]> granicus.if.org Git - postgresql/commit
Measure WaitLatch's timeout parameter in milliseconds, not microseconds.
authorTom Lane <tgl@sss.pgh.pa.us>
Tue, 9 Aug 2011 22:52:29 +0000 (18:52 -0400)
committerTom Lane <tgl@sss.pgh.pa.us>
Tue, 9 Aug 2011 22:52:29 +0000 (18:52 -0400)
commit9f17ffd8660243382809b8023ff3d38fbfac4e8b
treef58c4fa6f2509bb157dcad8ad4223673ed55d383
parent4e15a4db5e65e43271f8d20750d6500ab12632d0
Measure WaitLatch's timeout parameter in milliseconds, not microseconds.

The original definition had the problem that timeouts exceeding about 2100
seconds couldn't be specified on 32-bit machines.  Milliseconds seem like
sufficient resolution, and finer grain than that would be fantasy anyway
on many platforms.

Back-patch to 9.1 so that this aspect of the latch API won't change between
9.1 and later releases.

Peter Geoghegan
src/backend/access/transam/xlog.c
src/backend/port/unix_latch.c
src/backend/port/win32_latch.c
src/backend/postmaster/pgarch.c
src/backend/replication/syncrep.c
src/backend/replication/walsender.c