Re: [PATCH] perf record: mmap output file - v2

[Namhyung Kim]

Hi Ingo,

On Tue, 15 Oct 2013 08:02:00 +0200, Ingo Molnar wrote:
> * David Ahern <dsahern@xxxxxxxxx> wrote:
>
>> +	/* for MMAP based file writes */
>> +	void			*mmap_addr;
>> +	u64			bytes_at_mmap_start; /* bytes in file when mmap use starts */
>> +	u64			mmap_offset;    /* current location within mmap */
>> +	size_t			mmap_size;      /* size of mmap segments */
>> +	bool			use_mmap;
>
>> +	if (!rec->opts.pipe_output && stat(output_name, &st) == 0) {
>> +		rec->use_mmap = true;
>> +		rec->bytes_at_mmap_start = st.st_size - rec->bytes_written;
>> +	}
>
> 1)
>
> I think __cmd_record() has become way too large, nearly 300 lines of code. 
> It would be nice to split it into 2-3 helpers that operate on 'struct 
> perf_record' or so.

Agreed.

>
> 2)
>
> The stat() seems superfluous, here in __cmd_record() we've just checked 
> the output_name and made sure it exists. Can that stat() call ever fail?

AFAICS it's needed to check current file size.  But I think it's better
to use fstat().

>
> 3)
>
> The rec->bytes_at_mmap_start field feels a bit weird. If I read the code 
> correctly, in every 'perf record' invocation, rec->bytes_written starts at 
> 0 - i.e. we don't have repeat invocations of cmd_record().

rec->bytes_written is updated when it writes to the output file for
synthesizing COMM/MMAP events (this mmap output is not used at that time).

>
> That means that this:
>
> 		rec->bytes_at_mmap_start = st.st_size - rec->bytes_written;
>
> is really:
>
> 		rec->bytes_at_mmap_start = st.st_size;
>
> furthermore, since we don't allow appends anymore, st.st_size ought to be 
> zero as well.
>
> Which means that ->bytes_at_mmap_start is always zero - and could be 
> eliminated altogether.

No, st_size is bigger than rec->bytes_written due to the
perf_file_header which is written without updating rec->bytes_written.

Actually I worried about the mmap offset not being aligned to page
size.  But it seems that's not a problem.

Thanks,
Namhyung
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/