Bug 1025477

Summary:	VFIO MSI affinity
Product:	Red Hat Enterprise Linux 7	Reporter:	Alex Williamson <alex.williamson>
Component:	qemu-kvm	Assignee:	Alex Williamson <alex.williamson>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Virtualization Bugs <virt-bugs>
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	7.0	CC:	acathrow, bdas, chayang, hhuang, huding, juzhang, knoel, michen, virt-maint, xuhan
Target Milestone:	rc
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	qemu-kvm-1.5.3-17.el7	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2014-06-13 09:49:42 UTC	Type:	Bug
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:
Bug Blocks:	896690, 969793

Description Alex Williamson 2013-10-31 19:06:52 UTC

Description of problem:
VFIO is missing affinity support for MSI (MSI-X is already present)

Version-Release number of selected component (if applicable):


How reproducible:
100%

Steps to Reproduce:
1. Attempt to change IRQ affinity of a vfio-pci assigned device using MSI interrupts
2.
3.

Actual results:
interrupted vcpu unchanged

Expected results:
interrupted vcpu tracks user request

Additional info:

Comment 2 Miroslav Rezanina 2013-11-07 15:56:20 UTC

Fix included in qemu-kvm-1.5.3-17.el7

Comment 4 Xu Han 2013-11-18 06:51:22 UTC

Reproduce this bug with components:
qemu-kvm-1.5.3-16.el7.x86_64

Steps:
1. boot guest with vfio-pci assigned device using MSI interrupts(82579LM NIC).
/usr/libexec/qemu-kvm -nodefaults -M q35 -m 1G -cpu SandyBridge -smp 4,cores=2,threads=2,sockets=1 -monitor stdio -boot menu=on -rtc base=localtime,clock=host,driftfix=slew -qmp tcp:0:5555,server,nowait -vga qxl -drive file=/home/rhel7base.qcow2,format=qcow2,id=guest-img,if=none,werror=stop,rerror=stop -device virtio-blk-pci,scsi=off,drive=guest-img,id=os-disk -spice disable-ticketing,port=5931 \
-device vfio-pci,host=00:19.0,id=hostnet

2. get IRQ num and affinity of vfio-pci assigned device.
# cat /proc/interrupts | grep ens3
# cat /proc/irq/43/smp_affinity

3. change IRQ affinity of that device.
# echo 2 > /proc/irq/43/smp_affinity

Results:
After step1, check msi information on guest.
# lspci -vvv -s 00:03.0 | grep -i msi
	Capabilities: [d0] MSI: Enable+ Count=1/1 Maskable- 64bit+

After step2
# cat /proc/interrupts | grep ens3
 43:      50821          0          0          0   PCI-MSI-edge      ens3
# cat /proc/irq/43/smp_affinity
4

After step3, interrupted vcpu unchanged.
# cat /proc/irq/43/smp_affinity
4
# cat /proc/interrupts | grep ens3
 43:      51947          0          0          0   PCI-MSI-edge      ens3


Verify this bug with component:
qemu-kvm-1.5.3-19.el7.x86_64

Same steps as above.

Results:
After step2
# cat /proc/interrupts | grep ens3
 43:      43186          0        453          0   PCI-MSI-edge      ens3
# cat /proc/irq/43/smp_affinity
1

After step3, interrupted vcpu tracks user request.
# cat /proc/irq/43/smp_affinity
2
# cat /proc/interrupts | grep ens3
 43:      44167       2881        453          0   PCI-MSI-edge      ens3


Base on these test results above, this bug has been fixed.

Comment 6 Alex Williamson 2013-12-04 16:18:53 UTC

*** Bug 1004373 has been marked as a duplicate of this bug. ***

Comment 7 Ludek Smid 2014-06-13 09:49:42 UTC

This request was resolved in Red Hat Enterprise Linux 7.0.

Contact your manager or support representative in case you have further questions about the request.