Re: [tip:perf/urgent] perf_event: Eliminate raw->size

[OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>]

Xiao Guangrong <xiaoguangrong@cn.fujitsu.com> writes:

> Peter Zijlstra wrote:
>> On Mon, 2009-12-07 at 05:31 +0000, tip-bot for Xiao Guangrong wrote:
>>> Commit-ID:  f48f669d42e133db839af16656fd720107ef6742
>>> Gitweb:     http://git.kernel.org/tip/f48f669d42e133db839af16656fd720107ef6742
>>> Author:     Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
>>> AuthorDate: Mon, 7 Dec 2009 13:04:04 +0800
>>> Committer:  Ingo Molnar <mingo@elte.hu>
>>> CommitDate: Mon, 7 Dec 2009 06:26:25 +0100
>>>
>>> perf_event: Eliminate raw->size
>>>
>>> raw->size is not used, this patch just cleans it up.
>> 
>> This patch feels wrong.. I would suspect it to be used to validate the
>> data parsing, like do we consume past the end and did we consume
>> everything.
>> 
>
> But as i noticed in current code, it just use raw->data, maybe i miss something?
> what i do is just using 'data' instead of 'raw->data' :-(

Well, it's not user fault at all though. To pass "struct sample_raw_data"
instead of "void *" to raw_field_value/raw_field_ptr/etc., then bounds
check of sample_raw_data would be sane. It is what I intended...

Thanks.
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>