httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rainer Jung <>
Subject Re: Proxy regressions
Date Thu, 04 Nov 2010 08:12:49 GMT
On 04.11.2010 08:36, Ruediger Pluem wrote:
> On 11/04/2010 01:34 AM, Rainer Jung wrote:
>> On 04.11.2010 00:57, Graham Leggett wrote:
>>> On 03 Nov 2010, at 10:28 PM, Stefan Fritsch wrote:
>>>> Strange, I have these problems only with prefork, not with event.
>>>> But with event I get a segfault in the reslist cleanup code.
>>>> Can somebody cross-check this?
>>> This smelled like a pool lifetime issue, and looking closer it looked
>>> like we were cleaning up a brigade after we had returned the backend to
>>> the pool. The attached patch stops the crash, and it seems to run at a
>>> sensible speed again (I suspect the late cleanup of the brigade was
>>> sending the code into a spin). Can you give it a try?
>  From just looking at the code this seems to be the correct thing to do.

OK, I'll let Graham commit that one.

>> Your patch fixes the crashes for me. All tests pass if the following
>> patch is also applied. Otherwise worker hangs.
>> I don't know, whether that is the right thing to add, especially since
>> "cleared" now has a use very similar to "inreslist". The patch is an
>> addon to r1026665 where the "cleaned" flag was introduced, occassionally
>> set to "1" and tested against "0", but never reset.

> Yeah, this is correct. As I reviewed the patch initially I thought the reset
> would happen through an regular apr_pcalloc  for conn, but this is not the case.
> So the reset is the correct thing to do here.

Committed in r1030850.



View raw message