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;
MIME-Version: 1.0
From:   Yun Levi <ppbuk5246@gmail.com>
Date:   Fri, 10 Sep 2021 10:20:25 +0900
Message-ID: <CAM7-yPRHKg-=7ezm5PpxAvhmBx6_bXRkeLsmCNuAWH-DN4MKCA@mail.gmail.com>
Subject: About Full c3-stop status.
To:     Thomas Gleixner <tglx@linutronix.de>, preeti@linux.vnet.ibm.com,
        mingo@elte.hu, rmk+lkml@arm.linux.org.uk,
        dmitri.vorobiev@movial.com
Cc:     youngjun.park@ahnlab.com,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi all.

Sorry to interrupt you but I have a question while I am reading the
tick-broadcast codes.

When a cpu goes to c3-stop, I confirm it calls the
tick_broadcast_enter function with TICK_BROADCAST_ENTER.

While entering, it calls the broadcast_needs_cpu with current cpu.

In the last condition of broadcast_needs_cpu, it checks whether the
broadcast device is bound on this cpu.

However, except the ce_broadcast_hrtimer in tick-broadcast-hrtimer.c,
all other real devices set bound_on fields as 0, Therefore, it seems
that cpu 0 couldn't enter c3-stop though it has the feature.

My question is
     1. Couldn't cpu 0 enter c3-stop?
     2. If true, why don't we set bound_on fields specifically?

Thanks.

HTH.
Levi.