Horton

Re: [CLOBBER_CACHE]Server crashed with segfault 11 while executing clusterdb

Kyotaro Horiguchi <horikyota.ntt@gmail.com>

View thread View on postgr.es
From: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
To: tgl@sss.pgh.pa.us
Cc: sulamul@gmail.com, michael@paquier.xyz, noah@leadboat.com, neha.sharma@enterprisedb.com, amitlangote09@gmail.com, pgsql-hackers@postgresql.org
Date: 2021-03-25T06:40:34Z
Lists: pgsql-hackers 
Sorry for the bug.

At Thu, 25 Mar 2021 01:50:29 -0400, Tom Lane <tgl@sss.pgh.pa.us> wrote in 
> Amul Sul <sulamul@gmail.com> writes:
> > On Wed, Mar 24, 2021 at 8:09 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> >> static inline struct SMgrRelationData *
> >> RelationGetSmgr(Relation rel)
> >> {
> >>     if (unlikely(rel->rd_smgr == NULL))
> >>         RelationOpenSmgr(rel);
> >>     return rel->rd_smgr;
> >> }
> 
> > A quick question: Can't it be a macro instead of an inline function
> > like other macros we have in rel.h?
> 
> The multiple-evaluation hazard seems like an issue.  We've tolerated
> such hazards in the past, but mostly just because we weren't relying
> on static inlines being available, so there wasn't a good way around
> it.
> 
> Also, the conditional evaluation here would look rather ugly
> in a macro, I think, if indeed you could do it at all without
> provoking compiler warnings.

FWIW, +1 for the function as is.

regards.

-- 
Kyotaro Horiguchi
NTT Open Source Software Center

Commits

  1. Replace RelationOpenSmgr() with RelationGetSmgr().

  2. Avoid possible crash while finishing up a heap rewrite.

  3. Redefine pg_class.reltuples to be -1 before the first VACUUM or ANALYZE.