Re: scsi_reset requirement questions

Doug Ledford (dledford@redhat.com)
Fri, 13 Aug 1999 12:03:23 -0400

Alan Cox wrote:
>
> > > Is the io_request lock held for this case since it will be held calling
> > > old_scsi_done() ? It isnt held for the other cases of reset.
> >
> > It is held, and it is held in all other cases as well. The other cases all
> > originate from SCSI timeouts, check the function old_scsi_timesout() and
> > you'll see we grab the spinlock in there before we call the abort or reset
> > routines.
>
> Oh dear. Then we have a _lot_ of broken for SMP scsi drivers.

Huh? Explain please. When Linus and myself updated the various SCSI drivers
we took this into account. Most of the drivers never did a damn thing in the
reset routines, so those were easy:

return(SCSI_RESET_SNOOZE);

But in the other cases, we tried to make sure things were done right. What
exactly are you referring to?

-- 
  Doug Ledford   <dledford@redhat.com>
   Opinions expressed are my own, but
      they should be everybody's.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/