kernel/linux.git/tools/perf/util/data-convert-bt.c, branch linux-4.7.y

perf ctf: Convert invalid chars in a string before set value

2016-05-27T15:08:40+00:00

We observed some crazy apps on Android set their comm to unprintable string. For example: # cat /proc/10607/task/*/comm tencent.qqmusic ... Binder_2 日志输出线 <-- Chinese word 'log output thread' WifiManager ... 'perf data convert' fails to convert perf.data with such string to CTF format. For example: # cat << EOF > ./badguy.c #include int main(int argc, char *argv[]) { prctl(PR_SET_NAME, "\xe6\x97\xa5\xe5\xbf\x97\xe8\xbe\x93\xe5\x87\xba\xe7\xba\xbf"); while(1) sleep(1); return 0; } EOF # gcc ./badguy.c # perf record -e sched:* ./a.out # perf data convert --to-ctf ./bad.ctf CTF stream 4 flush failed [ perf data convert: Converted 'perf.data' into CTF data './bad.ctf' ] [ perf data convert: Converted and wrote 0.008 MB (78 samples) ] # babeltrace ./bad.ctf/ [error] Packet size (18446744073709551615 bits) is larger than remaining file size (262144 bits). [error] Stream index creation error. [error] Open file stream error. [warning] [Context] Cannot open_trace of format ctf at path ./bad.ctf. [warning] [Context] cannot open trace "./bad.ctf" from ./bad.ctf/ for reading. [error] Cannot open any trace for reading. [error] opening trace "./bad.ctf/" for reading. [error] none of the specified trace paths could be opened. This patch converts unprintable characters to hexadecimal word. After applying this patch the above test works correctly: # ~/perf data convert --to-ctf ./good.ctf [ perf data convert: Converted 'perf.data' into CTF data './good.ctf' ] [ perf data convert: Converted and wrote 0.008 MB (78 samples) ] # babeltrace ./good.ctf .. [23:14:35.491665268] (+0.000001100) sched:sched_wakeup: { cpu_id = 4 }, { perf_ip = 0xFFFFFFFF810AEF33, perf_tid = 0, perf_pid = 0, perf_id = 5123, perf_period = 1, common_type = 270, common_flags = 45, common_preempt_count = 4, common_pid = 0, comm = "\xe6\x97\xa5\xe5\xbf\x97\xe8\xbe\x93\xe5\x87\xba\xe7\xba\xbf", pid = 1057, prio = 120, success = 1, target_cpu = 4 } [23:14:35.491666230] (+0.000000962) sched:sched_wakeup: { cpu_id = 4 }, { perf_ip = 0xFFFFFFFF810AEF33, perf_tid = 0, perf_pid = 0, perf_id = 5122, perf_period = 1, common_type = 270, common_flags = 45, common_preempt_count = 4, common_pid = 0, comm = "\xe6\x97\xa5\xe5\xbf\x97\xe8\xbe\x93\xe5\x87\xba\xe7\xba\xbf", pid = 1057, prio = 120, success = 1, target_cpu = 4 } .. Committer note: To build perf with libabeltrace, use: $ mkdir -p /tmp/build/perf $ make LIBBABELTRACE=1 LIBBABELTRACE_DIR=/usr/local O=/tmp/build/perf -C tools/perf install-bin Or equivalent (no O=, fixup LIBBABELTRACE_DIR, etc). Signed-off-by: Wang Nan Tested-by: Arnaldo Carvalho de Melo Cc: Jiri Olsa Cc: pi3orama@163.com Link: http://lkml.kernel.org/r/1464348951-179595-1-git-send-email-wangnan0@huawei.com Signed-off-by: Arnaldo Carvalho de Melo

perf tools: Remove misplaced __maybe_unused

2016-03-23T15:03:04+00:00

All over the tree. Cc: David Ahern cc: Jiri Olsa Cc: Namhyung Kim Cc: Ravi Bangoria Link: http://lkml.kernel.org/n/tip-8nzhnokxyp8y4v7gf0j00oyb@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo

perf data: Explicitly set byte order for integer types

2016-03-03T14:10:34+00:00

After babeltrace commit 5cec03e402aa ("ir: copy variants and sequences when setting a field path"), 'perf data convert' gets incorrect result if there's bpf output data. For example: # perf data convert --to-ctf ./out.ctf # babeltrace ./out.ctf [10:44:31.186045346] (+?.?????????) evt: { cpu_id = 0 }, { perf_ip = 0xFFFFFFFF810E7DD1, perf_tid = 23819, perf_pid = 23819, perf_id = 518, raw_len = 3, raw_data = [ [0] = 0xC028E32F, [1] = 0x815D0100, [2] = 0x1000000 ] } [10:44:31.286101003] (+0.100055657) evt: { cpu_id = 0 }, { perf_ip = 0xFFFFFFFF8105B609, perf_tid = 23819, perf_pid = 23819, perf_id = 518, raw_len = 3, raw_data = [ [0] = 0x35D9F1EB, [1] = 0x15D81, [2] = 0x2 ] } The expected result of the first sample should be: raw_data = [ [0] = 0x2FE328C0, [1] = 0x15D81, [2] = 0x1 ] } however, 'perf data convert' output big endian value to resuling CTF file. The reason is a internal change (or a bug?) of babeltrace. Before this patch, at the first add_bpf_output_values(), byte order of all integer type is uncertain (is 0, neither 1234 (le) nor 4321 (be)). It would be fixed by: perf_evlist__deliver_sample -> process_sample_event -> ctf_stream ... ->bt_ctf_trace_add_stream_class ->bt_ctf_field_type_structure_set_byte_order ->bt_ctf_field_type_integer_set_byte_order during creating the stream. However, the babeltrace commit mentioned above duplicates types in sequence to prevent potential conflict in following call stack and link the newly allocated type into the 'raw_data' sequence: perf_evlist__deliver_sample -> process_sample_event -> ctf_stream ... -> bt_ctf_trace_add_stream_class -> bt_ctf_stream_class_resolve_types ... -> bt_ctf_field_type_sequence_copy ->bt_ctf_field_type_integer_copy This happens before byte order setting, so only the newly allocated type is initialized, the byte order of original type perf choose to create the first raw_data is still uncertain. Byte order in CTF output is not related to byte order in perf.data. Setting it to anything other than BT_CTF_BYTE_ORDER_NATIVE solves this problem (only BT_CTF_BYTE_ORDER_NATIVE needs to be fixed). To reduce behavior changing, set byte order according to compiling options. Signed-off-by: Wang Nan Cc: Jeremie Galarneau Cc: Alexei Starovoitov Cc: Brendan Gregg Cc: Jiri Olsa Cc: Jérémie Galarneau Cc: Li Zefan Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Zefan Li Cc: pi3orama@163.com Link: http://lkml.kernel.org/r/1456479154-136027-10-git-send-email-wangnan0@huawei.com Signed-off-by: Arnaldo Carvalho de Melo

perf data: Support converting data from bpf_perf_event_output()

2016-03-03T14:10:34+00:00

bpf_perf_event_output() outputs data through sample->raw_data. This patch adds support to convert those data into CTF. A python script then can be used to process output data from BPF programs. Test result: # cat ./test_bpf_output_2.c /************************ BEGIN **************************/ #include struct bpf_map_def { unsigned int type; unsigned int key_size; unsigned int value_size; unsigned int max_entries; }; #define SEC(NAME) __attribute__((section(NAME), used)) static u64 (*ktime_get_ns)(void) = (void *)BPF_FUNC_ktime_get_ns; static int (*trace_printk)(const char *fmt, int fmt_size, ...) = (void *)BPF_FUNC_trace_printk; static int (*get_smp_processor_id)(void) = (void *)BPF_FUNC_get_smp_processor_id; static int (*perf_event_output)(void *, struct bpf_map_def *, int, void *, unsigned long) = (void *)BPF_FUNC_perf_event_output; struct bpf_map_def SEC("maps") channel = { .type = BPF_MAP_TYPE_PERF_EVENT_ARRAY, .key_size = sizeof(int), .value_size = sizeof(u32), .max_entries = __NR_CPUS__, }; static inline int __attribute__((always_inline)) func(void *ctx, int type) { struct { u64 ktime; int type; } __attribute__((packed)) output_data; char error_data[] = "Error: failed to output\n"; int err; output_data.type = type; output_data.ktime = ktime_get_ns(); err = perf_event_output(ctx, &channel, get_smp_processor_id(), &output_data, sizeof(output_data)); if (err) trace_printk(error_data, sizeof(error_data)); return 0; } SEC("func_begin=sys_nanosleep") int func_begin(void *ctx) {return func(ctx, 1);} SEC("func_end=sys_nanosleep%return") int func_end(void *ctx) { return func(ctx, 2);} char _license[] SEC("license") = "GPL"; int _version SEC("version") = LINUX_VERSION_CODE; /************************* END ***************************/ # ./perf record -e bpf-output/no-inherit,name=evt/ \ -e ./test_bpf_output_2.c/map:channel.event=evt/ \ usleep 100000 [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.012 MB perf.data (2 samples) ] # ./perf script usleep 14942 92503.198504: evt: ffffffff810e0ba1 sys_nanosleep (/lib/modules/4.3.0.... usleep 14942 92503.298562: evt: ffffffff810585e9 kretprobe_trampoline_holder (/lib.... # ./perf data convert --to-ctf ./out.ctf [ perf data convert: Converted 'perf.data' into CTF data './out.ctf' ] [ perf data convert: Converted and wrote 0.000 MB (2 samples) ] # babeltrace ./out.ctf [01:41:43.198504134] (+?.?????????) evt: { cpu_id = 0 }, { perf_ip = 0xFFFFFFFF810E0BA1, perf_tid = 14942, perf_pid = 14942, perf_id = 1044, raw_len = 3, raw_data = [ [0] = 0x32C0C07B, [1] = 0x5421, [2] = 0x1 ] } [01:41:43.298562257] (+0.100058123) evt: { cpu_id = 0 }, { perf_ip = 0xFFFFFFFF810585E9, perf_tid = 14942, perf_pid = 14942, perf_id = 1044, raw_len = 3, raw_data = [ [0] = 0x38B77FAA, [1] = 0x5421, [2] = 0x2 ] } # cat ./test_bpf_output_2.py from babeltrace import TraceCollection tc = TraceCollection() tc.add_trace('./out.ctf', 'ctf') d = {1:[], 2:[]} for event in tc.events: if not event.name.startswith('evt'): continue raw_data = event['raw_data'] (time, type) = ((raw_data[0] + (raw_data[1] << 32)), raw_data[2]) d[type].append(time) print(list(map(lambda i: d[2][i] - d[1][i], range(len(d[1]))))); # python3 ./test_bpf_output_2.py [100056879] Committer note: Make sure you have python3-devel installed, not python-devel, which may be for python2, which will lead to some "PyInstance_Type" errors. Also make sure that you use the right libbabeltrace, because it is shipped in Fedora, for instance, but an older version. To build libbabeltrace's python binding one also needs to use: ./configure --enable-python-bindings And then set PYTHONPATH=/usr/local/lib64/python3.4/site-packages/. Signed-off-by: Wang Nan Tested-by: Arnaldo Carvalho de Melo Acked-by: Jiri Olsa Cc: Alexei Starovoitov Cc: Brendan Gregg Cc: Li Zefan Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Zefan Li Cc: pi3orama@163.com Link: http://lkml.kernel.org/r/1456479154-136027-9-git-send-email-wangnan0@huawei.com Signed-off-by: Arnaldo Carvalho de Melo

perf config: Bring perf_default_config to the very beginning at main()

2016-02-26T22:49:16+00:00

Before this patch each subcommand calls perf_config() by themself, reading the default configuration together with subcommand specific options. If a subcommand doesn't have it own options, it needs to call 'perf_config(perf_default_config, NULL)' to ensure .perfconfig is loaded. This patch brings perf_config(perf_default_config, NULL) to the very start of main(), so subcommands don't need to do it. After this patch, 'llvm.clang-path' works for 'perf trace'. Signed-off-by: Wang Nan Suggested-and-Tested-by: Arnaldo Carvalho de Melo Cc: Alexei Starovoitov Cc: Jiri Olsa Cc: Li Zefan Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: pi3orama@163.com Link: http://lkml.kernel.org/r/1456479154-136027-4-git-send-email-wangnan0@huawei.com Signed-off-by: Arnaldo Carvalho de Melo

perf data: Fix releasing event_class

2016-02-12T20:27:48+00:00

A new patch in libbabeltrace [1] reveals a object leak problem in 'perf data' CTF support: perf code never releases the event_class which is allocated in add_event() and stored in evsel's private field. If libbabeltrace has the above patch applied, leaking event_class prevents the writer from being destroyed and flushing metadata. For example: $ perf record ls perf.data [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.012 MB perf.data (12 samples) ] $ perf data convert --to-ctf ./out.ctf [ perf data convert: Converted 'perf.data' into CTF data './out.ctf' ] [ perf data convert: Converted and wrote 0.000 MB (12 samples) ] $ cat ./out.ctf/metadata $ ls -l ./out.ctf/metadata -rw-r----- 1 w00229757 mm 0 Jan 27 10:49 ./out.ctf/metadata The correct result should be: ... $ cat ./out.ctf/metadata /* CTF 1.8 */ trace { [SNIP] $ ls -l ./out.ctf/metadata -rw-r----- 1 w00229757 mm 2446 Jan 27 10:52 ./out.ctf/metadata The full story is: Patch [1] of babeltrace redesigns its reference counting scheme. In that patch: * writer <- trace (bt_ctf_writer_create) * trace <- stream_class (bt_ctf_trace_add_stream_class) * stream_class <- event_class (bt_ctf_stream_class_add_event_class) ('<-' means 'is a parent of') Holding of event_class causes reference count of corresponding 'writer' to increase through parent chain. Perf expects that 'writer' is released (so metadata is flushed) through bt_ctf_writer_put() in ctf_writer__cleanup(). However, since it never releases event_class, the reference of 'writer' won't be dropped, so bt_ctf_writer_put() won't lead to the release of writer. Before this CTF patch, !(writer <- trace). Even with event_class leaking, the writer ends up being released. [1] https://github.com/efficios/babeltrace/commit/e6a8e8e4744633807083a077ff9f101eb97d9801 Signed-off-by: Wang Nan Acked-by: Jiri Olsa Cc: Adrian Hunter Cc: Alexei Starovoitov Cc: Brendan Gregg Cc: Cody P Schafer Cc: He Kuang Cc: Jeremie Galarneau Cc: Kirill Smelkov Cc: Li Zefan Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Zefan Li Cc: pi3orama@163.com Link: http://lkml.kernel.org/r/1454680939-24963-6-git-send-email-wangnan0@huawei.com Signed-off-by: Arnaldo Carvalho de Melo

perf data: Add u32_hex data type

2015-12-11T12:12:09+00:00

Add hexadecimal u32 to base data type, which is useful for raw output because raw data is u32 aligned. Signed-off-by: Wang Nan Cc: David S. Miller Cc: Alexei Starovoitov Cc: Brendan Gregg Cc: David S. Miller Cc: Jiri Olsa Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Zefan Li Cc: pi3orama@163.com Link: http://lkml.kernel.org/r/1449541544-67621-12-git-send-email-wangnan0@huawei.com Signed-off-by: Arnaldo Carvalho de Melo

perf data: Fix signedness of value

2015-04-29T13:37:49+00:00

When converting int values, perf first extractes it to a ulonglong, then feeds it to babeltrace as a signed value. For negative 32 bit values (for example, return values of failed syscalls), the extracted data should be something like 0xfffffffe (-2). It becomes a large int64 value. Babeltrace denies to insert it with bt_ctf_field_signed_integer_set_value() because it is larger than 0x7fffffff, the largest positive value a 32 bit int can be. This patch introduces adjust_signedness(), which fills high bits of ulonglong with 1 if the value is negative. Signed-off-by: Wang Nan Cc: David Ahern Cc: Frederic Weisbecker Cc: He Kuang Cc: Jeremie Galarneau Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Sebastian Andrzej Siewior Cc: Tom Zanussi Link: http://lkml.kernel.org/r/1429372220-6406-8-git-send-email-jolsa@kernel.org Signed-off-by: Jiri Olsa [ s/signess/signedness/g ] Signed-off-by: Arnaldo Carvalho de Melo

perf data: Fix duplicate field names and avoid reserved keywords

2015-04-29T13:37:48+00:00

Some parameters of syscall tracepoints named as 'nr', 'event', etc. When dealing with them, perf convert to ctf meets some problem: 1. If a parameter with name 'nr', it will duplicate syscall's common field 'nr'. One such syscall is io_submit(). 2. If a parameter with name 'event', it is denied to be inserted because 'event' is a CTF spec keyword[1]. One such syscall is epoll_ctl. This patch appends '_dupl_X' suffix to avoid problem 1, prepend a '_' prefix to avoid problem 2. [1] http://diamon.org/docs/ctf/v1.8.2/#specC.1.2 Signed-off-by: Wang Nan Cc: David Ahern Cc: Frederic Weisbecker Cc: He Kuang Cc: Jeremie Galarneau Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Sebastian Andrzej Siewior Cc: Tom Zanussi Link: http://lkml.kernel.org/r/1429372220-6406-7-git-send-email-jolsa@kernel.org [ changed to use format_file::alias ] Signed-off-by: Jiri Olsa Signed-off-by: Arnaldo Carvalho de Melo

perf data: Add support for setting ordered_events queue size

2015-04-29T13:37:48+00:00

Adding support to limit the size of ordered_events queue, so we could control allocation size of perf data files without proper finished round events. Reviewed-by: David Ahern Signed-off-by: Jiri Olsa Acked-by: Namhyung Kim Cc: David Ahern Cc: Frederic Weisbecker Cc: He Kuang Cc: Jeremie Galarneau Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Sebastian Andrzej Siewior Cc: Tom Zanussi Cc: Wang Nan Link: http://lkml.kernel.org/r/1429372220-6406-5-git-send-email-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo