From 11d1578f9454159c43499d1d8fe8a7d728c176a3 Mon Sep 17 00:00:00 2001 From: Vince Weaver Date: Wed, 8 Jul 2009 17:46:14 -0400 Subject: perf_counter: Add P6 PMU support Add basic P6 PMU support. The P6 uses the EVNTSEL0 EN bit to enable/disable both its counters. We use this for the global enable/disable, and clear all config bits (except EN) to disable individual counters. Actual ia32 hardware doesn't support lfence, so use a locked op without side-effect to implement a full barrier. perf stat and perf record seem to function correctly. [a.p.zijlstra@chello.nl: cleanups and complete the enable/disable code] Signed-off-by: Vince Weaver Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_counter.c | 227 ++++++++++++++++++++++++++++++++++--- 1 file changed, 213 insertions(+), 14 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c index 36c3dc7b8991..1910f39ff19a 100644 --- a/arch/x86/kernel/cpu/perf_counter.c +++ b/arch/x86/kernel/cpu/perf_counter.c @@ -65,6 +65,44 @@ static DEFINE_PER_CPU(struct cpu_hw_counters, cpu_hw_counters) = { .enabled = 1, }; +/* + * Not sure about some of these + */ +static const u64 p6_perfmon_event_map[] = +{ + [PERF_COUNT_HW_CPU_CYCLES] = 0x0079, + [PERF_COUNT_HW_INSTRUCTIONS] = 0x00c0, + [PERF_COUNT_HW_CACHE_REFERENCES] = 0x0000, + [PERF_COUNT_HW_CACHE_MISSES] = 0x0000, + [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = 0x00c4, + [PERF_COUNT_HW_BRANCH_MISSES] = 0x00c5, + [PERF_COUNT_HW_BUS_CYCLES] = 0x0062, +}; + +static u64 p6_pmu_event_map(int event) +{ + return p6_perfmon_event_map[event]; +} + +static u64 p6_pmu_raw_event(u64 event) +{ +#define P6_EVNTSEL_EVENT_MASK 0x000000FFULL +#define P6_EVNTSEL_UNIT_MASK 0x0000FF00ULL +#define P6_EVNTSEL_EDGE_MASK 0x00040000ULL +#define P6_EVNTSEL_INV_MASK 0x00800000ULL +#define P6_EVNTSEL_COUNTER_MASK 0xFF000000ULL + +#define P6_EVNTSEL_MASK \ + (P6_EVNTSEL_EVENT_MASK | \ + P6_EVNTSEL_UNIT_MASK | \ + P6_EVNTSEL_EDGE_MASK | \ + P6_EVNTSEL_INV_MASK | \ + P6_EVNTSEL_COUNTER_MASK) + + return event & P6_EVNTSEL_MASK; +} + + /* * Intel PerfMon v3. Used on Core2 and later. */ @@ -726,6 +764,23 @@ static int __hw_perf_counter_init(struct perf_counter *counter) return 0; } +static void p6_pmu_disable_all(void) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + unsigned long val; + + if (!cpuc->enabled) + return; + + cpuc->enabled = 0; + barrier(); + + /* p6 only has one enable register */ + rdmsrl(MSR_P6_EVNTSEL0, val); + val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE; + wrmsrl(MSR_P6_EVNTSEL0, val); +} + static void intel_pmu_disable_all(void) { wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0); @@ -767,6 +822,23 @@ void hw_perf_disable(void) return x86_pmu.disable_all(); } +static void p6_pmu_enable_all(void) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + unsigned long val; + + if (cpuc->enabled) + return; + + cpuc->enabled = 1; + barrier(); + + /* p6 only has one enable register */ + rdmsrl(MSR_P6_EVNTSEL0, val); + val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + wrmsrl(MSR_P6_EVNTSEL0, val); +} + static void intel_pmu_enable_all(void) { wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl); @@ -819,16 +891,13 @@ static inline void intel_pmu_ack_status(u64 ack) static inline void x86_pmu_enable_counter(struct hw_perf_counter *hwc, int idx) { - int err; - err = checking_wrmsrl(hwc->config_base + idx, + (void)checking_wrmsrl(hwc->config_base + idx, hwc->config | ARCH_PERFMON_EVENTSEL0_ENABLE); } static inline void x86_pmu_disable_counter(struct hw_perf_counter *hwc, int idx) { - int err; - err = checking_wrmsrl(hwc->config_base + idx, - hwc->config); + (void)checking_wrmsrl(hwc->config_base + idx, hwc->config); } static inline void @@ -836,13 +905,24 @@ intel_pmu_disable_fixed(struct hw_perf_counter *hwc, int __idx) { int idx = __idx - X86_PMC_IDX_FIXED; u64 ctrl_val, mask; - int err; mask = 0xfULL << (idx * 4); rdmsrl(hwc->config_base, ctrl_val); ctrl_val &= ~mask; - err = checking_wrmsrl(hwc->config_base, ctrl_val); + (void)checking_wrmsrl(hwc->config_base, ctrl_val); +} + +static inline void +p6_pmu_disable_counter(struct hw_perf_counter *hwc, int idx) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + unsigned long val = ARCH_PERFMON_EVENTSEL0_ENABLE; + + if (!cpuc->enabled) + val = 0; + + (void)checking_wrmsrl(hwc->config_base + idx, val); } static inline void @@ -943,6 +1023,17 @@ intel_pmu_enable_fixed(struct hw_perf_counter *hwc, int __idx) err = checking_wrmsrl(hwc->config_base, ctrl_val); } +static void p6_pmu_enable_counter(struct hw_perf_counter *hwc, int idx) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + + if (cpuc->enabled) + x86_pmu_enable_counter(hwc, idx); + else + x86_pmu_disable_counter(hwc, idx); +} + + static void intel_pmu_enable_counter(struct hw_perf_counter *hwc, int idx) { if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) { @@ -1176,6 +1267,49 @@ static void intel_pmu_reset(void) local_irq_restore(flags); } +static int p6_pmu_handle_irq(struct pt_regs *regs) +{ + struct perf_sample_data data; + struct cpu_hw_counters *cpuc; + struct perf_counter *counter; + struct hw_perf_counter *hwc; + int idx, handled = 0; + u64 val; + + data.regs = regs; + data.addr = 0; + + cpuc = &__get_cpu_var(cpu_hw_counters); + + for (idx = 0; idx < x86_pmu.num_counters; idx++) { + if (!test_bit(idx, cpuc->active_mask)) + continue; + + counter = cpuc->counters[idx]; + hwc = &counter->hw; + + val = x86_perf_counter_update(counter, hwc, idx); + if (val & (1ULL << (x86_pmu.counter_bits - 1))) + continue; + + /* + * counter overflow + */ + handled = 1; + data.period = counter->hw.last_period; + + if (!x86_perf_counter_set_period(counter, hwc, idx)) + continue; + + if (perf_counter_overflow(counter, 1, &data)) + p6_pmu_disable_counter(hwc, idx); + } + + if (handled) + inc_irq_stat(apic_perf_irqs); + + return handled; +} /* * This handler is triggered by the local APIC, so the APIC IRQ handling @@ -1185,14 +1319,13 @@ static int intel_pmu_handle_irq(struct pt_regs *regs) { struct perf_sample_data data; struct cpu_hw_counters *cpuc; - int bit, cpu, loops; + int bit, loops; u64 ack, status; data.regs = regs; data.addr = 0; - cpu = smp_processor_id(); - cpuc = &per_cpu(cpu_hw_counters, cpu); + cpuc = &__get_cpu_var(cpu_hw_counters); perf_disable(); status = intel_pmu_get_status(); @@ -1249,14 +1382,13 @@ static int amd_pmu_handle_irq(struct pt_regs *regs) struct cpu_hw_counters *cpuc; struct perf_counter *counter; struct hw_perf_counter *hwc; - int cpu, idx, handled = 0; + int idx, handled = 0; u64 val; data.regs = regs; data.addr = 0; - cpu = smp_processor_id(); - cpuc = &per_cpu(cpu_hw_counters, cpu); + cpuc = &__get_cpu_var(cpu_hw_counters); for (idx = 0; idx < x86_pmu.num_counters; idx++) { if (!test_bit(idx, cpuc->active_mask)) @@ -1353,6 +1485,32 @@ static __read_mostly struct notifier_block perf_counter_nmi_notifier = { .priority = 1 }; +static struct x86_pmu p6_pmu = { + .name = "p6", + .handle_irq = p6_pmu_handle_irq, + .disable_all = p6_pmu_disable_all, + .enable_all = p6_pmu_enable_all, + .enable = p6_pmu_enable_counter, + .disable = p6_pmu_disable_counter, + .eventsel = MSR_P6_EVNTSEL0, + .perfctr = MSR_P6_PERFCTR0, + .event_map = p6_pmu_event_map, + .raw_event = p6_pmu_raw_event, + .max_events = ARRAY_SIZE(p6_perfmon_event_map), + .max_period = (1ULL << 31) - 1, + .version = 0, + .num_counters = 2, + /* + * Counters have 40 bits implemented. However they are designed such + * that bits [32-39] are sign extensions of bit 31. As such the + * effective width of a counter for P6-like PMU is 32 bits only. + * + * See IA-32 Intel Architecture Software developer manual Vol 3B + */ + .counter_bits = 32, + .counter_mask = (1ULL << 32) - 1, +}; + static struct x86_pmu intel_pmu = { .name = "Intel", .handle_irq = intel_pmu_handle_irq, @@ -1392,6 +1550,41 @@ static struct x86_pmu amd_pmu = { .max_period = (1ULL << 47) - 1, }; +static int p6_pmu_init(void) +{ + int high, low; + + switch (boot_cpu_data.x86_model) { + case 1: + case 3: /* Pentium Pro */ + case 5: + case 6: /* Pentium II */ + case 7: + case 8: + case 11: /* Pentium III */ + break; + case 9: + case 13: + /* for Pentium M, we need to check if PMU exist */ + rdmsr(MSR_IA32_MISC_ENABLE, low, high); + if (low & MSR_IA32_MISC_ENABLE_EMON) + break; + default: + pr_cont("unsupported p6 CPU model %d ", + boot_cpu_data.x86_model); + return -ENODEV; + } + + if (!cpu_has_apic) { + pr_info("no Local APIC, try rebooting with lapic"); + return -ENODEV; + } + + x86_pmu = p6_pmu; + + return 0; +} + static int intel_pmu_init(void) { union cpuid10_edx edx; @@ -1400,8 +1593,14 @@ static int intel_pmu_init(void) unsigned int ebx; int version; - if (!cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) + if (!cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) { + /* check for P6 processor family */ + if (boot_cpu_data.x86 == 6) { + return p6_pmu_init(); + } else { return -ENODEV; + } + } /* * Check whether the Architectural PerfMon supports -- cgit v1.2.2 From 9c74fb50867e8fb5f3be3be06716492c0f79309e Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 8 Jul 2009 10:21:41 +0200 Subject: perf_counter: Fix up P6 PMU details The P6 doesn't seem to support cache ref/hit/miss counts, so we extend the generic hardware event codes to have 0 and -1 mean the same thing as for the generic cache events. Furthermore, it turns out the 0 event does not count (that is, its reported that on PPro it actually does count something), therefore use a event configuration that's specified not to count to disable the counters. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_counter.c | 28 +++++++++++++++++++++++----- 1 file changed, 23 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c index 1910f39ff19a..c7cc6eac14ec 100644 --- a/arch/x86/kernel/cpu/perf_counter.c +++ b/arch/x86/kernel/cpu/perf_counter.c @@ -84,6 +84,14 @@ static u64 p6_pmu_event_map(int event) return p6_perfmon_event_map[event]; } +/* + * Counter setting that is specified not to count anything. + * We use this to effectively disable a counter. + * + * L2_RQSTS with 0 MESI unit mask. + */ +#define P6_NOP_COUNTER 0x0000002EULL + static u64 p6_pmu_raw_event(u64 event) { #define P6_EVNTSEL_EVENT_MASK 0x000000FFULL @@ -704,6 +712,7 @@ static int __hw_perf_counter_init(struct perf_counter *counter) { struct perf_counter_attr *attr = &counter->attr; struct hw_perf_counter *hwc = &counter->hw; + u64 config; int err; if (!x86_pmu_initialized()) @@ -756,10 +765,19 @@ static int __hw_perf_counter_init(struct perf_counter *counter) if (attr->config >= x86_pmu.max_events) return -EINVAL; + /* * The generic map: */ - hwc->config |= x86_pmu.event_map(attr->config); + config = x86_pmu.event_map(attr->config); + + if (config == 0) + return -ENOENT; + + if (config == -1LL) + return -EINVAL; + + hwc->config |= config; return 0; } @@ -767,7 +785,7 @@ static int __hw_perf_counter_init(struct perf_counter *counter) static void p6_pmu_disable_all(void) { struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); - unsigned long val; + u64 val; if (!cpuc->enabled) return; @@ -917,10 +935,10 @@ static inline void p6_pmu_disable_counter(struct hw_perf_counter *hwc, int idx) { struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); - unsigned long val = ARCH_PERFMON_EVENTSEL0_ENABLE; + u64 val = P6_NOP_COUNTER; - if (!cpuc->enabled) - val = 0; + if (cpuc->enabled) + val |= ARCH_PERFMON_EVENTSEL0_ENABLE; (void)checking_wrmsrl(hwc->config_base + idx, val); } -- cgit v1.2.2 From 984b838ce69c063a91b87550598ab7f3439dd94a Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 10 Jul 2009 09:59:56 +0200 Subject: perf_counter: Clean up global vs counter enable Ingo noticed that both AMD and P6 call x86_pmu_disable_counter() on *_pmu_enable_counter(). This is because we rely on the side effect of that call to program the event config but not touch the EN bit. We change that for AMD by having enable_all() simply write the full config in, and for P6 by explicitly coding it. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_counter.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c index c7cc6eac14ec..bed1c4c2f251 100644 --- a/arch/x86/kernel/cpu/perf_counter.c +++ b/arch/x86/kernel/cpu/perf_counter.c @@ -874,13 +874,13 @@ static void amd_pmu_enable_all(void) barrier(); for (idx = 0; idx < x86_pmu.num_counters; idx++) { + struct perf_counter *counter = cpuc->counters[idx]; u64 val; if (!test_bit(idx, cpuc->active_mask)) continue; - rdmsrl(MSR_K7_EVNTSEL0 + idx, val); - if (val & ARCH_PERFMON_EVENTSEL0_ENABLE) - continue; + + val = counter->hw.config; val |= ARCH_PERFMON_EVENTSEL0_ENABLE; wrmsrl(MSR_K7_EVNTSEL0 + idx, val); } @@ -1044,11 +1044,13 @@ intel_pmu_enable_fixed(struct hw_perf_counter *hwc, int __idx) static void p6_pmu_enable_counter(struct hw_perf_counter *hwc, int idx) { struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + u64 val; + val = hwc->config; if (cpuc->enabled) - x86_pmu_enable_counter(hwc, idx); - else - x86_pmu_disable_counter(hwc, idx); + val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + + (void)checking_wrmsrl(hwc->config_base + idx, val); } @@ -1068,8 +1070,6 @@ static void amd_pmu_enable_counter(struct hw_perf_counter *hwc, int idx) if (cpuc->enabled) x86_pmu_enable_counter(hwc, idx); - else - x86_pmu_disable_counter(hwc, idx); } static int -- cgit v1.2.2 From f1c6a58121f9846ac665b0fbd3cbab90ce8bcbac Mon Sep 17 00:00:00 2001 From: Daniel Qarras Date: Sun, 12 Jul 2009 04:32:40 -0700 Subject: perf_counter, x86: Extend perf_counter Pentium M support I've attached a patch to remove the Pentium M special casing of EMON and as noticed at least with my Pentium M the hardware PMU now works: Performance counter stats for '/bin/ls /var/tmp': 1.809988 task-clock-msecs # 0.125 CPUs 1 context-switches # 0.001 M/sec 0 CPU-migrations # 0.000 M/sec 224 page-faults # 0.124 M/sec 1425648 cycles # 787.656 M/sec 912755 instructions # 0.640 IPC Vince suggested that this code was trying to address erratum Y17 in Pentium-M's: http://download.intel.com/support/processors/mobile/pm/sb/25266532.pdf But that erratum (related to IA32_MISC_ENABLES.7) does not affect perfcounters as we dont use this toggle to disable RDPMC and WRMSR/RDMSR access to performance counters. We keep cr4's bit 8 (X86_CR4_PCE) clear so unprivileged RDPMC access is not allowed anyway. Cc: Vince Weaver Cc: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_counter.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c index bed1c4c2f251..7e346d4bc0fb 100644 --- a/arch/x86/kernel/cpu/perf_counter.c +++ b/arch/x86/kernel/cpu/perf_counter.c @@ -1583,10 +1583,8 @@ static int p6_pmu_init(void) break; case 9: case 13: - /* for Pentium M, we need to check if PMU exist */ - rdmsr(MSR_IA32_MISC_ENABLE, low, high); - if (low & MSR_IA32_MISC_ENABLE_EMON) - break; + /* Pentium M */ + break; default: pr_cont("unsupported p6 CPU model %d ", boot_cpu_data.x86_model); -- cgit v1.2.2 From 6aa542a694dc9ea4344a8a590d2628c33d1b9431 Mon Sep 17 00:00:00 2001 From: Alexey Fisher Date: Wed, 15 Jul 2009 14:16:09 +0200 Subject: x86: Add quirk for Intel DG45ID board to avoid low memory corruption AMI BIOS with low memory corruption was found on Intel DG45ID board (Bug 13710). Add this board to the blacklist - in the (somewhat optimistic) hope of future boards/BIOSes from Intel not having this bug. Also see: http://bugzilla.kernel.org/show_bug.cgi?id=13736 Signed-off-by: Alexey Fisher Cc: ykzhao Cc: alan@lxorguk.ukuu.org.uk Cc: LKML-Reference: <1247660169-4503-1-git-send-email-bug-track@fisher-privat.net> Signed-off-by: Ingo Molnar --- arch/x86/kernel/setup.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index de2cab132844..63f32d220ef2 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -672,6 +672,19 @@ static struct dmi_system_id __initdata bad_bios_dmi_table[] = { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies"), }, }, + { + /* + * AMI BIOS with low memory corruption was found on Intel DG45ID board. + * It hase different DMI_BIOS_VENDOR = "Intel Corp.", for now we will + * match only DMI_BOARD_NAME and see if there is more bad products + * with this vendor. + */ + .callback = dmi_low_memory_corruption, + .ident = "AMI BIOS", + .matches = { + DMI_MATCH(DMI_BOARD_NAME, "DG45ID"), + }, + }, #endif {} }; -- cgit v1.2.2 From 8bcdbe427924a1e4b4e4cf68020e92e9f93fe011 Mon Sep 17 00:00:00 2001 From: Catalin Marinas Date: Tue, 14 Jul 2009 10:52:55 +0100 Subject: x86: Include all of .data.* sections in _edata on 64-bit The .data.read_mostly and .data.cacheline_aligned sections aren't covered by the _sdata .. _edata range on x86-64. This affects kmemleak reporting leading to possible false positives by not scanning the whole data section. Signed-off-by: Catalin Marinas Tested-by: Alexey Fisher Acked-by: Sam Ravnborg Cc: Pekka Enberg LKML-Reference: <1247565175.28240.37.camel@pc1117.cambridge.arm.com> Signed-off-by: Ingo Molnar Cc: Sam Ravnborg --- arch/x86/kernel/vmlinux.lds.S | 7 ------- 1 file changed, 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S index 367e87882041..59f31d2dd435 100644 --- a/arch/x86/kernel/vmlinux.lds.S +++ b/arch/x86/kernel/vmlinux.lds.S @@ -112,11 +112,6 @@ SECTIONS _sdata = .; DATA_DATA CONSTRUCTORS - -#ifdef CONFIG_X86_64 - /* End of data section */ - _edata = .; -#endif } :data #ifdef CONFIG_X86_32 @@ -156,10 +151,8 @@ SECTIONS .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) { *(.data.read_mostly) -#ifdef CONFIG_X86_32 /* End of data section */ _edata = .; -#endif } #ifdef CONFIG_X86_64 -- cgit v1.2.2 From 77f32dfdd97857280ae1ebac64382ff082cd7ea0 Mon Sep 17 00:00:00 2001 From: Denis Turischev Date: Mon, 20 Jul 2009 18:48:17 +0300 Subject: x86: Add reboot fixup for SBC-fitPC2 The CompuLab SBC-fitPC2 board needs to reboot via BIOS. Signed-off-by: Denis Turischev Signed-off-by: Mike Rapoport Signed-off-by: H. Peter Anvin --- arch/x86/kernel/reboot.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index d2d1ce8170f0..508e982dd072 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -249,6 +249,14 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = { DMI_MATCH(DMI_PRODUCT_NAME, "VGN-Z540N"), }, }, + { /* Handle problems with rebooting on CompuLab SBC-FITPC2 */ + .callback = set_bios_reboot, + .ident = "CompuLab SBC-FITPC2", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "CompuLab"), + DMI_MATCH(DMI_PRODUCT_NAME, "SBC-FITPC2"), + }, + }, { } }; -- cgit v1.2.2 From 155b73529583c38f30fd394d692b15a893960782 Mon Sep 17 00:00:00 2001 From: Uros Bizjak Date: Sun, 19 Jul 2009 18:06:35 +0200 Subject: x86: Fix movq immediate operand constraints in uaccess_64.h arch/x86/include/asm/uaccess_64.h uses wrong asm operand constraint ("ir") for movq insn. Since movq sign-extends its immediate operand, "er" constraint should be used instead. Attached patch changes all uses of __put_user_asm in uaccess_64.h to use "er" when "q" insn suffix is involved. Patch was compile tested on x86_64 with defconfig. Signed-off-by: Uros Bizjak Signed-off-by: H. Peter Anvin Cc: stable@kernel.org --- arch/x86/include/asm/uaccess_64.h | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uaccess_64.h b/arch/x86/include/asm/uaccess_64.h index 8cc687326eb8..db24b215fc50 100644 --- a/arch/x86/include/asm/uaccess_64.h +++ b/arch/x86/include/asm/uaccess_64.h @@ -88,11 +88,11 @@ int __copy_to_user(void __user *dst, const void *src, unsigned size) ret, "l", "k", "ir", 4); return ret; case 8:__put_user_asm(*(u64 *)src, (u64 __user *)dst, - ret, "q", "", "ir", 8); + ret, "q", "", "er", 8); return ret; case 10: __put_user_asm(*(u64 *)src, (u64 __user *)dst, - ret, "q", "", "ir", 10); + ret, "q", "", "er", 10); if (unlikely(ret)) return ret; asm("":::"memory"); @@ -101,12 +101,12 @@ int __copy_to_user(void __user *dst, const void *src, unsigned size) return ret; case 16: __put_user_asm(*(u64 *)src, (u64 __user *)dst, - ret, "q", "", "ir", 16); + ret, "q", "", "er", 16); if (unlikely(ret)) return ret; asm("":::"memory"); __put_user_asm(1[(u64 *)src], 1 + (u64 __user *)dst, - ret, "q", "", "ir", 8); + ret, "q", "", "er", 8); return ret; default: return copy_user_generic((__force void *)dst, src, size); @@ -157,7 +157,7 @@ int __copy_in_user(void __user *dst, const void __user *src, unsigned size) ret, "q", "", "=r", 8); if (likely(!ret)) __put_user_asm(tmp, (u64 __user *)dst, - ret, "q", "", "ir", 8); + ret, "q", "", "er", 8); return ret; } default: -- cgit v1.2.2 From ebe119cd0929df4878f758ebf880cb435e4dcaaf Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Mon, 20 Jul 2009 23:27:39 -0700 Subject: x86: Fix movq immediate operand constraints in uaccess.h The movq instruction, generated by __put_user_asm() when used for 64-bit data, takes a sign-extended immediate ("e") not a zero-extended immediate ("Z"). Signed-off-by: H. Peter Anvin Cc: Uros Bizjak Cc: stable@kernel.org --- arch/x86/include/asm/uaccess.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uaccess.h b/arch/x86/include/asm/uaccess.h index 20e6a795e160..d2c6c930b491 100644 --- a/arch/x86/include/asm/uaccess.h +++ b/arch/x86/include/asm/uaccess.h @@ -212,9 +212,9 @@ extern int __get_user_bad(void); : "A" ((typeof(*(ptr)))(x)), "c" (ptr) : "ebx") #else #define __put_user_asm_u64(x, ptr, retval, errret) \ - __put_user_asm(x, ptr, retval, "q", "", "Zr", errret) + __put_user_asm(x, ptr, retval, "q", "", "er", errret) #define __put_user_asm_ex_u64(x, addr) \ - __put_user_asm_ex(x, addr, "q", "", "Zr") + __put_user_asm_ex(x, addr, "q", "", "er") #define __put_user_x8(x, ptr, __ret_pu) __put_user_x(8, x, ptr, __ret_pu) #endif -- cgit v1.2.2 From e9084ec98bb9aa3abc6cf73181177780ce7546f8 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 16 Jul 2009 09:45:11 +0100 Subject: x86, mce: Fix set_trigger() accessor Fix the condition checking the result of strchr() (which previously could result in an oops), and make the function return the number of bytes actively used. [ Impact: fix oops ] Signed-off-by: Jan Beulich Cc: Andi Kleen LKML-Reference: <4A5F04B7020000780000AB59@vpn.id2.novell.com> Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/mcheck/mce.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index 484c1e5f658e..1cfb623ce11c 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -1692,17 +1692,15 @@ static ssize_t set_trigger(struct sys_device *s, struct sysdev_attribute *attr, const char *buf, size_t siz) { char *p; - int len; strncpy(mce_helper, buf, sizeof(mce_helper)); mce_helper[sizeof(mce_helper)-1] = 0; - len = strlen(mce_helper); p = strchr(mce_helper, '\n'); - if (*p) + if (p) *p = 0; - return len; + return strlen(mce_helper) + !!p; } static ssize_t set_ignore_ce(struct sys_device *s, -- cgit v1.2.2 From 429b2b319af3987e808c18f6b81313104caf782c Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sat, 18 Jul 2009 08:56:57 +0200 Subject: x86-64: Fix bad_srat() to clear all state Need to clear both nodes and nodes_add state for start/end. Signed-off-by: Andi Kleen LKML-Reference: <20090718065657.GA2898@basil.fritz.box> Signed-off-by: H. Peter Anvin Cc: stable@kernel.org --- arch/x86/mm/srat_64.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/mm/srat_64.c b/arch/x86/mm/srat_64.c index 2dfcbf9df2ae..dbb5381f7b3b 100644 --- a/arch/x86/mm/srat_64.c +++ b/arch/x86/mm/srat_64.c @@ -79,8 +79,10 @@ static __init void bad_srat(void) acpi_numa = -1; for (i = 0; i < MAX_LOCAL_APIC; i++) apicid_to_node[i] = NUMA_NO_NODE; - for (i = 0; i < MAX_NUMNODES; i++) - nodes_add[i].start = nodes[i].end = 0; + for (i = 0; i < MAX_NUMNODES; i++) { + nodes[i].start = nodes[i].end = 0; + nodes_add[i].start = nodes_add[i].end = 0; + } remove_all_active_ranges(); } -- cgit v1.2.2 From 6effa8f6fc786f00e3a23eae605e0f2e8e748faa Mon Sep 17 00:00:00 2001 From: Hidehiro Kawai Date: Wed, 22 Jul 2009 11:56:20 +0900 Subject: x86, mce: Rename incorrect macro name "CONFIG_X86_THRESHOLD" CONFIG_X86_THRESHOLD used in arch/x86/kernel/irqinit.c is always undefined. Rename it to the correct name "CONFIG_X86_MCE_THRESHOLD". Signed-off-by: Hidehiro Kawai Reviewed-by: Hidetoshi Seto Cc: Andi Kleen LKML-Reference: <4A667FD4.3010509@hitachi.com> Signed-off-by: H. Peter Anvin --- arch/x86/kernel/irqinit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/irqinit.c b/arch/x86/kernel/irqinit.c index 696f0e475c2d..92b7703d3d58 100644 --- a/arch/x86/kernel/irqinit.c +++ b/arch/x86/kernel/irqinit.c @@ -187,7 +187,7 @@ static void __init apic_intr_init(void) #ifdef CONFIG_X86_THERMAL_VECTOR alloc_intr_gate(THERMAL_APIC_VECTOR, thermal_interrupt); #endif -#ifdef CONFIG_X86_THRESHOLD +#ifdef CONFIG_X86_MCE_THRESHOLD alloc_intr_gate(THRESHOLD_APIC_VECTOR, threshold_interrupt); #endif #if defined(CONFIG_X86_NEW_MCE) && defined(CONFIG_X86_LOCAL_APIC) -- cgit v1.2.2 From 9b7019ae6a542a801a80df6694c66e240e2c3e39 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 22 Jul 2009 16:31:36 +0200 Subject: perf_counter: Remove unused variables Fix a gcc unused variables warning. Signed-off-by: Peter Zijlstra --- arch/x86/kernel/cpu/perf_counter.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c index 7e346d4bc0fb..a7aa8f900954 100644 --- a/arch/x86/kernel/cpu/perf_counter.c +++ b/arch/x86/kernel/cpu/perf_counter.c @@ -1570,8 +1570,6 @@ static struct x86_pmu amd_pmu = { static int p6_pmu_init(void) { - int high, low; - switch (boot_cpu_data.x86_model) { case 1: case 3: /* Pentium Pro */ -- cgit v1.2.2 From 2cb078603abb612e3bcd428fb8122c3d39e08832 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Wed, 22 Jul 2009 09:59:35 -0700 Subject: x86, amd: Don't probe for extended APIC ID if APICs are disabled If we've logically disabled apics, don't probe the PCI space for the AMD extended APIC ID. [ Impact: prevent boot crash under Xen. ] Signed-off-by: Jeremy Fitzhardinge Reported-by: Bastian Blank Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/amd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 28e5f5956042..e2485b03f1cf 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -356,7 +356,7 @@ static void __cpuinit early_init_amd(struct cpuinfo_x86 *c) #endif #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_PCI) /* check CPU config space for extended APIC ID */ - if (c->x86 >= 0xf) { + if (cpu_has_apic && c->x86 >= 0xf) { unsigned int val; val = read_pci_config(0, 24, 0, 0x68); if ((val & ((1 << 17) | (1 << 18))) == ((1 << 17) | (1 << 18))) -- cgit v1.2.2 From d6c585a4342a2ff627a29f9aea77c5ed4cd76023 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 24 Jul 2009 08:34:59 +0200 Subject: x86: geode: Mark mfgpt irq IRQF_TIMER to prevent resume failure Timer interrupts are excluded from being disabled during suspend. The clock events code manages the disabling of clock events on its own because the timer interrupt needs to be functional before the resume code reenables the device interrupts. The mfgpt timer request its interrupt without setting the IRQF_TIMER flag so suspend_device_irqs() disables it as well which results in a fatal resume failure. Adding IRQF_TIMER to the interupt flags when requesting the mrgpt timer interrupt solves the problem. Signed-off-by: Thomas Gleixner LKML-Reference: Cc: Andres Salomon Cc: stable@kernel.org --- arch/x86/kernel/mfgpt_32.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/mfgpt_32.c b/arch/x86/kernel/mfgpt_32.c index 846510b78a09..2a62d843f015 100644 --- a/arch/x86/kernel/mfgpt_32.c +++ b/arch/x86/kernel/mfgpt_32.c @@ -347,7 +347,7 @@ static irqreturn_t mfgpt_tick(int irq, void *dev_id) static struct irqaction mfgptirq = { .handler = mfgpt_tick, - .flags = IRQF_DISABLED | IRQF_NOBALANCING, + .flags = IRQF_DISABLED | IRQF_NOBALANCING | IRQF_TIMER, .name = "mfgpt-timer" }; -- cgit v1.2.2 From 9e1b32caa525cb236e80e9c671e179bcecccc657 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Wed, 22 Jul 2009 15:44:28 +1000 Subject: mm: Pass virtual address to [__]p{te,ud,md}_free_tlb() mm: Pass virtual address to [__]p{te,ud,md}_free_tlb() Upcoming paches to support the new 64-bit "BookE" powerpc architecture will need to have the virtual address corresponding to PTE page when freeing it, due to the way the HW table walker works. Basically, the TLB can be loaded with "large" pages that cover the whole virtual space (well, sort-of, half of it actually) represented by a PTE page, and which contain an "indirect" bit indicating that this TLB entry RPN points to an array of PTEs from which the TLB can then create direct entries. Thus, in order to invalidate those when PTE pages are deleted, we need the virtual address to pass to tlbilx or tlbivax instructions. The old trick of sticking it somewhere in the PTE page struct page sucks too much, the address is almost readily available in all call sites and almost everybody implemets these as macros, so we may as well add the argument everywhere. I added it to the pmd and pud variants for consistency. Signed-off-by: Benjamin Herrenschmidt Acked-by: David Howells [MN10300 & FRV] Acked-by: Nick Piggin Acked-by: Martin Schwidefsky [s390] Signed-off-by: Linus Torvalds --- arch/x86/include/asm/pgalloc.h | 25 ++++++++++++++++++++++--- arch/x86/mm/pgtable.c | 6 +++--- 2 files changed, 25 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pgalloc.h b/arch/x86/include/asm/pgalloc.h index dd14c54ac718..0e8c2a0fd922 100644 --- a/arch/x86/include/asm/pgalloc.h +++ b/arch/x86/include/asm/pgalloc.h @@ -46,7 +46,13 @@ static inline void pte_free(struct mm_struct *mm, struct page *pte) __free_page(pte); } -extern void __pte_free_tlb(struct mmu_gather *tlb, struct page *pte); +extern void ___pte_free_tlb(struct mmu_gather *tlb, struct page *pte); + +static inline void __pte_free_tlb(struct mmu_gather *tlb, struct page *pte, + unsigned long address) +{ + ___pte_free_tlb(tlb, pte); +} static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd, pte_t *pte) @@ -78,7 +84,13 @@ static inline void pmd_free(struct mm_struct *mm, pmd_t *pmd) free_page((unsigned long)pmd); } -extern void __pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd); +extern void ___pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd); + +static inline void __pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd, + unsigned long adddress) +{ + ___pmd_free_tlb(tlb, pmd); +} #ifdef CONFIG_X86_PAE extern void pud_populate(struct mm_struct *mm, pud_t *pudp, pmd_t *pmd); @@ -108,7 +120,14 @@ static inline void pud_free(struct mm_struct *mm, pud_t *pud) free_page((unsigned long)pud); } -extern void __pud_free_tlb(struct mmu_gather *tlb, pud_t *pud); +extern void ___pud_free_tlb(struct mmu_gather *tlb, pud_t *pud); + +static inline void __pud_free_tlb(struct mmu_gather *tlb, pud_t *pud, + unsigned long address) +{ + ___pud_free_tlb(tlb, pud); +} + #endif /* PAGETABLE_LEVELS > 3 */ #endif /* PAGETABLE_LEVELS > 2 */ diff --git a/arch/x86/mm/pgtable.c b/arch/x86/mm/pgtable.c index 8e43bdd45456..af8f9650058c 100644 --- a/arch/x86/mm/pgtable.c +++ b/arch/x86/mm/pgtable.c @@ -25,7 +25,7 @@ pgtable_t pte_alloc_one(struct mm_struct *mm, unsigned long address) return pte; } -void __pte_free_tlb(struct mmu_gather *tlb, struct page *pte) +void ___pte_free_tlb(struct mmu_gather *tlb, struct page *pte) { pgtable_page_dtor(pte); paravirt_release_pte(page_to_pfn(pte)); @@ -33,14 +33,14 @@ void __pte_free_tlb(struct mmu_gather *tlb, struct page *pte) } #if PAGETABLE_LEVELS > 2 -void __pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd) +void ___pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd) { paravirt_release_pmd(__pa(pmd) >> PAGE_SHIFT); tlb_remove_page(tlb, virt_to_page(pmd)); } #if PAGETABLE_LEVELS > 3 -void __pud_free_tlb(struct mmu_gather *tlb, pud_t *pud) +void ___pud_free_tlb(struct mmu_gather *tlb, pud_t *pud) { paravirt_release_pud(__pa(pud) >> PAGE_SHIFT); tlb_remove_page(tlb, virt_to_page(pud)); -- cgit v1.2.2