Everhart, Glenn
From:	system@SendSpamHere.ORG
Sent:	Friday, August 28, 1998 7:04 AM
To:	Info-VAX@Mvb.Saic.Com
Subject:	Re: Solution: Why does Seagate ST19171N fail on INIT cmd?
In article <35E597EB.EEDA903@Mvb.Saic.Com>, Mark Berryman <Mark.Berryman@Mvb.Saic.Com> writes:
>Brian Schenkenberger, VAXman- wrote:
>> 
>> In article <6s45vt$96v$1@hecate.umd.edu>, bleau@umdsp.umd.edu (Lawrence Bleau) writes:
>> >Hello, folks!  I'm back, posting a solution, rather than a problem.
>
>[most of the description deleted...]
>
>> >I gave the drive to a colleague, who used a utility to turn off the two bits
>> >Stu mentioned.  It worked!  The drive can now be INITed, written, read,
>> >everything.
>> 
>> Glenn (Mr. SCSI) Everhart told me of this problem when I recently queried
>> him about a Seagate drive.
>> 
>> Could you get your "colleague" to explain how he went about disabling the
>> two bits: ARRE and AWRE on this drive?  It would make my like a whole lot
>> easier and less stressful.
>
>It should be pretty straight-forward (if you have a system recent enough
>to have the directory sys$etc and the utility contained therein called
>SCSI_MODE.EXE).
>
>Please note that all values given below are in hex which is the only
>radix understood by this utility.
>
>You will need DIAGNOSE and PHY_IO privilege.
>
>Issue a command as follows:
>
>$ mcr sys$etc:scsi_mode -devname dkb300 -page 1
>  (-devname specifies the disk to be read and page 1 is the
>   read-write error recovery page)
>
>This command will read the current value of page 1 and indicate which
>bits can be changed.  The resulting display should look something like
>this:
>
>$! Processing Page #1h
>$!
>$! Cur   00______ 04______ 08______ 0C______ 10______ 14______ 18______ 1C______
>$! 0000  17001008 001F4B2C 00000200 810AE408 78000000 08000000
>$!
>$! Chng  00______ 04______ 08______ 0C______ 10______ 14______ 18______ 1C______
>$! 0000  17001008 001F4B2C 00000200 810AFFFF FFFFFFFF FFFFFFFF
>$!
>
>The contents of page 1 are displayed beginning at offset 0C as follows:
>
>0C = 81 = page number with the high bit set (which means the page can be saved)
>0D = 0A = size of the rest of the page
>0E = E4 = the read-write recovery flags, the value you want to change
>          (and, as you can see from the change mask, the 1st byte that
>           can be changed)
>
>The AWRE and ARRE flags are the 2 high-order bits in the byte, thus - in
>this example - we want to change the E4 to 24.
>
>Issue the command: 
>
>mcr sys$etc:scsi_mode -devname dkb300 -page 1 -offset 0E 24 -devtyp "????"
>
>and it will display the change you are about to make and, after confirmation,
>make the change.  Note that the -devtyp switch is required as a sanity check
>and must equal the device-type string returned as the result of a SCSI INQUIRY
>command (in my case: rz26).  If you don't know the devtyp of the device in
>question, issue the command without the switch and it will tell you.
>
>Hope this helps,
>
>Mark Berryman
>Mark.Berryman@Mvb.Saic.Com

Thanks Mark,

I looked at the SCSI_INFO and SCSI_MODE stuff in SYS$ETC but there was no
clear documentation to indicate which bits needed diddling.  Thanks much
for pointing out which bits.

--
VAXman- OpenVMS APE certification number: AAA-0001          VAXman@TMESIS.COM