On Wed, Nov 17, 2010 at 02:58:17PM -0500, Stephen Gallagher wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 11/17/2010 02:38 PM, Sumit Bose wrote:
> On Wed, Nov 17, 2010 at 02:32:39PM -0500, Stephen Gallagher wrote:
> On 11/17/2010 04:22 AM, Sumit Bose wrote:
>>>>> Can you call 'client_ctx->auth_ctx->running--;'
directly after
>>>>> 'proxy_child_recv()' ?
>>>>
>
> Sure, I just moved the decrement and the creation of the immediate event
> to before the return value check, so now it will happen regardless of
> the result code. This also eliminates the code duplication in my
> previous patch for creating the immediate event.
>
>> if the proxy_child fails and tevent_create_immediate(), too,
>> proxy_reply() is not called.
>
Good catch. However, if we have ENOMEM here (the only way
tevent_create_immediate() can fail) and there are events on the queue,
then one of them may not fire. However, this is a hopefully impossible
situation to get into, since we're freeing req just above it. I've added
comments to that effect to the patch.
ACK
bye,
Sumit
- --
Stephen Gallagher
RHCE 804006346421761
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Fedora -
http://enigmail.mozdev.org/
iEYEARECAAYFAkzkM9kACgkQeiVVYja6o6Nu+wCfbHWGSi4s3Tl+b18LWQo5RrhH
2+8An2UY0I6aPWHr8Jgit39YCzL2a3K0
=lWnJ
-----END PGP SIGNATURE-----