Well, now we know that the workaround Dell gave us is truly a resolution too.
I hope this helps everyone.

-------- Original Message --------
Subject: Network issue on servers containing BCM 5709 / 5716 cards resolved -
Date:   Tue, 25 May 2010 23:10:44 +0530
From:   <[email protected]>
To:     <[email protected]>


Hello,

Connectivity loss was observed under some heavy traffic conditions on
RHEL 5.3/5.4/5.5 on servers containing BCM 5709/5716 cards.

Work around:

The issue is MSI-X specific. So passing "disable_msi=1" to the bnx2
driver would prevent the issue from happening.

Root cause details and fix:

The card was dropping the PCI writes issued to clear the mask bit in the
MSI-X vector control field as part of MSI-X handling. As a result, the
vector was lost and all the traffic associated with this vector would be
dropped.

This is fixed by increasing the GRC timeout value for the PCI
read/writes by setting a bit in the card's register.

Upstream fix:

Please find the upstream fix to this issue at this link -
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commi
t;h=c441b8d2cb2194b05550a558d6d95d8944e56a84

Kbase Link - http://kbase.redhat.com/faq/docs/DOC-26837 (Does not
contain the details about the fix)

RHEL update containing this fix -

The RHEL 5.5 errata kernel,  kernel-2.6.18-194.3.1.el5.x86_64.rpm
contains the fix for this issue.


With regards,
Narendra K
Linux Engineering
_______________________________________________
Linux-PowerEdge mailing list
[email protected]
https://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq

--
  <<  MCT>>    Michael C Tiernan.   xmpp:[email protected]
  MIT - Laboratory for Nuclear Science - http://www.lns.mit.edu
  High Perf Research Computing Facility at The Bates Linear Accelerator
  "Bit-smashing your bits better than anyone can!"


Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to