Re: 12.1 not useable: clientlib fails after a dozen queries (GSSAPI ?)
Tom Lane <tgl@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Peter <pmc@citylink.dinoex.sub.org>
Cc: pgsql-hackers@lists.postgresql.org, Stephen Frost <sfrost@snowman.net>
Date: 2020-01-11T21:42:31Z
Lists: pgsql-hackers, pgsql-general
Attachments
- 0003-buffer-allocation-1.patch (text/x-diff) patch 0003
I wrote: > (Still doesn't touch the static-buffer issue, though.) And here's a delta patch for that. This fixes an additional portability hazard, which is that there were various places doing stuff like input.length = ntohl(*(uint32 *) PqGSSRecvBuffer); That's a SIGBUS hazard on alignment-picky hardware, because there is no guarantee that a variable that's just declared "char ...[...]" will have any particular alignment. But malloc'ing the space will provide maxaligned storage. My FreeBSD testing has now given me enough confidence in these patches that I'm just going to go ahead and push them. But, if you'd like to do some more testing in advance of 12.2 release, please do. regards, tom lane
Commits
-
Extensive code review for GSSAPI encryption mechanism.
- fde155424f67 12.2 landed
- 2c0cdc818365 13.0 landed