+++ This bug was initially created as a clone of Bug #568601 +++
As requested by John, this BZ separates the AER feature from the general update of the bnx2 driver in bug 568601.
John has proposed a patch for the AER support. Broadcom QA is currently setting up an environment to test using aer-inject.
Created attachment 437288 [details]
[PATCH] bnx2: Add PCI Advanced Error Reporting.
This is the upstream patch we plan to send once net-next opens. It is largely based on John's original patch.
Michael, exactly when do you plan on submitting this?
Dave Miller is not accepting any new features right now until net-next-2.6 opens. It should open in about a week. I'll be on vacation soon and will ask a co-worker to send the patch for me. Thanks.
NOTE: this may not make RHEL 5.6 due to the development window for 5.6 Beta already being closed.
John, please also include this patch:
bnx2: Enable AER on PCIE devices only
in addition to the 1st patch:
bnx2: Add PCI Advanced Error Reporting support.
I encorporated both AER commits provided in comment #5 in a kernel build. See http://people.redhat.com/jfeeney/.rhel5.6/kernel-2.6.18-222.el5.bz617024.1*
There is a tool that can be used to twiddle this driver to inject the errors.
I would suggest downloading the code and reviewing
I noticed that you set the verified field today. Does this mean that you successfully tested the kernel I provided in comment #6?
Just making sure.
Thanks if you did.
[root@pilot ~]# uname -a
Linux pilot.hydra.ios 2.6.18-222.el5.bz617024.1 #1 SMP Wed Oct 13 16:36:45 EDT 2010 x86_64 x86_64 x86_64 GNU/Linux
I have been asked for the details of the testing you performed. Would it be possible to provide them here?
We have a script that injects each of the types of errors to a specific bus and function. We did this with the bnx2 device and then reviewed messages to verify:
1) The correct error status bits were being set.
2) No additional error status bits were being set between tests.
3) The link was coming back up after each correctable error.
You can download this test kernel (or newer) from http://people.redhat.com/jwilson/el5
Detailed testing feedback is always welcomed.
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.