From: Peter Zijlstra <peterz@infradead.org>
To: Boqun Feng <boqun.feng@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>,
Ingo Molnar <mingo@kernel.org>,
Juri Lelli <juri.lelli@redhat.com>,
Steven Rostedt <rostedt@goodmis.org>,
Davidlohr Bueso <dave@stgolabs.net>,
Will Deacon <will@kernel.org>, Waiman Long <longman@redhat.com>,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
Mike Galbraith <efault@gmx.de>,
Daniel Bristot de Oliveira <bristot@redhat.com>,
LKML <linux-kernel@vger.kernel.org>
Subject: Re: [RFC] locking: rwbase: Take care of ordering guarantee for fastpath reader
Date: Thu, 2 Sep 2021 13:55:29 +0200 [thread overview]
Message-ID: <YTC7sariSyBW48nh@hirez.programming.kicks-ass.net> (raw)
In-Reply-To: <20210901150627.620830-1-boqun.feng@gmail.com>
On Wed, Sep 01, 2021 at 11:06:27PM +0800, Boqun Feng wrote:
> Sorry I'm late for the party of PREEMPT_RT lock review. Just want to
> point the problem with this patch. Not even compile test, but show the
> idea and check if I'm missing something subtle.
No worries, glad you could have a look. I think you're right and we
missed this.
> diff --git a/kernel/locking/rwbase_rt.c b/kernel/locking/rwbase_rt.c
> index 4ba15088e640..a1886fd8bde6 100644
> --- a/kernel/locking/rwbase_rt.c
> +++ b/kernel/locking/rwbase_rt.c
> @@ -41,6 +41,12 @@
> * The risk of writer starvation is there, but the pathological use cases
> * which trigger it are not necessarily the typical RT workloads.
> *
> + * Fast-path orderings:
> + * The lock/unlock of readers can run in fast paths: lock and unlock are only
> + * atomic ops, and there is no inner lock to provide ACQUIRE and RELEASE
> + * semantics of rwbase_rt. Atomic ops then should be stronger than _acquire()
> + * and _release() to provide necessary ordering guarantee.
> + *
> * Common code shared between RT rw_semaphore and rwlock
> */
>
> @@ -53,6 +59,7 @@ static __always_inline int rwbase_read_trylock(struct rwbase_rt *rwb)
> * set.
> */
> for (r = atomic_read(&rwb->readers); r < 0;) {
> + /* Fully-ordered if cmpxchg() succeeds, provides ACQUIRE */
> if (likely(atomic_try_cmpxchg(&rwb->readers, &r, r + 1)))
> return 1;
> }
> @@ -162,6 +169,8 @@ static __always_inline void rwbase_read_unlock(struct rwbase_rt *rwb,
> /*
> * rwb->readers can only hit 0 when a writer is waiting for the
> * active readers to leave the critical section.
> + *
> + * dec_and_test() is fully ordered, provides RELEASE.
> */
> if (unlikely(atomic_dec_and_test(&rwb->readers)))
> __rwbase_read_unlock(rwb, state);
> @@ -172,7 +181,11 @@ static inline void __rwbase_write_unlock(struct rwbase_rt *rwb, int bias,
> {
> struct rt_mutex_base *rtm = &rwb->rtmutex;
>
> - atomic_add(READER_BIAS - bias, &rwb->readers);
> + /*
> + * _release() is needed in case that reader is in fast path, pairing
> + * with atomic_try_cmpxchg() in rwbase_read_trylock(), provides RELEASE
> + */
> + (void)atomic_add_return_release(READER_BIAS - bias, &rwb->readers);
Very narrow race with the unlock below, but yes agreed.
> raw_spin_unlock_irqrestore(&rtm->wait_lock, flags);
> rwbase_rtmutex_unlock(rtm);
> }
> @@ -216,8 +229,14 @@ static int __sched rwbase_write_lock(struct rwbase_rt *rwb,
> */
> rwbase_set_and_save_current_state(state);
>
> - /* Block until all readers have left the critical section. */
> - for (; atomic_read(&rwb->readers);) {
> + /*
> + * Block until all readers have left the critical section.
> + *
> + * _acqurie() is needed in case that the reader side runs in the fast
> + * path, pairing with the atomic_dec_and_test() in rwbase_read_unlock(),
> + * provides ACQUIRE.
> + */
> + for (; atomic_read_acquire(&rwb->readers);) {
> /* Optimized out for rwlocks */
> if (rwbase_signal_pending_state(state, current)) {
> __set_current_state(TASK_RUNNING);
I think we can restructure things to avoid this one, but yes. Suppose we
do:
readers = atomic_sub_return_relaxed(READER_BIAS, &rwb->readers);
/*
* These two provide either an smp_mb() or an UNLOCK+LOCK
* ordering, either is strong enough to provide ACQUIRE order
* for the above load of @readers.
*/
rwbase_set_and_save_current_state(state);
raw_spin_lock_irqsave(&rtm->wait_lock, flags);
while (readers) {
...
readers = atomic_read(&rwb->readers);
if (readers)
rwbase_schedule();
...
}
> @@ -229,6 +248,9 @@ static int __sched rwbase_write_lock(struct rwbase_rt *rwb,
> /*
> * Schedule and wait for the readers to leave the critical
> * section. The last reader leaving it wakes the waiter.
> + *
> + * _acquire() is not needed, because we can rely on the smp_mb()
> + * in set_current_state() to provide ACQUIRE.
> */
> if (atomic_read(&rwb->readers) != 0)
> rwbase_schedule();
> @@ -253,7 +275,11 @@ static inline int rwbase_write_trylock(struct rwbase_rt *rwb)
> atomic_sub(READER_BIAS, &rwb->readers);
>
> raw_spin_lock_irqsave(&rtm->wait_lock, flags);
> - if (!atomic_read(&rwb->readers)) {
> + /*
> + * _acquire() is needed in case reader is in the fast path, pairing with
> + * rwbase_read_unlock(), provides ACQUIRE.
> + */
> + if (!atomic_read_acquire(&rwb->readers)) {
Moo; the alternative is using dec_and_lock instead of dec_and_test, but
that's not going to be worth it.
> atomic_set(&rwb->readers, WRITER_BIAS);
> raw_spin_unlock_irqrestore(&rtm->wait_lock, flags);
> return 1;
> --
> 2.32.0
>
next prev parent reply other threads:[~2021-09-02 11:56 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-01 15:06 [RFC] locking: rwbase: Take care of ordering guarantee for fastpath reader Boqun Feng
2021-09-01 18:53 ` Waiman Long
2021-09-01 20:22 ` Davidlohr Bueso
2021-09-02 5:02 ` Boqun Feng
2021-09-02 11:55 ` Peter Zijlstra [this message]
2021-09-03 14:50 ` Boqun Feng
2021-09-04 10:12 ` Peter Zijlstra
2021-09-04 10:14 ` Peter Zijlstra
2021-09-04 10:19 ` Boqun Feng
2021-09-08 11:51 ` Peter Zijlstra
2021-09-08 12:14 ` Peter Zijlstra
2021-09-08 13:00 ` Boqun Feng
2021-09-08 13:08 ` Boqun Feng
2021-09-08 14:41 ` Peter Zijlstra
2021-09-08 14:49 ` Peter Zijlstra
2021-09-08 18:34 ` Davidlohr Bueso
2021-09-08 13:27 ` Boqun Feng
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=YTC7sariSyBW48nh@hirez.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=bigeasy@linutronix.de \
--cc=boqun.feng@gmail.com \
--cc=bristot@redhat.com \
--cc=dave@stgolabs.net \
--cc=efault@gmx.de \
--cc=juri.lelli@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=longman@redhat.com \
--cc=mingo@kernel.org \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=will@kernel.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;
as well as URLs for NNTP newsgroup(s).