[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YG3j1j/MNCszy12w@hirez.programming.kicks-ass.net>
Date: Wed, 7 Apr 2021 18:54:46 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Christoph Müllner <christophm30@...il.com>
Cc: Christoph Hellwig <hch@...radead.org>, Guo Ren <guoren@...nel.org>,
linux-riscv <linux-riscv@...ts.infradead.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
linux-csky@...r.kernel.org,
linux-arch <linux-arch@...r.kernel.org>,
Guo Ren <guoren@...ux.alibaba.com>,
Will Deacon <will@...nel.org>, Ingo Molnar <mingo@...hat.com>,
Waiman Long <longman@...hat.com>,
Arnd Bergmann <arnd@...db.de>, Anup Patel <anup@...infault.org>
Subject: Re: [PATCH v4 3/4] locking/qspinlock: Add
ARCH_USE_QUEUED_SPINLOCKS_XCHG32
On Wed, Apr 07, 2021 at 05:52:36PM +0200, Peter Zijlstra wrote:
> On Wed, Apr 07, 2021 at 04:29:12PM +0200, Christoph Müllner wrote:
> > The comparison with sparc64 is not applicable, as sparc64 does not
> > have LL/SC instructions.
>
> Sparc64 has CAS, without hardware fwd progress. It has to do software
> backoff for failed CAS in order to do software fwd progress.
Again, the longer answer is that this applies identically to LL/SC and
I'm sure we actually have (or had) an architecture in tree that did just
that. I just cannot remember which architecture that was.
Powered by blists - more mailing lists