[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <f3a36221-eb05-fe42-0208-d0268183c27e@gmail.com>
Date: Wed, 26 Apr 2017 23:56:02 +0900
From: Taeung Song <treeze.taeung@...il.com>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: linux-kernel@...r.kernel.org,
Federico Vaga <federico.vaga@...a.pv.it>
Subject: Re: [PATCH] trace-cmd: Fix a warning about incompatible pointer type
of load_plugin()
Hi Steven :)
On 04/26/2017 08:11 AM, Steven Rostedt wrote:
> On Mon, 24 Apr 2017 22:43:31 +0900
> Taeung Song <treeze.taeung@...il.com> wrote:
>
>> Currently the return type of load_plugin() in plugin_python.c is void type,
>> but it is different from the argument type of trace_util_load_plugins().
>> So fix it for the below warning.
>>
>> /home/taeung/git/opensource/trace-cmd/plugin_python.c: In function ‘pevent_plugin_loader’:
>> /home/taeung/git/opensource/trace-cmd/plugin_python.c:95:41: warning: passing argument 3 of ‘trace_util_load_plugins’ from incompatible pointer type [-Wincompatible-pointer-types]
>> trace_util_load_plugins(pevent, ".py", load_plugin, globals);
>> ^
>> In file included from /home/taeung/git/opensource/trace-cmd/plugin_python.c:3:0:
>> /home/taeung/git/opensource/trace-cmd/trace-cmd.h:300:5: note: expected ‘int (*)(struct pevent *, const char *, const char *, void *)’ but argument is of type ‘void (*)(struct pevent *, const char *, const char *, void *)’
>> int trace_util_load_plugins(struct pevent *pevent, const char *suffix,
>> ^
>>
>
> Thanks but Federico beat you to it.
>
> http://lkml.kernel.org/r/20170423102258.21609-2-federico.vaga@vaga.pv.it
>
> I'll be updating and pushing out a new trace-cmd this week.
>
>> Signed-off-by: Taeung Song <treeze.taeung@...il.com>
>> ---
>> plugin_python.c | 8 +++++---
>> 1 file changed, 5 insertions(+), 3 deletions(-)
>>
>> diff --git a/plugin_python.c b/plugin_python.c
>> index da07d27..d736f50 100644
>> --- a/plugin_python.c
>> +++ b/plugin_python.c
>> @@ -20,7 +20,7 @@ static const char pyload[] =
>> "finally:\n"
>> " file.close()\n";
>>
>> -static void load_plugin(struct pevent *pevent, const char *path,
>> +static int load_plugin(struct pevent *pevent, const char *path,
>> const char *name, void *data)
>> {
>> PyObject *globals = data;
>> @@ -32,7 +32,7 @@ static void load_plugin(struct pevent *pevent, const char *path,
>> PyObject *res;
>>
>> if (!full || !n)
>> - return;
>> + return -ENOMEM;
>>
>> strcpy(full, path);
>> strcat(full, "/");
>> @@ -43,16 +43,18 @@ static void load_plugin(struct pevent *pevent, const char *path,
>>
>> asprintf(&load, pyload, full, n);
>> if (!load)
>> - return;
>> + return -1;
>>
>> res = PyRun_String(load, Py_file_input, globals, globals);
>> if (!res) {
>> fprintf(stderr, "failed loading %s\n", full);
>> PyErr_Print();
>> + return -1;
>
> Although, Federico's version didn't have this. I may add this under
> your patch. Although, it requires freeing load. The better way would be
> to set a "ret" variable and return that.
>
> Want to send another patch on top of Federico's?
>
Yes! I want to do it.
After you update 'master' branch on your repo
(git.kernel.org/.../trace-cmd.git), I'll send a patch as you said !
And would you mind if I ask you to check another patch I sent ?
"[PATCH] Fix pevent_filter_clear_trivial()"
But I'll renewedly send the patch after I a bit modified it.
I'd appreciate it if you check the patch, too. :)
Thanks,
Taeung
Powered by blists - more mailing lists