Customers have noted a performance issue with reads on the QLGE driver for PPC machines. Development and testing of a patch is underway.
*** Bug 516051 has been marked as a duplicate of this bug. ***
Created attachment 357863 [details] qlge read perf and reboot hangs
IBM requests Red Hat consider inclusion of this patch for RHEL5.4 z-stream release due to significant read performance improvement on IBM HV4 PowerPc platforms. With the current 1.00.00.20 qlge driver read performance is 4.23Gbps, with this patch applied, read performance increases to 7.62 Gbps, an 80% increase. Testing was done using 4 netperf sessions running TCP_STREAM using a 64K message size with a 256K socket buffer size.
in kernel-2.6.18-165.el5 You can download this test kernel from http://people.redhat.com/dzickus/el5 Please do NOT transition this bugzilla state to VERIFIED until our QE team has sent specific instructions indicating when to do so. However feel free to provide a comment indicating that this fix has been verified.
Verified this patch is in -165 kernel and resolves the hang and performance issues.
------- Comment From sharyath.com 2009-11-17 11:55 EDT------- I am not sure why the Red Hat bugzilla information or not mirrored to this bugzilla. The patch is tested and found to fix the performance issue Moreover the patch is included in 5.4 kernel-2.6.18-165.el5 as requested by the IBM team So accordingly closing this bug Please do reopen in case further investigation is required
------- Comment From lxie.com 2010-01-26 14:36 EDT------- Luciano, Richard, Can you guys confirm if the rhel5.5 Alpha has fixed this bug? Thanks, Linda
------- Comment From lnx1138.ibm.com 2010-01-28 12:42 EDT------- Hi Linda, I believe Breno can test and confirm if it is fixed in the 5.5 Alpha.
------- Comment From rlary.com 2010-01-28 16:04 EDT------- (In reply to comment #12) > Hi Linda, I believe Breno can test and confirm if it is fixed in the 5.5 Alpha. Testing on RHEL5.5 alpha is blocked by Bug 59338 - RIT397463: manual install exited abnormally as loader received SIGSEGV
------- Comment From lxie.com 2010-02-09 10:52 EDT------- (In reply to comment #13) > (In reply to comment #12) > > Hi Linda, I believe Breno can test and confirm if it is fixed in the 5.5 Alpha. > > Testing on RHEL5.5 alpha is blocked by Bug 59338 - RIT397463: manual install > exited abnormally as loader received SIGSEGV the latest drop, beta which was released yesterday, has a fix for 59338.
------- Comment From lxie.com 2010-02-16 02:43 EDT------- (In reply to comment #13) > (In reply to comment #12) > > Hi Linda, I believe Breno can test and confirm if it is fixed in the 5.5 Alpha. > > Testing on RHEL5.5 alpha is blocked by Bug 59338 - RIT397463: manual install > exited abnormally as loader received SIGSEGV Richard, Bug 59338 was verified with beta, it's closed. Can you please test beta and close this bug if it's fixed? Thank you very much for your support. - Linda
------- Comment From leitao.ibm.com 2010-02-24 11:20 EDT------- I just tested this on Snap1 and the performance seems to be very fine, although I didn't test with Jumbo frames yet. I will consider this bug as closed for now. As soon as I test using jumbo frames and if I face any issue, I will reopen this bug. Thanks, Breno
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. http://rhn.redhat.com/errata/RHSA-2010-0178.html