On 2024-08-12 16:20, Stanislav Fomichev wrote:
On 08/12, Joe Damato wrote:
From: Martin Karsten <mkarsten@xxxxxxxxxxxx>
[snip]
@@ -2005,8 +2023,10 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
* trying again in search of more luck.
*/
res = ep_send_events(ep, events, maxevents);
- if (res)
+ if (res) {
+ ep_suspend_napi_irqs(ep);
Aren't we already doing defer in the busy_poll_stop? (or in napi_poll
when it's complete/done). Why do we need another rearming here?
If ep_poll finds data present due to previous softirq activity during a
sleep (or application bootstrap), the timer is armed with the shorter
gro-flush-timeout. This timer rearming with irq-suspend-timeout
kickstarts the suspend mechanism by injecting the larger timer.
Thanks,
Martin
