Horton

Re: Condition variable live lock

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

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Thomas Munro <thomas.munro@enterprisedb.com>
Cc: Andres Freund <andres@anarazel.de>, Robert Haas <robertmhaas@gmail.com>, Pg Hackers <pgsql-hackers@postgresql.org>
Date: 2018-01-05T15:28:10Z
Lists: pgsql-hackers 
Thomas Munro <thomas.munro@enterprisedb.com> writes:
> Could we install the sentinel and pop the first entry at the same
> time, so that we're not adding an extra spinlock acquire/release?

Hm, maybe.  Other ideas in that space:

* if queue is empty when we first acquire the spinlock, we don't
have to do anything at all.

* if queue is empty after we pop the first entry, we needn't bother
installing our sentinel, just signal that proc and we're done.

It's a question of how complicated you're willing to make this
logic, and whether you trust that we'll be able to test all the
code paths.

			regards, tom lane

Commits

  1. Add WL_EXIT_ON_PM_DEATH pseudo-event.

  2. While waiting for a condition variable, detect postmaster death.

  3. Fix race condition during replication origin drop.

  4. Allow ConditionVariable[PrepareTo]Sleep to auto-switch between CVs.

  5. Cosmetic improvements in condition_variable.[hc].

  6. Improve error detection capability in proclists.

  7. Remove return values of ConditionVariableSignal/Broadcast.

  8. Reorder steps in ConditionVariablePrepareToSleep for more safety.

  9. Rewrite ConditionVariableBroadcast() to avoid live-lock.

  10. Add parallel-aware hash joins.