On 10/18/11 05:15, Mike Snitzer wrote: > FYI, this patch from Tejun should also fix the concern I had relative to > mpath's underlying devices' request_queues: > https://lkml.org/lkml/2011/10/16/148 Oh, that's good information. Since the following commit, a queue can be accessed after the device being released: commit f758eeabeb96f878c860e8f110f94ec8820822a9 Author: Christoph Hellwig <hch@xxxxxxxxxxxxx> Date: Thu Apr 21 18:19:44 2011 -0600 writeback: split inode_wb_list_lock into bdi_writeback.list_lock I.e. __blkdev_put disk->fops->release() .. bdev_inode_switch_bdi() bdi_lock_two() [access to bdev->bd_inode->i_data.backing_dev_info] I think Tejun's patch is reasonable. Thanks, -- Jun'ichi Nomura, NEC Corporation -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
