Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Thomas Gleixner <tglx@linutronix.de>
To:     Ye Bin <yebin@huaweicloud.com>, dennis@kernel.org, tj@kernel.org,
        cl@linux.com, linux-mm@kvack.org, yury.norov@gmail.com,
        andriy.shevchenko@linux.intel.com, linux@rasmusvillemoes.dk
Cc:     linux-kernel@vger.kernel.org, dchinner@redhat.com,
        yebin10@huawei.com, yebin@huaweicloud.com,
        Peter Zijlstra <peterz@infradead.org>,
        Valentin Schneider <vschneid@redhat.com>,
        Linus Torvalds <torvalds@linuxfoundation.org>
Subject: Re: [PATCH v2 2/2] lib/percpu_counter: fix dying cpu compare race
In-Reply-To: <877cuj1mt1.ffs@tglx>
References: <20230406015629.1804722-1-yebin@huaweicloud.com>
 <20230406015629.1804722-3-yebin@huaweicloud.com> <877cuj1mt1.ffs@tglx>
Date:   Tue, 11 Apr 2023 08:56:48 +0200
Message-ID: <874jpm29fz.ffs@tglx>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

On Mon, Apr 10 2023 at 22:53, Thomas Gleixner wrote:
> On Thu, Apr 06 2023 at 09:56, Ye Bin wrote:
> cpu_dying_mask is a random number generator w/o cpu_hotplug_lock being
> held. And even with that lock held any cpumask operation on it is silly.
> The mask is a core detail:
>
>   commit e40f74c535b8 "cpumask: Introduce DYING mask"
>     
>     Introduce a cpumask that indicates (for each CPU) what direction the
>     CPU hotplug is currently going. Notably, it tracks rollbacks. Eg. when
>     an up fails and we do a roll-back down, it will accurately reflect the
>     direction.
>
> It does not tell anything to a user which is not aware of the actual
> hotplug state machine state.

Even if the mask is most of the time stable, it's a total disaster
performance wise. The bits in cpu_dying_mask are sticky until the next
online operation.

So for a system which has SMT enabled in BIOS, but SMT is disabled on
the kernel command line or later via the sysfs knob, this means that the
loop in __percpu_counter_sum() will iterate over all shutdown SMT
siblings forever. IOW, it will touch nr_of_shutdown_cpus() cachelines
for absolutely zero reason.

The same applies for the proposed counter.

Thanks,

        tglx