From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: switch blk-cgroup to work on gendisk Date: Thu, 19 Jan 2023 18:05:26 +0100 Message-ID: <20230119170526.GA5050@lst.de> References: <20230117081257.3089859-1-hch@lst.de> Mime-Version: 1.0 Return-path: Content-Disposition: inline In-Reply-To: List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Tejun Heo Cc: Christoph Hellwig , Jens Axboe , Josef Bacik , linux-block-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org On Thu, Jan 19, 2023 at 07:03:31AM -1000, Tejun Heo wrote: > On Tue, Jan 17, 2023 at 09:12:42AM +0100, Christoph Hellwig wrote: > > blk-cgroup works on only on live disks and "file system" I/O from bios. > > This all the information should be in the gendisk, and not the > > request_queue that also exists for pure passthrough request based > > devices. > > Can't blk-throttle be used w/ bio based ones tho? I always thought that was > the reason why we didn't move it into rq-qos framework. Yes, it can. Not sure if my sentence was unclear, but: - everything doing non-passthrough I/O only should be in the gendisk - everything related to blk-mq, including infrastruture for passthrough should remain in the request_queue The idea that the request_queue will eventually become a blk-mq only data structure and not exist (or just have a very leight weight stub) for bio based drivers.