[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87y19l2rpx.ffs@tglx>
Date: Wed, 10 Apr 2024 16:54:34 +0200
From: Thomas Gleixner <tglx@...utronix.de>
To: Ingo Molnar <mingo@...nel.org>, linux-kernel@...r.kernel.org
Cc: linux-tip-commits@...r.kernel.org, Mirsad Todorovac
<mirsad.todorovac@....unizg.hr>, Sean Anderson <sean.anderson@...o.com>,
x86@...nel.org
Subject: Re: [tip: timers/urgent] timekeeping: Use READ/WRITE_ONCE() for
tick_do_timer_cpu
On Wed, Apr 10 2024 at 10:15, Ingo Molnar wrote:
>> @@ -232,7 +233,7 @@ static void tick_setup_device(struct tick_device *td,
>> !tick_nohz_full_cpu(cpu)) {
>> tick_take_do_timer_from_boot();
>> tick_do_timer_boot_cpu = -1;
>> - WARN_ON(tick_do_timer_cpu != cpu);
>> + WARN_ON(READ_ON_ONCE(tick_do_timer_cpu) != cpu);
>
> So there's no such thing as READ_ON_ONCE() - I've amended the commit with
> the fix below.
Can you hand me a brown paper bag please?
Powered by blists - more mailing lists