On Wed, Jan 27, 2021 at 05:03:06PM +0000, Christoph Hellwig wrote: > On Mon, Jan 25, 2021 at 09:06:19PM -0800, Darrick J. Wong wrote: > > From: Darrick J. Wong <djwong@xxxxxxxxxx> > > > > When CONFIG_XFS_DEBUG=y, set WQ_SYSFS on all workqueues that we create > > so that we (developers) have a means to monitor cpu affinity and whatnot > > for background workers. In the next patchset we'll expose knobs for > > some of the workqueues publicly and document it, but not now. > > I don't really think this is a very good idea. If we want something like > this it should be kernel-wide and coordinated with the workqueue > maintainer, but I'm a little doubtful about the use case. I don't think it is particular useful kernel wide. If it was, the maintainer wouldn't have introduced a per-workqueue flag for this functionality. The reality is that very few workqueues in the system can expand out into running thousands of kworker threads like the XFS workqueues often do. And, really, there's nothing useful a typical user can do at this point with the workqueue knobs to "tune" the behaviour - the visibility/control the workqueue sysfs knobs provide at this point is really only useful to XFS developers running tests in controlled conditions... Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx
