mce.h 7.0 KB
Newer Older
H
H. Peter Anvin 已提交
1 2
#ifndef _ASM_X86_MCE_H
#define _ASM_X86_MCE_H
3

4
#include <linux/types.h>
5 6 7 8 9 10
#include <asm/ioctls.h>

/*
 * Machine Check support for x86
 */

11
#define MCG_BANKCNT_MASK	0xff         /* Number of Banks */
12
#define MCG_CTL_P		(1ULL<<8)    /* MCG_CTL register available */
13 14 15 16 17
#define MCG_EXT_P		(1ULL<<9)    /* Extended registers available */
#define MCG_CMCI_P		(1ULL<<10)   /* CMCI supported */
#define MCG_EXT_CNT_MASK	0xff0000     /* Number of Extended registers */
#define MCG_EXT_CNT_SHIFT	16
#define MCG_EXT_CNT(c)		(((c) & MCG_EXT_CNT_MASK) >> MCG_EXT_CNT_SHIFT)
A
Andi Kleen 已提交
18
#define MCG_SER_P	 	(1ULL<<24)   /* MCA recovery/new status bits */
19

20 21 22
#define MCG_STATUS_RIPV  (1ULL<<0)   /* restart ip valid */
#define MCG_STATUS_EIPV  (1ULL<<1)   /* ip points to correct instruction */
#define MCG_STATUS_MCIP  (1ULL<<2)   /* machine check in progress */
23

24 25 26 27 28 29 30
#define MCI_STATUS_VAL   (1ULL<<63)  /* valid error */
#define MCI_STATUS_OVER  (1ULL<<62)  /* previous errors lost */
#define MCI_STATUS_UC    (1ULL<<61)  /* uncorrected error */
#define MCI_STATUS_EN    (1ULL<<60)  /* error enabled */
#define MCI_STATUS_MISCV (1ULL<<59)  /* misc error reg. valid */
#define MCI_STATUS_ADDRV (1ULL<<58)  /* addr reg. valid */
#define MCI_STATUS_PCC   (1ULL<<57)  /* processor context corrupt */
A
Andi Kleen 已提交
31 32 33 34 35 36 37 38 39
#define MCI_STATUS_S	 (1ULL<<56)  /* Signaled machine check */
#define MCI_STATUS_AR	 (1ULL<<55)  /* Action required */

/* MISC register defines */
#define MCM_ADDR_SEGOFF  0	/* segment offset */
#define MCM_ADDR_LINEAR  1	/* linear address */
#define MCM_ADDR_PHYS	 2	/* physical address */
#define MCM_ADDR_MEM	 3	/* memory address */
#define MCM_ADDR_GENERIC 7	/* generic */
40

41 42 43 44 45 46
#define MCJ_CTX_MASK		3
#define MCJ_CTX(flags)		((flags) & MCJ_CTX_MASK)
#define MCJ_CTX_RANDOM		0    /* inject context: random */
#define MCJ_CTX_PROCESS		1    /* inject context: process */
#define MCJ_CTX_IRQ		2    /* inject context: IRQ */
#define MCJ_NMI_BROADCAST	4    /* do NMI broadcasting */
47
#define MCJ_EXCEPTION		8    /* raise as exception */
48

49 50 51 52 53 54
/* Fields are zero when not available */
struct mce {
	__u64 status;
	__u64 misc;
	__u64 addr;
	__u64 mcgstatus;
55
	__u64 ip;
56
	__u64 tsc;	/* cpu time stamp counter */
57 58
	__u64 time;	/* wall time_t when error was detected */
	__u8  cpuvendor;	/* cpu vendor as encoded in system.h */
59 60
	__u8  inject_flags;	/* software inject flags */
	__u16  pad;
61
	__u32 cpuid;	/* CPUID 1 EAX */
62 63
	__u8  cs;		/* code segment */
	__u8  bank;	/* machine check bank */
64
	__u8  cpu;	/* cpu number; obsolete; use extcpu now */
65
	__u8  finished;   /* entry is valid */
66
	__u32 extcpu;	/* linux cpu number that detected the error */
67 68 69
	__u32 socketid;	/* CPU socket ID */
	__u32 apicid;	/* CPU initial apic ID */
	__u64 mcgcap;	/* MCGCAP MSR: machine check capabilities of CPU */
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
};

/*
 * This structure contains all data related to the MCE log.  Also
 * carries a signature to make it easier to find from external
 * debugging tools.  Each entry is only valid when its finished flag
 * is set.
 */

#define MCE_LOG_LEN 32

struct mce_log {
	char signature[12]; /* "MACHINECHECK" */
	unsigned len;	    /* = MCE_LOG_LEN */
	unsigned next;
	unsigned flags;
86
	unsigned recordlen;	/* length of struct mce */
87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110
	struct mce entry[MCE_LOG_LEN];
};

#define MCE_OVERFLOW 0		/* bit 0 in flags means overflow */

#define MCE_LOG_SIGNATURE	"MACHINECHECK"

#define MCE_GET_RECORD_LEN   _IOR('M', 1, int)
#define MCE_GET_LOG_LEN      _IOR('M', 2, int)
#define MCE_GETCLEAR_FLAGS   _IOR('M', 3, int)

/* Software defined banks */
#define MCE_EXTENDED_BANK	128
#define MCE_THERMAL_BANK	MCE_EXTENDED_BANK + 0

#define K8_MCE_THRESHOLD_BASE      (MCE_EXTENDED_BANK + 1)      /* MCE_AMD */
#define K8_MCE_THRESHOLD_BANK_0    (MCE_THRESHOLD_BASE + 0 * 9)
#define K8_MCE_THRESHOLD_BANK_1    (MCE_THRESHOLD_BASE + 1 * 9)
#define K8_MCE_THRESHOLD_BANK_2    (MCE_THRESHOLD_BASE + 2 * 9)
#define K8_MCE_THRESHOLD_BANK_3    (MCE_THRESHOLD_BASE + 3 * 9)
#define K8_MCE_THRESHOLD_BANK_4    (MCE_THRESHOLD_BASE + 4 * 9)
#define K8_MCE_THRESHOLD_BANK_5    (MCE_THRESHOLD_BASE + 5 * 9)
#define K8_MCE_THRESHOLD_DRAM_ECC  (MCE_THRESHOLD_BANK_4 + 0)

111

112 113
#ifdef __KERNEL__

114 115
extern struct atomic_notifier_head x86_mce_decoder_chain;

H
Hidetoshi Seto 已提交
116 117 118 119
#include <linux/percpu.h>
#include <linux/init.h>
#include <asm/atomic.h>

120
extern int mce_disabled;
121
extern int mce_p5_enabled;
122

H
Hidetoshi Seto 已提交
123
#ifdef CONFIG_X86_MCE
124
int mcheck_init(void);
125
void mcheck_cpu_init(struct cpuinfo_x86 *c);
H
Hidetoshi Seto 已提交
126
#else
127
static inline int mcheck_init(void) { return 0; }
128
static inline void mcheck_cpu_init(struct cpuinfo_x86 *c) {}
H
Hidetoshi Seto 已提交
129 130
#endif

H
Hidetoshi Seto 已提交
131 132 133
#ifdef CONFIG_X86_ANCIENT_MCE
void intel_p5_mcheck_init(struct cpuinfo_x86 *c);
void winchip_mcheck_init(struct cpuinfo_x86 *c);
134
static inline void enable_p5_mce(void) { mce_p5_enabled = 1; }
H
Hidetoshi Seto 已提交
135 136 137
#else
static inline void intel_p5_mcheck_init(struct cpuinfo_x86 *c) {}
static inline void winchip_mcheck_init(struct cpuinfo_x86 *c) {}
138
static inline void enable_p5_mce(void) {}
H
Hidetoshi Seto 已提交
139 140
#endif

141 142
extern void (*x86_mce_decode_callback)(struct mce *m);

143
void mce_setup(struct mce *m);
144
void mce_log(struct mce *m);
I
Ingo Molnar 已提交
145
DECLARE_PER_CPU(struct sys_device, mce_dev);
146

A
Andi Kleen 已提交
147
/*
148 149 150
 * Maximum banks number.
 * This is the limit of the current register layout on
 * Intel CPUs.
A
Andi Kleen 已提交
151
 */
152
#define MAX_NR_BANKS 32
A
Andi Kleen 已提交
153

154
#ifdef CONFIG_X86_MCE_INTEL
155 156
extern int mce_cmci_disabled;
extern int mce_ignore_ce;
157
void mce_intel_feature_init(struct cpuinfo_x86 *c);
A
Andi Kleen 已提交
158 159 160 161
void cmci_clear(void);
void cmci_reenable(void);
void cmci_rediscover(int dying);
void cmci_recheck(void);
162 163
#else
static inline void mce_intel_feature_init(struct cpuinfo_x86 *c) { }
A
Andi Kleen 已提交
164 165 166 167
static inline void cmci_clear(void) {}
static inline void cmci_reenable(void) {}
static inline void cmci_rediscover(int dying) {}
static inline void cmci_recheck(void) {}
168 169 170 171 172 173 174 175
#endif

#ifdef CONFIG_X86_MCE_AMD
void mce_amd_feature_init(struct cpuinfo_x86 *c);
#else
static inline void mce_amd_feature_init(struct cpuinfo_x86 *c) { }
#endif

176
int mce_available(struct cpuinfo_x86 *c);
A
Andi Kleen 已提交
177

178
DECLARE_PER_CPU(unsigned, mce_exception_count);
179
DECLARE_PER_CPU(unsigned, mce_poll_count);
180

181 182
extern atomic_t mce_entry;

183 184 185
typedef DECLARE_BITMAP(mce_banks_t, MAX_NR_BANKS);
DECLARE_PER_CPU(mce_banks_t, mce_poll_banks);

186 187 188
enum mcp_flags {
	MCP_TIMESTAMP = (1 << 0),	/* log time stamp */
	MCP_UC = (1 << 1),		/* log uncorrected errors */
A
Andi Kleen 已提交
189
	MCP_DONTLOG = (1 << 2),		/* only clear, don't log */
190
};
191
void machine_check_poll(enum mcp_flags flags, mce_banks_t *b);
192

193
int mce_notify_irq(void);
194
void mce_notify_process(void);
195

196 197 198
DECLARE_PER_CPU(struct mce, injectm);
extern struct file_operations mce_chrdev_ops;

H
Hidetoshi Seto 已提交
199 200 201 202 203 204 205 206 207 208 209
/*
 * Exception handler
 */

/* Call the installed machine check handler for this CPU setup. */
extern void (*machine_check_vector)(struct pt_regs *, long error_code);
void do_machine_check(struct pt_regs *, long);

/*
 * Threshold handler
 */
210

211
extern void (*mce_threshold_vector)(void);
H
Hidetoshi Seto 已提交
212
extern void (*threshold_cpu_callback)(unsigned long action, unsigned int cpu);
213

214 215 216 217 218 219 220
/*
 * Thermal handler
 */

void intel_init_thermal(struct cpuinfo_x86 *c);

void mce_log_therm_throt_event(__u64 status);
221 222 223 224 225 226 227

#ifdef CONFIG_X86_THERMAL_VECTOR
extern void mcheck_intel_therm_init(void);
#else
static inline void mcheck_intel_therm_init(void) { }
#endif

228
#endif /* __KERNEL__ */
H
H. Peter Anvin 已提交
229
#endif /* _ASM_X86_MCE_H */