cvs commit: src/sys/cam/scsi scsi_da.c

[Paul Saab]

Nate Lawson (nate at root.org) wrote:
> Please place quirks in the proper section.  I believe you put that one in
> the USB section.  It would also help if you described what happens when
> they are sent a SYNC_CACHE (i.e. hang).  Most useful would be a PR entry
> with the dmesg and description of the behavior.  The reason for this is
> that especially in the USB case, many quirks had proliferated that were
> overly wildcarded and which weren't actually needed but there was no way
> to be certain which were needed and which weren't.  It would be nice to
> know when we can remove the quirk if this turns out to be an underlying
> driver problem and not a drive problem.

Sorry, I didn't notice this.. I was just taking the code from the Y! tree
and didn't realize it had been reorg'd in this way.

As for what I am trying to accomplish, the CISS controller doesn't support
SYNC_CACHE at all.  CISS only supports a small subset of the scsi spec.
The driver tries to fake it, but I have found it is unsafe to flush the
cache except on reboot.