From mboxrd@z Thu Jan  1 00:00:00 1970
From: hanjun.guo@linaro.org (Hanjun Guo)
Date: Mon, 10 Aug 2015 19:00:13 +0800
Subject: [PATCH v3 15/31] arm64: SMP support
In-Reply-To: <20150806095621.GB17691@e104818-lin.cambridge.arm.com>
References: <1347035226-18649-1-git-send-email-catalin.marinas@arm.com>
 <1347035226-18649-16-git-send-email-catalin.marinas@arm.com>
 <CAOZdJXVoDJM3y2mnKO+91yf2EHogKfMjumjhhi85REAfhcJM=Q@mail.gmail.com>
 <20150806095621.GB17691@e104818-lin.cambridge.arm.com>
Message-ID: <55C8843D.8050600@linaro.org>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Hi Timur,

On 08/06/2015 05:56 PM, Catalin Marinas wrote:
> On Wed, Aug 05, 2015 at 05:46:03PM -0700, Timur Tabi wrote:
>> Catalin,
>>
>> I know this patch is three years old, but I have a question about it:
>>
>> On Fri, Sep 7, 2012 at 9:26 AM, Catalin Marinas <catalin.marinas@arm.com> wrote:
>>> +static void ipi_cpu_stop(unsigned int cpu)
>>> +{
>>> +       if (system_state == SYSTEM_BOOTING ||
>>> +           system_state == SYSTEM_RUNNING) {
>>> +               raw_spin_lock(&stop_lock);
>>> +               pr_crit("CPU%u: stopping\n", cpu);
>>> +               dump_stack();
>>> +               raw_spin_unlock(&stop_lock);
>>> +       }
>>> +
>>> +       set_cpu_online(cpu, false);
>>> +
>>> +       local_fiq_disable();
>>> +       local_irq_disable();
>>> +
>>> +       while (1)
>>> +               cpu_relax();
>>> +}
>>
>> Why do you display a message and dump the stack when a CPU is stopped?
>>   I ask because emergency_restart() calls this function, and so
>> whenever I want to do an "emergency" reboot, it dumps the stack for
>> each core.  On an ARM64 system with lots of cores, this produces a ton
>> of output.  That makes emergency_restart() almost useless.
>
> This code was inherited from arch/arm which initially was always dumping
> the stack but it later gained the system_state check. Unless it's useful
> to anyone for debugging purposes, I'm ok with removing it.

Do you have patch for this one? My colleagues said that they met the
same problem :)

Thanks
Hanjun