Fw: [Bugme-new] [Bug 6657] New: Domain validation failures on aic7xxx

Fw: [Bugme-new] [Bug 6657] New: Domain validation failures on aic7xxx

[Andrew Morton]

Apparently a regression.

(I thought the plan was to make scsi_drivers-other@kernel-bugs.osdl.org
auto-forward stuff to linux-scsi?)


Begin forwarded message:

Date: Tue, 6 Jun 2006 08:53:16 -0700
From: bugme-daemon@bugzilla.kernel.org
To: bugme-new@lists.osdl.org
Subject: [Bugme-new] [Bug 6657] New: Domain validation failures on aic7xxx


http://bugzilla.kernel.org/show_bug.cgi?id=6657

           Summary: Domain validation failures on aic7xxx
    Kernel Version: 2.6.16
            Status: NEW
          Severity: high
             Owner: scsi_drivers-other@kernel-bugs.osdl.org
         Submitter: william@brodie-tyrrell.org


Most recent kernel where this bug did not occur: 2.6.10
Distribution: Gentoo
Hardware Environment: x86, AHA19160, WD18310
Software Environment:
Problem Description: 

In 2.6.14 and 2.6.16 I get domain validation errors on my WD18310 connected to a
19160, causing it to drop back to asynchronous rather than 40 or 80MHz wide.  

Domain validation reports parity errors, write buffer failures, performs resets
and generally stuffs around for a few minutes before deciding it will allow async.

I have 2 other discs (Seagate SX118202LS) attached to the same chain which still
work.  My scsi system works perfectly in 2.6.10 with no data corruption.

I have a very similar issue with a dual SYM53C896 in a different machine: it
works in 2.6.10 but produces occasional noise in dmesg:
sym0:9:0:phase change 6-7 11@17cd5f84 resid=6.
On newer (>= 2.6.14) kernels, it completely fails to boot, giving the same sort
of parity errors I'm having with the aic7xxx driver.  When I checked at the time
(a while ago now), there was NO change in the relevant driver between kernel
versions.  The only changes are to the scsi architecture.

This leads me to believe the bug is not in the aic7xxx driver but the new scsi
domain validation code that was overhauled somewhere around 2.6.13-14.

Some of my error messages look a lot like those in #5268.

Steps to reproduce: Use kernel >=2.6.14 with aic7892 or 53C896.

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

Re: Fw: [Bugme-new] [Bug 6657] New: Domain validation failures on aic7xxx

[James Bottomley]

On Tue, 2006-06-06 at 09:18 -0700, Andrew Morton wrote:
> Apparently a regression.
> 
> (I thought the plan was to make scsi_drivers-other@kernel-bugs.osdl.org
> auto-forward stuff to linux-scsi?)

It is ... let me go and give mbligh a swift kick [administered by this
email] to see if I can get him to move on this.

> In 2.6.14 and 2.6.16 I get domain validation errors on my WD18310 connected to a
> 19160, causing it to drop back to asynchronous rather than 40 or 80MHz wide.  
> 
> Domain validation reports parity errors, write buffer failures, performs resets
> and generally stuffs around for a few minutes before deciding it will allow async.
> 
> I have 2 other discs (Seagate SX118202LS) attached to the same chain which still
> work.  My scsi system works perfectly in 2.6.10 with no data corruption.

I'd say that's pretty conclusive. The theory being that the WD doesn't
have a functional echo buffer, or it lies about the size and in either
case we get errors back not because of a syndrome test failure but
because of a problem in the WD echo buffer.

Try this patch for diagnosis only ... what it's doing is disabling the
echo buffer tests globally.  If everything boots up OK (please send
dmesg output so I can see what the echo buffer length is), then we get
to try to figure out if the WD can do any echo buffer tests or if it has
to be globally blacklisted (not fun ... LVD really shouldn't be
configured without the echo buffer test of the transmission line).

James

diff --git a/drivers/scsi/scsi_transport_spi.c b/drivers/scsi/scsi_transport_spi.c
index 780aaed..88bb4ff 100644
--- a/drivers/scsi/scsi_transport_spi.c
+++ b/drivers/scsi/scsi_transport_spi.c
@@ -859,7 +859,9 @@ spi_dv_device_internal(struct scsi_devic
 	if (len == -1 && spi_dt(starget))
 		len = spi_dv_device_get_echo_buffer(sdev, buffer);
 
-	if (len <= 0) {
+	starget_printk(KERN_INFO, starget, "Echo buffer length %d\n", len);
+
+	if (1) {
 		starget_printk(KERN_INFO, starget, "Domain Validation skipping write tests\n");
 		return;
 	}

Re: Fw: [Bugme-new] [Bug 6657] New: Domain validation failures on aic7xxx

[Martin Bligh]

James Bottomley wrote:
> On Tue, 2006-06-06 at 09:18 -0700, Andrew Morton wrote:
> 
>>Apparently a regression.
>>
>>(I thought the plan was to make scsi_drivers-other@kernel-bugs.osdl.org
>>auto-forward stuff to linux-scsi?)
> 
> 
> It is ... let me go and give mbligh a swift kick [administered by this
> email] to see if I can get him to move on this.

It's not going to auto-forward, otherwise every update to bugs owned by
the virtual alias will come through. What I was planning was to cc
linux-scsi on emails to bugme-new (ie one email per new bug only) for
that category only. I'll try to get that going this morning ...

M.

Re: Fw: [Bugme-new] [Bug 6657] New: Domain validation failures on aic7xxx

[James Bottomley]

On Mon, 2006-06-12 at 09:47 -0700, Martin Bligh wrote:
> It's not going to auto-forward, otherwise every update to bugs owned
> by
> the virtual alias will come through. What I was planning was to cc
> linux-scsi on emails to bugme-new (ie one email per new bug only) for
> that category only. I'll try to get that going this morning ...

Yes ... that's what we want.  An email that can be replied to and will
go to the submitter and linux-scsi and be captured by bugzilla.

Thanks,

James

Re: Fw: [Bugme-new] [Bug 6657] New: Domain validation failures on aic7xxx

[Martin Bligh]

James Bottomley wrote:
> On Mon, 2006-06-12 at 09:47 -0700, Martin Bligh wrote:
> 
>>It's not going to auto-forward, otherwise every update to bugs owned
>>by
>>the virtual alias will come through. What I was planning was to cc
>>linux-scsi on emails to bugme-new (ie one email per new bug only) for
>>that category only. I'll try to get that going this morning ...
> 
> 
> Yes ... that's what we want.  An email that can be replied to and will
> go to the submitter and linux-scsi and be captured by bugzilla.

OK, set up. It should reply to bugme-daemon, which should result up
being back in the DB ... all emails should be tagged with [bugme-new]
in them, so its easy to filter if need be.

I cc'ed IO/storage-SCSI, and all the SCSI drivers ones. If that's not
correct, or you have other problems with it, let me know.

Hitting reply-all may result in mail to bugme-new as well as 
bugme-daemon ... a mild pain, but should get caught by the list
filtration, so no big deal, I think.

M.

Re: Fw: [Bugme-new] [Bug 6657] New: Domain validation failures on aic7xxx

[William Brodie-Tyrrell]

Hi,

I've applied your patch and rebuilt the kernel but it'll be a few days
before I can try it since the machine in question is Very Busy(tm)
right now.


thanks.


Stoned koala bears drooled eucalyptus spit in awe as James Bottomley
said:

> I'd say that's pretty conclusive. The theory being that the WD doesn't
> have a functional echo buffer, or it lies about the size and in either
> case we get errors back not because of a syndrome test failure but
> because of a problem in the WD echo buffer.
> 
> Try this patch for diagnosis only ... what it's doing is disabling the
> echo buffer tests globally.  If everything boots up OK (please send
> dmesg output so I can see what the echo buffer length is), then we get
> to try to figure out if the WD can do any echo buffer tests or if it has
> to be globally blacklisted (not fun ... LVD really shouldn't be
> configured without the echo buffer test of the transmission line).


-- 
William Brodie-Tyrrell

Carpe Diem - fish of the day.

<william@brodie-tyrrell.org>
http://www.brodie-tyrrell.org/