[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <185e1d7a-719d-4fec-8e4f-e7cf0bea1573@huawei.com>
Date: Tue, 14 May 2024 09:37:04 +0800
From: Wenchao Hao <haowenchao2@...wei.com>
To: Tejun Heo <tj@...nel.org>, Aaron Tomlin <atomlin@...mlin.com>
CC: Lai Jiangshan <jiangshanlai@...il.com>, <linux-kernel@...r.kernel.org>,
Wenchao Hao <haowenchao22@...il.com>, <neelx@...hat.com>
Subject: Re: [RFC PATCH] workqueue: Fix rescuer task's name truncated
On 2024/5/14 1:17, Tejun Heo wrote:
> Hello
>
> On Sat, May 11, 2024 at 03:15:11PM +0100, Aaron Tomlin wrote:
>>> @@ -5439,6 +5439,8 @@ static int init_rescuer(struct workqueue_struct *wq)
>>> return ret;
>>> }
>>>
>>> + snprintf(rescuer->desc, sizeof(rescuer->desc), "%s", wq->name);
>
> Can you please address the testbot reported warning?
>
Of course.
The warning is also reported here: https://lore.kernel.org/all/20240513030639.3772468-1-haowenchao2@huawei.com/
It's caused because commit 31c89007285d ("workqueue.c: Increase workqueue name length") increase
WQ_NAME_LEN from 24 to 32, but forget to increase WORKER_DESC_LEN which should equalto WQ_NAME_LEN.
Same usage can be found in process_one_work(), it called strscpy() which would not WARNING.
I sent a V2 patch with following changes appended for this warning, maybe you missed them.
diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h
index 158784dd189a..72031fa80414 100644
--- a/include/linux/workqueue.h
+++ b/include/linux/workqueue.h
@@ -92,7 +92,7 @@ enum wq_misc_consts {
WORK_BUSY_RUNNING = 1 << 1,
/* maximum string length for set_worker_desc() */
- WORKER_DESC_LEN = 24,
+ WORKER_DESC_LEN = 32,
};
/* Convenience constants - of type 'unsigned long', not 'enum'! */
>>> wq->rescuer = rescuer;
>>> if (wq->flags & WQ_UNBOUND)
>>> kthread_bind_mask(rescuer->task, wq_unbound_cpumask);
>>> @@ -6289,6 +6291,8 @@ void wq_worker_comm(char *buf, size_t size, struct task_struct *task)
>>> worker->desc);
>>> }
>>> raw_spin_unlock_irq(&pool->lock);
>>> + } else if (worker->desc[0] != '\0') {
>>> + scnprintf(buf + off, size - off, "-%s", worker->desc);
>>> }
>>> }
>>>
>>> --
>>
>> Reviewed-by: Aaron Tomlin <atomlin@...mlin.com>
>
> The patch looks fine to me otherwise.
>
> Thanks.
>
Powered by blists - more mailing lists