Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 23 Jun 2022 12:46:35 +0200
From:   Sebastian Andrzej Siewior <bigeasy@linutronix.de>
To:     Juri Lelli <juri.lelli@redhat.com>
Cc:     linux-rt-users <linux-rt-users@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>
Subject: Re: [RT] WARNING at hrtimer_start_range_ns
Message-ID: <YrREi3qSDfNek9S1@linutronix.de>
References: <YqnygxNWOztakt8+@localhost.localdomain>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <YqnygxNWOztakt8+@localhost.localdomain>
Precedence: bulk

On 2022-06-15 16:53:55 [+0200], Juri Lelli wrote:
> Hi,
Hi,

> My understanding is that we have
> 
> fs/aio.c::read_events
>   wait_event_interruptible_hrtimeout
>     __wait_event_hrtimeout
>       hrtimer_init_sleeper_on_stack <- this might mode |=HRTIMER_MODE_HARD on RT

will set HRTIMER_MODE_HARD  of might set HRTIMER_MODE_HARD if the task has elevated priority
(RT/DL).

>         hrtimer_start_range_ns
> 	  WARN_ON_ONCE(!(mode & HRTIMER_MODE_HARD) ^ !timer->is_hard)
> 	  fires since the latter doesn't see the change of mode done by
> 	  init_sleeper

correct.

> The attached patch seems to cure the problem, although I'm not entirely
> sure it's the cleanest fix (if I'm right at all about it :).

The problem seems to be that __wait_event_hrtimeout() does
hrtimer_start_range_ns() instead of hrtimer_sleeper_start_expires(). So
if you switch that, then it should go away.

> Thanks,
> Juri

Sebastian