All of lore.kernel.org
 help / color / mirror / Atom feed
From: Yoshihiro YUNOMAE <yoshihiro.yunomae.ez@hitachi.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com>,
	Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>,
	linux-kernel@vger.kernel.org, yrl.pp-manager.tt@hitachi.com
Subject: Re: Re: [PATCH V2 1/3] trace-cmd: Define general functions for outputting/inputting saved_cmdlines
Date: Tue, 23 Apr 2013 10:27:59 +0900	[thread overview]
Message-ID: <5175E39F.4020706@hitachi.com> (raw)
In-Reply-To: <1366661205.9609.133.camel@gandalf.local.home>

Hi Steven,

(2013/04/23 5:06), Steven Rostedt wrote:
> On Mon, 2013-04-22 at 18:43 +0900, Yoshihiro YUNOMAE wrote:
>> Currently, trace-cmd outputs data of saved_cmdlines to a trace.dat file
>> in create_file_fd() and inputs the data from the file in tracecmd_init_data().
>> On the other hand, trace-cmd will also output and input data of trace_clock in
>> those functions in the patch "trace-cmd: Add recording to trace_clock" and
>> "Add support for extracting trace_clock in report".
>>
>> The source code of the output/input of saved_cmdlines data can be reused when
>> extract trace_clock, so we define general functions for outputting/inputting a
>> file on debugfs.
>>
>> Signed-off-by: Yoshihiro YUNOMAE <yoshihiro.yunomae.ez@hitachi.com>
>> ---
>>   trace-input.c  |   46 +++++++++++++++++++++++++++++-------------
>>   trace-output.c |   62 ++++++++++++++++++++++++++++++++------------------------
>>   2 files changed, 67 insertions(+), 41 deletions(-)
>>
>> diff --git a/trace-input.c b/trace-input.c
>> index 56a8e8d..232015a 100644
>> --- a/trace-input.c
>> +++ b/trace-input.c
>> @@ -1870,6 +1870,37 @@ static int read_cpu_data(struct tracecmd_input *handle)
>>
>>   }
>>
>> +static int read_data_and_size(struct tracecmd_input *handle,
>> +				     char **data, unsigned long long *size)
>> +{
>> +	*size = read8(handle);
>> +	if (*size < 0)
>> +		return -1;
>> +	*data = malloc(*size + 1);
>> +	if (!*data)
>> +		return -1;
>> +	if (do_read_check(handle, *data, *size)) {
>> +		free(*data);
>> +		return -1;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static int read_and_parse_cmdlines(struct tracecmd_input *handle,
>> +							struct pevent *pevent)
>> +{
>> +	unsigned long long size;
>> +	char *cmdlines;
>> +
>> +	if (read_data_and_size(handle, &cmdlines, &size) < 0)
>> +		return -1;
>> +	cmdlines[size] = 0;
>> +	parse_cmdlines(pevent, cmdlines, size);
>> +	free(cmdlines);
>> +	return 0;
>> +}
>> +
>>   /**
>>    * tracecmd_init_data - prepare reading the data from trace.dat
>>    * @handle: input handle for the trace.dat file
>> @@ -1880,23 +1911,10 @@ static int read_cpu_data(struct tracecmd_input *handle)
>>   int tracecmd_init_data(struct tracecmd_input *handle)
>>   {
>>   	struct pevent *pevent = handle->pevent;
>> -	unsigned long long size;
>> -	char *cmdlines;
>>   	int ret;
>>
>> -	size = read8(handle);
>> -	if (size < 0)
>> -		return -1;
>> -	cmdlines = malloc(size + 1);
>> -	if (!cmdlines)
>> +	if (read_and_parse_cmdlines(handle, pevent) < 0)
>>   		return -1;
>> -	if (do_read_check(handle, cmdlines, size)) {
>> -		free(cmdlines);
>> -		return -1;
>> -	}
>> -	cmdlines[size] = 0;
>> -	parse_cmdlines(pevent, cmdlines, size);
>> -	free(cmdlines);
>>
>>   	handle->cpus = read4(handle);
>>   	if (handle->cpus < 0)
>> diff --git a/trace-output.c b/trace-output.c
>> index 460b773..8697976 100644
>> --- a/trace-output.c
>> +++ b/trace-output.c
>> @@ -687,6 +687,39 @@ static int read_ftrace_printk(struct tracecmd_output *handle)
>>   	return -1;
>>   }
>>
>> +static int save_tracing_file_data(struct tracecmd_output *handle,
>> +						const char *filename)
>> +{
>> +	unsigned long long endian8;
>> +	char *file = NULL;
>> +	struct stat st;
>> +	off64_t check_size;
>> +	off64_t size;
>> +	int ret;
>> +
>> +	file = get_tracing_file(handle, filename);
>> +	ret = stat(file, &st);
>> +	if (ret >= 0) {
>> +		size = get_size(file);
>> +		endian8 = convert_endian_8(handle, size);
>> +		if (do_write_check(handle, &endian8, 8))
>> +			return -1;
>
> Can't do a return here. The "get_tracing_file()" does an allocation that
> needs to be freed with "put_tracing_file()" below. You still need the
> goto out_free, or something.

Oh, I see.
I'll send a revised patch set soon.

Thanks,
Yoshihiro YUNOMAE

> You can initialize ret = -1; and then below have:
>
>> +		check_size = copy_file(handle, file);
>> +		if (size != check_size) {
>> +			errno = EINVAL;
>> +			warning("error in size of file '%s'", file);
>> +			return -1;
>> +		}
>> +	} else {
>> +		size = 0;
>> +		endian8 = convert_endian_8(handle, size);
>> +		if (do_write_check(handle, &endian8, 8))
>> +			return -1;
>> +	}
>
> 	ret = 0;
> out_free:
>
> -- Steve
>
>> +	put_tracing_file(file);
>> +	return 0;
>> +}
>> +
>>   static struct tracecmd_output *
>>   create_file_fd(int fd, struct tracecmd_input *ihandle,
>>   	       const char *tracing_dir,

-- 
Yoshihiro YUNOMAE
Software Platform Research Dept. Linux Technology Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: yoshihiro.yunomae.ez@hitachi.com



  reply	other threads:[~2013-04-23  1:28 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-04-22  9:43 [PATCH V2 0/3] trace-cmd: Support a raw format for outputting timestamp Yoshihiro YUNOMAE
2013-04-22  9:43 ` [PATCH V2 1/3] trace-cmd: Define general functions for outputting/inputting saved_cmdlines Yoshihiro YUNOMAE
2013-04-22 20:06   ` Steven Rostedt
2013-04-23  1:27     ` Yoshihiro YUNOMAE [this message]
2013-04-22  9:43 ` [PATCH V2 2/3] trace-cmd: Add recording to trace_clock Yoshihiro YUNOMAE
2013-04-22  9:43 ` [PATCH V2 3/3] trace-cmd: Add support for extracting trace_clock in report Yoshihiro YUNOMAE

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=5175E39F.4020706@hitachi.com \
    --to=yoshihiro.yunomae.ez@hitachi.com \
    --cc=hidehiro.kawai.ez@hitachi.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=masami.hiramatsu.pt@hitachi.com \
    --cc=rostedt@goodmis.org \
    --cc=yrl.pp-manager.tt@hitachi.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.