nathanw@MIT.EDU (Nathan J. Williams) writes:

> Given that we don't seem to have any devices with the opposite quirk,
> I kind of like Jason's idea....

It just occurred to me that since we only attempt to use large CDBs
for transfers of >255 blocks on disks >1G, we're unlikely to be aware
of old, small, crufty SCSI disks which barf on large CDBs. So making
this change may be more dangerous than I thought, and the number of
devices we currently have ONLYBIG quirks for isn't that large.

Maybe it would be safer if the umass driver could indicate its
"modernness" to the scsipi layer, perhaps by tweaking the
scsipi_bustype?

        - Nathan