[2272] in linux-scsi channel archive
Re: Where to find information?
daemon@ATHENA.MIT.EDU (Richard Waltham)
Sat Aug 9 08:04:37 1997
From: Richard Waltham <dormouse@farsrobt.demon.co.uk>
To: milliByte@deathsdoor.com
Date: Sat, 9 Aug 1997 13:01:31 +0100 (BST)
Cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <19970808062438.37842@infinity.milliByte.de> from "Dirk Foersterling" at Aug 8, 97 06:24:38 am
Dirk Foersterling wrote:
>
> My former question got no responses, so I'll ask another.
>
> Where to find information of the "phase change" outputs of the ncr53c8xx
> SCSI driver? I would like to decipher all the numbers within that
> message to find out what it want to tell me.
> In fact, I would like to find out if these messages have something to do
> with the file system and partition table corruption I encountered.
>
> Here's an example:
>
> ncr53c825-0-<4,0>: phase change 2-7 6@04a07430 resid=2.
>
> And another two:
>
> Aug 2 18:01:10 infinity kernel: ncr53c825-0-<4,0>: phase change 2-7 6@0562ec20 resid=2.
> Aug 2 18:01:10 infinity kernel: ncr53c825-0-<4,0>: phase change 2-7 10@05048038 resid=4.
>
>
> -dirk
>
> --
> D i r k F "o r s t e r l i n g
> milliByte@DeathsDoor.com - http://www.informatik.uni-frankfurt.de/~dirk
> -------------
> "mit ISDN kann man im Internet schneller warten"
>
Hopefully the following is accurate. If not Gerard will no doubt correct me
when he gets back from his vacation.
Looks like you have a problem with cabling/connectors/termination. The drive
is changing from command phase to message_in before all the command has been
sent.
0 - data_out
1 - data_in
2 - command
3 - status
4 - invalid
5 - invalid
6 - message_out
7 - message_in
6@xxxxxxxx - this is probably the number of bytes and address of the
command buffer.
resid is number of bytes left to send.
The drive is very likely getting a parity error part way through the command
phase and is changing to message_in to let the host know it's had an error.
I would expect the drive to retry the command phase and carry on with the
command in which case they should not cause problems but the drive will
usually give up after a number of retries if the error persists in which
casee these may lead to problems with file system errors.
A flaky scsi bus is likely to lead to problems in the end even if retries
are succeeding.
Richard
--
Richard Waltham | Work: richard@digtalinterface.demon.co.uk
At home in | Home: dormouse@farsrobt.demon.co.uk
Southampton UK | Compuserve 100421.1276
