2013-06-11 19:29:18 +04:00
# include <traceevent/event-parse.h>
2012-09-27 03:05:56 +04:00
# include "builtin.h"
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
# include "util/color.h"
2013-08-22 23:49:54 +04:00
# include "util/debug.h"
2012-09-27 03:05:56 +04:00
# include "util/evlist.h"
perf trace: Support 'strace' syscall event groups
I.e.:
$ cat ~/share/perf-core/strace/groups/file
access
chmod
creat
execve
faccessat
getcwd
lstat
mkdir
open
openat
quotactl
readlink
rename
rmdir
stat
statfs
symlink
unlink
$
Then, on a quiet desktop, try running this and then moving your mouse to
see the deluge of mouse related activity:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
# trace --ev probe:vfs_getname --filter-pids 2232 -e file
0.042 (0.042 ms): mousetweaks/2235 open(filename: 0x14e3910, mode: 438 ) ...
0.042 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/home/acme/.icons/Adwaita/cursors/xterm")
0.100 (0.100 ms): mousetweaks/2235 ... [continued]: open()) = -1 ENOENT No such file or directory
0.142 (0.018 ms): mousetweaks/2235 open(filename: 0x14c3c10, mode: 438 ) ...
0.142 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/home/acme/.icons/Adwaita/index.theme")
0.192 (0.069 ms): mousetweaks/2235 ... [continued]: open()) = -1 ENOENT No such file or directory
0.230 (0.017 ms): mousetweaks/2235 open(filename: 0x14c3c10, mode: 438 ) ...
0.230 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/usr/share/icons/Adwaita/cursors/xterm")
0.253 (0.041 ms): mousetweaks/2235 ... [continued]: open()) = 14
0.459 (0.008 ms): mousetweaks/2235 open(filename: 0x14e3910, mode: 438 ) ...
0.459 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/home/acme/.icons/Adwaita/cursors/left_side")
0.468 (0.017 ms): mousetweaks/2235 ... [continued]: open()) = -1 ENOENT No such file or directory
Need to combine that raw_syscalls:sys_enter(open) + probe:vfs_getname +
raw_syscalls:sys_exit(open) sequence...
Now, if you're bored, please write some more syscall groups, like the ones
in 'strace' and send it our way :-)
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-a42xklu59lcbxp7bbnic74a8@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-20 18:02:09 +03:00
# include "util/exec_cmd.h"
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
# include "util/machine.h"
2013-08-29 08:29:52 +04:00
# include "util/session.h"
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
# include "util/thread.h"
2012-09-27 03:05:56 +04:00
# include "util/parse-options.h"
2013-08-09 19:28:31 +04:00
# include "util/strlist.h"
2013-08-29 08:29:53 +04:00
# include "util/intlist.h"
2012-09-27 03:05:56 +04:00
# include "util/thread_map.h"
2013-10-09 07:26:53 +04:00
# include "util/stat.h"
2013-12-03 17:09:24 +04:00
# include "trace-event.h"
2013-12-05 06:41:39 +04:00
# include "util/parse-events.h"
2012-09-27 03:05:56 +04:00
# include <libaudit.h>
# include <stdlib.h>
perf trace: Add beautifier for mmap prot parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.984 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd62ae000
1.114 ( 0.016 ms): mmap(addr: 0, len: 125100, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xd628f000
1.252 ( 0.020 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: 2050, fd: 3, off: 0) = 0xc1600000
1.282 ( 0.024 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.315 ( 0.026 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: 2066, fd: 3, off: 1757184) = 0xc19ad000
1.352 ( 0.017 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: 50, fd: 4294967295, off: 0) = 0xc19b3000
1.415 ( 0.011 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628e000
1.440 ( 0.011 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628c000
1.569 ( 0.019 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.591 ( 0.017 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.616 ( 0.016 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
2.105 ( 0.018 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xcfe9c000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-21 00:44:42 +04:00
# include <sys/mman.h>
2013-09-02 20:46:44 +04:00
# include <linux/futex.h>
2012-09-27 03:05:56 +04:00
2013-09-12 17:29:00 +04:00
/* For older distros: */
# ifndef MAP_STACK
# define MAP_STACK 0x20000
# endif
# ifndef MADV_HWPOISON
# define MADV_HWPOISON 100
# endif
# ifndef MADV_MERGEABLE
# define MADV_MERGEABLE 12
# endif
# ifndef MADV_UNMERGEABLE
# define MADV_UNMERGEABLE 13
# endif
2014-02-06 05:00:35 +04:00
# ifndef EFD_SEMAPHORE
# define EFD_SEMAPHORE 1
# endif
2015-05-14 23:39:03 +03:00
# ifndef EFD_NONBLOCK
# define EFD_NONBLOCK 00004000
# endif
# ifndef EFD_CLOEXEC
# define EFD_CLOEXEC 02000000
# endif
# ifndef O_CLOEXEC
# define O_CLOEXEC 02000000
# endif
# ifndef SOCK_DCCP
# define SOCK_DCCP 6
# endif
# ifndef SOCK_CLOEXEC
# define SOCK_CLOEXEC 02000000
# endif
# ifndef SOCK_NONBLOCK
# define SOCK_NONBLOCK 00004000
# endif
# ifndef MSG_CMSG_CLOEXEC
# define MSG_CMSG_CLOEXEC 0x40000000
# endif
trace: Beautify perf_event_open syscall
Syswide tracing and then running 'stat' and 'trace':
$ perf trace -e perf_event_open
1034.649 (0.019 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1, flags: FD_CLOEXEC) = -1 EINVAL Invalid argument
1034.670 (0.008 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
1034.681 (0.007 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
1034.692 (0.007 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
9986.983 (0.014 ms): trace/6139 perf_event_open(attr_uptr: 0x7ffd9c629320, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.026 (0.016 ms): trace/6139 perf_event_open(attr_uptr: 0x37c7e70, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.041 (0.008 ms): trace/6139 perf_event_open(attr_uptr: 0x37c7e70, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.489 (0.092 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.536 (0.044 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 1, group_fd: -1, flags: FD_CLOEXEC) = 4
9987.580 (0.041 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 2, group_fd: -1, flags: FD_CLOEXEC) = 5
9987.620 (0.037 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 3, group_fd: -1, flags: FD_CLOEXEC) = 7
9987.659 (0.035 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, group_fd: -1, flags: FD_CLOEXEC) = 8
9987.692 (0.031 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 1, group_fd: -1, flags: FD_CLOEXEC) = 9
9987.727 (0.032 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 2, group_fd: -1, flags: FD_CLOEXEC) = 10
9987.761 (0.031 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 3, group_fd: -1, flags: FD_CLOEXEC) = 11
Need to intercept perf_copy_attr() with a kprobe or with eBPF...
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-njb105hab2i3t5dexym9lskl@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-12 04:47:54 +03:00
# ifndef PERF_FLAG_FD_NO_GROUP
# define PERF_FLAG_FD_NO_GROUP (1UL << 0)
# endif
# ifndef PERF_FLAG_FD_OUTPUT
# define PERF_FLAG_FD_OUTPUT (1UL << 1)
# endif
# ifndef PERF_FLAG_PID_CGROUP
# define PERF_FLAG_PID_CGROUP (1UL << 2) /* pid=cgroup id, per-cpu mode only */
# endif
# ifndef PERF_FLAG_FD_CLOEXEC
# define PERF_FLAG_FD_CLOEXEC (1UL << 3) /* O_CLOEXEC */
# endif
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
struct tp_field {
int offset ;
union {
u64 ( * integer ) ( struct tp_field * field , struct perf_sample * sample ) ;
void * ( * pointer ) ( struct tp_field * field , struct perf_sample * sample ) ;
} ;
} ;
# define TP_UINT_FIELD(bits) \
static u64 tp_field__u # # bits ( struct tp_field * field , struct perf_sample * sample ) \
{ \
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
u # # bits value ; \
memcpy ( & value , sample - > raw_data + field - > offset , sizeof ( value ) ) ; \
return value ; \
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
}
TP_UINT_FIELD ( 8 ) ;
TP_UINT_FIELD ( 16 ) ;
TP_UINT_FIELD ( 32 ) ;
TP_UINT_FIELD ( 64 ) ;
# define TP_UINT_FIELD__SWAPPED(bits) \
static u64 tp_field__swapped_u # # bits ( struct tp_field * field , struct perf_sample * sample ) \
{ \
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
u # # bits value ; \
memcpy ( & value , sample - > raw_data + field - > offset , sizeof ( value ) ) ; \
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
return bswap_ # # bits ( value ) ; \
}
TP_UINT_FIELD__SWAPPED ( 16 ) ;
TP_UINT_FIELD__SWAPPED ( 32 ) ;
TP_UINT_FIELD__SWAPPED ( 64 ) ;
static int tp_field__init_uint ( struct tp_field * field ,
struct format_field * format_field ,
bool needs_swap )
{
field - > offset = format_field - > offset ;
switch ( format_field - > size ) {
case 1 :
field - > integer = tp_field__u8 ;
break ;
case 2 :
field - > integer = needs_swap ? tp_field__swapped_u16 : tp_field__u16 ;
break ;
case 4 :
field - > integer = needs_swap ? tp_field__swapped_u32 : tp_field__u32 ;
break ;
case 8 :
field - > integer = needs_swap ? tp_field__swapped_u64 : tp_field__u64 ;
break ;
default :
return - 1 ;
}
return 0 ;
}
static void * tp_field__ptr ( struct tp_field * field , struct perf_sample * sample )
{
return sample - > raw_data + field - > offset ;
}
static int tp_field__init_ptr ( struct tp_field * field , struct format_field * format_field )
{
field - > offset = format_field - > offset ;
field - > pointer = tp_field__ptr ;
return 0 ;
}
struct syscall_tp {
struct tp_field id ;
union {
struct tp_field args , ret ;
} ;
} ;
static int perf_evsel__init_tp_uint_field ( struct perf_evsel * evsel ,
struct tp_field * field ,
const char * name )
{
struct format_field * format_field = perf_evsel__field ( evsel , name ) ;
if ( format_field = = NULL )
return - 1 ;
return tp_field__init_uint ( field , format_field , evsel - > needs_swap ) ;
}
# define perf_evsel__init_sc_tp_uint_field(evsel, name) \
( { struct syscall_tp * sc = evsel - > priv ; \
perf_evsel__init_tp_uint_field ( evsel , & sc - > name , # name ) ; } )
static int perf_evsel__init_tp_ptr_field ( struct perf_evsel * evsel ,
struct tp_field * field ,
const char * name )
{
struct format_field * format_field = perf_evsel__field ( evsel , name ) ;
if ( format_field = = NULL )
return - 1 ;
return tp_field__init_ptr ( field , format_field ) ;
}
# define perf_evsel__init_sc_tp_ptr_field(evsel, name) \
( { struct syscall_tp * sc = evsel - > priv ; \
perf_evsel__init_tp_ptr_field ( evsel , & sc - > name , # name ) ; } )
static void perf_evsel__delete_priv ( struct perf_evsel * evsel )
{
2013-12-27 00:41:15 +04:00
zfree ( & evsel - > priv ) ;
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
perf_evsel__delete ( evsel ) ;
}
2013-11-12 15:51:45 +04:00
static int perf_evsel__init_syscall_tp ( struct perf_evsel * evsel , void * handler )
{
evsel - > priv = malloc ( sizeof ( struct syscall_tp ) ) ;
if ( evsel - > priv ! = NULL ) {
if ( perf_evsel__init_sc_tp_uint_field ( evsel , id ) )
goto out_delete ;
evsel - > handler = handler ;
return 0 ;
}
return - ENOMEM ;
out_delete :
2013-12-27 00:41:15 +04:00
zfree ( & evsel - > priv ) ;
2013-11-12 15:51:45 +04:00
return - ENOENT ;
}
2013-11-07 23:41:19 +04:00
static struct perf_evsel * perf_evsel__syscall_newtp ( const char * direction , void * handler )
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
{
2013-11-07 23:41:19 +04:00
struct perf_evsel * evsel = perf_evsel__newtp ( " raw_syscalls " , direction ) ;
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
2013-12-05 06:41:39 +04:00
/* older kernel (e.g., RHEL6) use syscalls:{enter,exit} */
if ( evsel = = NULL )
evsel = perf_evsel__newtp ( " syscalls " , direction ) ;
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
if ( evsel ) {
2013-11-12 15:51:45 +04:00
if ( perf_evsel__init_syscall_tp ( evsel , handler ) )
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
goto out_delete ;
}
return evsel ;
out_delete :
perf_evsel__delete_priv ( evsel ) ;
return NULL ;
}
# define perf_evsel__sc_tp_uint(evsel, name, sample) \
( { struct syscall_tp * fields = evsel - > priv ; \
fields - > name . integer ( & fields - > name , sample ) ; } )
# define perf_evsel__sc_tp_ptr(evsel, name, sample) \
( { struct syscall_tp * fields = evsel - > priv ; \
fields - > name . pointer ( & fields - > name , sample ) ; } )
2013-09-03 19:20:12 +04:00
struct syscall_arg {
unsigned long val ;
2013-09-24 18:04:32 +04:00
struct thread * thread ;
struct trace * trace ;
2013-09-03 22:50:28 +04:00
void * parm ;
2013-09-03 19:20:12 +04:00
u8 idx ;
u8 mask ;
} ;
2013-09-03 22:50:28 +04:00
struct strarray {
2013-10-08 23:00:21 +04:00
int offset ;
2013-09-03 22:50:28 +04:00
int nr_entries ;
const char * * entries ;
} ;
# define DEFINE_STRARRAY(array) struct strarray strarray__##array = { \
. nr_entries = ARRAY_SIZE ( array ) , \
. entries = array , \
}
2013-10-08 23:00:21 +04:00
# define DEFINE_STRARRAY_OFFSET(array, off) struct strarray strarray__##array = { \
. offset = off , \
. nr_entries = ARRAY_SIZE ( array ) , \
. entries = array , \
}
2013-10-09 00:17:43 +04:00
static size_t __syscall_arg__scnprintf_strarray ( char * bf , size_t size ,
const char * intfmt ,
struct syscall_arg * arg )
2013-09-03 22:50:28 +04:00
{
struct strarray * sa = arg - > parm ;
2013-10-08 23:00:21 +04:00
int idx = arg - > val - sa - > offset ;
2013-09-03 22:50:28 +04:00
if ( idx < 0 | | idx > = sa - > nr_entries )
2013-10-09 00:17:43 +04:00
return scnprintf ( bf , size , intfmt , arg - > val ) ;
2013-09-03 22:50:28 +04:00
return scnprintf ( bf , size , " %s " , sa - > entries [ idx ] ) ;
}
2013-10-09 00:17:43 +04:00
static size_t syscall_arg__scnprintf_strarray ( char * bf , size_t size ,
struct syscall_arg * arg )
{
return __syscall_arg__scnprintf_strarray ( bf , size , " %d " , arg ) ;
}
2013-09-03 22:50:28 +04:00
# define SCA_STRARRAY syscall_arg__scnprintf_strarray
2014-02-10 21:09:48 +04:00
# if defined(__i386__) || defined(__x86_64__)
/*
* FIXME : Make this available to all arches as soon as the ioctl beautifier
* gets rewritten to support all arches .
*/
perf trace: Initial beautifier for ioctl's 'cmd' arg
[root@zoo linux]# trace -e ioctl | grep -v "cmd: 0x" | head -10
0.386 ( 0.001 ms): trace/1602 ioctl(fd: 1<pipe:[127057]>, cmd: TCGETS, arg: 0x7fff59fcb4d0 ) = -1 ENOTTY Inappropriate ioctl for device
1459.368 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.586 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.611 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
3740.526 ( 0.002 ms): awk/1612 ioctl(fd: 1<pipe:[128265]>, cmd: TCGETS, arg: 0x7fff4d166b90 ) = -1 ENOTTY Inappropriate ioctl for device
3740.704 ( 0.001 ms): awk/1612 ioctl(fd: 3</proc/meminfo>, cmd: TCGETS, arg: 0x7fff4d1669a0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.550 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.555 ( 0.003 ms): ps/1614 ioctl(fd: 2<socket:[19550]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.558 ( 0.002 ms): ps/1614 ioctl(cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.572 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TCGETS, arg: 0x7fff59176220 ) = -1 ENOTTY Inappropriate ioctl for device
[root@zoo linux]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-afajwap3mr60dfl4qpdl1pxn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-09 00:43:20 +04:00
static size_t syscall_arg__scnprintf_strhexarray ( char * bf , size_t size ,
struct syscall_arg * arg )
{
return __syscall_arg__scnprintf_strarray ( bf , size , " %#x " , arg ) ;
}
# define SCA_STRHEXARRAY syscall_arg__scnprintf_strhexarray
2014-02-10 21:09:48 +04:00
# endif /* defined(__i386__) || defined(__x86_64__) */
perf trace: Initial beautifier for ioctl's 'cmd' arg
[root@zoo linux]# trace -e ioctl | grep -v "cmd: 0x" | head -10
0.386 ( 0.001 ms): trace/1602 ioctl(fd: 1<pipe:[127057]>, cmd: TCGETS, arg: 0x7fff59fcb4d0 ) = -1 ENOTTY Inappropriate ioctl for device
1459.368 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.586 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.611 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
3740.526 ( 0.002 ms): awk/1612 ioctl(fd: 1<pipe:[128265]>, cmd: TCGETS, arg: 0x7fff4d166b90 ) = -1 ENOTTY Inappropriate ioctl for device
3740.704 ( 0.001 ms): awk/1612 ioctl(fd: 3</proc/meminfo>, cmd: TCGETS, arg: 0x7fff4d1669a0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.550 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.555 ( 0.003 ms): ps/1614 ioctl(fd: 2<socket:[19550]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.558 ( 0.002 ms): ps/1614 ioctl(cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.572 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TCGETS, arg: 0x7fff59176220 ) = -1 ENOTTY Inappropriate ioctl for device
[root@zoo linux]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-afajwap3mr60dfl4qpdl1pxn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-09 00:43:20 +04:00
2013-09-24 18:04:32 +04:00
static size_t syscall_arg__scnprintf_fd ( char * bf , size_t size ,
struct syscall_arg * arg ) ;
# define SCA_FD syscall_arg__scnprintf_fd
static size_t syscall_arg__scnprintf_fd_at ( char * bf , size_t size ,
struct syscall_arg * arg )
{
int fd = arg - > val ;
if ( fd = = AT_FDCWD )
return scnprintf ( bf , size , " CWD " ) ;
return syscall_arg__scnprintf_fd ( bf , size , arg ) ;
}
# define SCA_FDAT syscall_arg__scnprintf_fd_at
static size_t syscall_arg__scnprintf_close_fd ( char * bf , size_t size ,
struct syscall_arg * arg ) ;
# define SCA_CLOSE_FD syscall_arg__scnprintf_close_fd
2013-09-02 17:39:21 +04:00
static size_t syscall_arg__scnprintf_hex ( char * bf , size_t size ,
2013-09-03 19:20:12 +04:00
struct syscall_arg * arg )
2013-08-24 01:14:48 +04:00
{
2013-09-03 19:20:12 +04:00
return scnprintf ( bf , size , " %#lx " , arg - > val ) ;
2013-08-24 01:14:48 +04:00
}
2013-08-26 19:29:38 +04:00
# define SCA_HEX syscall_arg__scnprintf_hex
trace: Beautify perf_event_open syscall
Syswide tracing and then running 'stat' and 'trace':
$ perf trace -e perf_event_open
1034.649 (0.019 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1, flags: FD_CLOEXEC) = -1 EINVAL Invalid argument
1034.670 (0.008 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
1034.681 (0.007 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
1034.692 (0.007 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
9986.983 (0.014 ms): trace/6139 perf_event_open(attr_uptr: 0x7ffd9c629320, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.026 (0.016 ms): trace/6139 perf_event_open(attr_uptr: 0x37c7e70, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.041 (0.008 ms): trace/6139 perf_event_open(attr_uptr: 0x37c7e70, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.489 (0.092 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.536 (0.044 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 1, group_fd: -1, flags: FD_CLOEXEC) = 4
9987.580 (0.041 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 2, group_fd: -1, flags: FD_CLOEXEC) = 5
9987.620 (0.037 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 3, group_fd: -1, flags: FD_CLOEXEC) = 7
9987.659 (0.035 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, group_fd: -1, flags: FD_CLOEXEC) = 8
9987.692 (0.031 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 1, group_fd: -1, flags: FD_CLOEXEC) = 9
9987.727 (0.032 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 2, group_fd: -1, flags: FD_CLOEXEC) = 10
9987.761 (0.031 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 3, group_fd: -1, flags: FD_CLOEXEC) = 11
Need to intercept perf_copy_attr() with a kprobe or with eBPF...
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-njb105hab2i3t5dexym9lskl@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-12 04:47:54 +03:00
static size_t syscall_arg__scnprintf_int ( char * bf , size_t size ,
struct syscall_arg * arg )
{
return scnprintf ( bf , size , " %d " , arg - > val ) ;
}
# define SCA_INT syscall_arg__scnprintf_int
2013-09-02 17:39:21 +04:00
static size_t syscall_arg__scnprintf_mmap_prot ( char * bf , size_t size ,
2013-09-03 19:20:12 +04:00
struct syscall_arg * arg )
perf trace: Add beautifier for mmap prot parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.984 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd62ae000
1.114 ( 0.016 ms): mmap(addr: 0, len: 125100, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xd628f000
1.252 ( 0.020 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: 2050, fd: 3, off: 0) = 0xc1600000
1.282 ( 0.024 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.315 ( 0.026 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: 2066, fd: 3, off: 1757184) = 0xc19ad000
1.352 ( 0.017 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: 50, fd: 4294967295, off: 0) = 0xc19b3000
1.415 ( 0.011 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628e000
1.440 ( 0.011 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628c000
1.569 ( 0.019 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.591 ( 0.017 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.616 ( 0.016 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
2.105 ( 0.018 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xcfe9c000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-21 00:44:42 +04:00
{
2013-09-03 19:20:12 +04:00
int printed = 0 , prot = arg - > val ;
perf trace: Add beautifier for mmap prot parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.984 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd62ae000
1.114 ( 0.016 ms): mmap(addr: 0, len: 125100, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xd628f000
1.252 ( 0.020 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: 2050, fd: 3, off: 0) = 0xc1600000
1.282 ( 0.024 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.315 ( 0.026 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: 2066, fd: 3, off: 1757184) = 0xc19ad000
1.352 ( 0.017 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: 50, fd: 4294967295, off: 0) = 0xc19b3000
1.415 ( 0.011 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628e000
1.440 ( 0.011 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628c000
1.569 ( 0.019 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.591 ( 0.017 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.616 ( 0.016 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
2.105 ( 0.018 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xcfe9c000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-21 00:44:42 +04:00
if ( prot = = PROT_NONE )
return scnprintf ( bf , size , " NONE " ) ;
# define P_MMAP_PROT(n) \
if ( prot & PROT_ # # n ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , # n ) ; \
prot & = ~ PROT_ # # n ; \
}
P_MMAP_PROT ( EXEC ) ;
P_MMAP_PROT ( READ ) ;
P_MMAP_PROT ( WRITE ) ;
# ifdef PROT_SEM
P_MMAP_PROT ( SEM ) ;
# endif
P_MMAP_PROT ( GROWSDOWN ) ;
P_MMAP_PROT ( GROWSUP ) ;
# undef P_MMAP_PROT
if ( prot )
printed + = scnprintf ( bf + printed , size - printed , " %s%#x " , printed ? " | " : " " , prot ) ;
return printed ;
}
# define SCA_MMAP_PROT syscall_arg__scnprintf_mmap_prot
2013-09-02 17:39:21 +04:00
static size_t syscall_arg__scnprintf_mmap_flags ( char * bf , size_t size ,
2013-09-03 19:20:12 +04:00
struct syscall_arg * arg )
perf trace: Add beautifier for mmap flags parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.992 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa60be000
1.108 ( 0.012 ms): mmap(addr: 0, len: 125100, prot: READ, flags: PRIVATE, fd: 3, off: 0 ) = 0xa609f000
1.209 ( 0.014 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3, off: 0) = 0xc1600000
1.232 ( 0.018 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.255 ( 0.018 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3, off: 1757184) = 0xc19ad000
1.281 ( 0.011 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: 4294967295, off: 0) = 0xc19b3000
1.328 ( 0.008 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609e000
1.346 ( 0.008 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609c000
1.443 ( 0.013 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.459 ( 0.011 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.477 ( 0.011 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
1.855 ( 0.013 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: PRIVATE, fd: 3, off: 0) = 0x9fcac000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-23 17:48:33 +04:00
{
2013-09-03 19:20:12 +04:00
int printed = 0 , flags = arg - > val ;
perf trace: Add beautifier for mmap flags parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.992 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa60be000
1.108 ( 0.012 ms): mmap(addr: 0, len: 125100, prot: READ, flags: PRIVATE, fd: 3, off: 0 ) = 0xa609f000
1.209 ( 0.014 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3, off: 0) = 0xc1600000
1.232 ( 0.018 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.255 ( 0.018 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3, off: 1757184) = 0xc19ad000
1.281 ( 0.011 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: 4294967295, off: 0) = 0xc19b3000
1.328 ( 0.008 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609e000
1.346 ( 0.008 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609c000
1.443 ( 0.013 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.459 ( 0.011 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.477 ( 0.011 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
1.855 ( 0.013 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: PRIVATE, fd: 3, off: 0) = 0x9fcac000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-23 17:48:33 +04:00
# define P_MMAP_FLAG(n) \
if ( flags & MAP_ # # n ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , # n ) ; \
flags & = ~ MAP_ # # n ; \
}
P_MMAP_FLAG ( SHARED ) ;
P_MMAP_FLAG ( PRIVATE ) ;
2013-09-05 18:29:47 +04:00
# ifdef MAP_32BIT
perf trace: Add beautifier for mmap flags parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.992 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa60be000
1.108 ( 0.012 ms): mmap(addr: 0, len: 125100, prot: READ, flags: PRIVATE, fd: 3, off: 0 ) = 0xa609f000
1.209 ( 0.014 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3, off: 0) = 0xc1600000
1.232 ( 0.018 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.255 ( 0.018 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3, off: 1757184) = 0xc19ad000
1.281 ( 0.011 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: 4294967295, off: 0) = 0xc19b3000
1.328 ( 0.008 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609e000
1.346 ( 0.008 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609c000
1.443 ( 0.013 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.459 ( 0.011 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.477 ( 0.011 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
1.855 ( 0.013 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: PRIVATE, fd: 3, off: 0) = 0x9fcac000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-23 17:48:33 +04:00
P_MMAP_FLAG ( 32 BIT ) ;
2013-09-05 18:29:47 +04:00
# endif
perf trace: Add beautifier for mmap flags parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.992 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa60be000
1.108 ( 0.012 ms): mmap(addr: 0, len: 125100, prot: READ, flags: PRIVATE, fd: 3, off: 0 ) = 0xa609f000
1.209 ( 0.014 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3, off: 0) = 0xc1600000
1.232 ( 0.018 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.255 ( 0.018 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3, off: 1757184) = 0xc19ad000
1.281 ( 0.011 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: 4294967295, off: 0) = 0xc19b3000
1.328 ( 0.008 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609e000
1.346 ( 0.008 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609c000
1.443 ( 0.013 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.459 ( 0.011 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.477 ( 0.011 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
1.855 ( 0.013 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: PRIVATE, fd: 3, off: 0) = 0x9fcac000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-23 17:48:33 +04:00
P_MMAP_FLAG ( ANONYMOUS ) ;
P_MMAP_FLAG ( DENYWRITE ) ;
P_MMAP_FLAG ( EXECUTABLE ) ;
P_MMAP_FLAG ( FILE ) ;
P_MMAP_FLAG ( FIXED ) ;
P_MMAP_FLAG ( GROWSDOWN ) ;
2013-08-27 20:50:40 +04:00
# ifdef MAP_HUGETLB
perf trace: Add beautifier for mmap flags parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.992 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa60be000
1.108 ( 0.012 ms): mmap(addr: 0, len: 125100, prot: READ, flags: PRIVATE, fd: 3, off: 0 ) = 0xa609f000
1.209 ( 0.014 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3, off: 0) = 0xc1600000
1.232 ( 0.018 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.255 ( 0.018 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3, off: 1757184) = 0xc19ad000
1.281 ( 0.011 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: 4294967295, off: 0) = 0xc19b3000
1.328 ( 0.008 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609e000
1.346 ( 0.008 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609c000
1.443 ( 0.013 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.459 ( 0.011 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.477 ( 0.011 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
1.855 ( 0.013 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: PRIVATE, fd: 3, off: 0) = 0x9fcac000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-23 17:48:33 +04:00
P_MMAP_FLAG ( HUGETLB ) ;
2013-08-27 20:50:40 +04:00
# endif
perf trace: Add beautifier for mmap flags parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.992 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa60be000
1.108 ( 0.012 ms): mmap(addr: 0, len: 125100, prot: READ, flags: PRIVATE, fd: 3, off: 0 ) = 0xa609f000
1.209 ( 0.014 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3, off: 0) = 0xc1600000
1.232 ( 0.018 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.255 ( 0.018 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3, off: 1757184) = 0xc19ad000
1.281 ( 0.011 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: 4294967295, off: 0) = 0xc19b3000
1.328 ( 0.008 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609e000
1.346 ( 0.008 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609c000
1.443 ( 0.013 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.459 ( 0.011 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.477 ( 0.011 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
1.855 ( 0.013 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: PRIVATE, fd: 3, off: 0) = 0x9fcac000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-23 17:48:33 +04:00
P_MMAP_FLAG ( LOCKED ) ;
P_MMAP_FLAG ( NONBLOCK ) ;
P_MMAP_FLAG ( NORESERVE ) ;
P_MMAP_FLAG ( POPULATE ) ;
P_MMAP_FLAG ( STACK ) ;
# ifdef MAP_UNINITIALIZED
P_MMAP_FLAG ( UNINITIALIZED ) ;
# endif
# undef P_MMAP_FLAG
if ( flags )
printed + = scnprintf ( bf + printed , size - printed , " %s%#x " , printed ? " | " : " " , flags ) ;
return printed ;
}
# define SCA_MMAP_FLAGS syscall_arg__scnprintf_mmap_flags
perf trace: Add beautifier for mremap flags param
~/devel/kernel/tools/perf(branch:master*) » sudo ./perf trace ~/mremap_test
0.543 ( 0.003 ms): mprotect(start: 0x600000, len: 4096, prot: READ ) = 0
0.550 ( 0.003 ms): mprotect(start: 0x7f441260d000, len: 4096, prot: READ) = 0
0.561 ( 0.010 ms): munmap(addr: 0x7f44125e2000, len: 165572 ) = 0
0.595 ( 0.012 ms): mmap(len: 12288, prot: READ|WRITE, flags: SHARED|ANONYMOUS|LOCKED, fd: -1) = 0x12608000
0.603 ( 0.006 ms): mremap(addr: 0x7f4412608000, old_len: 4096, new_len: 4096, flags: MAYMOVE|FIXED, new_addr: 0x7f16da295000) = 0xda295000
0.608 ( 0.003 ms): mremap(addr: 0x7f441260a000, old_len: 4096, new_len: 4096, flags: MAYMOVE|FIXED, new_addr: 0x7f16da297000) = 0xda297000
0.612 ( 0.003 ms): mremap(addr: 0x7f4412609000, old_len: 4096, new_len: 4096, flags: MAYMOVE|FIXED, new_addr: 0x7f16da296000) = 0xda296000
0.619 ( 0.000 ms): exit_group(
Signed-off-by: Alex Snast <asnast@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1407944560-26924-1-git-send-email-asnast@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2014-08-13 19:42:40 +04:00
static size_t syscall_arg__scnprintf_mremap_flags ( char * bf , size_t size ,
struct syscall_arg * arg )
{
int printed = 0 , flags = arg - > val ;
# define P_MREMAP_FLAG(n) \
if ( flags & MREMAP_ # # n ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , # n ) ; \
flags & = ~ MREMAP_ # # n ; \
}
P_MREMAP_FLAG ( MAYMOVE ) ;
# ifdef MREMAP_FIXED
P_MREMAP_FLAG ( FIXED ) ;
# endif
# undef P_MREMAP_FLAG
if ( flags )
printed + = scnprintf ( bf + printed , size - printed , " %s%#x " , printed ? " | " : " " , flags ) ;
return printed ;
}
# define SCA_MREMAP_FLAGS syscall_arg__scnprintf_mremap_flags
2013-09-02 17:39:21 +04:00
static size_t syscall_arg__scnprintf_madvise_behavior ( char * bf , size_t size ,
2013-09-03 19:20:12 +04:00
struct syscall_arg * arg )
2013-08-23 17:06:41 +04:00
{
2013-09-03 19:20:12 +04:00
int behavior = arg - > val ;
2013-08-23 17:06:41 +04:00
switch ( behavior ) {
# define P_MADV_BHV(n) case MADV_##n: return scnprintf(bf, size, #n)
P_MADV_BHV ( NORMAL ) ;
P_MADV_BHV ( RANDOM ) ;
P_MADV_BHV ( SEQUENTIAL ) ;
P_MADV_BHV ( WILLNEED ) ;
P_MADV_BHV ( DONTNEED ) ;
P_MADV_BHV ( REMOVE ) ;
P_MADV_BHV ( DONTFORK ) ;
P_MADV_BHV ( DOFORK ) ;
P_MADV_BHV ( HWPOISON ) ;
# ifdef MADV_SOFT_OFFLINE
P_MADV_BHV ( SOFT_OFFLINE ) ;
# endif
P_MADV_BHV ( MERGEABLE ) ;
P_MADV_BHV ( UNMERGEABLE ) ;
2013-08-27 20:50:40 +04:00
# ifdef MADV_HUGEPAGE
2013-08-23 17:06:41 +04:00
P_MADV_BHV ( HUGEPAGE ) ;
2013-08-27 20:50:40 +04:00
# endif
# ifdef MADV_NOHUGEPAGE
2013-08-23 17:06:41 +04:00
P_MADV_BHV ( NOHUGEPAGE ) ;
2013-08-27 20:50:40 +04:00
# endif
2013-08-23 17:06:41 +04:00
# ifdef MADV_DONTDUMP
P_MADV_BHV ( DONTDUMP ) ;
# endif
# ifdef MADV_DODUMP
P_MADV_BHV ( DODUMP ) ;
# endif
# undef P_MADV_PHV
default : break ;
}
return scnprintf ( bf , size , " %#x " , behavior ) ;
}
# define SCA_MADV_BHV syscall_arg__scnprintf_madvise_behavior
2013-09-20 18:49:50 +04:00
static size_t syscall_arg__scnprintf_flock ( char * bf , size_t size ,
struct syscall_arg * arg )
{
int printed = 0 , op = arg - > val ;
if ( op = = 0 )
return scnprintf ( bf , size , " NONE " ) ;
# define P_CMD(cmd) \
if ( ( op & LOCK_ # # cmd ) = = LOCK_ # # cmd ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , # cmd ) ; \
op & = ~ LOCK_ # # cmd ; \
}
P_CMD ( SH ) ;
P_CMD ( EX ) ;
P_CMD ( NB ) ;
P_CMD ( UN ) ;
P_CMD ( MAND ) ;
P_CMD ( RW ) ;
P_CMD ( READ ) ;
P_CMD ( WRITE ) ;
# undef P_OP
if ( op )
printed + = scnprintf ( bf + printed , size - printed , " %s%#x " , printed ? " | " : " " , op ) ;
return printed ;
}
# define SCA_FLOCK syscall_arg__scnprintf_flock
2013-09-03 19:20:12 +04:00
static size_t syscall_arg__scnprintf_futex_op ( char * bf , size_t size , struct syscall_arg * arg )
2013-09-02 20:46:44 +04:00
{
enum syscall_futex_args {
SCF_UADDR = ( 1 < < 0 ) ,
SCF_OP = ( 1 < < 1 ) ,
SCF_VAL = ( 1 < < 2 ) ,
SCF_TIMEOUT = ( 1 < < 3 ) ,
SCF_UADDR2 = ( 1 < < 4 ) ,
SCF_VAL3 = ( 1 < < 5 ) ,
} ;
2013-09-03 19:20:12 +04:00
int op = arg - > val ;
2013-09-02 20:46:44 +04:00
int cmd = op & FUTEX_CMD_MASK ;
size_t printed = 0 ;
switch ( cmd ) {
# define P_FUTEX_OP(n) case FUTEX_##n: printed = scnprintf(bf, size, #n);
2013-09-03 19:20:12 +04:00
P_FUTEX_OP ( WAIT ) ; arg - > mask | = SCF_VAL3 | SCF_UADDR2 ; break ;
P_FUTEX_OP ( WAKE ) ; arg - > mask | = SCF_VAL3 | SCF_UADDR2 | SCF_TIMEOUT ; break ;
P_FUTEX_OP ( FD ) ; arg - > mask | = SCF_VAL3 | SCF_UADDR2 | SCF_TIMEOUT ; break ;
P_FUTEX_OP ( REQUEUE ) ; arg - > mask | = SCF_VAL3 | SCF_TIMEOUT ; break ;
P_FUTEX_OP ( CMP_REQUEUE ) ; arg - > mask | = SCF_TIMEOUT ; break ;
P_FUTEX_OP ( CMP_REQUEUE_PI ) ; arg - > mask | = SCF_TIMEOUT ; break ;
2013-09-02 20:46:44 +04:00
P_FUTEX_OP ( WAKE_OP ) ; break ;
2013-09-03 19:20:12 +04:00
P_FUTEX_OP ( LOCK_PI ) ; arg - > mask | = SCF_VAL3 | SCF_UADDR2 | SCF_TIMEOUT ; break ;
P_FUTEX_OP ( UNLOCK_PI ) ; arg - > mask | = SCF_VAL3 | SCF_UADDR2 | SCF_TIMEOUT ; break ;
P_FUTEX_OP ( TRYLOCK_PI ) ; arg - > mask | = SCF_VAL3 | SCF_UADDR2 ; break ;
P_FUTEX_OP ( WAIT_BITSET ) ; arg - > mask | = SCF_UADDR2 ; break ;
P_FUTEX_OP ( WAKE_BITSET ) ; arg - > mask | = SCF_UADDR2 ; break ;
2013-09-02 20:46:44 +04:00
P_FUTEX_OP ( WAIT_REQUEUE_PI ) ; break ;
default : printed = scnprintf ( bf , size , " %#x " , cmd ) ; break ;
}
if ( op & FUTEX_PRIVATE_FLAG )
printed + = scnprintf ( bf + printed , size - printed , " |PRIV " ) ;
if ( op & FUTEX_CLOCK_REALTIME )
printed + = scnprintf ( bf + printed , size - printed , " |CLKRT " ) ;
return printed ;
}
2013-09-03 23:15:12 +04:00
# define SCA_FUTEX_OP syscall_arg__scnprintf_futex_op
2013-10-08 23:00:21 +04:00
static const char * epoll_ctl_ops [ ] = { " ADD " , " DEL " , " MOD " , } ;
static DEFINE_STRARRAY_OFFSET ( epoll_ctl_ops , 1 ) ;
perf trace: Beautify epoll_ctl 'op' arg
[root@sandy ~]# perf trace -e epoll_ctl
2.490 (0.003 ms): systemd-logind/586 epoll_ctl(epfd: 10, op: ADD, fd: 24, event: 0x7fff22314ef0) = 0
2.621 (0.003 ms): systemd-logind/586 epoll_ctl(epfd: 10, op: DEL, fd: 24 ) = 0
2.833 (0.010 ms): systemd-logind/586 epoll_ctl(epfd: 10, op: ADD, fd: 24, event: 0x7fff22314cd0) = 0
2.953 (0.002 ms): systemd-logind/586 epoll_ctl(epfd: 10, op: DEL, fd: 24 ) = 0
3.118 (0.002 ms): systemd-logind/586 epoll_ctl(epfd: 10, op: ADD, fd: 24, event: 0x7fff22314d20) = 0
4.762 (0.002 ms): systemd-logind/586 epoll_ctl(epfd: 10, op: DEL, fd: 24 ) = 0
^C[root@sandy ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-88xz9phc8cbicnxonud6if8h@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-20 18:27:32 +04:00
2013-09-03 22:50:28 +04:00
static const char * itimers [ ] = { " REAL " , " VIRTUAL " , " PROF " , } ;
static DEFINE_STRARRAY ( itimers ) ;
2013-09-03 23:15:12 +04:00
static const char * whences [ ] = { " SET " , " CUR " , " END " ,
# ifdef SEEK_DATA
" DATA " ,
# endif
# ifdef SEEK_HOLE
" HOLE " ,
# endif
} ;
static DEFINE_STRARRAY ( whences ) ;
2013-09-02 20:46:44 +04:00
2013-09-03 23:28:58 +04:00
static const char * fcntl_cmds [ ] = {
" DUPFD " , " GETFD " , " SETFD " , " GETFL " , " SETFL " , " GETLK " , " SETLK " ,
" SETLKW " , " SETOWN " , " GETOWN " , " SETSIG " , " GETSIG " , " F_GETLK64 " ,
" F_SETLK64 " , " F_SETLKW64 " , " F_SETOWN_EX " , " F_GETOWN_EX " ,
" F_GETOWNER_UIDS " ,
} ;
static DEFINE_STRARRAY ( fcntl_cmds ) ;
2013-09-04 18:52:33 +04:00
static const char * rlimit_resources [ ] = {
" CPU " , " FSIZE " , " DATA " , " STACK " , " CORE " , " RSS " , " NPROC " , " NOFILE " ,
" MEMLOCK " , " AS " , " LOCKS " , " SIGPENDING " , " MSGQUEUE " , " NICE " , " RTPRIO " ,
" RTTIME " ,
} ;
static DEFINE_STRARRAY ( rlimit_resources ) ;
2013-09-03 23:37:46 +04:00
static const char * sighow [ ] = { " BLOCK " , " UNBLOCK " , " SETMASK " , } ;
static DEFINE_STRARRAY ( sighow ) ;
2013-09-23 05:45:00 +04:00
static const char * clockid [ ] = {
" REALTIME " , " MONOTONIC " , " PROCESS_CPUTIME_ID " , " THREAD_CPUTIME_ID " ,
" MONOTONIC_RAW " , " REALTIME_COARSE " , " MONOTONIC_COARSE " ,
} ;
static DEFINE_STRARRAY ( clockid ) ;
2013-09-04 17:27:41 +04:00
static const char * socket_families [ ] = {
" UNSPEC " , " LOCAL " , " INET " , " AX25 " , " IPX " , " APPLETALK " , " NETROM " ,
" BRIDGE " , " ATMPVC " , " X25 " , " INET6 " , " ROSE " , " DECnet " , " NETBEUI " ,
" SECURITY " , " KEY " , " NETLINK " , " PACKET " , " ASH " , " ECONET " , " ATMSVC " ,
" RDS " , " SNA " , " IRDA " , " PPPOX " , " WANPIPE " , " LLC " , " IB " , " CAN " , " TIPC " ,
" BLUETOOTH " , " IUCV " , " RXRPC " , " ISDN " , " PHONET " , " IEEE802154 " , " CAIF " ,
" ALG " , " NFC " , " VSOCK " ,
} ;
static DEFINE_STRARRAY ( socket_families ) ;
2013-09-04 18:00:44 +04:00
# ifndef SOCK_TYPE_MASK
# define SOCK_TYPE_MASK 0xf
# endif
static size_t syscall_arg__scnprintf_socket_type ( char * bf , size_t size ,
struct syscall_arg * arg )
{
size_t printed ;
int type = arg - > val ,
flags = type & ~ SOCK_TYPE_MASK ;
type & = SOCK_TYPE_MASK ;
/*
* Can ' t use a strarray , MIPS may override for ABI reasons .
*/
switch ( type ) {
# define P_SK_TYPE(n) case SOCK_##n: printed = scnprintf(bf, size, #n); break;
P_SK_TYPE ( STREAM ) ;
P_SK_TYPE ( DGRAM ) ;
P_SK_TYPE ( RAW ) ;
P_SK_TYPE ( RDM ) ;
P_SK_TYPE ( SEQPACKET ) ;
P_SK_TYPE ( DCCP ) ;
P_SK_TYPE ( PACKET ) ;
# undef P_SK_TYPE
default :
printed = scnprintf ( bf , size , " %#x " , type ) ;
}
# define P_SK_FLAG(n) \
if ( flags & SOCK_ # # n ) { \
printed + = scnprintf ( bf + printed , size - printed , " |%s " , # n ) ; \
flags & = ~ SOCK_ # # n ; \
}
P_SK_FLAG ( CLOEXEC ) ;
P_SK_FLAG ( NONBLOCK ) ;
# undef P_SK_FLAG
if ( flags )
printed + = scnprintf ( bf + printed , size - printed , " |%#x " , flags ) ;
return printed ;
}
# define SCA_SK_TYPE syscall_arg__scnprintf_socket_type
perf trace: Beautify send/recv syscall 'flags' arg
[root@sandy ~]# perf trace -a -e recvmmsg,recvmsg,recvfrom,sendto,sendmsg,sendmmsg
6.901 (0.002 ms): 589 recvmsg(fd: 51, msg: 0x7fff35673420, flags: CMSG_CLOEXEC) = -1 EAGAIN Resource temporarily unavailable
6.966 (0.008 ms): 589 sendmsg(fd: 50, msg: 0x7fff35673230, flags: NOSIGNAL ) = 961
6.984 (0.004 ms): 979 sendmsg(fd: 3, msg: 0x7fff5b484940, flags: NOSIGNAL ) = 945
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-h25k5k50nac0ej5cl5iwgvae@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-12 18:54:48 +04:00
# ifndef MSG_PROBE
# define MSG_PROBE 0x10
# endif
2013-09-23 05:44:56 +04:00
# ifndef MSG_WAITFORONE
# define MSG_WAITFORONE 0x10000
# endif
perf trace: Beautify send/recv syscall 'flags' arg
[root@sandy ~]# perf trace -a -e recvmmsg,recvmsg,recvfrom,sendto,sendmsg,sendmmsg
6.901 (0.002 ms): 589 recvmsg(fd: 51, msg: 0x7fff35673420, flags: CMSG_CLOEXEC) = -1 EAGAIN Resource temporarily unavailable
6.966 (0.008 ms): 589 sendmsg(fd: 50, msg: 0x7fff35673230, flags: NOSIGNAL ) = 961
6.984 (0.004 ms): 979 sendmsg(fd: 3, msg: 0x7fff5b484940, flags: NOSIGNAL ) = 945
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-h25k5k50nac0ej5cl5iwgvae@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-12 18:54:48 +04:00
# ifndef MSG_SENDPAGE_NOTLAST
# define MSG_SENDPAGE_NOTLAST 0x20000
# endif
# ifndef MSG_FASTOPEN
# define MSG_FASTOPEN 0x20000000
# endif
static size_t syscall_arg__scnprintf_msg_flags ( char * bf , size_t size ,
struct syscall_arg * arg )
{
int printed = 0 , flags = arg - > val ;
if ( flags = = 0 )
return scnprintf ( bf , size , " NONE " ) ;
# define P_MSG_FLAG(n) \
if ( flags & MSG_ # # n ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , # n ) ; \
flags & = ~ MSG_ # # n ; \
}
P_MSG_FLAG ( OOB ) ;
P_MSG_FLAG ( PEEK ) ;
P_MSG_FLAG ( DONTROUTE ) ;
P_MSG_FLAG ( TRYHARD ) ;
P_MSG_FLAG ( CTRUNC ) ;
P_MSG_FLAG ( PROBE ) ;
P_MSG_FLAG ( TRUNC ) ;
P_MSG_FLAG ( DONTWAIT ) ;
P_MSG_FLAG ( EOR ) ;
P_MSG_FLAG ( WAITALL ) ;
P_MSG_FLAG ( FIN ) ;
P_MSG_FLAG ( SYN ) ;
P_MSG_FLAG ( CONFIRM ) ;
P_MSG_FLAG ( RST ) ;
P_MSG_FLAG ( ERRQUEUE ) ;
P_MSG_FLAG ( NOSIGNAL ) ;
P_MSG_FLAG ( MORE ) ;
P_MSG_FLAG ( WAITFORONE ) ;
P_MSG_FLAG ( SENDPAGE_NOTLAST ) ;
P_MSG_FLAG ( FASTOPEN ) ;
P_MSG_FLAG ( CMSG_CLOEXEC ) ;
# undef P_MSG_FLAG
if ( flags )
printed + = scnprintf ( bf + printed , size - printed , " %s%#x " , printed ? " | " : " " , flags ) ;
return printed ;
}
# define SCA_MSG_FLAGS syscall_arg__scnprintf_msg_flags
2013-09-04 18:42:27 +04:00
static size_t syscall_arg__scnprintf_access_mode ( char * bf , size_t size ,
struct syscall_arg * arg )
{
size_t printed = 0 ;
int mode = arg - > val ;
if ( mode = = F_OK ) /* 0 */
return scnprintf ( bf , size , " F " ) ;
# define P_MODE(n) \
if ( mode & n # # _OK ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s " , # n ) ; \
mode & = ~ n # # _OK ; \
}
P_MODE ( R ) ;
P_MODE ( W ) ;
P_MODE ( X ) ;
# undef P_MODE
if ( mode )
printed + = scnprintf ( bf + printed , size - printed , " |%#x " , mode ) ;
return printed ;
}
# define SCA_ACCMODE syscall_arg__scnprintf_access_mode
perf trace: Deref sys_enter pointer args with contents from probe:vfs_getname
To work like strace and dereference syscall pointer args we need to
insert probes (or tracepoints) right after we copy those bytes from
userspace.
Since we're formatting the syscall args at raw_syscalls:sys_enter time,
we need to have a formatter that just stores the position where, later,
when we get the probe:vfs_getname, we can insert the pointer contents.
Now, if a probe:vfs_getname with this format is in place:
# perf probe -l
probe:vfs_getname (on getname_flags:72@/home/git/linux/fs/namei.c with pathname)
That was, in this case, put in place with:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
Then 'perf trace' will notice that and do the pointer -> contents
expansion:
# trace -e open touch /tmp/bla
0.165 (0.010 ms): touch/17752 open(filename: /etc/ld.so.cache, flags: CLOEXEC) = 3
0.195 (0.011 ms): touch/17752 open(filename: /lib64/libc.so.6, flags: CLOEXEC) = 3
0.512 (0.012 ms): touch/17752 open(filename: /usr/lib/locale/locale-archive, flags: CLOEXEC) = 3
0.582 (0.012 ms): touch/17752 open(filename: /tmp/bla, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
#
Roughly equivalent to strace's output:
# strace -rT -e open touch /tmp/bla
0.000000 open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3 <0.000039>
0.000317 open("/lib64/libc.so.6", O_RDONLY|O_CLOEXEC) = 3 <0.000102>
0.001461 open("/usr/lib/locale/locale-archive", O_RDONLY|O_CLOEXEC) = 3 <0.000072>
0.000405 open("/tmp/bla", O_WRONLY|O_CREAT|O_NOCTTY|O_NONBLOCK, 0666) = 3 <0.000055>
0.000641 +++ exited with 0 +++
#
Now we need to either look for at all syscalls that are marked as
pointers and have some well known names ("filename", "pathname", etc)
and set the arg formatter to the one used for the "open" syscall in this
patch.
This implementation works for syscalls with just a string being copied
from userspace, for matching syscalls with more than one string being
copied via the same probe/trace point (vfs_getname) we need to extend
the vfs_getname probe spec to include the pointer too, but there are
some problems with that in 'perf probe' or the kernel kprobes code, need
to investigate before considering supporting multiple strings per
syscall.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-xvuwx6nuj8cf389kf9s2ue2s@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-08-05 04:30:09 +03:00
static size_t syscall_arg__scnprintf_filename ( char * bf , size_t size ,
struct syscall_arg * arg ) ;
# define SCA_FILENAME syscall_arg__scnprintf_filename
2013-09-02 23:22:31 +04:00
static size_t syscall_arg__scnprintf_open_flags ( char * bf , size_t size ,
2013-09-03 19:20:12 +04:00
struct syscall_arg * arg )
2013-09-02 23:22:31 +04:00
{
2013-09-03 19:20:12 +04:00
int printed = 0 , flags = arg - > val ;
2013-09-02 23:22:31 +04:00
if ( ! ( flags & O_CREAT ) )
2013-09-03 19:20:12 +04:00
arg - > mask | = 1 < < ( arg - > idx + 1 ) ; /* Mask the mode parm */
2013-09-02 23:22:31 +04:00
if ( flags = = 0 )
return scnprintf ( bf , size , " RDONLY " ) ;
# define P_FLAG(n) \
if ( flags & O_ # # n ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , # n ) ; \
flags & = ~ O_ # # n ; \
}
P_FLAG ( APPEND ) ;
P_FLAG ( ASYNC ) ;
P_FLAG ( CLOEXEC ) ;
P_FLAG ( CREAT ) ;
P_FLAG ( DIRECT ) ;
P_FLAG ( DIRECTORY ) ;
P_FLAG ( EXCL ) ;
P_FLAG ( LARGEFILE ) ;
P_FLAG ( NOATIME ) ;
P_FLAG ( NOCTTY ) ;
# ifdef O_NONBLOCK
P_FLAG ( NONBLOCK ) ;
# elif O_NDELAY
P_FLAG ( NDELAY ) ;
# endif
# ifdef O_PATH
P_FLAG ( PATH ) ;
# endif
P_FLAG ( RDWR ) ;
# ifdef O_DSYNC
if ( ( flags & O_SYNC ) = = O_SYNC )
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , " SYNC " ) ;
else {
P_FLAG ( DSYNC ) ;
}
# else
P_FLAG ( SYNC ) ;
# endif
P_FLAG ( TRUNC ) ;
P_FLAG ( WRONLY ) ;
# undef P_FLAG
if ( flags )
printed + = scnprintf ( bf + printed , size - printed , " %s%#x " , printed ? " | " : " " , flags ) ;
return printed ;
}
# define SCA_OPEN_FLAGS syscall_arg__scnprintf_open_flags
trace: Beautify perf_event_open syscall
Syswide tracing and then running 'stat' and 'trace':
$ perf trace -e perf_event_open
1034.649 (0.019 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1, flags: FD_CLOEXEC) = -1 EINVAL Invalid argument
1034.670 (0.008 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
1034.681 (0.007 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
1034.692 (0.007 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
9986.983 (0.014 ms): trace/6139 perf_event_open(attr_uptr: 0x7ffd9c629320, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.026 (0.016 ms): trace/6139 perf_event_open(attr_uptr: 0x37c7e70, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.041 (0.008 ms): trace/6139 perf_event_open(attr_uptr: 0x37c7e70, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.489 (0.092 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.536 (0.044 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 1, group_fd: -1, flags: FD_CLOEXEC) = 4
9987.580 (0.041 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 2, group_fd: -1, flags: FD_CLOEXEC) = 5
9987.620 (0.037 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 3, group_fd: -1, flags: FD_CLOEXEC) = 7
9987.659 (0.035 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, group_fd: -1, flags: FD_CLOEXEC) = 8
9987.692 (0.031 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 1, group_fd: -1, flags: FD_CLOEXEC) = 9
9987.727 (0.032 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 2, group_fd: -1, flags: FD_CLOEXEC) = 10
9987.761 (0.031 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 3, group_fd: -1, flags: FD_CLOEXEC) = 11
Need to intercept perf_copy_attr() with a kprobe or with eBPF...
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-njb105hab2i3t5dexym9lskl@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-12 04:47:54 +03:00
static size_t syscall_arg__scnprintf_perf_flags ( char * bf , size_t size ,
struct syscall_arg * arg )
{
int printed = 0 , flags = arg - > val ;
if ( flags = = 0 )
return 0 ;
# define P_FLAG(n) \
if ( flags & PERF_FLAG_ # # n ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , # n ) ; \
flags & = ~ PERF_FLAG_ # # n ; \
}
P_FLAG ( FD_NO_GROUP ) ;
P_FLAG ( FD_OUTPUT ) ;
P_FLAG ( PID_CGROUP ) ;
P_FLAG ( FD_CLOEXEC ) ;
# undef P_FLAG
if ( flags )
printed + = scnprintf ( bf + printed , size - printed , " %s%#x " , printed ? " | " : " " , flags ) ;
return printed ;
}
# define SCA_PERF_FLAGS syscall_arg__scnprintf_perf_flags
2013-09-12 19:18:56 +04:00
static size_t syscall_arg__scnprintf_eventfd_flags ( char * bf , size_t size ,
struct syscall_arg * arg )
{
int printed = 0 , flags = arg - > val ;
if ( flags = = 0 )
return scnprintf ( bf , size , " NONE " ) ;
# define P_FLAG(n) \
if ( flags & EFD_ # # n ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , # n ) ; \
flags & = ~ EFD_ # # n ; \
}
P_FLAG ( SEMAPHORE ) ;
P_FLAG ( CLOEXEC ) ;
P_FLAG ( NONBLOCK ) ;
# undef P_FLAG
if ( flags )
printed + = scnprintf ( bf + printed , size - printed , " %s%#x " , printed ? " | " : " " , flags ) ;
return printed ;
}
# define SCA_EFD_FLAGS syscall_arg__scnprintf_eventfd_flags
2013-09-23 19:52:04 +04:00
static size_t syscall_arg__scnprintf_pipe_flags ( char * bf , size_t size ,
struct syscall_arg * arg )
{
int printed = 0 , flags = arg - > val ;
# define P_FLAG(n) \
if ( flags & O_ # # n ) { \
printed + = scnprintf ( bf + printed , size - printed , " %s%s " , printed ? " | " : " " , # n ) ; \
flags & = ~ O_ # # n ; \
}
P_FLAG ( CLOEXEC ) ;
P_FLAG ( NONBLOCK ) ;
# undef P_FLAG
if ( flags )
printed + = scnprintf ( bf + printed , size - printed , " %s%#x " , printed ? " | " : " " , flags ) ;
return printed ;
}
# define SCA_PIPE_FLAGS syscall_arg__scnprintf_pipe_flags
2013-09-04 00:17:15 +04:00
static size_t syscall_arg__scnprintf_signum ( char * bf , size_t size , struct syscall_arg * arg )
{
int sig = arg - > val ;
switch ( sig ) {
# define P_SIGNUM(n) case SIG##n: return scnprintf(bf, size, #n)
P_SIGNUM ( HUP ) ;
P_SIGNUM ( INT ) ;
P_SIGNUM ( QUIT ) ;
P_SIGNUM ( ILL ) ;
P_SIGNUM ( TRAP ) ;
P_SIGNUM ( ABRT ) ;
P_SIGNUM ( BUS ) ;
P_SIGNUM ( FPE ) ;
P_SIGNUM ( KILL ) ;
P_SIGNUM ( USR1 ) ;
P_SIGNUM ( SEGV ) ;
P_SIGNUM ( USR2 ) ;
P_SIGNUM ( PIPE ) ;
P_SIGNUM ( ALRM ) ;
P_SIGNUM ( TERM ) ;
P_SIGNUM ( CHLD ) ;
P_SIGNUM ( CONT ) ;
P_SIGNUM ( STOP ) ;
P_SIGNUM ( TSTP ) ;
P_SIGNUM ( TTIN ) ;
P_SIGNUM ( TTOU ) ;
P_SIGNUM ( URG ) ;
P_SIGNUM ( XCPU ) ;
P_SIGNUM ( XFSZ ) ;
P_SIGNUM ( VTALRM ) ;
P_SIGNUM ( PROF ) ;
P_SIGNUM ( WINCH ) ;
P_SIGNUM ( IO ) ;
P_SIGNUM ( PWR ) ;
P_SIGNUM ( SYS ) ;
2014-02-06 05:00:41 +04:00
# ifdef SIGEMT
P_SIGNUM ( EMT ) ;
# endif
# ifdef SIGSTKFLT
P_SIGNUM ( STKFLT ) ;
# endif
# ifdef SIGSWI
P_SIGNUM ( SWI ) ;
# endif
2013-09-04 00:17:15 +04:00
default : break ;
}
return scnprintf ( bf , size , " %#x " , sig ) ;
}
# define SCA_SIGNUM syscall_arg__scnprintf_signum
2014-02-10 21:09:48 +04:00
# if defined(__i386__) || defined(__x86_64__)
/*
* FIXME : Make this available to all arches .
*/
perf trace: Initial beautifier for ioctl's 'cmd' arg
[root@zoo linux]# trace -e ioctl | grep -v "cmd: 0x" | head -10
0.386 ( 0.001 ms): trace/1602 ioctl(fd: 1<pipe:[127057]>, cmd: TCGETS, arg: 0x7fff59fcb4d0 ) = -1 ENOTTY Inappropriate ioctl for device
1459.368 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.586 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.611 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
3740.526 ( 0.002 ms): awk/1612 ioctl(fd: 1<pipe:[128265]>, cmd: TCGETS, arg: 0x7fff4d166b90 ) = -1 ENOTTY Inappropriate ioctl for device
3740.704 ( 0.001 ms): awk/1612 ioctl(fd: 3</proc/meminfo>, cmd: TCGETS, arg: 0x7fff4d1669a0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.550 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.555 ( 0.003 ms): ps/1614 ioctl(fd: 2<socket:[19550]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.558 ( 0.002 ms): ps/1614 ioctl(cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.572 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TCGETS, arg: 0x7fff59176220 ) = -1 ENOTTY Inappropriate ioctl for device
[root@zoo linux]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-afajwap3mr60dfl4qpdl1pxn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-09 00:43:20 +04:00
# define TCGETS 0x5401
static const char * tioctls [ ] = {
" TCGETS " , " TCSETS " , " TCSETSW " , " TCSETSF " , " TCGETA " , " TCSETA " , " TCSETAW " ,
" TCSETAF " , " TCSBRK " , " TCXONC " , " TCFLSH " , " TIOCEXCL " , " TIOCNXCL " ,
" TIOCSCTTY " , " TIOCGPGRP " , " TIOCSPGRP " , " TIOCOUTQ " , " TIOCSTI " ,
" TIOCGWINSZ " , " TIOCSWINSZ " , " TIOCMGET " , " TIOCMBIS " , " TIOCMBIC " ,
" TIOCMSET " , " TIOCGSOFTCAR " , " TIOCSSOFTCAR " , " FIONREAD " , " TIOCLINUX " ,
" TIOCCONS " , " TIOCGSERIAL " , " TIOCSSERIAL " , " TIOCPKT " , " FIONBIO " ,
" TIOCNOTTY " , " TIOCSETD " , " TIOCGETD " , " TCSBRKP " , [ 0x27 ] = " TIOCSBRK " ,
" TIOCCBRK " , " TIOCGSID " , " TCGETS2 " , " TCSETS2 " , " TCSETSW2 " , " TCSETSF2 " ,
" TIOCGRS485 " , " TIOCSRS485 " , " TIOCGPTN " , " TIOCSPTLCK " ,
" TIOCGDEV||TCGETX " , " TCSETX " , " TCSETXF " , " TCSETXW " , " TIOCSIG " ,
" TIOCVHANGUP " , " TIOCGPKT " , " TIOCGPTLCK " , " TIOCGEXCL " ,
[ 0x50 ] = " FIONCLEX " , " FIOCLEX " , " FIOASYNC " , " TIOCSERCONFIG " ,
" TIOCSERGWILD " , " TIOCSERSWILD " , " TIOCGLCKTRMIOS " , " TIOCSLCKTRMIOS " ,
" TIOCSERGSTRUCT " , " TIOCSERGETLSR " , " TIOCSERGETMULTI " , " TIOCSERSETMULTI " ,
" TIOCMIWAIT " , " TIOCGICOUNT " , [ 0x60 ] = " FIOQSIZE " ,
} ;
static DEFINE_STRARRAY_OFFSET ( tioctls , 0x5401 ) ;
2014-02-10 21:09:48 +04:00
# endif /* defined(__i386__) || defined(__x86_64__) */
perf trace: Initial beautifier for ioctl's 'cmd' arg
[root@zoo linux]# trace -e ioctl | grep -v "cmd: 0x" | head -10
0.386 ( 0.001 ms): trace/1602 ioctl(fd: 1<pipe:[127057]>, cmd: TCGETS, arg: 0x7fff59fcb4d0 ) = -1 ENOTTY Inappropriate ioctl for device
1459.368 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.586 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.611 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
3740.526 ( 0.002 ms): awk/1612 ioctl(fd: 1<pipe:[128265]>, cmd: TCGETS, arg: 0x7fff4d166b90 ) = -1 ENOTTY Inappropriate ioctl for device
3740.704 ( 0.001 ms): awk/1612 ioctl(fd: 3</proc/meminfo>, cmd: TCGETS, arg: 0x7fff4d1669a0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.550 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.555 ( 0.003 ms): ps/1614 ioctl(fd: 2<socket:[19550]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.558 ( 0.002 ms): ps/1614 ioctl(cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.572 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TCGETS, arg: 0x7fff59176220 ) = -1 ENOTTY Inappropriate ioctl for device
[root@zoo linux]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-afajwap3mr60dfl4qpdl1pxn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-09 00:43:20 +04:00
2013-09-20 19:13:37 +04:00
# define STRARRAY(arg, name, array) \
. arg_scnprintf = { [ arg ] = SCA_STRARRAY , } , \
. arg_parm = { [ arg ] = & strarray__ # # array , }
2012-09-27 03:05:56 +04:00
static struct syscall_fmt {
const char * name ;
2012-09-27 20:16:00 +04:00
const char * alias ;
2013-09-03 19:20:12 +04:00
size_t ( * arg_scnprintf [ 6 ] ) ( char * bf , size_t size , struct syscall_arg * arg ) ;
2013-09-03 22:50:28 +04:00
void * arg_parm [ 6 ] ;
2012-09-27 03:05:56 +04:00
bool errmsg ;
bool timeout ;
2013-08-26 18:36:30 +04:00
bool hexret ;
2012-09-27 03:05:56 +04:00
} syscall_fmts [ ] = {
2013-09-04 18:42:27 +04:00
{ . name = " access " , . errmsg = true ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* filename */
[ 1 ] = SCA_ACCMODE , /* mode */ } , } ,
2012-09-27 20:16:00 +04:00
{ . name = " arch_prctl " , . errmsg = true , . alias = " prctl " , } ,
2013-08-26 19:29:38 +04:00
{ . name = " brk " , . hexret = true ,
. arg_scnprintf = { [ 0 ] = SCA_HEX , /* brk */ } , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " chdir " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* filename */ } , } ,
{ . name = " chmod " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* filename */ } , } ,
{ . name = " chroot " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* filename */ } , } ,
2013-09-23 05:45:00 +04:00
{ . name = " clock_gettime " , . errmsg = true , STRARRAY ( 0 , clk_id , clockid ) , } ,
2013-09-24 18:04:32 +04:00
{ . name = " close " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_CLOSE_FD , /* fd */ } , } ,
2013-07-30 23:38:23 +04:00
{ . name = " connect " , . errmsg = true , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " creat " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " dup " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " dup2 " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " dup3 " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-20 19:13:37 +04:00
{ . name = " epoll_ctl " , . errmsg = true , STRARRAY ( 1 , op , epoll_ctl_ops ) , } ,
2013-09-12 19:18:56 +04:00
{ . name = " eventfd2 " , . errmsg = true ,
. arg_scnprintf = { [ 1 ] = SCA_EFD_FLAGS , /* flags */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " faccessat " , . errmsg = true ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */
[ 1 ] = SCA_FILENAME , /* filename */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fadvise64 " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fallocate " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fchdir " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fchmod " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fchmodat " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fchown " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fchownat " , . errmsg = true ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* fd */
[ 1 ] = SCA_FILENAME , /* filename */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fcntl " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */
[ 1 ] = SCA_STRARRAY , /* cmd */ } ,
. arg_parm = { [ 1 ] = & strarray__fcntl_cmds , /* cmd */ } , } ,
{ . name = " fdatasync " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-20 18:49:50 +04:00
{ . name = " flock " , . errmsg = true ,
2013-09-24 18:04:32 +04:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */
[ 1 ] = SCA_FLOCK , /* cmd */ } , } ,
{ . name = " fsetxattr " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fstat " , . errmsg = true , . alias = " newfstat " ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fstatat " , . errmsg = true , . alias = " newfstatat " ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */
[ 1 ] = SCA_FILENAME , /* filename */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fstatfs " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " fsync " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " ftruncate " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-02 20:46:44 +04:00
{ . name = " futex " , . errmsg = true ,
. arg_scnprintf = { [ 1 ] = SCA_FUTEX_OP , /* op */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " futimesat " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " getdents " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " getdents64 " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-20 19:13:37 +04:00
{ . name = " getitimer " , . errmsg = true , STRARRAY ( 0 , which , itimers ) , } ,
{ . name = " getrlimit " , . errmsg = true , STRARRAY ( 0 , resource , rlimit_resources ) , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " getxattr " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
{ . name = " inotify_add_watch " , . errmsg = true ,
. arg_scnprintf = { [ 1 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-08-26 19:29:38 +04:00
{ . name = " ioctl " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */
2014-02-10 21:09:48 +04:00
# if defined(__i386__) || defined(__x86_64__)
/*
* FIXME : Make this available to all arches .
*/
perf trace: Initial beautifier for ioctl's 'cmd' arg
[root@zoo linux]# trace -e ioctl | grep -v "cmd: 0x" | head -10
0.386 ( 0.001 ms): trace/1602 ioctl(fd: 1<pipe:[127057]>, cmd: TCGETS, arg: 0x7fff59fcb4d0 ) = -1 ENOTTY Inappropriate ioctl for device
1459.368 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.586 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
1463.611 ( 0.002 ms): inotify_reader/10352 ioctl(fd: 18<anon_inode:inotify>, cmd: FIONREAD, arg: 0x7fb835228bcc ) = 0
3740.526 ( 0.002 ms): awk/1612 ioctl(fd: 1<pipe:[128265]>, cmd: TCGETS, arg: 0x7fff4d166b90 ) = -1 ENOTTY Inappropriate ioctl for device
3740.704 ( 0.001 ms): awk/1612 ioctl(fd: 3</proc/meminfo>, cmd: TCGETS, arg: 0x7fff4d1669a0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.550 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.555 ( 0.003 ms): ps/1614 ioctl(fd: 2<socket:[19550]>, cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.558 ( 0.002 ms): ps/1614 ioctl(cmd: TIOCGWINSZ, arg: 0x7fff591762b0 ) = -1 ENOTTY Inappropriate ioctl for device
3742.572 ( 0.002 ms): ps/1614 ioctl(fd: 1<pipe:[128266]>, cmd: TCGETS, arg: 0x7fff59176220 ) = -1 ENOTTY Inappropriate ioctl for device
[root@zoo linux]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-afajwap3mr60dfl4qpdl1pxn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-09 00:43:20 +04:00
[ 1 ] = SCA_STRHEXARRAY , /* cmd */
[ 2 ] = SCA_HEX , /* arg */ } ,
. arg_parm = { [ 1 ] = & strarray__tioctls , /* cmd */ } , } ,
2014-02-10 21:09:48 +04:00
# else
[ 2 ] = SCA_HEX , /* arg */ } , } ,
# endif
2013-09-04 00:17:15 +04:00
{ . name = " kill " , . errmsg = true ,
. arg_scnprintf = { [ 1 ] = SCA_SIGNUM , /* sig */ } , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " lchown " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* filename */ } , } ,
{ . name = " lgetxattr " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " linkat " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* fd */ } , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " listxattr " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " lseek " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */
[ 2 ] = SCA_STRARRAY , /* whence */ } ,
. arg_parm = { [ 2 ] = & strarray__whences , /* whence */ } , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " lsetxattr " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-08-26 19:21:41 +04:00
{ . name = " lstat " , . errmsg = true , . alias = " newlstat " , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " lsxattr " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-08-23 17:06:41 +04:00
{ . name = " madvise " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_HEX , /* start */
[ 2 ] = SCA_MADV_BHV , /* behavior */ } , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " mkdir " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " mkdirat " , . errmsg = true ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* fd */
[ 1 ] = SCA_FILENAME , /* pathname */ } , } ,
{ . name = " mknod " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* filename */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " mknodat " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* fd */ } , } ,
2013-09-24 07:09:38 +04:00
{ . name = " mlock " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_HEX , /* addr */ } , } ,
{ . name = " mlockall " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_HEX , /* addr */ } , } ,
2013-08-26 19:29:38 +04:00
{ . name = " mmap " , . hexret = true ,
perf trace: Add beautifier for mmap prot parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.984 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd62ae000
1.114 ( 0.016 ms): mmap(addr: 0, len: 125100, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xd628f000
1.252 ( 0.020 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: 2050, fd: 3, off: 0) = 0xc1600000
1.282 ( 0.024 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.315 ( 0.026 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: 2066, fd: 3, off: 1757184) = 0xc19ad000
1.352 ( 0.017 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: 50, fd: 4294967295, off: 0) = 0xc19b3000
1.415 ( 0.011 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628e000
1.440 ( 0.011 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628c000
1.569 ( 0.019 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.591 ( 0.017 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.616 ( 0.016 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
2.105 ( 0.018 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xcfe9c000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-21 00:44:42 +04:00
. arg_scnprintf = { [ 0 ] = SCA_HEX , /* addr */
perf trace: Add beautifier for mmap flags parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.992 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa60be000
1.108 ( 0.012 ms): mmap(addr: 0, len: 125100, prot: READ, flags: PRIVATE, fd: 3, off: 0 ) = 0xa609f000
1.209 ( 0.014 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3, off: 0) = 0xc1600000
1.232 ( 0.018 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.255 ( 0.018 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3, off: 1757184) = 0xc19ad000
1.281 ( 0.011 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: 4294967295, off: 0) = 0xc19b3000
1.328 ( 0.008 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609e000
1.346 ( 0.008 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: 4294967295, off: 0) = 0xa609c000
1.443 ( 0.013 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.459 ( 0.011 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.477 ( 0.011 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
1.855 ( 0.013 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: PRIVATE, fd: 3, off: 0) = 0x9fcac000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-23 17:48:33 +04:00
[ 2 ] = SCA_MMAP_PROT , /* prot */
2013-11-12 10:24:59 +04:00
[ 3 ] = SCA_MMAP_FLAGS , /* flags */
[ 4 ] = SCA_FD , /* fd */ } , } ,
2013-08-26 19:29:38 +04:00
{ . name = " mprotect " , . errmsg = true ,
perf trace: Add beautifier for mmap prot parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.984 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd62ae000
1.114 ( 0.016 ms): mmap(addr: 0, len: 125100, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xd628f000
1.252 ( 0.020 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: 2050, fd: 3, off: 0) = 0xc1600000
1.282 ( 0.024 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.315 ( 0.026 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: 2066, fd: 3, off: 1757184) = 0xc19ad000
1.352 ( 0.017 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: 50, fd: 4294967295, off: 0) = 0xc19b3000
1.415 ( 0.011 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628e000
1.440 ( 0.011 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628c000
1.569 ( 0.019 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.591 ( 0.017 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.616 ( 0.016 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
2.105 ( 0.018 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xcfe9c000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-21 00:44:42 +04:00
. arg_scnprintf = { [ 0 ] = SCA_HEX , /* start */
[ 2 ] = SCA_MMAP_PROT , /* prot */ } , } ,
{ . name = " mremap " , . hexret = true ,
. arg_scnprintf = { [ 0 ] = SCA_HEX , /* addr */
perf trace: Add beautifier for mremap flags param
~/devel/kernel/tools/perf(branch:master*) » sudo ./perf trace ~/mremap_test
0.543 ( 0.003 ms): mprotect(start: 0x600000, len: 4096, prot: READ ) = 0
0.550 ( 0.003 ms): mprotect(start: 0x7f441260d000, len: 4096, prot: READ) = 0
0.561 ( 0.010 ms): munmap(addr: 0x7f44125e2000, len: 165572 ) = 0
0.595 ( 0.012 ms): mmap(len: 12288, prot: READ|WRITE, flags: SHARED|ANONYMOUS|LOCKED, fd: -1) = 0x12608000
0.603 ( 0.006 ms): mremap(addr: 0x7f4412608000, old_len: 4096, new_len: 4096, flags: MAYMOVE|FIXED, new_addr: 0x7f16da295000) = 0xda295000
0.608 ( 0.003 ms): mremap(addr: 0x7f441260a000, old_len: 4096, new_len: 4096, flags: MAYMOVE|FIXED, new_addr: 0x7f16da297000) = 0xda297000
0.612 ( 0.003 ms): mremap(addr: 0x7f4412609000, old_len: 4096, new_len: 4096, flags: MAYMOVE|FIXED, new_addr: 0x7f16da296000) = 0xda296000
0.619 ( 0.000 ms): exit_group(
Signed-off-by: Alex Snast <asnast@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1407944560-26924-1-git-send-email-asnast@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2014-08-13 19:42:40 +04:00
[ 3 ] = SCA_MREMAP_FLAGS , /* flags */
perf trace: Add beautifier for mmap prot parm
[root@zoo ~]# perf trace -e mmap,mprotect sleep 1
0.984 ( 0.015 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd62ae000
1.114 ( 0.016 ms): mmap(addr: 0, len: 125100, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xd628f000
1.252 ( 0.020 ms): mmap(addr: 0x33c1600000, len: 3896312, prot: EXEC|READ, flags: 2050, fd: 3, off: 0) = 0xc1600000
1.282 ( 0.024 ms): mprotect(start: 0x33c17ad000, len: 2097152, prot: NONE ) = 0
1.315 ( 0.026 ms): mmap(addr: 0x33c19ad000, len: 24576, prot: READ|WRITE, flags: 2066, fd: 3, off: 1757184) = 0xc19ad000
1.352 ( 0.017 ms): mmap(addr: 0x33c19b3000, len: 17400, prot: READ|WRITE, flags: 50, fd: 4294967295, off: 0) = 0xc19b3000
1.415 ( 0.011 ms): mmap(addr: 0, len: 4096, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628e000
1.440 ( 0.011 ms): mmap(addr: 0, len: 8192, prot: READ|WRITE, flags: 34, fd: 4294967295, off: 0) = 0xd628c000
1.569 ( 0.019 ms): mprotect(start: 0x606000, len: 4096, prot: READ ) = 0
1.591 ( 0.017 ms): mprotect(start: 0x33c19ad000, len: 16384, prot: READ ) = 0
1.616 ( 0.016 ms): mprotect(start: 0x33c1420000, len: 4096, prot: READ ) = 0
2.105 ( 0.018 ms): mmap(addr: 0, len: 104789808, prot: READ, flags: 2, fd: 3, off: 0 ) = 0xcfe9c000
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-q1ubhdd9wigxneam616ggdsn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-08-21 00:44:42 +04:00
[ 4 ] = SCA_HEX , /* new_addr */ } , } ,
2013-09-24 07:09:38 +04:00
{ . name = " munlock " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_HEX , /* addr */ } , } ,
2013-08-26 19:29:38 +04:00
{ . name = " munmap " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_HEX , /* addr */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " name_to_handle_at " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " newfstatat " , . errmsg = true ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */
[ 1 ] = SCA_FILENAME , /* filename */ } , } ,
2013-09-02 23:22:31 +04:00
{ . name = " open " , . errmsg = true ,
perf trace: Deref sys_enter pointer args with contents from probe:vfs_getname
To work like strace and dereference syscall pointer args we need to
insert probes (or tracepoints) right after we copy those bytes from
userspace.
Since we're formatting the syscall args at raw_syscalls:sys_enter time,
we need to have a formatter that just stores the position where, later,
when we get the probe:vfs_getname, we can insert the pointer contents.
Now, if a probe:vfs_getname with this format is in place:
# perf probe -l
probe:vfs_getname (on getname_flags:72@/home/git/linux/fs/namei.c with pathname)
That was, in this case, put in place with:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
Then 'perf trace' will notice that and do the pointer -> contents
expansion:
# trace -e open touch /tmp/bla
0.165 (0.010 ms): touch/17752 open(filename: /etc/ld.so.cache, flags: CLOEXEC) = 3
0.195 (0.011 ms): touch/17752 open(filename: /lib64/libc.so.6, flags: CLOEXEC) = 3
0.512 (0.012 ms): touch/17752 open(filename: /usr/lib/locale/locale-archive, flags: CLOEXEC) = 3
0.582 (0.012 ms): touch/17752 open(filename: /tmp/bla, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
#
Roughly equivalent to strace's output:
# strace -rT -e open touch /tmp/bla
0.000000 open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3 <0.000039>
0.000317 open("/lib64/libc.so.6", O_RDONLY|O_CLOEXEC) = 3 <0.000102>
0.001461 open("/usr/lib/locale/locale-archive", O_RDONLY|O_CLOEXEC) = 3 <0.000072>
0.000405 open("/tmp/bla", O_WRONLY|O_CREAT|O_NOCTTY|O_NONBLOCK, 0666) = 3 <0.000055>
0.000641 +++ exited with 0 +++
#
Now we need to either look for at all syscalls that are marked as
pointers and have some well known names ("filename", "pathname", etc)
and set the arg formatter to the one used for the "open" syscall in this
patch.
This implementation works for syscalls with just a string being copied
from userspace, for matching syscalls with more than one string being
copied via the same probe/trace point (vfs_getname) we need to extend
the vfs_getname probe spec to include the pointer too, but there are
some problems with that in 'perf probe' or the kernel kprobes code, need
to investigate before considering supporting multiple strings per
syscall.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-xvuwx6nuj8cf389kf9s2ue2s@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-08-05 04:30:09 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* filename */
[ 1 ] = SCA_OPEN_FLAGS , /* flags */ } , } ,
2013-09-02 23:40:40 +04:00
{ . name = " open_by_handle_at " , . errmsg = true ,
2013-09-24 18:04:32 +04:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */
[ 2 ] = SCA_OPEN_FLAGS , /* flags */ } , } ,
2013-09-02 23:40:40 +04:00
{ . name = " openat " , . errmsg = true ,
2013-09-24 18:04:32 +04:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
[ 1 ] = SCA_FILENAME , /* filename */
2013-09-24 18:04:32 +04:00
[ 2 ] = SCA_OPEN_FLAGS , /* flags */ } , } ,
trace: Beautify perf_event_open syscall
Syswide tracing and then running 'stat' and 'trace':
$ perf trace -e perf_event_open
1034.649 (0.019 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1, flags: FD_CLOEXEC) = -1 EINVAL Invalid argument
1034.670 (0.008 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
1034.681 (0.007 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
1034.692 (0.007 ms): perf/6133 perf_event_open(attr_uptr: 0x36f0360, pid: 16134, cpu: -1, group_fd: -1) = -1 EINVAL Invalid argument
9986.983 (0.014 ms): trace/6139 perf_event_open(attr_uptr: 0x7ffd9c629320, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.026 (0.016 ms): trace/6139 perf_event_open(attr_uptr: 0x37c7e70, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.041 (0.008 ms): trace/6139 perf_event_open(attr_uptr: 0x37c7e70, pid: -1, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.489 (0.092 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, group_fd: -1, flags: FD_CLOEXEC) = 3
9987.536 (0.044 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 1, group_fd: -1, flags: FD_CLOEXEC) = 4
9987.580 (0.041 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 2, group_fd: -1, flags: FD_CLOEXEC) = 5
9987.620 (0.037 ms): trace/6139 perf_event_open(attr_uptr: 0x3795ee0, pid: 16140, cpu: 3, group_fd: -1, flags: FD_CLOEXEC) = 7
9987.659 (0.035 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, group_fd: -1, flags: FD_CLOEXEC) = 8
9987.692 (0.031 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 1, group_fd: -1, flags: FD_CLOEXEC) = 9
9987.727 (0.032 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 2, group_fd: -1, flags: FD_CLOEXEC) = 10
9987.761 (0.031 ms): trace/6139 perf_event_open(attr_uptr: 0x37975d0, pid: 16140, cpu: 3, group_fd: -1, flags: FD_CLOEXEC) = 11
Need to intercept perf_copy_attr() with a kprobe or with eBPF...
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-njb105hab2i3t5dexym9lskl@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-12 04:47:54 +03:00
{ . name = " perf_event_open " , . errmsg = true ,
. arg_scnprintf = { [ 1 ] = SCA_INT , /* pid */
[ 2 ] = SCA_INT , /* cpu */
[ 3 ] = SCA_FD , /* group_fd */
[ 4 ] = SCA_PERF_FLAGS , /* flags */ } , } ,
2013-09-23 19:52:04 +04:00
{ . name = " pipe2 " , . errmsg = true ,
. arg_scnprintf = { [ 1 ] = SCA_PIPE_FLAGS , /* flags */ } , } ,
2012-09-27 20:16:00 +04:00
{ . name = " poll " , . errmsg = true , . timeout = true , } ,
{ . name = " ppoll " , . errmsg = true , . timeout = true , } ,
2013-09-24 18:04:32 +04:00
{ . name = " pread " , . errmsg = true , . alias = " pread64 " ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " preadv " , . errmsg = true , . alias = " pread " ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-20 19:13:37 +04:00
{ . name = " prlimit64 " , . errmsg = true , STRARRAY ( 1 , resource , rlimit_resources ) , } ,
2013-09-24 18:04:32 +04:00
{ . name = " pwrite " , . errmsg = true , . alias = " pwrite64 " ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " pwritev " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " read " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " readlink " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* path */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " readlinkat " , . errmsg = true ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */
[ 1 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " readv " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
perf trace: Beautify send/recv syscall 'flags' arg
[root@sandy ~]# perf trace -a -e recvmmsg,recvmsg,recvfrom,sendto,sendmsg,sendmmsg
6.901 (0.002 ms): 589 recvmsg(fd: 51, msg: 0x7fff35673420, flags: CMSG_CLOEXEC) = -1 EAGAIN Resource temporarily unavailable
6.966 (0.008 ms): 589 sendmsg(fd: 50, msg: 0x7fff35673230, flags: NOSIGNAL ) = 961
6.984 (0.004 ms): 979 sendmsg(fd: 3, msg: 0x7fff5b484940, flags: NOSIGNAL ) = 945
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-h25k5k50nac0ej5cl5iwgvae@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-12 18:54:48 +04:00
{ . name = " recvfrom " , . errmsg = true ,
. arg_scnprintf = { [ 3 ] = SCA_MSG_FLAGS , /* flags */ } , } ,
{ . name = " recvmmsg " , . errmsg = true ,
. arg_scnprintf = { [ 3 ] = SCA_MSG_FLAGS , /* flags */ } , } ,
{ . name = " recvmsg " , . errmsg = true ,
. arg_scnprintf = { [ 2 ] = SCA_MSG_FLAGS , /* flags */ } , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " removexattr " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " renameat " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */ } , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " rmdir " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-09-04 00:17:15 +04:00
{ . name = " rt_sigaction " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_SIGNUM , /* sig */ } , } ,
2013-09-20 19:13:37 +04:00
{ . name = " rt_sigprocmask " , . errmsg = true , STRARRAY ( 0 , how , sighow ) , } ,
2013-09-04 00:17:15 +04:00
{ . name = " rt_sigqueueinfo " , . errmsg = true ,
. arg_scnprintf = { [ 1 ] = SCA_SIGNUM , /* sig */ } , } ,
{ . name = " rt_tgsigqueueinfo " , . errmsg = true ,
. arg_scnprintf = { [ 2 ] = SCA_SIGNUM , /* sig */ } , } ,
2012-09-27 20:16:00 +04:00
{ . name = " select " , . errmsg = true , . timeout = true , } ,
perf trace: Beautify send/recv syscall 'flags' arg
[root@sandy ~]# perf trace -a -e recvmmsg,recvmsg,recvfrom,sendto,sendmsg,sendmmsg
6.901 (0.002 ms): 589 recvmsg(fd: 51, msg: 0x7fff35673420, flags: CMSG_CLOEXEC) = -1 EAGAIN Resource temporarily unavailable
6.966 (0.008 ms): 589 sendmsg(fd: 50, msg: 0x7fff35673230, flags: NOSIGNAL ) = 961
6.984 (0.004 ms): 979 sendmsg(fd: 3, msg: 0x7fff5b484940, flags: NOSIGNAL ) = 945
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-h25k5k50nac0ej5cl5iwgvae@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-12 18:54:48 +04:00
{ . name = " sendmmsg " , . errmsg = true ,
. arg_scnprintf = { [ 3 ] = SCA_MSG_FLAGS , /* flags */ } , } ,
{ . name = " sendmsg " , . errmsg = true ,
. arg_scnprintf = { [ 2 ] = SCA_MSG_FLAGS , /* flags */ } , } ,
{ . name = " sendto " , . errmsg = true ,
. arg_scnprintf = { [ 3 ] = SCA_MSG_FLAGS , /* flags */ } , } ,
2013-09-20 19:13:37 +04:00
{ . name = " setitimer " , . errmsg = true , STRARRAY ( 0 , which , itimers ) , } ,
{ . name = " setrlimit " , . errmsg = true , STRARRAY ( 0 , resource , rlimit_resources ) , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " setxattr " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " shutdown " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-04 17:27:41 +04:00
{ . name = " socket " , . errmsg = true ,
2013-09-04 18:00:44 +04:00
. arg_scnprintf = { [ 0 ] = SCA_STRARRAY , /* family */
[ 1 ] = SCA_SK_TYPE , /* type */ } ,
2013-09-20 19:24:20 +04:00
. arg_parm = { [ 0 ] = & strarray__socket_families , /* family */ } , } ,
{ . name = " socketpair " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_STRARRAY , /* family */
[ 1 ] = SCA_SK_TYPE , /* type */ } ,
2013-09-04 17:27:41 +04:00
. arg_parm = { [ 0 ] = & strarray__socket_families , /* family */ } , } ,
2012-09-27 20:16:00 +04:00
{ . name = " stat " , . errmsg = true , . alias = " newstat " , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " statfs " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* pathname */ } , } ,
{ . name = " swapoff " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* specialfile */ } , } ,
{ . name = " swapon " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* specialfile */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " symlinkat " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */ } , } ,
2013-09-04 00:17:15 +04:00
{ . name = " tgkill " , . errmsg = true ,
. arg_scnprintf = { [ 2 ] = SCA_SIGNUM , /* sig */ } , } ,
{ . name = " tkill " , . errmsg = true ,
. arg_scnprintf = { [ 1 ] = SCA_SIGNUM , /* sig */ } , } ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
{ . name = " truncate " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* path */ } , } ,
2013-08-26 19:21:41 +04:00
{ . name = " uname " , . errmsg = true , . alias = " newuname " , } ,
2013-09-24 18:04:32 +04:00
{ . name = " unlinkat " , . errmsg = true ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dfd */
[ 1 ] = SCA_FILENAME , /* pathname */ } , } ,
{ . name = " utime " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* filename */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " utimensat " , . errmsg = true ,
perf trace: Use vfs_getname syscall arg beautifier in more syscalls
Those were covered and tested in this cset:
access, chdir, chmod, chown, chroot, creat, getxattr,
inotify_add_watch, lchown, lgetxattr, listxattr,
lsetxattr, mkdir, mkdirat, mknod, rmdir, faccessat,
newfstatat, openat, readlink, readlinkat, removexattr,
setxattr, statfs, swapon, swapoff, truncate, unlinkat,
utime, utimes, utimensat.
E.g.:
# trace -e statfs,access,mkdir mkdir /tmp/bla
0.285 (0.020 ms): mkdir/2799 access(filename: /etc/ld.so.preload, mode: R ) = -1 ENOENT No such file or directory
1.070 (0.032 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc930) = 0
1.087 (0.013 ms): mkdir/2799 statfs(pathname: /sys/fs/selinux, buf: 0x7ffeafbdc820) = 0
1.189 (0.014 ms): mkdir/2799 access(filename: /etc/selinux/config ) = 0
1.905 (0.610 ms): mkdir/2799 mkdir(pathname: /tmp/bla, mode: 511 ) = 0
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-wbqtnlktquun3wtpjdz3okul@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
and an empty message aborts the commit.
2015-08-05 05:31:25 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FDAT , /* dirfd */
[ 1 ] = SCA_FILENAME , /* filename */ } , } ,
{ . name = " utimes " , . errmsg = true ,
. arg_scnprintf = { [ 0 ] = SCA_FILENAME , /* filename */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " write " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2013-09-24 18:04:32 +04:00
{ . name = " writev " , . errmsg = true ,
2014-12-17 23:24:45 +03:00
. arg_scnprintf = { [ 0 ] = SCA_FD , /* fd */ } , } ,
2012-09-27 03:05:56 +04:00
} ;
static int syscall_fmt__cmp ( const void * name , const void * fmtp )
{
const struct syscall_fmt * fmt = fmtp ;
return strcmp ( name , fmt - > name ) ;
}
static struct syscall_fmt * syscall_fmt__find ( const char * name )
{
const int nmemb = ARRAY_SIZE ( syscall_fmts ) ;
return bsearch ( name , syscall_fmts , nmemb , sizeof ( struct syscall_fmt ) , syscall_fmt__cmp ) ;
}
struct syscall {
struct event_format * tp_format ;
2015-03-20 23:46:53 +03:00
int nr_args ;
struct format_field * args ;
2012-09-27 03:05:56 +04:00
const char * name ;
2014-06-17 21:29:24 +04:00
bool is_exit ;
2012-09-27 03:05:56 +04:00
struct syscall_fmt * fmt ;
2013-09-03 19:20:12 +04:00
size_t ( * * arg_scnprintf ) ( char * bf , size_t size , struct syscall_arg * arg ) ;
2013-09-03 22:50:28 +04:00
void * * arg_parm ;
2012-09-27 03:05:56 +04:00
} ;
perf trace: Add an event duration column
# perf trace usleep 1 | tail -10
0.453 ( 0.002 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763342848
0.456 ( 0.001 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763346944
0.459 ( 0.001 ms): arch_prctl(option: 4098, arg2: 140126839658240, arg3: 140126839652352, arg4: 34, arg5: 4294967295) = 0
0.473 ( 0.003 ms): mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
0.477 ( 0.003 ms): mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
0.483 ( 0.004 ms): munmap(addr: 140126839664640, len: 91882 ) = 0
0.540 ( 0.001 ms): brk(brk: 0 ) = 31928320
0.542 ( 0.002 ms): brk(brk: 32063488 ) = 32063488
1.456 ( 0.901 ms): nanosleep(rqtp: 140735472817168, rmtp: 0 ) = 0
1.462 ( 0.000 ms): exit_group(error_code: 0
#
This also comes from the tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-g9akh5hjw2kvjerpo9xror6f@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-24 23:24:47 +04:00
static size_t fprintf_duration ( unsigned long t , FILE * fp )
{
double duration = ( double ) t / NSEC_PER_MSEC ;
size_t printed = fprintf ( fp , " ( " ) ;
if ( duration > = 1.0 )
printed + = color_fprintf ( fp , PERF_COLOR_RED , " %6.3f ms " , duration ) ;
else if ( duration > = 0.01 )
printed + = color_fprintf ( fp , PERF_COLOR_YELLOW , " %6.3f ms " , duration ) ;
else
printed + = color_fprintf ( fp , PERF_COLOR_NORMAL , " %6.3f ms " , duration ) ;
2013-08-19 19:01:10 +04:00
return printed + fprintf ( fp , " ): " ) ;
perf trace: Add an event duration column
# perf trace usleep 1 | tail -10
0.453 ( 0.002 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763342848
0.456 ( 0.001 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763346944
0.459 ( 0.001 ms): arch_prctl(option: 4098, arg2: 140126839658240, arg3: 140126839652352, arg4: 34, arg5: 4294967295) = 0
0.473 ( 0.003 ms): mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
0.477 ( 0.003 ms): mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
0.483 ( 0.004 ms): munmap(addr: 140126839664640, len: 91882 ) = 0
0.540 ( 0.001 ms): brk(brk: 0 ) = 31928320
0.542 ( 0.002 ms): brk(brk: 32063488 ) = 32063488
1.456 ( 0.901 ms): nanosleep(rqtp: 140735472817168, rmtp: 0 ) = 0
1.462 ( 0.000 ms): exit_group(error_code: 0
#
This also comes from the tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-g9akh5hjw2kvjerpo9xror6f@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-24 23:24:47 +04:00
}
perf trace: Deref sys_enter pointer args with contents from probe:vfs_getname
To work like strace and dereference syscall pointer args we need to
insert probes (or tracepoints) right after we copy those bytes from
userspace.
Since we're formatting the syscall args at raw_syscalls:sys_enter time,
we need to have a formatter that just stores the position where, later,
when we get the probe:vfs_getname, we can insert the pointer contents.
Now, if a probe:vfs_getname with this format is in place:
# perf probe -l
probe:vfs_getname (on getname_flags:72@/home/git/linux/fs/namei.c with pathname)
That was, in this case, put in place with:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
Then 'perf trace' will notice that and do the pointer -> contents
expansion:
# trace -e open touch /tmp/bla
0.165 (0.010 ms): touch/17752 open(filename: /etc/ld.so.cache, flags: CLOEXEC) = 3
0.195 (0.011 ms): touch/17752 open(filename: /lib64/libc.so.6, flags: CLOEXEC) = 3
0.512 (0.012 ms): touch/17752 open(filename: /usr/lib/locale/locale-archive, flags: CLOEXEC) = 3
0.582 (0.012 ms): touch/17752 open(filename: /tmp/bla, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
#
Roughly equivalent to strace's output:
# strace -rT -e open touch /tmp/bla
0.000000 open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3 <0.000039>
0.000317 open("/lib64/libc.so.6", O_RDONLY|O_CLOEXEC) = 3 <0.000102>
0.001461 open("/usr/lib/locale/locale-archive", O_RDONLY|O_CLOEXEC) = 3 <0.000072>
0.000405 open("/tmp/bla", O_WRONLY|O_CREAT|O_NOCTTY|O_NONBLOCK, 0666) = 3 <0.000055>
0.000641 +++ exited with 0 +++
#
Now we need to either look for at all syscalls that are marked as
pointers and have some well known names ("filename", "pathname", etc)
and set the arg formatter to the one used for the "open" syscall in this
patch.
This implementation works for syscalls with just a string being copied
from userspace, for matching syscalls with more than one string being
copied via the same probe/trace point (vfs_getname) we need to extend
the vfs_getname probe spec to include the pointer too, but there are
some problems with that in 'perf probe' or the kernel kprobes code, need
to investigate before considering supporting multiple strings per
syscall.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-xvuwx6nuj8cf389kf9s2ue2s@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-08-05 04:30:09 +03:00
/**
* filename . ptr : The filename char pointer that will be vfs_getname ' d
* filename . entry_str_pos : Where to insert the string translated from
* filename . ptr by the vfs_getname tracepoint / kprobe .
*/
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
struct thread_trace {
u64 entry_time ;
u64 exit_time ;
bool entry_pending ;
2012-10-18 00:09:46 +04:00
unsigned long nr_events ;
2014-07-08 22:05:16 +04:00
unsigned long pfmaj , pfmin ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
char * entry_str ;
2012-10-18 00:13:12 +04:00
double runtime_ms ;
perf trace: Deref sys_enter pointer args with contents from probe:vfs_getname
To work like strace and dereference syscall pointer args we need to
insert probes (or tracepoints) right after we copy those bytes from
userspace.
Since we're formatting the syscall args at raw_syscalls:sys_enter time,
we need to have a formatter that just stores the position where, later,
when we get the probe:vfs_getname, we can insert the pointer contents.
Now, if a probe:vfs_getname with this format is in place:
# perf probe -l
probe:vfs_getname (on getname_flags:72@/home/git/linux/fs/namei.c with pathname)
That was, in this case, put in place with:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
Then 'perf trace' will notice that and do the pointer -> contents
expansion:
# trace -e open touch /tmp/bla
0.165 (0.010 ms): touch/17752 open(filename: /etc/ld.so.cache, flags: CLOEXEC) = 3
0.195 (0.011 ms): touch/17752 open(filename: /lib64/libc.so.6, flags: CLOEXEC) = 3
0.512 (0.012 ms): touch/17752 open(filename: /usr/lib/locale/locale-archive, flags: CLOEXEC) = 3
0.582 (0.012 ms): touch/17752 open(filename: /tmp/bla, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
#
Roughly equivalent to strace's output:
# strace -rT -e open touch /tmp/bla
0.000000 open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3 <0.000039>
0.000317 open("/lib64/libc.so.6", O_RDONLY|O_CLOEXEC) = 3 <0.000102>
0.001461 open("/usr/lib/locale/locale-archive", O_RDONLY|O_CLOEXEC) = 3 <0.000072>
0.000405 open("/tmp/bla", O_WRONLY|O_CREAT|O_NOCTTY|O_NONBLOCK, 0666) = 3 <0.000055>
0.000641 +++ exited with 0 +++
#
Now we need to either look for at all syscalls that are marked as
pointers and have some well known names ("filename", "pathname", etc)
and set the arg formatter to the one used for the "open" syscall in this
patch.
This implementation works for syscalls with just a string being copied
from userspace, for matching syscalls with more than one string being
copied via the same probe/trace point (vfs_getname) we need to extend
the vfs_getname probe spec to include the pointer too, but there are
some problems with that in 'perf probe' or the kernel kprobes code, need
to investigate before considering supporting multiple strings per
syscall.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-xvuwx6nuj8cf389kf9s2ue2s@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-08-05 04:30:09 +03:00
struct {
unsigned long ptr ;
int entry_str_pos ;
} filename ;
2013-09-24 18:04:32 +04:00
struct {
int max ;
char * * table ;
} paths ;
2013-10-09 07:26:53 +04:00
struct intlist * syscall_stats ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
} ;
static struct thread_trace * thread_trace__new ( void )
{
2013-09-24 18:04:32 +04:00
struct thread_trace * ttrace = zalloc ( sizeof ( struct thread_trace ) ) ;
if ( ttrace )
ttrace - > paths . max = - 1 ;
2013-10-09 07:26:53 +04:00
ttrace - > syscall_stats = intlist__new ( NULL ) ;
2013-09-24 18:04:32 +04:00
return ttrace ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
}
2013-08-19 19:01:10 +04:00
static struct thread_trace * thread__trace ( struct thread * thread , FILE * fp )
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
{
2012-10-18 00:09:46 +04:00
struct thread_trace * ttrace ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
if ( thread = = NULL )
goto fail ;
2014-10-06 04:46:03 +04:00
if ( thread__priv ( thread ) = = NULL )
thread__set_priv ( thread , thread_trace__new ( ) ) ;
2014-12-17 23:24:45 +03:00
2014-10-06 04:46:03 +04:00
if ( thread__priv ( thread ) = = NULL )
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
goto fail ;
2014-10-06 04:46:03 +04:00
ttrace = thread__priv ( thread ) ;
2012-10-18 00:09:46 +04:00
+ + ttrace - > nr_events ;
return ttrace ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
fail :
2013-08-19 19:01:10 +04:00
color_fprintf ( fp , PERF_COLOR_RED ,
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
" WARNING: not enough memory, dropping samples! \n " ) ;
return NULL ;
}
2014-06-26 20:14:25 +04:00
# define TRACE_PFMAJ (1 << 0)
# define TRACE_PFMIN (1 << 1)
2015-08-05 04:17:29 +03:00
static const size_t trace__entry_str_size = 2048 ;
2012-09-27 03:05:56 +04:00
struct trace {
2013-08-19 19:01:10 +04:00
struct perf_tool tool ;
2013-09-28 01:06:19 +04:00
struct {
int machine ;
int open_id ;
} audit ;
2012-09-27 03:05:56 +04:00
struct {
int max ;
struct syscall * table ;
2015-07-03 00:24:51 +03:00
struct {
2015-07-03 00:28:11 +03:00
struct perf_evsel * sys_enter ,
* sys_exit ;
2015-07-03 00:24:51 +03:00
} events ;
2012-09-27 03:05:56 +04:00
} syscalls ;
2013-12-19 21:43:45 +04:00
struct record_opts opts ;
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
struct perf_evlist * evlist ;
2013-09-28 23:13:00 +04:00
struct machine * host ;
perf trace: Handle multiple threads better wrt syscalls being intermixed
$ trace time taskset -c 0 usleep 1
0.845 ( 0.021 ms): time/16722 wait4(upid: 4294967295, stat_addr: 0x7fff17f443d4, ru: 0x7fff17f44438 ) ...
0.865 ( 0.008 ms): time/16723 execve(arg0: 140733595272004, arg1: 140733595272720, arg2: 140733595272768, arg3: 139755107218496, arg4: 7307199665339051828, arg5: 3) = -2
2.395 ( 1.523 ms): taskset/16723 execve(arg0: 140733595272013, arg1: 140733595272720, arg2: 140733595272768, arg3: 139755107218496, arg4: 7307199665339051828, arg5: 3) = 0
2.411 ( 0.002 ms): taskset/16723 brk( ) = 0x1915000
3.300 ( 0.058 ms): usleep/16723 nanosleep(rqtp: 0x7ffff4ada190 ) = 0
<SNIP>
3.305 ( 0.000 ms): usleep/16723 exit_group(
3.363 ( 2.539 ms): time/16722 ... [continued]: wait4()) = 16723
3.366 ( 0.001 ms): time/16722 rt_sigaction(sig: INT, act: 0x7fff17f44160, oact: 0x7fff17f44200, sigsetsize: 8) = 0
We we're not seeing this line:
0.845 ( 0.021 ms): time/16722 wait4(upid: 4294967295, stat_addr: 0x7fff17f443d4, ru: 0x7fff17f44438 ) ...
just the one when it finishes:
3.363 ( 2.539 ms): time/16722 ... [continued]: wait4()) = 16723
Still some issues left till we move to ordered_samples when multiple
CPUs/threads are involved...
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-zq9x30a1ky3djqewqn2v3ja3@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-13 19:22:21 +03:00
struct thread * current ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
u64 base_time ;
2013-08-19 19:01:10 +04:00
FILE * output ;
2012-10-18 00:09:46 +04:00
unsigned long nr_events ;
2013-08-21 19:56:21 +04:00
struct strlist * ev_qualifier ;
2015-07-03 00:28:11 +03:00
struct {
size_t nr ;
int * entries ;
} ev_qualifier_ids ;
2013-09-28 01:06:19 +04:00
const char * last_vfs_getname ;
2013-08-29 08:29:53 +04:00
struct intlist * tid_list ;
struct intlist * pid_list ;
perf trace: Introduce --filter-pids
When tracing in X we get event loops due to the tracing activity, i.e.
updates to a gnome-terminal that generate syscalls for X.org, etc.
To get a more useful view of what is happening, syscall wise, system
wide, we need to filter those, like in:
# ps ax|egrep '981|2296|1519' | grep -v egrep
981 tty1 Ss+ 5:40 /usr/bin/Xorg :0 -background none ...
1519 ? Sl 2:22 /usr/bin/gnome-shell
2296 ? Sl 4:16 /usr/libexec/gnome-terminal-server
#
# trace -e write --filter-pids 981,2296,1519
0.385 ( 0.021 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 136) = 136
0.922 ( 0.014 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 140) = 140
5006.525 ( 0.029 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 136) = 136
5007.235 ( 0.023 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 140) = 140
5177.646 ( 0.018 ms): rtkit-daemon/782 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7f7eea70be88, count: 8) = 8
8314.497 ( 0.004 ms): gsd-locate-poi/2084 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7fffe96af7b0, count: 8) = 8
8314.518 ( 0.002 ms): gsd-locate-poi/2084 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7fffe96af0e0, count: 8) = 8
^C#
When this option is used the tracer pid is also filtered.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-f5qmiyy7c0uxdm21ncatpeek@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-21 22:36:52 +03:00
struct {
size_t nr ;
pid_t * entries ;
} filter_pids ;
2014-01-06 22:43:02 +04:00
double duration_filter ;
double runtime_ms ;
struct {
u64 vfs_getname ,
proc_getname ;
} stats ;
bool not_ev_qualifier ;
bool live ;
bool full_time ;
2012-10-18 00:13:12 +04:00
bool sched ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
bool multiple_threads ;
2013-10-09 07:26:53 +04:00
bool summary ;
2013-11-12 20:31:15 +04:00
bool summary_only ;
perf trace: Add option to show process COMM
Enabled by default, disable with --no-comm, e.g.:
181.821 (0.001 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: PEEK|TRUNC|CMSG_CLOEXEC ) = 20
181.824 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.825 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.834 (0.002 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: CMSG_CLOEXEC ) = 20
181.836 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.838 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.705 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 1256
181.710 (0.002 ms): evolution-addr/10924 geteuid( ) = 1000
181.712 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.727 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 1256
181.731 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.734 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.908 (0.002 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 20
181.913 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.915 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.930 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 20
181.934 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.937 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
220.718 (0.010 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.741 (0.000 ms): dbus-daemon/10711 ... [continued]: epoll_wait()) = 1
220.759 (0.004 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.780 (0.002 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.788 (0.001 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = -1 EAGAIN Resource temporarily unavailable
220.760 (0.004 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.771 (0.023 ms): perf/26347 open(filename: 0xf2e780, mode: 15918976 ) = 19
220.850 (0.002 ms): perf/26347 close(fd: 19 ) = 0
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-6be5jvnkdzjptdrebfn5263n@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-12 19:35:21 +04:00
bool show_comm ;
2013-09-28 01:06:19 +04:00
bool show_tool_stats ;
2014-06-26 20:14:28 +04:00
bool trace_syscalls ;
perf trace: Support using -f to override perf.data file ownership
Enable perf trace to use perf.data when it is not owned by current user
or root.
Example:
# perf trace record ls
# chown Yunlong.Song:Yunlong.Song perf.data
# ls -al perf.data
-rw------- 1 Yunlong.Song Yunlong.Song 4153101 Apr 2 15:28 perf.data
# id
uid=0(root) gid=0(root) groups=0(root),64(pkcs11)
Before this patch:
# perf trace -i perf.data
File perf.data not owned by current user or root (use -f to override)
# perf trace -i perf.data -f
Error: unknown switch `f'
usage: perf trace [<options>] [<command>]
or: perf trace [<options>] -- <command> [<options>]
or: perf trace record [<options>] [<command>]
or: perf trace record [<options>] -- <command> [<options>]
--event <event> event selector. use 'perf list' to list
available events
--comm show the thread COMM next to its id
--tool_stats show tool stats
-e, --expr <expr> list of events to trace
-o, --output <file> output file name
-i, --input <file> Analyze events in file
-p, --pid <pid> trace events on existing process id
-t, --tid <tid> trace events on existing thread id
--filter-pids <float>
...
As shown above, the -f option does not work at all.
After this patch:
# perf trace -i perf.data
File perf.data not owned by current user or root (use -f to override)
# perf trace -i perf.data -f
0.056 ( 0.002 ms): ls/47325 brk( ...
0.108 ( 0.018 ms): ls/47325 mmap(len: 4096, prot: READ|WRITE, ...
0.145 ( 0.013 ms): ls/47325 access(filename: 0x7f31259a0eb0, ...
0.172 ( 0.008 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.180 ( 0.004 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.185 ( 0.004 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.189 ( 0.003 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.195 ( 0.004 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.199 ( 0.002 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.205 ( 0.004 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.211 ( 0.004 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.220 ( 0.007 ms): ls/47325 open(filename: 0x7f312599e8ff, ...
...
...
As shown above, the -f option really works now.
Signed-off-by: Yunlong Song <yunlong.song@huawei.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/1427982439-27388-10-git-send-email-yunlong.song@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-02 16:47:18 +03:00
bool force ;
2015-08-04 23:01:04 +03:00
bool vfs_getname ;
2014-06-26 20:14:25 +04:00
int trace_pgfaults ;
2012-09-27 03:05:56 +04:00
} ;
2013-09-28 00:34:10 +04:00
static int trace__set_fd_pathname ( struct thread * thread , int fd , const char * pathname )
2013-09-24 18:04:32 +04:00
{
2014-10-06 04:46:03 +04:00
struct thread_trace * ttrace = thread__priv ( thread ) ;
2013-09-24 18:04:32 +04:00
if ( fd > ttrace - > paths . max ) {
char * * npath = realloc ( ttrace - > paths . table , ( fd + 1 ) * sizeof ( char * ) ) ;
if ( npath = = NULL )
return - 1 ;
if ( ttrace - > paths . max ! = - 1 ) {
memset ( npath + ttrace - > paths . max + 1 , 0 ,
( fd - ttrace - > paths . max ) * sizeof ( char * ) ) ;
} else {
memset ( npath , 0 , ( fd + 1 ) * sizeof ( char * ) ) ;
}
ttrace - > paths . table = npath ;
ttrace - > paths . max = fd ;
}
ttrace - > paths . table [ fd ] = strdup ( pathname ) ;
return ttrace - > paths . table [ fd ] ! = NULL ? 0 : - 1 ;
}
2013-09-28 00:34:10 +04:00
static int thread__read_fd_path ( struct thread * thread , int fd )
{
char linkname [ PATH_MAX ] , pathname [ PATH_MAX ] ;
struct stat st ;
int ret ;
if ( thread - > pid_ = = thread - > tid ) {
scnprintf ( linkname , sizeof ( linkname ) ,
" /proc/%d/fd/%d " , thread - > pid_ , fd ) ;
} else {
scnprintf ( linkname , sizeof ( linkname ) ,
" /proc/%d/task/%d/fd/%d " , thread - > pid_ , thread - > tid , fd ) ;
}
if ( lstat ( linkname , & st ) < 0 | | st . st_size + 1 > ( off_t ) sizeof ( pathname ) )
return - 1 ;
ret = readlink ( linkname , pathname , sizeof ( pathname ) ) ;
if ( ret < 0 | | ret > st . st_size )
return - 1 ;
pathname [ ret ] = ' \0 ' ;
return trace__set_fd_pathname ( thread , fd , pathname ) ;
}
2013-09-28 01:06:19 +04:00
static const char * thread__fd_path ( struct thread * thread , int fd ,
struct trace * trace )
2013-09-24 18:04:32 +04:00
{
2014-10-06 04:46:03 +04:00
struct thread_trace * ttrace = thread__priv ( thread ) ;
2013-09-24 18:04:32 +04:00
if ( ttrace = = NULL )
return NULL ;
if ( fd < 0 )
return NULL ;
perf trace: Fix up fd -> pathname resolution
There was a brown paper bag bug in the patch that introduced a reference
implementation on using 'perf probe' made wannabe tracepoints that broke fd ->
pathname resolution, fix it:
[root@zoo ~]# perf probe 'vfs_getname=getname_flags:65 pathname=result->name:string'
Added new event:
probe:vfs_getname (on getname_flags:65 with pathname=result->name:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
[root@zoo ~]
Before:
[acme@zoo linux]$ trace touch -e open,fstat /tmp/b
1.159 ( 0.007 ms): open(filename: 0x7fd73f2fe088, flags: CLOEXEC ) = 3
1.163 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fff1b25e610 ) = 0
1.192 ( 0.009 ms): open(filename: 0x7fd73f4fedb8, flags: CLOEXEC ) = 3
1.201 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fff1b25e660 ) = 0
1.501 ( 0.013 ms): open(filename: 0x7fd73f0a1610, flags: CLOEXEC ) = 3
1.505 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fd73f2ddb60 ) = 0
1.581 ( 0.011 ms): open(filename: 0x7fff1b2603da, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
[acme@zoo linux]$
After:
[acme@zoo linux]$ trace touch -e open,fstat,dup2,mmap,close /tmp/b
1.105 ( 0.004 ms): mmap(len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: -1 ) = 0x2fbf000
1.136 ( 0.008 ms): open(filename: 0x7f8902dbc088, flags: CLOEXEC ) = 3
1.140 ( 0.002 ms): fstat(fd: 3</etc/ld.so.cache>, statbuf: 0x7fff19889ef0 ) = 0
1.146 ( 0.004 ms): mmap(len: 86079, prot: READ, flags: PRIVATE, fd: 3</etc/ld.so.cache> ) = 0x2fa9000
1.149 ( 0.001 ms): close(fd: 3</etc/ld.so.cache> ) = 0
1.170 ( 0.010 ms): open(filename: 0x7f8902fbcdb8, flags: CLOEXEC ) = 3
1.178 ( 0.002 ms): fstat(fd: 3</lib64/libc.so.6>, statbuf: 0x7fff19889f40 ) = 0
1.188 ( 0.006 ms): mmap(len: 3924576, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3</lib64/libc.so.6>) = 0x29e2000
1.207 ( 0.007 ms): mmap(addr: 0x7f8902d96000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3</lib64/libc.so.6>, off: 1785856) = 0x2d96000
1.217 ( 0.004 ms): mmap(addr: 0x7f8902d9c000, len: 16992, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: -1) = 0x2d9c000
1.228 ( 0.002 ms): close(fd: 3</lib64/libc.so.6> ) = 0
1.243 ( 0.003 ms): mmap(len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: -1 ) = 0x2fa8000
1.250 ( 0.003 ms): mmap(len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: -1 ) = 0x2fa6000
1.452 ( 0.010 ms): open(filename: 0x7f8902b5f610, flags: CLOEXEC ) = 3
1.455 ( 0.002 ms): fstat(fd: 3</usr/lib/locale/locale-archive>, statbuf: 0x7f8902d9bb60 ) = 0
1.461 ( 0.004 ms): mmap(len: 106070960, prot: READ, flags: PRIVATE, fd: 3</usr/lib/locale/locale-archive>) = 0xfc4b9000
1.469 ( 0.002 ms): close(fd: 3</usr/lib/locale/locale-archive> ) = 0
1.528 ( 0.010 ms): open(filename: 0x7fff1988c3da, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
1.532 ( 0.002 ms): dup2(oldfd: 3</tmp/b> ) = 0
1.535 ( 0.001 ms): close(fd: 3</tmp/b> ) = 0
1.544 ( 0.001 ms): close( ) = 0
1.555 ( 0.001 ms): close(fd: 1 ) = 0
1.558 ( 0.001 ms): close(fd: 2 ) = 0
[acme@zoo linux]$
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-vcm22xpjxc3j4hbyuzjzf7ik@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2014-06-10 23:00:18 +04:00
if ( ( fd > ttrace - > paths . max | | ttrace - > paths . table [ fd ] = = NULL ) ) {
2013-09-28 01:06:19 +04:00
if ( ! trace - > live )
return NULL ;
+ + trace - > stats . proc_getname ;
perf trace: Fix up fd -> pathname resolution
There was a brown paper bag bug in the patch that introduced a reference
implementation on using 'perf probe' made wannabe tracepoints that broke fd ->
pathname resolution, fix it:
[root@zoo ~]# perf probe 'vfs_getname=getname_flags:65 pathname=result->name:string'
Added new event:
probe:vfs_getname (on getname_flags:65 with pathname=result->name:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
[root@zoo ~]
Before:
[acme@zoo linux]$ trace touch -e open,fstat /tmp/b
1.159 ( 0.007 ms): open(filename: 0x7fd73f2fe088, flags: CLOEXEC ) = 3
1.163 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fff1b25e610 ) = 0
1.192 ( 0.009 ms): open(filename: 0x7fd73f4fedb8, flags: CLOEXEC ) = 3
1.201 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fff1b25e660 ) = 0
1.501 ( 0.013 ms): open(filename: 0x7fd73f0a1610, flags: CLOEXEC ) = 3
1.505 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fd73f2ddb60 ) = 0
1.581 ( 0.011 ms): open(filename: 0x7fff1b2603da, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
[acme@zoo linux]$
After:
[acme@zoo linux]$ trace touch -e open,fstat,dup2,mmap,close /tmp/b
1.105 ( 0.004 ms): mmap(len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: -1 ) = 0x2fbf000
1.136 ( 0.008 ms): open(filename: 0x7f8902dbc088, flags: CLOEXEC ) = 3
1.140 ( 0.002 ms): fstat(fd: 3</etc/ld.so.cache>, statbuf: 0x7fff19889ef0 ) = 0
1.146 ( 0.004 ms): mmap(len: 86079, prot: READ, flags: PRIVATE, fd: 3</etc/ld.so.cache> ) = 0x2fa9000
1.149 ( 0.001 ms): close(fd: 3</etc/ld.so.cache> ) = 0
1.170 ( 0.010 ms): open(filename: 0x7f8902fbcdb8, flags: CLOEXEC ) = 3
1.178 ( 0.002 ms): fstat(fd: 3</lib64/libc.so.6>, statbuf: 0x7fff19889f40 ) = 0
1.188 ( 0.006 ms): mmap(len: 3924576, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3</lib64/libc.so.6>) = 0x29e2000
1.207 ( 0.007 ms): mmap(addr: 0x7f8902d96000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3</lib64/libc.so.6>, off: 1785856) = 0x2d96000
1.217 ( 0.004 ms): mmap(addr: 0x7f8902d9c000, len: 16992, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: -1) = 0x2d9c000
1.228 ( 0.002 ms): close(fd: 3</lib64/libc.so.6> ) = 0
1.243 ( 0.003 ms): mmap(len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: -1 ) = 0x2fa8000
1.250 ( 0.003 ms): mmap(len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: -1 ) = 0x2fa6000
1.452 ( 0.010 ms): open(filename: 0x7f8902b5f610, flags: CLOEXEC ) = 3
1.455 ( 0.002 ms): fstat(fd: 3</usr/lib/locale/locale-archive>, statbuf: 0x7f8902d9bb60 ) = 0
1.461 ( 0.004 ms): mmap(len: 106070960, prot: READ, flags: PRIVATE, fd: 3</usr/lib/locale/locale-archive>) = 0xfc4b9000
1.469 ( 0.002 ms): close(fd: 3</usr/lib/locale/locale-archive> ) = 0
1.528 ( 0.010 ms): open(filename: 0x7fff1988c3da, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
1.532 ( 0.002 ms): dup2(oldfd: 3</tmp/b> ) = 0
1.535 ( 0.001 ms): close(fd: 3</tmp/b> ) = 0
1.544 ( 0.001 ms): close( ) = 0
1.555 ( 0.001 ms): close(fd: 1 ) = 0
1.558 ( 0.001 ms): close(fd: 2 ) = 0
[acme@zoo linux]$
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-vcm22xpjxc3j4hbyuzjzf7ik@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2014-06-10 23:00:18 +04:00
if ( thread__read_fd_path ( thread , fd ) )
2013-09-28 01:06:19 +04:00
return NULL ;
}
2013-09-24 18:04:32 +04:00
return ttrace - > paths . table [ fd ] ;
}
static size_t syscall_arg__scnprintf_fd ( char * bf , size_t size ,
struct syscall_arg * arg )
{
int fd = arg - > val ;
size_t printed = scnprintf ( bf , size , " %d " , fd ) ;
2013-09-28 01:06:19 +04:00
const char * path = thread__fd_path ( arg - > thread , fd , arg - > trace ) ;
2013-09-24 18:04:32 +04:00
if ( path )
printed + = scnprintf ( bf + printed , size - printed , " <%s> " , path ) ;
return printed ;
}
static size_t syscall_arg__scnprintf_close_fd ( char * bf , size_t size ,
struct syscall_arg * arg )
{
int fd = arg - > val ;
size_t printed = syscall_arg__scnprintf_fd ( bf , size , arg ) ;
2014-10-06 04:46:03 +04:00
struct thread_trace * ttrace = thread__priv ( arg - > thread ) ;
2013-09-24 18:04:32 +04:00
2013-12-27 00:41:15 +04:00
if ( ttrace & & fd > = 0 & & fd < = ttrace - > paths . max )
zfree ( & ttrace - > paths . table [ fd ] ) ;
2013-09-24 18:04:32 +04:00
return printed ;
}
perf trace: Deref sys_enter pointer args with contents from probe:vfs_getname
To work like strace and dereference syscall pointer args we need to
insert probes (or tracepoints) right after we copy those bytes from
userspace.
Since we're formatting the syscall args at raw_syscalls:sys_enter time,
we need to have a formatter that just stores the position where, later,
when we get the probe:vfs_getname, we can insert the pointer contents.
Now, if a probe:vfs_getname with this format is in place:
# perf probe -l
probe:vfs_getname (on getname_flags:72@/home/git/linux/fs/namei.c with pathname)
That was, in this case, put in place with:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
Then 'perf trace' will notice that and do the pointer -> contents
expansion:
# trace -e open touch /tmp/bla
0.165 (0.010 ms): touch/17752 open(filename: /etc/ld.so.cache, flags: CLOEXEC) = 3
0.195 (0.011 ms): touch/17752 open(filename: /lib64/libc.so.6, flags: CLOEXEC) = 3
0.512 (0.012 ms): touch/17752 open(filename: /usr/lib/locale/locale-archive, flags: CLOEXEC) = 3
0.582 (0.012 ms): touch/17752 open(filename: /tmp/bla, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
#
Roughly equivalent to strace's output:
# strace -rT -e open touch /tmp/bla
0.000000 open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3 <0.000039>
0.000317 open("/lib64/libc.so.6", O_RDONLY|O_CLOEXEC) = 3 <0.000102>
0.001461 open("/usr/lib/locale/locale-archive", O_RDONLY|O_CLOEXEC) = 3 <0.000072>
0.000405 open("/tmp/bla", O_WRONLY|O_CREAT|O_NOCTTY|O_NONBLOCK, 0666) = 3 <0.000055>
0.000641 +++ exited with 0 +++
#
Now we need to either look for at all syscalls that are marked as
pointers and have some well known names ("filename", "pathname", etc)
and set the arg formatter to the one used for the "open" syscall in this
patch.
This implementation works for syscalls with just a string being copied
from userspace, for matching syscalls with more than one string being
copied via the same probe/trace point (vfs_getname) we need to extend
the vfs_getname probe spec to include the pointer too, but there are
some problems with that in 'perf probe' or the kernel kprobes code, need
to investigate before considering supporting multiple strings per
syscall.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-xvuwx6nuj8cf389kf9s2ue2s@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-08-05 04:30:09 +03:00
static void thread__set_filename_pos ( struct thread * thread , const char * bf ,
unsigned long ptr )
{
struct thread_trace * ttrace = thread__priv ( thread ) ;
ttrace - > filename . ptr = ptr ;
ttrace - > filename . entry_str_pos = bf - ttrace - > entry_str ;
}
static size_t syscall_arg__scnprintf_filename ( char * bf , size_t size ,
struct syscall_arg * arg )
{
unsigned long ptr = arg - > val ;
if ( ! arg - > trace - > vfs_getname )
return scnprintf ( bf , size , " %#x " , ptr ) ;
thread__set_filename_pos ( arg - > thread , bf , ptr ) ;
return 0 ;
}
perf trace: Add duration filter
Example:
[acme@sandy linux]$ perf trace --duration 0.025 usleep 1
2.221 ( 0.958 ms): 6724 execve(arg0: 140733557168278, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = -2
3.690 ( 1.443 ms): 6724 execve(arg0: 140733557168295, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = 0
3.979 ( 0.048 ms): 6724 open(filename: 208733843841, flags: 0, mode: 1 ) = 3
4.071 ( 0.075 ms): 6724 open(filename: 139744419925673, flags: 0, mode: 0 ) = 3
4.318 ( 0.056 ms): 6724 nanosleep(rqtp: 140734030404608, rmtp: 0 ) = 0
[acme@sandy linux]$ perf trace --duration 0.100 usleep 1
1.143 ( 1.021 ms): 6726 execve(arg0: 140736323962279, arg1: 140736323972752, arg2: 34926752, arg3: 140736323961824, arg4: 7955998171588342573, arg5: 6725) = 0
[acme@sandy linux]$
Cherry picked from tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-oslw2j2958we9qf0ctra4whd@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-08 16:56:00 +04:00
static bool trace__filter_duration ( struct trace * trace , double t )
{
return t < ( trace - > duration_filter * NSEC_PER_MSEC ) ;
}
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
static size_t trace__fprintf_tstamp ( struct trace * trace , u64 tstamp , FILE * fp )
{
double ts = ( double ) ( tstamp - trace - > base_time ) / NSEC_PER_MSEC ;
perf trace: Add an event duration column
# perf trace usleep 1 | tail -10
0.453 ( 0.002 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763342848
0.456 ( 0.001 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763346944
0.459 ( 0.001 ms): arch_prctl(option: 4098, arg2: 140126839658240, arg3: 140126839652352, arg4: 34, arg5: 4294967295) = 0
0.473 ( 0.003 ms): mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
0.477 ( 0.003 ms): mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
0.483 ( 0.004 ms): munmap(addr: 140126839664640, len: 91882 ) = 0
0.540 ( 0.001 ms): brk(brk: 0 ) = 31928320
0.542 ( 0.002 ms): brk(brk: 32063488 ) = 32063488
1.456 ( 0.901 ms): nanosleep(rqtp: 140735472817168, rmtp: 0 ) = 0
1.462 ( 0.000 ms): exit_group(error_code: 0
#
This also comes from the tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-g9akh5hjw2kvjerpo9xror6f@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-24 23:24:47 +04:00
return fprintf ( fp , " %10.3f " , ts ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
}
2012-10-05 09:02:16 +04:00
static bool done = false ;
2013-10-16 18:57:33 +04:00
static bool interrupted = false ;
2012-10-05 09:02:16 +04:00
2013-10-16 18:57:33 +04:00
static void sig_handler ( int sig )
2012-10-05 09:02:16 +04:00
{
done = true ;
2013-10-16 18:57:33 +04:00
interrupted = sig = = SIGINT ;
2012-10-05 09:02:16 +04:00
}
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
static size_t trace__fprintf_entry_head ( struct trace * trace , struct thread * thread ,
perf trace: Add an event duration column
# perf trace usleep 1 | tail -10
0.453 ( 0.002 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763342848
0.456 ( 0.001 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763346944
0.459 ( 0.001 ms): arch_prctl(option: 4098, arg2: 140126839658240, arg3: 140126839652352, arg4: 34, arg5: 4294967295) = 0
0.473 ( 0.003 ms): mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
0.477 ( 0.003 ms): mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
0.483 ( 0.004 ms): munmap(addr: 140126839664640, len: 91882 ) = 0
0.540 ( 0.001 ms): brk(brk: 0 ) = 31928320
0.542 ( 0.002 ms): brk(brk: 32063488 ) = 32063488
1.456 ( 0.901 ms): nanosleep(rqtp: 140735472817168, rmtp: 0 ) = 0
1.462 ( 0.000 ms): exit_group(error_code: 0
#
This also comes from the tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-g9akh5hjw2kvjerpo9xror6f@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-24 23:24:47 +04:00
u64 duration , u64 tstamp , FILE * fp )
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
{
size_t printed = trace__fprintf_tstamp ( trace , tstamp , fp ) ;
perf trace: Add an event duration column
# perf trace usleep 1 | tail -10
0.453 ( 0.002 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763342848
0.456 ( 0.001 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763346944
0.459 ( 0.001 ms): arch_prctl(option: 4098, arg2: 140126839658240, arg3: 140126839652352, arg4: 34, arg5: 4294967295) = 0
0.473 ( 0.003 ms): mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
0.477 ( 0.003 ms): mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
0.483 ( 0.004 ms): munmap(addr: 140126839664640, len: 91882 ) = 0
0.540 ( 0.001 ms): brk(brk: 0 ) = 31928320
0.542 ( 0.002 ms): brk(brk: 32063488 ) = 32063488
1.456 ( 0.901 ms): nanosleep(rqtp: 140735472817168, rmtp: 0 ) = 0
1.462 ( 0.000 ms): exit_group(error_code: 0
#
This also comes from the tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-g9akh5hjw2kvjerpo9xror6f@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-24 23:24:47 +04:00
printed + = fprintf_duration ( duration , fp ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
perf trace: Add option to show process COMM
Enabled by default, disable with --no-comm, e.g.:
181.821 (0.001 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: PEEK|TRUNC|CMSG_CLOEXEC ) = 20
181.824 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.825 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.834 (0.002 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: CMSG_CLOEXEC ) = 20
181.836 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.838 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.705 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 1256
181.710 (0.002 ms): evolution-addr/10924 geteuid( ) = 1000
181.712 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.727 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 1256
181.731 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.734 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.908 (0.002 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 20
181.913 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.915 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.930 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 20
181.934 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.937 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
220.718 (0.010 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.741 (0.000 ms): dbus-daemon/10711 ... [continued]: epoll_wait()) = 1
220.759 (0.004 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.780 (0.002 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.788 (0.001 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = -1 EAGAIN Resource temporarily unavailable
220.760 (0.004 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.771 (0.023 ms): perf/26347 open(filename: 0xf2e780, mode: 15918976 ) = 19
220.850 (0.002 ms): perf/26347 close(fd: 19 ) = 0
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-6be5jvnkdzjptdrebfn5263n@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-12 19:35:21 +04:00
if ( trace - > multiple_threads ) {
if ( trace - > show_comm )
2013-09-11 18:56:44 +04:00
printed + = fprintf ( fp , " %.14s/ " , thread__comm_str ( thread ) ) ;
2013-07-04 17:20:31 +04:00
printed + = fprintf ( fp , " %d " , thread - > tid ) ;
perf trace: Add option to show process COMM
Enabled by default, disable with --no-comm, e.g.:
181.821 (0.001 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: PEEK|TRUNC|CMSG_CLOEXEC ) = 20
181.824 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.825 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.834 (0.002 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: CMSG_CLOEXEC ) = 20
181.836 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.838 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.705 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 1256
181.710 (0.002 ms): evolution-addr/10924 geteuid( ) = 1000
181.712 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.727 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 1256
181.731 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.734 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.908 (0.002 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 20
181.913 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.915 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.930 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 20
181.934 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.937 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
220.718 (0.010 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.741 (0.000 ms): dbus-daemon/10711 ... [continued]: epoll_wait()) = 1
220.759 (0.004 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.780 (0.002 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.788 (0.001 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = -1 EAGAIN Resource temporarily unavailable
220.760 (0.004 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.771 (0.023 ms): perf/26347 open(filename: 0xf2e780, mode: 15918976 ) = 19
220.850 (0.002 ms): perf/26347 close(fd: 19 ) = 0
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-6be5jvnkdzjptdrebfn5263n@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-12 19:35:21 +04:00
}
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
return printed ;
}
2013-08-19 19:01:10 +04:00
static int trace__process_event ( struct trace * trace , struct machine * machine ,
2013-09-11 18:18:24 +04:00
union perf_event * event , struct perf_sample * sample )
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
{
int ret = 0 ;
switch ( event - > header . type ) {
case PERF_RECORD_LOST :
2013-08-19 19:01:10 +04:00
color_fprintf ( trace - > output , PERF_COLOR_RED ,
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
" LOST % " PRIu64 " events! \n " , event - > lost . lost ) ;
2013-09-11 18:18:24 +04:00
ret = machine__process_lost_event ( machine , event , sample ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
default :
2013-09-11 18:18:24 +04:00
ret = machine__process_event ( machine , event , sample ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
break ;
}
return ret ;
}
2013-08-19 19:01:10 +04:00
static int trace__tool_process ( struct perf_tool * tool ,
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
union perf_event * event ,
2013-09-11 18:18:24 +04:00
struct perf_sample * sample ,
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
struct machine * machine )
{
2013-08-19 19:01:10 +04:00
struct trace * trace = container_of ( tool , struct trace , tool ) ;
2013-09-11 18:18:24 +04:00
return trace__process_event ( trace , machine , event , sample ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
}
static int trace__symbols_init ( struct trace * trace , struct perf_evlist * evlist )
{
2014-08-12 10:40:45 +04:00
int err = symbol__init ( NULL ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
if ( err )
return err ;
2013-09-28 23:13:00 +04:00
trace - > host = machine__new_host ( ) ;
if ( trace - > host = = NULL )
return - ENOMEM ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
perf python: Remove dependency on 'machine' methods
The python binding still doesn't provide symbol resolving facilities,
but the recent addition of the trace_event__register_resolver() function
made it add as a dependency the machine__resolve_kernel_addr() method,
that in turn drags all the symbol resolving code.
The problem:
[root@zoo ~]# perf test -v python
17: Try 'import perf' in python, checking link problems :
--- start ---
test child forked, pid 6853
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: /tmp/build/perf/python/perf.so: undefined symbol: machine__resolve_kernel_addr
test child finished with -1
---- end ----
Try 'import perf' in python, checking link problems: FAILED!
[root@zoo ~]#
Fix it by requiring this function to receive the resolver as a
parameter, just like pevent_register_function_resolver(), i.e. do
not explicitely refer to an object file not included in
tools/perf/util/python-ext-sources.
[root@zoo ~]# perf test python
17: Try 'import perf' in python, checking link problems : Ok
[root@zoo ~]#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Fixes: c3168b0db93a ("perf symbols: Provide libtraceevent callback to resolve kernel symbols")
Link: http://lkml.kernel.org/n/tip-vxlhh95v2em9zdbgj3jm7xi5@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-24 18:13:05 +03:00
if ( trace_event__register_resolver ( trace - > host , machine__resolve_kernel_addr ) < 0 )
2015-07-22 22:16:16 +03:00
return - errno ;
2013-11-11 18:36:12 +04:00
err = __machine__synthesize_threads ( trace - > host , & trace - > tool , & trace - > opts . target ,
2015-06-17 16:51:11 +03:00
evlist - > threads , trace__tool_process , false ,
trace - > opts . proc_map_timeout ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
if ( err )
symbol__exit ( ) ;
return err ;
}
2013-08-24 01:14:48 +04:00
static int syscall__set_arg_fmts ( struct syscall * sc )
{
struct format_field * field ;
int idx = 0 ;
2015-03-20 23:46:53 +03:00
sc - > arg_scnprintf = calloc ( sc - > nr_args , sizeof ( void * ) ) ;
2013-08-24 01:14:48 +04:00
if ( sc - > arg_scnprintf = = NULL )
return - 1 ;
2013-09-03 22:50:28 +04:00
if ( sc - > fmt )
sc - > arg_parm = sc - > fmt - > arg_parm ;
2015-03-20 23:46:53 +03:00
for ( field = sc - > args ; field ; field = field - > next ) {
2013-08-26 19:29:38 +04:00
if ( sc - > fmt & & sc - > fmt - > arg_scnprintf [ idx ] )
sc - > arg_scnprintf [ idx ] = sc - > fmt - > arg_scnprintf [ idx ] ;
else if ( field - > flags & FIELD_IS_POINTER )
2013-08-24 01:14:48 +04:00
sc - > arg_scnprintf [ idx ] = syscall_arg__scnprintf_hex ;
+ + idx ;
}
return 0 ;
}
2012-09-27 03:05:56 +04:00
static int trace__read_syscall_info ( struct trace * trace , int id )
{
char tp_name [ 128 ] ;
struct syscall * sc ;
2013-09-28 01:06:19 +04:00
const char * name = audit_syscall_to_name ( id , trace - > audit . machine ) ;
2012-10-20 19:39:03 +04:00
if ( name = = NULL )
return - 1 ;
2012-09-27 03:05:56 +04:00
if ( id > trace - > syscalls . max ) {
struct syscall * nsyscalls = realloc ( trace - > syscalls . table , ( id + 1 ) * sizeof ( * sc ) ) ;
if ( nsyscalls = = NULL )
return - 1 ;
if ( trace - > syscalls . max ! = - 1 ) {
memset ( nsyscalls + trace - > syscalls . max + 1 , 0 ,
( id - trace - > syscalls . max ) * sizeof ( * sc ) ) ;
} else {
memset ( nsyscalls , 0 , ( id + 1 ) * sizeof ( * sc ) ) ;
}
trace - > syscalls . table = nsyscalls ;
trace - > syscalls . max = id ;
}
sc = trace - > syscalls . table + id ;
2012-10-20 19:39:03 +04:00
sc - > name = name ;
2013-08-09 19:28:31 +04:00
2012-10-20 19:39:03 +04:00
sc - > fmt = syscall_fmt__find ( sc - > name ) ;
2012-09-27 03:05:56 +04:00
2012-09-27 20:16:00 +04:00
snprintf ( tp_name , sizeof ( tp_name ) , " sys_enter_%s " , sc - > name ) ;
2013-12-03 17:09:24 +04:00
sc - > tp_format = trace_event__tp_format ( " syscalls " , tp_name ) ;
2012-09-27 20:16:00 +04:00
if ( sc - > tp_format = = NULL & & sc - > fmt & & sc - > fmt - > alias ) {
snprintf ( tp_name , sizeof ( tp_name ) , " sys_enter_%s " , sc - > fmt - > alias ) ;
2013-12-03 17:09:24 +04:00
sc - > tp_format = trace_event__tp_format ( " syscalls " , tp_name ) ;
2012-09-27 20:16:00 +04:00
}
2012-09-27 03:05:56 +04:00
2013-08-24 01:14:48 +04:00
if ( sc - > tp_format = = NULL )
return - 1 ;
2015-03-20 23:46:53 +03:00
sc - > args = sc - > tp_format - > format . fields ;
sc - > nr_args = sc - > tp_format - > format . nr_fields ;
/* drop nr field - not relevant here; does not exist on older kernels */
if ( sc - > args & & strcmp ( sc - > args - > name , " nr " ) = = 0 ) {
sc - > args = sc - > args - > next ;
- - sc - > nr_args ;
}
2014-06-17 21:29:24 +04:00
sc - > is_exit = ! strcmp ( name , " exit_group " ) | | ! strcmp ( name , " exit " ) ;
2013-08-24 01:14:48 +04:00
return syscall__set_arg_fmts ( sc ) ;
2012-09-27 03:05:56 +04:00
}
perf trace: Validate syscall list passed via -e argument
The 'trace' tool was accepting any names passed and just looking if
syscalls returned via the raw_syscalls:* tracepoints were in that list,
leading to it accepting perf events and then never finding any, as those
are not valid syscall names, confusing users.
Fix it by checking each entry in the list using audit_name_to_syscall,
telling the user which entries are invalid and suggesting where to look
for valid syscall names.
E.g:
[root@zoo ~]# trace -e open,foo,bar,close,baz
Error: Invalid syscall bar, baz, foo
Hint: try 'perf list syscalls:sys_enter_*'
Hint: and: 'man syscalls'
[root@zoo ~]#
Reported-by: Flavio Leitner <fbl@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-4g1i3m1z6fzsrznn2umi02wa@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-26 01:32:33 +03:00
static int trace__validate_ev_qualifier ( struct trace * trace )
{
2015-07-03 00:28:11 +03:00
int err = 0 , i ;
perf trace: Validate syscall list passed via -e argument
The 'trace' tool was accepting any names passed and just looking if
syscalls returned via the raw_syscalls:* tracepoints were in that list,
leading to it accepting perf events and then never finding any, as those
are not valid syscall names, confusing users.
Fix it by checking each entry in the list using audit_name_to_syscall,
telling the user which entries are invalid and suggesting where to look
for valid syscall names.
E.g:
[root@zoo ~]# trace -e open,foo,bar,close,baz
Error: Invalid syscall bar, baz, foo
Hint: try 'perf list syscalls:sys_enter_*'
Hint: and: 'man syscalls'
[root@zoo ~]#
Reported-by: Flavio Leitner <fbl@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-4g1i3m1z6fzsrznn2umi02wa@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-26 01:32:33 +03:00
struct str_node * pos ;
2015-07-03 00:28:11 +03:00
trace - > ev_qualifier_ids . nr = strlist__nr_entries ( trace - > ev_qualifier ) ;
trace - > ev_qualifier_ids . entries = malloc ( trace - > ev_qualifier_ids . nr *
sizeof ( trace - > ev_qualifier_ids . entries [ 0 ] ) ) ;
if ( trace - > ev_qualifier_ids . entries = = NULL ) {
fputs ( " Error: \t Not enough memory for allocating events qualifier ids \n " ,
trace - > output ) ;
err = - EINVAL ;
goto out ;
}
i = 0 ;
perf trace: Validate syscall list passed via -e argument
The 'trace' tool was accepting any names passed and just looking if
syscalls returned via the raw_syscalls:* tracepoints were in that list,
leading to it accepting perf events and then never finding any, as those
are not valid syscall names, confusing users.
Fix it by checking each entry in the list using audit_name_to_syscall,
telling the user which entries are invalid and suggesting where to look
for valid syscall names.
E.g:
[root@zoo ~]# trace -e open,foo,bar,close,baz
Error: Invalid syscall bar, baz, foo
Hint: try 'perf list syscalls:sys_enter_*'
Hint: and: 'man syscalls'
[root@zoo ~]#
Reported-by: Flavio Leitner <fbl@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-4g1i3m1z6fzsrznn2umi02wa@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-26 01:32:33 +03:00
strlist__for_each ( pos , trace - > ev_qualifier ) {
const char * sc = pos - > s ;
2015-07-03 00:28:11 +03:00
int id = audit_name_to_syscall ( sc , trace - > audit . machine ) ;
perf trace: Validate syscall list passed via -e argument
The 'trace' tool was accepting any names passed and just looking if
syscalls returned via the raw_syscalls:* tracepoints were in that list,
leading to it accepting perf events and then never finding any, as those
are not valid syscall names, confusing users.
Fix it by checking each entry in the list using audit_name_to_syscall,
telling the user which entries are invalid and suggesting where to look
for valid syscall names.
E.g:
[root@zoo ~]# trace -e open,foo,bar,close,baz
Error: Invalid syscall bar, baz, foo
Hint: try 'perf list syscalls:sys_enter_*'
Hint: and: 'man syscalls'
[root@zoo ~]#
Reported-by: Flavio Leitner <fbl@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-4g1i3m1z6fzsrznn2umi02wa@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-26 01:32:33 +03:00
2015-07-03 00:28:11 +03:00
if ( id < 0 ) {
perf trace: Validate syscall list passed via -e argument
The 'trace' tool was accepting any names passed and just looking if
syscalls returned via the raw_syscalls:* tracepoints were in that list,
leading to it accepting perf events and then never finding any, as those
are not valid syscall names, confusing users.
Fix it by checking each entry in the list using audit_name_to_syscall,
telling the user which entries are invalid and suggesting where to look
for valid syscall names.
E.g:
[root@zoo ~]# trace -e open,foo,bar,close,baz
Error: Invalid syscall bar, baz, foo
Hint: try 'perf list syscalls:sys_enter_*'
Hint: and: 'man syscalls'
[root@zoo ~]#
Reported-by: Flavio Leitner <fbl@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-4g1i3m1z6fzsrznn2umi02wa@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-26 01:32:33 +03:00
if ( err = = 0 ) {
fputs ( " Error: \t Invalid syscall " , trace - > output ) ;
err = - EINVAL ;
} else {
fputs ( " , " , trace - > output ) ;
}
fputs ( sc , trace - > output ) ;
}
2015-07-03 00:28:11 +03:00
trace - > ev_qualifier_ids . entries [ i + + ] = id ;
perf trace: Validate syscall list passed via -e argument
The 'trace' tool was accepting any names passed and just looking if
syscalls returned via the raw_syscalls:* tracepoints were in that list,
leading to it accepting perf events and then never finding any, as those
are not valid syscall names, confusing users.
Fix it by checking each entry in the list using audit_name_to_syscall,
telling the user which entries are invalid and suggesting where to look
for valid syscall names.
E.g:
[root@zoo ~]# trace -e open,foo,bar,close,baz
Error: Invalid syscall bar, baz, foo
Hint: try 'perf list syscalls:sys_enter_*'
Hint: and: 'man syscalls'
[root@zoo ~]#
Reported-by: Flavio Leitner <fbl@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-4g1i3m1z6fzsrznn2umi02wa@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-26 01:32:33 +03:00
}
if ( err < 0 ) {
fputs ( " \n Hint: \t try 'perf list syscalls:sys_enter_*' "
" \n Hint: \t and: 'man syscalls' \n " , trace - > output ) ;
2015-07-03 00:28:11 +03:00
zfree ( & trace - > ev_qualifier_ids . entries ) ;
trace - > ev_qualifier_ids . nr = 0 ;
perf trace: Validate syscall list passed via -e argument
The 'trace' tool was accepting any names passed and just looking if
syscalls returned via the raw_syscalls:* tracepoints were in that list,
leading to it accepting perf events and then never finding any, as those
are not valid syscall names, confusing users.
Fix it by checking each entry in the list using audit_name_to_syscall,
telling the user which entries are invalid and suggesting where to look
for valid syscall names.
E.g:
[root@zoo ~]# trace -e open,foo,bar,close,baz
Error: Invalid syscall bar, baz, foo
Hint: try 'perf list syscalls:sys_enter_*'
Hint: and: 'man syscalls'
[root@zoo ~]#
Reported-by: Flavio Leitner <fbl@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-4g1i3m1z6fzsrznn2umi02wa@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-26 01:32:33 +03:00
}
2015-07-03 00:28:11 +03:00
out :
perf trace: Validate syscall list passed via -e argument
The 'trace' tool was accepting any names passed and just looking if
syscalls returned via the raw_syscalls:* tracepoints were in that list,
leading to it accepting perf events and then never finding any, as those
are not valid syscall names, confusing users.
Fix it by checking each entry in the list using audit_name_to_syscall,
telling the user which entries are invalid and suggesting where to look
for valid syscall names.
E.g:
[root@zoo ~]# trace -e open,foo,bar,close,baz
Error: Invalid syscall bar, baz, foo
Hint: try 'perf list syscalls:sys_enter_*'
Hint: and: 'man syscalls'
[root@zoo ~]#
Reported-by: Flavio Leitner <fbl@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-4g1i3m1z6fzsrznn2umi02wa@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-26 01:32:33 +03:00
return err ;
}
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
/*
* args is to be interpreted as a series of longs but we need to handle
* 8 - byte unaligned accesses . args points to raw_data within the event
* and raw_data is guaranteed to be 8 - byte unaligned because it is
* preceded by raw_size which is a u32 . So we need to copy args to a temp
* variable to read it . Most notably this avoids extended load instructions
* on unaligned addresses
*/
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
static size_t syscall__scnprintf_args ( struct syscall * sc , char * bf , size_t size ,
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
unsigned char * args , struct trace * trace ,
2013-09-24 18:04:32 +04:00
struct thread * thread )
2012-09-27 03:05:56 +04:00
{
size_t printed = 0 ;
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
unsigned char * p ;
unsigned long val ;
2012-09-27 03:05:56 +04:00
2015-03-20 23:46:53 +03:00
if ( sc - > args ! = NULL ) {
2012-09-27 03:05:56 +04:00
struct format_field * field ;
2013-09-03 19:20:12 +04:00
u8 bit = 1 ;
struct syscall_arg arg = {
2013-09-24 18:04:32 +04:00
. idx = 0 ,
. mask = 0 ,
. trace = trace ,
. thread = thread ,
2013-09-03 19:20:12 +04:00
} ;
2013-09-02 17:39:21 +04:00
2015-03-20 23:46:53 +03:00
for ( field = sc - > args ; field ;
2013-09-03 19:20:12 +04:00
field = field - > next , + + arg . idx , bit < < = 1 ) {
if ( arg . mask & bit )
2013-09-02 17:39:21 +04:00
continue ;
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
/* special care for unaligned accesses */
p = args + sizeof ( unsigned long ) * arg . idx ;
memcpy ( & val , p , sizeof ( val ) ) ;
2013-09-20 19:19:41 +04:00
/*
* Suppress this argument if its value is zero and
* and we don ' t have a string associated in an
* strarray for it .
*/
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
if ( val = = 0 & &
2013-09-20 19:19:41 +04:00
! ( sc - > arg_scnprintf & &
sc - > arg_scnprintf [ arg . idx ] = = SCA_STRARRAY & &
sc - > arg_parm [ arg . idx ] ) )
2013-09-12 18:27:34 +04:00
continue ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
printed + = scnprintf ( bf + printed , size - printed ,
2013-08-24 01:14:48 +04:00
" %s%s: " , printed ? " , " : " " , field - > name ) ;
2013-09-03 19:20:12 +04:00
if ( sc - > arg_scnprintf & & sc - > arg_scnprintf [ arg . idx ] ) {
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
arg . val = val ;
2013-09-03 22:50:28 +04:00
if ( sc - > arg_parm )
arg . parm = sc - > arg_parm [ arg . idx ] ;
2013-09-03 19:20:12 +04:00
printed + = sc - > arg_scnprintf [ arg . idx ] ( bf + printed ,
size - printed , & arg ) ;
2013-09-02 17:39:21 +04:00
} else {
2013-08-24 01:14:48 +04:00
printed + = scnprintf ( bf + printed , size - printed ,
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
" %ld " , val ) ;
2013-09-02 17:39:21 +04:00
}
2012-09-27 03:05:56 +04:00
}
} else {
2013-09-03 19:20:12 +04:00
int i = 0 ;
2012-09-27 03:05:56 +04:00
while ( i < 6 ) {
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
/* special care for unaligned accesses */
p = args + sizeof ( unsigned long ) * i ;
memcpy ( & val , p , sizeof ( val ) ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
printed + = scnprintf ( bf + printed , size - printed ,
" %sarg%d: %ld " ,
perf trace: Fix SIGBUS failures due to misaligned accesses
On Sparc64 perf-trace is failing in many spots due to extended load
instructions being used on misaligned accesses.
(gdb) run trace ls
Starting program: /tmp/perf/perf trace ls
[Thread debugging using libthread_db enabled]
Detaching after fork from child process 169460.
<ls output removed>
Program received signal SIGBUS, Bus error.
0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
warning: Source file is more recent than executable.
61 TP_UINT_FIELD(64);
(gdb) bt
0 0x000000000014f4dc in tp_field__u64 (field=0x4cc700, sample=0x7feffffa098) at builtin-trace.c:61
1 0x0000000000156ad4 in trace__sys_exit (trace=0x7feffffc268, evsel=0x4cc580, event=0xfffffc0104912000,
sample=0x7feffffa098) at builtin-trace.c:1701
2 0x0000000000158c14 in trace__run (trace=0x7feffffc268, argc=1, argv=0x7fefffff360) at builtin-trace.c:2160
3 0x000000000015b78c in cmd_trace (argc=1, argv=0x7fefffff360, prefix=0x0) at builtin-trace.c:2609
4 0x0000000000107d94 in run_builtin (p=0x4549c8, argc=2, argv=0x7fefffff360) at perf.c:341
5 0x0000000000108140 in handle_internal_command (argc=2, argv=0x7fefffff360) at perf.c:400
6 0x0000000000108308 in run_argv (argcp=0x7feffffef2c, argv=0x7feffffef20) at perf.c:444
7 0x0000000000108728 in main (argc=2, argv=0x7fefffff360) at perf.c:559
(gdb) p *sample
$1 = {ip = 4391276, pid = 169472, tid = 169472, time = 6303014583281250, addr = 0, id = 72082,
stream_id = 18446744073709551615, period = 1, weight = 0, transaction = 0, cpu = 73, raw_size = 36,
data_src = 84410401, flags = 0, insn_len = 0, raw_data = 0xfffffc010491203c, callchain = 0x0,
branch_stack = 0x0, user_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0},
intr_regs = {abi = 0, mask = 0, regs = 0x0, cache_regs = 0x7feffffa098, cache_mask = 0}, user_stack = {
offset = 0, size = 0, data = 0x0}, read = {time_enabled = 0, time_running = 0, {group = {nr = 0,
values = 0x0}, one = {value = 0, id = 0}}}}
(gdb) p *field
$2 = {offset = 16, {integer = 0x14f4a8 <tp_field__u64>, pointer = 0x14f4a8 <tp_field__u64>}}
sample->raw_data is guaranteed to not be 8-byte aligned because it is preceded
by the size as a u3. So accessing raw data with an extended load instruction causes
the SIGBUS. Resolve by using memcpy to a temporary variable of appropriate size.
Signed-off-by: David Ahern <david.ahern@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1424376022-140608-1-git-send-email-david.ahern@oracle.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-19 23:00:22 +03:00
printed ? " , " : " " , i , val ) ;
2012-09-27 03:05:56 +04:00
+ + i ;
}
}
return printed ;
}
2012-09-29 00:58:36 +04:00
typedef int ( * tracepoint_handler ) ( struct trace * trace , struct perf_evsel * evsel ,
2014-06-26 20:14:24 +04:00
union perf_event * event ,
2012-09-29 00:58:36 +04:00
struct perf_sample * sample ) ;
static struct syscall * trace__syscall_info ( struct trace * trace ,
2013-10-09 07:26:53 +04:00
struct perf_evsel * evsel , int id )
2012-09-29 00:58:36 +04:00
{
if ( id < 0 ) {
2013-08-23 00:55:25 +04:00
/*
* XXX : Noticed on x86_64 , reproduced as far back as 3.0 .36 , haven ' t tried
* before that , leaving at a higher verbosity level till that is
* explained . Reproduced with plain ftrace with :
*
* echo 1 > / t / events / raw_syscalls / sys_exit / enable
* grep " NR -1 " / t / trace_pipe
*
* After generating some load on the machine .
*/
if ( verbose > 1 ) {
static u64 n ;
fprintf ( trace - > output , " Invalid syscall %d id, skipping (%s, % " PRIu64 " ) ... \n " ,
id , perf_evsel__name ( evsel ) , + + n ) ;
}
2012-09-29 00:58:36 +04:00
return NULL ;
}
if ( ( id > trace - > syscalls . max | | trace - > syscalls . table [ id ] . name = = NULL ) & &
trace__read_syscall_info ( trace , id ) )
goto out_cant_read ;
if ( ( id > trace - > syscalls . max | | trace - > syscalls . table [ id ] . name = = NULL ) )
goto out_cant_read ;
return & trace - > syscalls . table [ id ] ;
out_cant_read :
2013-08-22 23:49:54 +04:00
if ( verbose ) {
fprintf ( trace - > output , " Problems reading syscall %d " , id ) ;
if ( id < = trace - > syscalls . max & & trace - > syscalls . table [ id ] . name ! = NULL )
fprintf ( trace - > output , " (%s) " , trace - > syscalls . table [ id ] . name ) ;
fputs ( " information \n " , trace - > output ) ;
}
2012-09-29 00:58:36 +04:00
return NULL ;
}
2013-10-09 07:26:53 +04:00
static void thread__update_stats ( struct thread_trace * ttrace ,
int id , struct perf_sample * sample )
{
struct int_node * inode ;
struct stats * stats ;
u64 duration = 0 ;
inode = intlist__findnew ( ttrace - > syscall_stats , id ) ;
if ( inode = = NULL )
return ;
stats = inode - > priv ;
if ( stats = = NULL ) {
stats = malloc ( sizeof ( struct stats ) ) ;
if ( stats = = NULL )
return ;
init_stats ( stats ) ;
inode - > priv = stats ;
}
if ( ttrace - > entry_time & & sample - > time > ttrace - > entry_time )
duration = sample - > time - ttrace - > entry_time ;
update_stats ( stats , duration ) ;
}
perf trace: Handle multiple threads better wrt syscalls being intermixed
$ trace time taskset -c 0 usleep 1
0.845 ( 0.021 ms): time/16722 wait4(upid: 4294967295, stat_addr: 0x7fff17f443d4, ru: 0x7fff17f44438 ) ...
0.865 ( 0.008 ms): time/16723 execve(arg0: 140733595272004, arg1: 140733595272720, arg2: 140733595272768, arg3: 139755107218496, arg4: 7307199665339051828, arg5: 3) = -2
2.395 ( 1.523 ms): taskset/16723 execve(arg0: 140733595272013, arg1: 140733595272720, arg2: 140733595272768, arg3: 139755107218496, arg4: 7307199665339051828, arg5: 3) = 0
2.411 ( 0.002 ms): taskset/16723 brk( ) = 0x1915000
3.300 ( 0.058 ms): usleep/16723 nanosleep(rqtp: 0x7ffff4ada190 ) = 0
<SNIP>
3.305 ( 0.000 ms): usleep/16723 exit_group(
3.363 ( 2.539 ms): time/16722 ... [continued]: wait4()) = 16723
3.366 ( 0.001 ms): time/16722 rt_sigaction(sig: INT, act: 0x7fff17f44160, oact: 0x7fff17f44200, sigsetsize: 8) = 0
We we're not seeing this line:
0.845 ( 0.021 ms): time/16722 wait4(upid: 4294967295, stat_addr: 0x7fff17f443d4, ru: 0x7fff17f44438 ) ...
just the one when it finishes:
3.363 ( 2.539 ms): time/16722 ... [continued]: wait4()) = 16723
Still some issues left till we move to ordered_samples when multiple
CPUs/threads are involved...
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-zq9x30a1ky3djqewqn2v3ja3@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-13 19:22:21 +03:00
static int trace__printf_interrupted_entry ( struct trace * trace , struct perf_sample * sample )
{
struct thread_trace * ttrace ;
u64 duration ;
size_t printed ;
if ( trace - > current = = NULL )
return 0 ;
ttrace = thread__priv ( trace - > current ) ;
if ( ! ttrace - > entry_pending )
return 0 ;
duration = sample - > time - ttrace - > entry_time ;
printed = trace__fprintf_entry_head ( trace , trace - > current , duration , sample - > time , trace - > output ) ;
printed + = fprintf ( trace - > output , " %-70s) ... \n " , ttrace - > entry_str ) ;
ttrace - > entry_pending = false ;
return printed ;
}
2012-09-29 00:58:36 +04:00
static int trace__sys_enter ( struct trace * trace , struct perf_evsel * evsel ,
2014-06-26 20:14:24 +04:00
union perf_event * event __maybe_unused ,
2012-09-29 00:58:36 +04:00
struct perf_sample * sample )
{
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
char * msg ;
2012-09-29 00:58:36 +04:00
void * args ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
size_t printed = 0 ;
2013-08-09 19:28:31 +04:00
struct thread * thread ;
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
int id = perf_evsel__sc_tp_uint ( evsel , id , sample ) , err = - 1 ;
2013-10-09 07:26:53 +04:00
struct syscall * sc = trace__syscall_info ( trace , evsel , id ) ;
2013-08-09 19:28:31 +04:00
struct thread_trace * ttrace ;
if ( sc = = NULL )
return - 1 ;
2012-09-29 00:58:36 +04:00
2013-09-28 23:13:00 +04:00
thread = machine__findnew_thread ( trace - > host , sample - > pid , sample - > tid ) ;
2013-08-19 19:01:10 +04:00
ttrace = thread__trace ( thread , trace - > output ) ;
2013-08-09 19:28:31 +04:00
if ( ttrace = = NULL )
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
goto out_put ;
2012-09-29 00:58:36 +04:00
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
args = perf_evsel__sc_tp_ptr ( evsel , args , sample ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
if ( ttrace - > entry_str = = NULL ) {
2015-08-05 04:17:29 +03:00
ttrace - > entry_str = malloc ( trace__entry_str_size ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
if ( ! ttrace - > entry_str )
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
goto out_put ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
}
2015-03-19 21:23:03 +03:00
if ( ! trace - > summary_only )
2015-03-26 00:01:15 +03:00
trace__printf_interrupted_entry ( trace , sample ) ;
perf trace: Handle multiple threads better wrt syscalls being intermixed
$ trace time taskset -c 0 usleep 1
0.845 ( 0.021 ms): time/16722 wait4(upid: 4294967295, stat_addr: 0x7fff17f443d4, ru: 0x7fff17f44438 ) ...
0.865 ( 0.008 ms): time/16723 execve(arg0: 140733595272004, arg1: 140733595272720, arg2: 140733595272768, arg3: 139755107218496, arg4: 7307199665339051828, arg5: 3) = -2
2.395 ( 1.523 ms): taskset/16723 execve(arg0: 140733595272013, arg1: 140733595272720, arg2: 140733595272768, arg3: 139755107218496, arg4: 7307199665339051828, arg5: 3) = 0
2.411 ( 0.002 ms): taskset/16723 brk( ) = 0x1915000
3.300 ( 0.058 ms): usleep/16723 nanosleep(rqtp: 0x7ffff4ada190 ) = 0
<SNIP>
3.305 ( 0.000 ms): usleep/16723 exit_group(
3.363 ( 2.539 ms): time/16722 ... [continued]: wait4()) = 16723
3.366 ( 0.001 ms): time/16722 rt_sigaction(sig: INT, act: 0x7fff17f44160, oact: 0x7fff17f44200, sigsetsize: 8) = 0
We we're not seeing this line:
0.845 ( 0.021 ms): time/16722 wait4(upid: 4294967295, stat_addr: 0x7fff17f443d4, ru: 0x7fff17f44438 ) ...
just the one when it finishes:
3.363 ( 2.539 ms): time/16722 ... [continued]: wait4()) = 16723
Still some issues left till we move to ordered_samples when multiple
CPUs/threads are involved...
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-zq9x30a1ky3djqewqn2v3ja3@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-13 19:22:21 +03:00
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
ttrace - > entry_time = sample - > time ;
msg = ttrace - > entry_str ;
2015-08-05 04:17:29 +03:00
printed + = scnprintf ( msg + printed , trace__entry_str_size - printed , " %s( " , sc - > name ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
2015-08-05 04:17:29 +03:00
printed + = syscall__scnprintf_args ( sc , msg + printed , trace__entry_str_size - printed ,
2013-09-24 18:04:32 +04:00
args , trace , thread ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
2014-06-17 21:29:24 +04:00
if ( sc - > is_exit ) {
2013-11-12 20:31:15 +04:00
if ( ! trace - > duration_filter & & ! trace - > summary_only ) {
2013-08-19 19:01:10 +04:00
trace__fprintf_entry_head ( trace , thread , 1 , sample - > time , trace - > output ) ;
fprintf ( trace - > output , " %-70s \n " , ttrace - > entry_str ) ;
perf trace: Add duration filter
Example:
[acme@sandy linux]$ perf trace --duration 0.025 usleep 1
2.221 ( 0.958 ms): 6724 execve(arg0: 140733557168278, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = -2
3.690 ( 1.443 ms): 6724 execve(arg0: 140733557168295, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = 0
3.979 ( 0.048 ms): 6724 open(filename: 208733843841, flags: 0, mode: 1 ) = 3
4.071 ( 0.075 ms): 6724 open(filename: 139744419925673, flags: 0, mode: 0 ) = 3
4.318 ( 0.056 ms): 6724 nanosleep(rqtp: 140734030404608, rmtp: 0 ) = 0
[acme@sandy linux]$ perf trace --duration 0.100 usleep 1
1.143 ( 1.021 ms): 6726 execve(arg0: 140736323962279, arg1: 140736323972752, arg2: 34926752, arg3: 140736323961824, arg4: 7955998171588342573, arg5: 6725) = 0
[acme@sandy linux]$
Cherry picked from tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-oslw2j2958we9qf0ctra4whd@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-08 16:56:00 +04:00
}
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
} else
ttrace - > entry_pending = true ;
2012-09-29 00:58:36 +04:00
2015-03-03 04:21:35 +03:00
if ( trace - > current ! = thread ) {
thread__put ( trace - > current ) ;
trace - > current = thread__get ( thread ) ;
}
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
err = 0 ;
out_put :
thread__put ( thread ) ;
return err ;
2012-09-29 00:58:36 +04:00
}
static int trace__sys_exit ( struct trace * trace , struct perf_evsel * evsel ,
2014-06-26 20:14:24 +04:00
union perf_event * event __maybe_unused ,
2012-09-29 00:58:36 +04:00
struct perf_sample * sample )
{
2014-09-26 16:54:01 +04:00
long ret ;
perf trace: Add an event duration column
# perf trace usleep 1 | tail -10
0.453 ( 0.002 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763342848
0.456 ( 0.001 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763346944
0.459 ( 0.001 ms): arch_prctl(option: 4098, arg2: 140126839658240, arg3: 140126839652352, arg4: 34, arg5: 4294967295) = 0
0.473 ( 0.003 ms): mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
0.477 ( 0.003 ms): mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
0.483 ( 0.004 ms): munmap(addr: 140126839664640, len: 91882 ) = 0
0.540 ( 0.001 ms): brk(brk: 0 ) = 31928320
0.542 ( 0.002 ms): brk(brk: 32063488 ) = 32063488
1.456 ( 0.901 ms): nanosleep(rqtp: 140735472817168, rmtp: 0 ) = 0
1.462 ( 0.000 ms): exit_group(error_code: 0
#
This also comes from the tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-g9akh5hjw2kvjerpo9xror6f@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-24 23:24:47 +04:00
u64 duration = 0 ;
2013-08-09 19:28:31 +04:00
struct thread * thread ;
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
int id = perf_evsel__sc_tp_uint ( evsel , id , sample ) , err = - 1 ;
2013-10-09 07:26:53 +04:00
struct syscall * sc = trace__syscall_info ( trace , evsel , id ) ;
2013-08-09 19:28:31 +04:00
struct thread_trace * ttrace ;
if ( sc = = NULL )
return - 1 ;
2012-09-29 00:58:36 +04:00
2013-09-28 23:13:00 +04:00
thread = machine__findnew_thread ( trace - > host , sample - > pid , sample - > tid ) ;
2013-08-19 19:01:10 +04:00
ttrace = thread__trace ( thread , trace - > output ) ;
2013-08-09 19:28:31 +04:00
if ( ttrace = = NULL )
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
goto out_put ;
2012-09-29 00:58:36 +04:00
2013-10-09 07:26:53 +04:00
if ( trace - > summary )
thread__update_stats ( ttrace , id , sample ) ;
perf trace: Don't relookup fields by name in each sample
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-11-06 23:35:57 +04:00
ret = perf_evsel__sc_tp_uint ( evsel , ret , sample ) ;
2012-09-29 00:58:36 +04:00
2013-09-28 01:06:19 +04:00
if ( id = = trace - > audit . open_id & & ret > = 0 & & trace - > last_vfs_getname ) {
trace__set_fd_pathname ( thread , ret , trace - > last_vfs_getname ) ;
trace - > last_vfs_getname = NULL ;
+ + trace - > stats . vfs_getname ;
}
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
ttrace - > exit_time = sample - > time ;
perf trace: Add duration filter
Example:
[acme@sandy linux]$ perf trace --duration 0.025 usleep 1
2.221 ( 0.958 ms): 6724 execve(arg0: 140733557168278, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = -2
3.690 ( 1.443 ms): 6724 execve(arg0: 140733557168295, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = 0
3.979 ( 0.048 ms): 6724 open(filename: 208733843841, flags: 0, mode: 1 ) = 3
4.071 ( 0.075 ms): 6724 open(filename: 139744419925673, flags: 0, mode: 0 ) = 3
4.318 ( 0.056 ms): 6724 nanosleep(rqtp: 140734030404608, rmtp: 0 ) = 0
[acme@sandy linux]$ perf trace --duration 0.100 usleep 1
1.143 ( 1.021 ms): 6726 execve(arg0: 140736323962279, arg1: 140736323972752, arg2: 34926752, arg3: 140736323961824, arg4: 7955998171588342573, arg5: 6725) = 0
[acme@sandy linux]$
Cherry picked from tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-oslw2j2958we9qf0ctra4whd@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-08 16:56:00 +04:00
if ( ttrace - > entry_time ) {
perf trace: Add an event duration column
# perf trace usleep 1 | tail -10
0.453 ( 0.002 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763342848
0.456 ( 0.001 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763346944
0.459 ( 0.001 ms): arch_prctl(option: 4098, arg2: 140126839658240, arg3: 140126839652352, arg4: 34, arg5: 4294967295) = 0
0.473 ( 0.003 ms): mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
0.477 ( 0.003 ms): mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
0.483 ( 0.004 ms): munmap(addr: 140126839664640, len: 91882 ) = 0
0.540 ( 0.001 ms): brk(brk: 0 ) = 31928320
0.542 ( 0.002 ms): brk(brk: 32063488 ) = 32063488
1.456 ( 0.901 ms): nanosleep(rqtp: 140735472817168, rmtp: 0 ) = 0
1.462 ( 0.000 ms): exit_group(error_code: 0
#
This also comes from the tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-g9akh5hjw2kvjerpo9xror6f@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-24 23:24:47 +04:00
duration = sample - > time - ttrace - > entry_time ;
perf trace: Add duration filter
Example:
[acme@sandy linux]$ perf trace --duration 0.025 usleep 1
2.221 ( 0.958 ms): 6724 execve(arg0: 140733557168278, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = -2
3.690 ( 1.443 ms): 6724 execve(arg0: 140733557168295, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = 0
3.979 ( 0.048 ms): 6724 open(filename: 208733843841, flags: 0, mode: 1 ) = 3
4.071 ( 0.075 ms): 6724 open(filename: 139744419925673, flags: 0, mode: 0 ) = 3
4.318 ( 0.056 ms): 6724 nanosleep(rqtp: 140734030404608, rmtp: 0 ) = 0
[acme@sandy linux]$ perf trace --duration 0.100 usleep 1
1.143 ( 1.021 ms): 6726 execve(arg0: 140736323962279, arg1: 140736323972752, arg2: 34926752, arg3: 140736323961824, arg4: 7955998171588342573, arg5: 6725) = 0
[acme@sandy linux]$
Cherry picked from tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-oslw2j2958we9qf0ctra4whd@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-08 16:56:00 +04:00
if ( trace__filter_duration ( trace , duration ) )
goto out ;
} else if ( trace - > duration_filter )
goto out ;
perf trace: Add an event duration column
# perf trace usleep 1 | tail -10
0.453 ( 0.002 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763342848
0.456 ( 0.001 ms): mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -763346944
0.459 ( 0.001 ms): arch_prctl(option: 4098, arg2: 140126839658240, arg3: 140126839652352, arg4: 34, arg5: 4294967295) = 0
0.473 ( 0.003 ms): mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
0.477 ( 0.003 ms): mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
0.483 ( 0.004 ms): munmap(addr: 140126839664640, len: 91882 ) = 0
0.540 ( 0.001 ms): brk(brk: 0 ) = 31928320
0.542 ( 0.002 ms): brk(brk: 32063488 ) = 32063488
1.456 ( 0.901 ms): nanosleep(rqtp: 140735472817168, rmtp: 0 ) = 0
1.462 ( 0.000 ms): exit_group(error_code: 0
#
This also comes from the tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-g9akh5hjw2kvjerpo9xror6f@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-24 23:24:47 +04:00
2013-11-12 20:31:15 +04:00
if ( trace - > summary_only )
goto out ;
2013-08-19 19:01:10 +04:00
trace__fprintf_entry_head ( trace , thread , duration , sample - > time , trace - > output ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
if ( ttrace - > entry_pending ) {
2013-08-19 19:01:10 +04:00
fprintf ( trace - > output , " %-70s " , ttrace - > entry_str ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
} else {
2013-08-19 19:01:10 +04:00
fprintf ( trace - > output , " ... [ " ) ;
color_fprintf ( trace - > output , PERF_COLOR_YELLOW , " continued " ) ;
fprintf ( trace - > output , " ]: %s() " , sc - > name ) ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
}
2013-08-26 18:28:34 +04:00
if ( sc - > fmt = = NULL ) {
signed_print :
2014-09-26 16:54:01 +04:00
fprintf ( trace - > output , " ) = %ld " , ret ) ;
2013-08-26 18:28:34 +04:00
} else if ( ret < 0 & & sc - > fmt - > errmsg ) {
2014-08-14 06:22:41 +04:00
char bf [ STRERR_BUFSIZE ] ;
2012-09-29 00:58:36 +04:00
const char * emsg = strerror_r ( - ret , bf , sizeof ( bf ) ) ,
* e = audit_errno_to_name ( - ret ) ;
2013-08-19 19:01:10 +04:00
fprintf ( trace - > output , " ) = -1 %s %s " , e , emsg ) ;
2013-08-26 18:28:34 +04:00
} else if ( ret = = 0 & & sc - > fmt - > timeout )
2013-08-19 19:01:10 +04:00
fprintf ( trace - > output , " ) = 0 Timeout " ) ;
2013-08-26 18:36:30 +04:00
else if ( sc - > fmt - > hexret )
2014-09-26 16:54:01 +04:00
fprintf ( trace - > output , " ) = %#lx " , ret ) ;
2012-09-29 00:58:36 +04:00
else
2013-08-26 18:28:34 +04:00
goto signed_print ;
2012-09-29 00:58:36 +04:00
2013-08-19 19:01:10 +04:00
fputc ( ' \n ' , trace - > output ) ;
perf trace: Add duration filter
Example:
[acme@sandy linux]$ perf trace --duration 0.025 usleep 1
2.221 ( 0.958 ms): 6724 execve(arg0: 140733557168278, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = -2
3.690 ( 1.443 ms): 6724 execve(arg0: 140733557168295, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = 0
3.979 ( 0.048 ms): 6724 open(filename: 208733843841, flags: 0, mode: 1 ) = 3
4.071 ( 0.075 ms): 6724 open(filename: 139744419925673, flags: 0, mode: 0 ) = 3
4.318 ( 0.056 ms): 6724 nanosleep(rqtp: 140734030404608, rmtp: 0 ) = 0
[acme@sandy linux]$ perf trace --duration 0.100 usleep 1
1.143 ( 1.021 ms): 6726 execve(arg0: 140736323962279, arg1: 140736323972752, arg2: 34926752, arg3: 140736323961824, arg4: 7955998171588342573, arg5: 6725) = 0
[acme@sandy linux]$
Cherry picked from tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-oslw2j2958we9qf0ctra4whd@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-08 16:56:00 +04:00
out :
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
ttrace - > entry_pending = false ;
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
err = 0 ;
out_put :
thread__put ( thread ) ;
return err ;
2012-09-29 00:58:36 +04:00
}
2013-09-28 01:06:19 +04:00
static int trace__vfs_getname ( struct trace * trace , struct perf_evsel * evsel ,
2014-06-26 20:14:24 +04:00
union perf_event * event __maybe_unused ,
2013-09-28 01:06:19 +04:00
struct perf_sample * sample )
{
perf trace: Deref sys_enter pointer args with contents from probe:vfs_getname
To work like strace and dereference syscall pointer args we need to
insert probes (or tracepoints) right after we copy those bytes from
userspace.
Since we're formatting the syscall args at raw_syscalls:sys_enter time,
we need to have a formatter that just stores the position where, later,
when we get the probe:vfs_getname, we can insert the pointer contents.
Now, if a probe:vfs_getname with this format is in place:
# perf probe -l
probe:vfs_getname (on getname_flags:72@/home/git/linux/fs/namei.c with pathname)
That was, in this case, put in place with:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
Then 'perf trace' will notice that and do the pointer -> contents
expansion:
# trace -e open touch /tmp/bla
0.165 (0.010 ms): touch/17752 open(filename: /etc/ld.so.cache, flags: CLOEXEC) = 3
0.195 (0.011 ms): touch/17752 open(filename: /lib64/libc.so.6, flags: CLOEXEC) = 3
0.512 (0.012 ms): touch/17752 open(filename: /usr/lib/locale/locale-archive, flags: CLOEXEC) = 3
0.582 (0.012 ms): touch/17752 open(filename: /tmp/bla, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
#
Roughly equivalent to strace's output:
# strace -rT -e open touch /tmp/bla
0.000000 open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3 <0.000039>
0.000317 open("/lib64/libc.so.6", O_RDONLY|O_CLOEXEC) = 3 <0.000102>
0.001461 open("/usr/lib/locale/locale-archive", O_RDONLY|O_CLOEXEC) = 3 <0.000072>
0.000405 open("/tmp/bla", O_WRONLY|O_CREAT|O_NOCTTY|O_NONBLOCK, 0666) = 3 <0.000055>
0.000641 +++ exited with 0 +++
#
Now we need to either look for at all syscalls that are marked as
pointers and have some well known names ("filename", "pathname", etc)
and set the arg formatter to the one used for the "open" syscall in this
patch.
This implementation works for syscalls with just a string being copied
from userspace, for matching syscalls with more than one string being
copied via the same probe/trace point (vfs_getname) we need to extend
the vfs_getname probe spec to include the pointer too, but there are
some problems with that in 'perf probe' or the kernel kprobes code, need
to investigate before considering supporting multiple strings per
syscall.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-xvuwx6nuj8cf389kf9s2ue2s@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-08-05 04:30:09 +03:00
struct thread * thread = machine__findnew_thread ( trace - > host , sample - > pid , sample - > tid ) ;
struct thread_trace * ttrace ;
size_t filename_len , entry_str_len , to_move ;
ssize_t remaining_space ;
char * pos ;
const char * filename ;
2013-09-28 01:06:19 +04:00
trace - > last_vfs_getname = perf_evsel__rawptr ( evsel , sample , " pathname " ) ;
perf trace: Deref sys_enter pointer args with contents from probe:vfs_getname
To work like strace and dereference syscall pointer args we need to
insert probes (or tracepoints) right after we copy those bytes from
userspace.
Since we're formatting the syscall args at raw_syscalls:sys_enter time,
we need to have a formatter that just stores the position where, later,
when we get the probe:vfs_getname, we can insert the pointer contents.
Now, if a probe:vfs_getname with this format is in place:
# perf probe -l
probe:vfs_getname (on getname_flags:72@/home/git/linux/fs/namei.c with pathname)
That was, in this case, put in place with:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
Then 'perf trace' will notice that and do the pointer -> contents
expansion:
# trace -e open touch /tmp/bla
0.165 (0.010 ms): touch/17752 open(filename: /etc/ld.so.cache, flags: CLOEXEC) = 3
0.195 (0.011 ms): touch/17752 open(filename: /lib64/libc.so.6, flags: CLOEXEC) = 3
0.512 (0.012 ms): touch/17752 open(filename: /usr/lib/locale/locale-archive, flags: CLOEXEC) = 3
0.582 (0.012 ms): touch/17752 open(filename: /tmp/bla, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3
#
Roughly equivalent to strace's output:
# strace -rT -e open touch /tmp/bla
0.000000 open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3 <0.000039>
0.000317 open("/lib64/libc.so.6", O_RDONLY|O_CLOEXEC) = 3 <0.000102>
0.001461 open("/usr/lib/locale/locale-archive", O_RDONLY|O_CLOEXEC) = 3 <0.000072>
0.000405 open("/tmp/bla", O_WRONLY|O_CREAT|O_NOCTTY|O_NONBLOCK, 0666) = 3 <0.000055>
0.000641 +++ exited with 0 +++
#
Now we need to either look for at all syscalls that are marked as
pointers and have some well known names ("filename", "pathname", etc)
and set the arg formatter to the one used for the "open" syscall in this
patch.
This implementation works for syscalls with just a string being copied
from userspace, for matching syscalls with more than one string being
copied via the same probe/trace point (vfs_getname) we need to extend
the vfs_getname probe spec to include the pointer too, but there are
some problems with that in 'perf probe' or the kernel kprobes code, need
to investigate before considering supporting multiple strings per
syscall.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-xvuwx6nuj8cf389kf9s2ue2s@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-08-05 04:30:09 +03:00
if ( ! thread )
goto out ;
ttrace = thread__priv ( thread ) ;
if ( ! ttrace )
goto out ;
if ( ! ttrace - > filename . ptr )
goto out ;
entry_str_len = strlen ( ttrace - > entry_str ) ;
remaining_space = trace__entry_str_size - entry_str_len - 1 ; /* \0 */
if ( remaining_space < = 0 )
goto out ;
filename = trace - > last_vfs_getname ;
filename_len = strlen ( filename ) ;
if ( filename_len > ( size_t ) remaining_space ) {
filename + = filename_len - remaining_space ;
filename_len = remaining_space ;
}
to_move = entry_str_len - ttrace - > filename . entry_str_pos + 1 ; /* \0 */
pos = ttrace - > entry_str + ttrace - > filename . entry_str_pos ;
memmove ( pos + filename_len , pos , to_move ) ;
memcpy ( pos , filename , filename_len ) ;
ttrace - > filename . ptr = 0 ;
ttrace - > filename . entry_str_pos = 0 ;
out :
2013-09-28 01:06:19 +04:00
return 0 ;
}
2012-10-18 00:13:12 +04:00
static int trace__sched_stat_runtime ( struct trace * trace , struct perf_evsel * evsel ,
2014-06-26 20:14:24 +04:00
union perf_event * event __maybe_unused ,
2012-10-18 00:13:12 +04:00
struct perf_sample * sample )
{
u64 runtime = perf_evsel__intval ( evsel , sample , " runtime " ) ;
double runtime_ms = ( double ) runtime / NSEC_PER_MSEC ;
2013-09-28 23:13:00 +04:00
struct thread * thread = machine__findnew_thread ( trace - > host ,
2013-08-27 12:23:03 +04:00
sample - > pid ,
sample - > tid ) ;
2013-08-19 19:01:10 +04:00
struct thread_trace * ttrace = thread__trace ( thread , trace - > output ) ;
2012-10-18 00:13:12 +04:00
if ( ttrace = = NULL )
goto out_dump ;
ttrace - > runtime_ms + = runtime_ms ;
trace - > runtime_ms + = runtime_ms ;
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
thread__put ( thread ) ;
2012-10-18 00:13:12 +04:00
return 0 ;
out_dump :
2013-08-19 19:01:10 +04:00
fprintf ( trace - > output , " %s: comm=%s,pid=%u,runtime=% " PRIu64 " ,vruntime=% " PRIu64 " ) \n " ,
2012-10-18 00:13:12 +04:00
evsel - > name ,
perf_evsel__strval ( evsel , sample , " comm " ) ,
( pid_t ) perf_evsel__intval ( evsel , sample , " pid " ) ,
runtime ,
perf_evsel__intval ( evsel , sample , " vruntime " ) ) ;
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
thread__put ( thread ) ;
2012-10-18 00:13:12 +04:00
return 0 ;
}
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
static int trace__event_handler ( struct trace * trace , struct perf_evsel * evsel ,
union perf_event * event __maybe_unused ,
struct perf_sample * sample )
{
trace__printf_interrupted_entry ( trace , sample ) ;
trace__fprintf_tstamp ( trace , sample - > time , trace - > output ) ;
2015-02-20 08:51:50 +03:00
if ( trace - > trace_syscalls )
fprintf ( trace - > output , " ( ): " ) ;
fprintf ( trace - > output , " %s: " , evsel - > name ) ;
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
if ( evsel - > tp_format ) {
event_format__fprintf ( evsel - > tp_format , sample - > cpu ,
sample - > raw_data , sample - > raw_size ,
trace - > output ) ;
}
fprintf ( trace - > output , " ) \n " ) ;
return 0 ;
}
2014-06-26 20:14:25 +04:00
static void print_location ( FILE * f , struct perf_sample * sample ,
struct addr_location * al ,
bool print_dso , bool print_sym )
{
if ( ( verbose | | print_dso ) & & al - > map )
fprintf ( f , " %s@ " , al - > map - > dso - > long_name ) ;
if ( ( verbose | | print_sym ) & & al - > sym )
2014-07-08 22:39:21 +04:00
fprintf ( f , " %s+0x% " PRIx64 , al - > sym - > name ,
2014-06-26 20:14:25 +04:00
al - > addr - al - > sym - > start ) ;
else if ( al - > map )
2014-07-08 22:39:21 +04:00
fprintf ( f , " 0x% " PRIx64 , al - > addr ) ;
2014-06-26 20:14:25 +04:00
else
2014-07-08 22:39:21 +04:00
fprintf ( f , " 0x% " PRIx64 , sample - > addr ) ;
2014-06-26 20:14:25 +04:00
}
static int trace__pgfault ( struct trace * trace ,
struct perf_evsel * evsel ,
union perf_event * event ,
struct perf_sample * sample )
{
struct thread * thread ;
u8 cpumode = event - > header . misc & PERF_RECORD_MISC_CPUMODE_MASK ;
struct addr_location al ;
char map_type = ' d ' ;
2014-07-08 22:05:16 +04:00
struct thread_trace * ttrace ;
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
int err = - 1 ;
2014-06-26 20:14:25 +04:00
thread = machine__findnew_thread ( trace - > host , sample - > pid , sample - > tid ) ;
2014-07-08 22:05:16 +04:00
ttrace = thread__trace ( thread , trace - > output ) ;
if ( ttrace = = NULL )
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
goto out_put ;
2014-07-08 22:05:16 +04:00
if ( evsel - > attr . config = = PERF_COUNT_SW_PAGE_FAULTS_MAJ )
ttrace - > pfmaj + + ;
else
ttrace - > pfmin + + ;
if ( trace - > summary_only )
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
goto out ;
2014-06-26 20:14:25 +04:00
2014-10-23 19:50:25 +04:00
thread__find_addr_location ( thread , cpumode , MAP__FUNCTION ,
2014-06-26 20:14:25 +04:00
sample - > ip , & al ) ;
trace__fprintf_entry_head ( trace , thread , 0 , sample - > time , trace - > output ) ;
fprintf ( trace - > output , " %sfault [ " ,
evsel - > attr . config = = PERF_COUNT_SW_PAGE_FAULTS_MAJ ?
" maj " : " min " ) ;
print_location ( trace - > output , sample , & al , false , true ) ;
fprintf ( trace - > output , " ] => " ) ;
2014-10-23 19:50:25 +04:00
thread__find_addr_location ( thread , cpumode , MAP__VARIABLE ,
2014-06-26 20:14:25 +04:00
sample - > addr , & al ) ;
if ( ! al . map ) {
2014-10-23 19:50:25 +04:00
thread__find_addr_location ( thread , cpumode ,
2014-06-26 20:14:25 +04:00
MAP__FUNCTION , sample - > addr , & al ) ;
if ( al . map )
map_type = ' x ' ;
else
map_type = ' ? ' ;
}
print_location ( trace - > output , sample , & al , true , false ) ;
fprintf ( trace - > output , " (%c%c) \n " , map_type , al . level ) ;
perf machine: Protect the machine->threads with a rwlock
In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-07 02:43:22 +03:00
out :
err = 0 ;
out_put :
thread__put ( thread ) ;
return err ;
2014-06-26 20:14:25 +04:00
}
2013-08-29 08:29:53 +04:00
static bool skip_sample ( struct trace * trace , struct perf_sample * sample )
{
if ( ( trace - > pid_list & & intlist__find ( trace - > pid_list , sample - > pid ) ) | |
( trace - > tid_list & & intlist__find ( trace - > tid_list , sample - > tid ) ) )
return false ;
if ( trace - > pid_list | | trace - > tid_list )
return true ;
return false ;
}
2013-08-29 08:29:52 +04:00
static int trace__process_sample ( struct perf_tool * tool ,
2014-06-26 20:14:24 +04:00
union perf_event * event ,
2013-08-29 08:29:52 +04:00
struct perf_sample * sample ,
struct perf_evsel * evsel ,
struct machine * machine __maybe_unused )
{
struct trace * trace = container_of ( tool , struct trace , tool ) ;
int err = 0 ;
2013-11-06 17:17:38 +04:00
tracepoint_handler handler = evsel - > handler ;
2013-08-29 08:29:52 +04:00
2013-08-29 08:29:53 +04:00
if ( skip_sample ( trace , sample ) )
return 0 ;
2013-09-04 22:37:43 +04:00
if ( ! trace - > full_time & & trace - > base_time = = 0 )
2013-08-29 08:29:52 +04:00
trace - > base_time = sample - > time ;
2013-12-05 06:41:41 +04:00
if ( handler ) {
+ + trace - > nr_events ;
2014-06-26 20:14:24 +04:00
handler ( trace , evsel , event , sample ) ;
2013-12-05 06:41:41 +04:00
}
2013-08-29 08:29:52 +04:00
return err ;
}
2013-08-29 08:29:53 +04:00
static int parse_target_str ( struct trace * trace )
{
if ( trace - > opts . target . pid ) {
trace - > pid_list = intlist__new ( trace - > opts . target . pid ) ;
if ( trace - > pid_list = = NULL ) {
pr_err ( " Error parsing process id string \n " ) ;
return - EINVAL ;
}
}
if ( trace - > opts . target . tid ) {
trace - > tid_list = intlist__new ( trace - > opts . target . tid ) ;
if ( trace - > tid_list = = NULL ) {
pr_err ( " Error parsing thread id string \n " ) ;
return - EINVAL ;
}
}
return 0 ;
}
2014-06-26 20:14:26 +04:00
static int trace__record ( struct trace * trace , int argc , const char * * argv )
2013-09-28 23:13:01 +04:00
{
unsigned int rec_argc , i , j ;
const char * * rec_argv ;
const char * const record_args [ ] = {
" record " ,
" -R " ,
" -m " , " 1024 " ,
" -c " , " 1 " ,
} ;
2014-06-26 20:14:26 +04:00
const char * const sc_args [ ] = { " -e " , } ;
unsigned int sc_args_nr = ARRAY_SIZE ( sc_args ) ;
const char * const majpf_args [ ] = { " -e " , " major-faults " } ;
unsigned int majpf_args_nr = ARRAY_SIZE ( majpf_args ) ;
const char * const minpf_args [ ] = { " -e " , " minor-faults " } ;
unsigned int minpf_args_nr = ARRAY_SIZE ( minpf_args ) ;
2013-12-05 06:41:39 +04:00
/* +1 is for the event string below */
2014-06-26 20:14:26 +04:00
rec_argc = ARRAY_SIZE ( record_args ) + sc_args_nr + 1 +
majpf_args_nr + minpf_args_nr + argc ;
2013-09-28 23:13:01 +04:00
rec_argv = calloc ( rec_argc + 1 , sizeof ( char * ) ) ;
if ( rec_argv = = NULL )
return - ENOMEM ;
2014-06-26 20:14:26 +04:00
j = 0 ;
2013-09-28 23:13:01 +04:00
for ( i = 0 ; i < ARRAY_SIZE ( record_args ) ; i + + )
2014-06-26 20:14:26 +04:00
rec_argv [ j + + ] = record_args [ i ] ;
2014-06-26 20:14:28 +04:00
if ( trace - > trace_syscalls ) {
for ( i = 0 ; i < sc_args_nr ; i + + )
rec_argv [ j + + ] = sc_args [ i ] ;
/* event string may be different for older kernels - e.g., RHEL6 */
if ( is_valid_tracepoint ( " raw_syscalls:sys_enter " ) )
rec_argv [ j + + ] = " raw_syscalls:sys_enter,raw_syscalls:sys_exit " ;
else if ( is_valid_tracepoint ( " syscalls:sys_enter " ) )
rec_argv [ j + + ] = " syscalls:sys_enter,syscalls:sys_exit " ;
else {
pr_err ( " Neither raw_syscalls nor syscalls events exist. \n " ) ;
return - 1 ;
}
2013-12-05 06:41:39 +04:00
}
2014-06-26 20:14:26 +04:00
if ( trace - > trace_pgfaults & TRACE_PFMAJ )
for ( i = 0 ; i < majpf_args_nr ; i + + )
rec_argv [ j + + ] = majpf_args [ i ] ;
if ( trace - > trace_pgfaults & TRACE_PFMIN )
for ( i = 0 ; i < minpf_args_nr ; i + + )
rec_argv [ j + + ] = minpf_args [ i ] ;
for ( i = 0 ; i < ( unsigned int ) argc ; i + + )
rec_argv [ j + + ] = argv [ i ] ;
2013-09-28 23:13:01 +04:00
2014-06-26 20:14:26 +04:00
return cmd_record ( j , rec_argv , NULL ) ;
2013-09-28 23:13:01 +04:00
}
2013-10-09 07:26:53 +04:00
static size_t trace__fprintf_thread_summary ( struct trace * trace , FILE * fp ) ;
2015-08-04 23:01:04 +03:00
static bool perf_evlist__add_vfs_getname ( struct perf_evlist * evlist )
2013-09-28 01:06:19 +04:00
{
2013-11-07 23:41:19 +04:00
struct perf_evsel * evsel = perf_evsel__newtp ( " probe " , " vfs_getname " ) ;
2013-09-28 01:06:19 +04:00
if ( evsel = = NULL )
2015-08-04 23:01:04 +03:00
return false ;
2013-09-28 01:06:19 +04:00
if ( perf_evsel__field ( evsel , " pathname " ) = = NULL ) {
perf_evsel__delete ( evsel ) ;
2015-08-04 23:01:04 +03:00
return false ;
2013-09-28 01:06:19 +04:00
}
2013-11-06 17:17:38 +04:00
evsel - > handler = trace__vfs_getname ;
2013-09-28 01:06:19 +04:00
perf_evlist__add ( evlist , evsel ) ;
2015-08-04 23:01:04 +03:00
return true ;
2013-09-28 01:06:19 +04:00
}
2014-06-26 20:14:25 +04:00
static int perf_evlist__add_pgfault ( struct perf_evlist * evlist ,
u64 config )
{
struct perf_evsel * evsel ;
struct perf_event_attr attr = {
. type = PERF_TYPE_SOFTWARE ,
. mmap_data = 1 ,
} ;
attr . config = config ;
2014-07-24 01:15:09 +04:00
attr . sample_period = 1 ;
2014-06-26 20:14:25 +04:00
event_attr_init ( & attr ) ;
evsel = perf_evsel__new ( & attr ) ;
if ( ! evsel )
return - ENOMEM ;
evsel - > handler = trace__pgfault ;
perf_evlist__add ( evlist , evsel ) ;
return 0 ;
}
2015-02-21 23:10:29 +03:00
static void trace__handle_event ( struct trace * trace , union perf_event * event , struct perf_sample * sample )
{
const u32 type = event - > header . type ;
struct perf_evsel * evsel ;
if ( ! trace - > full_time & & trace - > base_time = = 0 )
trace - > base_time = sample - > time ;
if ( type ! = PERF_RECORD_SAMPLE ) {
trace__process_event ( trace , trace - > host , event , sample ) ;
return ;
}
evsel = perf_evlist__id2evsel ( trace - > evlist , sample - > id ) ;
if ( evsel = = NULL ) {
fprintf ( trace - > output , " Unknown tp ID % " PRIu64 " , skipping... \n " , sample - > id ) ;
return ;
}
if ( evsel - > attr . type = = PERF_TYPE_TRACEPOINT & &
sample - > raw_data = = NULL ) {
fprintf ( trace - > output , " %s sample with no payload for tid: %d, cpu %d, raw_size=%d, skipping... \n " ,
perf_evsel__name ( evsel ) , sample - > tid ,
sample - > cpu , sample - > raw_size ) ;
} else {
tracepoint_handler handler = evsel - > handler ;
handler ( trace , evsel , event , sample ) ;
}
}
2015-07-03 00:24:51 +03:00
static int trace__add_syscall_newtp ( struct trace * trace )
{
int ret = - 1 ;
struct perf_evlist * evlist = trace - > evlist ;
struct perf_evsel * sys_enter , * sys_exit ;
sys_enter = perf_evsel__syscall_newtp ( " sys_enter " , trace__sys_enter ) ;
if ( sys_enter = = NULL )
goto out ;
if ( perf_evsel__init_sc_tp_ptr_field ( sys_enter , args ) )
goto out_delete_sys_enter ;
sys_exit = perf_evsel__syscall_newtp ( " sys_exit " , trace__sys_exit ) ;
if ( sys_exit = = NULL )
goto out_delete_sys_enter ;
if ( perf_evsel__init_sc_tp_uint_field ( sys_exit , ret ) )
goto out_delete_sys_exit ;
perf_evlist__add ( evlist , sys_enter ) ;
perf_evlist__add ( evlist , sys_exit ) ;
2015-07-03 00:28:11 +03:00
trace - > syscalls . events . sys_enter = sys_enter ;
trace - > syscalls . events . sys_exit = sys_exit ;
2015-07-03 00:24:51 +03:00
ret = 0 ;
out :
return ret ;
out_delete_sys_exit :
perf_evsel__delete_priv ( sys_exit ) ;
out_delete_sys_enter :
perf_evsel__delete_priv ( sys_enter ) ;
goto out ;
}
perf trace: Use event filters for the event qualifier list
We use raw_syscalls:sys_{enter,exit} events to show the syscalls, but were
using a rather lazy/inneficient way to implement our 'strace -e' equivalent:
filter out after reading the events in the ring buffer.
Deflect more work to the kernel by appending a filter expression for that,
that, together with the pid list, that is always present, if only to filter the
tracer itself, reduces pressure on the ring buffer and otherwise use
infrastructure already in place in the kernel to do early filtering.
If we use it with -v we can see the filter passed to the kernel,
for instance, for this contrieved case:
# trace -v -e \!open,close,write,poll,recvfrom,select,recvmsg,writev,sendmsg,read,futex,epoll_wait,ioctl,eventfd --filter-pids 2189,2566,1398,2692,4475,4532
<SNIP>
(common_pid != 2514 && common_pid != 1398 && common_pid != 2189 && common_pid != 2566 && common_pid != 2692 && common_pid != 4475 && common_pid != 4532) && (id != 3 && id != 232 && id != 284 && id != 202 && id != 16 && id != 2 && id != 7 && id != 0 && id != 45 && id != 47 && id != 23 && id != 46 && id != 1 && id != 20)
0.011 (0.011 ms): caribou/2295 eventfd2(flags: CLOEXEC|NONBLOCK) = 18
16.946 (0.019 ms): caribou/2295 eventfd2(flags: CLOEXEC|NONBLOCK) = 18
38.598 (0.167 ms): chronyd/794 socket(family: INET, type: DGRAM ) = 4
38.603 (0.002 ms): chronyd/794 fcntl(fd: 4<socket:[239307]>, cmd: GETFD) = 0
38.605 (0.001 ms): chronyd/794 fcntl(fd: 4<socket:[239307]>, cmd: SETFD, arg: 1) = 0
^C
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ti2tg18atproqpguc2moinp6@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-04 18:44:59 +03:00
static int trace__set_ev_qualifier_filter ( struct trace * trace )
{
int err = - 1 ;
char * filter = asprintf_expr_inout_ints ( " id " , ! trace - > not_ev_qualifier ,
trace - > ev_qualifier_ids . nr ,
trace - > ev_qualifier_ids . entries ) ;
if ( filter = = NULL )
goto out_enomem ;
if ( ! perf_evsel__append_filter ( trace - > syscalls . events . sys_enter , " && " , filter ) )
err = perf_evsel__append_filter ( trace - > syscalls . events . sys_exit , " && " , filter ) ;
free ( filter ) ;
out :
return err ;
out_enomem :
errno = ENOMEM ;
goto out ;
}
2015-07-03 00:24:51 +03:00
2012-10-05 09:02:16 +04:00
static int trace__run ( struct trace * trace , int argc , const char * * argv )
2012-09-27 03:05:56 +04:00
{
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
struct perf_evlist * evlist = trace - > evlist ;
perf evlist: Make perf_evlist__set_filter use perf_evsel__set_filter
Instead of calling perf_evsel__apply_filter straight away, so that
we can, in the next patches, expand the filter with more conditions
before actually calling the ioctl to pass the end result filter to
the kernel.
Now we need to call perf_evlist__apply_filters() after the filter
is completely setup, i.e. do the ioctl calls.
The perf_evlist__apply_filters() method was already in place, because
that is the model for the other tools that receives filters in the
command line: go on setting then in the evsel->filter and only at
the end, after parsing the whole command line, apply them.
We get, as a bonus, a more expressive message that states which
event, if any, failed to have the filter applied to, with an
error message stating what happened.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-f429pgz75ryz7tpe6v74etre@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-03 23:42:03 +03:00
struct perf_evsel * evsel ;
2012-10-18 00:09:46 +04:00
int err = - 1 , i ;
unsigned long before ;
2012-10-05 09:02:16 +04:00
const bool forks = argc > 0 ;
2014-09-22 21:39:48 +04:00
bool draining = false ;
2012-09-27 03:05:56 +04:00
2013-09-24 18:04:32 +04:00
trace - > live = true ;
2015-07-03 00:24:51 +03:00
if ( trace - > trace_syscalls & & trace__add_syscall_newtp ( trace ) )
2015-01-22 16:52:55 +03:00
goto out_error_raw_syscalls ;
2012-09-27 03:05:56 +04:00
2014-06-26 20:14:28 +04:00
if ( trace - > trace_syscalls )
2015-08-04 23:01:04 +03:00
trace - > vfs_getname = perf_evlist__add_vfs_getname ( evlist ) ;
2013-09-28 01:06:19 +04:00
2014-06-26 20:14:25 +04:00
if ( ( trace - > trace_pgfaults & TRACE_PFMAJ ) & &
2015-01-22 16:34:22 +03:00
perf_evlist__add_pgfault ( evlist , PERF_COUNT_SW_PAGE_FAULTS_MAJ ) ) {
2015-01-22 17:08:04 +03:00
goto out_error_mem ;
2015-01-22 16:34:22 +03:00
}
2014-06-26 20:14:25 +04:00
if ( ( trace - > trace_pgfaults & TRACE_PFMIN ) & &
perf_evlist__add_pgfault ( evlist , PERF_COUNT_SW_PAGE_FAULTS_MIN ) )
2015-01-22 17:08:04 +03:00
goto out_error_mem ;
2014-06-26 20:14:25 +04:00
2012-10-18 00:13:12 +04:00
if ( trace - > sched & &
2015-01-22 17:13:43 +03:00
perf_evlist__add_newtp ( evlist , " sched " , " sched_stat_runtime " ,
trace__sched_stat_runtime ) )
goto out_error_sched_stat_runtime ;
2012-10-18 00:13:12 +04:00
2012-09-27 03:05:56 +04:00
err = perf_evlist__create_maps ( evlist , & trace - > opts . target ) ;
if ( err < 0 ) {
2013-08-19 19:01:10 +04:00
fprintf ( trace - > output , " Problems parsing the target to trace, check your options! \n " ) ;
2012-09-27 03:05:56 +04:00
goto out_delete_evlist ;
}
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
err = trace__symbols_init ( trace , evlist ) ;
if ( err < 0 ) {
2013-08-19 19:01:10 +04:00
fprintf ( trace - > output , " Problems initializing symbol libraries! \n " ) ;
2014-01-03 22:56:06 +04:00
goto out_delete_evlist ;
perf trace: Support interrupted syscalls
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-07 01:43:19 +04:00
}
2012-12-10 23:41:31 +04:00
perf_evlist__config ( evlist , & trace - > opts ) ;
2012-09-27 03:05:56 +04:00
2012-10-05 09:02:16 +04:00
signal ( SIGCHLD , sig_handler ) ;
signal ( SIGINT , sig_handler ) ;
if ( forks ) {
2013-03-11 11:43:15 +04:00
err = perf_evlist__prepare_workload ( evlist , & trace - > opts . target ,
2014-01-03 21:56:49 +04:00
argv , false , NULL ) ;
2012-10-05 09:02:16 +04:00
if ( err < 0 ) {
2013-08-19 19:01:10 +04:00
fprintf ( trace - > output , " Couldn't run the workload! \n " ) ;
2014-01-03 22:56:06 +04:00
goto out_delete_evlist ;
2012-10-05 09:02:16 +04:00
}
}
2012-09-27 03:05:56 +04:00
err = perf_evlist__open ( evlist ) ;
perf trace: Improve messages related to /proc/sys/kernel/perf_event_paranoid
kernel/events/core.c has:
/*
* perf event paranoia level:
* -1 - not paranoid at all
* 0 - disallow raw tracepoint access for unpriv
* 1 - disallow cpu events for unpriv
* 2 - disallow kernel profiling for unpriv
*/
int sysctl_perf_event_paranoid __read_mostly = 1;
So, with the default being 1, a non-root user can trace his stuff:
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
1
[acme@zoo ~]$ yes > /dev/null &
[1] 15338
[acme@zoo ~]$ trace -p 15338 | head -5
0.005 ( 0.005 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.045 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.085 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.125 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.165 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
[acme@zoo ~]$
[acme@zoo ~]$ trace --duration 1 sleep 1
1002.148 (1001.218 ms): nanosleep(rqtp: 0x7fff46c79250 ) = 0
[acme@zoo ~]$
[acme@zoo ~]$ trace -- usleep 1 | tail -5
0.905 ( 0.002 ms): brk( ) = 0x1c82000
0.910 ( 0.003 ms): brk(brk: 0x1ca3000 ) = 0x1ca3000
0.913 ( 0.001 ms): brk( ) = 0x1ca3000
0.990 ( 0.059 ms): nanosleep(rqtp: 0x7fffe31a3280 ) = 0
0.995 ( 0.000 ms): exit_group(
[acme@zoo ~]$
But can't do system wide tracing:
[acme@zoo ~]$ trace
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 0
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
If the paranoid level is >= 2, i.e. turn this perf stuff off for !root users:
[acme@zoo ~]$ sudo sh -c 'echo 2 > /proc/sys/kernel/perf_event_paranoid'
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
2
[acme@zoo ~]$
[acme@zoo ~]$ trace usleep 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
If the user manages to get what he/she wants, convincing root not
to be paranoid at all...
[root@zoo ~]# echo -1 > /proc/sys/kernel/perf_event_paranoid
[root@zoo ~]# cat /proc/sys/kernel/perf_event_paranoid
-1
[root@zoo ~]#
[acme@zoo ~]$ ps -eo user,pid,comm | grep Xorg
root 729 Xorg
[acme@zoo ~]$
[acme@zoo ~]$ trace -a --duration 0.001 -e \!select,ioctl,writev | grep Xorg | head -5
23.143 ( 0.003 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.152 ( 0.004 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = 8
23.161 ( 0.002 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = -1 EAGAIN Resource temporarily unavailable
23.175 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.235 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
[acme@zoo ~]$
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-di28olfwd28rvkox7v3hqhu1@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-18 00:38:29 +04:00
if ( err < 0 )
goto out_error_open ;
2012-09-27 03:05:56 +04:00
2015-02-21 21:15:21 +03:00
/*
* Better not use ! target__has_task ( ) here because we need to cover the
* case where no threads were specified in the command line , but a
* workload was , and in that case we will fill in the thread_map when
* we fork the workload in perf_evlist__prepare_workload .
*/
perf trace: Introduce --filter-pids
When tracing in X we get event loops due to the tracing activity, i.e.
updates to a gnome-terminal that generate syscalls for X.org, etc.
To get a more useful view of what is happening, syscall wise, system
wide, we need to filter those, like in:
# ps ax|egrep '981|2296|1519' | grep -v egrep
981 tty1 Ss+ 5:40 /usr/bin/Xorg :0 -background none ...
1519 ? Sl 2:22 /usr/bin/gnome-shell
2296 ? Sl 4:16 /usr/libexec/gnome-terminal-server
#
# trace -e write --filter-pids 981,2296,1519
0.385 ( 0.021 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 136) = 136
0.922 ( 0.014 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 140) = 140
5006.525 ( 0.029 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 136) = 136
5007.235 ( 0.023 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 140) = 140
5177.646 ( 0.018 ms): rtkit-daemon/782 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7f7eea70be88, count: 8) = 8
8314.497 ( 0.004 ms): gsd-locate-poi/2084 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7fffe96af7b0, count: 8) = 8
8314.518 ( 0.002 ms): gsd-locate-poi/2084 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7fffe96af0e0, count: 8) = 8
^C#
When this option is used the tracer pid is also filtered.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-f5qmiyy7c0uxdm21ncatpeek@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-21 22:36:52 +03:00
if ( trace - > filter_pids . nr > 0 )
err = perf_evlist__set_filter_pids ( evlist , trace - > filter_pids . nr , trace - > filter_pids . entries ) ;
2015-06-23 01:36:02 +03:00
else if ( thread_map__pid ( evlist - > threads , 0 ) = = - 1 )
perf trace: Introduce --filter-pids
When tracing in X we get event loops due to the tracing activity, i.e.
updates to a gnome-terminal that generate syscalls for X.org, etc.
To get a more useful view of what is happening, syscall wise, system
wide, we need to filter those, like in:
# ps ax|egrep '981|2296|1519' | grep -v egrep
981 tty1 Ss+ 5:40 /usr/bin/Xorg :0 -background none ...
1519 ? Sl 2:22 /usr/bin/gnome-shell
2296 ? Sl 4:16 /usr/libexec/gnome-terminal-server
#
# trace -e write --filter-pids 981,2296,1519
0.385 ( 0.021 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 136) = 136
0.922 ( 0.014 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 140) = 140
5006.525 ( 0.029 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 136) = 136
5007.235 ( 0.023 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 140) = 140
5177.646 ( 0.018 ms): rtkit-daemon/782 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7f7eea70be88, count: 8) = 8
8314.497 ( 0.004 ms): gsd-locate-poi/2084 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7fffe96af7b0, count: 8) = 8
8314.518 ( 0.002 ms): gsd-locate-poi/2084 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7fffe96af0e0, count: 8) = 8
^C#
When this option is used the tracer pid is also filtered.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-f5qmiyy7c0uxdm21ncatpeek@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-21 22:36:52 +03:00
err = perf_evlist__set_filter_pid ( evlist , getpid ( ) ) ;
perf evlist: Make perf_evlist__set_filter use perf_evsel__set_filter
Instead of calling perf_evsel__apply_filter straight away, so that
we can, in the next patches, expand the filter with more conditions
before actually calling the ioctl to pass the end result filter to
the kernel.
Now we need to call perf_evlist__apply_filters() after the filter
is completely setup, i.e. do the ioctl calls.
The perf_evlist__apply_filters() method was already in place, because
that is the model for the other tools that receives filters in the
command line: go on setting then in the evsel->filter and only at
the end, after parsing the whole command line, apply them.
We get, as a bonus, a more expressive message that states which
event, if any, failed to have the filter applied to, with an
error message stating what happened.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-f429pgz75ryz7tpe6v74etre@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-03 23:42:03 +03:00
if ( err < 0 )
goto out_error_mem ;
perf trace: Use event filters for the event qualifier list
We use raw_syscalls:sys_{enter,exit} events to show the syscalls, but were
using a rather lazy/inneficient way to implement our 'strace -e' equivalent:
filter out after reading the events in the ring buffer.
Deflect more work to the kernel by appending a filter expression for that,
that, together with the pid list, that is always present, if only to filter the
tracer itself, reduces pressure on the ring buffer and otherwise use
infrastructure already in place in the kernel to do early filtering.
If we use it with -v we can see the filter passed to the kernel,
for instance, for this contrieved case:
# trace -v -e \!open,close,write,poll,recvfrom,select,recvmsg,writev,sendmsg,read,futex,epoll_wait,ioctl,eventfd --filter-pids 2189,2566,1398,2692,4475,4532
<SNIP>
(common_pid != 2514 && common_pid != 1398 && common_pid != 2189 && common_pid != 2566 && common_pid != 2692 && common_pid != 4475 && common_pid != 4532) && (id != 3 && id != 232 && id != 284 && id != 202 && id != 16 && id != 2 && id != 7 && id != 0 && id != 45 && id != 47 && id != 23 && id != 46 && id != 1 && id != 20)
0.011 (0.011 ms): caribou/2295 eventfd2(flags: CLOEXEC|NONBLOCK) = 18
16.946 (0.019 ms): caribou/2295 eventfd2(flags: CLOEXEC|NONBLOCK) = 18
38.598 (0.167 ms): chronyd/794 socket(family: INET, type: DGRAM ) = 4
38.603 (0.002 ms): chronyd/794 fcntl(fd: 4<socket:[239307]>, cmd: GETFD) = 0
38.605 (0.001 ms): chronyd/794 fcntl(fd: 4<socket:[239307]>, cmd: SETFD, arg: 1) = 0
^C
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ti2tg18atproqpguc2moinp6@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-04 18:44:59 +03:00
if ( trace - > ev_qualifier_ids . nr > 0 ) {
err = trace__set_ev_qualifier_filter ( trace ) ;
if ( err < 0 )
goto out_errno ;
2015-08-03 23:12:29 +03:00
pr_debug ( " event qualifier tracepoint filter: %s \n " ,
trace - > syscalls . events . sys_exit - > filter ) ;
}
perf trace: Use event filters for the event qualifier list
We use raw_syscalls:sys_{enter,exit} events to show the syscalls, but were
using a rather lazy/inneficient way to implement our 'strace -e' equivalent:
filter out after reading the events in the ring buffer.
Deflect more work to the kernel by appending a filter expression for that,
that, together with the pid list, that is always present, if only to filter the
tracer itself, reduces pressure on the ring buffer and otherwise use
infrastructure already in place in the kernel to do early filtering.
If we use it with -v we can see the filter passed to the kernel,
for instance, for this contrieved case:
# trace -v -e \!open,close,write,poll,recvfrom,select,recvmsg,writev,sendmsg,read,futex,epoll_wait,ioctl,eventfd --filter-pids 2189,2566,1398,2692,4475,4532
<SNIP>
(common_pid != 2514 && common_pid != 1398 && common_pid != 2189 && common_pid != 2566 && common_pid != 2692 && common_pid != 4475 && common_pid != 4532) && (id != 3 && id != 232 && id != 284 && id != 202 && id != 16 && id != 2 && id != 7 && id != 0 && id != 45 && id != 47 && id != 23 && id != 46 && id != 1 && id != 20)
0.011 (0.011 ms): caribou/2295 eventfd2(flags: CLOEXEC|NONBLOCK) = 18
16.946 (0.019 ms): caribou/2295 eventfd2(flags: CLOEXEC|NONBLOCK) = 18
38.598 (0.167 ms): chronyd/794 socket(family: INET, type: DGRAM ) = 4
38.603 (0.002 ms): chronyd/794 fcntl(fd: 4<socket:[239307]>, cmd: GETFD) = 0
38.605 (0.001 ms): chronyd/794 fcntl(fd: 4<socket:[239307]>, cmd: SETFD, arg: 1) = 0
^C
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ti2tg18atproqpguc2moinp6@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-04 18:44:59 +03:00
perf evlist: Make perf_evlist__set_filter use perf_evsel__set_filter
Instead of calling perf_evsel__apply_filter straight away, so that
we can, in the next patches, expand the filter with more conditions
before actually calling the ioctl to pass the end result filter to
the kernel.
Now we need to call perf_evlist__apply_filters() after the filter
is completely setup, i.e. do the ioctl calls.
The perf_evlist__apply_filters() method was already in place, because
that is the model for the other tools that receives filters in the
command line: go on setting then in the evsel->filter and only at
the end, after parsing the whole command line, apply them.
We get, as a bonus, a more expressive message that states which
event, if any, failed to have the filter applied to, with an
error message stating what happened.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-f429pgz75ryz7tpe6v74etre@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-03 23:42:03 +03:00
err = perf_evlist__apply_filters ( evlist , & evsel ) ;
if ( err < 0 )
goto out_error_apply_filters ;
2015-02-21 21:15:21 +03:00
2013-11-28 20:57:22 +04:00
err = perf_evlist__mmap ( evlist , trace - > opts . mmap_pages , false ) ;
2014-12-12 00:04:10 +03:00
if ( err < 0 )
goto out_error_mmap ;
2012-09-27 03:05:56 +04:00
2015-04-22 16:04:23 +03:00
if ( ! target__none ( & trace - > opts . target ) )
perf_evlist__enable ( evlist ) ;
2012-10-05 09:02:16 +04:00
if ( forks )
perf_evlist__start_workload ( evlist ) ;
2015-06-23 01:36:02 +03:00
trace - > multiple_threads = thread_map__pid ( evlist - > threads , 0 ) = = - 1 | |
2015-02-13 18:32:45 +03:00
evlist - > threads - > nr > 1 | |
perf_evlist__first ( evlist ) - > attr . inherit ;
2012-09-27 03:05:56 +04:00
again :
2012-10-18 00:09:46 +04:00
before = trace - > nr_events ;
2012-09-27 03:05:56 +04:00
for ( i = 0 ; i < evlist - > nr_mmaps ; i + + ) {
union perf_event * event ;
while ( ( event = perf_evlist__mmap_read ( evlist , i ) ) ! = NULL ) {
struct perf_sample sample ;
2012-10-18 00:09:46 +04:00
+ + trace - > nr_events ;
2012-09-27 03:05:56 +04:00
err = perf_evlist__parse_sample ( evlist , event , & sample ) ;
if ( err ) {
2013-08-19 19:01:10 +04:00
fprintf ( trace - > output , " Can't parse sample, err = %d, skipping... \n " , err ) ;
2013-10-24 11:43:33 +04:00
goto next_event ;
2012-09-27 03:05:56 +04:00
}
2015-02-21 23:10:29 +03:00
trace__handle_event ( trace , event , & sample ) ;
2013-10-24 11:43:33 +04:00
next_event :
perf_evlist__mmap_consume ( evlist , i ) ;
2013-09-03 18:55:07 +04:00
2013-10-16 18:57:33 +04:00
if ( interrupted )
goto out_disable ;
perf trace: Disable events and drain events when forked workload ends
We were not checking in the inner event processing loop if the forked workload
had finished, which, on a busy system, may make it take a long time trying to
drain events, entering a seemingly neverending loop, waiting for the system to
get idle enough to make it drain the buffers.
Fix it by disabling the events when 'done' is true, in the inner loop, to start
draining what is in the buffers.
Now:
[root@ssdandy ~]# time trace --filter-pids 14003 -a sleep 1 | tail
996.748 ( 0.002 ms): sh/30296 rt_sigprocmask(how: SETMASK, nset: 0x7ffc83418160, sigsetsize: 8) = 0
996.751 ( 0.002 ms): sh/30296 rt_sigprocmask(how: BLOCK, nset: 0x7ffc834181f0, oset: 0x7ffc83418270, sigsetsize: 8) = 0
996.755 ( 0.002 ms): sh/30296 rt_sigaction(sig: INT, act: 0x7ffc83417f50, oact: 0x7ffc83417ff0, sigsetsize: 8) = 0
1004.543 ( 0.362 ms): tail/30198 ... [continued]: read()) = 4096
1004.548 ( 7.791 ms): sh/30296 wait4(upid: -1, stat_addr: 0x7ffc834181a0) ...
1004.975 ( 0.427 ms): tail/30198 read(buf: 0x7633f0, count: 8192) = 4096
1005.390 ( 0.410 ms): tail/30198 read(buf: 0x765410, count: 8192) = 4096
1005.743 ( 0.348 ms): tail/30198 read(buf: 0x7633f0, count: 8192) = 4096
1006.197 ( 0.449 ms): tail/30198 read(buf: 0x765410, count: 8192) = 4096
1006.492 ( 0.290 ms): tail/30198 read(buf: 0x7633f0, count: 8192) = 4096
real 0m1.219s
user 0m0.704s
sys 0m0.331s
[root@ssdandy ~]#
Reported-by: Michael Petlan <mpetlan@redhat.com>
Suggested-by: Jiri Olsa <jolsa@redhat.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-p6kpn1b26qcbe47pufpw0tex@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-22 17:11:57 +03:00
if ( done & & ! draining ) {
perf_evlist__disable ( evlist ) ;
draining = true ;
}
2012-09-27 03:05:56 +04:00
}
}
2012-10-18 00:09:46 +04:00
if ( trace - > nr_events = = before ) {
2013-10-16 18:57:33 +04:00
int timeout = done ? 100 : - 1 ;
2012-10-05 09:02:16 +04:00
2014-09-22 21:39:48 +04:00
if ( ! draining & & perf_evlist__poll ( evlist , timeout ) > 0 ) {
if ( perf_evlist__filter_pollfd ( evlist , POLLERR | POLLHUP ) = = 0 )
draining = true ;
2013-10-16 18:57:33 +04:00
goto again ;
2014-09-22 21:39:48 +04:00
}
2013-10-16 18:57:33 +04:00
} else {
goto again ;
2012-10-05 09:02:16 +04:00
}
2013-10-16 18:57:33 +04:00
out_disable :
2015-03-03 04:21:35 +03:00
thread__zput ( trace - > current ) ;
2013-10-16 18:57:33 +04:00
perf_evlist__disable ( evlist ) ;
2012-09-27 03:05:56 +04:00
2013-09-28 01:06:19 +04:00
if ( ! err ) {
if ( trace - > summary )
trace__fprintf_thread_summary ( trace , trace - > output ) ;
if ( trace - > show_tool_stats ) {
fprintf ( trace - > output , " Stats: \n "
" vfs_getname : % " PRIu64 " \n "
" proc_getname: % " PRIu64 " \n " ,
trace - > stats . vfs_getname ,
trace - > stats . proc_getname ) ;
}
}
2013-10-09 07:26:53 +04:00
2012-09-27 03:05:56 +04:00
out_delete_evlist :
perf_evlist__delete ( evlist ) ;
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
trace - > evlist = NULL ;
2013-09-24 18:04:32 +04:00
trace - > live = false ;
2012-09-27 03:05:56 +04:00
return err ;
2013-10-17 19:07:58 +04:00
{
char errbuf [ BUFSIZ ] ;
perf trace: Improve messages related to /proc/sys/kernel/perf_event_paranoid
kernel/events/core.c has:
/*
* perf event paranoia level:
* -1 - not paranoid at all
* 0 - disallow raw tracepoint access for unpriv
* 1 - disallow cpu events for unpriv
* 2 - disallow kernel profiling for unpriv
*/
int sysctl_perf_event_paranoid __read_mostly = 1;
So, with the default being 1, a non-root user can trace his stuff:
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
1
[acme@zoo ~]$ yes > /dev/null &
[1] 15338
[acme@zoo ~]$ trace -p 15338 | head -5
0.005 ( 0.005 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.045 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.085 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.125 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.165 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
[acme@zoo ~]$
[acme@zoo ~]$ trace --duration 1 sleep 1
1002.148 (1001.218 ms): nanosleep(rqtp: 0x7fff46c79250 ) = 0
[acme@zoo ~]$
[acme@zoo ~]$ trace -- usleep 1 | tail -5
0.905 ( 0.002 ms): brk( ) = 0x1c82000
0.910 ( 0.003 ms): brk(brk: 0x1ca3000 ) = 0x1ca3000
0.913 ( 0.001 ms): brk( ) = 0x1ca3000
0.990 ( 0.059 ms): nanosleep(rqtp: 0x7fffe31a3280 ) = 0
0.995 ( 0.000 ms): exit_group(
[acme@zoo ~]$
But can't do system wide tracing:
[acme@zoo ~]$ trace
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 0
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
If the paranoid level is >= 2, i.e. turn this perf stuff off for !root users:
[acme@zoo ~]$ sudo sh -c 'echo 2 > /proc/sys/kernel/perf_event_paranoid'
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
2
[acme@zoo ~]$
[acme@zoo ~]$ trace usleep 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
If the user manages to get what he/she wants, convincing root not
to be paranoid at all...
[root@zoo ~]# echo -1 > /proc/sys/kernel/perf_event_paranoid
[root@zoo ~]# cat /proc/sys/kernel/perf_event_paranoid
-1
[root@zoo ~]#
[acme@zoo ~]$ ps -eo user,pid,comm | grep Xorg
root 729 Xorg
[acme@zoo ~]$
[acme@zoo ~]$ trace -a --duration 0.001 -e \!select,ioctl,writev | grep Xorg | head -5
23.143 ( 0.003 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.152 ( 0.004 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = 8
23.161 ( 0.002 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = -1 EAGAIN Resource temporarily unavailable
23.175 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.235 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
[acme@zoo ~]$
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-di28olfwd28rvkox7v3hqhu1@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-18 00:38:29 +04:00
2015-01-22 17:13:43 +03:00
out_error_sched_stat_runtime :
debugfs__strerror_open_tp ( errno , errbuf , sizeof ( errbuf ) , " sched " , " sched_stat_runtime " ) ;
goto out_error ;
2015-01-22 16:52:55 +03:00
out_error_raw_syscalls :
2015-01-22 17:13:43 +03:00
debugfs__strerror_open_tp ( errno , errbuf , sizeof ( errbuf ) , " raw_syscalls " , " sys_(enter|exit) " ) ;
perf trace: Improve messages related to /proc/sys/kernel/perf_event_paranoid
kernel/events/core.c has:
/*
* perf event paranoia level:
* -1 - not paranoid at all
* 0 - disallow raw tracepoint access for unpriv
* 1 - disallow cpu events for unpriv
* 2 - disallow kernel profiling for unpriv
*/
int sysctl_perf_event_paranoid __read_mostly = 1;
So, with the default being 1, a non-root user can trace his stuff:
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
1
[acme@zoo ~]$ yes > /dev/null &
[1] 15338
[acme@zoo ~]$ trace -p 15338 | head -5
0.005 ( 0.005 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.045 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.085 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.125 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.165 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
[acme@zoo ~]$
[acme@zoo ~]$ trace --duration 1 sleep 1
1002.148 (1001.218 ms): nanosleep(rqtp: 0x7fff46c79250 ) = 0
[acme@zoo ~]$
[acme@zoo ~]$ trace -- usleep 1 | tail -5
0.905 ( 0.002 ms): brk( ) = 0x1c82000
0.910 ( 0.003 ms): brk(brk: 0x1ca3000 ) = 0x1ca3000
0.913 ( 0.001 ms): brk( ) = 0x1ca3000
0.990 ( 0.059 ms): nanosleep(rqtp: 0x7fffe31a3280 ) = 0
0.995 ( 0.000 ms): exit_group(
[acme@zoo ~]$
But can't do system wide tracing:
[acme@zoo ~]$ trace
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 0
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
If the paranoid level is >= 2, i.e. turn this perf stuff off for !root users:
[acme@zoo ~]$ sudo sh -c 'echo 2 > /proc/sys/kernel/perf_event_paranoid'
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
2
[acme@zoo ~]$
[acme@zoo ~]$ trace usleep 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
If the user manages to get what he/she wants, convincing root not
to be paranoid at all...
[root@zoo ~]# echo -1 > /proc/sys/kernel/perf_event_paranoid
[root@zoo ~]# cat /proc/sys/kernel/perf_event_paranoid
-1
[root@zoo ~]#
[acme@zoo ~]$ ps -eo user,pid,comm | grep Xorg
root 729 Xorg
[acme@zoo ~]$
[acme@zoo ~]$ trace -a --duration 0.001 -e \!select,ioctl,writev | grep Xorg | head -5
23.143 ( 0.003 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.152 ( 0.004 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = 8
23.161 ( 0.002 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = -1 EAGAIN Resource temporarily unavailable
23.175 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.235 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
[acme@zoo ~]$
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-di28olfwd28rvkox7v3hqhu1@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-18 00:38:29 +04:00
goto out_error ;
2014-12-12 00:04:10 +03:00
out_error_mmap :
perf_evlist__strerror_mmap ( evlist , errno , errbuf , sizeof ( errbuf ) ) ;
goto out_error ;
perf trace: Improve messages related to /proc/sys/kernel/perf_event_paranoid
kernel/events/core.c has:
/*
* perf event paranoia level:
* -1 - not paranoid at all
* 0 - disallow raw tracepoint access for unpriv
* 1 - disallow cpu events for unpriv
* 2 - disallow kernel profiling for unpriv
*/
int sysctl_perf_event_paranoid __read_mostly = 1;
So, with the default being 1, a non-root user can trace his stuff:
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
1
[acme@zoo ~]$ yes > /dev/null &
[1] 15338
[acme@zoo ~]$ trace -p 15338 | head -5
0.005 ( 0.005 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.045 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.085 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.125 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.165 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
[acme@zoo ~]$
[acme@zoo ~]$ trace --duration 1 sleep 1
1002.148 (1001.218 ms): nanosleep(rqtp: 0x7fff46c79250 ) = 0
[acme@zoo ~]$
[acme@zoo ~]$ trace -- usleep 1 | tail -5
0.905 ( 0.002 ms): brk( ) = 0x1c82000
0.910 ( 0.003 ms): brk(brk: 0x1ca3000 ) = 0x1ca3000
0.913 ( 0.001 ms): brk( ) = 0x1ca3000
0.990 ( 0.059 ms): nanosleep(rqtp: 0x7fffe31a3280 ) = 0
0.995 ( 0.000 ms): exit_group(
[acme@zoo ~]$
But can't do system wide tracing:
[acme@zoo ~]$ trace
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 0
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
If the paranoid level is >= 2, i.e. turn this perf stuff off for !root users:
[acme@zoo ~]$ sudo sh -c 'echo 2 > /proc/sys/kernel/perf_event_paranoid'
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
2
[acme@zoo ~]$
[acme@zoo ~]$ trace usleep 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
If the user manages to get what he/she wants, convincing root not
to be paranoid at all...
[root@zoo ~]# echo -1 > /proc/sys/kernel/perf_event_paranoid
[root@zoo ~]# cat /proc/sys/kernel/perf_event_paranoid
-1
[root@zoo ~]#
[acme@zoo ~]$ ps -eo user,pid,comm | grep Xorg
root 729 Xorg
[acme@zoo ~]$
[acme@zoo ~]$ trace -a --duration 0.001 -e \!select,ioctl,writev | grep Xorg | head -5
23.143 ( 0.003 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.152 ( 0.004 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = 8
23.161 ( 0.002 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = -1 EAGAIN Resource temporarily unavailable
23.175 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.235 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
[acme@zoo ~]$
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-di28olfwd28rvkox7v3hqhu1@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-18 00:38:29 +04:00
out_error_open :
perf_evlist__strerror_open ( evlist , errno , errbuf , sizeof ( errbuf ) ) ;
out_error :
2013-10-17 19:07:58 +04:00
fprintf ( trace - > output , " %s \n " , errbuf ) ;
2013-10-04 09:17:31 +04:00
goto out_delete_evlist ;
perf evlist: Make perf_evlist__set_filter use perf_evsel__set_filter
Instead of calling perf_evsel__apply_filter straight away, so that
we can, in the next patches, expand the filter with more conditions
before actually calling the ioctl to pass the end result filter to
the kernel.
Now we need to call perf_evlist__apply_filters() after the filter
is completely setup, i.e. do the ioctl calls.
The perf_evlist__apply_filters() method was already in place, because
that is the model for the other tools that receives filters in the
command line: go on setting then in the evsel->filter and only at
the end, after parsing the whole command line, apply them.
We get, as a bonus, a more expressive message that states which
event, if any, failed to have the filter applied to, with an
error message stating what happened.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-f429pgz75ryz7tpe6v74etre@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-03 23:42:03 +03:00
out_error_apply_filters :
fprintf ( trace - > output ,
" Failed to set filter \" %s \" on event %s with %d (%s) \n " ,
evsel - > filter , perf_evsel__name ( evsel ) , errno ,
strerror_r ( errno , errbuf , sizeof ( errbuf ) ) ) ;
goto out_delete_evlist ;
2012-09-27 03:05:56 +04:00
}
2015-01-22 17:08:04 +03:00
out_error_mem :
fprintf ( trace - > output , " Not enough memory to run! \n " ) ;
goto out_delete_evlist ;
perf trace: Use event filters for the event qualifier list
We use raw_syscalls:sys_{enter,exit} events to show the syscalls, but were
using a rather lazy/inneficient way to implement our 'strace -e' equivalent:
filter out after reading the events in the ring buffer.
Deflect more work to the kernel by appending a filter expression for that,
that, together with the pid list, that is always present, if only to filter the
tracer itself, reduces pressure on the ring buffer and otherwise use
infrastructure already in place in the kernel to do early filtering.
If we use it with -v we can see the filter passed to the kernel,
for instance, for this contrieved case:
# trace -v -e \!open,close,write,poll,recvfrom,select,recvmsg,writev,sendmsg,read,futex,epoll_wait,ioctl,eventfd --filter-pids 2189,2566,1398,2692,4475,4532
<SNIP>
(common_pid != 2514 && common_pid != 1398 && common_pid != 2189 && common_pid != 2566 && common_pid != 2692 && common_pid != 4475 && common_pid != 4532) && (id != 3 && id != 232 && id != 284 && id != 202 && id != 16 && id != 2 && id != 7 && id != 0 && id != 45 && id != 47 && id != 23 && id != 46 && id != 1 && id != 20)
0.011 (0.011 ms): caribou/2295 eventfd2(flags: CLOEXEC|NONBLOCK) = 18
16.946 (0.019 ms): caribou/2295 eventfd2(flags: CLOEXEC|NONBLOCK) = 18
38.598 (0.167 ms): chronyd/794 socket(family: INET, type: DGRAM ) = 4
38.603 (0.002 ms): chronyd/794 fcntl(fd: 4<socket:[239307]>, cmd: GETFD) = 0
38.605 (0.001 ms): chronyd/794 fcntl(fd: 4<socket:[239307]>, cmd: SETFD, arg: 1) = 0
^C
#
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ti2tg18atproqpguc2moinp6@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-04 18:44:59 +03:00
out_errno :
fprintf ( trace - > output , " errno=%d,%s \n " , errno , strerror ( errno ) ) ;
goto out_delete_evlist ;
perf trace: Improve messages related to /proc/sys/kernel/perf_event_paranoid
kernel/events/core.c has:
/*
* perf event paranoia level:
* -1 - not paranoid at all
* 0 - disallow raw tracepoint access for unpriv
* 1 - disallow cpu events for unpriv
* 2 - disallow kernel profiling for unpriv
*/
int sysctl_perf_event_paranoid __read_mostly = 1;
So, with the default being 1, a non-root user can trace his stuff:
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
1
[acme@zoo ~]$ yes > /dev/null &
[1] 15338
[acme@zoo ~]$ trace -p 15338 | head -5
0.005 ( 0.005 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.045 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.085 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.125 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
0.165 ( 0.001 ms): write(fd: 1</dev/null>, buf: 0x7fe6db765000, count: 4096 ) = 4096
[acme@zoo ~]$
[acme@zoo ~]$ trace --duration 1 sleep 1
1002.148 (1001.218 ms): nanosleep(rqtp: 0x7fff46c79250 ) = 0
[acme@zoo ~]$
[acme@zoo ~]$ trace -- usleep 1 | tail -5
0.905 ( 0.002 ms): brk( ) = 0x1c82000
0.910 ( 0.003 ms): brk(brk: 0x1ca3000 ) = 0x1ca3000
0.913 ( 0.001 ms): brk( ) = 0x1ca3000
0.990 ( 0.059 ms): nanosleep(rqtp: 0x7fffe31a3280 ) = 0
0.995 ( 0.000 ms): exit_group(
[acme@zoo ~]$
But can't do system wide tracing:
[acme@zoo ~]$ trace
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 0
Error: Operation not permitted.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 1.
[acme@zoo ~]$
If the paranoid level is >= 2, i.e. turn this perf stuff off for !root users:
[acme@zoo ~]$ sudo sh -c 'echo 2 > /proc/sys/kernel/perf_event_paranoid'
[acme@zoo ~]$ cat /proc/sys/kernel/perf_event_paranoid
2
[acme@zoo ~]$
[acme@zoo ~]$ trace usleep 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
[acme@zoo ~]$ trace --cpu 1
Error: Permission denied.
Hint: Check /proc/sys/kernel/perf_event_paranoid setting.
Hint: For your workloads it needs to be <= 1
Hint: For system wide tracing it needs to be set to -1.
Hint: The current value is 2.
[acme@zoo ~]$
If the user manages to get what he/she wants, convincing root not
to be paranoid at all...
[root@zoo ~]# echo -1 > /proc/sys/kernel/perf_event_paranoid
[root@zoo ~]# cat /proc/sys/kernel/perf_event_paranoid
-1
[root@zoo ~]#
[acme@zoo ~]$ ps -eo user,pid,comm | grep Xorg
root 729 Xorg
[acme@zoo ~]$
[acme@zoo ~]$ trace -a --duration 0.001 -e \!select,ioctl,writev | grep Xorg | head -5
23.143 ( 0.003 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.152 ( 0.004 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = 8
23.161 ( 0.002 ms): Xorg/729 read(fd: 31, buf: 0x2544af0, count: 4096 ) = -1 EAGAIN Resource temporarily unavailable
23.175 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
23.235 ( 0.002 ms): Xorg/729 setitimer(which: REAL, value: 0x7fffaadf16e0 ) = 0
[acme@zoo ~]$
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-di28olfwd28rvkox7v3hqhu1@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-10-18 00:38:29 +04:00
}
2012-09-27 03:05:56 +04:00
2013-08-29 08:29:52 +04:00
static int trace__replay ( struct trace * trace )
{
const struct perf_evsel_str_handler handlers [ ] = {
2013-09-28 01:06:19 +04:00
{ " probe:vfs_getname " , trace__vfs_getname , } ,
2013-08-29 08:29:52 +04:00
} ;
2013-10-15 18:27:32 +04:00
struct perf_data_file file = {
. path = input_name ,
. mode = PERF_DATA_MODE_READ ,
perf trace: Support using -f to override perf.data file ownership
Enable perf trace to use perf.data when it is not owned by current user
or root.
Example:
# perf trace record ls
# chown Yunlong.Song:Yunlong.Song perf.data
# ls -al perf.data
-rw------- 1 Yunlong.Song Yunlong.Song 4153101 Apr 2 15:28 perf.data
# id
uid=0(root) gid=0(root) groups=0(root),64(pkcs11)
Before this patch:
# perf trace -i perf.data
File perf.data not owned by current user or root (use -f to override)
# perf trace -i perf.data -f
Error: unknown switch `f'
usage: perf trace [<options>] [<command>]
or: perf trace [<options>] -- <command> [<options>]
or: perf trace record [<options>] [<command>]
or: perf trace record [<options>] -- <command> [<options>]
--event <event> event selector. use 'perf list' to list
available events
--comm show the thread COMM next to its id
--tool_stats show tool stats
-e, --expr <expr> list of events to trace
-o, --output <file> output file name
-i, --input <file> Analyze events in file
-p, --pid <pid> trace events on existing process id
-t, --tid <tid> trace events on existing thread id
--filter-pids <float>
...
As shown above, the -f option does not work at all.
After this patch:
# perf trace -i perf.data
File perf.data not owned by current user or root (use -f to override)
# perf trace -i perf.data -f
0.056 ( 0.002 ms): ls/47325 brk( ...
0.108 ( 0.018 ms): ls/47325 mmap(len: 4096, prot: READ|WRITE, ...
0.145 ( 0.013 ms): ls/47325 access(filename: 0x7f31259a0eb0, ...
0.172 ( 0.008 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.180 ( 0.004 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.185 ( 0.004 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.189 ( 0.003 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.195 ( 0.004 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.199 ( 0.002 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.205 ( 0.004 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.211 ( 0.004 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.220 ( 0.007 ms): ls/47325 open(filename: 0x7f312599e8ff, ...
...
...
As shown above, the -f option really works now.
Signed-off-by: Yunlong Song <yunlong.song@huawei.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/1427982439-27388-10-git-send-email-yunlong.song@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-02 16:47:18 +03:00
. force = trace - > force ,
2013-10-15 18:27:32 +04:00
} ;
2013-08-29 08:29:52 +04:00
struct perf_session * session ;
2013-11-12 10:25:00 +04:00
struct perf_evsel * evsel ;
2013-08-29 08:29:52 +04:00
int err = - 1 ;
trace - > tool . sample = trace__process_sample ;
trace - > tool . mmap = perf_event__process_mmap ;
2013-09-23 05:44:58 +04:00
trace - > tool . mmap2 = perf_event__process_mmap2 ;
2013-08-29 08:29:52 +04:00
trace - > tool . comm = perf_event__process_comm ;
trace - > tool . exit = perf_event__process_exit ;
trace - > tool . fork = perf_event__process_fork ;
trace - > tool . attr = perf_event__process_attr ;
trace - > tool . tracing_data = perf_event__process_tracing_data ;
trace - > tool . build_id = perf_event__process_build_id ;
2014-07-06 16:18:21 +04:00
trace - > tool . ordered_events = true ;
2013-08-29 08:29:52 +04:00
trace - > tool . ordering_requires_timestamps = true ;
/* add tid to output */
trace - > multiple_threads = true ;
2013-10-15 18:27:32 +04:00
session = perf_session__new ( & file , false , & trace - > tool ) ;
2013-08-29 08:29:52 +04:00
if ( session = = NULL )
2014-09-24 05:33:37 +04:00
return - 1 ;
2013-08-29 08:29:52 +04:00
2014-08-12 10:40:45 +04:00
if ( symbol__init ( & session - > header . env ) < 0 )
2014-08-12 10:40:44 +04:00
goto out ;
2013-09-28 23:13:00 +04:00
trace - > host = & session - > machines . host ;
2013-08-29 08:29:52 +04:00
err = perf_session__set_tracepoints_handlers ( session , handlers ) ;
if ( err )
goto out ;
2013-11-12 10:25:00 +04:00
evsel = perf_evlist__find_tracepoint_by_name ( session - > evlist ,
" raw_syscalls:sys_enter " ) ;
2013-12-05 06:41:39 +04:00
/* older kernels have syscalls tp versus raw_syscalls */
if ( evsel = = NULL )
evsel = perf_evlist__find_tracepoint_by_name ( session - > evlist ,
" syscalls:sys_enter " ) ;
2013-11-12 10:25:00 +04:00
2014-06-26 20:14:28 +04:00
if ( evsel & &
( perf_evsel__init_syscall_tp ( evsel , trace__sys_enter ) < 0 | |
perf_evsel__init_sc_tp_ptr_field ( evsel , args ) ) ) {
2013-11-12 10:25:00 +04:00
pr_err ( " Error during initialize raw_syscalls:sys_enter event \n " ) ;
goto out ;
}
evsel = perf_evlist__find_tracepoint_by_name ( session - > evlist ,
" raw_syscalls:sys_exit " ) ;
2013-12-05 06:41:39 +04:00
if ( evsel = = NULL )
evsel = perf_evlist__find_tracepoint_by_name ( session - > evlist ,
" syscalls:sys_exit " ) ;
2014-06-26 20:14:28 +04:00
if ( evsel & &
( perf_evsel__init_syscall_tp ( evsel , trace__sys_exit ) < 0 | |
perf_evsel__init_sc_tp_uint_field ( evsel , ret ) ) ) {
2013-11-12 10:25:00 +04:00
pr_err ( " Error during initialize raw_syscalls:sys_exit event \n " ) ;
2013-08-29 08:29:52 +04:00
goto out ;
}
2014-06-26 20:14:26 +04:00
evlist__for_each ( session - > evlist , evsel ) {
if ( evsel - > attr . type = = PERF_TYPE_SOFTWARE & &
( evsel - > attr . config = = PERF_COUNT_SW_PAGE_FAULTS_MAJ | |
evsel - > attr . config = = PERF_COUNT_SW_PAGE_FAULTS_MIN | |
evsel - > attr . config = = PERF_COUNT_SW_PAGE_FAULTS ) )
evsel - > handler = trace__pgfault ;
}
2013-08-29 08:29:53 +04:00
err = parse_target_str ( trace ) ;
if ( err ! = 0 )
goto out ;
2013-08-29 08:29:52 +04:00
setup_pager ( ) ;
2015-03-03 17:58:45 +03:00
err = perf_session__process_events ( session ) ;
2013-08-29 08:29:52 +04:00
if ( err )
pr_err ( " Failed to process events, error %d " , err ) ;
2013-10-09 07:26:53 +04:00
else if ( trace - > summary )
trace__fprintf_thread_summary ( trace , trace - > output ) ;
2013-08-29 08:29:52 +04:00
out :
perf_session__delete ( session ) ;
return err ;
}
2012-10-18 00:13:12 +04:00
static size_t trace__fprintf_threads_header ( FILE * fp )
{
size_t printed ;
2013-11-12 18:42:14 +04:00
printed = fprintf ( fp , " \n Summary of events: \n \n " ) ;
2013-10-09 07:26:53 +04:00
return printed ;
}
static size_t thread__dump_stats ( struct thread_trace * ttrace ,
struct trace * trace , FILE * fp )
{
struct stats * stats ;
size_t printed = 0 ;
struct syscall * sc ;
struct int_node * inode = intlist__first ( ttrace - > syscall_stats ) ;
if ( inode = = NULL )
return 0 ;
printed + = fprintf ( fp , " \n " ) ;
2013-11-13 16:21:48 +04:00
printed + = fprintf ( fp , " syscall calls min avg max stddev \n " ) ;
printed + = fprintf ( fp , " (msec) (msec) (msec) (%%) \n " ) ;
printed + = fprintf ( fp , " --------------- -------- --------- --------- --------- ------ \n " ) ;
2013-11-12 18:42:14 +04:00
2013-10-09 07:26:53 +04:00
/* each int_node is a syscall */
while ( inode ) {
stats = inode - > priv ;
if ( stats ) {
double min = ( double ) ( stats - > min ) / NSEC_PER_MSEC ;
double max = ( double ) ( stats - > max ) / NSEC_PER_MSEC ;
double avg = avg_stats ( stats ) ;
double pct ;
u64 n = ( u64 ) stats - > n ;
pct = avg ? 100.0 * stddev_stats ( stats ) / avg : 0.0 ;
avg / = NSEC_PER_MSEC ;
sc = & trace - > syscalls . table [ inode - > i ] ;
2013-11-12 18:42:14 +04:00
printed + = fprintf ( fp , " %-15s " , sc - > name ) ;
2013-11-13 16:21:48 +04:00
printed + = fprintf ( fp , " %8 " PRIu64 " %9.3f %9.3f " ,
2013-11-12 18:10:10 +04:00
n , min , avg ) ;
2013-11-13 16:21:48 +04:00
printed + = fprintf ( fp , " %9.3f %9.2f%% \n " , max , pct ) ;
2013-10-09 07:26:53 +04:00
}
inode = intlist__next ( inode ) ;
}
printed + = fprintf ( fp , " \n \n " ) ;
2012-10-18 00:13:12 +04:00
return printed ;
}
2013-09-28 23:12:59 +04:00
/* struct used to pass data to per-thread function */
struct summary_data {
FILE * fp ;
struct trace * trace ;
size_t printed ;
} ;
static int trace__fprintf_one_thread ( struct thread * thread , void * priv )
{
struct summary_data * data = priv ;
FILE * fp = data - > fp ;
size_t printed = data - > printed ;
struct trace * trace = data - > trace ;
2014-10-06 04:46:03 +04:00
struct thread_trace * ttrace = thread__priv ( thread ) ;
2013-09-28 23:12:59 +04:00
double ratio ;
if ( ttrace = = NULL )
return 0 ;
ratio = ( double ) ttrace - > nr_events / trace - > nr_events * 100.0 ;
2013-11-14 20:43:30 +04:00
printed + = fprintf ( fp , " %s (%d), " , thread__comm_str ( thread ) , thread - > tid ) ;
2013-11-12 18:42:14 +04:00
printed + = fprintf ( fp , " %lu events, " , ttrace - > nr_events ) ;
2013-11-14 20:43:30 +04:00
printed + = fprintf ( fp , " %.1f%% " , ratio ) ;
2014-07-08 22:05:16 +04:00
if ( ttrace - > pfmaj )
printed + = fprintf ( fp , " , %lu majfaults " , ttrace - > pfmaj ) ;
if ( ttrace - > pfmin )
printed + = fprintf ( fp , " , %lu minfaults " , ttrace - > pfmin ) ;
2013-11-12 18:42:14 +04:00
printed + = fprintf ( fp , " , %.3f msec \n " , ttrace - > runtime_ms ) ;
2013-10-09 07:26:53 +04:00
printed + = thread__dump_stats ( ttrace , trace , fp ) ;
2013-09-28 23:12:59 +04:00
data - > printed + = printed ;
return 0 ;
}
2012-10-18 00:13:12 +04:00
static size_t trace__fprintf_thread_summary ( struct trace * trace , FILE * fp )
{
2013-09-28 23:12:59 +04:00
struct summary_data data = {
. fp = fp ,
. trace = trace
} ;
data . printed = trace__fprintf_threads_header ( fp ) ;
2012-10-18 00:13:12 +04:00
2013-09-28 23:12:59 +04:00
machine__for_each_thread ( trace - > host , trace__fprintf_one_thread , & data ) ;
return data . printed ;
2012-10-18 00:13:12 +04:00
}
perf trace: Add duration filter
Example:
[acme@sandy linux]$ perf trace --duration 0.025 usleep 1
2.221 ( 0.958 ms): 6724 execve(arg0: 140733557168278, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = -2
3.690 ( 1.443 ms): 6724 execve(arg0: 140733557168295, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = 0
3.979 ( 0.048 ms): 6724 open(filename: 208733843841, flags: 0, mode: 1 ) = 3
4.071 ( 0.075 ms): 6724 open(filename: 139744419925673, flags: 0, mode: 0 ) = 3
4.318 ( 0.056 ms): 6724 nanosleep(rqtp: 140734030404608, rmtp: 0 ) = 0
[acme@sandy linux]$ perf trace --duration 0.100 usleep 1
1.143 ( 1.021 ms): 6726 execve(arg0: 140736323962279, arg1: 140736323972752, arg2: 34926752, arg3: 140736323961824, arg4: 7955998171588342573, arg5: 6725) = 0
[acme@sandy linux]$
Cherry picked from tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-oslw2j2958we9qf0ctra4whd@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-08 16:56:00 +04:00
static int trace__set_duration ( const struct option * opt , const char * str ,
int unset __maybe_unused )
{
struct trace * trace = opt - > value ;
trace - > duration_filter = atof ( str ) ;
return 0 ;
}
perf trace: Introduce --filter-pids
When tracing in X we get event loops due to the tracing activity, i.e.
updates to a gnome-terminal that generate syscalls for X.org, etc.
To get a more useful view of what is happening, syscall wise, system
wide, we need to filter those, like in:
# ps ax|egrep '981|2296|1519' | grep -v egrep
981 tty1 Ss+ 5:40 /usr/bin/Xorg :0 -background none ...
1519 ? Sl 2:22 /usr/bin/gnome-shell
2296 ? Sl 4:16 /usr/libexec/gnome-terminal-server
#
# trace -e write --filter-pids 981,2296,1519
0.385 ( 0.021 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 136) = 136
0.922 ( 0.014 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 140) = 140
5006.525 ( 0.029 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 136) = 136
5007.235 ( 0.023 ms): goa-daemon/2061 write(fd: 1</dev/null>, buf: 0x7fbeb017b000, count: 140) = 140
5177.646 ( 0.018 ms): rtkit-daemon/782 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7f7eea70be88, count: 8) = 8
8314.497 ( 0.004 ms): gsd-locate-poi/2084 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7fffe96af7b0, count: 8) = 8
8314.518 ( 0.002 ms): gsd-locate-poi/2084 write(fd: 5<anon_inode:[eventfd]>, buf: 0x7fffe96af0e0, count: 8) = 8
^C#
When this option is used the tracer pid is also filtered.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-f5qmiyy7c0uxdm21ncatpeek@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-21 22:36:52 +03:00
static int trace__set_filter_pids ( const struct option * opt , const char * str ,
int unset __maybe_unused )
{
int ret = - 1 ;
size_t i ;
struct trace * trace = opt - > value ;
/*
* FIXME : introduce a intarray class , plain parse csv and create a
* { int nr , int entries [ ] } struct . . .
*/
struct intlist * list = intlist__new ( str ) ;
if ( list = = NULL )
return - 1 ;
i = trace - > filter_pids . nr = intlist__nr_entries ( list ) + 1 ;
trace - > filter_pids . entries = calloc ( i , sizeof ( pid_t ) ) ;
if ( trace - > filter_pids . entries = = NULL )
goto out ;
trace - > filter_pids . entries [ 0 ] = getpid ( ) ;
for ( i = 1 ; i < trace - > filter_pids . nr ; + + i )
trace - > filter_pids . entries [ i ] = intlist__entry ( list , i - 1 ) - > i ;
intlist__delete ( list ) ;
ret = 0 ;
out :
return ret ;
}
2013-08-19 19:01:10 +04:00
static int trace__open_output ( struct trace * trace , const char * filename )
{
struct stat st ;
if ( ! stat ( filename , & st ) & & st . st_size ) {
char oldname [ PATH_MAX ] ;
scnprintf ( oldname , sizeof ( oldname ) , " %s.old " , filename ) ;
unlink ( oldname ) ;
rename ( filename , oldname ) ;
}
trace - > output = fopen ( filename , " w " ) ;
return trace - > output = = NULL ? - errno : 0 ;
}
2014-06-26 20:14:25 +04:00
static int parse_pagefaults ( const struct option * opt , const char * str ,
int unset __maybe_unused )
{
int * trace_pgfaults = opt - > value ;
if ( strcmp ( str , " all " ) = = 0 )
* trace_pgfaults | = TRACE_PFMAJ | TRACE_PFMIN ;
else if ( strcmp ( str , " maj " ) = = 0 )
* trace_pgfaults | = TRACE_PFMAJ ;
else if ( strcmp ( str , " min " ) = = 0 )
* trace_pgfaults | = TRACE_PFMIN ;
else
return - 1 ;
return 0 ;
}
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
static void evlist__set_evsel_handler ( struct perf_evlist * evlist , void * handler )
{
struct perf_evsel * evsel ;
evlist__for_each ( evlist , evsel )
evsel - > handler = handler ;
}
2012-09-27 03:05:56 +04:00
int cmd_trace ( int argc , const char * * argv , const char * prefix __maybe_unused )
{
2015-03-18 16:35:57 +03:00
const char * trace_usage [ ] = {
2012-10-05 09:02:16 +04:00
" perf trace [<options>] [<command>] " ,
" perf trace [<options>] -- <command> [<options>] " ,
2013-09-28 23:13:01 +04:00
" perf trace record [<options>] [<command>] " ,
" perf trace record [<options>] -- <command> [<options>] " ,
2012-09-27 03:05:56 +04:00
NULL
} ;
struct trace trace = {
2013-09-28 01:06:19 +04:00
. audit = {
. machine = audit_detect_machine ( ) ,
. open_id = audit_name_to_syscall ( " open " , trace . audit . machine ) ,
} ,
2012-09-27 03:05:56 +04:00
. syscalls = {
. max = - 1 ,
} ,
. opts = {
. target = {
. uid = UINT_MAX ,
. uses_mmap = true ,
} ,
. user_freq = UINT_MAX ,
. user_interval = ULLONG_MAX ,
2014-01-15 00:52:14 +04:00
. no_buffering = true ,
2014-12-12 23:28:32 +03:00
. mmap_pages = UINT_MAX ,
2015-06-17 16:51:11 +03:00
. proc_map_timeout = 500 ,
2012-09-27 03:05:56 +04:00
} ,
2013-08-19 19:01:10 +04:00
. output = stdout ,
perf trace: Add option to show process COMM
Enabled by default, disable with --no-comm, e.g.:
181.821 (0.001 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: PEEK|TRUNC|CMSG_CLOEXEC ) = 20
181.824 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.825 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.834 (0.002 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: CMSG_CLOEXEC ) = 20
181.836 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.838 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.705 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 1256
181.710 (0.002 ms): evolution-addr/10924 geteuid( ) = 1000
181.712 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.727 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 1256
181.731 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.734 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.908 (0.002 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 20
181.913 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.915 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.930 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 20
181.934 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.937 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
220.718 (0.010 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.741 (0.000 ms): dbus-daemon/10711 ... [continued]: epoll_wait()) = 1
220.759 (0.004 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.780 (0.002 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.788 (0.001 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = -1 EAGAIN Resource temporarily unavailable
220.760 (0.004 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.771 (0.023 ms): perf/26347 open(filename: 0xf2e780, mode: 15918976 ) = 19
220.850 (0.002 ms): perf/26347 close(fd: 19 ) = 0
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-6be5jvnkdzjptdrebfn5263n@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-12 19:35:21 +04:00
. show_comm = true ,
2014-06-26 20:14:28 +04:00
. trace_syscalls = true ,
2012-09-27 03:05:56 +04:00
} ;
2013-08-19 19:01:10 +04:00
const char * output_name = NULL ;
2013-08-09 19:28:31 +04:00
const char * ev_qualifier_str = NULL ;
2012-09-27 03:05:56 +04:00
const struct option trace_options [ ] = {
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
OPT_CALLBACK ( 0 , " event " , & trace . evlist , " event " ,
" event selector. use 'perf list' to list available events " ,
parse_events_option ) ,
perf trace: Add option to show process COMM
Enabled by default, disable with --no-comm, e.g.:
181.821 (0.001 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: PEEK|TRUNC|CMSG_CLOEXEC ) = 20
181.824 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.825 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.834 (0.002 ms): deja-dup-monit/10784 recvmsg(fd: 8, msg: 0x7fff4342baf0, flags: CMSG_CLOEXEC ) = 20
181.836 (0.001 ms): deja-dup-monit/10784 geteuid( ) = 1000
181.838 (0.001 ms): deja-dup-monit/10784 getegid( ) = 1000
181.705 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 1256
181.710 (0.002 ms): evolution-addr/10924 geteuid( ) = 1000
181.712 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.727 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 1256
181.731 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.734 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.908 (0.002 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: PEEK|TRUNC|CMSG_CLOEXEC) = 20
181.913 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.915 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
181.930 (0.003 ms): evolution-addr/10924 recvmsg(fd: 10, msg: 0x7fff17dc6990, flags: CMSG_CLOEXEC ) = 20
181.934 (0.001 ms): evolution-addr/10924 geteuid( ) = 1000
181.937 (0.001 ms): evolution-addr/10924 getegid( ) = 1000
220.718 (0.010 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.741 (0.000 ms): dbus-daemon/10711 ... [continued]: epoll_wait()) = 1
220.759 (0.004 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.780 (0.002 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = 200
220.788 (0.001 ms): dbus-daemon/10711 recvmsg(fd: 11, msg: 0x7ffff94594d0, flags: CMSG_CLOEXEC ) = -1 EAGAIN Resource temporarily unavailable
220.760 (0.004 ms): at-spi2-regist/10715 sendmsg(fd: 3, msg: 0x7fffdb8756c0, flags: NOSIGNAL ) = 200
220.771 (0.023 ms): perf/26347 open(filename: 0xf2e780, mode: 15918976 ) = 19
220.850 (0.002 ms): perf/26347 close(fd: 19 ) = 0
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-6be5jvnkdzjptdrebfn5263n@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-09-12 19:35:21 +04:00
OPT_BOOLEAN ( 0 , " comm " , & trace . show_comm ,
" show the thread COMM next to its id " ) ,
2013-09-28 01:06:19 +04:00
OPT_BOOLEAN ( 0 , " tool_stats " , & trace . show_tool_stats , " show tool stats " ) ,
2015-04-23 18:02:07 +03:00
OPT_STRING ( ' e ' , " expr " , & ev_qualifier_str , " expr " , " list of syscalls to trace " ) ,
2013-08-19 19:01:10 +04:00
OPT_STRING ( ' o ' , " output " , & output_name , " file " , " output file name " ) ,
2013-08-29 08:29:52 +04:00
OPT_STRING ( ' i ' , " input " , & input_name , " file " , " Analyze events in file " ) ,
2012-09-27 03:05:56 +04:00
OPT_STRING ( ' p ' , " pid " , & trace . opts . target . pid , " pid " ,
" trace events on existing process id " ) ,
2013-08-20 21:15:45 +04:00
OPT_STRING ( ' t ' , " tid " , & trace . opts . target . tid , " tid " ,
2012-09-27 03:05:56 +04:00
" trace events on existing thread id " ) ,
2015-04-23 17:59:20 +03:00
OPT_CALLBACK ( 0 , " filter-pids " , & trace , " CSV list of pids " ,
" pids to filter (by the kernel) " , trace__set_filter_pids ) ,
2013-08-20 21:15:45 +04:00
OPT_BOOLEAN ( ' a ' , " all-cpus " , & trace . opts . target . system_wide ,
2012-09-27 03:05:56 +04:00
" system-wide collection from all CPUs " ) ,
2013-08-20 21:15:45 +04:00
OPT_STRING ( ' C ' , " cpu " , & trace . opts . target . cpu_list , " cpu " ,
2012-09-27 03:05:56 +04:00
" list of cpus to monitor " ) ,
2013-08-29 08:29:52 +04:00
OPT_BOOLEAN ( 0 , " no-inherit " , & trace . opts . no_inherit ,
2012-09-27 03:05:56 +04:00
" child tasks do not inherit counters " ) ,
2013-09-01 14:36:12 +04:00
OPT_CALLBACK ( ' m ' , " mmap-pages " , & trace . opts . mmap_pages , " pages " ,
" number of mmap data pages " ,
perf_evlist__parse_mmap_pages ) ,
2013-08-20 21:15:45 +04:00
OPT_STRING ( ' u ' , " uid " , & trace . opts . target . uid_str , " user " ,
2012-09-27 03:05:56 +04:00
" user to profile " ) ,
perf trace: Add duration filter
Example:
[acme@sandy linux]$ perf trace --duration 0.025 usleep 1
2.221 ( 0.958 ms): 6724 execve(arg0: 140733557168278, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = -2
3.690 ( 1.443 ms): 6724 execve(arg0: 140733557168295, arg1: 140733557178768, arg2: 16134304, arg3: 140733557167840, arg4: 7955998171588342573, arg5: 6723) = 0
3.979 ( 0.048 ms): 6724 open(filename: 208733843841, flags: 0, mode: 1 ) = 3
4.071 ( 0.075 ms): 6724 open(filename: 139744419925673, flags: 0, mode: 0 ) = 3
4.318 ( 0.056 ms): 6724 nanosleep(rqtp: 140734030404608, rmtp: 0 ) = 0
[acme@sandy linux]$ perf trace --duration 0.100 usleep 1
1.143 ( 1.021 ms): 6726 execve(arg0: 140736323962279, arg1: 140736323972752, arg2: 34926752, arg3: 140736323961824, arg4: 7955998171588342573, arg5: 6725) = 0
[acme@sandy linux]$
Cherry picked from tmp.perf/trace2 branch.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/tip-oslw2j2958we9qf0ctra4whd@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-10-08 16:56:00 +04:00
OPT_CALLBACK ( 0 , " duration " , & trace , " float " ,
" show only events with duration > N.M ms " ,
trace__set_duration ) ,
2012-10-18 00:13:12 +04:00
OPT_BOOLEAN ( 0 , " sched " , & trace . sched , " show blocking scheduler events " ) ,
2013-08-22 23:49:54 +04:00
OPT_INCR ( ' v ' , " verbose " , & verbose , " be more verbose " ) ,
2013-09-04 22:37:43 +04:00
OPT_BOOLEAN ( ' T ' , " time " , & trace . full_time ,
" Show full timestamp, not time relative to first start " ) ,
2013-11-12 20:31:15 +04:00
OPT_BOOLEAN ( ' s ' , " summary " , & trace . summary_only ,
" Show only syscall summary with statistics " ) ,
OPT_BOOLEAN ( ' S ' , " with-summary " , & trace . summary ,
" Show all syscalls and summary with statistics " ) ,
2014-06-26 20:14:25 +04:00
OPT_CALLBACK_DEFAULT ( ' F ' , " pf " , & trace . trace_pgfaults , " all|maj|min " ,
" Trace pagefaults " , parse_pagefaults , " maj " ) ,
2014-06-26 20:14:28 +04:00
OPT_BOOLEAN ( 0 , " syscalls " , & trace . trace_syscalls , " Trace syscalls " ) ,
perf trace: Support using -f to override perf.data file ownership
Enable perf trace to use perf.data when it is not owned by current user
or root.
Example:
# perf trace record ls
# chown Yunlong.Song:Yunlong.Song perf.data
# ls -al perf.data
-rw------- 1 Yunlong.Song Yunlong.Song 4153101 Apr 2 15:28 perf.data
# id
uid=0(root) gid=0(root) groups=0(root),64(pkcs11)
Before this patch:
# perf trace -i perf.data
File perf.data not owned by current user or root (use -f to override)
# perf trace -i perf.data -f
Error: unknown switch `f'
usage: perf trace [<options>] [<command>]
or: perf trace [<options>] -- <command> [<options>]
or: perf trace record [<options>] [<command>]
or: perf trace record [<options>] -- <command> [<options>]
--event <event> event selector. use 'perf list' to list
available events
--comm show the thread COMM next to its id
--tool_stats show tool stats
-e, --expr <expr> list of events to trace
-o, --output <file> output file name
-i, --input <file> Analyze events in file
-p, --pid <pid> trace events on existing process id
-t, --tid <tid> trace events on existing thread id
--filter-pids <float>
...
As shown above, the -f option does not work at all.
After this patch:
# perf trace -i perf.data
File perf.data not owned by current user or root (use -f to override)
# perf trace -i perf.data -f
0.056 ( 0.002 ms): ls/47325 brk( ...
0.108 ( 0.018 ms): ls/47325 mmap(len: 4096, prot: READ|WRITE, ...
0.145 ( 0.013 ms): ls/47325 access(filename: 0x7f31259a0eb0, ...
0.172 ( 0.008 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.180 ( 0.004 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.185 ( 0.004 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.189 ( 0.003 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.195 ( 0.004 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.199 ( 0.002 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.205 ( 0.004 ms): ls/47325 open(filename: 0x7fffeb9a0d00, ...
0.211 ( 0.004 ms): ls/47325 stat(filename: 0x7fffeb9a0d00, ...
0.220 ( 0.007 ms): ls/47325 open(filename: 0x7f312599e8ff, ...
...
...
As shown above, the -f option really works now.
Signed-off-by: Yunlong Song <yunlong.song@huawei.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/1427982439-27388-10-git-send-email-yunlong.song@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-02 16:47:18 +03:00
OPT_BOOLEAN ( ' f ' , " force " , & trace . force , " don't complain, do it " ) ,
2015-06-17 16:51:11 +03:00
OPT_UINTEGER ( 0 , " proc-map-timeout " , & trace . opts . proc_map_timeout ,
" per thread proc mmap processing timeout in ms " ) ,
2012-09-27 03:05:56 +04:00
OPT_END ( )
} ;
2015-03-18 16:35:57 +03:00
const char * const trace_subcommands [ ] = { " record " , NULL } ;
2012-09-27 03:05:56 +04:00
int err ;
2012-10-05 09:02:13 +04:00
char bf [ BUFSIZ ] ;
2012-09-27 03:05:56 +04:00
2015-02-24 21:35:55 +03:00
signal ( SIGSEGV , sighandler_dump_stack ) ;
signal ( SIGFPE , sighandler_dump_stack ) ;
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
trace . evlist = perf_evlist__new ( ) ;
if ( trace . evlist = = NULL ) {
pr_err ( " Not enough memory to run! \n " ) ;
2015-05-11 15:28:36 +03:00
err = - ENOMEM ;
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
goto out ;
}
2015-03-18 16:35:57 +03:00
argc = parse_options_subcommand ( argc , argv , trace_options , trace_subcommands ,
trace_usage , PARSE_OPT_STOP_AT_NON_OPTION ) ;
2013-11-12 20:31:15 +04:00
2014-06-26 20:14:25 +04:00
if ( trace . trace_pgfaults ) {
trace . opts . sample_address = true ;
trace . opts . sample_time = true ;
}
perf trace: Allow mixing with other events
Basically adopting 'perf record' --event command line argument syntax:
# trace -e \!mprotect,mmap,munmap,open,close,read,fstat,access,arch_prctl --event sched:*switch,sched:*exec,sched:*exit usleep 1
0.048 ( ): sched:sched_process_exec:filename=/bin/usleep pid=24732 old_pid=24732)
0.078 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.430 (0.002 ms): usleep/24732 brk( ) = 0x78f000
0.434 (0.003 ms): usleep/24732 brk(brk: 0x7b0000 ) = 0x7b0000
0.438 (0.001 ms): usleep/24732 brk( ) = 0x7b0000
0.460 (0.004 ms): usleep/24732 nanosleep(rqtp: 0x7ffff3696a40) ...
0.460 ( ): sched:sched_switch:prev_comm=usleep prev_pid=24732 prev_prio=120 prev_state=S ==> next_comm=swapper/1 next_pid=0 next_prio=120)
0.515 (0.058 ms): usleep/24732 ... [continued]: nanosleep()) = 0
0.520 (0.000 ms): usleep/24732 exit_group(
0.550 ( ): sched:sched_process_exit:comm=usleep pid=24732 prio=120)
#
Next steps, probably in this order:
1) Use ordered_events code, the logic in trace needs the events to be
time ordered when needed, i.e. when multiple CPUs are involved.
2) Callchains!
3) Automatically account for interruptions when saying how long things
took.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gpst8mph575yb4wgf91qibyb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-02-03 18:58:57 +03:00
if ( trace . evlist - > nr_entries > 0 )
evlist__set_evsel_handler ( trace . evlist , trace__event_handler ) ;
2014-06-26 20:14:26 +04:00
if ( ( argc > = 1 ) & & ( strcmp ( argv [ 0 ] , " record " ) = = 0 ) )
return trace__record ( & trace , argc - 1 , & argv [ 1 ] ) ;
/* summary_only implies summary option, but don't overwrite summary if set */
if ( trace . summary_only )
trace . summary = trace . summary_only ;
2015-02-06 12:16:45 +03:00
if ( ! trace . trace_syscalls & & ! trace . trace_pgfaults & &
trace . evlist - > nr_entries = = 0 /* Was --events used? */ ) {
2014-06-26 20:14:28 +04:00
pr_err ( " Please specify something to trace. \n " ) ;
return - 1 ;
}
2013-08-19 19:01:10 +04:00
if ( output_name ! = NULL ) {
err = trace__open_output ( & trace , output_name ) ;
if ( err < 0 ) {
perror ( " failed to create output file " ) ;
goto out ;
}
}
2013-08-09 19:28:31 +04:00
if ( ev_qualifier_str ! = NULL ) {
2013-08-21 19:56:21 +04:00
const char * s = ev_qualifier_str ;
perf trace: Support 'strace' syscall event groups
I.e.:
$ cat ~/share/perf-core/strace/groups/file
access
chmod
creat
execve
faccessat
getcwd
lstat
mkdir
open
openat
quotactl
readlink
rename
rmdir
stat
statfs
symlink
unlink
$
Then, on a quiet desktop, try running this and then moving your mouse to
see the deluge of mouse related activity:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
# trace --ev probe:vfs_getname --filter-pids 2232 -e file
0.042 (0.042 ms): mousetweaks/2235 open(filename: 0x14e3910, mode: 438 ) ...
0.042 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/home/acme/.icons/Adwaita/cursors/xterm")
0.100 (0.100 ms): mousetweaks/2235 ... [continued]: open()) = -1 ENOENT No such file or directory
0.142 (0.018 ms): mousetweaks/2235 open(filename: 0x14c3c10, mode: 438 ) ...
0.142 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/home/acme/.icons/Adwaita/index.theme")
0.192 (0.069 ms): mousetweaks/2235 ... [continued]: open()) = -1 ENOENT No such file or directory
0.230 (0.017 ms): mousetweaks/2235 open(filename: 0x14c3c10, mode: 438 ) ...
0.230 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/usr/share/icons/Adwaita/cursors/xterm")
0.253 (0.041 ms): mousetweaks/2235 ... [continued]: open()) = 14
0.459 (0.008 ms): mousetweaks/2235 open(filename: 0x14e3910, mode: 438 ) ...
0.459 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/home/acme/.icons/Adwaita/cursors/left_side")
0.468 (0.017 ms): mousetweaks/2235 ... [continued]: open()) = -1 ENOENT No such file or directory
Need to combine that raw_syscalls:sys_enter(open) + probe:vfs_getname +
raw_syscalls:sys_exit(open) sequence...
Now, if you're bored, please write some more syscall groups, like the ones
in 'strace' and send it our way :-)
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-a42xklu59lcbxp7bbnic74a8@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-20 18:02:09 +03:00
struct strlist_config slist_config = {
. dirname = system_path ( STRACE_GROUPS_DIR ) ,
} ;
2013-08-21 19:56:21 +04:00
trace . not_ev_qualifier = * s = = ' ! ' ;
if ( trace . not_ev_qualifier )
+ + s ;
perf trace: Support 'strace' syscall event groups
I.e.:
$ cat ~/share/perf-core/strace/groups/file
access
chmod
creat
execve
faccessat
getcwd
lstat
mkdir
open
openat
quotactl
readlink
rename
rmdir
stat
statfs
symlink
unlink
$
Then, on a quiet desktop, try running this and then moving your mouse to
see the deluge of mouse related activity:
# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
Added new event:
probe:vfs_getname (on getname_flags:72 with pathname=filename:string)
You can now use it in all perf tools, such as:
perf record -e probe:vfs_getname -aR sleep 1
#
# trace --ev probe:vfs_getname --filter-pids 2232 -e file
0.042 (0.042 ms): mousetweaks/2235 open(filename: 0x14e3910, mode: 438 ) ...
0.042 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/home/acme/.icons/Adwaita/cursors/xterm")
0.100 (0.100 ms): mousetweaks/2235 ... [continued]: open()) = -1 ENOENT No such file or directory
0.142 (0.018 ms): mousetweaks/2235 open(filename: 0x14c3c10, mode: 438 ) ...
0.142 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/home/acme/.icons/Adwaita/index.theme")
0.192 (0.069 ms): mousetweaks/2235 ... [continued]: open()) = -1 ENOENT No such file or directory
0.230 (0.017 ms): mousetweaks/2235 open(filename: 0x14c3c10, mode: 438 ) ...
0.230 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/usr/share/icons/Adwaita/cursors/xterm")
0.253 (0.041 ms): mousetweaks/2235 ... [continued]: open()) = 14
0.459 (0.008 ms): mousetweaks/2235 open(filename: 0x14e3910, mode: 438 ) ...
0.459 ( ): probe:vfs_getname:(ffffffff812230bc) pathname="/home/acme/.icons/Adwaita/cursors/left_side")
0.468 (0.017 ms): mousetweaks/2235 ... [continued]: open()) = -1 ENOENT No such file or directory
Need to combine that raw_syscalls:sys_enter(open) + probe:vfs_getname +
raw_syscalls:sys_exit(open) sequence...
Now, if you're bored, please write some more syscall groups, like the ones
in 'strace' and send it our way :-)
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Milian Wolff <mail@milianw.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-a42xklu59lcbxp7bbnic74a8@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-07-20 18:02:09 +03:00
trace . ev_qualifier = strlist__new ( s , & slist_config ) ;
2013-08-09 19:28:31 +04:00
if ( trace . ev_qualifier = = NULL ) {
2013-08-19 19:01:10 +04:00
fputs ( " Not enough memory to parse event qualifier " ,
trace . output ) ;
err = - ENOMEM ;
goto out_close ;
2013-08-09 19:28:31 +04:00
}
perf trace: Validate syscall list passed via -e argument
The 'trace' tool was accepting any names passed and just looking if
syscalls returned via the raw_syscalls:* tracepoints were in that list,
leading to it accepting perf events and then never finding any, as those
are not valid syscall names, confusing users.
Fix it by checking each entry in the list using audit_name_to_syscall,
telling the user which entries are invalid and suggesting where to look
for valid syscall names.
E.g:
[root@zoo ~]# trace -e open,foo,bar,close,baz
Error: Invalid syscall bar, baz, foo
Hint: try 'perf list syscalls:sys_enter_*'
Hint: and: 'man syscalls'
[root@zoo ~]#
Reported-by: Flavio Leitner <fbl@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-4g1i3m1z6fzsrznn2umi02wa@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-06-26 01:32:33 +03:00
err = trace__validate_ev_qualifier ( & trace ) ;
if ( err )
goto out_close ;
2013-08-09 19:28:31 +04:00
}
2013-11-12 23:46:16 +04:00
err = target__validate ( & trace . opts . target ) ;
2012-10-05 09:02:13 +04:00
if ( err ) {
2013-11-12 23:46:16 +04:00
target__strerror ( & trace . opts . target , err , bf , sizeof ( bf ) ) ;
2013-08-19 19:01:10 +04:00
fprintf ( trace . output , " %s " , bf ) ;
goto out_close ;
2012-10-05 09:02:13 +04:00
}
2013-11-12 23:46:16 +04:00
err = target__parse_uid ( & trace . opts . target ) ;
2012-09-27 03:05:56 +04:00
if ( err ) {
2013-11-12 23:46:16 +04:00
target__strerror ( & trace . opts . target , err , bf , sizeof ( bf ) ) ;
2013-08-19 19:01:10 +04:00
fprintf ( trace . output , " %s " , bf ) ;
goto out_close ;
2012-09-27 03:05:56 +04:00
}
2013-11-12 23:46:16 +04:00
if ( ! argc & & target__none ( & trace . opts . target ) )
2012-10-05 09:02:14 +04:00
trace . opts . target . system_wide = true ;
2013-08-29 08:29:52 +04:00
if ( input_name )
err = trace__replay ( & trace ) ;
else
err = trace__run ( & trace , argc , argv ) ;
2012-10-18 00:13:12 +04:00
2013-08-19 19:01:10 +04:00
out_close :
if ( output_name ! = NULL )
fclose ( trace . output ) ;
out :
2012-10-18 00:13:12 +04:00
return err ;
2012-09-27 03:05:56 +04:00
}