On Wed, Sep 25, 2024 at 8:30 AM Andres Freund <andres@anarazel.de> wrote:
> Just ran that. There probably is a performance difference, but it's small
> (<0.5%) making it somewhat hard to be certain. It looks like the main reason
> for that is ConditionVariableBroadcast() on the iocv shows up even though
> nobody is waiting.
. o O { Gotta fix that. Memory barriers might be enough to check for
empty wait list?, and even in the slow path, atomic wait lists or
something better than spinlocks... }
> However, our habit of modifying buffers while IO is going on is
> causing issues with filesystem level checksums as well, as evidenced by the
> fact that debug_io_direct = data on btrfs causes filesystem corruption. So I
> tend to think it'd be better to just stop doing that alltogether (we also do
> that for WAL, when writing out a partial page, but a potential fix there would
> be different, I think).
+many. Interesting point re the WAL variant. For the record, here's
some discussion and a repro for that problem, which Andrew currently
works around in a build farm animal with mount options:
https://www.postgresql.org/message-id/CA%2BhUKGKSBaz78Fw3WTF3Q8ArqKCz1GgsTfRFiDPbu-j9OFz-jw%40mail.gmail.com
