Hi Andrew,
May I ask you to remove "epoll: ensure ep_poll() doesn't miss wakeup
events" from your -mm queue? Jason lately found out that the patch
does not fully solve the problem and this one patch is a second
attempt to do things correctly in a different way (namely to do
the final check under the lock). Previous changes are not needed.
Thanks.
--
Roman
On 2020-05-05 10:40, Roman Penyaev wrote:
> The original problem was described here:
> https://lkml.org/lkml/2020/4/27/1121
>
> There is a possible race when ep_scan_ready_list() leaves ->rdllist
> and ->obflist empty for a short period of time although some events
> are pending. It is quite likely that ep_events_available() observes
> empty lists and goes to sleep. Since 339ddb53d373 ("fs/epoll: remove
> unnecessary wakeups of nested epoll") we are conservative in wakeups
> (there is only one place for wakeup and this is ep_poll_callback()),
> thus ep_events_available() must always observe correct state of
> two lists. The easiest and correct way is to do the final check
> under the lock. This does not impact the performance, since lock
> is taken anyway for adding a wait entry to the wait queue.
>
> In this patch barrierless __set_current_state() is used. This is
> safe since waitqueue_active() is called under the same lock on wakeup
> side.
>
> Short-circuit for fatal signals (i.e. fatal_signal_pending() check)
> is moved to the line just before actual events harvesting routine.
> This is fully compliant to what is said in the comment of the patch
> where the actual fatal_signal_pending() check was added:
> c257a340ede0 ("fs, epoll: short circuit fetching events if thread
> has been killed").
>
> Signed-off-by: Roman Penyaev <[email protected]>
> Reported-by: Jason Baron <[email protected]>
> Cc: Andrew Morton <[email protected]>
> Cc: Khazhismel Kumykov <[email protected]>
> Cc: Alexander Viro <[email protected]>
> Cc: [email protected]
> Cc: [email protected]
> Cc: [email protected]
> ---
> fs/eventpoll.c | 48 ++++++++++++++++++++++++++++--------------------
> 1 file changed, 28 insertions(+), 20 deletions(-)
>
> diff --git a/fs/eventpoll.c b/fs/eventpoll.c
> index aba03ee749f8..8453e5403283 100644
> --- a/fs/eventpoll.c
> +++ b/fs/eventpoll.c
> @@ -1879,34 +1879,33 @@ static int ep_poll(struct eventpoll *ep,
> struct epoll_event __user *events,
> * event delivery.
> */
> init_wait(&wait);
> - write_lock_irq(&ep->lock);
> - __add_wait_queue_exclusive(&ep->wq, &wait);
> - write_unlock_irq(&ep->lock);
>
> + write_lock_irq(&ep->lock);
> /*
> - * We don't want to sleep if the ep_poll_callback() sends us
> - * a wakeup in between. That's why we set the task state
> - * to TASK_INTERRUPTIBLE before doing the checks.
> + * Barrierless variant, waitqueue_active() is called under
> + * the same lock on wakeup ep_poll_callback() side, so it
> + * is safe to avoid an explicit barrier.
> */
> - set_current_state(TASK_INTERRUPTIBLE);
> + __set_current_state(TASK_INTERRUPTIBLE);
> +
> /*
> - * Always short-circuit for fatal signals to allow
> - * threads to make a timely exit without the chance of
> - * finding more events available and fetching
> - * repeatedly.
> + * Do the final check under the lock. ep_scan_ready_list()
> + * plays with two lists (->rdllist and ->ovflist) and there
> + * is always a race when both lists are empty for short
> + * period of time although events are pending, so lock is
> + * important.
> */
> - if (fatal_signal_pending(current)) {
> - res = -EINTR;
> - break;
> + eavail = ep_events_available(ep);
> + if (!eavail) {
> + if (signal_pending(current))
> + res = -EINTR;
> + else
> + __add_wait_queue_exclusive(&ep->wq, &wait);
> }
> + write_unlock_irq(&ep->lock);
>
> - eavail = ep_events_available(ep);
> - if (eavail)
> - break;
> - if (signal_pending(current)) {
> - res = -EINTR;
> + if (eavail || res)
> break;
> - }
>
> if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS)) {
> timed_out = 1;
> @@ -1927,6 +1926,15 @@ static int ep_poll(struct eventpoll *ep, struct
> epoll_event __user *events,
> }
>
> send_events:
> + if (fatal_signal_pending(current))
> + /*
> + * Always short-circuit for fatal signals to allow
> + * threads to make a timely exit without the chance of
> + * finding more events available and fetching
> + * repeatedly.
> + */
> + res = -EINTR;
> +
> /*
> * Try to transfer events to user space. In case we get 0 events and
> * there's still timeout left over, we go trying again in search of
On Tue, 05 May 2020 10:42:05 +0200 Roman Penyaev <[email protected]> wrote:
> May I ask you to remove "epoll: ensure ep_poll() doesn't miss wakeup
> events" from your -mm queue? Jason lately found out that the patch
> does not fully solve the problem and this one patch is a second
> attempt to do things correctly in a different way (namely to do
> the final check under the lock). Previous changes are not needed.
Where do we stand with Khazhismel's "eventpoll: fix missing wakeup for
ovflist in ep_poll_callback"?
http://lkml.kernel.org/r/[email protected]
On Tue, May 5, 2020 at 1:04 PM Andrew Morton <[email protected]> wrote:
>
> On Tue, 05 May 2020 10:42:05 +0200 Roman Penyaev <[email protected]> wrote:
>
> > May I ask you to remove "epoll: ensure ep_poll() doesn't miss wakeup
> > events" from your -mm queue? Jason lately found out that the patch
> > does not fully solve the problem and this one patch is a second
> > attempt to do things correctly in a different way (namely to do
> > the final check under the lock). Previous changes are not needed.
>
> Where do we stand with Khazhismel's "eventpoll: fix missing wakeup for
> ovflist in ep_poll_callback"?
>
> http://lkml.kernel.org/r/[email protected]
>
My understanding is - we need the ep_poll_callback fix on a logical
level (ovfllist was never triggering wakeup), and the two follow-ups
close races - in both how we add/remove from the wait queue, and how
we observe the ready list, which are needed if we only wake when we
add events, where before we were also waking when we were splicing
ovflist events when reading the ready list. As well, the first two
together are needed for epoll60 to pass in my testing.
Khazhy
On 2020-05-05 22:03, Andrew Morton wrote:
> On Tue, 05 May 2020 10:42:05 +0200 Roman Penyaev <[email protected]>
> wrote:
>
>> May I ask you to remove "epoll: ensure ep_poll() doesn't miss wakeup
>> events" from your -mm queue? Jason lately found out that the patch
>> does not fully solve the problem and this one patch is a second
>> attempt to do things correctly in a different way (namely to do
>> the final check under the lock). Previous changes are not needed.
>
> Where do we stand with Khazhismel's "eventpoll: fix missing wakeup for
> ovflist in ep_poll_callback"?
>
> http://lkml.kernel.org/r/[email protected]
This one from Khazhismel is needed. Others are complementary to the
Khazhismel's, except the "epoll: ensure ep_poll() doesn't miss
wakeup events", which you've already removed.
Thanks.
--
Roman