When 'perf record' write headers, it calls write_xxx in
tools/perf/util/header.c, and check return value. It rolls back all
working only when return value is negative.
This patch ensures write_cpudesc() and write_total_mem() return negative number
when error. Without this patch, headers reported by 'perf report' header is
error in some platform. Following output is caputured on ARM, which doesn't
contain "Processor" field in /proc/cpuinfo. See "cpudesc", "total memory" and
"cmdline" field.
bash-4.2# perf record ls
...
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.001 MB perf.data (~36 samples) ]
bash-4.2# perf report --stdio --header
Error:
The perf.data file has no samples!
# ========
# captured on: Fri Sep 12 10:09:10 2014
# hostname : arma15el
# os release : 3.17.0+
# perf version : 3.10.53
# arch : armv7l
# nrcpus online : 4
# nrcpus avail : 1
# cpudesc : (null)
# total memory : 0 kB
# cmdline :
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, excl_host = 0, excl_guest = 1, precise_ip = 0
# pmu mappings: not available
# ========
#
Signed-off-by: Wang Nan <[email protected]>
---
tools/perf/util/header.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
index ce0de00..39b80ac 100644
--- a/tools/perf/util/header.c
+++ b/tools/perf/util/header.c
@@ -605,8 +605,10 @@ static int write_cpudesc(int fd, struct perf_header *h __maybe_unused,
break;
}
- if (ret)
+ if (ret) {
+ ret = -1;
goto done;
+ }
s = buf;
@@ -950,7 +952,8 @@ static int write_total_mem(int fd, struct perf_header *h __maybe_unused,
n = sscanf(buf, "%*s %"PRIu64, &mem);
if (n == 1)
ret = do_write(fd, &mem, sizeof(mem));
- }
+ } else
+ ret = -1;
free(buf);
fclose(fp);
return ret;
--
1.8.4
On Thu, 16 Oct 2014 11:08:29 +0800, Wang Nan wrote:
> When 'perf record' write headers, it calls write_xxx in
> tools/perf/util/header.c, and check return value. It rolls back all
> working only when return value is negative.
>
> This patch ensures write_cpudesc() and write_total_mem() return negative number
> when error. Without this patch, headers reported by 'perf report' header is
> error in some platform. Following output is caputured on ARM, which doesn't
> contain "Processor" field in /proc/cpuinfo. See "cpudesc", "total memory" and
> "cmdline" field.
>
> bash-4.2# perf record ls
> ...
> [ perf record: Woken up 1 times to write data ]
> [ perf record: Captured and wrote 0.001 MB perf.data (~36 samples) ]
> bash-4.2# perf report --stdio --header
> Error:
> The perf.data file has no samples!
> # ========
> # captured on: Fri Sep 12 10:09:10 2014
> # hostname : arma15el
> # os release : 3.17.0+
> # perf version : 3.10.53
> # arch : armv7l
> # nrcpus online : 4
> # nrcpus avail : 1
> # cpudesc : (null)
> # total memory : 0 kB
> # cmdline :
> # event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, excl_host = 0, excl_guest = 1, precise_ip = 0
> # pmu mappings: not available
> # ========
> #
>
> Signed-off-by: Wang Nan <[email protected]>
I guess the total_memory, cmdline and pmu mappings are somehow affected
by the broken cpudesc. Do they have their own problem on ARM?
Anyway I think it's good to check the result properly, so
Acked-by: Namhyung Kim <[email protected]>
Thanks,
Namhyung
> ---
> tools/perf/util/header.c | 7 +++++--
> 1 file changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
> index ce0de00..39b80ac 100644
> --- a/tools/perf/util/header.c
> +++ b/tools/perf/util/header.c
> @@ -605,8 +605,10 @@ static int write_cpudesc(int fd, struct perf_header *h __maybe_unused,
> break;
> }
>
> - if (ret)
> + if (ret) {
> + ret = -1;
> goto done;
> + }
>
> s = buf;
>
> @@ -950,7 +952,8 @@ static int write_total_mem(int fd, struct perf_header *h __maybe_unused,
> n = sscanf(buf, "%*s %"PRIu64, &mem);
> if (n == 1)
> ret = do_write(fd, &mem, sizeof(mem));
> - }
> + } else
> + ret = -1;
> free(buf);
> fclose(fp);
> return ret;
On 2014/10/22 15:00, Namhyung Kim wrote:
> On Thu, 16 Oct 2014 11:08:29 +0800, Wang Nan wrote:
>> When 'perf record' write headers, it calls write_xxx in
>> tools/perf/util/header.c, and check return value. It rolls back all
>> working only when return value is negative.
>>
>> This patch ensures write_cpudesc() and write_total_mem() return negative number
>> when error. Without this patch, headers reported by 'perf report' header is
>> error in some platform. Following output is caputured on ARM, which doesn't
>> contain "Processor" field in /proc/cpuinfo. See "cpudesc", "total memory" and
>> "cmdline" field.
>>
>> bash-4.2# perf record ls
>> ...
>> [ perf record: Woken up 1 times to write data ]
>> [ perf record: Captured and wrote 0.001 MB perf.data (~36 samples) ]
>> bash-4.2# perf report --stdio --header
>> Error:
>> The perf.data file has no samples!
>> # ========
>> # captured on: Fri Sep 12 10:09:10 2014
>> # hostname : arma15el
>> # os release : 3.17.0+
>> # perf version : 3.10.53
>> # arch : armv7l
>> # nrcpus online : 4
>> # nrcpus avail : 1
>> # cpudesc : (null)
>> # total memory : 0 kB
>> # cmdline :
>> # event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, excl_host = 0, excl_guest = 1, precise_ip = 0
>> # pmu mappings: not available
>> # ========
>> #
>>
>> Signed-off-by: Wang Nan <[email protected]>
>
> I guess the total_memory, cmdline and pmu mappings are somehow affected
> by the broken cpudesc. Do they have their own problem on ARM?
>
The buggy output is caused by cpudesc's problem. I didn't trigger their own problem on ARM because
'total memory' and cmdline are always success. I find the error in write_total_mem() by checking
code by hand.
Thanks.
> Anyway I think it's good to check the result properly, so
>
> Acked-by: Namhyung Kim <[email protected]>
>
> Thanks,
> Namhyung
>
>
>> ---
>> tools/perf/util/header.c | 7 +++++--
>> 1 file changed, 5 insertions(+), 2 deletions(-)
>>
>> diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
>> index ce0de00..39b80ac 100644
>> --- a/tools/perf/util/header.c
>> +++ b/tools/perf/util/header.c
>> @@ -605,8 +605,10 @@ static int write_cpudesc(int fd, struct perf_header *h __maybe_unused,
>> break;
>> }
>>
>> - if (ret)
>> + if (ret) {
>> + ret = -1;
>> goto done;
>> + }
>>
>> s = buf;
>>
>> @@ -950,7 +952,8 @@ static int write_total_mem(int fd, struct perf_header *h __maybe_unused,
>> n = sscanf(buf, "%*s %"PRIu64, &mem);
>> if (n == 1)
>> ret = do_write(fd, &mem, sizeof(mem));
>> - }
>> + } else
>> + ret = -1;
>> free(buf);
>> fclose(fp);
>> return ret;
Hi Ingo,
Could you please collect this patch which fixes a perf problem?
Thanks.
On 2014/10/22 15:00, Namhyung Kim wrote:
> On Thu, 16 Oct 2014 11:08:29 +0800, Wang Nan wrote:
>> When 'perf record' write headers, it calls write_xxx in
>> tools/perf/util/header.c, and check return value. It rolls back all
>> working only when return value is negative.
>>
>> This patch ensures write_cpudesc() and write_total_mem() return negative number
>> when error. Without this patch, headers reported by 'perf report' header is
>> error in some platform. Following output is caputured on ARM, which doesn't
>> contain "Processor" field in /proc/cpuinfo. See "cpudesc", "total memory" and
>> "cmdline" field.
>>
>> bash-4.2# perf record ls
>> ...
>> [ perf record: Woken up 1 times to write data ]
>> [ perf record: Captured and wrote 0.001 MB perf.data (~36 samples) ]
>> bash-4.2# perf report --stdio --header
>> Error:
>> The perf.data file has no samples!
>> # ========
>> # captured on: Fri Sep 12 10:09:10 2014
>> # hostname : arma15el
>> # os release : 3.17.0+
>> # perf version : 3.10.53
>> # arch : armv7l
>> # nrcpus online : 4
>> # nrcpus avail : 1
>> # cpudesc : (null)
>> # total memory : 0 kB
>> # cmdline :
>> # event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, excl_host = 0, excl_guest = 1, precise_ip = 0
>> # pmu mappings: not available
>> # ========
>> #
>>
>> Signed-off-by: Wang Nan <[email protected]>
>
> I guess the total_memory, cmdline and pmu mappings are somehow affected
> by the broken cpudesc. Do they have their own problem on ARM?
>
> Anyway I think it's good to check the result properly, so
>
> Acked-by: Namhyung Kim <[email protected]>
>
> Thanks,
> Namhyung
>
>
>> ---
>> tools/perf/util/header.c | 7 +++++--
>> 1 file changed, 5 insertions(+), 2 deletions(-)
>>
>> diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
>> index ce0de00..39b80ac 100644
>> --- a/tools/perf/util/header.c
>> +++ b/tools/perf/util/header.c
>> @@ -605,8 +605,10 @@ static int write_cpudesc(int fd, struct perf_header *h __maybe_unused,
>> break;
>> }
>>
>> - if (ret)
>> + if (ret) {
>> + ret = -1;
>> goto done;
>> + }
>>
>> s = buf;
>>
>> @@ -950,7 +952,8 @@ static int write_total_mem(int fd, struct perf_header *h __maybe_unused,
>> n = sscanf(buf, "%*s %"PRIu64, &mem);
>> if (n == 1)
>> ret = do_write(fd, &mem, sizeof(mem));
>> - }
>> + } else
>> + ret = -1;
>> free(buf);
>> fclose(fp);
>> return ret;
Em Wed, Oct 22, 2014 at 04:00:12PM +0900, Namhyung Kim escreveu:
> On Thu, 16 Oct 2014 11:08:29 +0800, Wang Nan wrote:
> > When 'perf record' write headers, it calls write_xxx in
> > tools/perf/util/header.c, and check return value. It rolls back all
> > working only when return value is negative.
> >
> > This patch ensures write_cpudesc() and write_total_mem() return negative number
> > when error. Without this patch, headers reported by 'perf report' header is
> > error in some platform. Following output is caputured on ARM, which doesn't
> > contain "Processor" field in /proc/cpuinfo. See "cpudesc", "total memory" and
> > "cmdline" field.
> >
> > bash-4.2# perf record ls
> > ...
> > [ perf record: Woken up 1 times to write data ]
> > [ perf record: Captured and wrote 0.001 MB perf.data (~36 samples) ]
> > bash-4.2# perf report --stdio --header
> > Error:
> > The perf.data file has no samples!
> > # ========
> > # captured on: Fri Sep 12 10:09:10 2014
> > # hostname : arma15el
> > # os release : 3.17.0+
> > # perf version : 3.10.53
> > # arch : armv7l
> > # nrcpus online : 4
> > # nrcpus avail : 1
> > # cpudesc : (null)
> > # total memory : 0 kB
> > # cmdline :
> > # event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, excl_host = 0, excl_guest = 1, precise_ip = 0
> > # pmu mappings: not available
> > # ========
> > #
> >
> > Signed-off-by: Wang Nan <[email protected]>
>
> I guess the total_memory, cmdline and pmu mappings are somehow affected
> by the broken cpudesc. Do they have their own problem on ARM?
>
> Anyway I think it's good to check the result properly, so
>
> Acked-by: Namhyung Kim <[email protected]>
Thanks, applied
- Arnaldo
Commit-ID: ed3077585f2f041e0db0fc41060b69673e98963b
Gitweb: http://git.kernel.org/tip/ed3077585f2f041e0db0fc41060b69673e98963b
Author: Wang Nan <[email protected]>
AuthorDate: Thu, 16 Oct 2014 11:08:29 +0800
Committer: Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Wed, 29 Oct 2014 10:32:48 -0200
perf tools: Ensure return negative value when write header error
When 'perf record' write headers, it calls write_xxx in
tools/perf/util/header.c, and check return value. It rolls back all
working only when return value is negative.
This patch ensures write_cpudesc() and write_total_mem() return negative number
when error. Without this patch, headers reported by 'perf report' header is
error in some platform. Following output is caputured on ARM, which doesn't
contain "Processor" field in /proc/cpuinfo. See "cpudesc", "total memory" and
"cmdline" field.
bash-4.2# perf record ls
...
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.001 MB perf.data (~36 samples) ]
bash-4.2# perf report --stdio --header
Error:
The perf.data file has no samples!
# ========
# captured on: Fri Sep 12 10:09:10 2014
# hostname : arma15el
# os release : 3.17.0+
# perf version : 3.10.53
# arch : armv7l
# nrcpus online : 4
# nrcpus avail : 1
# cpudesc : (null)
# total memory : 0 kB
# cmdline :
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, excl_host = 0, excl_guest = 1, precise_ip = 0
# pmu mappings: not available
# ========
#
Cc: Adrian Hunter <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Li Zefan <[email protected]>
Cc: Namhyung Kim <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Waiman Long <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Wang Nan <[email protected]>
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
---
tools/perf/util/header.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
index 26f5b2f..0ecf4a3 100644
--- a/tools/perf/util/header.c
+++ b/tools/perf/util/header.c
@@ -601,8 +601,10 @@ static int __write_cpudesc(int fd, const char *cpuinfo_proc)
break;
}
- if (ret)
+ if (ret) {
+ ret = -1;
goto done;
+ }
s = buf;
@@ -965,7 +967,8 @@ static int write_total_mem(int fd, struct perf_header *h __maybe_unused,
n = sscanf(buf, "%*s %"PRIu64, &mem);
if (n == 1)
ret = do_write(fd, &mem, sizeof(mem));
- }
+ } else
+ ret = -1;
free(buf);
fclose(fp);
return ret;