> No quite my plan is:
>
> 1. Rip it off.
> 2. Reimplement stuff if and only if someone really shows pressure
> for using it.
>
> The "command parsing" excess is certainly going to go.
Its maybe handy actually. Without command parsing I can tell the drive to
do anything without good control - you know say like all the upcoming SSSCA
encrypt chunks of your harddisk so you can never get them back stuff.
The important bit is that for each command you must know the sequence of
phases. Get it wrong and your storage system goes off to visit undefined
states. I don't like my disks in undefined states because it tends to leave
them with undefined content.
Two things I do think wants considering
#1 Can the same thing be done by passing the command and sequence of
transitions from user space (scsi generic takes that approach but
scsi is a little more forgiving since the bogus transition will
screw your command in a "oh whoops" detectable manner). IDE
has a nice habit of explaining you screwed up by scribbling on
the disk and/or locking solid
#2 Shoot all the little routines and make them into a table.
That would tidy it no end.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Mar 07 2002 - 21:00:51 EST