Bug 16016

Summary:	cam/scsi/scsi_da.c: Fujitsu M2952 doesn't like synch cache either
Product:	Base System	Reporter:	klh <klh>
Component:	kern	Assignee:	freebsd-bugs (Nobody) <bugs>
Status:	Closed FIXED
Severity:	Affects Only Me
Priority:	Normal
Version:	3.3-RELEASE
Hardware:	Any
OS:	Any

Description klh 2000-01-09 22:00:01 UTC

From system log:

Jan  9 00:17:44 bohica /kernel: (da2:amd0:0:2:0): Synchronize cache failed, status == 0xc0, scsi status == 0x2

This only happens for da2, which is a Fujitsu M2952ESP.  Other 2 drives
on system (IBM DPES and Seagate ST32550W) seem OK.

Fix: 

Examination of scsi_da.c suggests that the M2952 should be added to
da_quirk_table to join its close cousin, the M2954.

If more info or verification is desired, let me know.
How-To-Repeat: With nothing mounted on the da2 Fujitsu:

# dd if=/dev/rda2s1 of=/dev/null bs=10b

Let it run for several seconds, then interrupt out.  Error message
pops up on console.

Comment 1 Justin T. Gibbs freebsd_committer

2000-01-09 22:12:04 UTC

>>Fix:
>Examination of scsi_da.c suggests that the M2952 should be added to
>da_quirk_table to join its close cousin, the M2954.
>
>If more info or verification is desired, let me know.

Since you have hardware to test this, can you determine why the code
to weed out devices that return ILLEGAL_REQUEST fails to work for
your device?  We should be able to completely avoid quirk entries
for this particular issue.

--
Justin

Comment 2 Matt Jacob freebsd_committer

2001-01-16 07:57:46 UTC

State Changed
From-To: open->feedback

Per Justin's request, can you supply the requested info or state 
that this is no longer a problem?

Comment 3 Ken Harrenstien 2001-02-09 07:25:25 UTC

Sorry, didn't realize this was still pending.  The audit trail does
not reflect several messages that I exchanged with Justin, in which
he diagnosed the code and made a fix to pci/amd.c which resolved the
problem (completely, as far as I could tell).

I just checked and the fix is there as of rev 1.3 (14-Jan-2000); has
been out there for a while now.

Appreciate the cleanup you are doing!

--Ken

Comment 4 dwmalone freebsd_committer

2001-02-09 13:24:49 UTC

State Changed
From-To: feedback->closed

Submitter indicates that problem has been fixed.