On Wed, Jun 30, 2021 at 03:08:11PM -0700, Bart Van Assche wrote: > On 6/30/21 9:54 AM, Martin Wilck wrote: > > @Martin, @Bart, do you have a suggestion for me? > > The code in block/scsi_ioctl.c exists in the block layer since until > recently most of it was used by both the IDE and SCSI code. Now that > drivers/ide is gone (thanks Christoph!), how about moving block/bsg.c > and block/scsi_ioctl.c to drivers/scsi/? As far as I can see the BSG > code is only used by SCSI drivers: I have a series to clean this mess up: 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.
