<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<TITLE>RE: ofw Digest, Vol 11, Issue 28</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<BR>
<P><FONT SIZE=2>Q<BR>
<BR>
Ken Dieter<BR>
kedieter@cisco.com<BR>
630-701-5809<BR>
(sent from handheld PDA)<BR>
<BR>
-----Original Message-----<BR>
From: ofw-request@lists.openfabrics.org [<A HREF="mailto:ofw-request@lists.openfabrics.org">mailto:ofw-request@lists.openfabrics.org</A>]<BR>
Sent: Tuesday, December 25, 2007 03:00 PM Eastern Standard Time<BR>
To: ofw@lists.openfabrics.org<BR>
Subject: ofw Digest, Vol 11, Issue 28<BR>
<BR>
Send ofw mailing list submissions to<BR>
ofw@lists.openfabrics.org<BR>
<BR>
To subscribe or unsubscribe via the World Wide Web, visit<BR>
<A HREF="http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw">http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw</A><BR>
or, via email, send a message with subject or body 'help' to<BR>
ofw-request@lists.openfabrics.org<BR>
<BR>
You can reach the person managing the list at<BR>
ofw-owner@lists.openfabrics.org<BR>
<BR>
When replying, please edit your Subject line so it is more specific<BR>
than "Re: Contents of ofw digest..."<BR>
<BR>
<BR>
Today's Topics:<BR>
<BR>
1. RE: SRP bug? (Bill Boas)<BR>
<BR>
<BR>
----------------------------------------------------------------------<BR>
<BR>
Message: 1<BR>
Date: Mon, 24 Dec 2007 14:06:17 -0800<BR>
From: "Bill Boas" <bboas@systemfabricworks.com><BR>
Subject: RE: [ofw] SRP bug?<BR>
To: "'Randy Kreiser'" <rkreiser@datadirectnet.com>, "'Leonid Keller'"<BR>
<leonid@mellanox.co.il>, <rob@systemfabricworks.com>,<BR>
<ofw@lists.openfabrics.org><BR>
Message-ID: <008501c84679$35d15ec0$6401a8c0@YOURCB10AA3FFD><BR>
Content-Type: text/plain; charset="us-ascii"<BR>
<BR>
Randy, thanks for responding to Leonid's questions.<BR>
<BR>
Leonid,<BR>
<BR>
The US Gov customers are anxious to learn if we are making progress<BR>
understanding the cause(s) of this bug.<BR>
<BR>
Do you have access to suitable hardware and software in the Mellanox<BR>
facilities where you are - (Yokneam)? To duplicate this bug and run further<BR>
tests to diagnose the root causes?<BR>
<BR>
Bill.<BR>
<BR>
Bill Boas<BR>
VP, Business Development<BR>
System Fabric Works<BR>
510-375-8840<BR>
bboas@systemfabricworks.com<BR>
www.systemfabricworks.com<BR>
<BR>
<BR>
-----Original Message-----<BR>
From: ofw-bounces@lists.openfabrics.org<BR>
[<A HREF="mailto:ofw-bounces@lists.openfabrics.org">mailto:ofw-bounces@lists.openfabrics.org</A>] On Behalf Of Randy Kreiser<BR>
Sent: Monday, December 24, 2007 10:39 AM<BR>
To: 'Leonid Keller'; rob@systemfabricworks.com; ofw@lists.openfabrics.org<BR>
Subject: RE: [ofw] SRP bug?<BR>
<BR>
HI Leonid, answers are below!<BR>
<BR>
Randy<BR>
<BR>
-----Original Message-----<BR>
From: Leonid Keller [<A HREF="mailto:leonid@mellanox.co.il">mailto:leonid@mellanox.co.il</A>]<BR>
Sent: Sunday, December 23, 2007 3:53 AM<BR>
To: Randy Kreiser; rob@systemfabricworks.com; ofw@lists.openfabrics.org<BR>
Subject: RE: [ofw] SRP bug?<BR>
<BR>
Hi Randy,<BR>
<BR>
Thank you for the reply.<BR>
A bit more questions:<BR>
Does it fail with ModeFlags=0 (a default value) ?<BR>
<BR>
A) Yes, it fails with settings of 0,1 and 3<BR>
<BR>
What SW run on the target side ?<BR>
<BR>
A) We are running a target driver DDN version 3.08<BR>
<BR>
What kind of device is that appliance (from SRP point of view) ?<BR>
<BR>
A) Dumb block device (RAID controller with 8 luns).<BR>
<BR>
Does data transfer work in raw mode (without formatting) ?<BR>
<BR>
A) Yes, we setup a CXFS client running windows and it reads and<BR>
writes until your heart is content!<BR>
<BR>
(you can check that with Iometer)<BR>
TIA<BR>
<BR>
Leonid<BR>
<BR>
> -----Original Message-----<BR>
> From: Randy Kreiser [<A HREF="mailto:rkreiser@datadirectnet.com">mailto:rkreiser@datadirectnet.com</A>]<BR>
> Sent: Friday, December 21, 2007 4:53 PM<BR>
> To: Leonid Keller; rob@systemfabricworks.com;<BR>
> ofw@lists.openfabrics.org<BR>
> Subject: RE: [ofw] SRP bug?<BR>
><BR>
> Leonid, set the register you wanted to a "1" and it fails much quicker<BR>
> but that was the only change I saw as it still fails the format.<BR>
><BR>
> Randy<BR>
><BR>
> -----Original Message-----<BR>
> From: Leonid Keller [<A HREF="mailto:leonid@mellanox.co.il">mailto:leonid@mellanox.co.il</A>]<BR>
> Sent: Thursday, December 20, 2007 4:50 AM<BR>
> To: rob@systemfabricworks.com; ofw@lists.openfabrics.org<BR>
> Cc: Randy Kreiser<BR>
> Subject: RE: [ofw] SRP bug?<BR>
><BR>
> Hi Rob,<BR>
><BR>
> Thank you for the elaborate analysis. It seems right.<BR>
> I'd like to get some more information, maybe you or someone else can<BR>
> help.<BR>
><BR>
> Did this trace come from an IB sniffer ?<BR>
> (Otherwise we can't be sure that the corruption happens at Initator's<BR>
> side.)<BR>
><BR>
> How often it happens ?<BR>
><BR>
> How can one reproduce it ?<BR>
><BR>
> What SRP target is being used ?<BR>
><BR>
> Could we ask (and whom) to perform experiments ?<BR>
> For example, I'd suggest to set ModeFlags to 1 in<BR>
> HKLM\SYSTEM\CurrentControlSet\Services\ibsrp\parameters,<BR>
> restart SRP driver and rerun the test.<BR>
><BR>
> Leonid<BR>
><BR>
><BR>
><BR>
> > -----Original Message-----<BR>
> > From: ofw-bounces@lists.openfabrics.org<BR>
> > [<A HREF="mailto:ofw-bounces@lists.openfabrics.org">mailto:ofw-bounces@lists.openfabrics.org</A>] On Behalf Of Robert H.B.<BR>
> > Netzer<BR>
> > Sent: Tuesday, December 18, 2007 8:38 PM<BR>
> > To: ofw@lists.openfabrics.org<BR>
> > Cc: 'Randy Kreiser'<BR>
> > Subject: [ofw] SRP bug?<BR>
> ><BR>
> > I have recently been shown a trace of an SRP session<BR>
> between the WinOF<BR>
> > 1.0.1 SRP initiator and a DDN S2A9550 storage appliance<BR>
> that has the<BR>
> > following suspicious SRP_CMD. It seems to contain a bad virtual<BR>
> > address. Here is the payload of the send from the initiator to the<BR>
> > appliance (this is a few hundred cmds into the stream):<BR>
> ><BR>
> > 02000000 00200100 EF010000 00000000<BR>
> > 00000000 00000000 00000000 00000000<BR>
> > 2A000064 00220000 20000000 00000000<BR>
> > 00000000 05A83364 A8002201 00000010<BR>
> > 00004000 03006209 04006309 AA002301<BR>
> > 00004000<BR>
> ><BR>
> > Consulting the SRP and SCSI specs and decoding this:<BR>
> ><BR>
> > The first row indicates that it's an SRP_CMD, that there is one<BR>
> > data-out buffer descriptor, and that it's an "indirect data buffer<BR>
> > descriptor" (type 2h, encoded in the high nibble of the sixth byte<BR>
> > above).<BR>
> ><BR>
> > The SCSI CBD starts in the third row and is a write<BR>
> (10-byte CDB). The<BR>
> > length is 20h blocks (16k bytes).<BR>
> ><BR>
> > The data-out buffer descriptor starts at byte 48 (fourth row) and<BR>
> > consists of a 16-byte "indirect table memory descriptor", a<BR>
> four-byte<BR>
> > total length (00004000), and one 16-byte "partial memory<BR>
> descriptor"<BR>
> > (there is one of these because the data-out buffer<BR>
> descriptor count,<BR>
> > the 7th byte in the SRP_CMD, is 1).<BR>
> ><BR>
> > The suspicious part is the partial memory descriptor, which is this<BR>
> > (copying the last four words from above): 03006209<BR>
> > 04006309 AA002301 00004000. This is a virtual address of<BR>
> > 03006209 04006309, a memory handle (AA002301) that looks like the<BR>
> > other ones earlier in the trace, and a data length of 16k.<BR>
> ><BR>
> > The SRP stream gets into trouble when the target does an RDMA Read<BR>
> > Request using this virtual address -- it looks bogus.<BR>
> ><BR>
> > I'm hoping that someone can double-check my decoding of<BR>
> this packet,<BR>
> > and perhaps Tzachi could take a look.<BR>
> ><BR>
> > Rob Netzer<BR>
> > System Fabric Works, Inc.<BR>
> ><BR>
> ><BR>
> > _______________________________________________<BR>
> > ofw mailing list<BR>
> > ofw@lists.openfabrics.org<BR>
> > <A HREF="http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw">http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw</A><BR>
> ><BR>
><BR>
><BR>
><BR>
><BR>
<BR>
<BR>
<BR>
_______________________________________________<BR>
ofw mailing list<BR>
ofw@lists.openfabrics.org<BR>
<A HREF="http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw">http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw</A><BR>
<BR>
<BR>
------------------------------<BR>
<BR>
_______________________________________________<BR>
ofw mailing list<BR>
ofw@lists.openfabrics.org<BR>
<A HREF="http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw">http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw</A><BR>
<BR>
End of ofw Digest, Vol 11, Issue 28<BR>
***********************************<BR>
</FONT>
</P>
</BODY>
</HTML>