mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Justin T. Gibbs" <gibbs@scsiguy.com>
To: Krzysztof Halasa <khc@pm.waw.pl>
Cc: linux-kernel@vger.kernel.org
Subject: Re: new aic7xxx bug, 2.4.13/6.2.4
Date: Fri, 02 Nov 2001 07:55:30 -0700	[thread overview]
Message-ID: <200111021455.fA2EtVY46425@aslan.scsiguy.com> (raw)
In-Reply-To: Your message of "02 Nov 2001 15:03:09 +0100." <m3n1259so2.fsf@defiant.pm.waw.pl>

>Jason Lunz <j@falooley.org> writes:
>
>> I'm having troubles with the last few revisions of the new Gibbs aic7xxx
>> driver that until now has served me well. This is kernel
>> 2.4.13-preempt-lvmrc4 with the 6.2.4 scsi driver, but I saw it happen on
>> 2.4.12 with 6.2.1. I haven't tried older versions with this CD.
>
>Something similar here, but I have disks on this SCSI :-(

Actually not that similar.

>scsi0:A:0: parity error detected in Message-in phase. SEQADDR(0x1b6) SCSIRATE(
>0x0) 

Here's your first clue.  Parity errors occur on "broken" SCSI busses.
Your cabling or terminator is bad, you have a bent pin, or just too
much noise in and around your cabling.

>scsi0: Unexpected busfree in Message-out phase 
>SEQADDR == 0x166 

When we tried to tell the target about the parity error during message-in
phase, the target was unable to see our message without parity errors.
So, it had to go bus free to terminate the transaction.  It is a bug
here that we didn't properly identify the active transaction and abort
it here due to some overly protective code, but that can be readily
fixed.

>scsi0:0:0:0: Attempting to queue an ABORT message 

Because we failed to abort the command above, the mid-layer
timesout the command and we start recovery.

>scsi0: Dumping Card State while idle, at SEQADDR 0x7 
>ACCUM = 0x33, SINDEX = 0x7, DINDEX = 0x8c, ARG_2 = 0x0 
>HCNT = 0x0 
>SCSISEQ = 0x12, SBLKCTL = 0x2 
> DFCNTRL = 0x0, DFSTATUS = 0x29 
>LASTPHASE = 0x1, SCSISIGI = 0x0, SXFRCTL0 = 0x80 
>SSTAT0 = 0x5, SSTAT1 = 0xa 
>STACK == 0x3, 0x105, 0x160, 0xe4 
>SCB count = 20 
>Kernel NEXTQSCB = 19 
>Card NEXTQSCB = 19 
>QINFIFO entries:  
>Waiting Queue entries:  
>Disconnected Queue entries: 7:5  
>QOUTFIFO entries:  
>Sequencer Free SCB List: 8 11 4 5 10 15 2 3 1 0 9 13 14 6 12  
>Pending list: 5 
>Kernel Free SCB list: 7 12 0 4 10 6 9 14 3 1 8 11 15 13 2 18 17 16  
>DevQ(0:0:0): 0 waiting 
>DevQ(0:1:0): 0 waiting 
>(scsi0:A:0:0): Queuing a recovery SCB 
>scsi0:0:0:0: Device is disconnected, re-queuing SCB 
>Recovery code sleeping 
>(scsi0:A:0:0): Abort Tag Message Sent 
>(scsi0:A:0:0): SCB 5 - Abort Tag Completed. 
>Recovery SCB completes 
>Recovery code awake 
>aic7xxx_abort returns 0x2002 

And we abort the transaction successfully.

>scsi0:A:0: parity error detected in Message-in phase. SEQADDR(0x1b6) SCSIRATE(
>0x0) 
>Kernel panic: HOST_MSG_LOOP with invalid SCB ff 

We get another parity error in a state without a proper connection (probably
on the identify message after a reselection) and some more defensive
programming prevents us from handling it correctly. 8-(

The end result is that you need to fix your SCSI bus to be more reliable.
I will fix the issues you've discovered in parity error and unexpected
busfree handling in the next driver release.  If your bus is working
properly, you shouldn't see these errors.

--
Justin

  reply	other threads:[~2001-11-02 14:56 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2001-11-02  3:24 Jason Lunz
2001-11-02 14:03 ` Krzysztof Halasa
2001-11-02 14:55   ` Justin T. Gibbs [this message]
2001-11-02 14:43 ` Justin T. Gibbs
2001-11-02 19:35   ` Jason Lunz
2001-11-02 20:03     ` Justin T. Gibbs
2001-11-02 20:17     ` Richard B. Johnson
2001-11-12 18:45     ` Hans-Peter Jansen
2001-11-12 22:00       ` Jason Lunz

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200111021455.fA2EtVY46425@aslan.scsiguy.com \
    --to=gibbs@scsiguy.com \
    --cc=khc@pm.waw.pl \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®