This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH v4] Make bindresvport() function to multithread-safe
On Fri, Sep 28, 2012 at 12:34 PM, Roland McGrath <roland@hack.frob.com> wrote:
>> The fact that Peng is submitting the patch and working with the
>> community is sufficient evidence for me.
>>
>> How much evidence do we need?
>
> People often propose changes that are not actually worthwhile.
> The good intentions of the submitter are no evidence at all of
> the worthiness of the change.
Roland,
I agree with the general statement you're making. I guess
where I have a hard time is deciding what constitutes worthwhile,
and how hard to push for such a rationalization.
Truth be told I got caught up in the interesting technical details
and forgot to push Peng on your original question. Let me do that
now.
Peng,
There are really three things every high quality patch needs:
(a) Rationale for the change.
(b) Performance analysis e.g. How does the patch effect the
performance of the function?
(c) The change itself.
You have worked hard to provide (b) and (c), but have not yet provided (a).
Could you please answer Roland's question here:
http://sourceware.org/ml/libc-alpha/2012-09/msg00465.html
"I don't see the rationale for requiring bindresvport to be thread-safe now."
More specifically can you explain *why* you need bindresvport to be
thread-safe?
Cheers,
Carlos.