From: Thomas Gleixner <tglx@linutronix.de>
To: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: linux-kernel@vger.kernel.org,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>, Darren Hart <darren@dvhart.com>,
Steven Rostedt <rostedt@goodmis.org>,
fredrik.markstrom@windriver.com,
Davidlohr Bueso <davidlohr@hp.com>,
Manfred Spraul <manfred@colorfullife.com>
Subject: Re: [PATCH 2/3] futex: avoid double wake up in futex_wake() on -RT
Date: Tue, 7 Apr 2015 21:47:13 +0200 (CEST) [thread overview]
Message-ID: <alpine.DEB.2.11.1504072131400.3845@nanos> (raw)
In-Reply-To: <1428419030-20030-3-git-send-email-bigeasy@linutronix.de>
On Tue, 7 Apr 2015, Sebastian Andrzej Siewior wrote:
> futex_wake() wakes the waiter while holding the hb->lock. This leads to
> a similar double wake up on -RT if the waiter has a higher priority than
> the process perfroming the wake up.
Well, the non pi wakeup is designed in a way that the waiter side does
not take the hash bucket lock.
So if you observe that hb->lock contention and the resulting PI
boosting dance on RT, then it's not the wakeup/waiter exit path. That
must be something like this:
T1 T2
wakeup()
-->preemption
sys_exit(futex_wait)
...
sys_enter(futex_wake);
lock(hb->lock);
> -/*
> - * The hash bucket lock must be held when this is called.
> - * Afterwards, the futex_q must not be accessed.
> - */
The comment should stay here, because you are not allowed to access
the futex_q after q->lock_ptr has been set to NULL and that happens in
this function.
> -static void wake_futex(struct futex_q *q)
> +static struct task_struct *__wake_futex(struct futex_q *q)
> {
> struct task_struct *p = q->task;
>
> if (WARN(q->pi_state || q->rt_waiter, "refusing to wake PI futex\n"))
> - return;
> + return NULL;
>
> /*
> * We set q->lock_ptr = NULL _before_ we wake up the task. If
> @@ -1117,6 +1113,19 @@ static void wake_futex(struct futex_q *q)
> */
> smp_wmb();
> q->lock_ptr = NULL;
> + return p;
...
> @@ -1256,14 +1266,23 @@ futex_wake(u32 __user *uaddr, unsigned int flags, int nr_wake, u32 bitset)
> if (!(this->bitset & bitset))
> continue;
>
> - wake_futex(this);
> + if (nr_wake == 1)
> + waiter = __wake_futex(this);
> + else
> + wake_futex(this);
> if (++ret >= nr_wake)
> break;
> }
> }
>
> spin_unlock(&hb->lock);
> +
> out_put_key:
> + if (waiter) {
> + wake_up_state(waiter, TASK_NORMAL);
> + put_task_struct(waiter);
> + }
This should go before out_put_key, because none of the other code
pathes which jump there can set waiter.
Thanks,
tglx
next prev parent reply other threads:[~2015-04-07 19:46 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-04-07 15:03 improve futex on -RT by avoiding the double wake-up Sebastian Andrzej Siewior
2015-04-07 15:03 ` [PATCH 1/3] futex: avoid double wake up in PI futex wait / wake on -RT Sebastian Andrzej Siewior
2015-04-07 18:41 ` Thomas Gleixner
2015-04-10 14:42 ` [PATCH 1/3 v2] " Sebastian Andrzej Siewior
2015-04-07 15:03 ` [PATCH 2/3] futex: avoid double wake up in futex_wake() " Sebastian Andrzej Siewior
2015-04-07 19:47 ` Thomas Gleixner [this message]
2015-04-10 16:11 ` [PATCH 2/3 v2] " Sebastian Andrzej Siewior
2015-04-13 3:02 ` Davidlohr Bueso
2015-04-16 5:09 ` Davidlohr Bueso
2015-04-16 9:19 ` Thomas Gleixner
2015-04-16 10:16 ` Peter Zijlstra
2015-04-16 10:49 ` Thomas Gleixner
2015-04-16 14:42 ` Davidlohr Bueso
2015-04-16 15:54 ` Peter Zijlstra
2015-04-16 16:22 ` Davidlohr Bueso
2015-04-07 15:03 ` [PATCH 3/3] ipc/mqueue: remove STATE_PENDING Sebastian Andrzej Siewior
2015-04-07 17:48 ` Manfred Spraul
2015-04-07 18:28 ` Thomas Gleixner
2015-04-10 14:37 ` [PATCH v2] " Sebastian Andrzej Siewior
2015-04-23 22:18 ` Thomas Gleixner
2015-04-28 3:24 ` Davidlohr Bueso
2015-04-28 12:37 ` Peter Zijlstra
2015-04-28 16:36 ` Davidlohr Bueso
2015-04-28 16:43 ` Peter Zijlstra
2015-04-28 16:59 ` Davidlohr Bueso
2015-04-29 19:44 ` Manfred Spraul
2015-04-30 18:46 ` Davidlohr Bueso
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=alpine.DEB.2.11.1504072131400.3845@nanos \
--to=tglx@linutronix.de \
--cc=bigeasy@linutronix.de \
--cc=darren@dvhart.com \
--cc=davidlohr@hp.com \
--cc=fredrik.markstrom@windriver.com \
--cc=linux-kernel@vger.kernel.org \
--cc=manfred@colorfullife.com \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox