linux-kernel - Re: [PATCH] drm/panthor: fix queue_reset_timeout

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20251202190525.60973e38@fedora>
Date: Tue, 2 Dec 2025 19:05:25 +0100
From: Boris Brezillon <boris.brezillon@...labora.com>
To: Chia-I Wu <olvaffe@...il.com>
Cc: Steven Price <steven.price@....com>, Liviu Dudau <liviu.dudau@....com>,
 Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>, Maxime Ripard
 <mripard@...nel.org>, Thomas Zimmermann <tzimmermann@...e.de>, David Airlie
 <airlied@...il.com>, Simona Vetter <simona@...ll.ch>, Grant Likely
 <grant.likely@...aro.org>, Heiko Stuebner <heiko@...ech.de>,
 dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
 ashley.smith@...labora.com
Subject: Re: [PATCH] drm/panthor: fix queue_reset_timeout_locked

On Tue,  2 Dec 2025 09:40:28 -0800
Chia-I Wu <olvaffe@...il.com> wrote:

> queue_check_job_completion calls queue_reset_timeout_locked to reset the
> timeout when progress is made. We want the reset to happen when the
> timeout is running, not when it is suspended.
> 
> Fixes: 345c5b7cc0f85 ("drm/panthor: Make the timeout per-queue instead of per-job")
> Signed-off-by: Chia-I Wu <olvaffe@...il.com>
> ---
>  drivers/gpu/drm/panthor/panthor_sched.c | 24 ++++++++++++------------
>  1 file changed, 12 insertions(+), 12 deletions(-)
> 
> diff --git a/drivers/gpu/drm/panthor/panthor_sched.c b/drivers/gpu/drm/panthor/panthor_sched.c
> index 389d508b3848e..47e81fc45fb56 100644
> --- a/drivers/gpu/drm/panthor/panthor_sched.c
> +++ b/drivers/gpu/drm/panthor/panthor_sched.c
> @@ -1052,18 +1052,6 @@ group_is_idle(struct panthor_group *group)
>  	return hweight32(inactive_queues) == group->queue_count;
>  }
>  
> -static void
> -queue_reset_timeout_locked(struct panthor_queue *queue)
> -{
> -	lockdep_assert_held(&queue->fence_ctx.lock);
> -
> -	if (queue->timeout.remaining != MAX_SCHEDULE_TIMEOUT) {
> -		mod_delayed_work(queue->scheduler.timeout_wq,
> -				 &queue->timeout.work,
> -				 msecs_to_jiffies(JOB_TIMEOUT_MS));
> -	}
> -}
> -
>  static bool
>  group_can_run(struct panthor_group *group)
>  {
> @@ -1080,6 +1068,18 @@ queue_timeout_is_suspended(struct panthor_queue *queue)
>  	return queue->timeout.remaining != MAX_SCHEDULE_TIMEOUT;
>  }
>  
> +static void
> +queue_reset_timeout_locked(struct panthor_queue *queue)
> +{
> +	lockdep_assert_held(&queue->fence_ctx.lock);
> +
> +	if (!queue_timeout_is_suspended(queue)) {
> +		mod_delayed_work(queue->scheduler.timeout_wq,
> +				 &queue->timeout.work,
> +				 msecs_to_jiffies(JOB_TIMEOUT_MS));
> +	}

Oops. If queue_reset_timeout_locked() is called in a context where the
timeout is suspended, shouldn't we adjust the remaining value?

static void
queue_reset_timeout_locked(struct panthor_queue *queue)
{
        lockdep_assert_held(&queue->fence_ctx.lock);

        if (queue_timeout_is_suspended(queue)) { 
                queue->timeout.remaining = msecs_to_jiffies(JOB_TIMEOUT_MS);
	} else {
                mod_delayed_work(queue->scheduler.timeout_wq,
                                 &queue->timeout.work,
                                 msecs_to_jiffies(JOB_TIMEOUT_MS));
        }
}


> +}
> +
>  static void
>  queue_suspend_timeout_locked(struct panthor_queue *queue)
>  {