This patch series eliminates ctx->notifier EventNotifier activity when
aio_poll() is in polling mode. There is no need to use the EventNotifier since
a polling handler can detect that aio_notify() has been called by monitoring a
field in memory instead.

Optimizing out the EventNotifier calls improves null-co random read 4KB
iodepth=1 IOPS by 18%.

aio_compute_timeout() is now called twice if aio_poll() needs to block, which
means an extra qemu_clock_get_ns() call is made when there is an active timer.
An alternative would be to set timeout = 0 if ctx->notified is true before
blocking, but going around the event loop again could slow things down more.

I have not modified docs/spin/aio_notify*.promela because I'm not familiar with
the SPIN model checker.

Stefan Hajnoczi (3):
  async: rename event_notifier_dummy_cb/poll()
  async: always set ctx->notified in aio_notify()
  aio-posix: keep aio_notify_me disabled during polling

 util/aio-posix.c | 56 ++++++++++++++++++++++++------------------------
 util/async.c     | 22 +++++++++----------
 2 files changed, 38 insertions(+), 40 deletions(-)

-- 
2.26.2

On 04/08/20 07:28, Stefan Hajnoczi wrote:
> This patch series eliminates ctx->notifier EventNotifier activity when
> aio_poll() is in polling mode. There is no need to use the EventNotifier since
> a polling handler can detect that aio_notify() has been called by monitoring a
> field in memory instead.
> 
> Optimizing out the EventNotifier calls improves null-co random read 4KB
> iodepth=1 IOPS by 18%.
> 
> aio_compute_timeout() is now called twice if aio_poll() needs to block, which
> means an extra qemu_clock_get_ns() call is made when there is an active timer.
> An alternative would be to set timeout = 0 if ctx->notified is true before
> blocking, but going around the event loop again could slow things down more.
> 
> I have not modified docs/spin/aio_notify*.promela because I'm not familiar with
> the SPIN model checker.

I'll take a look.  Looks good apart from more comments, on which I
commented.

Paolo