Horton

Re: Why our Valgrind reports suck

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Andres Freund <andres@anarazel.de>
Cc: pgsql-hackers@lists.postgresql.org
Date: 2025-05-11T19:33:30Z
Lists: pgsql-hackers 
I wrote:
> Okay, here is a patch series that updates the
> 0001-Make-memory-contexts-themselves-more-visible-to-valg.patch
> patch you posted in that thread,

I forgot to mention that I did try to implement the "two-level
pool" scheme that the Valgrind documentation talks about, and
could not make it work.  There seem to be undocumented interactions
between the outer and inner chunks, and it's not real clear to me
that there's not outright bugs.  Anyway, AFAICS that scheme
would bring us no immediate advantages anyway, compared to the
flat approach of just adding mempool chunks for the allocators'
management areas.

			regards, tom lane

Commits

  1. Undo thinko in commit e78d1d6d4.

  2. Avoid leakage of zero-length arrays in partition_bounds_copy().

  3. Fix MemoryContextAllocAligned's interaction with Valgrind.

  4. Fix assorted pretty-trivial memory leaks in the backend.

  5. Improve our support for Valgrind's leak tracking.

  6. Reduce leakage during PL/pgSQL function compilation.

  7. Silence Valgrind leakage complaints in more-or-less-hackish ways.

  8. Silence complaints about leaks in PlanCacheComputeResultDesc.

  9. Suppress complaints about leaks in TS dictionary loading.

  10. Suppress complaints about leaks in function cache loading.

  11. Fix per-relation memory leakage in autovacuum.

  12. Fix AlignedAllocRealloc to cope sanely with OOM.