2015-04-02 06:05:17

by Wang Nan

Subject: [PATCH] perf: annotate: make it respect -i option.

There is a bug in perf annotate that it doesn't respect user provided
'-i'/'--input' option:

# perf record ls
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.001 MB perf.data (8 samples) ]
# mv ./perf.data ./perf.data.new
# perf annotate -i ./perf.data.new --stdio
failed to open perf.data: No such file or directory (try 'perf record' first)

This patch fix it by setting file path after option parsing, like
what 'perf report' does.

Signed-off-by: Wang Nan <[email protected]>
---
tools/perf/builtin-annotate.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 71bf745..929f83c 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -283,7 +283,6 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __maybe_unused)
},
};
struct perf_data_file file = {
- .path = input_name,
.mode = PERF_DATA_MODE_READ,
};
const struct option options[] = {
@@ -342,6 +341,7 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __maybe_unused)

setup_browser(true);

+ file.path = input_name,
annotate.session = perf_session__new(&file, false, &annotate.tool);
if (annotate.session == NULL)
return -1;
--
1.8.3.4

2015-04-02 08:19:08

by Namhyung Kim

Subject: Re: [PATCH] perf: annotate: make it respect -i option.

On Thu, Apr 02, 2015 at 06:04:52AM +0000, Wang Nan wrote:
> There is a bug in perf annotate that it doesn't respect user provided
> '-i'/'--input' option:
>
> # perf record ls
> [ perf record: Woken up 1 times to write data ]
> [ perf record: Captured and wrote 0.001 MB perf.data (8 samples) ]
> # mv ./perf.data ./perf.data.new
> # perf annotate -i ./perf.data.new --stdio
> failed to open perf.data: No such file or directory (try 'perf record' first)
>
> This patch fix it by setting file path after option parsing, like
> what 'perf report' does.
>
> Signed-off-by: Wang Nan <[email protected]>

I guess other commands are also suffered from this bug.. anyway,

Acked-by: Namhyung Kim <[email protected]>

Thanks,
Namhyung

2015-04-02 08:24:32

by Jiri Olsa

Subject: [PATCH] perf kmem: Respect -i option

On Thu, Apr 02, 2015 at 06:04:52AM +0000, Wang Nan wrote:
> There is a bug in perf annotate that it doesn't respect user provided
> '-i'/'--input' option:
>
> # perf record ls
> [ perf record: Woken up 1 times to write data ]
> [ perf record: Captured and wrote 0.001 MB perf.data (8 samples) ]
> # mv ./perf.data ./perf.data.new
> # perf annotate -i ./perf.data.new --stdio
> failed to open perf.data: No such file or directory (try 'perf record' first)
>
> This patch fix it by setting file path after option parsing, like
> what 'perf report' does.
>
> Signed-off-by: Wang Nan <[email protected]>

Acked-by: Jiri Olsa <[email protected]>

plus found another in kmem.. ;-)

thanks,
jirka


---
Currently the perf kmem does not respect -i option.
Initializing the file.path properly after options
get parsed.

Signed-off-by: Jiri Olsa <[email protected]>
---
tools/perf/builtin-kmem.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 64d3623d45a0..c81e24870e7b 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -684,7 +684,6 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
};
struct perf_session *session;
struct perf_data_file file = {
- .path = input_name,
.mode = PERF_DATA_MODE_READ,
};
int ret = -1;
@@ -700,6 +699,8 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
return __cmd_record(argc, argv);
}

+ file.path = input_name;
+
session = perf_session__new(&file, false, &perf_kmem);
if (session == NULL)
return -1;
--
1.9.3

2015-04-21 03:11:56

by Wang Nan

Subject: Re: [PATCH] perf: annotate: make it respect -i option.

On 2015/4/2 16:12, Namhyung Kim wrote:
> On Thu, Apr 02, 2015 at 06:04:52AM +0000, Wang Nan wrote:
>> There is a bug in perf annotate that it doesn't respect user provided
>> '-i'/'--input' option:
>>
>> # perf record ls
>> [ perf record: Woken up 1 times to write data ]
>> [ perf record: Captured and wrote 0.001 MB perf.data (8 samples) ]
>> # mv ./perf.data ./perf.data.new
>> # perf annotate -i ./perf.data.new --stdio
>> failed to open perf.data: No such file or directory (try 'perf record' first)
>>
>> This patch fix it by setting file path after option parsing, like
>> what 'perf report' does.
>>
>> Signed-off-by: Wang Nan <[email protected]>
>
> I guess other commands are also suffered from this bug.. anyway,
>
> Acked-by: Namhyung Kim <[email protected]>
>
> Thanks,
> Namhyung
>

Hi,

Looks like the next patch 'perf kmem: Respect -i option' has already been collected
by tip/master, but this patch is lost. Is there any problem?

Thank you!