT_FEAT_DUMMY; /* Setup the lapic or request the broadcast */ setup_APIC_timer(); @@ -1387,7 +1382,6 @@ void __cpuinit end_local_APIC_setup(void) } #endif - setup_apic_nmi_watchdog(NULL); apic_pm_activate(); } @@ -1750,17 +1744,10 @@ int __init APIC_init_uniprocessor(void) setup_IO_APIC(); else { nr_ioapics = 0; - localise_nmi_watchdog(); } -#else - localise_nmi_watchdog(); #endif x86_init.timers.setup_percpu_clockev(); -#ifdef CONFIG_X86_64 - check_nmi_watchdog(); -#endif - return 0; } diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c index b68b17460016..3e25afe9a62a 100644 --- a/arch/x86/kernel/apic/hw_nmi.c +++ b/arch/x86/kernel/apic/hw_nmi.c @@ -94,14 +94,4 @@ early_initcall(register_trigger_all_cpu_backtrace); #endif /* STUB calls to mimic old nmi_watchdog behaviour */ -#if defined(CONFIG_X86_LOCAL_APIC) -unsigned int nmi_watchdog = NMI_NONE; -EXPORT_SYMBOL(nmi_watchdog); -#endif -atomic_t nmi_active = ATOMIC_INIT(0); /* oprofile uses this */ -EXPORT_SYMBOL(nmi_active); int unknown_nmi_panic; -void cpu_nmi_set_wd_enabled(void) { return; } -void stop_apic_nmi_watchdog(void *unused) { return; } -void setup_apic_nmi_watchdog(void *unused) { return; } -int __init check_nmi_watchdog(void) { return 0; } diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 7cc0a721f628..e4a040c28de1 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -54,7 +54,6 @@ #include #include #include -#include #include #include #include @@ -2643,24 +2642,6 @@ static void lapic_register_intr(int irq) "edge"); } -static void __init setup_nmi(void) -{ - /* - * Dirty trick to enable the NMI watchdog ... - * We put the 8259A master into AEOI mode and - * unmask on all local APICs LVT0 as NMI. - * - * The idea to use the 8259A in AEOI mode ('8259A Virtual Wire') - * is from Maciej W. Rozycki - so we do not have to EOI from - * the NMI handler or the timer interrupt. - */ - apic_printk(APIC_VERBOSE, KERN_INFO "activating NMI Watchdog ..."); - - enable_NMI_through_LVT0(); - - apic_printk(APIC_VERBOSE, " done.\n"); -} - /* * This looks a bit hackish but it's about the only one way of sending * a few INTA cycles to 8259As and any associated glue logic. ICR does @@ -2766,15 +2747,6 @@ static inline void __init check_timer(void) */ apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT); legacy_pic->init(1); -#ifdef CONFIG_X86_32 - { - unsigned int ver; - - ver = apic_read(APIC_LVR); - ver = GET_APIC_VERSION(ver); - timer_ack = (nmi_watchdog == NMI_IO_APIC && !APIC_INTEGRATED(ver)); - } -#endif pin1 = find_isa_irq_pin(0, mp_INT); apic1 = find_isa_irq_apic(0, mp_INT); @@ -2822,10 +2794,6 @@ static inline void __init check_timer(void) unmask_ioapic(cfg); } if (timer_irq_works()) { - if (nmi_watchdog == NMI_IO_APIC) { - setup_nmi(); - legacy_pic->unmask(0); - } if (disable_timer_pin_1 > 0) clear_IO_APIC_pin(0, pin1); goto out; @@ -2851,11 +2819,6 @@ static inline void __init check_timer(void) if (timer_irq_works()) { apic_printk(APIC_QUIET, KERN_INFO "....... works.\n"); timer_through_8259 = 1; - if (nmi_watchdog == NMI_IO_APIC) { - legacy_pic->mask(0); - setup_nmi(); - legacy_pic->unmask(0); - } goto out; } /* @@ -2867,15 +2830,6 @@ static inline void __init check_timer(void) apic_printk(APIC_QUIET, KERN_INFO "....... failed.\n"); } - if (nmi_watchdog == NMI_IO_APIC) { - apic_printk(APIC_QUIET, KERN_WARNING "timer doesn't work " - "through the IO-APIC - disabling NMI Watchdog!\n"); - nmi_watchdog = NMI_NONE; - } -#ifdef CONFIG_X86_32 - timer_ack = 0; -#endif - apic_printk(APIC_QUIET, KERN_INFO "...trying to set up timer as Virtual Wire IRQ...\n"); diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index ed6310183efb..1f129a14e3a2 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -330,9 +330,6 @@ static bool reserve_pmc_hardware(void) { int i; - if (nmi_watchdog == NMI_LOCAL_APIC) - disable_lapic_nmi_watchdog(); - for (i = 0; i < x86_pmu.num_counters; i++) { if (!reserve_perfctr_nmi(x86_pmu.perfctr + i)) goto perfctr_fail; @@ -355,9 +352,6 @@ perfctr_fail: for (i--; i >= 0; i--) release_perfctr_nmi(x86_pmu.perfctr + i); - if (nmi_watchdog == NMI_LOCAL_APIC) - enable_lapic_nmi_watchdog(); - return false; } @@ -369,9 +363,6 @@ static void release_pmc_hardware(void) release_perfctr_nmi(x86_pmu.perfctr + i); release_evntsel_nmi(x86_pmu.eventsel + i); } - - if (nmi_watchdog == NMI_LOCAL_APIC) - enable_lapic_nmi_watchdog(); } #else diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c index d9f4ff8fcd69..14d45928c282 100644 --- a/arch/x86/kernel/cpu/perfctr-watchdog.c +++ b/arch/x86/kernel/cpu/perfctr-watchdog.c @@ -22,26 +22,6 @@ #include #include -struct nmi_watchdog_ctlblk { - unsigned int cccr_msr; - unsigned int perfctr_msr; /* the MSR to reset in NMI handler */ - unsigned int evntsel_msr; /* the MSR to select the events to handle */ -}; - -/* Interface defining a CPU specific perfctr watchdog */ -struct wd_ops { - int (*reserve)(void); - void (*unreserve)(void); - int (*setup)(unsigned nmi_hz); - void (*rearm)(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz); - void (*stop)(void); - unsigned perfctr; - unsigned evntsel; - u64 checkbit; -}; - -static const struct wd_ops *wd_ops; - /* * this number is calculated from Intel's MSR_P4_CRU_ESCR5 register and it's * offset from MSR_P4_BSU_ESCR0. @@ -60,8 +40,6 @@ static const struct wd_ops *wd_ops; static DECLARE_BITMAP(perfctr_nmi_owner, NMI_MAX_COUNTER_BITS); static DECLARE_BITMAP(evntsel_nmi_owner, NMI_MAX_COUNTER_BITS); -static DEFINE_PER_CPU(struct nmi_watchdog_ctlblk, nmi_watchdog_ctlblk); - /* converts an msr to an appropriate reservation bit */ static inline unsigned int nmi_perfctr_msr_to_bit(unsigned int msr) { @@ -172,623 +150,3 @@ void release_evntsel_nmi(unsigned int msr) clear_bit(counter, evntsel_nmi_owner); } EXPORT_SYMBOL(release_evntsel_nmi); - -void disable_lapic_nmi_watchdog(void) -{ - BUG_ON(nmi_watchdog != NMI_LOCAL_APIC); - - if (atomic_read(&nmi_active) <= 0) - return; - - on_each_cpu(stop_apic_nmi_watchdog, NULL, 1); - - if (wd_ops) - wd_ops->unreserve(); - - BUG_ON(atomic_read(&nmi_active) != 0); -} - -void enable_lapic_nmi_watchdog(void) -{ - BUG_ON(nmi_watchdog != NMI_LOCAL_APIC); - - /* are we already enabled */ - if (atomic_read(&nmi_active) != 0) - return; - - /* are we lapic aware */ - if (!wd_ops) - return; - if (!wd_ops->reserve()) { - printk(KERN_ERR "NMI watchdog: cannot reserve perfctrs\n"); - return; - } - - on_each_cpu(setup_apic_nmi_watchdog, NULL, 1); - touch_nmi_watchdog(); -} - -/* - * Activate the NMI watchdog via the local APIC. - */ - -static unsigned int adjust_for_32bit_ctr(unsigned int hz) -{ - u64 counter_val; - unsigned int retval = hz; - - /* - * On Intel CPUs with P6/ARCH_PERFMON only 32 bits in the counter - * are writable, with higher bits sign extending from bit 31. - * So, we can only program the counter with 31 bit values and - * 32nd bit should be 1, for 33.. to be 1. - * Find the appropriate nmi_hz - */ - counter_val = (u64)cpu_khz * 1000; - do_div(counter_val, retval); - if (counter_val > 0x7fffffffULL) { - u64 count = (u64)cpu_khz * 1000; - do_div(count, 0x7fffffffUL); - retval = count + 1; - } - return retval; -} - -static void write_watchdog_counter(unsigned int perfctr_msr, - const char *descr, unsigned nmi_hz) -{ - u64 count = (u64)cpu_khz * 1000; - - do_div(count, nmi_hz); - if (descr) - pr_debug("setting %s to -0x%08Lx\n", descr, count); - wrmsrl(perfctr_msr, 0 - count); -} - -static void write_watchdog_counter32(unsigned int perfctr_msr, - const char *descr, unsigned nmi_hz) -{ - u64 count = (u64)cpu_khz * 1000; - - do_div(count, nmi_hz); - if (descr) - pr_debug("setting %s to -0x%08Lx\n", descr, count); - wrmsr(perfctr_msr, (u32)(-count), 0); -} - -/* - * AMD K7/K8/Family10h/Family11h support. - * AMD keeps this interface nicely stable so there is not much variety - */ -#define K7_EVNTSEL_ENABLE (1 << 22) -#define K7_EVNTSEL_INT (1 << 20) -#define K7_EVNTSEL_OS (1 << 17) -#define K7_EVNTSEL_USR (1 << 16) -#define K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING 0x76 -#define K7_NMI_EVENT K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING - -static int setup_k7_watchdog(unsigned nmi_hz) -{ - unsigned int perfctr_msr, evntsel_msr; - unsigned int evntsel; - struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk); - - perfctr_msr = wd_ops->perfctr; - evntsel_msr = wd_ops->evntsel; - - wrmsrl(perfctr_msr, 0UL); - - evntsel = K7_EVNTSEL_INT - | K7_EVNTSEL_OS - | K7_EVNTSEL_USR - | K7_NMI_EVENT; - - /* setup the timer */ - wrmsr(evntsel_msr, evntsel, 0); - write_watchdog_counter(perfctr_msr, "K7_PERFCTR0", nmi_hz); - - /* initialize the wd struct before enabling */ - wd->perfctr_msr = perfctr_msr; - wd->evntsel_msr = evntsel_msr; - wd->cccr_msr = 0; /* unused */ - - /* ok, everything is initialized, announce that we're set */ - cpu_nmi_set_wd_enabled(); - - apic_write(APIC_LVTPC, APIC_DM_NMI); - evntsel |= K7_EVNTSEL_ENABLE; - wrmsr(evntsel_msr, evntsel, 0); - - return 1; -} - -static void single_msr_stop_watchdog(void) -{ - struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk); - - wrmsr(wd->evntsel_msr, 0, 0); -} - -static int single_msr_reserve(void) -{ - if (!reserve_perfctr_nmi(wd_ops->perfctr)) - return 0; - - if (!reserve_evntsel_nmi(wd_ops->evntsel)) { - release_perfctr_nmi(wd_ops->perfctr); - return 0; - } - return 1; -} - -static void single_msr_unreserve(void) -{ - release_evntsel_nmi(wd_ops->evntsel); - release_perfctr_nmi(wd_ops->perfctr); -} - -static void __kprobes -single_msr_rearm(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz) -{ - /* start the cycle over again */ - write_watchdog_counter(wd->perfctr_msr, NULL, nmi_hz); -} - -static const struct wd_ops k7_wd_ops = { - .reserve = single_msr_reserve, - .unreserve = single_msr_unreserve, - .setup = setup_k7_watchdog, - .rearm = single_msr_rearm, - .stop = single_msr_stop_watchdog, - .perfctr = MSR_K7_PERFCTR0, - .evntsel = MSR_K7_EVNTSEL0, - .checkbit = 1ULL << 47, -}; - -/* - * Intel Model 6 (PPro+,P2,P3,P-M,Core1) - */ -#define P6_EVNTSEL0_ENABLE (1 << 22) -#define P6_EVNTSEL_INT (1 << 20) -#define P6_EVNTSEL_OS (1 << 17) -#define P6_EVNTSEL_USR (1 << 16) -#define P6_EVENT_CPU_CLOCKS_NOT_HALTED 0x79 -#define P6_NMI_EVENT P6_EVENT_CPU_CLOCKS_NOT_HALTED - -static int setup_p6_watchdog(unsigned nmi_hz) -{ - unsigned int perfctr_msr, evntsel_msr; - unsigned int evntsel; - struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk); - - perfctr_msr = wd_ops->perfctr; - evntsel_msr = wd_ops->evntsel; - - /* KVM doesn't implement this MSR */ - if (wrmsr_safe(perfctr_msr, 0, 0) < 0) - return 0; - - evntsel = P6_EVNTSEL_INT - | P6_EVNTSEL_OS - | P6_EVNTSEL_USR - | P6_NMI_EVENT; - - /* setup the timer */ - wrmsr(evntsel_msr, evntsel, 0); - nmi_hz = adjust_for_32bit_ctr(nmi_hz); - write_watchdog_counter32(perfctr_msr, "P6_PERFCTR0", nmi_hz); - - /* initialize the wd struct before enabling */ - wd->perfctr_msr = perfctr_msr; - wd->evntsel_msr = evntsel_msr; - wd->cccr_msr = 0; /* unused */ - - /* ok, everything is initialized, announce that we're set */ - cpu_nmi_set_wd_enabled(); - - apic_write(APIC_LVTPC, APIC_DM_NMI); - evntsel |= P6_EVNTSEL0_ENABLE; - wrmsr(evntsel_msr, evntsel, 0); - - return 1; -} - -static void __kprobes p6_rearm(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz) -{ - /* - * P6 based Pentium M need to re-unmask - * the apic vector but it doesn't hurt - * other P6 variant. - * ArchPerfom/Core Duo also needs this - */ - apic_write(APIC_LVTPC, APIC_DM_NMI); - - /* P6/ARCH_PERFMON has 32 bit counter write */ - write_watchdog_counter32(wd->perfctr_msr, NULL, nmi_hz); -} - -static const struct wd_ops p6_wd_ops = { - .reserve = single_msr_reserve, - .unreserve = single_msr_unreserve, - .setup = setup_p6_watchdog, - .rearm = p6_rearm, - .stop = single_msr_stop_watchdog, - .perfctr = MSR_P6_PERFCTR0, - .evntsel = MSR_P6_EVNTSEL0, - .checkbit = 1ULL << 39, -}; - -/* - * Intel P4 performance counters. - * By far the most complicated of all. - */ -#define MSR_P4_MISC_ENABLE_PERF_AVAIL (1 << 7) -#define P4_ESCR_EVENT_SELECT(N) ((N) << 25) -#define P4_ESCR_OS (1 << 3) -#define P4_ESCR_USR (1 << 2) -#define P4_CCCR_OVF_PMI0 (1 << 26) -#define P4_CCCR_OVF_PMI1 (1 << 27) -#define P4_CCCR_THRESHOLD(N) ((N) << 20) -#define P4_CCCR_COMPLEMENT (1 << 19) -#define P4_CCCR_COMPARE (1 << 18) -#define P4_CCCR_REQUIRED (3 << 16) -#define P4_CCCR_ESCR_SELECT(N) ((N) << 13) -#define P4_CCCR_ENABLE (1 << 12) -#define P4_CCCR_OVF (1 << 31) - -#define P4_CONTROLS 18 -static unsigned int p4_controls[18] = { - MSR_P4_BPU_CCCR0, - MSR_P4_BPU_CCCR1, - MSR_P4_BPU_CCCR2, - MSR_P4_BPU_CCCR3, - MSR_P4_MS_CCCR0, - MSR_P4_MS_CCCR1, - MSR_P4_MS_CCCR2, - MSR_P4_MS_CCCR3, - MSR_P4_FLAME_CCCR0, - MSR_P4_FLAME_CCCR1, - MSR_P4_FLAME_CCCR2, - MSR_P4_FLAME_CCCR3, - MSR_P4_IQ_CCCR0, - MSR_P4_IQ_CCCR1, - MSR_P4_IQ_CCCR2, - MSR_P4_IQ_CCCR3, - MSR_P4_IQ_CCCR4, - MSR_P4_IQ_CCCR5, -}; -/* - * Set up IQ_COUNTER0 to behave like a clock, by having IQ_CCCR0 filter - * CRU_ESCR0 (with any non-null event selector) through a complemented - * max threshold. [IA32-Vol3, Section 14.9.9] - */ -static int setup_p4_watchdog(unsigned nmi_hz) -{ - unsigned int perfctr_msr, evntsel_msr, cccr_msr; - unsigned int evntsel, cccr_val; - unsigned int misc_enable, dummy; - unsigned int ht_num; - struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk); - - rdmsr(MSR_IA32_MISC_ENABLE, misc_enable, dummy); - if (!(misc_enable & MSR_P4_MISC_ENABLE_PERF_AVAIL)) - return 0; - -#ifdef CONFIG_SMP - /* detect which hyperthread we are on */ - if (smp_num_siblings == 2) { - unsigned int ebx, apicid; - - ebx = cpuid_ebx(1); - apicid = (ebx >> 24) & 0xff; - ht_num = apicid & 1; - } else -#endif - ht_num = 0; - - /* - * performance counters are shared resources - * assign each hyperthread its own set - * (re-use the ESCR0 register, seems safe - * and keeps the cccr_val the same) - */ - if (!ht_num) { - /* logical cpu 0 */ - perfctr_msr = MSR_P4_IQ_PERFCTR0; - evntsel_msr = MSR_P4_CRU_ESCR0; - cccr_msr = MSR_P4_IQ_CCCR0; - cccr_val = P4_CCCR_OVF_PMI0 | P4_CCCR_ESCR_SELECT(4); - - /* - * If we're on the kdump kernel or other situation, we may - * still have other performance counter registers set to - * interrupt and they'll keep interrupting forever because - * of the P4_CCCR_OVF quirk. So we need to ACK all the - * pending interrupts and disable all the registers here, - * before reenabling the NMI delivery. Refer to p4_rearm() - * about the P4_CCCR_OVF quirk. - */ - if (reset_devices) { - unsigned int low, high; - int i; - - for (i = 0; i < P4_CONTROLS; i++) { - rdmsr(p4_controls[i], low, high); - low &= ~(P4_CCCR_ENABLE | P4_CCCR_OVF); - wrmsr(p4_controls[i], low, high); - } - } - } else { - /* logical cpu 1 */ - perfctr_msr = MSR_P4_IQ_PERFCTR1; - evntsel_msr = MSR_P4_CRU_ESCR0; - cccr_msr = MSR_P4_IQ_CCCR1; - - /* Pentium 4 D processors don't support P4_CCCR_OVF_PMI1 */ - if (boot_cpu_data.x86_model == 4 && boot_cpu_data.x86_mask == 4) - cccr_val = P4_CCCR_OVF_PMI0; - else - cccr_val = P4_CCCR_OVF_PMI1; - cccr_val |= P4_CCCR_ESCR_SELECT(4); - } - - evntsel = P4_ESCR_EVENT_SELECT(0x3F) - | P4_ESCR_OS - | P4_ESCR_USR; - - cccr_val |= P4_CCCR_THRESHOLD(15) - | P4_CCCR_COMPLEMENT - | P4_CCCR_COMPARE - | P4_CCCR_REQUIRED; - - wrmsr(evntsel_msr, evntsel, 0); - wrmsr(cccr_msr, cccr_val, 0); - write_watchdog_counter(perfctr_msr, "P4_IQ_COUNTER0", nmi_hz); - - wd->perfctr_msr = perfctr_msr; - wd->evntsel_msr = evntsel_msr; - wd->cccr_msr = cccr_msr; - - /* ok, everything is initialized, announce that we're set */ - cpu_nmi_set_wd_enabled(); - - apic_write(APIC_LVTPC, APIC_DM_NMI); - cccr_val |= P4_CCCR_ENABLE; - wrmsr(cccr_msr, cccr_val, 0); - return 1; -} - -static void stop_p4_watchdog(void) -{ - struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk); - wrmsr(wd->cccr_msr, 0, 0); - wrmsr(wd->evntsel_msr, 0, 0); -} - -static int p4_reserve(void) -{ - if (!reserve_perfctr_nmi(MSR_P4_IQ_PERFCTR0)) - return 0; -#ifdef CONFIG_SMP - if (smp_num_siblings > 1 && !reserve_perfctr_nmi(MSR_P4_IQ_PERFCTR1)) - goto fail1; -#endif - if (!reserve_evntsel_nmi(MSR_P4_CRU_ESCR0)) - goto fail2; - /* RED-PEN why is ESCR1 not reserved here? */ - return 1; - fail2: -#ifdef CONFIG_SMP - if (smp_num_siblings > 1) - release_perfctr_nmi(MSR_P4_IQ_PERFCTR1); - fail1: -#endif - release_perfctr_nmi(MSR_P4_IQ_PERFCTR0); - return 0; -} - -static void p4_unreserve(void) -{ -#ifdef CONFIG_SMP - if (smp_num_siblings > 1) - release_perfctr_nmi(MSR_P4_IQ_PERFCTR1); -#endif - release_evntsel_nmi(MSR_P4_CRU_ESCR0); - release_perfctr_nmi(MSR_P4_IQ_PERFCTR0); -} - -static void __kprobes p4_rearm(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz) -{ - unsigned dummy; - /* - * P4 quirks: - * - An overflown perfctr will assert its interrupt - * until the OVF flag in its CCCR is cleared. - * - LVTPC is masked on interrupt and must be - * unmasked by the LVTPC handler. - */ - rdmsrl(wd->cccr_msr, dummy); - dummy &= ~P4_CCCR_OVF; - wrmsrl(wd->cccr_msr, dummy); - apic_write(APIC_LVTPC, APIC_DM_NMI); - /* start the cycle over again */ - write_watchdog_counter(wd->perfctr_msr, NULL, nmi_hz); -} - -static const struct wd_ops p4_wd_ops = { - .reserve = p4_reserve, - .unreserve = p4_unreserve, - .setup = setup_p4_watchdog, - .rearm = p4_rearm, - .stop = stop_p4_watchdog, - /* RED-PEN this is wrong for the other sibling */ - .perfctr = MSR_P4_BPU_PERFCTR0, - .evntsel = MSR_P4_BSU_ESCR0, - .checkbit = 1ULL << 39, -}; - -/* - * Watchdog using the Intel architected PerfMon. - * Used for Core2 and hopefully all future Intel CPUs. - */ -#define ARCH_PERFMON_NMI_EVENT_SEL ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL -#define ARCH_PERFMON_NMI_EVENT_UMASK ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK - -static struct wd_ops intel_arch_wd_ops; - -static int setup_intel_arch_watchdog(unsigned nmi_hz) -{ - unsigned int ebx; - union cpuid10_eax eax; - unsigned int unused; - unsigned int perfctr_msr, evntsel_msr; - unsigned int evntsel; - struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk); - - /* - * Check whether the Architectural PerfMon supports - * Unhalted Core Cycles Event or not. - * NOTE: Corresponding bit = 0 in ebx indicates event present. - */ - cpuid(10, &(eax.full), &ebx, &unused, &unused); - if ((eax.split.mask_length < - (ARCH_PERFMON_UNHALTED_CORE_CYCLES_INDEX+1)) || - (ebx & ARCH_PERFMON_UNHALTED_CORE_CYCLES_PRESENT)) - return 0; - - perfctr_msr = wd_ops->perfctr; - evntsel_msr = wd_ops->evntsel; - - wrmsrl(perfctr_msr, 0UL); - - evntsel = ARCH_PERFMON_EVENTSEL_INT - | ARCH_PERFMON_EVENTSEL_OS - | ARCH_PERFMON_EVENTSEL_USR - | ARCH_PERFMON_NMI_EVENT_SEL - | ARCH_PERFMON_NMI_EVENT_UMASK; - - /* setup the timer */ - wrmsr(evntsel_msr, evntsel, 0); - nmi_hz = adjust_for_32bit_ctr(nmi_hz); - write_watchdog_counter32(perfctr_msr, "INTEL_ARCH_PERFCTR0", nmi_hz); - - wd->perfctr_msr = perfctr_msr; - wd->evntsel_msr = evntsel_msr; - wd->cccr_msr = 0; /* unused */ - - /* ok, everything is initialized, announce that we're set */ - cpu_nmi_set_wd_enabled(); - - apic_write(APIC_LVTPC, APIC_DM_NMI); - evntsel |= ARCH_PERFMON_EVENTSEL_ENABLE; - wrmsr(evntsel_msr, evntsel, 0); - intel_arch_wd_ops.checkbit = 1ULL << (eax.split.bit_width - 1); - return 1; -} - -static struct wd_ops intel_arch_wd_ops __read_mostly = { - .reserve = single_msr_reserve, - .unreserve = single_msr_unreserve, - .setup = setup_intel_arch_watchdog, - .rearm = p6_rearm, - .stop = single_msr_stop_watchdog, - .perfctr = MSR_ARCH_PERFMON_PERFCTR1, - .evntsel = MSR_ARCH_PERFMON_EVENTSEL1, -}; - -static void probe_nmi_watchdog(void) -{ - switch (boot_cpu_data.x86_vendor) { - case X86_VENDOR_AMD: - if (boot_cpu_data.x86 == 6 || - (boot_cpu_data.x86 >= 0xf && boot_cpu_data.x86 <= 0x15)) - wd_ops = &k7_wd_ops; - return; - case X86_VENDOR_INTEL: - /* Work around where perfctr1 doesn't have a working enable - * bit as described in the following errata: - * AE49 Core Duo and Intel Core Solo 65 nm - * AN49 Intel Pentium Dual-Core - * AF49 Dual-Core Intel Xeon Processor LV - */ - if ((boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model == 14) || - ((boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model == 15 && - boot_cpu_data.x86_mask == 4))) { - intel_arch_wd_ops.perfctr = MSR_ARCH_PERFMON_PERFCTR0; - intel_arch_wd_ops.evntsel = MSR_ARCH_PERFMON_EVENTSEL0; - } - if (cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) { - wd_ops = &intel_arch_wd_ops; - break; - } - switch (boot_cpu_data.x86) { - case 6: - if (boot_cpu_data.x86_model > 13) - return; - - wd_ops = &p6_wd_ops; - break; - case 15: - wd_ops = &p4_wd_ops; - break; - default: - return; - } - break; - } -} - -/* Interface to nmi.c */ - -int lapic_watchdog_init(unsigned nmi_hz) -{ - if (!wd_ops) { - probe_nmi_watchdog(); - if (!wd_ops) { - printk(KERN_INFO "NMI watchdog: CPU not supported\n"); - return -1; - } - - if (!wd_ops->reserve()) { - printk(KERN_ERR - "NMI watchdog: cannot reserve perfctrs\n"); - return -1; - } - } - - if (!(wd_ops->setup(nmi_hz))) { - printk(KERN_ERR "Cannot setup NMI watchdog on CPU %d\n", - raw_smp_processor_id()); - return -1; - } - - return 0; -} - -void lapic_watchdog_stop(void) -{ - if (wd_ops) - wd_ops->stop(); -} - -unsigned lapic_adjust_nmi_hz(unsigned hz) -{ - struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk); - if (wd->perfctr_msr == MSR_P6_PERFCTR0 || - wd->perfctr_msr == MSR_ARCH_PERFMON_PERFCTR1) - hz = adjust_for_32bit_ctr(hz); - return hz; -} - -int __kprobes lapic_wd_event(unsigned nmi_hz) -{ - struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk); - u64 ctr; - - rdmsrl(wd->perfctr_msr, ctr); - if (ctr & wd_ops->checkbit) /* perfctr still running? */ - return 0; - - wd_ops->rearm(wd, nmi_hz); - return 1; -} diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 083e99d1b7df..f0a0624eea55 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -316,12 +316,6 @@ notrace static void __cpuinit start_secondary(void *unused) */ check_tsc_sync_target(); - if (nmi_watchdog == NMI_IO_APIC) { - legacy_pic->mask(0); - enable_NMI_through_LVT0(); - legacy_pic->unmask(0); - } - /* This must be done before setting cpu_online_mask */ set_cpu_sibling_map(raw_smp_processor_id()); wmb(); @@ -1061,8 +1055,6 @@ static int __init smp_sanity_check(unsigned max_cpus) printk(KERN_INFO "SMP mode deactivated.\n"); smpboot_clear_io_apic(); - localise_nmi_watchdog(); - connect_bsp_APIC(); setup_local_APIC(); end_local_APIC_setup(); @@ -1196,7 +1188,6 @@ void __init native_smp_cpus_done(unsigned int max_cpus) #ifdef CONFIG_X86_IO_APIC setup_ioapic_dest(); #endif - check_nmi_watchdog(); mtrr_aps_init(); } @@ -1341,8 +1332,6 @@ int native_cpu_disable(void) if (cpu == 0) return -EBUSY; - if (nmi_watchdog == NMI_LOCAL_APIC) - stop_apic_nmi_watchdog(NULL); clear_local_APIC(); cpu_disable_common(); diff --git a/arch/x86/kernel/time.c b/arch/x86/kernel/time.c index fb5cc5e14cfa..25a28a245937 100644 --- a/arch/x86/kernel/time.c +++ b/arch/x86/kernel/time.c @@ -22,10 +22,6 @@ #include #include -#if defined(CONFIG_X86_32) && defined(CONFIG_X86_IO_APIC) -int timer_ack; -#endif - #ifdef CONFIG_X86_64 volatile unsigned long __jiffies __section_jiffies = INITIAL_JIFFIES; #endif @@ -63,20 +59,6 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id) /* Keep nmi watchdog up to date */ inc_irq_stat(irq0_irqs); - /* Optimized out for !IO_APIC and x86_64 */ - if (timer_ack) { - /* - * Subtle, when I/O APICs are used we have to ack timer IRQ - * manually to deassert NMI lines for the watchdog if run - * on an 82489DX-based system. - */ - raw_spin_lock(&i8259A_lock); - outb(0x0c, PIC_MASTER_OCW3); - /* Ack the IRQ; AEOI will end it automatically. */ - inb(PIC_MASTER_POLL); - raw_spin_unlock(&i8259A_lock); - } - global_clock_event->event_handler(global_clock_event); /* MCA bus quirk: Acknowledge irq0 by setting bit 7 in port 0x61 */ diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index db30d9cb9dd6..f02c179c2552 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -437,14 +437,12 @@ do_nmi(struct pt_regs *regs, long error_code) void stop_nmi(void) { - acpi_nmi_disable(); ignore_nmis++; } void restart_nmi(void) { ignore_nmis--; - acpi_nmi_enable(); } /* May run on IST stack. */ diff --git a/arch/x86/oprofile/nmi_timer_int.c b/arch/x86/oprofile/nmi_timer_int.c index e3ecb71b5790..0636dd93cef8 100644 --- a/arch/x86/oprofile/nmi_timer_int.c +++ b/arch/x86/oprofile/nmi_timer_int.c @@ -58,9 +58,6 @@ static void timer_stop(void) int __init op_nmi_timer_init(struct oprofile_operations *ops) { - if ((nmi_watchdog != NMI_IO_APIC) || (atomic_read(&nmi_active) <= 0)) - return -ENODEV; - ops->start = timer_start; ops->stop = timer_stop; ops->cpu_type = "timer"; diff --git a/drivers/acpi/acpica/nsinit.c b/drivers/acpi/acpica/nsinit.c index 660a2728908d..0cac7ec0d2ec 100644 --- a/drivers/acpi/acpica/nsinit.c +++ b/drivers/acpi/acpica/nsinit.c @@ -577,9 +577,7 @@ acpi_ns_init_one_device(acpi_handle obj_handle, * as possible (without an NMI being received in the middle of * this) - so disable NMIs and initialize the device: */ - acpi_nmi_disable(); status = acpi_ns_evaluate(info); - acpi_nmi_enable(); if (ACPI_SUCCESS(status)) { walk_info->num_INI++; diff --git a/drivers/watchdog/hpwdt.c b/drivers/watchdog/hpwdt.c index 3d77116e4634..c19f4a20794a 100644 --- a/drivers/watchdog/hpwdt.c +++ b/drivers/watchdog/hpwdt.c @@ -649,12 +649,7 @@ static void __devinit hpwdt_check_nmi_decoding(struct pci_dev *dev) * If nmi_watchdog is turned off then we can turn on * our nmi decoding capability. */ - if (!nmi_watchdog_active()) - hpwdt_nmi_decoding = 1; - else - dev_warn(&dev->dev, "NMI decoding is disabled. To enable this " - "functionality you must reboot with nmi_watchdog=0 " - "and load the hpwdt driver with priority=1.\n"); + hpwdt_nmi_decoding = 1; } #else static void __devinit hpwdt_check_nmi_decoding(struct pci_dev *dev) diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 0cb3e5c246d0..1c451e6ecc17 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -25,8 +25,6 @@ static inline void touch_nmi_watchdog(void) #else extern void touch_nmi_watchdog(void); #endif -static inline void acpi_nmi_disable(void) { } -static inline void acpi_nmi_enable(void) { } /* * Create trigger_all_cpu_backtrace() out of the arch-provided -- cgit v1.2.3 From 8191c9f69202d4dbc66063cb92059b8a58640d34 Mon Sep 17 00:00:00 2001 From: Dimitri Sivanich Date: Tue, 16 Nov 2010 16:23:52 -0600 Subject: x86: UV: Address interrupt/IO port operation conflict This patch for SGI UV systems addresses a problem whereby interrupt transactions being looped back from a local IOH, through the hub to a local CPU can (erroneously) conflict with IO port operations and other transactions. To workaound this we set a high bit in the APIC IDs used for interrupts. This bit appears to be ignored by the sockets, but it avoids the conflict in the hub. Signed-off-by: Dimitri Sivanich LKML-Reference: <20101116222352.GA8155@sgi.com> Signed-off-by: Ingo Molnar ___ arch/x86/include/asm/uv/uv_hub.h | 4 ++++ arch/x86/include/asm/uv/uv_mmrs.h | 19 ++++++++++++++++++- arch/x86/kernel/apic/x2apic_uv_x.c | 25 +++++++++++++++++++++++-- arch/x86/platform/uv/tlb_uv.c | 2 +- arch/x86/platform/uv/uv_time.c | 4 +++- 5 files changed, 49 insertions(+), 5 deletions(-) --- arch/x86/include/asm/uv/uv_hub.h | 4 ++++ arch/x86/include/asm/uv/uv_mmrs.h | 19 ++++++++++++++++++- arch/x86/kernel/apic/x2apic_uv_x.c | 25 +++++++++++++++++++++++-- arch/x86/platform/uv/tlb_uv.c | 2 +- arch/x86/platform/uv/uv_time.c | 4 +++- 5 files changed, 49 insertions(+), 5 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h index e969f691cbfd..a501741c2335 100644 --- a/arch/x86/include/asm/uv/uv_hub.h +++ b/arch/x86/include/asm/uv/uv_hub.h @@ -199,6 +199,8 @@ union uvh_apicid { #define UVH_APICID 0x002D0E00L #define UV_APIC_PNODE_SHIFT 6 +#define UV_APICID_HIBIT_MASK 0xffff0000 + /* Local Bus from cpu's perspective */ #define LOCAL_BUS_BASE 0x1c00000 #define LOCAL_BUS_SIZE (4 * 1024 * 1024) @@ -491,8 +493,10 @@ static inline void uv_set_cpu_scir_bits(int cpu, unsigned char value) } } +extern unsigned int uv_apicid_hibits; static unsigned long uv_hub_ipi_value(int apicid, int vector, int mode) { + apicid |= uv_apicid_hibits; return (1UL << UVH_IPI_INT_SEND_SHFT) | ((apicid) << UVH_IPI_INT_APIC_ID_SHFT) | (mode << UVH_IPI_INT_DELIVERY_MODE_SHFT) | diff --git a/arch/x86/include/asm/uv/uv_mmrs.h b/arch/x86/include/asm/uv/uv_mmrs.h index 6d90adf4428a..20cafeac7455 100644 --- a/arch/x86/include/asm/uv/uv_mmrs.h +++ b/arch/x86/include/asm/uv/uv_mmrs.h @@ -5,7 +5,7 @@ * * SGI UV MMR definitions * - * Copyright (C) 2007-2008 Silicon Graphics, Inc. All rights reserved. + * Copyright (C) 2007-2010 Silicon Graphics, Inc. All rights reserved. */ #ifndef _ASM_X86_UV_UV_MMRS_H @@ -753,6 +753,23 @@ union uvh_lb_bau_sb_descriptor_base_u { } s; }; +/* ========================================================================= */ +/* UVH_LB_TARGET_PHYSICAL_APIC_ID_MASK */ +/* ========================================================================= */ +#define UVH_LB_TARGET_PHYSICAL_APIC_ID_MASK 0x320130UL +#define UVH_LB_TARGET_PHYSICAL_APIC_ID_MASK_32 0x009f0 + +#define UVH_LB_TARGET_PHYSICAL_APIC_ID_MASK_BIT_ENABLES_SHFT 0 +#define UVH_LB_TARGET_PHYSICAL_APIC_ID_MASK_BIT_ENABLES_MASK 0x00000000ffffffffUL + +union uvh_lb_target_physical_apic_id_mask_u { + unsigned long v; + struct uvh_lb_target_physical_apic_id_mask_s { + unsigned long bit_enables : 32; /* RW */ + unsigned long rsvd_32_63 : 32; /* */ + } s; +}; + /* ========================================================================= */ /* UVH_NODE_ID */ /* ========================================================================= */ diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index 194539aea175..c1c52c341f40 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -44,6 +44,8 @@ static u64 gru_start_paddr, gru_end_paddr; static union uvh_apicid uvh_apicid; int uv_min_hub_revision_id; EXPORT_SYMBOL_GPL(uv_min_hub_revision_id); +unsigned int uv_apicid_hibits; +EXPORT_SYMBOL_GPL(uv_apicid_hibits); static DEFINE_SPINLOCK(uv_nmi_lock); static inline bool is_GRU_range(u64 start, u64 end) @@ -85,6 +87,23 @@ static void __init early_get_apic_pnode_shift(void) uvh_apicid.s.pnode_shift = UV_APIC_PNODE_SHIFT; } +/* + * Add an extra bit as dictated by bios to the destination apicid of + * interrupts potentially passing through the UV HUB. This prevents + * a deadlock between interrupts and IO port operations. + */ +static void __init uv_set_apicid_hibit(void) +{ + union uvh_lb_target_physical_apic_id_mask_u apicid_mask; + unsigned long *mmr; + + mmr = early_ioremap(UV_LOCAL_MMR_BASE | + UVH_LB_TARGET_PHYSICAL_APIC_ID_MASK, sizeof(*mmr)); + apicid_mask.v = *mmr; + early_iounmap(mmr, sizeof(*mmr)); + uv_apicid_hibits = apicid_mask.s.bit_enables & UV_APICID_HIBIT_MASK; +} + static int __init uv_acpi_madt_oem_check(char *oem_id, char *oem_table_id) { int nodeid; @@ -102,6 +121,7 @@ static int __init uv_acpi_madt_oem_check(char *oem_id, char *oem_table_id) __get_cpu_var(x2apic_extra_bits) = nodeid << (uvh_apicid.s.pnode_shift - 1); uv_system_type = UV_NON_UNIQUE_APIC; + uv_set_apicid_hibit(); return 1; } } @@ -155,6 +175,7 @@ static int __cpuinit uv_wakeup_secondary(int phys_apicid, unsigned long start_ri int pnode; pnode = uv_apicid_to_pnode(phys_apicid); + phys_apicid |= uv_apicid_hibits; val = (1UL << UVH_IPI_INT_SEND_SHFT) | (phys_apicid << UVH_IPI_INT_APIC_ID_SHFT) | ((start_rip << UVH_IPI_INT_VECTOR_SHFT) >> 12) | @@ -236,7 +257,7 @@ static unsigned int uv_cpu_mask_to_apicid(const struct cpumask *cpumask) int cpu = cpumask_first(cpumask); if ((unsigned)cpu < nr_cpu_ids) - return per_cpu(x86_cpu_to_apicid, cpu); + return per_cpu(x86_cpu_to_apicid, cpu) | uv_apicid_hibits; else return BAD_APICID; } @@ -255,7 +276,7 @@ uv_cpu_mask_to_apicid_and(const struct cpumask *cpumask, if (cpumask_test_cpu(cpu, cpu_online_mask)) break; } - return per_cpu(x86_cpu_to_apicid, cpu); + return per_cpu(x86_cpu_to_apicid, cpu) | uv_apicid_hibits; } static unsigned int x2apic_get_apic_id(unsigned long x) diff --git a/arch/x86/platform/uv/tlb_uv.c b/arch/x86/platform/uv/tlb_uv.c index a318194002b5..ba9caa808a9c 100644 --- a/arch/x86/platform/uv/tlb_uv.c +++ b/arch/x86/platform/uv/tlb_uv.c @@ -1455,7 +1455,7 @@ static void __init uv_init_uvhub(int uvhub, int vector) * the below initialization can't be in firmware because the * messaging IRQ will be determined by the OS */ - apicid = uvhub_to_first_apicid(uvhub); + apicid = uvhub_to_first_apicid(uvhub) | uv_apicid_hibits; uv_write_global_mmr64(pnode, UVH_BAU_DATA_CONFIG, ((apicid << 32) | vector)); } diff --git a/arch/x86/platform/uv/uv_time.c b/arch/x86/platform/uv/uv_time.c index 56e421bc379b..9daf5d1af9f1 100644 --- a/arch/x86/platform/uv/uv_time.c +++ b/arch/x86/platform/uv/uv_time.c @@ -89,6 +89,7 @@ static void uv_rtc_send_IPI(int cpu) apicid = cpu_physical_id(cpu); pnode = uv_apicid_to_pnode(apicid); + apicid |= uv_apicid_hibits; val = (1UL << UVH_IPI_INT_SEND_SHFT) | (apicid << UVH_IPI_INT_APIC_ID_SHFT) | (X86_PLATFORM_IPI_VECTOR << UVH_IPI_INT_VECTOR_SHFT); @@ -107,6 +108,7 @@ static int uv_intr_pending(int pnode) static int uv_setup_intr(int cpu, u64 expires) { u64 val; + unsigned long apicid = cpu_physical_id(cpu) | uv_apicid_hibits; int pnode = uv_cpu_to_pnode(cpu); uv_write_global_mmr64(pnode, UVH_RTC1_INT_CONFIG, @@ -117,7 +119,7 @@ static int uv_setup_intr(int cpu, u64 expires) UVH_EVENT_OCCURRED0_RTC1_MASK); val = (X86_PLATFORM_IPI_VECTOR << UVH_RTC1_INT_CONFIG_VECTOR_SHFT) | - ((u64)cpu_physical_id(cpu) << UVH_RTC1_INT_CONFIG_APIC_ID_SHFT); + ((u64)apicid << UVH_RTC1_INT_CONFIG_APIC_ID_SHFT); /* Set configuration */ uv_write_global_mmr64(pnode, UVH_RTC1_INT_CONFIG, val); -- cgit v1.2.3 From 5bd5a452662bc37c54fb6828db1a3faf87e6511c Mon Sep 17 00:00:00 2001 From: Matthieu Castet Date: Tue, 16 Nov 2010 22:31:26 +0100 Subject: x86: Add NX protection for kernel data This patch expands functionality of CONFIG_DEBUG_RODATA to set main (static) kernel data area as NX. The following steps are taken to achieve this: 1. Linker script is adjusted so .text always starts and ends on a page bound 2. Linker script is adjusted so .rodata always start and end on a page boundary 3. NX is set for all pages from _etext through _end in mark_rodata_ro. 4. free_init_pages() sets released memory NX in arch/x86/mm/init.c 5. bios rom is set to x when pcibios is used. The results of patch application may be observed in the diff of kernel page table dumps: pcibios: -- data_nx_pt_before.txt 2009-10-13 07:48:59.000000000 -0400 ++ data_nx_pt_after.txt 2009-10-13 07:26:46.000000000 -0400 0x00000000-0xc0000000 3G pmd ---[ Kernel Mapping ]--- -0xc0000000-0xc0100000 1M RW GLB x pte +0xc0000000-0xc00a0000 640K RW GLB NX pte +0xc00a0000-0xc0100000 384K RW GLB x pte -0xc0100000-0xc03d7000 2908K ro GLB x pte +0xc0100000-0xc0318000 2144K ro GLB x pte +0xc0318000-0xc03d7000 764K ro GLB NX pte -0xc03d7000-0xc0600000 2212K RW GLB x pte +0xc03d7000-0xc0600000 2212K RW GLB NX pte 0xc0600000-0xf7a00000 884M RW PSE GLB NX pmd 0xf7a00000-0xf7bfe000 2040K RW GLB NX pte 0xf7bfe000-0xf7c00000 8K pte No pcibios: -- data_nx_pt_before.txt 2009-10-13 07:48:59.000000000 -0400 ++ data_nx_pt_after.txt 2009-10-13 07:26:46.000000000 -0400 0x00000000-0xc0000000 3G pmd ---[ Kernel Mapping ]--- -0xc0000000-0xc0100000 1M RW GLB x pte +0xc0000000-0xc0100000 1M RW GLB NX pte -0xc0100000-0xc03d7000 2908K ro GLB x pte +0xc0100000-0xc0318000 2144K ro GLB x pte +0xc0318000-0xc03d7000 764K ro GLB NX pte -0xc03d7000-0xc0600000 2212K RW GLB x pte +0xc03d7000-0xc0600000 2212K RW GLB NX pte 0xc0600000-0xf7a00000 884M RW PSE GLB NX pmd 0xf7a00000-0xf7bfe000 2040K RW GLB NX pte 0xf7bfe000-0xf7c00000 8K pte The patch has been originally developed for Linux 2.6.34-rc2 x86 by Siarhei Liakh and Xuxian Jiang . -v1: initial patch for 2.6.30 -v2: patch for 2.6.31-rc7 -v3: moved all code into arch/x86, adjusted credits -v4: fixed ifdef, removed credits from CREDITS -v5: fixed an address calculation bug in mark_nxdata_nx() -v6: added acked-by and PT dump diff to commit log -v7: minor adjustments for -tip -v8: rework with the merge of "Set first MB as RW+NX" Signed-off-by: Siarhei Liakh Signed-off-by: Xuxian Jiang Signed-off-by: Matthieu CASTET Cc: Arjan van de Ven Cc: James Morris Cc: Andi Kleen Cc: Rusty Russell Cc: Stephen Rothwell Cc: Dave Jones Cc: Kees Cook Cc: Linus Torvalds LKML-Reference: <4CE2F82E.60601@free.fr> [ minor cleanliness edits ] Signed-off-by: Ingo Molnar --- arch/x86/include/asm/pci.h | 1 + arch/x86/kernel/vmlinux.lds.S | 8 ++++++-- arch/x86/mm/init.c | 3 ++- arch/x86/mm/init_32.c | 20 +++++++++++++++++++- arch/x86/mm/init_64.c | 3 ++- arch/x86/mm/pageattr.c | 5 ++++- arch/x86/pci/pcbios.c | 23 +++++++++++++++++++++++ 7 files changed, 57 insertions(+), 6 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/pci.h b/arch/x86/include/asm/pci.h index ca0437c714b2..676129229630 100644 --- a/arch/x86/include/asm/pci.h +++ b/arch/x86/include/asm/pci.h @@ -65,6 +65,7 @@ extern unsigned long pci_mem_start; #define PCIBIOS_MIN_CARDBUS_IO 0x4000 +extern int pcibios_enabled; void pcibios_config_init(void); struct pci_bus *pcibios_scan_root(int bus); diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S index e03530aebfd0..bf4700755184 100644 --- a/arch/x86/kernel/vmlinux.lds.S +++ b/arch/x86/kernel/vmlinux.lds.S @@ -69,7 +69,7 @@ jiffies_64 = jiffies; PHDRS { text PT_LOAD FLAGS(5); /* R_E */ - data PT_LOAD FLAGS(7); /* RWE */ + data PT_LOAD FLAGS(6); /* RW_ */ #ifdef CONFIG_X86_64 user PT_LOAD FLAGS(5); /* R_E */ #ifdef CONFIG_SMP @@ -116,6 +116,10 @@ SECTIONS EXCEPTION_TABLE(16) :text = 0x9090 +#if defined(CONFIG_DEBUG_RODATA) + /* .text should occupy whole number of pages */ + . = ALIGN(PAGE_SIZE); +#endif X64_ALIGN_DEBUG_RODATA_BEGIN RO_DATA(PAGE_SIZE) X64_ALIGN_DEBUG_RODATA_END @@ -335,7 +339,7 @@ SECTIONS __bss_start = .; *(.bss..page_aligned) *(.bss) - . = ALIGN(4); + . = ALIGN(PAGE_SIZE); __bss_stop = .; } diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c index c0e28a13de7d..947f42abe820 100644 --- a/arch/x86/mm/init.c +++ b/arch/x86/mm/init.c @@ -364,8 +364,9 @@ void free_init_pages(char *what, unsigned long begin, unsigned long end) /* * We just marked the kernel text read only above, now that * we are going to free part of that, we need to make that - * writeable first. + * writeable and non-executable first. */ + set_memory_nx(begin, (end - begin) >> PAGE_SHIFT); set_memory_rw(begin, (end - begin) >> PAGE_SHIFT); printk(KERN_INFO "Freeing %s: %luk freed\n", what, (end - begin) >> 10); diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index 0e969f9f401b..f89b5bb4e93f 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -226,7 +226,7 @@ page_table_range_init(unsigned long start, unsigned long end, pgd_t *pgd_base) static inline int is_kernel_text(unsigned long addr) { - if (addr >= PAGE_OFFSET && addr <= (unsigned long)__init_end) + if (addr >= (unsigned long)_text && addr <= (unsigned long)__init_end) return 1; return 0; } @@ -912,6 +912,23 @@ void set_kernel_text_ro(void) set_pages_ro(virt_to_page(start), size >> PAGE_SHIFT); } +static void mark_nxdata_nx(void) +{ + /* + * When this called, init has already been executed and released, + * so everything past _etext sould be NX. + */ + unsigned long start = PFN_ALIGN(_etext); + /* + * This comes from is_kernel_text upper limit. Also HPAGE where used: + */ + unsigned long size = (((unsigned long)__init_end + HPAGE_SIZE) & HPAGE_MASK) - start; + + if (__supported_pte_mask & _PAGE_NX) + printk(KERN_INFO "NX-protecting the kernel data: %luk\n", size >> 10); + set_pages_nx(virt_to_page(start), size >> PAGE_SHIFT); +} + void mark_rodata_ro(void) { unsigned long start = PFN_ALIGN(_text); @@ -946,6 +963,7 @@ void mark_rodata_ro(void) printk(KERN_INFO "Testing CPA: write protecting again\n"); set_pages_ro(virt_to_page(start), size >> PAGE_SHIFT); #endif + mark_nxdata_nx(); } #endif diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index 71a59296af80..ce59c05cae12 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -788,6 +788,7 @@ void mark_rodata_ro(void) unsigned long rodata_start = ((unsigned long)__start_rodata + PAGE_SIZE - 1) & PAGE_MASK; unsigned long end = (unsigned long) &__end_rodata_hpage_align; + unsigned long kernel_end = (((unsigned long)&__init_end + HPAGE_SIZE) & HPAGE_MASK); unsigned long text_end = PAGE_ALIGN((unsigned long) &__stop___ex_table); unsigned long rodata_end = PAGE_ALIGN((unsigned long) &__end_rodata); unsigned long data_start = (unsigned long) &_sdata; @@ -802,7 +803,7 @@ void mark_rodata_ro(void) * The rodata section (but not the kernel text!) should also be * not-executable. */ - set_memory_nx(rodata_start, (end - rodata_start) >> PAGE_SHIFT); + set_memory_nx(rodata_start, (kernel_end - rodata_start) >> PAGE_SHIFT); rodata_test(); diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c index 6f2a6b6deb6b..8b830ca14ac4 100644 --- a/arch/x86/mm/pageattr.c +++ b/arch/x86/mm/pageattr.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #include @@ -261,8 +262,10 @@ static inline pgprot_t static_protections(pgprot_t prot, unsigned long address, * The BIOS area between 640k and 1Mb needs to be executable for * PCI BIOS based config access (CONFIG_PCI_GOBIOS) support. */ - if (within(pfn, BIOS_BEGIN >> PAGE_SHIFT, BIOS_END >> PAGE_SHIFT)) +#ifdef CONFIG_PCI_BIOS + if (pcibios_enabled && within(pfn, BIOS_BEGIN >> PAGE_SHIFT, BIOS_END >> PAGE_SHIFT)) pgprot_val(forbidden) |= _PAGE_NX; +#endif /* * The kernel text needs to be executable for obvious reasons diff --git a/arch/x86/pci/pcbios.c b/arch/x86/pci/pcbios.c index 2492d165096a..a5f7d0d63de0 100644 --- a/arch/x86/pci/pcbios.c +++ b/arch/x86/pci/pcbios.c @@ -9,6 +9,7 @@ #include #include #include +#include /* BIOS32 signature: "_32_" */ #define BIOS32_SIGNATURE (('_' << 0) + ('3' << 8) + ('2' << 16) + ('_' << 24)) @@ -25,6 +26,27 @@ #define PCIBIOS_HW_TYPE1_SPEC 0x10 #define PCIBIOS_HW_TYPE2_SPEC 0x20 +int pcibios_enabled; + +/* According to the BIOS specification at: + * http://members.datafast.net.au/dft0802/specs/bios21.pdf, we could + * restrict the x zone to some pages and make it ro. But this may be + * broken on some bios, complex to handle with static_protections. + * We could make the 0xe0000-0x100000 range rox, but this can break + * some ISA mapping. + * + * So we let's an rw and x hole when pcibios is used. This shouldn't + * happen for modern system with mmconfig, and if you don't want it + * you could disable pcibios... + */ +static inline void set_bios_x(void) +{ + pcibios_enabled = 1; + set_memory_x(PAGE_OFFSET + BIOS_BEGIN, (BIOS_END - BIOS_BEGIN) >> PAGE_SHIFT); + if (__supported_pte_mask & _PAGE_NX) + printk(KERN_INFO "PCI : PCI BIOS aera is rw and x. Use pci=nobios if you want it NX.\n"); +} + /* * This is the standard structure used to identify the entry point * to the BIOS32 Service Directory, as documented in @@ -332,6 +354,7 @@ static struct pci_raw_ops * __devinit pci_find_bios(void) DBG("PCI: BIOS32 Service Directory entry at 0x%lx\n", bios32_entry); bios32_indirect.address = bios32_entry + PAGE_OFFSET; + set_bios_x(); if (check_pcibios()) return &pci_bios_access; } -- cgit v1.2.3 From 37db6c8f1d0c4b8f01dc049f3a893b725288660f Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 16 Nov 2010 08:25:08 +0000 Subject: x86-64: Fix and clean up AMD Fam10 MMCONF enabling Candidate memory ranges were not calculated properly (start addresses got needlessly rounded down, and end addresses didn't get rounded up at all), address comparison for secondary CPUs was done on only part of the address, and disabled status wasn't tracked properly. Signed-off-by: Jan Beulich Acked-by: Yinghai Lu Acked-by: Andreas Herrmann LKML-Reference: <4CE24DF40200007800022737@vpn.id2.novell.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/msr-index.h | 2 +- arch/x86/kernel/mmconf-fam10h_64.c | 64 ++++++++++++++++++-------------------- 2 files changed, 31 insertions(+), 35 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h index 3ea3dc487047..6b89f5e86021 100644 --- a/arch/x86/include/asm/msr-index.h +++ b/arch/x86/include/asm/msr-index.h @@ -128,7 +128,7 @@ #define FAM10H_MMIO_CONF_ENABLE (1<<0) #define FAM10H_MMIO_CONF_BUSRANGE_MASK 0xf #define FAM10H_MMIO_CONF_BUSRANGE_SHIFT 2 -#define FAM10H_MMIO_CONF_BASE_MASK 0xfffffff +#define FAM10H_MMIO_CONF_BASE_MASK 0xfffffffULL #define FAM10H_MMIO_CONF_BASE_SHIFT 20 #define MSR_FAM10H_NODE_ID 0xc001100c diff --git a/arch/x86/kernel/mmconf-fam10h_64.c b/arch/x86/kernel/mmconf-fam10h_64.c index 6da143c2a6b8..ac861b8348e2 100644 --- a/arch/x86/kernel/mmconf-fam10h_64.c +++ b/arch/x86/kernel/mmconf-fam10h_64.c @@ -25,7 +25,6 @@ struct pci_hostbridge_probe { }; static u64 __cpuinitdata fam10h_pci_mmconf_base; -static int __cpuinitdata fam10h_pci_mmconf_base_status; static struct pci_hostbridge_probe pci_probes[] __cpuinitdata = { { 0, 0x18, PCI_VENDOR_ID_AMD, 0x1200 }, @@ -44,10 +43,12 @@ static int __cpuinit cmp_range(const void *x1, const void *x2) return start1 - start2; } -/*[47:0] */ -/* need to avoid (0xfd<<32) and (0xfe<<32), ht used space */ +#define MMCONF_UNIT (1ULL << FAM10H_MMIO_CONF_BASE_SHIFT) +#define MMCONF_MASK (~(MMCONF_UNIT - 1)) +#define MMCONF_SIZE (MMCONF_UNIT << 8) +/* need to avoid (0xfd<<32), (0xfe<<32), and (0xff<<32), ht used space */ #define FAM10H_PCI_MMCONF_BASE (0xfcULL<<32) -#define BASE_VALID(b) ((b != (0xfdULL << 32)) && (b != (0xfeULL << 32))) +#define BASE_VALID(b) ((b) + MMCONF_SIZE <= (0xfdULL<<32) || (b) >= (1ULL<<40)) static void __cpuinit get_fam10h_pci_mmconf_base(void) { int i; @@ -64,12 +65,11 @@ static void __cpuinit get_fam10h_pci_mmconf_base(void) struct range range[8]; /* only try to get setting from BSP */ - /* -1 or 1 */ - if (fam10h_pci_mmconf_base_status) + if (fam10h_pci_mmconf_base) return; if (!early_pci_allowed()) - goto fail; + return; found = 0; for (i = 0; i < ARRAY_SIZE(pci_probes); i++) { @@ -91,7 +91,7 @@ static void __cpuinit get_fam10h_pci_mmconf_base(void) } if (!found) - goto fail; + return; /* SYS_CFG */ address = MSR_K8_SYSCFG; @@ -99,16 +99,16 @@ static void __cpuinit get_fam10h_pci_mmconf_base(void) /* TOP_MEM2 is not enabled? */ if (!(val & (1<<21))) { - tom2 = 0; + tom2 = 1ULL << 32; } else { /* TOP_MEM2 */ address = MSR_K8_TOP_MEM2; rdmsrl(address, val); - tom2 = val & (0xffffULL<<32); + tom2 = max(val & 0xffffff800000ULL, 1ULL << 32); } if (base <= tom2) - base = tom2 + (1ULL<<32); + base = (tom2 + 2 * MMCONF_UNIT - 1) & MMCONF_MASK; /* * need to check if the range is in the high mmio range that is @@ -123,11 +123,11 @@ static void __cpuinit get_fam10h_pci_mmconf_base(void) if (!(reg & 3)) continue; - start = (((u64)reg) << 8) & (0xffULL << 32); /* 39:16 on 31:8*/ + start = (u64)(reg & 0xffffff00) << 8; /* 39:16 on 31:8*/ reg = read_pci_config(bus, slot, 1, 0x84 + (i << 3)); - end = (((u64)reg) << 8) & (0xffULL << 32); /* 39:16 on 31:8*/ + end = ((u64)(reg & 0xffffff00) << 8) | 0xffff; /* 39:16 on 31:8*/ - if (!end) + if (end < tom2) continue; range[hi_mmio_num].start = start; @@ -143,32 +143,27 @@ static void __cpuinit get_fam10h_pci_mmconf_base(void) if (range[hi_mmio_num - 1].end < base) goto out; - if (range[0].start > base) + if (range[0].start > base + MMCONF_SIZE) goto out; /* need to find one window */ - base = range[0].start - (1ULL << 32); + base = (range[0].start & MMCONF_MASK) - MMCONF_UNIT; if ((base > tom2) && BASE_VALID(base)) goto out; - base = range[hi_mmio_num - 1].end + (1ULL << 32); - if ((base > tom2) && BASE_VALID(base)) + base = (range[hi_mmio_num - 1].end + MMCONF_UNIT) & MMCONF_MASK; + if (BASE_VALID(base)) goto out; /* need to find window between ranges */ - if (hi_mmio_num > 1) - for (i = 0; i < hi_mmio_num - 1; i++) { - if (range[i + 1].start > (range[i].end + (1ULL << 32))) { - base = range[i].end + (1ULL << 32); - if ((base > tom2) && BASE_VALID(base)) - goto out; - } + for (i = 1; i < hi_mmio_num; i++) { + base = (range[i - 1].end + MMCONF_UNIT) & MMCONF_MASK; + val = range[i].start & MMCONF_MASK; + if (val >= base + MMCONF_SIZE && BASE_VALID(base)) + goto out; } - -fail: - fam10h_pci_mmconf_base_status = -1; return; + out: fam10h_pci_mmconf_base = base; - fam10h_pci_mmconf_base_status = 1; } void __cpuinit fam10h_check_enable_mmcfg(void) @@ -190,11 +185,10 @@ void __cpuinit fam10h_check_enable_mmcfg(void) /* only trust the one handle 256 buses, if acpi=off */ if (!acpi_pci_disabled || busnbits >= 8) { - u64 base; - base = val & (0xffffULL << 32); - if (fam10h_pci_mmconf_base_status <= 0) { + u64 base = val & MMCONF_MASK; + + if (!fam10h_pci_mmconf_base) { fam10h_pci_mmconf_base = base; - fam10h_pci_mmconf_base_status = 1; return; } else if (fam10h_pci_mmconf_base == base) return; @@ -206,8 +200,10 @@ void __cpuinit fam10h_check_enable_mmcfg(void) * with 256 buses */ get_fam10h_pci_mmconf_base(); - if (fam10h_pci_mmconf_base_status <= 0) + if (!fam10h_pci_mmconf_base) { + pci_probe &= ~PCI_CHECK_ENABLE_AMD_MMCONF; return; + } printk(KERN_INFO "Enable MMCONFIG on AMD Family 10h\n"); val &= ~((FAM10H_MMIO_CONF_BASE_MASK< Date: Fri, 5 Nov 2010 05:59:39 -0400 Subject: x86: Eliminate bp argument from the stack tracing routines The various stack tracing routines take a 'bp' argument in which the caller is supposed to provide the base pointer to use, or 0 if doesn't have one. Since bp is garbage whenever CONFIG_FRAME_POINTER is not defined, this means all callers in principle should either always pass 0, or be conditional on CONFIG_FRAME_POINTER. However, there are only really three use cases for stack tracing: (a) Trace the current task, including IRQ stack if any (b) Trace the current task, but skip IRQ stack (c) Trace some other task In all cases, if CONFIG_FRAME_POINTER is not defined, bp should just be 0. If it _is_ defined, then - in case (a) bp should be gotten directly from the CPU's register, so the caller should pass NULL for regs, - in case (b) the caller should should pass the IRQ registers to dump_trace(), - in case (c) bp should be gotten from the top of the task's stack, so the caller should pass NULL for regs. Hence, the bp argument is not necessary because the combination of task and regs is sufficient to determine an appropriate value for bp. This patch introduces a new inline function stack_frame(task, regs) that computes the desired bp. This function is then called from the two versions of dump_stack(). Signed-off-by: Soren Sandmann Acked-by: Steven Rostedt Cc: Thomas Gleixner Cc: Ingo Molnar Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Arjan van de Ven , Cc: Frederic Weisbecker , Cc: Arnaldo Carvalho de Melo , LKML-Reference: > Signed-off-by: Frederic Weisbecker --- arch/x86/include/asm/kdebug.h | 2 +- arch/x86/include/asm/stacktrace.h | 33 ++++++++++++++++++++++++++++++--- arch/x86/kernel/cpu/perf_event.c | 2 +- arch/x86/kernel/dumpstack.c | 12 ++++++------ arch/x86/kernel/dumpstack_32.c | 25 +++++++------------------ arch/x86/kernel/dumpstack_64.c | 24 +++++++----------------- arch/x86/kernel/process.c | 3 +-- arch/x86/kernel/stacktrace.c | 8 ++++---- arch/x86/mm/kmemcheck/error.c | 2 +- arch/x86/oprofile/backtrace.c | 2 +- include/linux/stacktrace.h | 4 +++- 11 files changed, 62 insertions(+), 55 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/kdebug.h b/arch/x86/include/asm/kdebug.h index 5bdfca86581b..f23eb2528464 100644 --- a/arch/x86/include/asm/kdebug.h +++ b/arch/x86/include/asm/kdebug.h @@ -28,7 +28,7 @@ extern void die(const char *, struct pt_regs *,long); extern int __must_check __die(const char *, struct pt_regs *, long); extern void show_registers(struct pt_regs *regs); extern void show_trace(struct task_struct *t, struct pt_regs *regs, - unsigned long *sp, unsigned long bp); + unsigned long *sp); extern void __show_regs(struct pt_regs *regs, int all); extern void show_regs(struct pt_regs *regs); extern unsigned long oops_begin(void); diff --git a/arch/x86/include/asm/stacktrace.h b/arch/x86/include/asm/stacktrace.h index 2b16a2ad23dc..52b5c7ed3608 100644 --- a/arch/x86/include/asm/stacktrace.h +++ b/arch/x86/include/asm/stacktrace.h @@ -7,6 +7,7 @@ #define _ASM_X86_STACKTRACE_H #include +#include extern int kstack_depth_to_print; @@ -46,7 +47,7 @@ struct stacktrace_ops { }; void dump_trace(struct task_struct *tsk, struct pt_regs *regs, - unsigned long *stack, unsigned long bp, + unsigned long *stack, const struct stacktrace_ops *ops, void *data); #ifdef CONFIG_X86_32 @@ -57,13 +58,39 @@ void dump_trace(struct task_struct *tsk, struct pt_regs *regs, #define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :) #endif +#ifdef CONFIG_FRAME_POINTER +static inline unsigned long +stack_frame(struct task_struct *task, struct pt_regs *regs) +{ + unsigned long bp; + + if (regs) + return regs->bp; + + if (task == current) { + /* Grab bp right from our regs */ + get_bp(bp); + return bp; + } + + /* bp is the last reg pushed by switch_to */ + return *(unsigned long *)task->thread.sp; +} +#else +static inline unsigned long +stack_frame(struct task_struct *task, struct pt_regs *regs) +{ + return 0; +} +#endif + extern void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp, char *log_lvl); + unsigned long *stack, 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); + unsigned long *sp, char *log_lvl); extern unsigned int code_bytes; diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index ed6310183efb..461a85dcaba4 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1666,7 +1666,7 @@ perf_callchain_kernel(struct perf_callchain_entry *entry, struct pt_regs *regs) perf_callchain_store(entry, regs->ip); - dump_trace(NULL, regs, NULL, regs->bp, &backtrace_ops, entry); + dump_trace(NULL, regs, NULL, &backtrace_ops, entry); } #ifdef CONFIG_COMPAT diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c index 6e8752c1bd52..8474c998cbd4 100644 --- a/arch/x86/kernel/dumpstack.c +++ b/arch/x86/kernel/dumpstack.c @@ -175,21 +175,21 @@ static const struct stacktrace_ops print_trace_ops = { void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp, char *log_lvl) + unsigned long *stack, char *log_lvl) { printk("%sCall Trace:\n", log_lvl); - dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl); + dump_trace(task, regs, stack, &print_trace_ops, log_lvl); } void show_trace(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp) + unsigned long *stack) { - show_trace_log_lvl(task, regs, stack, bp, ""); + show_trace_log_lvl(task, regs, stack, ""); } void show_stack(struct task_struct *task, unsigned long *sp) { - show_stack_log_lvl(task, NULL, sp, 0, ""); + show_stack_log_lvl(task, NULL, sp, ""); } /* @@ -210,7 +210,7 @@ void dump_stack(void) init_utsname()->release, (int)strcspn(init_utsname()->version, " "), init_utsname()->version); - show_trace(NULL, NULL, &stack, bp); + show_trace(NULL, NULL, &stack); } EXPORT_SYMBOL(dump_stack); diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index 1bc7f75a5bda..74cc1eda384b 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -17,11 +17,12 @@ #include -void dump_trace(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp, +void dump_trace(struct task_struct *task, + struct pt_regs *regs, unsigned long *stack, const struct stacktrace_ops *ops, void *data) { int graph = 0; + unsigned long bp; if (!task) task = current; @@ -34,18 +35,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, stack = (unsigned long *)task->thread.sp; } -#ifdef CONFIG_FRAME_POINTER - if (!bp) { - if (task == current) { - /* Grab bp right from our regs */ - get_bp(bp); - } else { - /* bp is the last reg pushed by switch_to */ - bp = *(unsigned long *) task->thread.sp; - } - } -#endif - + bp = stack_frame(task, regs); for (;;) { struct thread_info *context; @@ -65,7 +55,7 @@ EXPORT_SYMBOL(dump_trace); void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, - unsigned long *sp, unsigned long bp, char *log_lvl) + unsigned long *sp, char *log_lvl) { unsigned long *stack; int i; @@ -87,7 +77,7 @@ show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, touch_nmi_watchdog(); } printk(KERN_CONT "\n"); - show_trace_log_lvl(task, regs, sp, bp, log_lvl); + show_trace_log_lvl(task, regs, sp, log_lvl); } @@ -112,8 +102,7 @@ void show_registers(struct pt_regs *regs) u8 *ip; printk(KERN_EMERG "Stack:\n"); - show_stack_log_lvl(NULL, regs, ®s->sp, - 0, KERN_EMERG); + show_stack_log_lvl(NULL, regs, ®s->sp, KERN_EMERG); printk(KERN_EMERG "Code: "); diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 6a340485249a..64101335de19 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -139,8 +139,8 @@ fixup_bp_irq_link(unsigned long bp, unsigned long *stack, * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack */ -void dump_trace(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp, +void dump_trace(struct task_struct *task, + struct pt_regs *regs, unsigned long *stack, const struct stacktrace_ops *ops, void *data) { const unsigned cpu = get_cpu(); @@ -149,6 +149,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned used = 0; struct thread_info *tinfo; int graph = 0; + unsigned long bp; if (!task) task = current; @@ -160,18 +161,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, stack = (unsigned long *)task->thread.sp; } -#ifdef CONFIG_FRAME_POINTER - if (!bp) { - if (task == current) { - /* Grab bp right from our regs */ - get_bp(bp); - } else { - /* bp is the last reg pushed by switch_to */ - bp = *(unsigned long *) task->thread.sp; - } - } -#endif - + bp = stack_frame(task, regs); /* * Print function call entries in all stacks, starting at the * current stack address. If the stacks consist of nested @@ -235,7 +225,7 @@ EXPORT_SYMBOL(dump_trace); void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, - unsigned long *sp, unsigned long bp, char *log_lvl) + unsigned long *sp, char *log_lvl) { unsigned long *irq_stack_end; unsigned long *irq_stack; @@ -279,7 +269,7 @@ show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, preempt_enable(); printk(KERN_CONT "\n"); - show_trace_log_lvl(task, regs, sp, bp, log_lvl); + show_trace_log_lvl(task, regs, sp, log_lvl); } void show_registers(struct pt_regs *regs) @@ -308,7 +298,7 @@ void show_registers(struct pt_regs *regs) printk(KERN_EMERG "Stack:\n"); show_stack_log_lvl(NULL, regs, (unsigned long *)sp, - regs->bp, KERN_EMERG); + KERN_EMERG); printk(KERN_EMERG "Code: "); diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index 57d1868a86aa..96ed1aac543a 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -91,8 +91,7 @@ void exit_thread(void) void show_regs(struct pt_regs *regs) { show_registers(regs); - show_trace(NULL, regs, (unsigned long *)kernel_stack_pointer(regs), - regs->bp); + show_trace(NULL, regs, (unsigned long *)kernel_stack_pointer(regs)); } void show_regs_common(void) diff --git a/arch/x86/kernel/stacktrace.c b/arch/x86/kernel/stacktrace.c index b53c525368a7..938c8e10a19a 100644 --- a/arch/x86/kernel/stacktrace.c +++ b/arch/x86/kernel/stacktrace.c @@ -73,22 +73,22 @@ static const struct stacktrace_ops save_stack_ops_nosched = { */ void save_stack_trace(struct stack_trace *trace) { - dump_trace(current, NULL, NULL, 0, &save_stack_ops, trace); + dump_trace(current, NULL, NULL, &save_stack_ops, trace); if (trace->nr_entries < trace->max_entries) trace->entries[trace->nr_entries++] = ULONG_MAX; } EXPORT_SYMBOL_GPL(save_stack_trace); -void save_stack_trace_bp(struct stack_trace *trace, unsigned long bp) +void save_stack_trace_regs(struct stack_trace *trace, struct pt_regs *regs) { - dump_trace(current, NULL, NULL, bp, &save_stack_ops, trace); + dump_trace(current, regs, NULL, &save_stack_ops, trace); if (trace->nr_entries < trace->max_entries) trace->entries[trace->nr_entries++] = ULONG_MAX; } void save_stack_trace_tsk(struct task_struct *tsk, struct stack_trace *trace) { - dump_trace(tsk, NULL, NULL, 0, &save_stack_ops_nosched, trace); + dump_trace(tsk, NULL, NULL, &save_stack_ops_nosched, trace); if (trace->nr_entries < trace->max_entries) trace->entries[trace->nr_entries++] = ULONG_MAX; } diff --git a/arch/x86/mm/kmemcheck/error.c b/arch/x86/mm/kmemcheck/error.c index af3b6c8a436f..704a37cedddb 100644 --- a/arch/x86/mm/kmemcheck/error.c +++ b/arch/x86/mm/kmemcheck/error.c @@ -185,7 +185,7 @@ void kmemcheck_error_save(enum kmemcheck_shadow state, e->trace.entries = e->trace_entries; e->trace.max_entries = ARRAY_SIZE(e->trace_entries); e->trace.skip = 0; - save_stack_trace_bp(&e->trace, regs->bp); + save_stack_trace_regs(&e->trace, regs); /* Round address down to nearest 16 bytes */ shadow_copy = kmemcheck_shadow_lookup(address diff --git a/arch/x86/oprofile/backtrace.c b/arch/x86/oprofile/backtrace.c index 2d49d4e19a36..72cbec14d783 100644 --- a/arch/x86/oprofile/backtrace.c +++ b/arch/x86/oprofile/backtrace.c @@ -126,7 +126,7 @@ x86_backtrace(struct pt_regs * const regs, unsigned int depth) if (!user_mode_vm(regs)) { unsigned long stack = kernel_stack_pointer(regs); if (depth) - dump_trace(NULL, regs, (unsigned long *)stack, 0, + dump_trace(NULL, regs, (unsigned long *)stack, &backtrace_ops, &depth); return; } diff --git a/include/linux/stacktrace.h b/include/linux/stacktrace.h index 51efbef38fb0..25310f1d7f37 100644 --- a/include/linux/stacktrace.h +++ b/include/linux/stacktrace.h @@ -2,6 +2,7 @@ #define __LINUX_STACKTRACE_H struct task_struct; +struct pt_regs; #ifdef CONFIG_STACKTRACE struct task_struct; @@ -13,7 +14,8 @@ struct stack_trace { }; extern void save_stack_trace(struct stack_trace *trace); -extern void save_stack_trace_bp(struct stack_trace *trace, unsigned long bp); +extern void save_stack_trace_regs(struct stack_trace *trace, + struct pt_regs *regs); extern void save_stack_trace_tsk(struct task_struct *tsk, struct stack_trace *trace); -- cgit v1.2.3 From 91d95fda8594ce5e0ccd81381ee7b956cf513c59 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Wed, 24 Nov 2010 12:57:18 -0800 Subject: arch/x86/include/asm/fixmap.h: mark __set_fixmap_offset as __always_inline When compiling arch/x86/kernel/early_printk_mrst.c with i386 allmodconfig, gcc-4.1.0 generates an out-of-line copy of __set_fixmap_offset() which contains a reference to __this_fixmap_does_not_exist which the compiler cannot elide. Marking __set_fixmap_offset() as __always_inline prevents this. Cc: Thomas Gleixner Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: Feng Tang Acked-by: Randy Dunlap Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/include/asm/fixmap.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/fixmap.h b/arch/x86/include/asm/fixmap.h index 4d293dced62f..9479a037419f 100644 --- a/arch/x86/include/asm/fixmap.h +++ b/arch/x86/include/asm/fixmap.h @@ -216,8 +216,8 @@ static inline unsigned long virt_to_fix(const unsigned long vaddr) } /* Return an pointer with offset calculated */ -static inline unsigned long __set_fixmap_offset(enum fixed_addresses idx, - phys_addr_t phys, pgprot_t flags) +static __always_inline unsigned long +__set_fixmap_offset(enum fixed_addresses idx, phys_addr_t phys, pgprot_t flags) { __set_fixmap(idx, phys, flags); return fix_to_virt(idx) + (phys & (PAGE_SIZE - 1)); -- cgit v1.2.3 From 004417a6d468e24399e383645c068b498eed84ad Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 25 Nov 2010 18:38:29 +0100 Subject: perf, arch: Cleanup perf-pmu init vs lockup-detector The perf hardware pmu got initialized at various points in the boot, some before early_initcall() some after (notably arch_initcall). The problem is that the NMI lockup detector is ran from early_initcall() and expects the hardware pmu to be present. Sanitize this by moving all architecture hardware pmu implementations to initialize at early_initcall() and move the lockup detector to an explicit initcall right after that. Cc: paulus Cc: davem Cc: Michael Cree Cc: Deng-Cheng Zhu Acked-by: Paul Mundt Acked-by: Will Deacon Signed-off-by: Peter Zijlstra LKML-Reference: <1290707759.2145.119.camel@laptop> Signed-off-by: Ingo Molnar --- arch/alpha/include/asm/perf_event.h | 6 ------ arch/alpha/kernel/irq_alpha.c | 2 -- arch/alpha/kernel/perf_event.c | 9 ++++++--- arch/arm/kernel/perf_event.c | 2 +- arch/mips/kernel/perf_event_mipsxx.c | 2 +- arch/powerpc/kernel/e500-pmu.c | 2 +- arch/powerpc/kernel/mpc7450-pmu.c | 2 +- arch/powerpc/kernel/power4-pmu.c | 2 +- arch/powerpc/kernel/power5+-pmu.c | 2 +- arch/powerpc/kernel/power5-pmu.c | 2 +- arch/powerpc/kernel/power6-pmu.c | 2 +- arch/powerpc/kernel/power7-pmu.c | 2 +- arch/powerpc/kernel/ppc970-pmu.c | 2 +- arch/sh/kernel/cpu/sh4/perf_event.c | 2 +- arch/sh/kernel/cpu/sh4a/perf_event.c | 2 +- arch/sparc/include/asm/perf_event.h | 4 ---- arch/sparc/kernel/nmi.c | 2 -- arch/sparc/kernel/perf_event.c | 7 +++++-- arch/x86/include/asm/perf_event.h | 2 -- arch/x86/kernel/cpu/common.c | 1 - arch/x86/kernel/cpu/perf_event.c | 11 +++++++---- include/linux/sched.h | 4 ++++ init/main.c | 1 + kernel/watchdog.c | 7 +++---- 24 files changed, 38 insertions(+), 42 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/alpha/include/asm/perf_event.h b/arch/alpha/include/asm/perf_event.h index fe792ca818f6..5996e7a6757e 100644 --- a/arch/alpha/include/asm/perf_event.h +++ b/arch/alpha/include/asm/perf_event.h @@ -1,10 +1,4 @@ #ifndef __ASM_ALPHA_PERF_EVENT_H #define __ASM_ALPHA_PERF_EVENT_H -#ifdef CONFIG_PERF_EVENTS -extern void init_hw_perf_events(void); -#else -static inline void init_hw_perf_events(void) { } -#endif - #endif /* __ASM_ALPHA_PERF_EVENT_H */ diff --git a/arch/alpha/kernel/irq_alpha.c b/arch/alpha/kernel/irq_alpha.c index 5f77afb88e89..4c8bb374eb0a 100644 --- a/arch/alpha/kernel/irq_alpha.c +++ b/arch/alpha/kernel/irq_alpha.c @@ -112,8 +112,6 @@ init_IRQ(void) wrent(entInt, 0); alpha_mv.init_irq(); - - init_hw_perf_events(); } /* diff --git a/arch/alpha/kernel/perf_event.c b/arch/alpha/kernel/perf_event.c index 1cc49683fb69..3283059b6e85 100644 --- a/arch/alpha/kernel/perf_event.c +++ b/arch/alpha/kernel/perf_event.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include @@ -863,13 +864,13 @@ static void alpha_perf_event_irq_handler(unsigned long la_ptr, /* * Init call to initialise performance events at kernel startup. */ -void __init init_hw_perf_events(void) +int __init init_hw_perf_events(void) { pr_info("Performance events: "); if (!supported_cpu()) { pr_cont("No support for your CPU.\n"); - return; + return 0; } pr_cont("Supported CPU type!\n"); @@ -882,5 +883,7 @@ void __init init_hw_perf_events(void) alpha_pmu = &ev67_pmu; perf_pmu_register(&pmu); -} + return 0; +} +early_initcall(init_hw_perf_events); diff --git a/arch/arm/kernel/perf_event.c b/arch/arm/kernel/perf_event.c index 07a50357492a..d45f70e5f2ee 100644 --- a/arch/arm/kernel/perf_event.c +++ b/arch/arm/kernel/perf_event.c @@ -3038,7 +3038,7 @@ init_hw_perf_events(void) return 0; } -arch_initcall(init_hw_perf_events); +early_initcall(init_hw_perf_events); /* * Callchain handling code. diff --git a/arch/mips/kernel/perf_event_mipsxx.c b/arch/mips/kernel/perf_event_mipsxx.c index 5c7c6fc07565..183e0d226669 100644 --- a/arch/mips/kernel/perf_event_mipsxx.c +++ b/arch/mips/kernel/perf_event_mipsxx.c @@ -1047,6 +1047,6 @@ init_hw_perf_events(void) return 0; } -arch_initcall(init_hw_perf_events); +early_initcall(init_hw_perf_events); #endif /* defined(CONFIG_CPU_MIPS32)... */ diff --git a/arch/powerpc/kernel/e500-pmu.c b/arch/powerpc/kernel/e500-pmu.c index 7c07de0d8943..b150b510510f 100644 --- a/arch/powerpc/kernel/e500-pmu.c +++ b/arch/powerpc/kernel/e500-pmu.c @@ -126,4 +126,4 @@ static int init_e500_pmu(void) return register_fsl_emb_pmu(&e500_pmu); } -arch_initcall(init_e500_pmu); +early_initcall(init_e500_pmu); diff --git a/arch/powerpc/kernel/mpc7450-pmu.c b/arch/powerpc/kernel/mpc7450-pmu.c index 09d72028f317..2cc5e0301d0b 100644 --- a/arch/powerpc/kernel/mpc7450-pmu.c +++ b/arch/powerpc/kernel/mpc7450-pmu.c @@ -414,4 +414,4 @@ static int init_mpc7450_pmu(void) return register_power_pmu(&mpc7450_pmu); } -arch_initcall(init_mpc7450_pmu); +early_initcall(init_mpc7450_pmu); diff --git a/arch/powerpc/kernel/power4-pmu.c b/arch/powerpc/kernel/power4-pmu.c index 2a361cdda635..ead8b3c2649e 100644 --- a/arch/powerpc/kernel/power4-pmu.c +++ b/arch/powerpc/kernel/power4-pmu.c @@ -613,4 +613,4 @@ static int init_power4_pmu(void) return register_power_pmu(&power4_pmu); } -arch_initcall(init_power4_pmu); +early_initcall(init_power4_pmu); diff --git a/arch/powerpc/kernel/power5+-pmu.c b/arch/powerpc/kernel/power5+-pmu.c index 199de527d411..eca0ac595cb6 100644 --- a/arch/powerpc/kernel/power5+-pmu.c +++ b/arch/powerpc/kernel/power5+-pmu.c @@ -682,4 +682,4 @@ static int init_power5p_pmu(void) return register_power_pmu(&power5p_pmu); } -arch_initcall(init_power5p_pmu); +early_initcall(init_power5p_pmu); diff --git a/arch/powerpc/kernel/power5-pmu.c b/arch/powerpc/kernel/power5-pmu.c index 98b6a729a9dd..d5ff0f64a5e6 100644 --- a/arch/powerpc/kernel/power5-pmu.c +++ b/arch/powerpc/kernel/power5-pmu.c @@ -621,4 +621,4 @@ static int init_power5_pmu(void) return register_power_pmu(&power5_pmu); } -arch_initcall(init_power5_pmu); +early_initcall(init_power5_pmu); diff --git a/arch/powerpc/kernel/power6-pmu.c b/arch/powerpc/kernel/power6-pmu.c index 84a607bda8fb..31603927e376 100644 --- a/arch/powerpc/kernel/power6-pmu.c +++ b/arch/powerpc/kernel/power6-pmu.c @@ -544,4 +544,4 @@ static int init_power6_pmu(void) return register_power_pmu(&power6_pmu); } -arch_initcall(init_power6_pmu); +early_initcall(init_power6_pmu); diff --git a/arch/powerpc/kernel/power7-pmu.c b/arch/powerpc/kernel/power7-pmu.c index 852f7b7f6b40..593740fcb799 100644 --- a/arch/powerpc/kernel/power7-pmu.c +++ b/arch/powerpc/kernel/power7-pmu.c @@ -369,4 +369,4 @@ static int init_power7_pmu(void) return register_power_pmu(&power7_pmu); } -arch_initcall(init_power7_pmu); +early_initcall(init_power7_pmu); diff --git a/arch/powerpc/kernel/ppc970-pmu.c b/arch/powerpc/kernel/ppc970-pmu.c index 3fee685de4df..9a6e093858fe 100644 --- a/arch/powerpc/kernel/ppc970-pmu.c +++ b/arch/powerpc/kernel/ppc970-pmu.c @@ -494,4 +494,4 @@ static int init_ppc970_pmu(void) return register_power_pmu(&ppc970_pmu); } -arch_initcall(init_ppc970_pmu); +early_initcall(init_ppc970_pmu); diff --git a/arch/sh/kernel/cpu/sh4/perf_event.c b/arch/sh/kernel/cpu/sh4/perf_event.c index dbf3b4bb71fe..748955df018d 100644 --- a/arch/sh/kernel/cpu/sh4/perf_event.c +++ b/arch/sh/kernel/cpu/sh4/perf_event.c @@ -250,4 +250,4 @@ static int __init sh7750_pmu_init(void) return register_sh_pmu(&sh7750_pmu); } -arch_initcall(sh7750_pmu_init); +early_initcall(sh7750_pmu_init); diff --git a/arch/sh/kernel/cpu/sh4a/perf_event.c b/arch/sh/kernel/cpu/sh4a/perf_event.c index 580276525731..17e6bebfede0 100644 --- a/arch/sh/kernel/cpu/sh4a/perf_event.c +++ b/arch/sh/kernel/cpu/sh4a/perf_event.c @@ -284,4 +284,4 @@ static int __init sh4a_pmu_init(void) return register_sh_pmu(&sh4a_pmu); } -arch_initcall(sh4a_pmu_init); +early_initcall(sh4a_pmu_init); diff --git a/arch/sparc/include/asm/perf_event.h b/arch/sparc/include/asm/perf_event.h index 6e8bfa1786da..4d3dbe3703e9 100644 --- a/arch/sparc/include/asm/perf_event.h +++ b/arch/sparc/include/asm/perf_event.h @@ -4,8 +4,6 @@ #ifdef CONFIG_PERF_EVENTS #include -extern void init_hw_perf_events(void); - #define perf_arch_fetch_caller_regs(regs, ip) \ do { \ unsigned long _pstate, _asi, _pil, _i7, _fp; \ @@ -26,8 +24,6 @@ do { \ (regs)->u_regs[UREG_I6] = _fp; \ (regs)->u_regs[UREG_I7] = _i7; \ } while (0) -#else -static inline void init_hw_perf_events(void) { } #endif #endif diff --git a/arch/sparc/kernel/nmi.c b/arch/sparc/kernel/nmi.c index a4bd7ba74c89..300f810142f5 100644 --- a/arch/sparc/kernel/nmi.c +++ b/arch/sparc/kernel/nmi.c @@ -270,8 +270,6 @@ int __init nmi_init(void) atomic_set(&nmi_active, -1); } } - if (!err) - init_hw_perf_events(); return err; } diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c index 0d6deb55a2ae..75c5b1263970 100644 --- a/arch/sparc/kernel/perf_event.c +++ b/arch/sparc/kernel/perf_event.c @@ -1307,20 +1307,23 @@ static bool __init supported_pmu(void) return false; } -void __init init_hw_perf_events(void) +int __init init_hw_perf_events(void) { pr_info("Performance events: "); if (!supported_pmu()) { pr_cont("No support for PMU type '%s'\n", sparc_pmu_type); - return; + return 0; } pr_cont("Supported PMU type is '%s'\n", sparc_pmu_type); perf_pmu_register(&pmu); register_die_notifier(&perf_event_nmi_notifier); + + return 0; } +early_initcall(init_hw_perf_event); void perf_callchain_kernel(struct perf_callchain_entry *entry, struct pt_regs *regs) diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h index 550e26b1dbb3..d9d4dae305f6 100644 --- a/arch/x86/include/asm/perf_event.h +++ b/arch/x86/include/asm/perf_event.h @@ -125,7 +125,6 @@ union cpuid10_edx { #define IBS_OP_MAX_CNT_EXT 0x007FFFFFULL /* not a register bit mask */ #ifdef CONFIG_PERF_EVENTS -extern void init_hw_perf_events(void); extern void perf_events_lapic_init(void); #define PERF_EVENT_INDEX_OFFSET 0 @@ -156,7 +155,6 @@ extern unsigned long perf_misc_flags(struct pt_regs *regs); } #else -static inline void init_hw_perf_events(void) { } static inline void perf_events_lapic_init(void) { } #endif diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 4b68bda30938..1d59834396bd 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -894,7 +894,6 @@ void __init identify_boot_cpu(void) #else vgetcpu_set_mode(); #endif - init_hw_perf_events(); } void __cpuinit identify_secondary_cpu(struct cpuinfo_x86 *c) diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index c01dfec635db..817d2b195e8e 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1353,7 +1353,7 @@ static void __init pmu_check_apic(void) pr_info("no hardware sampling interrupt available.\n"); } -void __init init_hw_perf_events(void) +int __init init_hw_perf_events(void) { struct event_constraint *c; int err; @@ -1368,11 +1368,11 @@ void __init init_hw_perf_events(void) err = amd_pmu_init(); break; default: - return; + return 0; } if (err != 0) { pr_cont("no PMU driver, software events only.\n"); - return; + return 0; } pmu_check_apic(); @@ -1380,7 +1380,7 @@ void __init init_hw_perf_events(void) /* sanity check that the hardware exists or is emulated */ if (!check_hw_exists()) { pr_cont("Broken PMU hardware detected, software events only.\n"); - return; + return 0; } pr_cont("%s PMU driver.\n", x86_pmu.name); @@ -1431,7 +1431,10 @@ void __init init_hw_perf_events(void) perf_pmu_register(&pmu); perf_cpu_notifier(x86_pmu_notifier); + + return 0; } +early_initcall(init_hw_perf_events); static inline void x86_pmu_read(struct perf_event *event) { diff --git a/include/linux/sched.h b/include/linux/sched.h index 2c79e921a68b..d2e63d1e725c 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -316,6 +316,7 @@ extern int proc_dowatchdog_thresh(struct ctl_table *table, int write, size_t *lenp, loff_t *ppos); extern unsigned int softlockup_panic; extern int softlockup_thresh; +void lockup_detector_init(void); #else static inline void touch_softlockup_watchdog(void) { @@ -326,6 +327,9 @@ static inline void touch_softlockup_watchdog_sync(void) static inline void touch_all_softlockup_watchdogs(void) { } +static inline void lockup_detector_init(void) +{ +} #endif #ifdef CONFIG_DETECT_HUNG_TASK diff --git a/init/main.c b/init/main.c index 8646401f7a0e..261ad7b3fe0b 100644 --- a/init/main.c +++ b/init/main.c @@ -882,6 +882,7 @@ static int __init kernel_init(void * unused) smp_prepare_cpus(setup_max_cpus); do_pre_smp_initcalls(); + lockup_detector_init(); smp_init(); sched_init_smp(); diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 6e3c41a4024c..cad4e42060a9 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -547,13 +547,13 @@ static struct notifier_block __cpuinitdata cpu_nfb = { .notifier_call = cpu_callback }; -static int __init spawn_watchdog_task(void) +void __init lockup_detector_init(void) { void *cpu = (void *)(long)smp_processor_id(); int err; if (no_watchdog) - return 0; + return; err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); WARN_ON(notifier_to_errno(err)); @@ -561,6 +561,5 @@ static int __init spawn_watchdog_task(void) cpu_callback(&cpu_nfb, CPU_ONLINE, cpu); register_cpu_notifier(&cpu_nfb); - return 0; + return; } -early_initcall(spawn_watchdog_task); -- cgit v1.2.3 From af86da5318136eb49c0453c2e2be3280ee5d18d9 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Fri, 26 Nov 2010 14:32:09 +0300 Subject: perf, x86: P4 PMU - describe config format Add description of .config in a sake of RAW events. At least this should bring some light to those who will be reading this code. Signed-off-by: Cyrill Gorcunov Reviewed-by: Stephane Eranian Cc: Lin Ming Signed-off-by: Ingo Molnar --- arch/x86/include/asm/perf_event_p4.h | 63 +++++++++++++++++++++++++++++++----- 1 file changed, 55 insertions(+), 8 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/perf_event_p4.h b/arch/x86/include/asm/perf_event_p4.h index a70cd216be5d..295e2ff18a6a 100644 --- a/arch/x86/include/asm/perf_event_p4.h +++ b/arch/x86/include/asm/perf_event_p4.h @@ -744,14 +744,6 @@ enum P4_ESCR_EMASKS { }; /* - * P4 PEBS specifics (Replay Event only) - * - * Format (bits): - * 0-6: metric from P4_PEBS_METRIC enum - * 7 : reserved - * 8 : reserved - * 9-11 : reserved - * * Note we have UOP and PEBS bits reserved for now * just in case if we will need them once */ @@ -788,5 +780,60 @@ enum P4_PEBS_METRIC { P4_PEBS_METRIC__max }; +/* + * Notes on internal configuration of ESCR+CCCR tuples + * + * Since P4 has quite the different architecture of + * performance registers in compare with "architectural" + * once and we have on 64 bits to keep configuration + * of performance event, the following trick is used. + * + * 1) Since both ESCR and CCCR registers have only low + * 32 bits valuable, we pack them into a single 64 bit + * configuration. Low 32 bits of such config correspond + * to low 32 bits of CCCR register and high 32 bits + * correspond to low 32 bits of ESCR register. + * + * 2) The meaning of every bit of such config field can + * be found in Intel SDM but it should be noted that + * we "borrow" some reserved bits for own usage and + * clean them or set to a proper value when we do + * a real write to hardware registers. + * + * 3) The format of bits of config is the following + * and should be either 0 or set to some predefined + * values: + * + * Low 32 bits + * ----------- + * 0-6: P4_PEBS_METRIC enum + * 7-11: reserved + * 12: reserved (Enable) + * 13-15: reserved (ESCR select) + * 16-17: Active Thread + * 18: Compare + * 19: Complement + * 20-23: Threshold + * 24: Edge + * 25: reserved (FORCE_OVF) + * 26: reserved (OVF_PMI_T0) + * 27: reserved (OVF_PMI_T1) + * 28-29: reserved + * 30: reserved (Cascade) + * 31: reserved (OVF) + * + * High 32 bits + * ------------ + * 0: reserved (T1_USR) + * 1: reserved (T1_OS) + * 2: reserved (T0_USR) + * 3: reserved (T0_OS) + * 4: Tag Enable + * 5-8: Tag Value + * 9-24: Event Mask (may use P4_ESCR_EMASK_BIT helper) + * 25-30: enum P4_EVENTS + * 31: reserved (HT thread) + */ + #endif /* PERF_EVENT_P4_H */ -- cgit v1.2.3 From e7a3481c0246c8e45e79c629efd63b168e91fcda Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Mon, 25 Oct 2010 16:53:46 -0700 Subject: x86/pvclock: Zero last_value on resume If the guest domain has been suspend/resumed or migrated, then the system clock backing the pvclock clocksource may revert to a smaller value (ie, can be non-monotonic across the migration/save-restore). Make sure we zero last_value in that case so that the domain continues to see clock updates. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/x86/include/asm/pvclock.h | 1 + arch/x86/kernel/pvclock.c | 5 +++++ arch/x86/xen/time.c | 2 ++ 3 files changed, 8 insertions(+) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/pvclock.h b/arch/x86/include/asm/pvclock.h index 7f7e577a0e39..31d84acc1512 100644 --- a/arch/x86/include/asm/pvclock.h +++ b/arch/x86/include/asm/pvclock.h @@ -11,6 +11,7 @@ unsigned long pvclock_tsc_khz(struct pvclock_vcpu_time_info *src); void pvclock_read_wallclock(struct pvclock_wall_clock *wall, struct pvclock_vcpu_time_info *vcpu, struct timespec *ts); +void pvclock_resume(void); /* * Scale a 64-bit delta by scaling and multiplying by a 32-bit fraction, diff --git a/arch/x86/kernel/pvclock.c b/arch/x86/kernel/pvclock.c index 008b91eefa18..42eb3300dfc6 100644 --- a/arch/x86/kernel/pvclock.c +++ b/arch/x86/kernel/pvclock.c @@ -83,6 +83,11 @@ unsigned long pvclock_tsc_khz(struct pvclock_vcpu_time_info *src) static atomic64_t last_value = ATOMIC64_INIT(0); +void pvclock_resume(void) +{ + atomic64_set(&last_value, 0); +} + cycle_t pvclock_clocksource_read(struct pvclock_vcpu_time_info *src) { struct pvclock_shadow_time shadow; diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c index b2bb5aa3b054..5da5e53fb94c 100644 --- a/arch/x86/xen/time.c +++ b/arch/x86/xen/time.c @@ -426,6 +426,8 @@ void xen_timer_resume(void) { int cpu; + pvclock_resume(); + if (xen_clockevent != &xen_vcpuop_clockevent) return; -- cgit v1.2.3 From 7deb18dcf0478940ac979de002db1ed8ba6531dc Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Fri, 3 Dec 2010 18:54:22 +0900 Subject: x86: Introduce text_poke_smp_batch() for batch-code modifying Introduce text_poke_smp_batch(). This function modifies several text areas with one stop_machine() on SMP. Because calling stop_machine() is heavy task, it is better to aggregate text_poke requests. ( Note: I've talked with Rusty about this interface, and he would not like to expand stop_machine() interface, since it is not for generic use. ) Signed-off-by: Masami Hiramatsu Acked-by: Steven Rostedt Cc: Rusty Russell Cc: Frederic Weisbecker Cc: Ananth N Mavinakayanahalli Cc: Jason Baron Cc: Mathieu Desnoyers Cc: Jan Beulich Cc: 2nddept-manager@sdl.hitachi.co.jp LKML-Reference: <20101203095422.2961.51217.stgit@ltc236.sdl.hitachi.co.jp> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/alternative.h | 7 ++++++ arch/x86/kernel/alternative.c | 49 +++++++++++++++++++++++++++++++------- 2 files changed, 47 insertions(+), 9 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/alternative.h b/arch/x86/include/asm/alternative.h index 76561d20ea2f..4a2adaa9aefc 100644 --- a/arch/x86/include/asm/alternative.h +++ b/arch/x86/include/asm/alternative.h @@ -180,8 +180,15 @@ extern void *text_poke_early(void *addr, const void *opcode, size_t len); * On the local CPU you need to be protected again NMI or MCE handlers seeing an * inconsistent instruction while you patch. */ +struct text_poke_param { + void *addr; + const void *opcode; + size_t len; +}; + extern void *text_poke(void *addr, const void *opcode, size_t len); extern void *text_poke_smp(void *addr, const void *opcode, size_t len); +extern void text_poke_smp_batch(struct text_poke_param *params, int n); #if defined(CONFIG_DYNAMIC_FTRACE) || defined(HAVE_JUMP_LABEL) #define IDEAL_NOP_SIZE_5 5 diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index 5079f24c955a..553d0b0d639b 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -591,17 +591,21 @@ static atomic_t stop_machine_first; static int wrote_text; struct text_poke_params { - void *addr; - const void *opcode; - size_t len; + struct text_poke_param *params; + int nparams; }; static int __kprobes stop_machine_text_poke(void *data) { struct text_poke_params *tpp = data; + struct text_poke_param *p; + int i; if (atomic_dec_and_test(&stop_machine_first)) { - text_poke(tpp->addr, tpp->opcode, tpp->len); + for (i = 0; i < tpp->nparams; i++) { + p = &tpp->params[i]; + text_poke(p->addr, p->opcode, p->len); + } smp_wmb(); /* Make sure other cpus see that this has run */ wrote_text = 1; } else { @@ -610,8 +614,12 @@ static int __kprobes stop_machine_text_poke(void *data) smp_mb(); /* Load wrote_text before following execution */ } - flush_icache_range((unsigned long)tpp->addr, - (unsigned long)tpp->addr + tpp->len); + for (i = 0; i < tpp->nparams; i++) { + p = &tpp->params[i]; + flush_icache_range((unsigned long)p->addr, + (unsigned long)p->addr + p->len); + } + return 0; } @@ -631,10 +639,13 @@ static int __kprobes stop_machine_text_poke(void *data) void *__kprobes text_poke_smp(void *addr, const void *opcode, size_t len) { struct text_poke_params tpp; + struct text_poke_param p; - tpp.addr = addr; - tpp.opcode = opcode; - tpp.len = len; + p.addr = addr; + p.opcode = opcode; + p.len = len; + tpp.params = &p; + tpp.nparams = 1; atomic_set(&stop_machine_first, 1); wrote_text = 0; /* Use __stop_machine() because the caller already got online_cpus. */ @@ -642,6 +653,26 @@ void *__kprobes text_poke_smp(void *addr, const void *opcode, size_t len) return addr; } +/** + * text_poke_smp_batch - Update instructions on a live kernel on SMP + * @params: an array of text_poke parameters + * @n: the number of elements in params. + * + * Modify multi-byte instruction by using stop_machine() on SMP. Since the + * stop_machine() is heavy task, it is better to aggregate text_poke requests + * and do it once if possible. + * + * Note: Must be called under get_online_cpus() and text_mutex. + */ +void __kprobes text_poke_smp_batch(struct text_poke_param *params, int n) +{ + struct text_poke_params tpp = {.params = params, .nparams = n}; + + atomic_set(&stop_machine_first, 1); + wrote_text = 0; + stop_machine(stop_machine_text_poke, (void *)&tpp, NULL); +} + #if defined(CONFIG_DYNAMIC_FTRACE) || defined(HAVE_JUMP_LABEL) #ifdef CONFIG_X86_64 -- cgit v1.2.3 From 73c1160ce377d8fc6d84cb630ebf9658808bec49 Mon Sep 17 00:00:00 2001 From: Andre Przywara Date: Wed, 1 Dec 2010 12:17:44 +0100 Subject: KVM: enlarge number of possible CPUID leaves Currently the number of CPUID leaves KVM handles is limited to 40. My desktop machine (AthlonII) already has 35 and future CPUs will expand this well beyond the limit. Extend the limit to 80 to make room for future processors. KVM-Stable-Tag. Signed-off-by: Andre Przywara Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 9e6fe391094e..f702f82aa1eb 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -79,7 +79,7 @@ #define KVM_NUM_MMU_PAGES (1 << KVM_MMU_HASH_SHIFT) #define KVM_MIN_FREE_MMU_PAGES 5 #define KVM_REFILL_PAGES 25 -#define KVM_MAX_CPUID_ENTRIES 40 +#define KVM_MAX_CPUID_ENTRIES 80 #define KVM_NR_FIXED_MTRR_REGION 88 #define KVM_NR_VAR_MTRR 8 -- cgit v1.2.3 From 3fb82d56ad003e804923185316236f26b30dfdd5 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Tue, 23 Nov 2010 16:11:40 -0800 Subject: x86, suspend: Avoid unnecessary smp alternatives switch during suspend/resume During suspend, we disable all the non boot cpus. And during resume we bring them all back again. So no need to do alternatives_smp_switch() in between. On my core 2 based laptop, this speeds up the suspend path by 15msec and the resume path by 5 msec (suspend/resume speed up differences can be attributed to the different P-states that the cpu is in during suspend/resume). Signed-off-by: Suresh Siddha LKML-Reference: <1290557500.4946.8.camel@sbsiddha-MOBL3.sc.intel.com> Cc: Rafael J. Wysocki Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/alternative.h | 1 + arch/x86/kernel/alternative.c | 3 ++- arch/x86/kernel/smpboot.c | 14 ++++++++++++++ kernel/cpu.c | 11 +++++++++++ 4 files changed, 28 insertions(+), 1 deletion(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/alternative.h b/arch/x86/include/asm/alternative.h index 76561d20ea2f..01171f6c2c37 100644 --- a/arch/x86/include/asm/alternative.h +++ b/arch/x86/include/asm/alternative.h @@ -66,6 +66,7 @@ extern void alternatives_smp_module_add(struct module *mod, char *name, extern void alternatives_smp_module_del(struct module *mod); extern void alternatives_smp_switch(int smp); extern int alternatives_text_reserved(void *start, void *end); +extern bool skip_smp_alternatives; #else static inline void alternatives_smp_module_add(struct module *mod, char *name, void *locks, void *locks_end, diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index 5079f24c955a..9f98eb400fef 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -353,6 +353,7 @@ void __init_or_module alternatives_smp_module_del(struct module *mod) mutex_unlock(&smp_alt); } +bool skip_smp_alternatives; void alternatives_smp_switch(int smp) { struct smp_alt_module *mod; @@ -368,7 +369,7 @@ void alternatives_smp_switch(int smp) printk("lockdep: fixing up alternatives.\n"); #endif - if (noreplace_smp || smp_alt_once) + if (noreplace_smp || smp_alt_once || skip_smp_alternatives) return; BUG_ON(!smp && (num_online_cpus() > 1)); diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 083e99d1b7df..837c81e99edf 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1166,6 +1166,20 @@ out: preempt_enable(); } +void arch_disable_nonboot_cpus_begin(void) +{ + /* + * Avoid the smp alternatives switch during the disable_nonboot_cpus(). + * In the suspend path, we will be back in the SMP mode shortly anyways. + */ + skip_smp_alternatives = true; +} + +void arch_disable_nonboot_cpus_end(void) +{ + skip_smp_alternatives = false; +} + void arch_enable_nonboot_cpus_begin(void) { set_mtrr_aps_delayed_init(); diff --git a/kernel/cpu.c b/kernel/cpu.c index f6e726f18491..8ccc182069ec 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -386,6 +386,14 @@ out: #ifdef CONFIG_PM_SLEEP_SMP static cpumask_var_t frozen_cpus; +void __weak arch_disable_nonboot_cpus_begin(void) +{ +} + +void __weak arch_disable_nonboot_cpus_end(void) +{ +} + int disable_nonboot_cpus(void) { int cpu, first_cpu, error = 0; @@ -397,6 +405,7 @@ int disable_nonboot_cpus(void) * with the userspace trying to use the CPU hotplug at the same time */ cpumask_clear(frozen_cpus); + arch_disable_nonboot_cpus_begin(); printk("Disabling non-boot CPUs ...\n"); for_each_online_cpu(cpu) { @@ -412,6 +421,8 @@ int disable_nonboot_cpus(void) } } + arch_disable_nonboot_cpus_end(); + if (!error) { BUG_ON(num_online_cpus() > 1); /* Make sure the CPUs won't be enabled by someone else */ -- cgit v1.2.3 From a2c606d53ab71dee6410f10ef0adf67321d60e06 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Thu, 16 Dec 2010 10:39:02 -0700 Subject: x86: avoid high BIOS area when allocating address space This prevents allocation of the last 2MB before 4GB. The experiment described here shows Windows 7 ignoring the last 1MB: https://bugzilla.kernel.org/show_bug.cgi?id=23542#c27 This patch ignores the top 2MB instead of just 1MB because H. Peter Anvin says "There will be ROM at the top of the 32-bit address space; it's a fact of the architecture, and on at least older systems it was common to have a shadow 1 MiB below." Acked-by: H. Peter Anvin Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/include/asm/e820.h | 3 +++ arch/x86/kernel/resource.c | 3 ++- 2 files changed, 5 insertions(+), 1 deletion(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/e820.h b/arch/x86/include/asm/e820.h index 5be1542fbfaf..e99d55d74df5 100644 --- a/arch/x86/include/asm/e820.h +++ b/arch/x86/include/asm/e820.h @@ -72,6 +72,9 @@ struct e820map { #define BIOS_BEGIN 0x000a0000 #define BIOS_END 0x00100000 +#define BIOS_ROM_BASE 0xffe00000 +#define BIOS_ROM_END 0xffffffff + #ifdef __KERNEL__ /* see comment in arch/x86/kernel/e820.c */ extern struct e820map e820; diff --git a/arch/x86/kernel/resource.c b/arch/x86/kernel/resource.c index 89638af2ff19..2a26819bb6a8 100644 --- a/arch/x86/kernel/resource.c +++ b/arch/x86/kernel/resource.c @@ -37,10 +37,11 @@ static void remove_e820_regions(struct resource *avail) void arch_remove_reservations(struct resource *avail) { - /* Trim out BIOS area (low 1MB) and E820 regions */ + /* Trim out BIOS areas (low 1MB and high 2MB) and E820 regions */ if (avail->flags & IORESOURCE_MEM) { if (avail->start < BIOS_END) avail->start = BIOS_END; + resource_clip(avail, BIOS_ROM_BASE, BIOS_ROM_END); remove_e820_regions(avail); } -- cgit v1.2.3 From da169f5df2764a6a937cb3b07562e269edfb1c0e Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 24 Sep 2010 15:54:43 +0200 Subject: oprofile, x86: Add support for 6 counters (AMD family 15h) This patch adds support for up to 6 hardware counters for AMD family 15h cpus. There is a new MSR range for hardware counters beginning at MSRC001_0200 Performance Event Select (PERF_CTL0). Signed-off-by: Robert Richter --- arch/x86/include/asm/msr-index.h | 4 +++ arch/x86/oprofile/op_model_amd.c | 54 ++++++++++++++++++++++++++-------------- 2 files changed, 40 insertions(+), 18 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h index 6b89f5e86021..86030f63ba02 100644 --- a/arch/x86/include/asm/msr-index.h +++ b/arch/x86/include/asm/msr-index.h @@ -123,6 +123,10 @@ #define MSR_AMD64_IBSCTL 0xc001103a #define MSR_AMD64_IBSBRTARGET 0xc001103b +/* Fam 15h MSRs */ +#define MSR_F15H_PERF_CTL 0xc0010200 +#define MSR_F15H_PERF_CTR 0xc0010201 + /* Fam 10h MSRs */ #define MSR_FAM10H_MMIO_CONF_BASE 0xc0010058 #define FAM10H_MMIO_CONF_ENABLE (1<<0) diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index a011bcc0f943..f2984d43a6b3 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -29,11 +29,12 @@ #include "op_x86_model.h" #include "op_counter.h" -#define NUM_COUNTERS 4 +#define NUM_COUNTERS 4 +#define NUM_COUNTERS_F15H 6 #ifdef CONFIG_OPROFILE_EVENT_MULTIPLEX -#define NUM_VIRT_COUNTERS 32 +#define NUM_VIRT_COUNTERS 32 #else -#define NUM_VIRT_COUNTERS NUM_COUNTERS +#define NUM_VIRT_COUNTERS 0 #endif #define OP_EVENT_MASK 0x0FFF @@ -41,7 +42,8 @@ #define MSR_AMD_EVENTSEL_RESERVED ((0xFFFFFCF0ULL<<32)|(1ULL<<21)) -static unsigned long reset_value[NUM_VIRT_COUNTERS]; +static int num_counters; +static unsigned long reset_value[OP_MAX_COUNTER]; #define IBS_FETCH_SIZE 6 #define IBS_OP_SIZE 12 @@ -387,7 +389,7 @@ static void op_mux_switch_ctrl(struct op_x86_model_spec const *model, int i; /* enable active counters */ - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { int virt = op_x86_phys_to_virt(i); if (!reset_value[virt]) continue; @@ -406,7 +408,7 @@ static void op_amd_shutdown(struct op_msrs const * const msrs) { int i; - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { if (!msrs->counters[i].addr) continue; release_perfctr_nmi(MSR_K7_PERFCTR0 + i); @@ -418,7 +420,7 @@ static int op_amd_fill_in_addresses(struct op_msrs * const msrs) { int i; - for (i = 0; i < NUM_COUNTERS; i++) { + for (i = 0; i < num_counters; i++) { if (!reserve_perfctr_nmi(MSR_K7_PERFCTR0 + i)) goto fail; if (!reserve_evntsel_nmi(MSR_K7_EVNTSEL0 + i)) { @@ -426,8 +428,13 @@ static int op_amd_fill_in_addresses(struct op_msrs * const msrs) goto fail; } /* both registers must be reserved */ - msrs->counters[i].addr = MSR_K7_PERFCTR0 + i; - msrs->controls[i].addr = MSR_K7_EVNTSEL0 + i; + if (num_counters == NUM_COUNTERS_F15H) { + msrs->counters[i].addr = MSR_F15H_PERF_CTR + (i << 1); + msrs->controls[i].addr = MSR_F15H_PERF_CTL + (i << 1); + } else { + msrs->controls[i].addr = MSR_K7_EVNTSEL0 + i; + msrs->counters[i].addr = MSR_K7_PERFCTR0 + i; + } continue; fail: if (!counter_config[i].enabled) @@ -447,7 +454,7 @@ static void op_amd_setup_ctrs(struct op_x86_model_spec const *model, int i; /* setup reset_value */ - for (i = 0; i < NUM_VIRT_COUNTERS; ++i) { + for (i = 0; i < OP_MAX_COUNTER; ++i) { if (counter_config[i].enabled && msrs->counters[op_x86_virt_to_phys(i)].addr) reset_value[i] = counter_config[i].count; @@ -456,7 +463,7 @@ static void op_amd_setup_ctrs(struct op_x86_model_spec const *model, } /* clear all counters */ - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { if (!msrs->controls[i].addr) continue; rdmsrl(msrs->controls[i].addr, val); @@ -472,7 +479,7 @@ static void op_amd_setup_ctrs(struct op_x86_model_spec const *model, } /* enable active counters */ - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { int virt = op_x86_phys_to_virt(i); if (!reset_value[virt]) continue; @@ -503,7 +510,7 @@ static int op_amd_check_ctrs(struct pt_regs * const regs, u64 val; int i; - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { int virt = op_x86_phys_to_virt(i); if (!reset_value[virt]) continue; @@ -526,7 +533,7 @@ static void op_amd_start(struct op_msrs const * const msrs) u64 val; int i; - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { if (!reset_value[op_x86_phys_to_virt(i)]) continue; rdmsrl(msrs->controls[i].addr, val); @@ -546,7 +553,7 @@ static void op_amd_stop(struct op_msrs const * const msrs) * Subtle: stop on all counters to avoid race with setting our * pm callback */ - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { if (!reset_value[op_x86_phys_to_virt(i)]) continue; rdmsrl(msrs->controls[i].addr, val); @@ -698,18 +705,29 @@ static int setup_ibs_files(struct super_block *sb, struct dentry *root) return 0; } +struct op_x86_model_spec op_amd_spec; + static int op_amd_init(struct oprofile_operations *ops) { init_ibs(); create_arch_files = ops->create_files; ops->create_files = setup_ibs_files; + + if (boot_cpu_data.x86 == 0x15) { + num_counters = NUM_COUNTERS_F15H; + } else { + num_counters = NUM_COUNTERS; + } + + op_amd_spec.num_counters = num_counters; + op_amd_spec.num_controls = num_counters; + op_amd_spec.num_virt_counters = max(num_counters, NUM_VIRT_COUNTERS); + return 0; } struct op_x86_model_spec op_amd_spec = { - .num_counters = NUM_COUNTERS, - .num_controls = NUM_COUNTERS, - .num_virt_counters = NUM_VIRT_COUNTERS, + /* num_counters/num_controls filled in at runtime */ .reserved = MSR_AMD_EVENTSEL_RESERVED, .event_mask = OP_EVENT_MASK, .init = op_amd_init, -- cgit v1.2.3 From 4a7863cc2eb5f9804f1c4e9156619a801cd7f14f Mon Sep 17 00:00:00 2001 From: Don Zickus Date: Wed, 22 Dec 2010 14:00:03 -0500 Subject: x86, nmi_watchdog: Remove ARCH_HAS_NMI_WATCHDOG and rely on CONFIG_HARDLOCKUP_DETECTOR The x86 arch has shifted its use of the nmi_watchdog from a local implementation to the global one provide by kernel/watchdog.c. This shift has caused a whole bunch of compile problems under different config options. I attempt to simplify things with the patch below. In order to simplify things, I had to come to terms with the meaning of two terms ARCH_HAS_NMI_WATCHDOG and CONFIG_HARDLOCKUP_DETECTOR. Basically they mean the same thing, the former on a local level and the latter on a global level. With the old x86 nmi watchdog gone, there is no need to rely on defining the ARCH_HAS_NMI_WATCHDOG variable because it doesn't make sense any more. x86 will now use the global implementation. The changes below do a few things. First it changes the few places that relied on ARCH_HAS_NMI_WATCHDOG to use CONFIG_X86_LOCAL_APIC (the former was an alias for the latter anyway, so nothing unusual here). Those pieces of code were relying more on local apic functionality the nmi watchdog functionality, so the change should make sense. Second, I removed the x86 implementation of touch_nmi_watchdog(). It isn't need now, instead x86 will rely on kernel/watchdog.c's implementation. Third, I removed the #define ARCH_HAS_NMI_WATCHDOG itself from x86. And tweaked the include/linux/nmi.h file to tell users to look for an externally defined touch_nmi_watchdog in the case of ARCH_HAS_NMI_WATCHDOG _or_ CONFIG_HARDLOCKUP_DETECTOR. This changes removes some of the ugliness in that file. Finally, I added a Kconfig dependency for CONFIG_HARDLOCKUP_DETECTOR that said you can't have ARCH_HAS_NMI_WATCHDOG _and_ CONFIG_HARDLOCKUP_DETECTOR. You can only have one nmi_watchdog. Tested with ARCH=i386: allnoconfig, defconfig, allyesconfig, (various broken configs) ARCH=x86_64: allnoconfig, defconfig, allyesconfig, (various broken configs) Hopefully, after this patch I won't get any more compile broken emails. :-) v3: changed a couple of 'linux/nmi.h' -> 'asm/nmi.h' to pick-up correct function prototypes when CONFIG_HARDLOCKUP_DETECTOR is not set. Signed-off-by: Don Zickus Cc: Peter Zijlstra Cc: fweisbec@gmail.com LKML-Reference: <1293044403-14117-1-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/irq.h | 4 ---- arch/x86/include/asm/nmi.h | 2 +- arch/x86/kernel/apic/hw_nmi.c | 10 ---------- arch/x86/kernel/cpu/perfctr-watchdog.c | 2 +- arch/x86/oprofile/op_model_p4.c | 2 +- drivers/watchdog/hpwdt.c | 4 ++-- include/linux/nmi.h | 6 +----- lib/Kconfig.debug | 3 ++- 8 files changed, 8 insertions(+), 25 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/irq.h b/arch/x86/include/asm/irq.h index 13b0ebaa512f..ba870bb6dd8e 100644 --- a/arch/x86/include/asm/irq.h +++ b/arch/x86/include/asm/irq.h @@ -15,10 +15,6 @@ static inline int irq_canonicalize(int irq) return ((irq == 2) ? 9 : irq); } -#ifdef CONFIG_X86_LOCAL_APIC -# define ARCH_HAS_NMI_WATCHDOG -#endif - #ifdef CONFIG_X86_32 extern void irq_ctx_init(int cpu); #else diff --git a/arch/x86/include/asm/nmi.h b/arch/x86/include/asm/nmi.h index 3545838cddeb..c4021b953510 100644 --- a/arch/x86/include/asm/nmi.h +++ b/arch/x86/include/asm/nmi.h @@ -5,7 +5,7 @@ #include #include -#ifdef ARCH_HAS_NMI_WATCHDOG +#ifdef CONFIG_X86_LOCAL_APIC extern void die_nmi(char *str, struct pt_regs *regs, int do_panic); extern int avail_to_resrv_perfctr_nmi_bit(unsigned int); diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c index 93da91df5b38..c57d0b599448 100644 --- a/arch/x86/kernel/apic/hw_nmi.c +++ b/arch/x86/kernel/apic/hw_nmi.c @@ -17,7 +17,6 @@ #include #include -#ifdef ARCH_HAS_NMI_WATCHDOG #ifdef CONFIG_HARDLOCKUP_DETECTOR u64 hw_nmi_get_sample_period(void) { @@ -25,15 +24,6 @@ u64 hw_nmi_get_sample_period(void) } #endif -#ifndef CONFIG_HARDLOCKUP_DETECTOR -void touch_nmi_watchdog(void) -{ - touch_softlockup_watchdog(); -} -EXPORT_SYMBOL(touch_nmi_watchdog); -#endif -#endif - #ifdef arch_trigger_all_cpu_backtrace /* For reliability, we're prepared to waste bits here. */ static DECLARE_BITMAP(backtrace_mask, NR_CPUS) __read_mostly; diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c index 14d45928c282..d5a236615501 100644 --- a/arch/x86/kernel/cpu/perfctr-watchdog.c +++ b/arch/x86/kernel/cpu/perfctr-watchdog.c @@ -16,7 +16,7 @@ #include #include #include -#include +#include #include #include diff --git a/arch/x86/oprofile/op_model_p4.c b/arch/x86/oprofile/op_model_p4.c index 182558dd5515..9fadec074142 100644 --- a/arch/x86/oprofile/op_model_p4.c +++ b/arch/x86/oprofile/op_model_p4.c @@ -11,7 +11,7 @@ #include #include #include -#include +#include #include #include #include diff --git a/drivers/watchdog/hpwdt.c b/drivers/watchdog/hpwdt.c index c19f4a20794a..dea7b5bf6e2c 100644 --- a/drivers/watchdog/hpwdt.c +++ b/drivers/watchdog/hpwdt.c @@ -642,7 +642,7 @@ static struct notifier_block die_notifier = { */ #ifdef CONFIG_HPWDT_NMI_DECODING -#ifdef ARCH_HAS_NMI_WATCHDOG +#ifdef CONFIG_X86_LOCAL_APIC static void __devinit hpwdt_check_nmi_decoding(struct pci_dev *dev) { /* @@ -657,7 +657,7 @@ static void __devinit hpwdt_check_nmi_decoding(struct pci_dev *dev) dev_warn(&dev->dev, "NMI decoding is disabled. " "Your kernel does not support a NMI Watchdog.\n"); } -#endif /* ARCH_HAS_NMI_WATCHDOG */ +#endif /* CONFIG_X86_LOCAL_APIC */ static int __devinit hpwdt_init_nmi_decoding(struct pci_dev *dev) { diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 17ccf44e7dcb..c536f8545f74 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -14,18 +14,14 @@ * may be used to reset the timeout - for code which intentionally * disables interrupts for a long time. This call is stateless. */ -#ifdef ARCH_HAS_NMI_WATCHDOG +#if defined(ARCH_HAS_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR) #include extern void touch_nmi_watchdog(void); #else -#ifndef CONFIG_HARDLOCKUP_DETECTOR static inline void touch_nmi_watchdog(void) { touch_softlockup_watchdog(); } -#else -extern void touch_nmi_watchdog(void); -#endif #endif /* diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 28b42b9274d0..2d05adb98401 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -173,7 +173,8 @@ config LOCKUP_DETECTOR An NMI is generated every 60 seconds or so to check for hardlockups. config HARDLOCKUP_DETECTOR - def_bool LOCKUP_DETECTOR && PERF_EVENTS && HAVE_PERF_EVENTS_NMI + def_bool LOCKUP_DETECTOR && PERF_EVENTS && HAVE_PERF_EVENTS_NMI && \ + !ARCH_HAS_NMI_WATCHDOG config BOOTPARAM_SOFTLOCKUP_PANIC bool "Panic (Reboot) On Soft Lockups" -- cgit v1.2.3 From 56d91f132c9be66e98cce1b1e77a28027048bb26 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Thu, 16 Dec 2010 19:09:24 -0800 Subject: x86, acpi: Add MAX_LOCAL_APIC for 32bit We should use MAX_LOCAL_APIC for max apic ids and MAX_APICS as number of local apics. Also apic_version[] array should use MAX_LOCAL_APICs. Signed-off-by: Yinghai Lu LKML-Reference: <4D0AD464.2020408@kernel.org> Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/apicdef.h | 1 + arch/x86/include/asm/mpspec.h | 2 +- arch/x86/kernel/apic/apic.c | 2 +- 3 files changed, 3 insertions(+), 2 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/apicdef.h b/arch/x86/include/asm/apicdef.h index a859ca461fb0..47a30ff8e517 100644 --- a/arch/x86/include/asm/apicdef.h +++ b/arch/x86/include/asm/apicdef.h @@ -145,6 +145,7 @@ #ifdef CONFIG_X86_32 # define MAX_IO_APICS 64 +# define MAX_LOCAL_APIC 256 #else # define MAX_IO_APICS 128 # define MAX_LOCAL_APIC 32768 diff --git a/arch/x86/include/asm/mpspec.h b/arch/x86/include/asm/mpspec.h index c82868e9f905..7c1aebf8fcbf 100644 --- a/arch/x86/include/asm/mpspec.h +++ b/arch/x86/include/asm/mpspec.h @@ -6,7 +6,7 @@ #include #include -extern int apic_version[MAX_APICS]; +extern int apic_version[]; extern int pic_mode; #ifdef CONFIG_X86_32 diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 463839645f9b..0d5d07f2253e 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -1707,7 +1707,7 @@ void __init init_apic_mappings(void) * This initializes the IO-APIC and APIC hardware if this is * a UP kernel. */ -int apic_version[MAX_APICS]; +int apic_version[MAX_LOCAL_APIC]; int __init APIC_init_uniprocessor(void) { -- cgit v1.2.3 From c8217b8305e5e75c23617f2f4cd262527d952c0a Mon Sep 17 00:00:00 2001 From: Cliff Wickman Date: Mon, 13 Dec 2010 10:51:57 -0600 Subject: x86, paravirt: Use native_halt on a halt, not native_safe_halt halt() should use native_halt() safe_halt() uses native_safe_halt() If CONFIG_PARAVIRT=y, halt() is defined in arch/x86/include/asm/paravirt.h as static inline void halt(void) { PVOP_VCALL0(pv_irq_ops.safe_halt); } Otherwise (no CONFIG_PARAVIRT) halt() in arch/x86/include/asm/irqflags.h is static inline void halt(void) { native_halt(); } So it looks to me like the CONFIG_PARAVIRT case of using native_safe_halt() for a halt() is an oversight. Am I missing something? It probably hasn't shown up as a problem because the local apic is disabled on a shutdown or restart. But if we disable interrupts and call halt() we shouldn't expect that the halt() will re-enable interrupts. Signed-off-by: Cliff Wickman LKML-Reference: Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/paravirt.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h index ef9975812c77..7709c12431b8 100644 --- a/arch/x86/include/asm/paravirt.h +++ b/arch/x86/include/asm/paravirt.h @@ -112,7 +112,7 @@ static inline void arch_safe_halt(void) static inline void halt(void) { - PVOP_VCALL0(pv_irq_ops.safe_halt); + PVOP_VCALL0(pv_irq_ops.halt); } static inline void wbinvd(void) -- cgit v1.2.3 From 9e76a97efd31a08cb19d0ba12013b8fb4ad3e474 Mon Sep 17 00:00:00 2001 From: "R, Durgadoss" Date: Mon, 3 Jan 2011 17:22:04 +0530 Subject: x86, hwmon: Add core threshold notification to therm_throt.c This patch adds code to therm_throt.c to notify core thermal threshold events. These thresholds are supported by the IA32_THERM_INTERRUPT register. The status/log for the same is monitored using the IA32_THERM_STATUS register. The necessary #defines are in msr-index.h. A call back is added to mce.h, to further notify the thermal stack, about the threshold events. Signed-off-by: Durgadoss R LKML-Reference: Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/mce.h | 3 +++ arch/x86/include/asm/msr-index.h | 12 ++++++++++ arch/x86/kernel/cpu/mcheck/therm_throt.c | 40 ++++++++++++++++++++++++++++++++ 3 files changed, 55 insertions(+) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/mce.h b/arch/x86/include/asm/mce.h index c62c13cb9788..eb16e94ae04f 100644 --- a/arch/x86/include/asm/mce.h +++ b/arch/x86/include/asm/mce.h @@ -223,6 +223,9 @@ void intel_init_thermal(struct cpuinfo_x86 *c); void mce_log_therm_throt_event(__u64 status); +/* Interrupt Handler for core thermal thresholds */ +extern int (*platform_thermal_notify)(__u64 msr_val); + #ifdef CONFIG_X86_THERMAL_VECTOR extern void mcheck_intel_therm_init(void); #else diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h index 6b89f5e86021..622c80b7dbee 100644 --- a/arch/x86/include/asm/msr-index.h +++ b/arch/x86/include/asm/msr-index.h @@ -253,6 +253,18 @@ #define PACKAGE_THERM_INT_LOW_ENABLE (1 << 1) #define PACKAGE_THERM_INT_PLN_ENABLE (1 << 24) +/* Thermal Thresholds Support */ +#define THERM_INT_THRESHOLD0_ENABLE (1 << 15) +#define THERM_SHIFT_THRESHOLD0 8 +#define THERM_MASK_THRESHOLD0 (0x7f << THERM_SHIFT_THRESHOLD0) +#define THERM_INT_THRESHOLD1_ENABLE (1 << 23) +#define THERM_SHIFT_THRESHOLD1 16 +#define THERM_MASK_THRESHOLD1 (0x7f << THERM_SHIFT_THRESHOLD1) +#define THERM_STATUS_THRESHOLD0 (1 << 6) +#define THERM_LOG_THRESHOLD0 (1 << 7) +#define THERM_STATUS_THRESHOLD1 (1 << 8) +#define THERM_LOG_THRESHOLD1 (1 << 9) + /* MISC_ENABLE bits: architectural */ #define MSR_IA32_MISC_ENABLE_FAST_STRING (1ULL << 0) #define MSR_IA32_MISC_ENABLE_TCC (1ULL << 1) diff --git a/arch/x86/kernel/cpu/mcheck/therm_throt.c b/arch/x86/kernel/cpu/mcheck/therm_throt.c index 4b683267eca5..e12246ff5aa6 100644 --- a/arch/x86/kernel/cpu/mcheck/therm_throt.c +++ b/arch/x86/kernel/cpu/mcheck/therm_throt.c @@ -53,8 +53,13 @@ struct thermal_state { struct _thermal_state core_power_limit; struct _thermal_state package_throttle; struct _thermal_state package_power_limit; + struct _thermal_state core_thresh0; + struct _thermal_state core_thresh1; }; +/* Callback to handle core threshold interrupts */ +int (*platform_thermal_notify)(__u64 msr_val); + static DEFINE_PER_CPU(struct thermal_state, thermal_state); static atomic_t therm_throt_en = ATOMIC_INIT(0); @@ -200,6 +205,22 @@ static int therm_throt_process(bool new_event, int event, int level) return 0; } +static int thresh_event_valid(int event) +{ + struct _thermal_state *state; + unsigned int this_cpu = smp_processor_id(); + struct thermal_state *pstate = &per_cpu(thermal_state, this_cpu); + u64 now = get_jiffies_64(); + + state = (event == 0) ? &pstate->core_thresh0 : &pstate->core_thresh1; + + if (time_before64(now, state->next_check)) + return 0; + + state->next_check = now + CHECK_INTERVAL; + return 1; +} + #ifdef CONFIG_SYSFS /* Add/Remove thermal_throttle interface for CPU device: */ static __cpuinit int thermal_throttle_add_dev(struct sys_device *sys_dev, @@ -313,6 +334,22 @@ device_initcall(thermal_throttle_init_device); #define PACKAGE_THROTTLED ((__u64)2 << 62) #define PACKAGE_POWER_LIMIT ((__u64)3 << 62) +static void notify_thresholds(__u64 msr_val) +{ + /* check whether the interrupt handler is defined; + * otherwise simply return + */ + if (!platform_thermal_notify) + return; + + /* lower threshold reached */ + if ((msr_val & THERM_LOG_THRESHOLD0) && thresh_event_valid(0)) + platform_thermal_notify(msr_val); + /* higher threshold reached */ + if ((msr_val & THERM_LOG_THRESHOLD1) && thresh_event_valid(1)) + platform_thermal_notify(msr_val); +} + /* Thermal transition interrupt handler */ static void intel_thermal_interrupt(void) { @@ -321,6 +358,9 @@ static void intel_thermal_interrupt(void) rdmsrl(MSR_IA32_THERM_STATUS, msr_val); + /* Check for violation of core thermal thresholds*/ + notify_thresholds(msr_val); + if (therm_throt_process(msr_val & THERM_STATUS_PROCHOT, THERMAL_THROTTLING_EVENT, CORE_LEVEL) != 0) -- cgit v1.2.3 From cfa60917f0ba6eca83f41aef3cb4a7dd7736ac9f Mon Sep 17 00:00:00 2001 From: Cliff Wickman Date: Mon, 3 Jan 2011 12:03:53 -0600 Subject: x86, UV, BAU: Extend for more than 16 cpus per socket Fix a hard-coded limit of a maximum of 16 cpu's per socket. The UV Broadcast Assist Unit code initializes by scanning the cpu topology of the system and assigning a master cpu for each socket and UV hub. That scan had an assumption of a limit of 16 cpus per socket. With Westmere we are going over that limit. The UV hub hardware will allow up to 32. If the scan finds the system has gone over that limit it returns an error and we print a warning and fall back to doing TLB shootdowns without the BAU. Signed-off-by: Cliff Wickman Cc: # .37.x LKML-Reference: Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uv/uv_bau.h | 9 +++++---- arch/x86/platform/uv/tlb_uv.c | 22 ++++++++++++++++++---- 2 files changed, 23 insertions(+), 8 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/uv/uv_bau.h b/arch/x86/include/asm/uv/uv_bau.h index 42d412fd8b02..ce1d54c8a433 100644 --- a/arch/x86/include/asm/uv/uv_bau.h +++ b/arch/x86/include/asm/uv/uv_bau.h @@ -26,20 +26,22 @@ * BAU_SB_DESCRIPTOR_BASE register, set 1 is located at BASE + 512, * set 2 is at BASE + 2*512, set 3 at BASE + 3*512, and so on. * - * We will use 31 sets, one for sending BAU messages from each of the 32 + * We will use one set for sending BAU messages from each of the * cpu's on the uvhub. * * TLB shootdown will use the first of the 8 descriptors of each set. * Each of the descriptors is 64 bytes in size (8*64 = 512 bytes in a set). */ +#define MAX_CPUS_PER_UVHUB 64 +#define MAX_CPUS_PER_SOCKET 32 +#define UV_ADP_SIZE 64 /* hardware-provided max. */ +#define UV_CPUS_PER_ACT_STATUS 32 /* hardware-provided max. */ #define UV_ITEMS_PER_DESCRIPTOR 8 /* the 'throttle' to prevent the hardware stay-busy bug */ #define MAX_BAU_CONCURRENT 3 -#define UV_CPUS_PER_ACT_STATUS 32 #define UV_ACT_STATUS_MASK 0x3 #define UV_ACT_STATUS_SIZE 2 -#define UV_ADP_SIZE 32 #define UV_DISTRIBUTION_SIZE 256 #define UV_SW_ACK_NPENDING 8 #define UV_NET_ENDPOINT_INTD 0x38 @@ -100,7 +102,6 @@ * number of destination side software ack resources */ #define DEST_NUM_RESOURCES 8 -#define MAX_CPUS_PER_NODE 32 /* * completion statuses for sending a TLB flush message */ diff --git a/arch/x86/platform/uv/tlb_uv.c b/arch/x86/platform/uv/tlb_uv.c index ba9caa808a9c..df58e9cad96a 100644 --- a/arch/x86/platform/uv/tlb_uv.c +++ b/arch/x86/platform/uv/tlb_uv.c @@ -1341,7 +1341,7 @@ uv_activation_descriptor_init(int node, int pnode) /* * each bau_desc is 64 bytes; there are 8 (UV_ITEMS_PER_DESCRIPTOR) - * per cpu; and up to 32 (UV_ADP_SIZE) cpu's per uvhub + * per cpu; and one per cpu on the uvhub (UV_ADP_SIZE) */ bau_desc = kmalloc_node(sizeof(struct bau_desc) * UV_ADP_SIZE * UV_ITEMS_PER_DESCRIPTOR, GFP_KERNEL, node); @@ -1490,7 +1490,7 @@ calculate_destination_timeout(void) /* * initialize the bau_control structure for each cpu */ -static void __init uv_init_per_cpu(int nuvhubs) +static int __init uv_init_per_cpu(int nuvhubs) { int i; int cpu; @@ -1507,7 +1507,7 @@ static void __init uv_init_per_cpu(int nuvhubs) struct bau_control *smaster = NULL; struct socket_desc { short num_cpus; - short cpu_number[16]; + short cpu_number[MAX_CPUS_PER_SOCKET]; }; struct uvhub_desc { unsigned short socket_mask; @@ -1540,6 +1540,10 @@ static void __init uv_init_per_cpu(int nuvhubs) sdp = &bdp->socket[socket]; sdp->cpu_number[sdp->num_cpus] = cpu; sdp->num_cpus++; + if (sdp->num_cpus > MAX_CPUS_PER_SOCKET) { + printk(KERN_EMERG "%d cpus per socket invalid\n", sdp->num_cpus); + return 1; + } } for (uvhub = 0; uvhub < nuvhubs; uvhub++) { if (!(*(uvhub_mask + (uvhub/8)) & (1 << (uvhub%8)))) @@ -1570,6 +1574,12 @@ static void __init uv_init_per_cpu(int nuvhubs) bcp->uvhub_master = hmaster; bcp->uvhub_cpu = uv_cpu_hub_info(cpu)-> blade_processor_id; + if (bcp->uvhub_cpu >= MAX_CPUS_PER_UVHUB) { + printk(KERN_EMERG + "%d cpus per uvhub invalid\n", + bcp->uvhub_cpu); + return 1; + } } nextsocket: socket++; @@ -1595,6 +1605,7 @@ nextsocket: bcp->congested_reps = congested_reps; bcp->congested_period = congested_period; } + return 0; } /* @@ -1625,7 +1636,10 @@ static int __init uv_bau_init(void) spin_lock_init(&disable_lock); congested_cycles = microsec_2_cycles(congested_response_us); - uv_init_per_cpu(nuvhubs); + if (uv_init_per_cpu(nuvhubs)) { + nobau = 1; + return 0; + } uv_partition_base_pnode = 0x7fffffff; for (uvhub = 0; uvhub < nuvhubs; uvhub++) -- cgit v1.2.3 From cb2ded37fd2e1039f96c8c892da024a8f033add5 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Tue, 4 Jan 2011 16:38:52 -0800 Subject: x86: Fix APIC ID sizing bug on larger systems, clean up MAX_APICS confusion Found one x2apic pre-enabled system, x2apic_mode suddenly get corrupted after register some cpus, when compiled CONFIG_NR_CPUS=255 instead of 512. It turns out that generic_processor_info() ==> phyid_set(apicid, phys_cpu_present_map) causes the problem. phys_cpu_present_map is sized by MAX_APICS bits, and pre-enabled system some cpus have an apic id > 255. The variable after phys_cpu_present_map may get corrupted silently: ffffffff828e8420 B phys_cpu_present_map ffffffff828e8440 B apic_verbosity ffffffff828e8444 B local_apic_timer_c2_ok ffffffff828e8448 B disable_apic ffffffff828e844c B x2apic_mode ffffffff828e8450 B x2apic_disabled ffffffff828e8454 B num_processors ... Actually phys_cpu_present_map is referenced via apic id, instead index. We should use MAX_LOCAL_APIC instead MAX_APICS. For 64-bit it will be 32768 in all cases. BSS will increase by 4k bytes on 64-bit: text data bss dec filename 21696943 4193748 12787712 38678403 vmlinux.before 21696943 4193748 12791808 38682499 vmlinux.after No change on 32bit. Finally we can remove MAX_APCIS that was rather confusing. Signed-off-by: Yinghai Lu Cc: H. Peter Anvin Cc: "Eric W. Biederman" LKML-Reference: <4D23BD9C.3070102@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/mpspec.h | 29 +++++++++++------------------ arch/x86/include/asm/mpspec_def.h | 7 ------- arch/x86/kernel/acpi/boot.c | 6 +++--- arch/x86/kernel/apic/io_apic.c | 3 ++- arch/x86/platform/sfi/sfi.c | 4 ++-- arch/x86/platform/visws/visws_quirks.c | 2 +- 6 files changed, 19 insertions(+), 32 deletions(-) (limited to 'arch/x86/include') diff --git a/arch/x86/include/asm/mpspec.h b/arch/x86/include/asm/mpspec.h index 7c1aebf8fcbf..0c90dd9f0505 100644 --- a/arch/x86/include/asm/mpspec.h +++ b/arch/x86/include/asm/mpspec.h @@ -5,6 +5,7 @@ #include #include +#include extern int apic_version[]; extern int pic_mode; @@ -107,7 +108,7 @@ extern int mp_register_gsi(struct device *dev, u32 gsi, int edge_level, int active_high_low); #endif /* CONFIG_ACPI */ -#define PHYSID_ARRAY_SIZE BITS_TO_LONGS(MAX_APICS) +#define PHYSID_ARRAY_SIZE BITS_TO_LONGS(MAX_LOCAL_APIC) struct physid_mask { unsigned long mask[PHYSID_ARRAY_SIZE]; @@ -122,31 +123,31 @@ typedef struct physid_mask physid_mask_t; test_and_set_bit(physid, (map).mask) #define physids_and(dst, src1, src2) \ - bitmap_and((dst).mask, (src1).mask, (src2).mask, MAX_APICS) + bitmap_and((dst).mask, (src1).mask, (src2).mask, MAX_LOCAL_APIC) #define physids_or(dst, src1, src2) \ - bitmap_or((dst).mask, (src1).mask, (src2).mask, MAX_APICS) + bitmap_or((dst).mask, (src1).mask, (src2).mask, MAX_LOCAL_APIC) #define physids_clear(map) \ - bitmap_zero((map).mask, MAX_APICS) + bitmap_zero((map).mask, MAX_LOCAL_APIC) #define physids_complement(dst, src) \ - bitmap_complement((dst).mask, (src).mask, MAX_APICS) + bitmap_complement((dst).mask, (src).mask, MAX_LOCAL_APIC) #define physids_empty(map) \ - bitmap_empty((map).mask, MAX_APICS) + bitmap_empty((map).mask, MAX_LOCAL_APIC) #define physids_equal(map1, map2) \ - bitmap_equal((map1).mask, (map2).mask, MAX_APICS) + bitmap_equal((map1).mask, (map2).mask, MAX_LOCAL_APIC) #define physids_weight(map) \ - bitmap_weight((map).mask, MAX_APICS) + bitmap_weight((map).mask, MAX_LOCAL_APIC) #define physids_shift_right(d, s, n) \ - bitmap_shift_right((d).mask, (s).mask, n, MAX_APICS) + bitmap_shift_right((d).mask, (s).mask, n, MAX_LOCAL_APIC) #define physids_shift_left(d, s, n) \ - bitmap_shift_left((d).mask, (s).mask, n, MAX_APICS) + bitmap_shift_left((d).mask, (s).mask, n, MAX_LOCAL_APIC) static inline unsigned long physids_coerce(physid_mask_t *map) { @@ -159,14 +160,6 @@ static inline void physids_promote(unsigned long physids, physid_mask_t *map) map->mask[0] = physids; } -/* Note: will create very large stack frames if physid_mask_t is big */ -#define physid_mask_of_physid(physid) \ - ({ \ - physid_mask_t __physid_mask = PHYSID_MASK_NONE; \ - physid_set(physid, __physid_mask); \ - __physid_mask; \ - }) - static inline void physid_set_mask_of_physid(int physid, physid_mask_t *map) { physids_clear(*map); diff --git a/arch/x86/include/asm/mpspec_def.h b/arch/x86/include/asm/mpspec_def.h index 4a7f96d7c188..c0a955a9a087 100644 --- a/arch/x86/include/asm/mpspec_def.h +++ b/arch/x86/include/asm/mpspec_def.h @@ -15,13 +15,6 @@ #ifdef CONFIG_X86_32 # define MAX_MPC_ENTRY 1024 -# define MAX_APICS 256 -#else -# if NR_CPUS <= 255 -# define MAX_APICS 255 -# else -# define MAX_APICS 32768 -# endif #endif /* Intel MP Floating Pointer Structure */ diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 7235e5fbdb6d..17c8090fabd4 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -915,13 +915,13 @@ static int __init acpi_parse_madt_lapic_entries(void) acpi_register_lapic_address(acpi_lapic_addr); count = acpi_table_parse_madt(ACPI_MADT_TYPE_LOCAL_SAPIC, - acpi_parse_sapic, MAX_APICS); + acpi_parse_sapic, MAX_LOCAL_APIC); if (!count) { x2count = acpi_table_parse_madt(ACPI_MADT_TYPE_LOCAL_X2APIC, - acpi_parse_x2apic, MAX_APICS); + acpi_parse_x2apic, MAX_LOCAL_APIC); count = acpi_table_parse_madt(ACPI_MADT_TYPE_LOCAL_APIC, - acpi_parse_lapic, MAX_APICS); + acpi_parse_lapic, MAX_LOCAL_APIC); } if (!count && !x2count) { printk(KERN_ERR PREFIX "No LAPIC entries present\n"); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 943d814ef8e4..2fc696e4d565 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -4109,7 +4109,8 @@ void __init pre_init_apic_IRQ0(void) printk(KERN_INFO "Early APIC setup for system timer0\n"); #ifndef CONFIG_SMP - phys_cpu_present_map = physid_mask_of_physid(boot_cpu_physical_apicid); + physid_set_mask_of_physid(boot_cpu_physical_apicid, + &phys_cpu_present_map); #endif /* Make sure the irq descriptor is set up */ cfg = alloc_irq_and_cfg_at(0, 0); diff --git a/arch/x86/platform/sfi/sfi.c b/arch/x86/platform/sfi/sfi.c index dd4c281ffe57..ca54875ac795 100644 --- a/arch/x86/platform/sfi/sfi.c +++ b/arch/x86/platform/sfi/sfi.c @@ -48,9 +48,9 @@ static void __init mp_sfi_register_lapic_address(unsigned long address) /* All CPUs enumerated by SFI must be present and enabled */ static void __cpuinit mp_sfi_register_lapic(u8 id) { - if (MAX_APICS - id <= 0) { + if (MAX_LOCAL_APIC - id <= 0) { pr_warning("Processor #%d invalid (max %d)\n", - id, MAX_APICS); + id, MAX_LOCAL_APIC); return; } diff --git a/arch/x86/platform/visws/visws_quirks.c b/arch/x86/platform/visws/visws_quirks.c index 3371bd053b89..632037671746 100644 --- a/arch/x86/platform/visws/visws_quirks.c +++ b/arch/x86/platform/visws/visws_quirks.c @@ -171,7 +171,7 @@ static void __init MP_processor_info(struct mpc_cpu *m) ver = m->apicver; if ((ver >= 0x14 && m->apicid >= 0xff) || m->apicid >= 0xf) { printk(KERN_ERR "Processor #%d INVALID. (Max ID: %d).\n", - m->apicid, MAX_APICS); + m->apicid, MAX_LOCAL_APIC); return; } -- cgit v1.2.3