提交 81faaae4 编写于 作者: I Ingo Molnar

Merge branch 'x86/pebs' into x86/unify-cpu-detect

Conflicts:
	arch/x86/Kconfig.cpu
	include/asm-x86/ds.h
Signed-off-by: NIngo Molnar <mingo@elte.hu>
......@@ -477,3 +477,21 @@ config CPU_SUP_UMC_32
depends on !64BIT
help
This enables extended support for UMC processors
config X86_DS
bool "Debug Store support"
default y
help
Add support for Debug Store.
This allows the kernel to provide a memory buffer to the hardware
to store various profiling and tracing events.
config X86_PTRACE_BTS
bool "ptrace interface to Branch Trace Store"
default y
depends on (X86_DS && X86_DEBUGCTLMSR)
help
Add a ptrace interface to allow collecting an execution trace
of the traced task.
This collects control flow changes in a (cyclic) buffer and allows
debuggers to fill in the gaps and show an execution trace of the debuggee.
......@@ -221,10 +221,11 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c)
set_cpu_cap(c, X86_FEATURE_BTS);
if (!(l1 & (1<<12)))
set_cpu_cap(c, X86_FEATURE_PEBS);
ds_init_intel(c);
}
if (cpu_has_bts)
ds_init_intel(c);
ptrace_bts_init_intel(c);
/*
* See if we have a good local APIC by checking for buggy Pentia,
......
此差异已折叠。
......@@ -277,6 +277,14 @@ void exit_thread(void)
tss->x86_tss.io_bitmap_base = INVALID_IO_BITMAP_OFFSET;
put_cpu();
}
#ifdef CONFIG_X86_DS
/* Free any DS contexts that have not been properly released. */
if (unlikely(current->thread.ds_ctx)) {
/* we clear debugctl to make sure DS is not used. */
update_debugctlmsr(0);
ds_free(current->thread.ds_ctx);
}
#endif /* CONFIG_X86_DS */
}
void flush_thread(void)
......@@ -438,6 +446,35 @@ int set_tsc_mode(unsigned int val)
return 0;
}
#ifdef CONFIG_X86_DS
static int update_debugctl(struct thread_struct *prev,
struct thread_struct *next, unsigned long debugctl)
{
unsigned long ds_prev = 0;
unsigned long ds_next = 0;
if (prev->ds_ctx)
ds_prev = (unsigned long)prev->ds_ctx->ds;
if (next->ds_ctx)
ds_next = (unsigned long)next->ds_ctx->ds;
if (ds_next != ds_prev) {
/* we clear debugctl to make sure DS
* is not in use when we change it */
debugctl = 0;
update_debugctlmsr(0);
wrmsr(MSR_IA32_DS_AREA, ds_next, 0);
}
return debugctl;
}
#else
static int update_debugctl(struct thread_struct *prev,
struct thread_struct *next, unsigned long debugctl)
{
return debugctl;
}
#endif /* CONFIG_X86_DS */
static noinline void
__switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p,
struct tss_struct *tss)
......@@ -448,14 +485,7 @@ __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p,
prev = &prev_p->thread;
next = &next_p->thread;
debugctl = prev->debugctlmsr;
if (next->ds_area_msr != prev->ds_area_msr) {
/* we clear debugctl to make sure DS
* is not in use when we change it */
debugctl = 0;
update_debugctlmsr(0);
wrmsr(MSR_IA32_DS_AREA, next->ds_area_msr, 0);
}
debugctl = update_debugctl(prev, next, prev->debugctlmsr);
if (next->debugctlmsr != debugctl)
update_debugctlmsr(next->debugctlmsr);
......@@ -479,13 +509,13 @@ __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p,
hard_enable_TSC();
}
#ifdef X86_BTS
#ifdef CONFIG_X86_PTRACE_BTS
if (test_tsk_thread_flag(prev_p, TIF_BTS_TRACE_TS))
ptrace_bts_take_timestamp(prev_p, BTS_TASK_DEPARTS);
if (test_tsk_thread_flag(next_p, TIF_BTS_TRACE_TS))
ptrace_bts_take_timestamp(next_p, BTS_TASK_ARRIVES);
#endif
#endif /* CONFIG_X86_PTRACE_BTS */
if (!test_tsk_thread_flag(next_p, TIF_IO_BITMAP)) {
......
......@@ -239,6 +239,14 @@ void exit_thread(void)
t->io_bitmap_max = 0;
put_cpu();
}
#ifdef CONFIG_X86_DS
/* Free any DS contexts that have not been properly released. */
if (unlikely(t->ds_ctx)) {
/* we clear debugctl to make sure DS is not used. */
update_debugctlmsr(0);
ds_free(t->ds_ctx);
}
#endif /* CONFIG_X86_DS */
}
void flush_thread(void)
......@@ -472,13 +480,27 @@ static inline void __switch_to_xtra(struct task_struct *prev_p,
next = &next_p->thread;
debugctl = prev->debugctlmsr;
if (next->ds_area_msr != prev->ds_area_msr) {
/* we clear debugctl to make sure DS
* is not in use when we change it */
debugctl = 0;
update_debugctlmsr(0);
wrmsrl(MSR_IA32_DS_AREA, next->ds_area_msr);
#ifdef CONFIG_X86_DS
{
unsigned long ds_prev = 0, ds_next = 0;
if (prev->ds_ctx)
ds_prev = (unsigned long)prev->ds_ctx->ds;
if (next->ds_ctx)
ds_next = (unsigned long)next->ds_ctx->ds;
if (ds_next != ds_prev) {
/*
* We clear debugctl to make sure DS
* is not in use when we change it:
*/
debugctl = 0;
update_debugctlmsr(0);
wrmsrl(MSR_IA32_DS_AREA, ds_next);
}
}
#endif /* CONFIG_X86_DS */
if (next->debugctlmsr != debugctl)
update_debugctlmsr(next->debugctlmsr);
......@@ -516,13 +538,13 @@ static inline void __switch_to_xtra(struct task_struct *prev_p,
memset(tss->io_bitmap, 0xff, prev->io_bitmap_max);
}
#ifdef X86_BTS
#ifdef CONFIG_X86_PTRACE_BTS
if (test_tsk_thread_flag(prev_p, TIF_BTS_TRACE_TS))
ptrace_bts_take_timestamp(prev_p, BTS_TASK_DEPARTS);
if (test_tsk_thread_flag(next_p, TIF_BTS_TRACE_TS))
ptrace_bts_take_timestamp(next_p, BTS_TASK_ARRIVES);
#endif
#endif /* CONFIG_X86_PTRACE_BTS */
}
/*
......
......@@ -554,45 +554,115 @@ static int ptrace_set_debugreg(struct task_struct *child,
return 0;
}
#ifdef X86_BTS
#ifdef CONFIG_X86_PTRACE_BTS
/*
* The configuration for a particular BTS hardware implementation.
*/
struct bts_configuration {
/* the size of a BTS record in bytes; at most BTS_MAX_RECORD_SIZE */
unsigned char sizeof_bts;
/* the size of a field in the BTS record in bytes */
unsigned char sizeof_field;
/* a bitmask to enable/disable BTS in DEBUGCTL MSR */
unsigned long debugctl_mask;
};
static struct bts_configuration bts_cfg;
#define BTS_MAX_RECORD_SIZE (8 * 3)
/*
* Branch Trace Store (BTS) uses the following format. Different
* architectures vary in the size of those fields.
* - source linear address
* - destination linear address
* - flags
*
* Later architectures use 64bit pointers throughout, whereas earlier
* architectures use 32bit pointers in 32bit mode.
*
* We compute the base address for the first 8 fields based on:
* - the field size stored in the DS configuration
* - the relative field position
*
* In order to store additional information in the BTS buffer, we use
* a special source address to indicate that the record requires
* special interpretation.
*
* Netburst indicated via a bit in the flags field whether the branch
* was predicted; this is ignored.
*/
enum bts_field {
bts_from = 0,
bts_to,
bts_flags,
bts_escape = (unsigned long)-1,
bts_qual = bts_to,
bts_jiffies = bts_flags
};
static int ptrace_bts_get_size(struct task_struct *child)
static inline unsigned long bts_get(const char *base, enum bts_field field)
{
if (!child->thread.ds_area_msr)
return -ENXIO;
base += (bts_cfg.sizeof_field * field);
return *(unsigned long *)base;
}
static inline void bts_set(char *base, enum bts_field field, unsigned long val)
{
base += (bts_cfg.sizeof_field * field);;
(*(unsigned long *)base) = val;
}
return ds_get_bts_index((void *)child->thread.ds_area_msr);
/*
* Translate a BTS record from the raw format into the bts_struct format
*
* out (out): bts_struct interpretation
* raw: raw BTS record
*/
static void ptrace_bts_translate_record(struct bts_struct *out, const void *raw)
{
memset(out, 0, sizeof(*out));
if (bts_get(raw, bts_from) == bts_escape) {
out->qualifier = bts_get(raw, bts_qual);
out->variant.jiffies = bts_get(raw, bts_jiffies);
} else {
out->qualifier = BTS_BRANCH;
out->variant.lbr.from_ip = bts_get(raw, bts_from);
out->variant.lbr.to_ip = bts_get(raw, bts_to);
}
}
static int ptrace_bts_read_record(struct task_struct *child,
long index,
static int ptrace_bts_read_record(struct task_struct *child, size_t index,
struct bts_struct __user *out)
{
struct bts_struct ret;
int retval;
int bts_end;
int bts_index;
const void *bts_record;
size_t bts_index, bts_end;
int error;
if (!child->thread.ds_area_msr)
return -ENXIO;
error = ds_get_bts_end(child, &bts_end);
if (error < 0)
return error;
if (index < 0)
return -EINVAL;
bts_end = ds_get_bts_end((void *)child->thread.ds_area_msr);
if (bts_end <= index)
return -EINVAL;
error = ds_get_bts_index(child, &bts_index);
if (error < 0)
return error;
/* translate the ptrace bts index into the ds bts index */
bts_index = ds_get_bts_index((void *)child->thread.ds_area_msr);
bts_index -= (index + 1);
if (bts_index < 0)
bts_index += bts_end;
bts_index += bts_end - (index + 1);
if (bts_end <= bts_index)
bts_index -= bts_end;
error = ds_access_bts(child, bts_index, &bts_record);
if (error < 0)
return error;
retval = ds_read_bts((void *)child->thread.ds_area_msr,
bts_index, &ret);
if (retval < 0)
return retval;
ptrace_bts_translate_record(&ret, bts_record);
if (copy_to_user(out, &ret, sizeof(ret)))
return -EFAULT;
......@@ -600,101 +670,106 @@ static int ptrace_bts_read_record(struct task_struct *child,
return sizeof(ret);
}
static int ptrace_bts_clear(struct task_struct *child)
{
if (!child->thread.ds_area_msr)
return -ENXIO;
return ds_clear((void *)child->thread.ds_area_msr);
}
static int ptrace_bts_drain(struct task_struct *child,
long size,
struct bts_struct __user *out)
{
int end, i;
void *ds = (void *)child->thread.ds_area_msr;
if (!ds)
return -ENXIO;
struct bts_struct ret;
const unsigned char *raw;
size_t end, i;
int error;
end = ds_get_bts_index(ds);
if (end <= 0)
return end;
error = ds_get_bts_index(child, &end);
if (error < 0)
return error;
if (size < (end * sizeof(struct bts_struct)))
return -EIO;
for (i = 0; i < end; i++, out++) {
struct bts_struct ret;
int retval;
error = ds_access_bts(child, 0, (const void **)&raw);
if (error < 0)
return error;
retval = ds_read_bts(ds, i, &ret);
if (retval < 0)
return retval;
for (i = 0; i < end; i++, out++, raw += bts_cfg.sizeof_bts) {
ptrace_bts_translate_record(&ret, raw);
if (copy_to_user(out, &ret, sizeof(ret)))
return -EFAULT;
}
ds_clear(ds);
error = ds_clear_bts(child);
if (error < 0)
return error;
return end;
}
static void ptrace_bts_ovfl(struct task_struct *child)
{
send_sig(child->thread.bts_ovfl_signal, child, 0);
}
static int ptrace_bts_config(struct task_struct *child,
long cfg_size,
const struct ptrace_bts_config __user *ucfg)
{
struct ptrace_bts_config cfg;
int bts_size, ret = 0;
void *ds;
int error = 0;
error = -EOPNOTSUPP;
if (!bts_cfg.sizeof_bts)
goto errout;
error = -EIO;
if (cfg_size < sizeof(cfg))
return -EIO;
goto errout;
error = -EFAULT;
if (copy_from_user(&cfg, ucfg, sizeof(cfg)))
return -EFAULT;
goto errout;
if ((int)cfg.size < 0)
return -EINVAL;
error = -EINVAL;
if ((cfg.flags & PTRACE_BTS_O_SIGNAL) &&
!(cfg.flags & PTRACE_BTS_O_ALLOC))
goto errout;
bts_size = 0;
ds = (void *)child->thread.ds_area_msr;
if (ds) {
bts_size = ds_get_bts_size(ds);
if (bts_size < 0)
return bts_size;
}
cfg.size = PAGE_ALIGN(cfg.size);
if (cfg.flags & PTRACE_BTS_O_ALLOC) {
ds_ovfl_callback_t ovfl = 0;
unsigned int sig = 0;
/* we ignore the error in case we were not tracing child */
(void)ds_release_bts(child);
if (cfg.flags & PTRACE_BTS_O_SIGNAL) {
if (!cfg.signal)
goto errout;
sig = cfg.signal;
ovfl = ptrace_bts_ovfl;
}
if (bts_size != cfg.size) {
ret = ptrace_bts_realloc(child, cfg.size,
cfg.flags & PTRACE_BTS_O_CUT_SIZE);
if (ret < 0)
error = ds_request_bts(child, /* base = */ 0, cfg.size, ovfl);
if (error < 0)
goto errout;
ds = (void *)child->thread.ds_area_msr;
child->thread.bts_ovfl_signal = sig;
}
if (cfg.flags & PTRACE_BTS_O_SIGNAL)
ret = ds_set_overflow(ds, DS_O_SIGNAL);
else
ret = ds_set_overflow(ds, DS_O_WRAP);
if (ret < 0)
error = -EINVAL;
if (!child->thread.ds_ctx && cfg.flags)
goto errout;
if (cfg.flags & PTRACE_BTS_O_TRACE)
child->thread.debugctlmsr |= ds_debugctl_mask();
child->thread.debugctlmsr |= bts_cfg.debugctl_mask;
else
child->thread.debugctlmsr &= ~ds_debugctl_mask();
child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask;
if (cfg.flags & PTRACE_BTS_O_SCHED)
set_tsk_thread_flag(child, TIF_BTS_TRACE_TS);
else
clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS);
ret = sizeof(cfg);
error = sizeof(cfg);
out:
if (child->thread.debugctlmsr)
......@@ -702,10 +777,10 @@ static int ptrace_bts_config(struct task_struct *child,
else
clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR);
return ret;
return error;
errout:
child->thread.debugctlmsr &= ~ds_debugctl_mask();
child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask;
clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS);
goto out;
}
......@@ -714,29 +789,40 @@ static int ptrace_bts_status(struct task_struct *child,
long cfg_size,
struct ptrace_bts_config __user *ucfg)
{
void *ds = (void *)child->thread.ds_area_msr;
struct ptrace_bts_config cfg;
size_t end;
const void *base, *max;
int error;
if (cfg_size < sizeof(cfg))
return -EIO;
memset(&cfg, 0, sizeof(cfg));
error = ds_get_bts_end(child, &end);
if (error < 0)
return error;
if (ds) {
cfg.size = ds_get_bts_size(ds);
error = ds_access_bts(child, /* index = */ 0, &base);
if (error < 0)
return error;
if (ds_get_overflow(ds) == DS_O_SIGNAL)
cfg.flags |= PTRACE_BTS_O_SIGNAL;
error = ds_access_bts(child, /* index = */ end, &max);
if (error < 0)
return error;
if (test_tsk_thread_flag(child, TIF_DEBUGCTLMSR) &&
child->thread.debugctlmsr & ds_debugctl_mask())
cfg.flags |= PTRACE_BTS_O_TRACE;
memset(&cfg, 0, sizeof(cfg));
cfg.size = (max - base);
cfg.signal = child->thread.bts_ovfl_signal;
cfg.bts_size = sizeof(struct bts_struct);
if (test_tsk_thread_flag(child, TIF_BTS_TRACE_TS))
cfg.flags |= PTRACE_BTS_O_SCHED;
}
if (cfg.signal)
cfg.flags |= PTRACE_BTS_O_SIGNAL;
cfg.bts_size = sizeof(struct bts_struct);
if (test_tsk_thread_flag(child, TIF_DEBUGCTLMSR) &&
child->thread.debugctlmsr & bts_cfg.debugctl_mask)
cfg.flags |= PTRACE_BTS_O_TRACE;
if (test_tsk_thread_flag(child, TIF_BTS_TRACE_TS))
cfg.flags |= PTRACE_BTS_O_SCHED;
if (copy_to_user(ucfg, &cfg, sizeof(cfg)))
return -EFAULT;
......@@ -744,89 +830,38 @@ static int ptrace_bts_status(struct task_struct *child,
return sizeof(cfg);
}
static int ptrace_bts_write_record(struct task_struct *child,
const struct bts_struct *in)
{
int retval;
unsigned char bts_record[BTS_MAX_RECORD_SIZE];
if (!child->thread.ds_area_msr)
return -ENXIO;
BUG_ON(BTS_MAX_RECORD_SIZE < bts_cfg.sizeof_bts);
retval = ds_write_bts((void *)child->thread.ds_area_msr, in);
if (retval)
return retval;
memset(bts_record, 0, bts_cfg.sizeof_bts);
switch (in->qualifier) {
case BTS_INVALID:
break;
return sizeof(*in);
}
case BTS_BRANCH:
bts_set(bts_record, bts_from, in->variant.lbr.from_ip);
bts_set(bts_record, bts_to, in->variant.lbr.to_ip);
break;
static int ptrace_bts_realloc(struct task_struct *child,
int size, int reduce_size)
{
unsigned long rlim, vm;
int ret, old_size;
case BTS_TASK_ARRIVES:
case BTS_TASK_DEPARTS:
bts_set(bts_record, bts_from, bts_escape);
bts_set(bts_record, bts_qual, in->qualifier);
bts_set(bts_record, bts_jiffies, in->variant.jiffies);
break;
if (size < 0)
default:
return -EINVAL;
old_size = ds_get_bts_size((void *)child->thread.ds_area_msr);
if (old_size < 0)
return old_size;
ret = ds_free((void **)&child->thread.ds_area_msr);
if (ret < 0)
goto out;
size >>= PAGE_SHIFT;
old_size >>= PAGE_SHIFT;
current->mm->total_vm -= old_size;
current->mm->locked_vm -= old_size;
if (size == 0)
goto out;
rlim = current->signal->rlim[RLIMIT_AS].rlim_cur >> PAGE_SHIFT;
vm = current->mm->total_vm + size;
if (rlim < vm) {
ret = -ENOMEM;
if (!reduce_size)
goto out;
size = rlim - current->mm->total_vm;
if (size <= 0)
goto out;
}
rlim = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT;
vm = current->mm->locked_vm + size;
if (rlim < vm) {
ret = -ENOMEM;
if (!reduce_size)
goto out;
size = rlim - current->mm->locked_vm;
if (size <= 0)
goto out;
}
ret = ds_allocate((void **)&child->thread.ds_area_msr,
size << PAGE_SHIFT);
if (ret < 0)
goto out;
current->mm->total_vm += size;
current->mm->locked_vm += size;
out:
if (child->thread.ds_area_msr)
set_tsk_thread_flag(child, TIF_DS_AREA_MSR);
else
clear_tsk_thread_flag(child, TIF_DS_AREA_MSR);
return ret;
/* The writing task will be the switched-to task on a context
* switch. It needs to write into the switched-from task's BTS
* buffer. */
return ds_unchecked_write_bts(child, bts_record, bts_cfg.sizeof_bts);
}
void ptrace_bts_take_timestamp(struct task_struct *tsk,
......@@ -839,7 +874,66 @@ void ptrace_bts_take_timestamp(struct task_struct *tsk,
ptrace_bts_write_record(tsk, &rec);
}
#endif /* X86_BTS */
static const struct bts_configuration bts_cfg_netburst = {
.sizeof_bts = sizeof(long) * 3,
.sizeof_field = sizeof(long),
.debugctl_mask = (1<<2)|(1<<3)|(1<<5)
};
static const struct bts_configuration bts_cfg_pentium_m = {
.sizeof_bts = sizeof(long) * 3,
.sizeof_field = sizeof(long),
.debugctl_mask = (1<<6)|(1<<7)
};
static const struct bts_configuration bts_cfg_core2 = {
.sizeof_bts = 8 * 3,
.sizeof_field = 8,
.debugctl_mask = (1<<6)|(1<<7)|(1<<9)
};
static inline void bts_configure(const struct bts_configuration *cfg)
{
bts_cfg = *cfg;
}
void __cpuinit ptrace_bts_init_intel(struct cpuinfo_x86 *c)
{
switch (c->x86) {
case 0x6:
switch (c->x86_model) {
case 0xD:
case 0xE: /* Pentium M */
bts_configure(&bts_cfg_pentium_m);
break;
case 0xF: /* Core2 */
case 0x1C: /* Atom */
bts_configure(&bts_cfg_core2);
break;
default:
/* sorry, don't know about them */
break;
}
break;
case 0xF:
switch (c->x86_model) {
case 0x0:
case 0x1:
case 0x2: /* Netburst */
bts_configure(&bts_cfg_netburst);
break;
default:
/* sorry, don't know about them */
break;
}
break;
default:
/* sorry, don't know about them */
break;
}
}
#endif /* CONFIG_X86_PTRACE_BTS */
/*
* Called by kernel/ptrace.c when detaching..
......@@ -852,15 +946,15 @@ void ptrace_disable(struct task_struct *child)
#ifdef TIF_SYSCALL_EMU
clear_tsk_thread_flag(child, TIF_SYSCALL_EMU);
#endif
if (child->thread.ds_area_msr) {
#ifdef X86_BTS
ptrace_bts_realloc(child, 0, 0);
#endif
child->thread.debugctlmsr &= ~ds_debugctl_mask();
if (!child->thread.debugctlmsr)
clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR);
clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS);
}
#ifdef CONFIG_X86_PTRACE_BTS
(void)ds_release_bts(child);
child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask;
if (!child->thread.debugctlmsr)
clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR);
clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS);
#endif /* CONFIG_X86_PTRACE_BTS */
}
#if defined CONFIG_X86_32 || defined CONFIG_IA32_EMULATION
......@@ -980,7 +1074,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data)
/*
* These bits need more cooking - not enabled yet:
*/
#ifdef X86_BTS
#ifdef CONFIG_X86_PTRACE_BTS
case PTRACE_BTS_CONFIG:
ret = ptrace_bts_config
(child, data, (struct ptrace_bts_config __user *)addr);
......@@ -992,7 +1086,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data)
break;
case PTRACE_BTS_SIZE:
ret = ptrace_bts_get_size(child);
ret = ds_get_bts_index(child, /* pos = */ 0);
break;
case PTRACE_BTS_GET:
......@@ -1001,14 +1095,14 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data)
break;
case PTRACE_BTS_CLEAR:
ret = ptrace_bts_clear(child);
ret = ds_clear_bts(child);
break;
case PTRACE_BTS_DRAIN:
ret = ptrace_bts_drain
(child, data, (struct bts_struct __user *) addr);
break;
#endif
#endif /* CONFIG_X86_PTRACE_BTS */
default:
ret = ptrace_request(child, request, addr, data);
......
......@@ -2,71 +2,237 @@
* Debug Store (DS) support
*
* This provides a low-level interface to the hardware's Debug Store
* feature that is used for last branch recording (LBR) and
* feature that is used for branch trace store (BTS) and
* precise-event based sampling (PEBS).
*
* Different architectures use a different DS layout/pointer size.
* The below functions therefore work on a void*.
* It manages:
* - per-thread and per-cpu allocation of BTS and PEBS
* - buffer memory allocation (optional)
* - buffer overflow handling
* - buffer access
*
* It assumes:
* - get_task_struct on all parameter tasks
* - current is allowed to trace parameter tasks
*
* Since there is no user for PEBS, yet, only LBR (or branch
* trace store, BTS) is supported.
*
*
* Copyright (C) 2007 Intel Corporation.
* Markus Metzger <markus.t.metzger@intel.com>, Dec 2007
* Copyright (C) 2007-2008 Intel Corporation.
* Markus Metzger <markus.t.metzger@intel.com>, 2007-2008
*/
#ifndef ASM_X86__DS_H
#define ASM_X86__DS_H
#ifdef CONFIG_X86_DS
#include <linux/types.h>
#include <linux/init.h>
struct cpuinfo_x86;
struct task_struct;
/* a branch trace record entry
/*
* Request BTS or PEBS
*
* Due to alignement constraints, the actual buffer may be slightly
* smaller than the requested or provided buffer.
*
* In order to unify the interface between various processor versions,
* we use the below data structure for all processors.
* Returns 0 on success; -Eerrno otherwise
*
* task: the task to request recording for;
* NULL for per-cpu recording on the current cpu
* base: the base pointer for the (non-pageable) buffer;
* NULL if buffer allocation requested
* size: the size of the requested or provided buffer
* ovfl: pointer to a function to be called on buffer overflow;
* NULL if cyclic buffer requested
*/
enum bts_qualifier {
BTS_INVALID = 0,
BTS_BRANCH,
BTS_TASK_ARRIVES,
BTS_TASK_DEPARTS
};
typedef void (*ds_ovfl_callback_t)(struct task_struct *);
extern int ds_request_bts(struct task_struct *task, void *base, size_t size,
ds_ovfl_callback_t ovfl);
extern int ds_request_pebs(struct task_struct *task, void *base, size_t size,
ds_ovfl_callback_t ovfl);
/*
* Release BTS or PEBS resources
*
* Frees buffers allocated on ds_request.
*
* Returns 0 on success; -Eerrno otherwise
*
* task: the task to release resources for;
* NULL to release resources for the current cpu
*/
extern int ds_release_bts(struct task_struct *task);
extern int ds_release_pebs(struct task_struct *task);
/*
* Return the (array) index of the write pointer.
* (assuming an array of BTS/PEBS records)
*
* Returns -Eerrno on error
*
* task: the task to access;
* NULL to access the current cpu
* pos (out): if not NULL, will hold the result
*/
extern int ds_get_bts_index(struct task_struct *task, size_t *pos);
extern int ds_get_pebs_index(struct task_struct *task, size_t *pos);
/*
* Return the (array) index one record beyond the end of the array.
* (assuming an array of BTS/PEBS records)
*
* Returns -Eerrno on error
*
* task: the task to access;
* NULL to access the current cpu
* pos (out): if not NULL, will hold the result
*/
extern int ds_get_bts_end(struct task_struct *task, size_t *pos);
extern int ds_get_pebs_end(struct task_struct *task, size_t *pos);
/*
* Provide a pointer to the BTS/PEBS record at parameter index.
* (assuming an array of BTS/PEBS records)
*
* The pointer points directly into the buffer. The user is
* responsible for copying the record.
*
* Returns the size of a single record on success; -Eerrno on error
*
* task: the task to access;
* NULL to access the current cpu
* index: the index of the requested record
* record (out): pointer to the requested record
*/
extern int ds_access_bts(struct task_struct *task,
size_t index, const void **record);
extern int ds_access_pebs(struct task_struct *task,
size_t index, const void **record);
/*
* Write one or more BTS/PEBS records at the write pointer index and
* advance the write pointer.
*
* If size is not a multiple of the record size, trailing bytes are
* zeroed out.
*
* May result in one or more overflow notifications.
*
* If called during overflow handling, that is, with index >=
* interrupt threshold, the write will wrap around.
*
* An overflow notification is given if and when the interrupt
* threshold is reached during or after the write.
*
* Returns the number of bytes written or -Eerrno.
*
* task: the task to access;
* NULL to access the current cpu
* buffer: the buffer to write
* size: the size of the buffer
*/
extern int ds_write_bts(struct task_struct *task,
const void *buffer, size_t size);
extern int ds_write_pebs(struct task_struct *task,
const void *buffer, size_t size);
/*
* Same as ds_write_bts/pebs, but omit ownership checks.
*
* This is needed to have some other task than the owner of the
* BTS/PEBS buffer or the parameter task itself write into the
* respective buffer.
*/
extern int ds_unchecked_write_bts(struct task_struct *task,
const void *buffer, size_t size);
extern int ds_unchecked_write_pebs(struct task_struct *task,
const void *buffer, size_t size);
/*
* Reset the write pointer of the BTS/PEBS buffer.
*
* Returns 0 on success; -Eerrno on error
*
* task: the task to access;
* NULL to access the current cpu
*/
extern int ds_reset_bts(struct task_struct *task);
extern int ds_reset_pebs(struct task_struct *task);
/*
* Clear the BTS/PEBS buffer and reset the write pointer.
* The entire buffer will be zeroed out.
*
* Returns 0 on success; -Eerrno on error
*
* task: the task to access;
* NULL to access the current cpu
*/
extern int ds_clear_bts(struct task_struct *task);
extern int ds_clear_pebs(struct task_struct *task);
/*
* Provide the PEBS counter reset value.
*
* Returns 0 on success; -Eerrno on error
*
* task: the task to access;
* NULL to access the current cpu
* value (out): the counter reset value
*/
extern int ds_get_pebs_reset(struct task_struct *task, u64 *value);
/*
* Set the PEBS counter reset value.
*
* Returns 0 on success; -Eerrno on error
*
* task: the task to access;
* NULL to access the current cpu
* value: the new counter reset value
*/
extern int ds_set_pebs_reset(struct task_struct *task, u64 value);
/*
* Initialization
*/
struct cpuinfo_x86;
extern void __cpuinit ds_init_intel(struct cpuinfo_x86 *);
struct bts_struct {
u64 qualifier;
union {
/* BTS_BRANCH */
struct {
u64 from_ip;
u64 to_ip;
} lbr;
/* BTS_TASK_ARRIVES or
BTS_TASK_DEPARTS */
u64 jiffies;
} variant;
/*
* The DS context - part of struct thread_struct.
*/
struct ds_context {
/* pointer to the DS configuration; goes into MSR_IA32_DS_AREA */
unsigned char *ds;
/* the owner of the BTS and PEBS configuration, respectively */
struct task_struct *owner[2];
/* buffer overflow notification function for BTS and PEBS */
ds_ovfl_callback_t callback[2];
/* the original buffer address */
void *buffer[2];
/* the number of allocated pages for on-request allocated buffers */
unsigned int pages[2];
/* use count */
unsigned long count;
/* a pointer to the context location inside the thread_struct
* or the per_cpu context array */
struct ds_context **this;
/* a pointer to the task owning this context, or NULL, if the
* context is owned by a cpu */
struct task_struct *task;
};
/* Overflow handling mechanisms */
#define DS_O_SIGNAL 1 /* send overflow signal */
#define DS_O_WRAP 2 /* wrap around */
extern int ds_allocate(void **, size_t);
extern int ds_free(void **);
extern int ds_get_bts_size(void *);
extern int ds_get_bts_end(void *);
extern int ds_get_bts_index(void *);
extern int ds_set_overflow(void *, int);
extern int ds_get_overflow(void *);
extern int ds_clear(void *);
extern int ds_read_bts(void *, int, struct bts_struct *);
extern int ds_write_bts(void *, const struct bts_struct *);
extern unsigned long ds_debugctl_mask(void);
extern void __cpuinit ds_init_intel(struct cpuinfo_x86 *c);
/* called by exit_thread() to free leftover contexts */
extern void ds_free(struct ds_context *context);
#else /* CONFIG_X86_DS */
#define ds_init_intel(config) do {} while (0)
#endif /* CONFIG_X86_DS */
#endif /* ASM_X86__DS_H */
......@@ -20,6 +20,7 @@ struct mm_struct;
#include <asm/msr.h>
#include <asm/desc_defs.h>
#include <asm/nops.h>
#include <asm/ds.h>
#include <linux/personality.h>
#include <linux/cpumask.h>
......@@ -430,9 +431,14 @@ struct thread_struct {
unsigned io_bitmap_max;
/* MSR_IA32_DEBUGCTLMSR value to switch in if TIF_DEBUGCTLMSR is set. */
unsigned long debugctlmsr;
/* Debug Store - if not 0 points to a DS Save Area configuration;
* goes into MSR_IA32_DS_AREA */
unsigned long ds_area_msr;
#ifdef CONFIG_X86_DS
/* Debug Store context; see include/asm-x86/ds.h; goes into MSR_IA32_DS_AREA */
struct ds_context *ds_ctx;
#endif /* CONFIG_X86_DS */
#ifdef CONFIG_X86_PTRACE_BTS
/* the signal to send on a bts buffer overflow */
unsigned int bts_ovfl_signal;
#endif /* CONFIG_X86_PTRACE_BTS */
};
static inline unsigned long native_get_debugreg(int regno)
......
......@@ -80,8 +80,9 @@
#define PTRACE_SINGLEBLOCK 33 /* resume execution until next branch */
#ifndef __ASSEMBLY__
#ifdef CONFIG_X86_PTRACE_BTS
#ifndef __ASSEMBLY__
#include <asm/types.h>
/* configuration/status structure used in PTRACE_BTS_CONFIG and
......@@ -97,20 +98,20 @@ struct ptrace_bts_config {
/* actual size of bts_struct in bytes */
__u32 bts_size;
};
#endif
#endif /* __ASSEMBLY__ */
#define PTRACE_BTS_O_TRACE 0x1 /* branch trace */
#define PTRACE_BTS_O_SCHED 0x2 /* scheduling events w/ jiffies */
#define PTRACE_BTS_O_SIGNAL 0x4 /* send SIG<signal> on buffer overflow
instead of wrapping around */
#define PTRACE_BTS_O_CUT_SIZE 0x8 /* cut requested size to max available
instead of failing */
#define PTRACE_BTS_O_ALLOC 0x8 /* (re)allocate buffer */
#define PTRACE_BTS_CONFIG 40
/* Configure branch trace recording.
ADDR points to a struct ptrace_bts_config.
DATA gives the size of that buffer.
A new buffer is allocated, iff the size changes.
A new buffer is allocated, if requested in the flags.
An overflow signal may only be requested for new buffers.
Returns the number of bytes read.
*/
#define PTRACE_BTS_STATUS 41
......@@ -119,7 +120,7 @@ struct ptrace_bts_config {
Returns the number of bytes written.
*/
#define PTRACE_BTS_SIZE 42
/* Return the number of available BTS records.
/* Return the number of available BTS records for draining.
DATA and ADDR are ignored.
*/
#define PTRACE_BTS_GET 43
......@@ -139,5 +140,6 @@ struct ptrace_bts_config {
BTS records are read from oldest to newest.
Returns number of BTS records drained.
*/
#endif /* CONFIG_X86_PTRACE_BTS */
#endif /* ASM_X86__PTRACE_ABI_H */
......@@ -127,14 +127,48 @@ struct pt_regs {
#endif /* __KERNEL__ */
#endif /* !__i386__ */
#ifdef CONFIG_X86_PTRACE_BTS
/* a branch trace record entry
*
* In order to unify the interface between various processor versions,
* we use the below data structure for all processors.
*/
enum bts_qualifier {
BTS_INVALID = 0,
BTS_BRANCH,
BTS_TASK_ARRIVES,
BTS_TASK_DEPARTS
};
struct bts_struct {
__u64 qualifier;
union {
/* BTS_BRANCH */
struct {
__u64 from_ip;
__u64 to_ip;
} lbr;
/* BTS_TASK_ARRIVES or
BTS_TASK_DEPARTS */
__u64 jiffies;
} variant;
};
#endif /* CONFIG_X86_PTRACE_BTS */
#ifdef __KERNEL__
/* the DS BTS struct is used for ptrace as well */
#include <asm/ds.h>
#include <linux/init.h>
struct cpuinfo_x86;
struct task_struct;
#ifdef CONFIG_X86_PTRACE_BTS
extern void __cpuinit ptrace_bts_init_intel(struct cpuinfo_x86 *);
extern void ptrace_bts_take_timestamp(struct task_struct *, enum bts_qualifier);
#else
#define ptrace_bts_init_intel(config) do {} while (0)
#endif /* CONFIG_X86_PTRACE_BTS */
extern unsigned long profile_pc(struct pt_regs *regs);
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册