Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
Message-ID: <15f8fbd8-2dbc-3df7-c748-c76e3479e227@I-love.SAKURA.ne.jp>
Date:   Wed, 13 Sep 2023 23:30:23 +0900
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101
 Thunderbird/102.15.0
Subject: Re: BUG: soft lockup in smp_call_function
Content-Language: en-US
To:     Hillf Danton <hdanton@sina.com>,
        Sanan Hasanov <Sanan.Hasanov@ucf.edu>,
        Thomas Gleixner <tglx@linutronix.de>, peterz@infradead.org
Cc:     Linus Torvalds <torvalds@linux-foundation.org>,
        syzkaller@googlegroups.com, LKML <linux-kernel@vger.kernel.org>
References: <20230913110709.6684-1-hdanton@sina.com>
From:   Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
In-Reply-To: <20230913110709.6684-1-hdanton@sina.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2023/09/13 20:07, Hillf Danton wrote:
> 
> 	cpu1			cpu4 (see below)
> 	====			====
> 	drm_crtc_vblank_off	__run_hrtimer
> 	spin_lock_irq(&dev->event_lock);
> 	...
> 				drm_handle_vblank
> 	hrtimer_cancel		spin_lock_irqsave(&dev->event_lock, irqflags);
> 
> 
> Deadlock should have been reported instead provided the lockdep_map in
> struct timer_list were added also to hrtimer, so it is highly appreciated
> if Tetsuo or Thomas adds it before 6.8 or 6.10.

Not me. ;-)

Since hrtimer_cancel() retries forever until lock_hrtimer_base() succeeds,
we want to add a lockdep annotation into hrtimer_cancel() so that we can
detect this type of deadlock?