Re: [PATCH] workqueue: add workqueue.mayday_initial_timeout

[Lai Jiangshan]

Hello, Chen

Thanks for trying to improve.

On Tue, Nov 11, 2025 at 10:55 AM Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
>
> On Tue, Nov 11, 2025 at 10:52:44AM +0800, ying chen wrote:
> > If creating a new worker takes longer than MAYDAY_INITIAL_TIMEOUT,
> > the rescuer thread will be woken up to process works scheduled on
> > @pool, resulting in sequential execution of all works. This may lead
> > to a situation where one work blocks others. However, the initial
> > rescue timeout defaults to 10 milliseconds, which can easily be
> > triggered in heavy-load environments.


This is what it was intended for. A workqueue equipped with a rescuer
is considered a single-threaded workqueue, where deadlock can occur
when there are works depending on each other.

It is recommended that the user modify the works to eliminate any
dependencies or consider using multiple workqueues. Increasing the
timeout doesn’t solve the problem; it merely hides it.

Changing the workqueue implementation to support concurrency for
the rescuer would complicate the design unnecessarily. The use of a
rescuer is only for memory reclamation, where the works should be
simple, focusing on freeing memory rather than waiting each other.

Thanks
Lai

>
> Then we should tune it automatically, not blame the admin for not tuning
> this parameter.