guest, defer disabling poll
mode to when the event loop is about to be blocked.
With this patch applied, FIO seq-read performance (bs=4k, iodepth=64,
cache=writethrough) in VMs increases from 330K/s to 413K/s IOPS.
Suggested-by: Stefan Hajnoczi
Signed-off-by: Chao Gao
---
util/aio-posix.c | 21
mon_ops->need_wait(ctx)) {
>+if (poll_set_started(ctx, &ready_list, false)) {
>+timeout = 0;
>+progress = true;
In this case, is it ok to skip the call of ->wait() below? If yes, maybe put
the call in the "else" path.
>+}
>+
>
On Wed, Jul 06, 2022 at 12:59:29PM +0100, Stefan Hajnoczi wrote:
>On Fri, Jul 01, 2022 at 05:13:48PM +0800, Chao Gao wrote:
>> When measuring FIO read performance (cache=writethrough, bs=4k, iodepth=64)
>> in
>> VMs, we observe ~80K/s notifications (e.g., EPT_MISCONFIG) f
151 3.60% 2.79% 0.40us 52.07us
1.30us ( +- 31.44% )
Signed-off-by: Chao Gao
---
util/aio-posix.c | 16
1 file changed, 12 insertions(+), 4 deletions(-)
diff --git a/util/aio-posix.c b/util/aio-posix.c
index 731f3826c0..bd2076145b 100644
--- a/util/aio-posix
