On Tue, Jun 26, 2018 at 12:53:32PM +0200, Douglas Gilbert wrote:
> The SCSI subsystem may not be the primary storage medium on a system, so
> bringing down the system because the SCSI mid-level has an unrecoverable
> error may be overkill. Think an embedded system with the rootfs on a
> /dev/mmcblk<n>p<m> device with the only "SCSI" device being an external
> USB port for a memory key (I was fighting with a 3D printer was just that
> config a few days ago). Then the user plugs in the world's worst USB
> key (billion of candidates). Isn't one error strategy just to report,
> ignore and continue?
Exactly and even if SCSI is the primary storage medium BUG_ON()s are
just bad, sorry. It's always better to do error recovery then hard
crashing a machine as it may be able to write out logs when doing
error recovery so one actually has a chance debugging the reasons why
it happened.
Byte,
Johannes
--
Johannes Thumshirn Storage
[email protected] +49 911 74053 689
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (AG Nürnberg)
Key fingerprint = EC38 9CAB C2C4 F25D 8600 D0D0 0393 969D 2D76 0850
