postgresql/src
Tom Lane 37e76546a2 Avoid improbable PANIC during heap_update.
heap_update needs to clear any existing "all visible" flag on
the old tuple's page (and on the new page too, if different).
Per coding rules, to do this it must acquire pin on the appropriate
visibility-map page while not holding exclusive buffer lock;
which creates a race condition since someone else could set the
flag whenever we're not holding the buffer lock.  The code is
supposed to handle that by re-checking the flag after acquiring
buffer lock and retrying if it became set.  However, one code
path through heap_update itself, as well as one in its subroutine
RelationGetBufferForTuple, failed to do this.  The end result,
in the unlikely event that a concurrent VACUUM did set the flag
while we're transiently not holding lock, is a non-recurring
"PANIC: wrong buffer passed to visibilitymap_clear" failure.

This has been seen a few times in the buildfarm since recent VACUUM
changes that added code paths that could set the all-visible flag
while holding only exclusive buffer lock.  Previously, the flag
was (usually?) set only after doing LockBufferForCleanup, which
would insist on buffer pin count zero, thus preventing the flag
from becoming set partway through heap_update.  However, it's
clear that it's heap_update not VACUUM that's at fault here.

What's less clear is whether there is any hazard from these bugs
in released branches.  heap_update is certainly violating API
expectations, but if there is no code path that can set all-visible
without a cleanup lock then it's only a latent bug.  That's not
100% certain though, besides which we should worry about extensions
or future back-patch fixes that could introduce such code paths.

I chose to back-patch to v12.  Fixing RelationGetBufferForTuple
before that would require also back-patching portions of older
fixes (notably 0d1fe9f74), which is more code churn than seems
prudent to fix a hypothetical issue.

Discussion: https://postgr.es/m/2247102.1618008027@sss.pgh.pa.us
2021-04-13 12:17:24 -04:00
..
backend Avoid improbable PANIC during heap_update. 2021-04-13 12:17:24 -04:00
bin pg_checksums: Fix progress reporting. 2021-04-03 00:07:49 +09:00
common Second thoughts on TOAST decompression. 2020-11-02 11:25:18 -05:00
fe_utils psql: call clearerr() just before printing 2021-03-29 18:34:39 -03:00
include Use correct spelling of statistics kind 2021-03-23 05:00:19 +01:00
interfaces Use macro MONTHS_PER_YEAR instead of '12' in /ecpg/pgtypeslib 2021-04-02 16:42:29 -04:00
makefiles Remove libpq.rc, use win32ver.rc for libpq 2020-01-15 15:06:12 +01:00
pl Port regress-python3-mangle.mk to Solaris "sed". 2021-04-12 19:24:24 -07:00
port pg_attribute_no_sanitize_alignment() macro 2021-02-13 17:49:08 -05:00
template Further tweaking of PG_SYSROOT heuristics for macOS. 2021-01-20 12:07:31 -05:00
test Use "-I." in directories holding Bison parsers, for Oracle compilers. 2021-04-12 19:24:58 -07:00
timezone Update time zone data files to tzdata release 2021a. 2021-01-24 16:29:47 -05:00
tools backpatch "jit: Add support for LLVM 12." 2020-12-07 18:32:32 -08:00
tutorial tutorial: land height is "elevation", not "altitude" 2021-03-10 20:25:18 -05:00
.gitignore
DEVELOPERS
Makefile Fix partial-build problems introduced by having more generated headers. 2018-04-09 16:42:10 -04:00
Makefile.global.in Update Unicode data to Unicode 13.0.0 and CLDR 37 2020-04-24 09:52:59 +02:00
Makefile.shlib Add PostgreSQL home page to --help output 2020-02-28 13:12:21 +01:00
nls-global.mk NLS: Fix backend gettext triggers 2019-09-23 09:04:20 +02:00