[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20181123105132.47ee57ad@vmware.local.home>
Date: Fri, 23 Nov 2018 10:51:32 -0500
From: Steven Rostedt <rostedt@...dmis.org>
To: Catalin Marinas <catalin.marinas@....com>
Cc: Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
Andrea Parri <andrea.parri@...rulasolutions.com>,
Peter Zijlstra <peterz@...radead.org>, zhe.he@...driver.com,
tglx@...utronix.de, linux-mm@...ck.org,
linux-kernel@...r.kernel.org, linux-rt-users@...r.kernel.org,
boqun.feng@...il.com
Subject: Re: [PATCH v2] kmemleak: Turn kmemleak_lock to raw spinlock on RT
On Fri, 23 Nov 2018 11:31:31 +0000
Catalin Marinas <catalin.marinas@....com> wrote:
> With qwrlocks, the readers will normally block if there is a pending
> writer (to avoid starving the writer), unless in_interrupt() when the
> readers are allowed to starve a pending writer.
>
> TLA+/PlusCal model here: ;)
>
> https://git.kernel.org/pub/scm/linux/kernel/git/cmarinas/kernel-tla.git/tree/qrwlock.tla
>
And the code appears to confirm it too:
void queued_read_lock_slowpath(struct qrwlock *lock)
{
/*
* Readers come here when they cannot get the lock without waiting
*/
if (unlikely(in_interrupt())) {
/*
* Readers in interrupt context will get the lock immediately
* if the writer is just waiting (not holding the lock yet),
* so spin with ACQUIRE semantics until the lock is available
* without waiting in the queue.
*/
atomic_cond_read_acquire(&lock->cnts, !(VAL & _QW_LOCKED));
return;
}
atomic_sub(_QR_BIAS, &lock->cnts);
/*
* Put the reader into the wait queue
*/
arch_spin_lock(&lock->wait_lock);
atomic_add(_QR_BIAS, &lock->cnts);
-- Steve
Powered by blists - more mailing lists