On 12/6/18 12:15 PM, Christoph Hellwig wrote: > On Thu, Dec 06, 2018 at 12:14:29PM -0700, Jens Axboe wrote: >> On 12/6/18 12:11 PM, Jeff Moyer wrote: >>> Jens Axboe <axboe@xxxxxxxxx> writes: >>> >>>> From: Christoph Hellwig <hch@xxxxxx> >>>> >>>> Just call blk_poll on the iocb cookie, we can derive the block device >>>> from the inode trivially. >>> >>> Does this work for multi-device file systems? >> >> It should, that's the whole purpose of having fops->iopoll. You should >> be able to derive it from inode + cookie. > > It still assumes the whole I/O will got to a single device. For > XFS this is always tree, but for btrfs this might mean I/O could > need splitting if ->iopoll was to be supported there. Right, we can only support one target for one particular piece of IO, but multi-device works fine in general. -- Jens Axboe
