[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <5d834f7a-dc62-308c-1017-1fb92c0efc62@redhat.com>
Date: Tue, 25 Oct 2022 21:44:15 -0400
From: Waiman Long <longman@...hat.com>
To: Peter Zijlstra <peterz@...radead.org>
Cc: Ingo Molnar <mingo@...hat.com>, Will Deacon <will@...nel.org>,
Boqun Feng <boqun.feng@...il.com>,
linux-kernel@...r.kernel.org, john.p.donnelly@...cle.com,
Hillf Danton <hdanton@...a.com>,
Mukesh Ojha <quic_mojha@...cinc.com>,
Ting11 Wang 王婷 <wangting11@...omi.com>
Subject: Re: [PATCH v3 2/5] locking/rwsem: Limit # of null owner retries for
handoff writer
On 10/25/22 16:14, Peter Zijlstra wrote:
> On Tue, Oct 25, 2022 at 03:55:09PM -0400, Waiman Long wrote:
>
>> That is quite a number of changes spread over many different functions. That
>> is the kind of changes that may make it harder to backport to stable
>> releases.
> Yeah; don't care. it's the right thing to do. It also doesn't need to be
> reverted since it's a sane and good property for lock-ops to have.
I am sorry to confuse you. What I am saying is the original patch 2 of
the series, not what you have proposed. I am fine about disabling
preemption while running in the core rwsem code.
As a counter proposal, I will suggest modifying patch 2 to disable
preemption for the reader rwsem code as the writer code already have
preemption disabled in the most critical parts. Then I add another patch
to complete the writer side conversion for symmetry. Are you OK with that?
Cheers,
Longman
Powered by blists - more mailing lists