On Mon, May 11, 2015 at 04:52:14PM +0200, Hannes Reinecke wrote: > Sigh. This is one of the sore topics with the SCSI stack. > > The default sense code handling is correct only for filesystem > I/O; BLOCK_PC callers are expected to handle all errors themselves. > Which typically is a pain as one always forgets the one or the > other issue. > > The device handlers have a callout into that generic function > to handle and device handler specific sense codes. > > So with that I do agree that calling alua_check_sense() here > is dubious as it should have been run from the generic path already. > > Will be checking and fixing it up. What I meant is that we really shouldn't handle the sense codes in the ALUA handler - they are generic SCSІ sense codes and we'd better handle them in the core code, ditto for the other device handlers actually. Now the problem of BLOCK_PC ignoring the sense handling is a different one, but why don't we export scsi_check_sense and allow BLOCK_PC callers like the device handlers reuse the logic instead of duplicating it? -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
