Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Thomas Gleixner <tglx@linutronix.de>
To:     Yunfeng Ye <yeyunfeng@huawei.com>
Cc:     linux-kernel@vger.kernel.org, Shiyuan Hu <hushiyuan@huawei.com>,
        Hewenliang <hewenliang4@huawei.com>,
        Marc Zyngier <maz@kernel.org>
Subject: Re: [PATCH] genirq/proc: Show percpu irq affinity
In-Reply-To: <aa537126-4646-08e2-6bea-833f71b20ade@huawei.com>
Date:   Sat, 22 Aug 2020 13:35:10 +0200
Message-ID: <87mu2moqvl.fsf@nanos.tec.linutronix.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Sat, Aug 22 2020 at 17:33, Yunfeng Ye wrote:
> On 2020/8/13 16:27, Thomas Gleixner wrote:
> For example, the cmdline "irqaffinity=3D0,1,126,127" on the 128 cores sys=
tem=EF=BC=9A
>
> [root@localhost ~]# cat /proc/irq/4/smp_affinity_list
> 0-1,126-127
>
> The irq 4 is "arch_timer" interrupt, which is a percpu interrupt.
>
> So is it necessary to show the percpu irq affinity correct?

Yes, it makes sense to do so, but you used the wrong check. The correct
one is:

irq_settings_is_per_cpu_devid()

which will not wreckage the output for other per cpu marked interrupts
which never set the percpu_affinity pointer with the obvious
consequences... The pointer would need a NULL check in any case, but it
might be more straight forward to update affinity when percpu_affinity
is initialized. Haven't looked in detail though.

Thanks,

        tglx