[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200820062024.GB6447@lst.de>
Date: Thu, 20 Aug 2020 08:20:24 +0200
From: Christoph Hellwig <hch@....de>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Peter Zijlstra <peterz@...radead.org>,
Christoph Hellwig <hch@....de>, Ingo Molnar <mingo@...nel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Will Deacon <will@...nel.org>,
"Paul E. McKenney" <paulmck@...nel.org>,
Jens Axboe <axboe@...nel.dk>,
Chris Wilson <chris@...is-wilson.co.uk>,
David Miller <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>,
Frédéric Weisbecker <fweisbec@...il.com>,
Oleg Nesterov <oleg@...hat.com>,
Vincent Guittot <vincent.guittot@...aro.org>
Subject: Re: [RFC][PATCH v2 08/10] smp,irq_work: Use the new irq_work API
On Wed, Aug 19, 2020 at 11:50:55AM -0700, Linus Torvalds wrote:
> On Wed, Aug 19, 2020 at 12:22 AM <peterz@...radead.org> wrote:
> >
> > That is, the external serialization comes from the non-atomic
> > test-and-set they both have. This works nicely when there is external
> > state that already serializes things, but totally comes apart (and
> > causes trivial list corruption) when you get it wrong.
>
> Quite often, there just isn't any *need* for serialization, because
> there is only ever one op active.
Yes, that's pretty much the block use case. The request gets completed
once, and and the IPI is used to bounce it back to the issuing cpu.
Powered by blists - more mailing lists