[Pgpool-hackers] 3.0 Segfault when processing front-end response CLOSE

[Toshihiro Kitagawa]

Hi,

> Thanks for the report and fix!
> Kitagawa is in charge of this part... Kitagawa, what do you think?

This patch is right. Thanks.
I just commited the patch that includes a fix of similar issue that was reported in pgpool-general-jp.

--
Toshihiro Kitagawa
SRA OSS, Inc. Japan

> --
> Tatsuo Ishii
> SRA OSS, Inc. Japan
> English: http://www.sraoss.co.jp/index_en.php
> Japanese: http://www.sraoss.co.jp
> 
> > I recently upgraded to 3.0 and ran into the following issue...
> > 
> > Program terminated with signal 11, Segmentation fault.
> > #0  Close (frontend=<value optimized out>, backend=<value optimized out>, len=5, contents=0xebdbe0 "PC_4") at pool_proto_modules.c:1017
> > 
> > 
> > It looks like pstmt never gets set on "*contents == 'P'" calls; so attempts to "query_context = pstmt->qctxt" fail.  Adding the following patch helped me get past the issue.  
> > 
> > # diff -c pool_proto_modules.c pool_proto_modules.c.patch
> > *** pool_proto_modules.c        2010-10-08 19:25:02.000000000 +0000
> > --- pool_proto_modules.c.patch  2010-10-08 19:23:31.000000000 +0000
> > ***************
> > *** 1007,1012 ****
> > --- 1007,1013 ----
> >                 }
> > 
> >                 session_context->pending_portal = portal;
> > +                 pstmt = portal->pstmt;
> >         }
> >         else
> >         {
> > 
> > 
> > 
> > Dustin Land
> > id software
> > _______________________________________________
> > Pgpool-hackers mailing list
> > Pgpool-hackers at pgfoundry.org
> > http://pgfoundry.org/mailman/listinfo/pgpool-hackers
> _______________________________________________
> Pgpool-hackers mailing list
> Pgpool-hackers at pgfoundry.org
> http://pgfoundry.org/mailman/listinfo/pgpool-hackers
>