Messages in this thread | | | From | "Xue, Zhan" <> | Subject | RE: Regarding WQ_MEM_RECLAIM | Date | Fri, 26 Aug 2022 07:07:57 +0000 |
| |
Hi Tejun, The brief point seems the answer to one concern about " workqueue: WQ_MEM_RECLAIM (current wq) is flushing !WQ_MEM_RECLAIM events:(target wq)". If current wq is for memory reclaim , the target wq should be marked with WQ_MEM_RECLAIM as well in case that the flushing target wq (its work items) in the context of current wq is inevitable.
BR Xuezhan -----Original Message----- From: Tejun Heo <htejun@gmail.com> On Behalf Of tj@kernel.org Sent: Friday, August 26, 2022 2:03 AM To: Xue, Zhan <zhan.xue@intel.com> Cc: florian@mickler.org; LKML <linux-kernel@vger.kernel.org> Subject: Re: Regarding WQ_MEM_RECLAIM
On Wed, Aug 24, 2022 at 06:10:56AM +0000, Xue, Zhan wrote: > Convert to plain text..
The email's formatting is too painful to reply directly.
Here are two brief points:
* Just don't share the same workqueue between work items which need forward progress guarantee and ones which don't.
* If something can block memory reclaim, it is in the memory reclaim path by definition.
Thanks.
-- tejun
| |