Subject: Re: problems with ahc vs. format command
To: Manuel Bouyer <>
From: Chuck Silvers <>
List: tech-kern
Date: 06/12/2001 22:57:19

commenting out that line gives a better result.  scsictl format still spits
out an error immediately:

sd2(esp0:0:2:0): esp0: timed out [ecb 0xf087a038 (flags 0x101, dleft 0, stat 0)$

but then it appeared to continue with the format operation.
I'm now back to the state where I can read and write the beginning of
the disk, but all the bad sectors from before are still there (oh well).
"scsictl reassign" doesn't seem to work either:

15 sparc5:~ # scsictl sd2 reassign 8566
/dev/rsd2c: SCSI command timed out
/dev/rsd2c: device is busy
/dev/rsd2c: Check Condition on CDB: 07 00 00 00 00 00
    SENSE KEY: Media Error
   INFO FIELD: 8544
     ASC/ASCQ: Unrecovered Read Error

and on the console:

sd2(esp0:0:2:0):  Check Condition on CDB: 0x08 00 21 00 80 00
    SENSE KEY:  Media Error
   INFO FIELD:  8544
     ASC/ASCQ:  Unrecovered Read Error

there are bunch of sectors in the 85xx range that have problems even after
several reassigns, so this disk looks like it's just too damaged to salvage.


On Tue, Jun 12, 2001 at 04:58:10PM +0200, Manuel Bouyer wrote:
> On Tue, Jun 12, 2001 at 11:39:46PM +0900, Izumi Tsutsui wrote:
> > According to OpenBSD iha(4) man page, no.
> > (Though it has some structures for tags..)
> OK, so it's possible that this command doesn't work with tagged queueing.
> Chuck, can you try disabling tagged queuing on the ahc or ncr53c9x driver
> (just comment out the
> xm.xm_mode |= PERIPH_CAP_TQING
> part in the relevant driver) and see if it helps ?
> If so, then we'll have to rethink about how tags are allocated in the
> mid layer.
> --
> Manuel Bouyer, LIP6, Universite Paris VI. 
> --