Bug 604159

Summary: virtio-blk: Avoid zeroing every request structure
Product: Red Hat Enterprise Linux 5 Reporter: RHEL Program Management <pm-rhel>
Component: kvmAssignee: Virtualization Maintenance <virt-maint>
Status: CLOSED ERRATA QA Contact: Virtualization Bugs <virt-bugs>
Severity: medium Docs Contact:
Priority: high    
Version: 5.6CC: cpelland, ehabkost, gyue, Jes.Sorensen, llim, mjenner, mkenneth, pm-eus, tburke, virt-maint, ykaul
Target Milestone: rcKeywords: ZStream
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: kvm-83-164.el5_5.13 Doc Type: Bug Fix
Doc Text:
The virtio-blk back end for qemu-kvm performed unnecessary zeroing of memory on every I/O request, which reduced virtual guest performance. With this update, the virtio-blk back end avoids spurious zeroing of request structures on each I/O request, with the result that guest performance is increased.
Story Points: ---
Clone Of: Environment:
Last Closed: 2010-07-29 05:49:59 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On: 598042    
Bug Blocks:    
Attachments:
Description Flags
test result
none
test result with raw none

Description RHEL Program Management 2010-06-15 14:22:17 UTC
This bug has been copied from bug #598042 and has been proposed
to be backported to 5.5 z-stream (EUS).

Comment 6 Golita Yue 2010-06-24 09:43:30 UTC
Created attachment 426506 [details]
test result

Comment 8 Golita Yue 2010-06-25 11:04:16 UTC
Created attachment 426846 [details]
test result with raw

Comment 9 Golita Yue 2010-07-05 08:48:19 UTC
Tested in kvm-83-164.el5_5.11 & kvm-83-164.el5_5.13.(#comment5,6,7,8)

The performance boost of 7-29% in kvm-83-164.el5_5.13. So this bug fixed.

Comment 13 Eduardo Habkost 2010-07-23 18:14:44 UTC
Technical note added. If any revisions are required, please edit the "Technical Notes" field
accordingly. All revisions will be proofread by the Engineering Content Services team.

New Contents:
Cause: the virtio-blk backend on qemu-kvm was doing unnecessary zeroing of memory on every I/O request.

Consequence: performance impact due to unnecessary memory writes.

Fix: avoid zeroing whole requests structures on every I/O requests.

Consequence: performance boosts on virtio-blk.

Comment 14 Douglas Silas 2010-07-28 15:40:56 UTC
Technical note updated. If any revisions are required, please edit the "Technical Notes" field
accordingly. All revisions will be proofread by the Engineering Content Services team.

Diffed Contents:
@@ -1,7 +1 @@
-Cause: the virtio-blk backend on qemu-kvm was doing unnecessary zeroing of memory on every I/O request.
+The virtio-blk back end for qemu-kvm performed unnecessary zeroing of memory on every I/O request, which reduced virtual guest performance. With this update, the virtio-blk back end avoids spurious zeroing of request structures on each I/O request, with the result that guest performance is increased.-
-Consequence: performance impact due to unnecessary memory writes.
-
-Fix: avoid zeroing whole requests structures on every I/O requests.
-
-Consequence: performance boosts on virtio-blk.

Comment 16 errata-xmlrpc 2010-07-29 05:49:59 UTC
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/RHBA-2010-0536.html