Date:   Sun, 26 Feb 2023 17:51:07 +0100
From:   Peter Zijlstra <peterz@infradead.org>
To:     longman@redhat.com, mingo@redhat.com, will@kernel.org
Cc:     linux-kernel@vger.kernel.org, boqun.feng@gmail.com
Subject: Re: [PATCH 3/6] locking/rwsem: Rework writer wakeup
Message-ID: <Y/uN+89FlTw45uiA@hirez.programming.kicks-ass.net>
References: <20230223122642.491637862@infradead.org>
 <20230223123319.487908155@infradead.org>
 <Y/t1AwGC9OoN/lFc@hirez.programming.kicks-ass.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Y/t1AwGC9OoN/lFc@hirez.programming.kicks-ass.net>
Precedence: bulk

On Sun, Feb 26, 2023 at 04:04:35PM +0100, Peter Zijlstra wrote:
> On Thu, Feb 23, 2023 at 01:26:45PM +0100, Peter Zijlstra wrote:
> > @@ -1072,7 +1067,7 @@ rwsem_down_read_slowpath(struct rw_semap
> >  	for (;;) {
> >  		set_current_state(state);
> >  		if (!smp_load_acquire(&waiter.task)) {
> > -			/* Matches rwsem_mark_wake()'s smp_store_release(). */
> > +			/* Matches rwsem_waiter_wake()'s smp_store_release(). */
> >  			break;
> >  		}
> >  		if (signal_pending_state(state, current)) {
> > @@ -1143,54 +1138,36 @@ rwsem_down_write_slowpath(struct rw_sema
> >  	} else {
> >  		atomic_long_or(RWSEM_FLAG_WAITERS, &sem->count);
> 
> Found it; if we remove the try_write_lock below, then at least this
> new-waiter path needs to still do a trylock.
> 
> Let me go test the other patches on top of all this and push out a fresh
> set if that all still works.

queue.git locking/core

We'll see what the robots make of it.