Description of problem: When doing IO larger than 32kb on raw or ocfs, the kernel splits it up into 32kb size chunks and submits them seperately. there is no attempt made to merge in the IO's. this is a performance hit Version-Release number of selected component (if applicable): 2.4.21-27 How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Created attachment 111115 [details] patch to merge IOs patch on top of 2.4.21-27.0.2
The attached patch attempts to fix this.
We have already committed larger superbh support as part of bug #131391. Is there really demand for merging too? That will require an entirely separate set of testing, and there is no way it will make U5.
Any word on this?
With the modified superbh support, this should matter less. But we haven't been able to get some comparison data yet.
The patch referred to in bug 131391 has been released as part of U5. Closing this bug for inactivity; please reopen if you want this issue pursued.
The fix referred to by Stephen was committed to the RHEL3 U5 patch pool on 23-Feb-2005 (in kernel version 2.4.21-28.EL). Following is the U5 advisory message from the Errata System: 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-2005-294.html