Bug 146460

Created attachment 110354 [details]
sourceforge ipmi version 3.3 patch ported to RHEL3 U4

Patch is from Jack Vogel, should apply cleanly against the U4 base kernel,
2.4.21-27.EL.  Patch updates the ipmi driver to version 3.3.

diffstat 1106704617-ipmi-v3.3-patch
 Documentation/Configure.help        |   25
 Documentation/IPMI.txt              |  168 ++
 Makefile                            |    2
 drivers/char/Config.in              |    3
 drivers/char/ipmi/Makefile          |    8
 drivers/char/ipmi/ipmi_bt_sm.c      |  513 +++++++
 drivers/char/ipmi/ipmi_devintf.c    |  172 +-
 drivers/char/ipmi/ipmi_kcs_intf.c   |  393 +++--
 drivers/char/ipmi/ipmi_kcs_sm.c     |  153 +-
 drivers/char/ipmi/ipmi_kcs_sm.h     |   70 -
 drivers/char/ipmi/ipmi_msghandler.c | 1541 ++++++++++++++++++++---
 drivers/char/ipmi/ipmi_poweroff.c   |  543 ++++++++
 drivers/char/ipmi/ipmi_si.c         | 2361
++++++++++++++++++++++++++++++++++++
 drivers/char/ipmi/ipmi_si_sm.h      |  126 +
 drivers/char/ipmi/ipmi_smic_sm.c    |  599 +++++++++
 drivers/char/ipmi/ipmi_watchdog.c   |   27
 include/linux/ipmi.h                |  147 ++
 include/linux/ipmi_msgdefs.h        |   57
 include/linux/ipmi_smi.h            |   18
 kernel/panic.c                      |    2
 20 files changed, 6398 insertions(+), 530 deletions(-)

Patch has been tested by IBM on x343, e325, x336, x366, HS20, and x346. 

Has the patch been tested at any customer sites?  If we generate a
hotfix, or test, kernel are there any customers interested in testing
the impi driver?

IBM Crit Sit manager has a meeting scheduled for 4pm (Eastern) where
they will propose testing hotfix test kernel to customer--should have
answer to question in comment #12 at that time, if not before.

IBM customer has indicated that they would gladly take the hot fix. 
If Red Hat will branch CVS and build src and bin kernel rpms and post
externally, IBM will deploy at customer site asap.

Will review the updated patch and integrate with the patch used to
generate the test kernels.

Waiting on final test results.

Patch has been tested and verified to work on all systems that were previously 
not working, except for the IBM e325. A recent change in the BIOS has caused a 
potential regression. We are currently discussing with the BIOS guys. It 
appears as though the changes made in the BIOS are compliant with the IPMI 
specification. However, this still needs further analysis.

Chris any update with your dealings with the BIOS group regarding the e325 system?  

PM ACK for U6

Peter, cc'ed Jack Vogel, so he can respond to your question, as he has been
working with the e325 BIOS guys and I have been on vacation, so I'm somewhat out
of touch here. 

BTW, time has passed, there have been a couple subsequent fixes to the
driver that are important enough to pick up, so when the time comes to
take code can we try and do that?

As long as we rerun all of the tests it should not be a problem.  

Updated request for v35 of the open ipmi driver.

Jack - what is the determination on the e325 BIOS issue?  Move ahead with the
v35 version of the driver?  Let me know so I can get some test kernels created
for you, (and other sites). 

Working on the v35 version for inclusion in RHEL3 U6.

Patch posted for inclusion in RHEL3 U6.

The OpenIPMI driver update has just been committed to the RHEL3 U6
patch pool this evening (in kernel version 2.4.21-33.EL).

Bugzilla to track fix per comment #45 is Bug 165467.

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 the 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.

http://rhn.redhat.com/errata/RHSA-2005-663.html