aboutsummaryrefslogtreecommitdiff
path: root/arch/x86/kernel/cpu
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86/kernel/cpu')
-rw-r--r--arch/x86/kernel/cpu/Makefile2
-rw-r--r--arch/x86/kernel/cpu/amd.c52
-rw-r--r--arch/x86/kernel/cpu/common.c6
-rwxr-xr-xarch/x86/kernel/cpu/cpu_debug.c785
-rw-r--r--arch/x86/kernel/cpu/intel.c25
-rw-r--r--arch/x86/kernel/cpu/mcheck/Makefile1
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_32.c14
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_64.c530
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_amd_64.c22
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_intel_64.c207
-rw-r--r--arch/x86/kernel/cpu/mcheck/threshold.c29
11 files changed, 1506 insertions, 167 deletions
diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile
index 82db7f45e2d..d4356f8b752 100644
--- a/arch/x86/kernel/cpu/Makefile
+++ b/arch/x86/kernel/cpu/Makefile
@@ -14,6 +14,8 @@ obj-y += vmware.o hypervisor.o
obj-$(CONFIG_X86_32) += bugs.o cmpxchg.o
obj-$(CONFIG_X86_64) += bugs_64.o
+obj-$(CONFIG_X86_CPU_DEBUG) += cpu_debug.o
+
obj-$(CONFIG_CPU_SUP_INTEL) += intel.o
obj-$(CONFIG_CPU_SUP_AMD) += amd.o
obj-$(CONFIG_CPU_SUP_CYRIX_32) += cyrix.o
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index 25423a5b80e..f47df59016c 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -5,6 +5,7 @@
#include <asm/io.h>
#include <asm/processor.h>
#include <asm/apic.h>
+#include <asm/cpu.h>
#ifdef CONFIG_X86_64
# include <asm/numa_64.h>
@@ -141,6 +142,55 @@ static void __cpuinit init_amd_k6(struct cpuinfo_x86 *c)
}
}
+static void __cpuinit amd_k7_smp_check(struct cpuinfo_x86 *c)
+{
+#ifdef CONFIG_SMP
+ /* calling is from identify_secondary_cpu() ? */
+ if (c->cpu_index == boot_cpu_id)
+ return;
+
+ /*
+ * Certain Athlons might work (for various values of 'work') in SMP
+ * but they are not certified as MP capable.
+ */
+ /* Athlon 660/661 is valid. */
+ if ((c->x86_model == 6) && ((c->x86_mask == 0) ||
+ (c->x86_mask == 1)))
+ goto valid_k7;
+
+ /* Duron 670 is valid */
+ if ((c->x86_model == 7) && (c->x86_mask == 0))
+ goto valid_k7;
+
+ /*
+ * Athlon 662, Duron 671, and Athlon >model 7 have capability
+ * bit. It's worth noting that the A5 stepping (662) of some
+ * Athlon XP's have the MP bit set.
+ * See http://www.heise.de/newsticker/data/jow-18.10.01-000 for
+ * more.
+ */
+ if (((c->x86_model == 6) && (c->x86_mask >= 2)) ||
+ ((c->x86_model == 7) && (c->x86_mask >= 1)) ||
+ (c->x86_model > 7))
+ if (cpu_has_mp)
+ goto valid_k7;
+
+ /* If we get here, not a certified SMP capable AMD system. */
+
+ /*
+ * Don't taint if we are running SMP kernel on a single non-MP
+ * approved Athlon
+ */
+ WARN_ONCE(1, "WARNING: This combination of AMD"
+ "processors is not suitable for SMP.\n");
+ if (!test_taint(TAINT_UNSAFE_SMP))
+ add_taint(TAINT_UNSAFE_SMP);
+
+valid_k7:
+ ;
+#endif
+}
+
static void __cpuinit init_amd_k7(struct cpuinfo_x86 *c)
{
u32 l, h;
@@ -175,6 +225,8 @@ static void __cpuinit init_amd_k7(struct cpuinfo_x86 *c)
}
set_cpu_cap(c, X86_FEATURE_K7);
+
+ amd_k7_smp_check(c);
}
#endif
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 826d5c87627..f8869978bbb 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -1078,8 +1078,7 @@ void __cpuinit cpu_init(void)
atomic_inc(&init_mm.mm_count);
me->active_mm = &init_mm;
- if (me->mm)
- BUG();
+ BUG_ON(me->mm);
enter_lazy_tlb(&init_mm, me);
load_sp0(t, &current->thread);
@@ -1145,8 +1144,7 @@ void __cpuinit cpu_init(void)
*/
atomic_inc(&init_mm.mm_count);
curr->active_mm = &init_mm;
- if (curr->mm)
- BUG();
+ BUG_ON(curr->mm);
enter_lazy_tlb(&init_mm, curr);
load_sp0(t, thread);
diff --git a/arch/x86/kernel/cpu/cpu_debug.c b/arch/x86/kernel/cpu/cpu_debug.c
new file mode 100755
index 00000000000..9abbcbd933c
--- /dev/null
+++ b/arch/x86/kernel/cpu/cpu_debug.c
@@ -0,0 +1,785 @@
+/*
+ * CPU x86 architecture debug code
+ *
+ * Copyright(C) 2009 Jaswinder Singh Rajput
+ *
+ * For licencing details see kernel-base/COPYING
+ */
+
+#include <linux/interrupt.h>
+#include <linux/compiler.h>
+#include <linux/seq_file.h>
+#include <linux/debugfs.h>
+#include <linux/kprobes.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/percpu.h>
+#include <linux/signal.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/types.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/smp.h>
+
+#include <asm/cpu_debug.h>
+#include <asm/paravirt.h>
+#include <asm/system.h>
+#include <asm/traps.h>
+#include <asm/apic.h>
+#include <asm/desc.h>
+
+static DEFINE_PER_CPU(struct cpu_cpuX_base, cpu_arr[CPU_REG_ALL_BIT]);
+static DEFINE_PER_CPU(struct cpu_private *, priv_arr[MAX_CPU_FILES]);
+static DEFINE_PER_CPU(unsigned, cpu_modelflag);
+static DEFINE_PER_CPU(int, cpu_priv_count);
+static DEFINE_PER_CPU(unsigned, cpu_model);
+
+static DEFINE_MUTEX(cpu_debug_lock);
+
+static struct dentry *cpu_debugfs_dir;
+
+static struct cpu_debug_base cpu_base[] = {
+ { "mc", CPU_MC }, /* Machine Check */
+ { "monitor", CPU_MONITOR }, /* Monitor */
+ { "time", CPU_TIME }, /* Time */
+ { "pmc", CPU_PMC }, /* Performance Monitor */
+ { "platform", CPU_PLATFORM }, /* Platform */
+ { "apic", CPU_APIC }, /* APIC */
+ { "poweron", CPU_POWERON }, /* Power-on */
+ { "control", CPU_CONTROL }, /* Control */
+ { "features", CPU_FEATURES }, /* Features control */
+ { "lastbranch", CPU_LBRANCH }, /* Last Branch */
+ { "bios", CPU_BIOS }, /* BIOS */
+ { "freq", CPU_FREQ }, /* Frequency */
+ { "mtrr", CPU_MTRR }, /* MTRR */
+ { "perf", CPU_PERF }, /* Performance */
+ { "cache", CPU_CACHE }, /* Cache */
+ { "sysenter", CPU_SYSENTER }, /* Sysenter */
+ { "therm", CPU_THERM }, /* Thermal */
+ { "misc", CPU_MISC }, /* Miscellaneous */
+ { "debug", CPU_DEBUG }, /* Debug */
+ { "pat", CPU_PAT }, /* PAT */
+ { "vmx", CPU_VMX }, /* VMX */
+ { "call", CPU_CALL }, /* System Call */
+ { "base", CPU_BASE }, /* BASE Address */
+ { "smm", CPU_SMM }, /* System mgmt mode */
+ { "svm", CPU_SVM }, /*Secure Virtial Machine*/
+ { "osvm", CPU_OSVM }, /* OS-Visible Workaround*/
+ { "tss", CPU_TSS }, /* Task Stack Segment */
+ { "cr", CPU_CR }, /* Control Registers */
+ { "dt", CPU_DT }, /* Descriptor Table */
+ { "registers", CPU_REG_ALL }, /* Select all Registers */
+};
+
+static struct cpu_file_base cpu_file[] = {
+ { "index", CPU_REG_ALL }, /* index */
+ { "value", CPU_REG_ALL }, /* value */
+};
+
+/* Intel Registers Range */
+static struct cpu_debug_range cpu_intel_range[] = {
+ { 0x00000000, 0x00000001, CPU_MC, CPU_INTEL_ALL },
+ { 0x00000006, 0x00000007, CPU_MONITOR, CPU_CX_AT_XE },
+ { 0x00000010, 0x00000010, CPU_TIME, CPU_INTEL_ALL },
+ { 0x00000011, 0x00000013, CPU_PMC, CPU_INTEL_PENTIUM },
+ { 0x00000017, 0x00000017, CPU_PLATFORM, CPU_PX_CX_AT_XE },
+ { 0x0000001B, 0x0000001B, CPU_APIC, CPU_P6_CX_AT_XE },
+
+ { 0x0000002A, 0x0000002A, CPU_POWERON, CPU_PX_CX_AT_XE },
+ { 0x0000002B, 0x0000002B, CPU_POWERON, CPU_INTEL_XEON },
+ { 0x0000002C, 0x0000002C, CPU_FREQ, CPU_INTEL_XEON },
+ { 0x0000003A, 0x0000003A, CPU_CONTROL, CPU_CX_AT_XE },
+
+ { 0x00000040, 0x00000043, CPU_LBRANCH, CPU_PM_CX_AT_XE },
+ { 0x00000044, 0x00000047, CPU_LBRANCH, CPU_PM_CO_AT },
+ { 0x00000060, 0x00000063, CPU_LBRANCH, CPU_C2_AT },
+ { 0x00000064, 0x00000067, CPU_LBRANCH, CPU_INTEL_ATOM },
+
+ { 0x00000079, 0x00000079, CPU_BIOS, CPU_P6_CX_AT_XE },
+ { 0x00000088, 0x0000008A, CPU_CACHE, CPU_INTEL_P6 },
+ { 0x0000008B, 0x0000008B, CPU_BIOS, CPU_P6_CX_AT_XE },
+ { 0x0000009B, 0x0000009B, CPU_MONITOR, CPU_INTEL_XEON },
+
+ { 0x000000C1, 0x000000C2, CPU_PMC, CPU_P6_CX_AT },
+ { 0x000000CD, 0x000000CD, CPU_FREQ, CPU_CX_AT },
+ { 0x000000E7, 0x000000E8, CPU_PERF, CPU_CX_AT },
+ { 0x000000FE, 0x000000FE, CPU_MTRR, CPU_P6_CX_XE },
+
+ { 0x00000116, 0x00000116, CPU_CACHE, CPU_INTEL_P6 },
+ { 0x00000118, 0x00000118, CPU_CACHE, CPU_INTEL_P6 },
+ { 0x00000119, 0x00000119, CPU_CACHE, CPU_INTEL_PX },
+ { 0x0000011A, 0x0000011B, CPU_CACHE, CPU_INTEL_P6 },
+ { 0x0000011E, 0x0000011E, CPU_CACHE, CPU_PX_CX_AT },
+
+ { 0x00000174, 0x00000176, CPU_SYSENTER, CPU_P6_CX_AT_XE },
+ { 0x00000179, 0x0000017A, CPU_MC, CPU_PX_CX_AT_XE },
+ { 0x0000017B, 0x0000017B, CPU_MC, CPU_P6_XE },
+ { 0x00000186, 0x00000187, CPU_PMC, CPU_P6_CX_AT },
+ { 0x00000198, 0x00000199, CPU_PERF, CPU_PM_CX_AT_XE },
+ { 0x0000019A, 0x0000019A, CPU_TIME, CPU_PM_CX_AT_XE },
+ { 0x0000019B, 0x0000019D, CPU_THERM, CPU_PM_CX_AT_XE },
+ { 0x000001A0, 0x000001A0, CPU_MISC, CPU_PM_CX_AT_XE },
+
+ { 0x000001C9, 0x000001C9, CPU_LBRANCH, CPU_PM_CX_AT },
+ { 0x000001D7, 0x000001D8, CPU_LBRANCH, CPU_INTEL_XEON },
+ { 0x000001D9, 0x000001D9, CPU_DEBUG, CPU_CX_AT_XE },
+ { 0x000001DA, 0x000001DA, CPU_LBRANCH, CPU_INTEL_XEON },
+ { 0x000001DB, 0x000001DB, CPU_LBRANCH, CPU_P6_XE },
+ { 0x000001DC, 0x000001DC, CPU_LBRANCH, CPU_INTEL_P6 },
+ { 0x000001DD, 0x000001DE, CPU_LBRANCH, CPU_PX_CX_AT_XE },
+ { 0x000001E0, 0x000001E0, CPU_LBRANCH, CPU_INTEL_P6 },
+
+ { 0x00000200, 0x0000020F, CPU_MTRR, CPU_P6_CX_XE },
+ { 0x00000250, 0x00000250, CPU_MTRR, CPU_P6_CX_XE },
+ { 0x00000258, 0x00000259, CPU_MTRR, CPU_P6_CX_XE },
+ { 0x00000268, 0x0000026F, CPU_MTRR, CPU_P6_CX_XE },
+ { 0x00000277, 0x00000277, CPU_PAT, CPU_C2_AT_XE },
+ { 0x000002FF, 0x000002FF, CPU_MTRR, CPU_P6_CX_XE },
+
+ { 0x00000300, 0x00000308, CPU_PMC, CPU_INTEL_XEON },
+ { 0x00000309, 0x0000030B, CPU_PMC, CPU_C2_AT_XE },
+ { 0x0000030C, 0x00000311, CPU_PMC, CPU_INTEL_XEON },
+ { 0x00000345, 0x00000345, CPU_PMC, CPU_C2_AT },
+ { 0x00000360, 0x00000371, CPU_PMC, CPU_INTEL_XEON },
+ { 0x0000038D, 0x00000390, CPU_PMC, CPU_C2_AT },
+ { 0x000003A0, 0x000003BE, CPU_PMC, CPU_INTEL_XEON },
+ { 0x000003C0, 0x000003CD, CPU_PMC, CPU_INTEL_XEON },
+ { 0x000003E0, 0x000003E1, CPU_PMC, CPU_INTEL_XEON },
+ { 0x000003F0, 0x000003F0, CPU_PMC, CPU_INTEL_XEON },
+ { 0x000003F1, 0x000003F1, CPU_PMC, CPU_C2_AT_XE },
+ { 0x000003F2, 0x000003F2, CPU_PMC, CPU_INTEL_XEON },
+
+ { 0x00000400, 0x00000402, CPU_MC, CPU_PM_CX_AT_XE },
+ { 0x00000403, 0x00000403, CPU_MC, CPU_INTEL_XEON },
+ { 0x00000404, 0x00000406, CPU_MC, CPU_PM_CX_AT_XE },
+ { 0x00000407, 0x00000407, CPU_MC, CPU_INTEL_XEON },
+ { 0x00000408, 0x0000040A, CPU_MC, CPU_PM_CX_AT_XE },
+ { 0x0000040B, 0x0000040B, CPU_MC, CPU_INTEL_XEON },
+ { 0x0000040C, 0x0000040E, CPU_MC, CPU_PM_CX_XE },
+ { 0x0000040F, 0x0000040F, CPU_MC, CPU_INTEL_XEON },
+ { 0x00000410, 0x00000412, CPU_MC, CPU_PM_CX_AT_XE },
+ { 0x00000413, 0x00000417, CPU_MC, CPU_CX_AT_XE },
+ { 0x00000480, 0x0000048B, CPU_VMX, CPU_CX_AT_XE },
+
+ { 0x00000600, 0x00000600, CPU_DEBUG, CPU_PM_CX_AT_XE },
+ { 0x00000680, 0x0000068F, CPU_LBRANCH, CPU_INTEL_XEON },
+ { 0x000006C0, 0x000006CF, CPU_LBRANCH, CPU_INTEL_XEON },
+
+ { 0x000107CC, 0x000107D3, CPU_PMC, CPU_INTEL_XEON_MP },
+
+ { 0xC0000080, 0xC0000080, CPU_FEATURES, CPU_INTEL_XEON },
+ { 0xC0000081, 0xC0000082, CPU_CALL, CPU_INTEL_XEON },
+ { 0xC0000084, 0xC0000084, CPU_CALL, CPU_INTEL_XEON },
+ { 0xC0000100, 0xC0000102, CPU_BASE, CPU_INTEL_XEON },
+};
+
+/* AMD Registers Range */
+static struct cpu_debug_range cpu_amd_range[] = {
+ { 0x00000010, 0x00000010, CPU_TIME, CPU_ALL, },
+ { 0x0000001B, 0x0000001B, CPU_APIC, CPU_ALL, },
+ { 0x000000FE, 0x000000FE, CPU_MTRR, CPU_ALL, },
+
+ { 0x00000174, 0x00000176, CPU_SYSENTER, CPU_ALL, },
+ { 0x00000179, 0x0000017A, CPU_MC, CPU_ALL, },
+ { 0x0000017B, 0x0000017B, CPU_MC, CPU_ALL, },
+ { 0x000001D9, 0x000001D9, CPU_DEBUG, CPU_ALL, },
+ { 0x000001DB, 0x000001DE, CPU_LBRANCH, CPU_ALL, },
+
+ { 0x00000200, 0x0000020F, CPU_MTRR, CPU_ALL, },
+ { 0x00000250, 0x00000250, CPU_MTRR, CPU_ALL, },
+ { 0x00000258, 0x00000259, CPU_MTRR, CPU_ALL, },
+ { 0x00000268, 0x0000026F, CPU_MTRR, CPU_ALL, },
+ { 0x00000277, 0x00000277, CPU_PAT, CPU_ALL, },
+ { 0x000002FF, 0x000002FF, CPU_MTRR, CPU_ALL, },
+
+ { 0x00000400, 0x00000417, CPU_MC, CPU_ALL, },
+
+ { 0xC0000080, 0xC0000080, CPU_FEATURES, CPU_ALL, },
+ { 0xC0000081, 0xC0000084, CPU_CALL, CPU_ALL, },
+ { 0xC0000100, 0xC0000102, CPU_BASE, CPU_ALL, },
+ { 0xC0000103, 0xC0000103, CPU_TIME, CPU_ALL, },
+
+ { 0xC0000408, 0xC000040A, CPU_MC, CPU_ALL, },
+
+ { 0xc0010000, 0xc0010007, CPU_PMC, CPU_ALL, },
+ { 0xc0010010, 0xc0010010, CPU_MTRR, CPU_ALL, },
+ { 0xc0010016, 0xc001001A, CPU_MTRR, CPU_ALL, },
+ { 0xc001001D, 0xc001001D, CPU_MTRR, CPU_ALL, },
+ { 0xc0010030, 0xc0010035, CPU_BIOS, CPU_ALL, },
+ { 0xc0010056, 0xc0010056, CPU_SMM, CPU_ALL, },
+ { 0xc0010061, 0xc0010063, CPU_SMM, CPU_ALL, },
+ { 0xc0010074, 0xc0010074, CPU_MC, CPU_ALL, },
+ { 0xc0010111, 0xc0010113, CPU_SMM, CPU_ALL, },
+ { 0xc0010114, 0xc0010118, CPU_SVM, CPU_ALL, },
+ { 0xc0010119, 0xc001011A, CPU_SMM, CPU_ALL, },
+ { 0xc0010140, 0xc0010141, CPU_OSVM, CPU_ALL, },
+ { 0xc0010156, 0xc0010156, CPU_SMM, CPU_ALL, },
+};
+
+
+static int get_cpu_modelflag(unsigned cpu)
+{
+ int flag;
+
+ switch (per_cpu(cpu_model, cpu)) {
+ /* Intel */
+ case 0x0501:
+ case 0x0502:
+ case 0x0504:
+ flag = CPU_INTEL_PENTIUM;
+ break;
+ case 0x0601:
+ case 0x0603:
+ case 0x0605:
+ case 0x0607:
+ case 0x0608:
+ case 0x060A:
+ case 0x060B:
+ flag = CPU_INTEL_P6;
+ break;
+ case 0x0609:
+ case 0x060D:
+ flag = CPU_INTEL_PENTIUM_M;
+ break;
+ case 0x060E:
+ flag = CPU_INTEL_CORE;
+ break;
+ case 0x060F:
+ case 0x0617:
+ flag = CPU_INTEL_CORE2;
+ break;
+ case 0x061C:
+ flag = CPU_INTEL_ATOM;
+ break;
+ case 0x0F00:
+ case 0x0F01:
+ case 0x0F02:
+ case 0x0F03:
+ case 0x0F04:
+ flag = CPU_INTEL_XEON_P4;
+ break;
+ case 0x0F06:
+ flag = CPU_INTEL_XEON_MP;
+ break;
+ default:
+ flag = CPU_NONE;
+ break;
+ }
+
+ return flag;
+}
+
+static int get_cpu_range_count(unsigned cpu)
+{
+ int index;
+
+ switch (per_cpu(cpu_model, cpu) >> 16) {
+ case X86_VENDOR_INTEL:
+ index = ARRAY_SIZE(cpu_intel_range);
+ break;
+ case X86_VENDOR_AMD:
+ index = ARRAY_SIZE(cpu_amd_range);
+ break;
+ default:
+ index = 0;
+ break;
+ }
+
+ return index;
+}
+
+static int is_typeflag_valid(unsigned cpu, unsigned flag)
+{
+ unsigned vendor, modelflag;
+ int i, index;
+
+ /* Standard Registers should be always valid */
+ if (flag >= CPU_TSS)
+ return 1;
+
+ modelflag = per_cpu(cpu_modelflag, cpu);
+ vendor = per_cpu(cpu_model, cpu) >> 16;
+ index = get_cpu_range_count(cpu);
+
+ for (i = 0; i < index; i++) {
+ switch (vendor) {
+ case X86_VENDOR_INTEL:
+ if ((cpu_intel_range[i].model & modelflag) &&
+ (cpu_intel_range[i].flag & flag))
+ return 1;
+ break;
+ case X86_VENDOR_AMD:
+ if (cpu_amd_range[i].flag & flag)
+ return 1;
+ break;
+ }
+ }
+
+ /* Invalid */
+ return 0;
+}
+
+static unsigned get_cpu_range(unsigned cpu, unsigned *min, unsigned *max,
+ int index, unsigned flag)
+{
+ unsigned modelflag;
+
+ modelflag = per_cpu(cpu_modelflag, cpu);
+ *max = 0;
+ switch (per_cpu(cpu_model, cpu) >> 16) {
+ case X86_VENDOR_INTEL:
+ if ((cpu_intel_range[index].model & modelflag) &&
+ (cpu_intel_range[index].flag & flag)) {
+ *min = cpu_intel_range[index].min;
+ *max = cpu_intel_range[index].max;
+ }
+ break;
+ case X86_VENDOR_AMD:
+ if (cpu_amd_range[index].flag & flag) {
+ *min = cpu_amd_range[index].min;
+ *max = cpu_amd_range[index].max;
+ }
+ break;
+ }
+
+ return *max;
+}
+
+/* This function can also be called with seq = NULL for printk */
+static void print_cpu_data(struct seq_file *seq, unsigned type,
+ u32 low, u32 high)
+{
+ struct cpu_private *priv;
+ u64 val = high;
+
+ if (seq) {
+ priv = seq->private;
+ if (priv->file) {
+ val = (val << 32) | low;
+ seq_printf(seq, "0x%llx\n", val);
+ } else
+ seq_printf(seq, " %08x: %08x_%08x\n",
+ type, high, low);
+ } else
+ printk(KERN_INFO " %08x: %08x_%08x\n", type, high, low);
+}
+
+/* This function can also be called with seq = NULL for printk */
+static void print_msr(struct seq_file *seq, unsigned cpu, unsigned flag)
+{
+ unsigned msr, msr_min, msr_max;
+ struct cpu_private *priv;
+ u32 low, high;
+ int i, range;
+
+ if (seq) {
+ priv = seq->private;
+ if (priv->file) {
+ if (!rdmsr_safe_on_cpu(priv->cpu, priv->reg,
+ &low, &high))
+ print_cpu_data(seq, priv->reg, low, high);
+ return;
+ }
+ }
+
+ range = get_cpu_range_count(cpu);
+
+ for (i = 0; i < range; i++) {
+ if (!get_cpu_range(cpu, &msr_min, &msr_max, i, flag))
+ continue;
+
+ for (msr = msr_min; msr <= msr_max; msr++) {
+ if (rdmsr_safe_on_cpu(cpu, msr, &low, &high))
+ continue;
+ print_cpu_data(seq, msr, low, high);
+ }
+ }
+}
+
+static void print_tss(void *arg)
+{
+ struct pt_regs *regs = task_pt_regs(current);
+ struct seq_file *seq = arg;
+ unsigned int seg;
+
+ seq_printf(seq, " RAX\t: %016lx\n", regs->ax);
+ seq_printf(seq, " RBX\t: %016lx\n", regs->bx);
+ seq_printf(seq, " RCX\t: %016lx\n", regs->cx);
+ seq_printf(seq, " RDX\t: %016lx\n", regs->dx);
+
+ seq_printf(seq, " RSI\t: %016lx\n", regs->si);
+ seq_printf(seq, " RDI\t: %016lx\n", regs->di);
+ seq_printf(seq, " RBP\t: %016lx\n", regs->bp);
+ seq_printf(seq, " ESP\t: %016lx\n", regs->sp);
+
+#ifdef CONFIG_X86_64
+ seq_printf(seq, " R08\t: %016lx\n", regs->r8);
+ seq_printf(seq, " R09\t: %016lx\n", regs->r9);
+ seq_printf(seq, " R10\t: %016lx\n", regs->r10);
+ seq_printf(seq, " R11\t: %016lx\n", regs->r11);
+ seq_printf(seq, " R12\t: %016lx\n", regs->r12);
+ seq_printf(seq, " R13\t: %016lx\n", regs->r13);
+ seq_printf(seq, " R14\t: %016lx\n", regs->r14);
+ seq_printf(seq, " R15\t: %016lx\n", regs->r15);
+#endif
+
+ asm("movl %%cs,%0" : "=r" (seg));
+ seq_printf(seq, " CS\t: %04x\n", seg);
+ asm("movl %%ds,%0" : "=r" (seg));
+ seq_printf(seq, " DS\t: %04x\n", seg);
+ seq_printf(seq, " SS\t: %04lx\n", regs->ss & 0xffff);
+ asm("movl %%es,%0" : "=r" (seg));
+ seq_printf(seq, " ES\t: %04x\n", seg);
+ asm("movl %%fs,%0" : "=r" (seg));
+ seq_printf(seq, " FS\t: %04x\n", seg);
+ asm("movl %%gs,%0" : "=r" (seg));
+ seq_printf(seq, " GS\t: %04x\n", seg);
+
+ seq_printf(seq, " EFLAGS\t: %016lx\n", regs->flags);
+
+ seq_printf(seq, " EIP\t: %016lx\n", regs->ip);
+}
+
+static void print_cr(void *arg)
+{
+ struct seq_file *seq = arg;
+
+ seq_printf(seq, " cr0\t: %016lx\n", read_cr0());
+ seq_printf(seq, " cr2\t: %016lx\n", read_cr2());
+ seq_printf(seq, " cr3\t: %016lx\n", read_cr3());
+ seq_printf(seq, " cr4\t: %016lx\n", read_cr4_safe());
+#ifdef CONFIG_X86_64
+ seq_printf(seq, " cr8\t: %016lx\n", read_cr8());
+#endif
+}
+
+static void print_desc_ptr(char *str, struct seq_file *seq, struct desc_ptr dt)
+{
+ seq_printf(seq, " %s\t: %016llx\n", str, (u64)(dt.address | dt.size));
+}
+
+static void print_dt(void *seq)
+{
+ struct desc_ptr dt;
+ unsigned long ldt;
+
+ /* IDT */
+ store_idt((struct desc_ptr *)&dt);
+ print_desc_ptr("IDT", seq, dt);
+
+ /* GDT */
+ store_gdt((struct desc_ptr *)&dt);
+ print_desc_ptr("GDT", seq, dt);
+
+ /* LDT */
+ store_ldt(ldt);
+ seq_printf(seq, " LDT\t: %016lx\n", ldt);
+
+ /* TR */
+ store_tr(ldt);
+ seq_printf(seq, " TR\t: %016lx\n", ldt);
+}
+
+static void print_dr(void *arg)
+{
+ struct seq_file *seq = arg;
+ unsigned long dr;
+ int i;
+
+ for (i = 0; i < 8; i++) {
+ /* Ignore db4, db5 */
+ if ((i == 4) || (i == 5))
+ continue;
+ get_debugreg(dr, i);
+ seq_printf(seq, " dr%d\t: %016lx\n", i, dr);
+ }
+
+ seq_printf(seq, "\n MSR\t:\n");
+}
+
+static void print_apic(void *arg)
+{
+ struct seq_file *seq = arg;
+
+#ifdef CONFIG_X86_LOCAL_APIC
+ seq_printf(seq, " LAPIC\t:\n");
+ seq_printf(seq, " ID\t\t: %08x\n", apic_read(APIC_ID) >> 24);
+ seq_printf(seq, " LVR\t\t: %08x\n", apic_read(APIC_LVR));
+ seq_printf(seq, " TASKPRI\t: %08x\n", apic_read(APIC_TASKPRI));
+ seq_printf(seq, " ARBPRI\t\t: %08x\n", apic_read(APIC_ARBPRI));
+ seq_printf(seq, " PROCPRI\t: %08x\n", apic_read(APIC_PROCPRI));
+ seq_printf(seq, " LDR\t\t: %08x\n", apic_read(APIC_LDR));
+ seq_printf(seq, " DFR\t\t: %08x\n", apic_read(APIC_DFR));
+ seq_printf(seq, " SPIV\t\t: %08x\n", apic_read(APIC_SPIV));
+ seq_printf(seq, " ISR\t\t: %08x\n", apic_read(APIC_ISR));
+ seq_printf(seq, " ESR\t\t: %08x\n", apic_read(APIC_ESR));
+ seq_printf(seq, " ICR\t\t: %08x\n", apic_read(APIC_ICR));
+ seq_printf(seq, " ICR2\t\t: %08x\n", apic_read(APIC_ICR2));
+ seq_printf(seq, " LVTT\t\t: %08x\n", apic_read(APIC_LVTT));
+ seq_printf(seq, " LVTTHMR\t: %08x\n", apic_read(APIC_LVTTHMR));
+ seq_printf(seq, " LVTPC\t\t: %08x\n", apic_read(APIC_LVTPC));
+ seq_printf(seq, " LVT0\t\t: %08x\n", apic_read(APIC_LVT0));
+ seq_printf(seq, " LVT1\t\t: %08x\n", apic_read(APIC_LVT1));
+ seq_printf(seq, " LVTERR\t\t: %08x\n", apic_read(APIC_LVTERR));
+ seq_printf(seq, " TMICT\t\t: %08x\n", apic_read(APIC_TMICT));
+ seq_printf(seq, " TMCCT\t\t: %08x\n", apic_read(APIC_TMCCT));
+ seq_printf(seq, " TDCR\t\t: %08x\n", apic_read(APIC_TDCR));
+#endif /* CONFIG_X86_LOCAL_APIC */
+
+ seq_printf(seq, "\n MSR\t:\n");
+}
+
+static int cpu_seq_show(struct seq_file *seq, void *v)
+{
+ struct cpu_private *priv = seq->private;
+
+ if (priv == NULL)
+ return -EINVAL;
+
+ switch (cpu_base[priv->type].flag) {
+ case CPU_TSS:
+ smp_call_function_single(priv->cpu, print_tss, seq, 1);
+ break;
+ case CPU_CR:
+ smp_call_function_single(priv->cpu, print_cr, seq, 1);
+ break;
+ case CPU_DT:
+ smp_call_function_single(priv->cpu, print_dt, seq, 1);
+ break;
+ case CPU_DEBUG:
+ if (priv->file == CPU_INDEX_BIT)
+ smp_call_function_single(priv->cpu, print_dr, seq, 1);
+ print_msr(seq, priv->cpu, cpu_base[priv->type].flag);
+ break;
+ case CPU_APIC:
+ if (priv->file == CPU_INDEX_BIT)
+ smp_call_function_single(priv->cpu, print_apic, seq, 1);
+ print_msr(seq, priv->cpu, cpu_base[priv->type].flag);
+ break;
+
+ default:
+ print_msr(seq, priv->cpu, cpu_base[priv->type].flag);
+ break;
+ }
+ seq_printf(seq, "\n");
+
+ return 0;
+}
+
+static void *cpu_seq_start(struct seq_file *seq, loff_t *pos)
+{
+ if (*pos == 0) /* One time is enough ;-) */
+ return seq;
+
+ return NULL;
+}
+
+static void *cpu_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+ (*pos)++;
+
+ return cpu_seq_start(seq, pos);
+}
+
+static void cpu_seq_stop(struct seq_file *seq, void *v)
+{
+}
+
+static const struct seq_operations cpu_seq_ops = {
+ .start = cpu_seq_start,
+ .next = cpu_seq_next,
+ .stop = cpu_seq_stop,
+ .show = cpu_seq_show,
+};
+
+static int cpu_seq_open(struct inode *inode, struct file *file)
+{
+ struct cpu_private *priv = inode->i_private;
+ struct seq_file *seq;
+ int err;
+
+ err = seq_open(file, &cpu_seq_ops);
+ if (!err) {
+ seq = file->private_data;
+ seq->private = priv;
+ }
+
+ return err;
+}
+
+static const struct file_operations cpu_fops = {
+ .open = cpu_seq_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = seq_release,
+};
+
+static int cpu_create_file(unsigned cpu, unsigned type, unsigned reg,
+ unsigned file, struct dentry *dentry)
+{
+ struct cpu_private *priv = NULL;
+
+ /* Already intialized */
+ if (file == CPU_INDEX_BIT)
+ if (per_cpu(cpu_arr[type].init, cpu))
+ return 0;
+
+ priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+ if (priv == NULL)
+ return -ENOMEM;
+
+ priv->cpu = cpu;
+ priv->type = type;
+ priv->reg = reg;
+ priv->file = file;
+ mutex_lock(&cpu_debug_lock);
+ per_cpu(priv_arr[type], cpu) = priv;
+ per_cpu(cpu_priv_count, cpu)++;
+ mutex_unlock(&cpu_debug_lock);
+
+ if (file)
+ debugfs_create_file(cpu_file[file].name, S_IRUGO,
+ dentry, (void *)priv, &cpu_fops);
+ else {
+ debugfs_create_file(cpu_base[type].name, S_IRUGO,
+ per_cpu(cpu_arr[type].dentry, cpu),
+ (void *)priv, &cpu_fops);
+ mutex_lock(&cpu_debug_lock);
+ per_cpu(cpu_arr[type].init, cpu) = 1;
+ mutex_unlock(&cpu_debug_lock);
+ }
+
+ return 0;
+}
+
+static int cpu_init_regfiles(unsigned cpu, unsigned int type, unsigned reg,
+ struct dentry *dentry)
+{
+ unsigned file;
+ int err = 0;
+
+ for (file = 0; file < ARRAY_SIZE(cpu_file); file++) {
+ err = cpu_create_file(cpu, type, reg, file, dentry);
+ if (err)
+ return err;
+ }
+
+ return err;
+}
+
+static int cpu_init_msr(unsigned cpu, unsigned type, struct dentry *dentry)
+{
+ struct dentry *cpu_dentry = NULL;
+ unsigned reg, reg_min, reg_max;
+ int i, range, err = 0;
+ char reg_dir[12];
+ u32 low, high;
+
+ range = get_cpu_range_count(cpu);
+
+ for (i = 0; i < range; i++) {
+ if (!get_cpu_range(cpu, &reg_min, &reg_max, i,
+ cpu_base[type].flag))
+ continue;
+
+ for (reg = reg_min; reg <= reg_max; reg++) {
+ if (rdmsr_safe_on_cpu(cpu, reg, &low, &high))
+ continue;
+
+ sprintf(reg_dir, "0x%x", reg);
+ cpu_dentry = debugfs_create_dir(reg_dir, dentry);
+ err = cpu_init_regfiles(cpu, type, reg, cpu_dentry);
+ if (err)
+ return err;
+ }
+ }
+
+ return err;
+}
+
+static int cpu_init_allreg(unsigned cpu, struct dentry *dentry)
+{
+ struct dentry *cpu_dentry = NULL;
+ unsigned type;
+ int err = 0;
+
+ for (type = 0; type < ARRAY_SIZE(cpu_base) - 1; type++) {
+ if (!is_typeflag_valid(cpu, cpu_base[type].flag))
+ continue;
+ cpu_dentry = debugfs_create_dir(cpu_base[type].name, dentry);
+ per_cpu(cpu_arr[type].dentry, cpu) = cpu_dentry;
+
+ if (type < CPU_TSS_BIT)
+ err = cpu_init_msr(cpu, type, cpu_dentry);
+ else
+ err = cpu_create_file(cpu, type, 0, CPU_INDEX_BIT,
+ cpu_dentry);
+ if (err)
+ return err;
+ }
+
+ return err;
+}
+
+static int cpu_init_cpu(void)
+{
+ struct dentry *cpu_dentry = NULL;
+ struct cpuinfo_x86 *cpui;
+ char cpu_dir[12];
+ unsigned cpu;
+ int err = 0;
+
+ for (cpu = 0; cpu < nr_cpu_ids; cpu++) {
+ cpui = &cpu_data(cpu);
+ if (!cpu_has(cpui, X86_FEATURE_MSR))
+ continue;
+ per_cpu(cpu_model, cpu) = ((cpui->x86_vendor << 16) |
+ (cpui->x86 << 8) |
+ (cpui->x86_model));
+ per_cpu(cpu_modelflag, cpu) = get_cpu_modelflag(cpu);
+
+ sprintf(cpu_dir, "cpu%d", cpu);
+ cpu_dentry = debugfs_create_dir(cpu_dir, cpu_debugfs_dir);
+ err = cpu_init_allreg(cpu, cpu_dentry);
+
+ pr_info("cpu%d(%d) debug files %d\n",
+ cpu, nr_cpu_ids, per_cpu(cpu_priv_count, cpu));
+ if (per_cpu(cpu_priv_count, cpu) > MAX_CPU_FILES) {
+ pr_err("Register files count %d exceeds limit %d\n",
+ per_cpu(cpu_priv_count, cpu), MAX_CPU_FILES);
+ per_cpu(cpu_priv_count, cpu) = MAX_CPU_FILES;
+ err = -ENFILE;
+ }
+ if (err)
+ return err;
+ }
+
+ return err;
+}
+
+static int __init cpu_debug_init(void)
+{
+ cpu_debugfs_dir = debugfs_create_dir("cpu", arch_debugfs_dir);
+
+ return cpu_init_cpu();
+}
+
+static void __exit cpu_debug_exit(void)
+{
+ int i, cpu;
+
+ if (cpu_debugfs_dir)
+ debugfs_remove_recursive(cpu_debugfs_dir);
+
+ for (cpu = 0; cpu < nr_cpu_ids; cpu++)
+ for (i = 0; i < per_cpu(cpu_priv_count, cpu); i++)
+ kfree(per_cpu(priv_arr[i], cpu));
+}
+
+module_init(cpu_debug_init);
+module_exit(cpu_debug_exit);
+
+MODULE_AUTHOR("Jaswinder Singh Rajput");
+MODULE_DESCRIPTION("CPU Debug module");
+MODULE_LICENSE("GPL");
diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c
index 25c559ba8d5..191117f1ad5 100644
--- a/arch/x86/kernel/cpu/intel.c
+++ b/arch/x86/kernel/cpu/intel.c
@@ -13,6 +13,7 @@
#include <asm/uaccess.h>
#include <asm/ds.h>
#include <asm/bugs.h>
+#include <asm/cpu.h>
#ifdef CONFIG_X86_64
#include <asm/topology.h>
@@ -110,6 +111,28 @@ static void __cpuinit trap_init_f00f_bug(void)
}
#endif
+static void __cpuinit intel_smp_check(struct cpuinfo_x86 *c)
+{
+#ifdef CONFIG_SMP
+ /* calling is from identify_secondary_cpu() ? */
+ if (c->cpu_index == boot_cpu_id)
+ return;
+
+ /*
+ * Mask B, Pentium, but not Pentium MMX
+ */
+ if (c->x86 == 5 &&
+ c->x86_mask >= 1 && c->x86_mask <= 4 &&
+ c->x86_model <= 3) {
+ /*
+ * Remember we have B step Pentia with bugs
+ */
+ WARN_ONCE(1, "WARNING: SMP operation may be unreliable"
+ "with B stepping processors.\n");
+ }
+#endif
+}
+
static void __cpuinit intel_workarounds(struct cpuinfo_x86 *c)
{
unsigned long lo, hi;
@@ -186,6 +209,8 @@ static void __cpuinit intel_workarounds(struct cpuinfo_x86 *c)
#ifdef CONFIG_X86_NUMAQ
numaq_tsc_disable();
#endif
+
+ intel_smp_check(c);
}
#else
static void __cpuinit intel_workarounds(struct cpuinfo_x86 *c)
diff --git a/arch/x86/kernel/cpu/mcheck/Makefile b/arch/x86/kernel/cpu/mcheck/Makefile
index d7d2323bbb6..b2f89829bbe 100644
--- a/arch/x86/kernel/cpu/mcheck/Makefile
+++ b/arch/x86/kernel/cpu/mcheck/Makefile
@@ -4,3 +4,4 @@ obj-$(CONFIG_X86_32) += k7.o p4.o p5.o p6.o winchip.o
obj-$(CONFIG_X86_MCE_INTEL) += mce_intel_64.o
obj-$(CONFIG_X86_MCE_AMD) += mce_amd_64.o
obj-$(CONFIG_X86_MCE_NONFATAL) += non-fatal.o
+obj-$(CONFIG_X86_MCE_THRESHOLD) += threshold.o
diff --git a/arch/x86/kernel/cpu/mcheck/mce_32.c b/arch/x86/kernel/cpu/mcheck/mce_32.c
index dfaebce3633..3552119b091 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_32.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_32.c
@@ -60,20 +60,6 @@ void mcheck_init(struct cpuinfo_x86 *c)
}
}
-static unsigned long old_cr4 __initdata;
-
-void __init stop_mce(void)
-{
- old_cr4 = read_cr4();
- clear_in_cr4(X86_CR4_MCE);
-}
-
-void __init restart_mce(void)
-{
- if (old_cr4 & X86_CR4_MCE)
- set_in_cr4(X86_CR4_MCE);
-}
-
static int __init mcheck_disable(char *str)
{
mce_disabled = 1;
diff --git a/arch/x86/kernel/cpu/mcheck/mce_64.c b/arch/x86/kernel/cpu/mcheck/mce_64.c
index fe79985ce0f..ca14604611e 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_64.c
@@ -3,6 +3,8 @@
* K8 parts Copyright 2002,2003 Andi Kleen, SuSE Labs.
* Rest from unknown author(s).
* 2004 Andi Kleen. Rewrote most of it.
+ * Copyright 2008 Intel Corporation
+ * Author: Andi Kleen
*/
#include <linux/init.h>
@@ -24,6 +26,9 @@
#include <linux/ctype.h>
#include <linux/kmod.h>
#include <linux/kdebug.h>
+#include <linux/kobject.h>
+#include <linux/sysfs.h>
+#include <linux/ratelimit.h>
#include <asm/processor.h>
#include <asm/msr.h>
#include <asm/mce.h>
@@ -32,7 +37,6 @@
#include <asm/idle.h>
#define MISC_MCELOG_MINOR 227
-#define NR_SYSFS_BANKS 6
atomic_t mce_entry;
@@ -47,7 +51,7 @@ static int mce_dont_init;
*/
static int tolerant = 1;
static int banks;
-static unsigned long bank[NR_SYSFS_BANKS] = { [0 ... NR_SYSFS_BANKS-1] = ~0UL };
+static u64 *bank;
static unsigned long notify_user;
static int rip_msr;
static int mce_bootlog = -1;
@@ -58,6 +62,19 @@ static char *trigger_argv[2] = { trigger, NULL };
static DECLARE_WAIT_QUEUE_HEAD(mce_wait);
+/* MCA banks polled by the period polling timer for corrected events */
+DEFINE_PER_CPU(mce_banks_t, mce_poll_banks) = {
+ [0 ... BITS_TO_LONGS(MAX_NR_BANKS)-1] = ~0UL
+};
+
+/* Do initial initialization of a struct mce */
+void mce_setup(struct mce *m)
+{
+ memset(m, 0, sizeof(struct mce));
+ m->cpu = smp_processor_id();
+ rdtscll(m->tsc);
+}
+
/*
* Lockless MCE logging infrastructure.
* This avoids deadlocks on printk locks without having to break locks. Also
@@ -119,11 +136,11 @@ static void print_mce(struct mce *m)
print_symbol("{%s}", m->ip);
printk("\n");
}
- printk(KERN_EMERG "TSC %Lx ", m->tsc);
+ printk(KERN_EMERG "TSC %llx ", m->tsc);
if (m->addr)
- printk("ADDR %Lx ", m->addr);
+ printk("ADDR %llx ", m->addr);
if (m->misc)
- printk("MISC %Lx ", m->misc);
+ printk("MISC %llx ", m->misc);
printk("\n");
printk(KERN_EMERG "This is not a software problem!\n");
printk(KERN_EMERG "Run through mcelog --ascii to decode "
@@ -149,8 +166,10 @@ static void mce_panic(char *msg, struct mce *backup, unsigned long start)
panic(msg);
}
-static int mce_available(struct cpuinfo_x86 *c)
+int mce_available(struct cpuinfo_x86 *c)
{
+ if (mce_dont_init)
+ return 0;
return cpu_has(c, X86_FEATURE_MCE) && cpu_has(c, X86_FEATURE_MCA);
}
@@ -172,7 +191,77 @@ static inline void mce_get_rip(struct mce *m, struct pt_regs *regs)
}
/*
- * The actual machine check handler
+ * Poll for corrected events or events that happened before reset.
+ * Those are just logged through /dev/mcelog.
+ *
+ * This is executed in standard interrupt context.
+ */
+void machine_check_poll(enum mcp_flags flags, mce_banks_t *b)
+{
+ struct mce m;
+ int i;
+
+ mce_setup(&m);
+
+ rdmsrl(MSR_IA32_MCG_STATUS, m.mcgstatus);
+ for (i = 0; i < banks; i++) {
+ if (!bank[i] || !test_bit(i, *b))
+ continue;
+
+ m.misc = 0;
+ m.addr = 0;
+ m.bank = i;
+ m.tsc = 0;
+
+ barrier();
+ rdmsrl(MSR_IA32_MC0_STATUS + i*4, m.status);
+ if (!(m.status & MCI_STATUS_VAL))
+ continue;
+
+ /*
+ * Uncorrected events are handled by the exception handler
+ * when it is enabled. But when the exception is disabled log
+ * everything.
+ *
+ * TBD do the same check for MCI_STATUS_EN here?
+ */
+ if ((m.status & MCI_STATUS_UC) && !(flags & MCP_UC))
+ continue;
+
+ if (m.status & MCI_STATUS_MISCV)
+ rdmsrl(MSR_IA32_MC0_MISC + i*4, m.misc);
+ if (m.status & MCI_STATUS_ADDRV)
+ rdmsrl(MSR_IA32_MC0_ADDR + i*4, m.addr);
+
+ if (!(flags & MCP_TIMESTAMP))
+ m.tsc = 0;
+ /*
+ * Don't get the IP here because it's unlikely to
+ * have anything to do with the actual error location.
+ */
+
+ mce_log(&m);
+ add_taint(TAINT_MACHINE_CHECK);
+
+ /*
+ * Clear state for this bank.
+ */
+ wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0);
+ }
+
+ /*
+ * Don't clear MCG_STATUS here because it's only defined for
+ * exceptions.
+ */
+}
+
+/*
+ * The actual machine check handler. This only handles real
+ * exceptions when something got corrupted coming in through int 18.
+ *
+ * This is executed in NMI context not subject to normal locking rules. This
+ * implies that most kernel services cannot be safely used. Don't even
+ * think about putting a printk in there!
*/
void do_machine_check(struct pt_regs * regs, long error_code)
{
@@ -190,17 +279,18 @@ void do_machine_check(struct pt_regs * regs, long error_code)
* error.
*/
int kill_it = 0;
+ DECLARE_BITMAP(toclear, MAX_NR_BANKS);
atomic_inc(&mce_entry);
- if ((regs
- && notify_die(DIE_NMI, "machine check", regs, error_code,
+ if (notify_die(DIE_NMI, "machine check", regs, error_code,
18, SIGKILL) == NOTIFY_STOP)
- || !banks)
+ goto out2;
+ if (!banks)
goto out2;
- memset(&m, 0, sizeof(struct mce));
- m.cpu = smp_processor_id();
+ mce_setup(&m);
+
rdmsrl(MSR_IA32_MCG_STATUS, m.mcgstatus);
/* if the restart IP is not valid, we're done for */
if (!(m.mcgstatus & MCG_STATUS_RIPV))
@@ -210,18 +300,32 @@ void do_machine_check(struct pt_regs * regs, long error_code)
barrier();
for (i = 0; i < banks; i++) {
- if (i < NR_SYSFS_BANKS && !bank[i])
+ __clear_bit(i, toclear);
+ if (!bank[i])
continue;
m.misc = 0;
m.addr = 0;
m.bank = i;
- m.tsc = 0;
rdmsrl(MSR_IA32_MC0_STATUS + i*4, m.status);
if ((m.status & MCI_STATUS_VAL) == 0)
continue;
+ /*
+ * Non uncorrected errors are handled by machine_check_poll
+ * Leave them alone.
+ */
+ if ((m.status & MCI_STATUS_UC) == 0)
+ continue;
+
+ /*
+ * Set taint even when machine check was not enabled.
+ */
+ add_taint(TAINT_MACHINE_CHECK);
+
+ __set_bit(i, toclear);
+
if (m.status & MCI_STATUS_EN) {
/* if PCC was set, there's no way out */
no_way_out |= !!(m.status & MCI_STATUS_PCC);
@@ -235,6 +339,12 @@ void do_machine_check(struct pt_regs * regs, long error_code)
no_way_out = 1;
kill_it = 1;
}
+ } else {
+ /*
+ * Machine check event was not enabled. Clear, but
+ * ignore.
+ */
+ continue;
}
if (m.status & MCI_STATUS_MISCV)
@@ -243,10 +353,7 @@ void do_machine_check(struct pt_regs * regs, long error_code)
rdmsrl(MSR_IA32_MC0_ADDR + i*4, m.addr);
mce_get_rip(&m, regs);
- if (error_code >= 0)
- rdtscll(m.tsc);
- if (error_code != -2)
- mce_log(&m);
+ mce_log(&m);
/* Did this bank cause the exception? */
/* Assume that the bank with uncorrectable errors did it,
@@ -255,14 +362,8 @@ void do_machine_check(struct pt_regs * regs, long error_code)
panicm = m;
panicm_found = 1;
}
-
- add_taint(TAINT_MACHINE_CHECK);
}
- /* Never do anything final in the polling timer */
- if (!regs)
- goto out;
-
/* If we didn't find an uncorrectable error, pick
the last one (shouldn't happen, just being safe). */
if (!panicm_found)
@@ -309,10 +410,11 @@ void do_machine_check(struct pt_regs * regs, long error_code)
/* notify userspace ASAP */
set_thread_flag(TIF_MCE_NOTIFY);
- out:
/* the last thing we do is clear state */
- for (i = 0; i < banks; i++)
- wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0);
+ for (i = 0; i < banks; i++) {
+ if (test_bit(i, toclear))
+ wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0);
+ }
wrmsrl(MSR_IA32_MCG_STATUS, 0);
out2:
atomic_dec(&mce_entry);
@@ -332,15 +434,13 @@ void do_machine_check(struct pt_regs * regs, long error_code)
* and historically has been the register value of the
* MSR_IA32_THERMAL_STATUS (Intel) msr.
*/
-void mce_log_therm_throt_event(unsigned int cpu, __u64 status)
+void mce_log_therm_throt_event(__u64 status)
{
struct mce m;
- memset(&m, 0, sizeof(m));
- m.cpu = cpu;
+ mce_setup(&m);
m.bank = MCE_THERMAL_BANK;
m.status = status;
- rdtscll(m.tsc);
mce_log(&m);
}
#endif /* CONFIG_X86_MCE_INTEL */
@@ -353,18 +453,18 @@ void mce_log_therm_throt_event(unsigned int cpu, __u64 status)
static int check_interval = 5 * 60; /* 5 minutes */
static int next_interval; /* in jiffies */
-static void mcheck_timer(struct work_struct *work);
-static DECLARE_DELAYED_WORK(mcheck_work, mcheck_timer);
+static void mcheck_timer(unsigned long);
+static DEFINE_PER_CPU(struct timer_list, mce_timer);
-static void mcheck_check_cpu(void *info)
+static void mcheck_timer(unsigned long data)
{
- if (mce_available(&current_cpu_data))
- do_machine_check(NULL, 0);
-}
+ struct timer_list *t = &per_cpu(mce_timer, data);
-static void mcheck_timer(struct work_struct *work)
-{
- on_each_cpu(mcheck_check_cpu, NULL, 1);
+ WARN_ON(smp_processor_id() != data);
+
+ if (mce_available(&current_cpu_data))
+ machine_check_poll(MCP_TIMESTAMP,
+ &__get_cpu_var(mce_poll_banks));
/*
* Alert userspace if needed. If we logged an MCE, reduce the
@@ -377,31 +477,41 @@ static void mcheck_timer(struct work_struct *work)
(int)round_jiffies_relative(check_interval*HZ));
}
- schedule_delayed_work(&mcheck_work, next_interval);
+ t->expires = jiffies + next_interval;
+ add_timer(t);
+}
+
+static void mce_do_trigger(struct work_struct *work)
+{
+ call_usermodehelper(trigger, trigger_argv, NULL, UMH_NO_WAIT);
}
+static DECLARE_WORK(mce_trigger_work, mce_do_trigger);
+
/*
- * This is only called from process context. This is where we do
- * anything we need to alert userspace about new MCEs. This is called
- * directly from the poller and also from entry.S and idle, thanks to
- * TIF_MCE_NOTIFY.
+ * Notify the user(s) about new machine check events.
+ * Can be called from interrupt context, but not from machine check/NMI
+ * context.
*/
int mce_notify_user(void)
{
+ /* Not more than two messages every minute */
+ static DEFINE_RATELIMIT_STATE(ratelimit, 60*HZ, 2);
+
clear_thread_flag(TIF_MCE_NOTIFY);
if (test_and_clear_bit(0, &notify_user)) {
- static unsigned long last_print;
- unsigned long now = jiffies;
-
wake_up_interruptible(&mce_wait);
- if (trigger[0])
- call_usermodehelper(trigger, trigger_argv, NULL,
- UMH_NO_WAIT);
- if (time_after_eq(now, last_print + (check_interval*HZ))) {
- last_print = now;
+ /*
+ * There is no risk of missing notifications because
+ * work_pending is always cleared before the function is
+ * executed.
+ */
+ if (trigger[0] && !work_pending(&mce_trigger_work))
+ schedule_work(&mce_trigger_work);
+
+ if (__ratelimit(&ratelimit))
printk(KERN_INFO "Machine check events logged\n");
- }
return 1;
}
@@ -425,63 +535,78 @@ static struct notifier_block mce_idle_notifier = {
static __init int periodic_mcheck_init(void)
{
- next_interval = check_interval * HZ;
- if (next_interval)
- schedule_delayed_work(&mcheck_work,
- round_jiffies_relative(next_interval));
- idle_notifier_register(&mce_idle_notifier);
- return 0;
+ idle_notifier_register(&mce_idle_notifier);
+ return 0;
}
__initcall(periodic_mcheck_init);
-
/*
* Initialize Machine Checks for a CPU.
*/
-static void mce_init(void *dummy)
+static int mce_cap_init(void)
{
u64 cap;
- int i;
+ unsigned b;
rdmsrl(MSR_IA32_MCG_CAP, cap);
- banks = cap & 0xff;
- if (banks > MCE_EXTENDED_BANK) {
- banks = MCE_EXTENDED_BANK;
- printk(KERN_INFO "MCE: warning: using only %d banks\n",
- MCE_EXTENDED_BANK);
+ b = cap & 0xff;
+ if (b > MAX_NR_BANKS) {
+ printk(KERN_WARNING
+ "MCE: Using only %u machine check banks out of %u\n",
+ MAX_NR_BANKS, b);
+ b = MAX_NR_BANKS;
}
+
+ /* Don't support asymmetric configurations today */
+ WARN_ON(banks != 0 && b != banks);
+ banks = b;
+ if (!bank) {
+ bank = kmalloc(banks * sizeof(u64), GFP_KERNEL);
+ if (!bank)
+ return -ENOMEM;
+ memset(bank, 0xff, banks * sizeof(u64));
+ }
+
/* Use accurate RIP reporting if available. */
if ((cap & (1<<9)) && ((cap >> 16) & 0xff) >= 9)
rip_msr = MSR_IA32_MCG_EIP;
- /* Log the machine checks left over from the previous reset.
- This also clears all registers */
- do_machine_check(NULL, mce_bootlog ? -1 : -2);
+ return 0;
+}
+
+static void mce_init(void *dummy)
+{
+ u64 cap;
+ int i;
+ mce_banks_t all_banks;
+
+ /*
+ * Log the machine checks left over from the previous reset.
+ */
+ bitmap_fill(all_banks, MAX_NR_BANKS);
+ machine_check_poll(MCP_UC, &all_banks);
set_in_cr4(X86_CR4_MCE);
+ rdmsrl(MSR_IA32_MCG_CAP, cap);
if (cap & MCG_CTL_P)
wrmsr(MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff);
for (i = 0; i < banks; i++) {
- if (i < NR_SYSFS_BANKS)
- wrmsrl(MSR_IA32_MC0_CTL+4*i, bank[i]);
- else
- wrmsrl(MSR_IA32_MC0_CTL+4*i, ~0UL);
-
+ wrmsrl(MSR_IA32_MC0_CTL+4*i, bank[i]);
wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0);
}
}
/* Add per CPU specific workarounds here */
-static void __cpuinit mce_cpu_quirks(struct cpuinfo_x86 *c)
+static void mce_cpu_quirks(struct cpuinfo_x86 *c)
{
/* This should be disabled by the BIOS, but isn't always */
if (c->x86_vendor == X86_VENDOR_AMD) {
- if(c->x86 == 15)
+ if (c->x86 == 15 && banks > 4)
/* disable GART TBL walk error reporting, which trips off
incorrectly with the IOMMU & 3ware & Cerberus. */
- clear_bit(10, &bank[4]);
+ clear_bit(10, (unsigned long *)&bank[4]);
if(c->x86 <= 17 && mce_bootlog < 0)
/* Lots of broken BIOS around that don't clear them
by default and leave crap in there. Don't log. */
@@ -504,20 +629,38 @@ static void mce_cpu_features(struct cpuinfo_x86 *c)
}
}
+static void mce_init_timer(void)
+{
+ struct timer_list *t = &__get_cpu_var(mce_timer);
+
+ /* data race harmless because everyone sets to the same value */
+ if (!next_interval)
+ next_interval = check_interval * HZ;
+ if (!next_interval)
+ return;
+ setup_timer(t, mcheck_timer, smp_processor_id());
+ t->expires = round_jiffies(jiffies + next_interval);
+ add_timer(t);
+}
+
/*
* Called for each booted CPU to set up machine checks.
* Must be called with preempt off.
*/
void __cpuinit mcheck_init(struct cpuinfo_x86 *c)
{
- mce_cpu_quirks(c);
+ if (!mce_available(c))
+ return;
- if (mce_dont_init ||
- !mce_available(c))
+ if (mce_cap_init() < 0) {
+ mce_dont_init = 1;
return;
+ }
+ mce_cpu_quirks(c);
mce_init(NULL);
mce_cpu_features(c);
+ mce_init_timer();
}
/*
@@ -573,7 +716,7 @@ static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize,
{
unsigned long *cpu_tsc;
static DEFINE_MUTEX(mce_read_mutex);
- unsigned next;
+ unsigned prev, next;
char __user *buf = ubuf;
int i, err;
@@ -592,25 +735,32 @@ static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize,
}
err = 0;
- for (i = 0; i < next; i++) {
- unsigned long start = jiffies;
-
- while (!mcelog.entry[i].finished) {
- if (time_after_eq(jiffies, start + 2)) {
- memset(mcelog.entry + i,0, sizeof(struct mce));
- goto timeout;
+ prev = 0;
+ do {
+ for (i = prev; i < next; i++) {
+ unsigned long start = jiffies;
+
+ while (!mcelog.entry[i].finished) {
+ if (time_after_eq(jiffies, start + 2)) {
+ memset(mcelog.entry + i, 0,
+ sizeof(struct mce));
+ goto timeout;
+ }
+ cpu_relax();
}
- cpu_relax();
+ smp_rmb();
+ err |= copy_to_user(buf, mcelog.entry + i,
+ sizeof(struct mce));
+ buf += sizeof(struct mce);
+timeout:
+ ;
}
- smp_rmb();
- err |= copy_to_user(buf, mcelog.entry + i, sizeof(struct mce));
- buf += sizeof(struct mce);
- timeout:
- ;
- }
- memset(mcelog.entry, 0, next * sizeof(struct mce));
- mcelog.next = 0;
+ memset(mcelog.entry + prev, 0,
+ (next - prev) * sizeof(struct mce));
+ prev = next;
+ next = cmpxchg(&mcelog.next, prev, 0);
+ } while (next != prev);
synchronize_sched();
@@ -680,20 +830,6 @@ static struct miscdevice mce_log_device = {
&mce_chrdev_ops,
};
-static unsigned long old_cr4 __initdata;
-
-void __init stop_mce(void)
-{
- old_cr4 = read_cr4();
- clear_in_cr4(X86_CR4_MCE);
-}
-
-void __init restart_mce(void)
-{
- if (old_cr4 & X86_CR4_MCE)
- set_in_cr4(X86_CR4_MCE);
-}
-
/*
* Old style boot options parsing. Only for compatibility.
*/
@@ -703,8 +839,7 @@ static int __init mcheck_disable(char *str)
return 1;
}
-/* mce=off disables machine check. Note you can re-enable it later
- using sysfs.
+/* mce=off disables machine check.
mce=TOLERANCELEVEL (number, see above)
mce=bootlog Log MCEs from before booting. Disabled by default on AMD.
mce=nobootlog Don't log MCEs from before booting. */
@@ -728,6 +863,29 @@ __setup("mce=", mcheck_enable);
* Sysfs support
*/
+/*
+ * Disable machine checks on suspend and shutdown. We can't really handle
+ * them later.
+ */
+static int mce_disable(void)
+{
+ int i;
+
+ for (i = 0; i < banks; i++)
+ wrmsrl(MSR_IA32_MC0_CTL + i*4, 0);
+ return 0;
+}
+
+static int mce_suspend(struct sys_device *dev, pm_message_t state)
+{
+ return mce_disable();
+}
+
+static int mce_shutdown(struct sys_device *dev)
+{
+ return mce_disable();
+}
+
/* On resume clear all MCE state. Don't want to see leftovers from the BIOS.
Only one CPU is active at this time, the others get readded later using
CPU hotplug. */
@@ -738,20 +896,24 @@ static int mce_resume(struct sys_device *dev)
return 0;
}
+static void mce_cpu_restart(void *data)
+{
+ del_timer_sync(&__get_cpu_var(mce_timer));
+ if (mce_available(&current_cpu_data))
+ mce_init(NULL);
+ mce_init_timer();
+}
+
/* Reinit MCEs after user configuration changes */
static void mce_restart(void)
{
- if (next_interval)
- cancel_delayed_work(&mcheck_work);
- /* Timer race is harmless here */
- on_each_cpu(mce_init, NULL, 1);
next_interval = check_interval * HZ;
- if (next_interval)
- schedule_delayed_work(&mcheck_work,
- round_jiffies_relative(next_interval));
+ on_each_cpu(mce_cpu_restart, NULL, 1);
}
static struct sysdev_class mce_sysclass = {
+ .suspend = mce_suspend,
+ .shutdown = mce_shutdown,
.resume = mce_resume,
.name = "machinecheck",
};
@@ -778,16 +940,26 @@ void (*threshold_cpu_callback)(unsigned long action, unsigned int cpu) __cpuinit
} \
static SYSDEV_ATTR(name, 0644, show_ ## name, set_ ## name);
-/*
- * TBD should generate these dynamically based on number of available banks.
- * Have only 6 contol banks in /sysfs until then.
- */
-ACCESSOR(bank0ctl,bank[0],mce_restart())
-ACCESSOR(bank1ctl,bank[1],mce_restart())
-ACCESSOR(bank2ctl,bank[2],mce_restart())
-ACCESSOR(bank3ctl,bank[3],mce_restart())
-ACCESSOR(bank4ctl,bank[4],mce_restart())
-ACCESSOR(bank5ctl,bank[5],mce_restart())
+static struct sysdev_attribute *bank_attrs;
+
+static ssize_t show_bank(struct sys_device *s, struct sysdev_attribute *attr,
+ char *buf)
+{
+ u64 b = bank[attr - bank_attrs];
+ return sprintf(buf, "%llx\n", b);
+}
+
+static ssize_t set_bank(struct sys_device *s, struct sysdev_attribute *attr,
+ const char *buf, size_t siz)
+{
+ char *end;
+ u64 new = simple_strtoull(buf, &end, 0);
+ if (end == buf)
+ return -EINVAL;
+ bank[attr - bank_attrs] = new;
+ mce_restart();
+ return end-buf;
+}
static ssize_t show_trigger(struct sys_device *s, struct sysdev_attribute *attr,
char *buf)
@@ -814,8 +986,6 @@ static SYSDEV_ATTR(trigger, 0644, show_trigger, set_trigger);
static SYSDEV_INT_ATTR(tolerant, 0644, tolerant);
ACCESSOR(check_interval,check_interval,mce_restart())
static struct sysdev_attribute *mce_attributes[] = {
- &attr_bank0ctl, &attr_bank1ctl, &attr_bank2ctl,
- &attr_bank3ctl, &attr_bank4ctl, &attr_bank5ctl,
&attr_tolerant.attr, &attr_check_interval, &attr_trigger,
NULL
};
@@ -845,11 +1015,22 @@ static __cpuinit int mce_create_device(unsigned int cpu)
if (err)
goto error;
}
+ for (i = 0; i < banks; i++) {
+ err = sysdev_create_file(&per_cpu(device_mce, cpu),
+ &bank_attrs[i]);
+ if (err)
+ goto error2;
+ }
cpu_set(cpu, mce_device_initialized);
return 0;
+error2:
+ while (--i >= 0) {
+ sysdev_remove_file(&per_cpu(device_mce, cpu),
+ &bank_attrs[i]);
+ }
error:
- while (i--) {
+ while (--i >= 0) {
sysdev_remove_file(&per_cpu(device_mce,cpu),
mce_attributes[i]);
}
@@ -868,15 +1049,46 @@ static __cpuinit void mce_remove_device(unsigned int cpu)
for (i = 0; mce_attributes[i]; i++)
sysdev_remove_file(&per_cpu(device_mce,cpu),
mce_attributes[i]);
+ for (i = 0; i < banks; i++)
+ sysdev_remove_file(&per_cpu(device_mce, cpu),
+ &bank_attrs[i]);
sysdev_unregister(&per_cpu(device_mce,cpu));
cpu_clear(cpu, mce_device_initialized);
}
+/* Make sure there are no machine checks on offlined CPUs. */
+static void mce_disable_cpu(void *h)
+{
+ int i;
+ unsigned long action = *(unsigned long *)h;
+
+ if (!mce_available(&current_cpu_data))
+ return;
+ if (!(action & CPU_TASKS_FROZEN))
+ cmci_clear();
+ for (i = 0; i < banks; i++)
+ wrmsrl(MSR_IA32_MC0_CTL + i*4, 0);
+}
+
+static void mce_reenable_cpu(void *h)
+{
+ int i;
+ unsigned long action = *(unsigned long *)h;
+
+ if (!mce_available(&current_cpu_data))
+ return;
+ if (!(action & CPU_TASKS_FROZEN))
+ cmci_reenable();
+ for (i = 0; i < banks; i++)
+ wrmsrl(MSR_IA32_MC0_CTL + i*4, bank[i]);
+}
+
/* Get notified when a cpu comes on/off. Be hotplug friendly. */
static int __cpuinit mce_cpu_callback(struct notifier_block *nfb,
unsigned long action, void *hcpu)
{
unsigned int cpu = (unsigned long)hcpu;
+ struct timer_list *t = &per_cpu(mce_timer, cpu);
switch (action) {
case CPU_ONLINE:
@@ -891,6 +1103,21 @@ static int __cpuinit mce_cpu_callback(struct notifier_block *nfb,
threshold_cpu_callback(action, cpu);
mce_remove_device(cpu);
break;
+ case CPU_DOWN_PREPARE:
+ case CPU_DOWN_PREPARE_FROZEN:
+ del_timer_sync(t);
+ smp_call_function_single(cpu, mce_disable_cpu, &action, 1);
+ break;
+ case CPU_DOWN_FAILED:
+ case CPU_DOWN_FAILED_FROZEN:
+ t->expires = round_jiffies(jiffies + next_interval);
+ add_timer_on(t, cpu);
+ smp_call_function_single(cpu, mce_reenable_cpu, &action, 1);
+ break;
+ case CPU_POST_DEAD:
+ /* intentionally ignoring frozen here */
+ cmci_rediscover(cpu);
+ break;
}
return NOTIFY_OK;
}
@@ -899,6 +1126,34 @@ static struct notifier_block mce_cpu_notifier __cpuinitdata = {
.notifier_call = mce_cpu_callback,
};
+static __init int mce_init_banks(void)
+{
+ int i;
+
+ bank_attrs = kzalloc(sizeof(struct sysdev_attribute) * banks,
+ GFP_KERNEL);
+ if (!bank_attrs)
+ return -ENOMEM;
+
+ for (i = 0; i < banks; i++) {
+ struct sysdev_attribute *a = &bank_attrs[i];
+ a->attr.name = kasprintf(GFP_KERNEL, "bank%d", i);
+ if (!a->attr.name)
+ goto nomem;
+ a->attr.mode = 0644;
+ a->show = show_bank;
+ a->store = set_bank;
+ }
+ return 0;
+
+nomem:
+ while (--i >= 0)
+ kfree(bank_attrs[i].attr.name);
+ kfree(bank_attrs);
+ bank_attrs = NULL;
+ return -ENOMEM;
+}
+
static __init int mce_init_device(void)
{
int err;
@@ -906,6 +1161,11 @@ static __init int mce_init_device(void)
if (!mce_available(&boot_cpu_data))
return -EIO;
+
+ err = mce_init_banks();
+ if (err)
+ return err;
+
err = sysdev_class_register(&mce_sysclass);
if (err)
return err;
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
index 9817506dd46..c5a32f92d07 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
@@ -79,6 +79,8 @@ static unsigned char shared_bank[NR_BANKS] = {
static DEFINE_PER_CPU(unsigned char, bank_map); /* see which banks are on */
+static void amd_threshold_interrupt(void);
+
/*
* CPU Initialization
*/
@@ -174,6 +176,8 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c)
tr.reset = 0;
tr.old_limit = 0;
threshold_restart_bank(&tr);
+
+ mce_threshold_vector = amd_threshold_interrupt;
}
}
}
@@ -187,19 +191,13 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c)
* the interrupt goes off when error_count reaches threshold_limit.
* the handler will simply log mcelog w/ software defined bank number.
*/
-asmlinkage void mce_threshold_interrupt(void)
+static void amd_threshold_interrupt(void)
{
unsigned int bank, block;
struct mce m;
u32 low = 0, high = 0, address = 0;
- ack_APIC_irq();
- exit_idle();
- irq_enter();
-
- memset(&m, 0, sizeof(m));
- rdtscll(m.tsc);
- m.cpu = smp_processor_id();
+ mce_setup(&m);
/* assume first bank caused it */
for (bank = 0; bank < NR_BANKS; ++bank) {
@@ -233,7 +231,8 @@ asmlinkage void mce_threshold_interrupt(void)
/* Log the machine check that caused the threshold
event. */
- do_machine_check(NULL, 0);
+ machine_check_poll(MCP_TIMESTAMP,
+ &__get_cpu_var(mce_poll_banks));
if (high & MASK_OVERFLOW_HI) {
rdmsrl(address, m.misc);
@@ -243,13 +242,10 @@ asmlinkage void mce_threshold_interrupt(void)
+ bank * NR_BLOCKS
+ block;
mce_log(&m);
- goto out;
+ return;
}
}
}
-out:
- inc_irq_stat(irq_threshold_count);
- irq_exit();
}
/*
diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c
index aa5e287c98e..aaa7d973093 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c
@@ -1,6 +1,8 @@
/*
* Intel specific MCE features.
* Copyright 2004 Zwane Mwaikambo <zwane@linuxpower.ca>
+ * Copyright (C) 2008, 2009 Intel Corporation
+ * Author: Andi Kleen
*/
#include <linux/init.h>
@@ -13,6 +15,7 @@
#include <asm/hw_irq.h>
#include <asm/idle.h>
#include <asm/therm_throt.h>
+#include <asm/apic.h>
asmlinkage void smp_thermal_interrupt(void)
{
@@ -25,7 +28,7 @@ asmlinkage void smp_thermal_interrupt(void)
rdmsrl(MSR_IA32_THERM_STATUS, msr_val);
if (therm_throt_process(msr_val & 1))
- mce_log_therm_throt_event(smp_processor_id(), msr_val);
+ mce_log_therm_throt_event(msr_val);
inc_irq_stat(irq_thermal_count);
irq_exit();
@@ -85,7 +88,209 @@ static void intel_init_thermal(struct cpuinfo_x86 *c)
return;
}
+/*
+ * Support for Intel Correct Machine Check Interrupts. This allows
+ * the CPU to raise an interrupt when a corrected machine check happened.
+ * Normally we pick those up using a regular polling timer.
+ * Also supports reliable discovery of shared banks.
+ */
+
+static DEFINE_PER_CPU(mce_banks_t, mce_banks_owned);
+
+/*
+ * cmci_discover_lock protects against parallel discovery attempts
+ * which could race against each other.
+ */
+static DEFINE_SPINLOCK(cmci_discover_lock);
+
+#define CMCI_THRESHOLD 1
+
+static int cmci_supported(int *banks)
+{
+ u64 cap;
+
+ /*
+ * Vendor check is not strictly needed, but the initial
+ * initialization is vendor keyed and this
+ * makes sure none of the backdoors are entered otherwise.
+ */
+ if (boot_cpu_data.x86_vendor != X86_VENDOR_INTEL)
+ return 0;
+ if (!cpu_has_apic || lapic_get_maxlvt() < 6)
+ return 0;
+ rdmsrl(MSR_IA32_MCG_CAP, cap);
+ *banks = min_t(unsigned, MAX_NR_BANKS, cap & 0xff);
+ return !!(cap & MCG_CMCI_P);
+}
+
+/*
+ * The interrupt handler. This is called on every event.
+ * Just call the poller directly to log any events.
+ * This could in theory increase the threshold under high load,
+ * but doesn't for now.
+ */
+static void intel_threshold_interrupt(void)
+{
+ machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_banks_owned));
+ mce_notify_user();
+}
+
+static void print_update(char *type, int *hdr, int num)
+{
+ if (*hdr == 0)
+ printk(KERN_INFO "CPU %d MCA banks", smp_processor_id());
+ *hdr = 1;
+ printk(KERN_CONT " %s:%d", type, num);
+}
+
+/*
+ * Enable CMCI (Corrected Machine Check Interrupt) for available MCE banks
+ * on this CPU. Use the algorithm recommended in the SDM to discover shared
+ * banks.
+ */
+static void cmci_discover(int banks, int boot)
+{
+ unsigned long *owned = (void *)&__get_cpu_var(mce_banks_owned);
+ int hdr = 0;
+ int i;
+
+ spin_lock(&cmci_discover_lock);
+ for (i = 0; i < banks; i++) {
+ u64 val;
+
+ if (test_bit(i, owned))
+ continue;
+
+ rdmsrl(MSR_IA32_MC0_CTL2 + i, val);
+
+ /* Already owned by someone else? */
+ if (val & CMCI_EN) {
+ if (test_and_clear_bit(i, owned) || boot)
+ print_update("SHD", &hdr, i);
+ __clear_bit(i, __get_cpu_var(mce_poll_banks));
+ continue;
+ }
+
+ val |= CMCI_EN | CMCI_THRESHOLD;
+ wrmsrl(MSR_IA32_MC0_CTL2 + i, val);
+ rdmsrl(MSR_IA32_MC0_CTL2 + i, val);
+
+ /* Did the enable bit stick? -- the bank supports CMCI */
+ if (val & CMCI_EN) {
+ if (!test_and_set_bit(i, owned) || boot)
+ print_update("CMCI", &hdr, i);
+ __clear_bit(i, __get_cpu_var(mce_poll_banks));
+ } else {
+ WARN_ON(!test_bit(i, __get_cpu_var(mce_poll_banks)));
+ }
+ }
+ spin_unlock(&cmci_discover_lock);
+ if (hdr)
+ printk(KERN_CONT "\n");
+}
+
+/*
+ * Just in case we missed an event during initialization check
+ * all the CMCI owned banks.
+ */
+void cmci_recheck(void)
+{
+ unsigned long flags;
+ int banks;
+
+ if (!mce_available(&current_cpu_data) || !cmci_supported(&banks))
+ return;
+ local_irq_save(flags);
+ machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_banks_owned));
+ local_irq_restore(flags);
+}
+
+/*
+ * Disable CMCI on this CPU for all banks it owns when it goes down.
+ * This allows other CPUs to claim the banks on rediscovery.
+ */
+void cmci_clear(void)
+{
+ int i;
+ int banks;
+ u64 val;
+
+ if (!cmci_supported(&banks))
+ return;
+ spin_lock(&cmci_discover_lock);
+ for (i = 0; i < banks; i++) {
+ if (!test_bit(i, __get_cpu_var(mce_banks_owned)))
+ continue;
+ /* Disable CMCI */
+ rdmsrl(MSR_IA32_MC0_CTL2 + i, val);
+ val &= ~(CMCI_EN|CMCI_THRESHOLD_MASK);
+ wrmsrl(MSR_IA32_MC0_CTL2 + i, val);
+ __clear_bit(i, __get_cpu_var(mce_banks_owned));
+ }
+ spin_unlock(&cmci_discover_lock);
+}
+
+/*
+ * After a CPU went down cycle through all the others and rediscover
+ * Must run in process context.
+ */
+void cmci_rediscover(int dying)
+{
+ int banks;
+ int cpu;
+ cpumask_var_t old;
+
+ if (!cmci_supported(&banks))
+ return;
+ if (!alloc_cpumask_var(&old, GFP_KERNEL))
+ return;
+ cpumask_copy(old, &current->cpus_allowed);
+
+ for_each_online_cpu (cpu) {
+ if (cpu == dying)
+ continue;
+ if (set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu)))
+ continue;
+ /* Recheck banks in case CPUs don't all have the same */
+ if (cmci_supported(&banks))
+ cmci_discover(banks, 0);
+ }
+
+ set_cpus_allowed_ptr(current, old);
+ free_cpumask_var(old);
+}
+
+/*
+ * Reenable CMCI on this CPU in case a CPU down failed.
+ */
+void cmci_reenable(void)
+{
+ int banks;
+ if (cmci_supported(&banks))
+ cmci_discover(banks, 0);
+}
+
+static __cpuinit void intel_init_cmci(void)
+{
+ int banks;
+
+ if (!cmci_supported(&banks))
+ return;
+
+ mce_threshold_vector = intel_threshold_interrupt;
+ cmci_discover(banks, 1);
+ /*
+ * For CPU #0 this runs with still disabled APIC, but that's
+ * ok because only the vector is set up. We still do another
+ * check for the banks later for CPU #0 just to make sure
+ * to not miss any events.
+ */
+ apic_write(APIC_LVTCMCI, THRESHOLD_APIC_VECTOR|APIC_DM_FIXED);
+ cmci_recheck();
+}
+
void mce_intel_feature_init(struct cpuinfo_x86 *c)
{
intel_init_thermal(c);
+ intel_init_cmci();
}
diff --git a/arch/x86/kernel/cpu/mcheck/threshold.c b/arch/x86/kernel/cpu/mcheck/threshold.c
new file mode 100644
index 00000000000..23ee9e730f7
--- /dev/null
+++ b/arch/x86/kernel/cpu/mcheck/threshold.c
@@ -0,0 +1,29 @@
+/*
+ * Common corrected MCE threshold handler code:
+ */
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+
+#include <asm/irq_vectors.h>
+#include <asm/apic.h>
+#include <asm/idle.h>
+#include <asm/mce.h>
+
+static void default_threshold_interrupt(void)
+{
+ printk(KERN_ERR "Unexpected threshold interrupt at vector %x\n",
+ THRESHOLD_APIC_VECTOR);
+}
+
+void (*mce_threshold_vector)(void) = default_threshold_interrupt;
+
+asmlinkage void mce_threshold_interrupt(void)
+{
+ exit_idle();
+ irq_enter();
+ inc_irq_stat(irq_threshold_count);
+ mce_threshold_vector();
+ irq_exit();
+ /* Ack only at the end to avoid potential reentry */
+ ack_APIC_irq();
+}
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-29 09:27:07 +0000