On 01/07/21 08:19, Christoph Hellwig wrote:
http://git.infradead.org/users/hch/block.git/shortlog/refs/heads/scsi-ioctl But more importantly, dm has no business interpreting the errors. Just like how SG_IO processing generally does not look at the error and just returns it to userspace and leaves error handling to the caller so should be the decision to resubmit it in a multi-path setup.
The problem is that userspace does not have a way to direct the command to a different path in the resubmission. It may not even have permission to issue DM_TABLE_STATUS, or to access the /dev nodes for the underlying paths, so without Martin's patches SG_IO on dm-mpath is basically unreliable by design.
Paolo
