2009-05-26 17:10:09 +08:00
|
|
|
|
|
|
|
#include "../perf.h"
|
|
|
|
#include "util.h"
|
|
|
|
#include "parse-options.h"
|
|
|
|
#include "parse-events.h"
|
|
|
|
#include "exec_cmd.h"
|
2009-06-02 04:50:19 +08:00
|
|
|
#include "string.h"
|
2009-07-22 02:16:29 +08:00
|
|
|
#include "cache.h"
|
2009-05-26 17:10:09 +08:00
|
|
|
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
extern char *strcasestr(const char *haystack, const char *needle);
|
|
|
|
|
2009-06-06 15:58:57 +08:00
|
|
|
int nr_counters;
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-06-06 15:58:57 +08:00
|
|
|
struct perf_counter_attr attrs[MAX_COUNTERS];
|
2009-05-26 17:10:09 +08:00
|
|
|
|
|
|
|
struct event_symbol {
|
perf_counter tools: Define and use our own u64, s64 etc. definitions
On 64-bit powerpc, __u64 is defined to be unsigned long rather than
unsigned long long. This causes compiler warnings every time we
print a __u64 value with %Lx.
Rather than changing __u64, we define our own u64 to be unsigned long
long on all architectures, and similarly s64 as signed long long.
For consistency we also define u32, s32, u16, s16, u8 and s8. These
definitions are put in a new header, types.h, because these definitions
are needed in util/string.h and util/symbol.h.
The main change here is the mechanical change of __[us]{64,32,16,8}
to remove the "__". The other changes are:
* Create types.h
* Include types.h in perf.h, util/string.h and util/symbol.h
* Add types.h to the LIB_H definition in Makefile
* Added (u64) casts in process_overflow_event() and print_sym_table()
to kill two remaining warnings.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: benh@kernel.crashing.org
LKML-Reference: <19003.33494.495844.956580@cargo.ozlabs.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-19 20:21:42 +08:00
|
|
|
u8 type;
|
|
|
|
u64 config;
|
2009-06-06 15:58:57 +08:00
|
|
|
char *symbol;
|
2009-06-22 19:14:28 +08:00
|
|
|
char *alias;
|
2009-05-26 17:10:09 +08:00
|
|
|
};
|
|
|
|
|
2009-07-22 02:16:29 +08:00
|
|
|
char debugfs_path[MAXPATHLEN];
|
|
|
|
|
2009-06-22 19:13:14 +08:00
|
|
|
#define CHW(x) .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_##x
|
|
|
|
#define CSW(x) .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_##x
|
2009-06-06 15:58:57 +08:00
|
|
|
|
2009-05-26 17:10:09 +08:00
|
|
|
static struct event_symbol event_symbols[] = {
|
2009-06-22 19:14:28 +08:00
|
|
|
{ CHW(CPU_CYCLES), "cpu-cycles", "cycles" },
|
|
|
|
{ CHW(INSTRUCTIONS), "instructions", "" },
|
|
|
|
{ CHW(CACHE_REFERENCES), "cache-references", "" },
|
|
|
|
{ CHW(CACHE_MISSES), "cache-misses", "" },
|
|
|
|
{ CHW(BRANCH_INSTRUCTIONS), "branch-instructions", "branches" },
|
|
|
|
{ CHW(BRANCH_MISSES), "branch-misses", "" },
|
|
|
|
{ CHW(BUS_CYCLES), "bus-cycles", "" },
|
|
|
|
|
|
|
|
{ CSW(CPU_CLOCK), "cpu-clock", "" },
|
|
|
|
{ CSW(TASK_CLOCK), "task-clock", "" },
|
2009-06-22 23:17:26 +08:00
|
|
|
{ CSW(PAGE_FAULTS), "page-faults", "faults" },
|
2009-06-22 19:14:28 +08:00
|
|
|
{ CSW(PAGE_FAULTS_MIN), "minor-faults", "" },
|
|
|
|
{ CSW(PAGE_FAULTS_MAJ), "major-faults", "" },
|
|
|
|
{ CSW(CONTEXT_SWITCHES), "context-switches", "cs" },
|
|
|
|
{ CSW(CPU_MIGRATIONS), "cpu-migrations", "migrations" },
|
2009-05-26 17:10:09 +08:00
|
|
|
};
|
|
|
|
|
2009-05-26 15:17:18 +08:00
|
|
|
#define __PERF_COUNTER_FIELD(config, name) \
|
|
|
|
((config & PERF_COUNTER_##name##_MASK) >> PERF_COUNTER_##name##_SHIFT)
|
|
|
|
|
|
|
|
#define PERF_COUNTER_RAW(config) __PERF_COUNTER_FIELD(config, RAW)
|
|
|
|
#define PERF_COUNTER_CONFIG(config) __PERF_COUNTER_FIELD(config, CONFIG)
|
|
|
|
#define PERF_COUNTER_TYPE(config) __PERF_COUNTER_FIELD(config, TYPE)
|
|
|
|
#define PERF_COUNTER_ID(config) __PERF_COUNTER_FIELD(config, EVENT)
|
|
|
|
|
|
|
|
static char *hw_event_names[] = {
|
2009-06-06 19:58:12 +08:00
|
|
|
"cycles",
|
2009-05-26 15:17:18 +08:00
|
|
|
"instructions",
|
2009-06-06 19:58:12 +08:00
|
|
|
"cache-references",
|
|
|
|
"cache-misses",
|
2009-05-26 15:17:18 +08:00
|
|
|
"branches",
|
2009-06-06 19:58:12 +08:00
|
|
|
"branch-misses",
|
|
|
|
"bus-cycles",
|
2009-05-26 15:17:18 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static char *sw_event_names[] = {
|
2009-06-13 19:35:00 +08:00
|
|
|
"cpu-clock-msecs",
|
|
|
|
"task-clock-msecs",
|
2009-06-06 19:58:12 +08:00
|
|
|
"page-faults",
|
|
|
|
"context-switches",
|
|
|
|
"CPU-migrations",
|
|
|
|
"minor-faults",
|
|
|
|
"major-faults",
|
2009-05-26 15:17:18 +08:00
|
|
|
};
|
|
|
|
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
#define MAX_ALIASES 8
|
|
|
|
|
2009-06-22 23:17:26 +08:00
|
|
|
static char *hw_cache[][MAX_ALIASES] = {
|
2009-07-06 20:01:31 +08:00
|
|
|
{ "L1-dcache", "l1-d", "l1d", "L1-data", },
|
|
|
|
{ "L1-icache", "l1-i", "l1i", "L1-instruction", },
|
2009-06-25 20:55:22 +08:00
|
|
|
{ "LLC", "L2" },
|
|
|
|
{ "dTLB", "d-tlb", "Data-TLB", },
|
|
|
|
{ "iTLB", "i-tlb", "Instruction-TLB", },
|
|
|
|
{ "branch", "branches", "bpu", "btb", "bpc", },
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
};
|
|
|
|
|
2009-06-22 23:17:26 +08:00
|
|
|
static char *hw_cache_op[][MAX_ALIASES] = {
|
2009-06-25 20:55:22 +08:00
|
|
|
{ "load", "loads", "read", },
|
|
|
|
{ "store", "stores", "write", },
|
|
|
|
{ "prefetch", "prefetches", "speculative-read", "speculative-load", },
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
};
|
|
|
|
|
2009-06-22 23:17:26 +08:00
|
|
|
static char *hw_cache_result[][MAX_ALIASES] = {
|
2009-06-25 20:55:22 +08:00
|
|
|
{ "refs", "Reference", "ops", "access", },
|
|
|
|
{ "misses", "miss", },
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
};
|
|
|
|
|
2009-06-25 19:46:07 +08:00
|
|
|
#define C(x) PERF_COUNT_HW_CACHE_##x
|
|
|
|
#define CACHE_READ (1 << C(OP_READ))
|
|
|
|
#define CACHE_WRITE (1 << C(OP_WRITE))
|
|
|
|
#define CACHE_PREFETCH (1 << C(OP_PREFETCH))
|
|
|
|
#define COP(x) (1 << x)
|
|
|
|
|
|
|
|
/*
|
|
|
|
* cache operartion stat
|
|
|
|
* L1I : Read and prefetch only
|
|
|
|
* ITLB and BPU : Read-only
|
|
|
|
*/
|
|
|
|
static unsigned long hw_cache_stat[C(MAX)] = {
|
|
|
|
[C(L1D)] = (CACHE_READ | CACHE_WRITE | CACHE_PREFETCH),
|
|
|
|
[C(L1I)] = (CACHE_READ | CACHE_PREFETCH),
|
|
|
|
[C(LL)] = (CACHE_READ | CACHE_WRITE | CACHE_PREFETCH),
|
|
|
|
[C(DTLB)] = (CACHE_READ | CACHE_WRITE | CACHE_PREFETCH),
|
|
|
|
[C(ITLB)] = (CACHE_READ),
|
|
|
|
[C(BPU)] = (CACHE_READ),
|
|
|
|
};
|
|
|
|
|
2009-07-22 00:20:22 +08:00
|
|
|
#define for_each_subsystem(sys_dir, sys_dirent, sys_next, file, st) \
|
|
|
|
while (!readdir_r(sys_dir, &sys_dirent, &sys_next) && sys_next) \
|
2009-07-22 02:16:29 +08:00
|
|
|
if (snprintf(file, MAXPATHLEN, "%s/%s", debugfs_path, \
|
|
|
|
sys_dirent.d_name) && \
|
2009-07-22 00:20:22 +08:00
|
|
|
(!stat(file, &st)) && (S_ISDIR(st.st_mode)) && \
|
|
|
|
(strcmp(sys_dirent.d_name, ".")) && \
|
|
|
|
(strcmp(sys_dirent.d_name, "..")))
|
|
|
|
|
|
|
|
#define for_each_event(sys_dirent, evt_dir, evt_dirent, evt_next, file, st) \
|
|
|
|
while (!readdir_r(evt_dir, &evt_dirent, &evt_next) && evt_next) \
|
2009-07-22 02:16:29 +08:00
|
|
|
if (snprintf(file, MAXPATHLEN, "%s/%s/%s", debugfs_path, \
|
|
|
|
sys_dirent.d_name, evt_dirent.d_name) && \
|
2009-07-22 00:20:22 +08:00
|
|
|
(!stat(file, &st)) && (S_ISDIR(st.st_mode)) && \
|
|
|
|
(strcmp(evt_dirent.d_name, ".")) && \
|
|
|
|
(strcmp(evt_dirent.d_name, "..")))
|
|
|
|
|
|
|
|
#define MAX_EVENT_LENGTH 30
|
|
|
|
|
2009-07-22 02:16:29 +08:00
|
|
|
int valid_debugfs_mount(const char *debugfs)
|
2009-07-22 00:20:22 +08:00
|
|
|
{
|
|
|
|
struct statfs st_fs;
|
|
|
|
|
2009-07-22 02:16:29 +08:00
|
|
|
if (statfs(debugfs, &st_fs) < 0)
|
2009-07-22 00:20:22 +08:00
|
|
|
return -ENOENT;
|
|
|
|
else if (st_fs.f_type != (long) DEBUGFS_MAGIC)
|
|
|
|
return -ENOENT;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static char *tracepoint_id_to_name(u64 config)
|
|
|
|
{
|
|
|
|
static char tracepoint_name[2 * MAX_EVENT_LENGTH];
|
|
|
|
DIR *sys_dir, *evt_dir;
|
|
|
|
struct dirent *sys_next, *evt_next, sys_dirent, evt_dirent;
|
|
|
|
struct stat st;
|
|
|
|
char id_buf[4];
|
|
|
|
int fd;
|
|
|
|
u64 id;
|
|
|
|
char evt_path[MAXPATHLEN];
|
|
|
|
|
2009-07-22 02:16:29 +08:00
|
|
|
if (valid_debugfs_mount(debugfs_path))
|
2009-07-22 00:20:22 +08:00
|
|
|
return "unkown";
|
|
|
|
|
2009-07-22 02:16:29 +08:00
|
|
|
sys_dir = opendir(debugfs_path);
|
2009-07-22 00:20:22 +08:00
|
|
|
if (!sys_dir)
|
|
|
|
goto cleanup;
|
|
|
|
|
|
|
|
for_each_subsystem(sys_dir, sys_dirent, sys_next, evt_path, st) {
|
|
|
|
evt_dir = opendir(evt_path);
|
|
|
|
if (!evt_dir)
|
|
|
|
goto cleanup;
|
|
|
|
for_each_event(sys_dirent, evt_dir, evt_dirent, evt_next,
|
|
|
|
evt_path, st) {
|
|
|
|
snprintf(evt_path, MAXPATHLEN, "%s/%s/%s/id",
|
2009-07-22 02:16:29 +08:00
|
|
|
debugfs_path, sys_dirent.d_name,
|
2009-07-22 00:20:22 +08:00
|
|
|
evt_dirent.d_name);
|
|
|
|
fd = open(evt_path, O_RDONLY);
|
|
|
|
if (fd < 0)
|
|
|
|
continue;
|
|
|
|
if (read(fd, id_buf, sizeof(id_buf)) < 0) {
|
|
|
|
close(fd);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
close(fd);
|
|
|
|
id = atoll(id_buf);
|
|
|
|
if (id == config) {
|
|
|
|
closedir(evt_dir);
|
|
|
|
closedir(sys_dir);
|
|
|
|
snprintf(tracepoint_name, 2 * MAX_EVENT_LENGTH,
|
|
|
|
"%s:%s", sys_dirent.d_name,
|
|
|
|
evt_dirent.d_name);
|
|
|
|
return tracepoint_name;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
closedir(evt_dir);
|
|
|
|
}
|
|
|
|
|
|
|
|
cleanup:
|
|
|
|
closedir(sys_dir);
|
|
|
|
return "unkown";
|
|
|
|
}
|
|
|
|
|
2009-06-25 19:46:07 +08:00
|
|
|
static int is_cache_op_valid(u8 cache_type, u8 cache_op)
|
|
|
|
{
|
|
|
|
if (hw_cache_stat[cache_type] & COP(cache_op))
|
|
|
|
return 1; /* valid */
|
|
|
|
else
|
|
|
|
return 0; /* invalid */
|
|
|
|
}
|
|
|
|
|
2009-06-25 20:55:22 +08:00
|
|
|
static char *event_cache_name(u8 cache_type, u8 cache_op, u8 cache_result)
|
|
|
|
{
|
|
|
|
static char name[50];
|
|
|
|
|
|
|
|
if (cache_result) {
|
|
|
|
sprintf(name, "%s-%s-%s", hw_cache[cache_type][0],
|
|
|
|
hw_cache_op[cache_op][0],
|
|
|
|
hw_cache_result[cache_result][0]);
|
|
|
|
} else {
|
|
|
|
sprintf(name, "%s-%s", hw_cache[cache_type][0],
|
|
|
|
hw_cache_op[cache_op][1]);
|
|
|
|
}
|
|
|
|
|
|
|
|
return name;
|
|
|
|
}
|
|
|
|
|
2009-06-06 15:58:57 +08:00
|
|
|
char *event_name(int counter)
|
2009-05-26 15:17:18 +08:00
|
|
|
{
|
perf_counter tools: Define and use our own u64, s64 etc. definitions
On 64-bit powerpc, __u64 is defined to be unsigned long rather than
unsigned long long. This causes compiler warnings every time we
print a __u64 value with %Lx.
Rather than changing __u64, we define our own u64 to be unsigned long
long on all architectures, and similarly s64 as signed long long.
For consistency we also define u32, s32, u16, s16, u8 and s8. These
definitions are put in a new header, types.h, because these definitions
are needed in util/string.h and util/symbol.h.
The main change here is the mechanical change of __[us]{64,32,16,8}
to remove the "__". The other changes are:
* Create types.h
* Include types.h in perf.h, util/string.h and util/symbol.h
* Add types.h to the LIB_H definition in Makefile
* Added (u64) casts in process_overflow_event() and print_sym_table()
to kill two remaining warnings.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: benh@kernel.crashing.org
LKML-Reference: <19003.33494.495844.956580@cargo.ozlabs.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-19 20:21:42 +08:00
|
|
|
u64 config = attrs[counter].config;
|
2009-06-06 15:58:57 +08:00
|
|
|
int type = attrs[counter].type;
|
2009-05-26 15:17:18 +08:00
|
|
|
static char buf[32];
|
|
|
|
|
2009-06-06 15:58:57 +08:00
|
|
|
if (attrs[counter].type == PERF_TYPE_RAW) {
|
|
|
|
sprintf(buf, "raw 0x%llx", config);
|
2009-05-26 15:17:18 +08:00
|
|
|
return buf;
|
|
|
|
}
|
|
|
|
|
|
|
|
switch (type) {
|
|
|
|
case PERF_TYPE_HARDWARE:
|
2009-06-11 20:06:28 +08:00
|
|
|
if (config < PERF_COUNT_HW_MAX)
|
2009-06-06 15:58:57 +08:00
|
|
|
return hw_event_names[config];
|
2009-05-26 15:17:18 +08:00
|
|
|
return "unknown-hardware";
|
|
|
|
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
case PERF_TYPE_HW_CACHE: {
|
perf_counter tools: Define and use our own u64, s64 etc. definitions
On 64-bit powerpc, __u64 is defined to be unsigned long rather than
unsigned long long. This causes compiler warnings every time we
print a __u64 value with %Lx.
Rather than changing __u64, we define our own u64 to be unsigned long
long on all architectures, and similarly s64 as signed long long.
For consistency we also define u32, s32, u16, s16, u8 and s8. These
definitions are put in a new header, types.h, because these definitions
are needed in util/string.h and util/symbol.h.
The main change here is the mechanical change of __[us]{64,32,16,8}
to remove the "__". The other changes are:
* Create types.h
* Include types.h in perf.h, util/string.h and util/symbol.h
* Add types.h to the LIB_H definition in Makefile
* Added (u64) casts in process_overflow_event() and print_sym_table()
to kill two remaining warnings.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: benh@kernel.crashing.org
LKML-Reference: <19003.33494.495844.956580@cargo.ozlabs.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-19 20:21:42 +08:00
|
|
|
u8 cache_type, cache_op, cache_result;
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
|
|
|
|
cache_type = (config >> 0) & 0xff;
|
|
|
|
if (cache_type > PERF_COUNT_HW_CACHE_MAX)
|
|
|
|
return "unknown-ext-hardware-cache-type";
|
|
|
|
|
|
|
|
cache_op = (config >> 8) & 0xff;
|
2009-06-06 19:58:12 +08:00
|
|
|
if (cache_op > PERF_COUNT_HW_CACHE_OP_MAX)
|
|
|
|
return "unknown-ext-hardware-cache-op";
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
|
|
|
|
cache_result = (config >> 16) & 0xff;
|
2009-06-06 19:58:12 +08:00
|
|
|
if (cache_result > PERF_COUNT_HW_CACHE_RESULT_MAX)
|
|
|
|
return "unknown-ext-hardware-cache-result";
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
|
2009-06-25 19:46:07 +08:00
|
|
|
if (!is_cache_op_valid(cache_type, cache_op))
|
|
|
|
return "invalid-cache";
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
|
2009-06-25 20:55:22 +08:00
|
|
|
return event_cache_name(cache_type, cache_op, cache_result);
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
}
|
|
|
|
|
2009-05-26 15:17:18 +08:00
|
|
|
case PERF_TYPE_SOFTWARE:
|
2009-06-11 20:06:28 +08:00
|
|
|
if (config < PERF_COUNT_SW_MAX)
|
2009-06-06 15:58:57 +08:00
|
|
|
return sw_event_names[config];
|
2009-05-26 15:17:18 +08:00
|
|
|
return "unknown-software";
|
|
|
|
|
2009-07-22 00:20:22 +08:00
|
|
|
case PERF_TYPE_TRACEPOINT:
|
|
|
|
return tracepoint_id_to_name(config);
|
|
|
|
|
2009-05-26 15:17:18 +08:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return "unknown";
|
|
|
|
}
|
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
static int parse_aliases(const char **str, char *names[][MAX_ALIASES], int size)
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
{
|
|
|
|
int i, j;
|
2009-07-01 11:04:34 +08:00
|
|
|
int n, longest = -1;
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
|
|
|
|
for (i = 0; i < size; i++) {
|
2009-07-01 11:04:34 +08:00
|
|
|
for (j = 0; j < MAX_ALIASES && names[i][j]; j++) {
|
|
|
|
n = strlen(names[i][j]);
|
|
|
|
if (n > longest && !strncasecmp(*str, names[i][j], n))
|
|
|
|
longest = n;
|
|
|
|
}
|
|
|
|
if (longest > 0) {
|
|
|
|
*str += longest;
|
|
|
|
return i;
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-06-07 03:04:17 +08:00
|
|
|
return -1;
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
}
|
|
|
|
|
2009-06-22 23:17:26 +08:00
|
|
|
static int
|
2009-07-01 11:04:34 +08:00
|
|
|
parse_generic_hw_event(const char **str, struct perf_counter_attr *attr)
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
{
|
2009-07-01 11:04:34 +08:00
|
|
|
const char *s = *str;
|
|
|
|
int cache_type = -1, cache_op = -1, cache_result = -1;
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
cache_type = parse_aliases(&s, hw_cache, PERF_COUNT_HW_CACHE_MAX);
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
/*
|
|
|
|
* No fallback - if we cannot get a clear cache type
|
|
|
|
* then bail out:
|
|
|
|
*/
|
|
|
|
if (cache_type == -1)
|
2009-07-01 11:04:34 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
while ((cache_op == -1 || cache_result == -1) && *s == '-') {
|
|
|
|
++s;
|
|
|
|
|
|
|
|
if (cache_op == -1) {
|
|
|
|
cache_op = parse_aliases(&s, hw_cache_op,
|
|
|
|
PERF_COUNT_HW_CACHE_OP_MAX);
|
|
|
|
if (cache_op >= 0) {
|
|
|
|
if (!is_cache_op_valid(cache_type, cache_op))
|
|
|
|
return 0;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (cache_result == -1) {
|
|
|
|
cache_result = parse_aliases(&s, hw_cache_result,
|
|
|
|
PERF_COUNT_HW_CACHE_RESULT_MAX);
|
|
|
|
if (cache_result >= 0)
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Can't parse this as a cache op or result, so back up
|
|
|
|
* to the '-'.
|
|
|
|
*/
|
|
|
|
--s;
|
|
|
|
break;
|
|
|
|
}
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Fall back to reads:
|
|
|
|
*/
|
2009-06-07 03:04:17 +08:00
|
|
|
if (cache_op == -1)
|
|
|
|
cache_op = PERF_COUNT_HW_CACHE_OP_READ;
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Fall back to accesses:
|
|
|
|
*/
|
|
|
|
if (cache_result == -1)
|
|
|
|
cache_result = PERF_COUNT_HW_CACHE_RESULT_ACCESS;
|
|
|
|
|
|
|
|
attr->config = cache_type | (cache_op << 8) | (cache_result << 16);
|
|
|
|
attr->type = PERF_TYPE_HW_CACHE;
|
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
*str = s;
|
|
|
|
return 1;
|
perf_counter: Implement generalized cache event types
Extend generic event enumeration with the PERF_TYPE_HW_CACHE
method.
This is a 3-dimensional space:
{ L1-D, L1-I, L2, ITLB, DTLB, BPU } x
{ load, store, prefetch } x
{ accesses, misses }
User-space passes in the 3 coordinates and the kernel provides
a counter. (if the hardware supports that type and if the
combination makes sense.)
Combinations that make no sense produce a -EINVAL.
Combinations that are not supported by the hardware produce -ENOTSUP.
Extend the tools to deal with this, and rewrite the event symbol
parsing code with various popular aliases for the units and
access methods above. So 'l1-cache-miss' and 'l1d-read-ops' are
both valid aliases.
( x86 is supported for now, with the Nehalem event table filled in,
and with Core2 and Atom having placeholder tables. )
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-06 02:22:46 +08:00
|
|
|
}
|
|
|
|
|
2009-07-22 00:20:22 +08:00
|
|
|
static int parse_tracepoint_event(const char **strp,
|
|
|
|
struct perf_counter_attr *attr)
|
|
|
|
{
|
|
|
|
const char *evt_name;
|
|
|
|
char sys_name[MAX_EVENT_LENGTH];
|
|
|
|
char id_buf[4];
|
|
|
|
int fd;
|
|
|
|
unsigned int sys_length, evt_length;
|
|
|
|
u64 id;
|
|
|
|
char evt_path[MAXPATHLEN];
|
|
|
|
|
2009-07-22 02:16:29 +08:00
|
|
|
if (valid_debugfs_mount(debugfs_path))
|
2009-07-22 00:20:22 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
evt_name = strchr(*strp, ':');
|
|
|
|
if (!evt_name)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
sys_length = evt_name - *strp;
|
|
|
|
if (sys_length >= MAX_EVENT_LENGTH)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
strncpy(sys_name, *strp, sys_length);
|
|
|
|
sys_name[sys_length] = '\0';
|
|
|
|
evt_name = evt_name + 1;
|
|
|
|
evt_length = strlen(evt_name);
|
|
|
|
if (evt_length >= MAX_EVENT_LENGTH)
|
|
|
|
return 0;
|
|
|
|
|
2009-07-22 02:16:29 +08:00
|
|
|
snprintf(evt_path, MAXPATHLEN, "%s/%s/%s/id", debugfs_path,
|
|
|
|
sys_name, evt_name);
|
2009-07-22 00:20:22 +08:00
|
|
|
fd = open(evt_path, O_RDONLY);
|
|
|
|
if (fd < 0)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (read(fd, id_buf, sizeof(id_buf)) < 0) {
|
|
|
|
close(fd);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
close(fd);
|
|
|
|
id = atoll(id_buf);
|
|
|
|
attr->config = id;
|
|
|
|
attr->type = PERF_TYPE_TRACEPOINT;
|
|
|
|
*strp = evt_name + evt_length;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2009-06-22 19:14:28 +08:00
|
|
|
static int check_events(const char *str, unsigned int i)
|
|
|
|
{
|
2009-07-01 11:04:34 +08:00
|
|
|
int n;
|
2009-06-22 19:14:28 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
n = strlen(event_symbols[i].symbol);
|
|
|
|
if (!strncmp(str, event_symbols[i].symbol, n))
|
|
|
|
return n;
|
|
|
|
|
|
|
|
n = strlen(event_symbols[i].alias);
|
|
|
|
if (n)
|
|
|
|
if (!strncmp(str, event_symbols[i].alias, n))
|
|
|
|
return n;
|
2009-06-22 19:14:28 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
static int
|
|
|
|
parse_symbolic_event(const char **strp, struct perf_counter_attr *attr)
|
2009-05-26 17:10:09 +08:00
|
|
|
{
|
2009-07-01 11:04:34 +08:00
|
|
|
const char *str = *strp;
|
2009-05-26 17:10:09 +08:00
|
|
|
unsigned int i;
|
2009-07-01 11:04:34 +08:00
|
|
|
int n;
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
for (i = 0; i < ARRAY_SIZE(event_symbols); i++) {
|
|
|
|
n = check_events(str, i);
|
|
|
|
if (n > 0) {
|
|
|
|
attr->type = event_symbols[i].type;
|
|
|
|
attr->config = event_symbols[i].config;
|
|
|
|
*strp = str + n;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int parse_raw_event(const char **strp, struct perf_counter_attr *attr)
|
|
|
|
{
|
|
|
|
const char *str = *strp;
|
|
|
|
u64 config;
|
|
|
|
int n;
|
2009-06-06 15:58:57 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
if (*str != 'r')
|
2009-06-06 15:58:57 +08:00
|
|
|
return 0;
|
2009-07-01 11:04:34 +08:00
|
|
|
n = hex2u64(str + 1, &config);
|
|
|
|
if (n > 0) {
|
|
|
|
*strp = str + n + 1;
|
|
|
|
attr->type = PERF_TYPE_RAW;
|
|
|
|
attr->config = config;
|
|
|
|
return 1;
|
2009-06-06 15:58:57 +08:00
|
|
|
}
|
2009-07-01 11:04:34 +08:00
|
|
|
return 0;
|
|
|
|
}
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
static int
|
|
|
|
parse_numeric_event(const char **strp, struct perf_counter_attr *attr)
|
|
|
|
{
|
|
|
|
const char *str = *strp;
|
|
|
|
char *endp;
|
|
|
|
unsigned long type;
|
|
|
|
u64 config;
|
|
|
|
|
|
|
|
type = strtoul(str, &endp, 0);
|
|
|
|
if (endp > str && type < PERF_TYPE_MAX && *endp == ':') {
|
|
|
|
str = endp + 1;
|
|
|
|
config = strtoul(str, &endp, 0);
|
|
|
|
if (endp > str) {
|
|
|
|
attr->type = type;
|
|
|
|
attr->config = config;
|
|
|
|
*strp = endp;
|
|
|
|
return 1;
|
2009-06-02 04:50:19 +08:00
|
|
|
}
|
2009-07-01 11:04:34 +08:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
parse_event_modifier(const char **strp, struct perf_counter_attr *attr)
|
|
|
|
{
|
|
|
|
const char *str = *strp;
|
|
|
|
int eu = 1, ek = 1, eh = 1;
|
2009-06-06 15:58:57 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
if (*str++ != ':')
|
2009-06-06 15:58:57 +08:00
|
|
|
return 0;
|
2009-07-01 11:04:34 +08:00
|
|
|
while (*str) {
|
|
|
|
if (*str == 'u')
|
|
|
|
eu = 0;
|
|
|
|
else if (*str == 'k')
|
|
|
|
ek = 0;
|
|
|
|
else if (*str == 'h')
|
|
|
|
eh = 0;
|
|
|
|
else
|
|
|
|
break;
|
|
|
|
++str;
|
2009-05-26 15:17:18 +08:00
|
|
|
}
|
2009-07-01 11:04:34 +08:00
|
|
|
if (str >= *strp + 2) {
|
|
|
|
*strp = str;
|
|
|
|
attr->exclude_user = eu;
|
|
|
|
attr->exclude_kernel = ek;
|
|
|
|
attr->exclude_hv = eh;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
/*
|
|
|
|
* Each event can have multiple symbolic names.
|
|
|
|
* Symbolic names are (almost) exactly matched.
|
|
|
|
*/
|
|
|
|
static int parse_event_symbols(const char **str, struct perf_counter_attr *attr)
|
|
|
|
{
|
2009-07-22 00:20:22 +08:00
|
|
|
if (!(parse_tracepoint_event(str, attr) ||
|
|
|
|
parse_raw_event(str, attr) ||
|
2009-07-01 11:04:34 +08:00
|
|
|
parse_numeric_event(str, attr) ||
|
|
|
|
parse_symbolic_event(str, attr) ||
|
|
|
|
parse_generic_hw_event(str, attr)))
|
|
|
|
return 0;
|
2009-06-06 15:58:57 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
parse_event_modifier(str, attr);
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
return 1;
|
2009-05-26 17:10:09 +08:00
|
|
|
}
|
|
|
|
|
2009-07-01 18:37:06 +08:00
|
|
|
int parse_events(const struct option *opt __used, const char *str, int unset __used)
|
2009-05-26 17:10:09 +08:00
|
|
|
{
|
2009-06-06 15:58:57 +08:00
|
|
|
struct perf_counter_attr attr;
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
for (;;) {
|
|
|
|
if (nr_counters == MAX_COUNTERS)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
memset(&attr, 0, sizeof(attr));
|
|
|
|
if (!parse_event_symbols(&str, &attr))
|
|
|
|
return -1;
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
if (!(*str == 0 || *str == ',' || isspace(*str)))
|
|
|
|
return -1;
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
attrs[nr_counters] = attr;
|
|
|
|
nr_counters++;
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-07-01 11:04:34 +08:00
|
|
|
if (*str == 0)
|
|
|
|
break;
|
|
|
|
if (*str == ',')
|
|
|
|
++str;
|
|
|
|
while (isspace(*str))
|
|
|
|
++str;
|
2009-05-26 17:10:09 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-06-06 18:24:17 +08:00
|
|
|
static const char * const event_type_descriptors[] = {
|
|
|
|
"",
|
|
|
|
"Hardware event",
|
|
|
|
"Software event",
|
|
|
|
"Tracepoint event",
|
|
|
|
"Hardware cache event",
|
|
|
|
};
|
|
|
|
|
2009-07-22 00:20:22 +08:00
|
|
|
/*
|
|
|
|
* Print the events from <debugfs_mount_point>/tracing/events
|
|
|
|
*/
|
|
|
|
|
|
|
|
static void print_tracepoint_events(void)
|
|
|
|
{
|
|
|
|
DIR *sys_dir, *evt_dir;
|
|
|
|
struct dirent *sys_next, *evt_next, sys_dirent, evt_dirent;
|
|
|
|
struct stat st;
|
|
|
|
char evt_path[MAXPATHLEN];
|
|
|
|
|
2009-07-22 02:16:29 +08:00
|
|
|
if (valid_debugfs_mount(debugfs_path))
|
2009-07-22 00:20:22 +08:00
|
|
|
return;
|
|
|
|
|
2009-07-22 02:16:29 +08:00
|
|
|
sys_dir = opendir(debugfs_path);
|
2009-07-22 00:20:22 +08:00
|
|
|
if (!sys_dir)
|
|
|
|
goto cleanup;
|
|
|
|
|
|
|
|
for_each_subsystem(sys_dir, sys_dirent, sys_next, evt_path, st) {
|
|
|
|
evt_dir = opendir(evt_path);
|
|
|
|
if (!evt_dir)
|
|
|
|
goto cleanup;
|
|
|
|
for_each_event(sys_dirent, evt_dir, evt_dirent, evt_next,
|
|
|
|
evt_path, st) {
|
|
|
|
snprintf(evt_path, MAXPATHLEN, "%s:%s",
|
|
|
|
sys_dirent.d_name, evt_dirent.d_name);
|
|
|
|
fprintf(stderr, " %-40s [%s]\n", evt_path,
|
|
|
|
event_type_descriptors[PERF_TYPE_TRACEPOINT+1]);
|
|
|
|
}
|
|
|
|
closedir(evt_dir);
|
|
|
|
}
|
|
|
|
|
|
|
|
cleanup:
|
|
|
|
closedir(sys_dir);
|
|
|
|
}
|
|
|
|
|
2009-05-26 17:10:09 +08:00
|
|
|
/*
|
2009-06-06 18:24:17 +08:00
|
|
|
* Print the help text for the event symbols:
|
2009-05-26 17:10:09 +08:00
|
|
|
*/
|
2009-06-06 18:24:17 +08:00
|
|
|
void print_events(void)
|
2009-05-26 17:10:09 +08:00
|
|
|
{
|
2009-06-06 18:24:17 +08:00
|
|
|
struct event_symbol *syms = event_symbols;
|
2009-07-01 21:06:18 +08:00
|
|
|
unsigned int i, type, op, prev_type = -1;
|
2009-06-22 19:14:28 +08:00
|
|
|
char name[40];
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-06-06 18:24:17 +08:00
|
|
|
fprintf(stderr, "\n");
|
|
|
|
fprintf(stderr, "List of pre-defined events (to be used in -e):\n");
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-06-06 18:24:17 +08:00
|
|
|
for (i = 0; i < ARRAY_SIZE(event_symbols); i++, syms++) {
|
|
|
|
type = syms->type + 1;
|
2009-07-13 08:25:47 +08:00
|
|
|
if (type >= ARRAY_SIZE(event_type_descriptors))
|
2009-06-06 18:24:17 +08:00
|
|
|
type = 0;
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-06-06 18:24:17 +08:00
|
|
|
if (type != prev_type)
|
|
|
|
fprintf(stderr, "\n");
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-06-22 19:14:28 +08:00
|
|
|
if (strlen(syms->alias))
|
|
|
|
sprintf(name, "%s OR %s", syms->symbol, syms->alias);
|
|
|
|
else
|
|
|
|
strcpy(name, syms->symbol);
|
|
|
|
fprintf(stderr, " %-40s [%s]\n", name,
|
2009-06-06 18:24:17 +08:00
|
|
|
event_type_descriptors[type]);
|
2009-05-26 17:10:09 +08:00
|
|
|
|
2009-06-06 18:24:17 +08:00
|
|
|
prev_type = type;
|
2009-05-26 17:10:09 +08:00
|
|
|
}
|
|
|
|
|
2009-07-01 21:06:18 +08:00
|
|
|
fprintf(stderr, "\n");
|
|
|
|
for (type = 0; type < PERF_COUNT_HW_CACHE_MAX; type++) {
|
|
|
|
for (op = 0; op < PERF_COUNT_HW_CACHE_OP_MAX; op++) {
|
|
|
|
/* skip invalid cache type */
|
|
|
|
if (!is_cache_op_valid(type, op))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
for (i = 0; i < PERF_COUNT_HW_CACHE_RESULT_MAX; i++) {
|
|
|
|
fprintf(stderr, " %-40s [%s]\n",
|
|
|
|
event_cache_name(type, op, i),
|
|
|
|
event_type_descriptors[4]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-06-06 18:24:17 +08:00
|
|
|
fprintf(stderr, "\n");
|
2009-06-22 19:14:28 +08:00
|
|
|
fprintf(stderr, " %-40s [raw hardware event descriptor]\n",
|
2009-06-06 18:24:17 +08:00
|
|
|
"rNNN");
|
|
|
|
fprintf(stderr, "\n");
|
|
|
|
|
2009-07-22 00:20:22 +08:00
|
|
|
print_tracepoint_events();
|
|
|
|
|
2009-06-06 18:24:17 +08:00
|
|
|
exit(129);
|
2009-05-26 17:10:09 +08:00
|
|
|
}
|