Re: [PATCH v2] x86/mce: Fix endless loop when run task works after #MC

From: "Luck, Tony" <tony.luck@intel.com>
To: Ding Hui <dinghui@sangfor.com.cn>
Cc: bp@alien8.de, bp@suse.de, naoya.horiguchi@nec.com,
	osalvador@suse.de, peterz@infradead.org,
	linux-edac@vger.kernel.org, linux-kernel@vger.kernel.org,
	tglx@linutronix.de, mingo@redhat.com, x86@kernel.org,
	hpa@zytor.com, youquan.song@intel.com, huangcun@sangfor.com.cn,
	stable@vger.kernel.org
Subject: Re: [PATCH v2] x86/mce: Fix endless loop when run task works after #MC
Date: Tue, 6 Jul 2021 09:44:51 -0700	[thread overview]
Message-ID: <20210706164451.GA1289248@agluck-desk2.amr.corp.intel.com> (raw)
In-Reply-To: <20210706121606.15864-1-dinghui@sangfor.com.cn>

On Tue, Jul 06, 2021 at 08:16:06PM +0800, Ding Hui wrote:
> Recently we encounter multi #MC on the same task when it's
> task_work_run() has not been called, current->mce_kill_me was
> added to task_works list more than once, that make a circular
> linked task_works, so task_work_run() will do a endless loop.

I saw the same and posted a similar fix a while back:

https://www.spinics.net/lists/linux-mm/msg251006.html

It didn't get merged because some validation tests began failing
around the same time.  I'm now pretty sure I understand what happened
with those other tests.

I'll post my updated version (second patch in a three part series)
later today.

> diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c

> +	if (!cmpxchg(&current->mce_kill_me.func, NULL, ch.func)) {
> +		current->mce_addr = m->addr;
> +		current->mce_kflags = m->kflags;
> +		current->mce_ripv = !!(m->mcgstatus & MCG_STATUS_RIPV);
> +		current->mce_whole_page = whole_page(m);

You don't need an atomic cmpxchg here (nor the WRITE_ONCE() to clear it).
The task is operating on its own task_struct. Nobody else should touch
the mce_kill_me field.

-Tony