Re: [Patch] proc: check error pointer returned by m_start()

[Cong Wang <amwang@redhat.com>]

于 2011年03月28日 11:58, Linus Torvalds 写道:
> On Sun, Mar 27, 2011 at 8:46 PM, Amerigo Wang<amwang@redhat.com>  wrote:
>>
>> This is true, that commit changed the return value of m_start(),
>> which will return an error pointer on failure, but Al forgot
>> to check the error pointer in m_stop() which will be called
>> when m_start() fails. This patches fixes it.
>
> I did this slightly differently, and put the check in m_stop()
> instead, because I felt that matched the logic of m_start, while
> vma_stop() is more of an internal helper thing.
>

Ok, I am fine with this, will send an updated patch.

> I dunno. I don't think it matters. But one thing I reacted to was that
> when I was walking through the logic, I really wanted to say "seq_file
> is wrong to call m_stop if m_start returned an error code". I really
> felt like "hwy, if ->start fails, we damn well shouldn't have called
> ->stop".


This is a good point and makes prefect sense.

>
> But I guess we're stuck with that particular semantic for seq_files by now.
>

Yup, I guess there are some seq_file users still rely on this behavior,
we can fix them all later.

Thanks.