Bug 696576

Summary: [PV Xen] Mouse stuck after save/restore of PV guest on RHEL5.6
Product: [Fedora] Fedora Reporter: Igor Mammedov <imammedo>
Component: kernelAssignee: Justin M. Forbes <jforbes>
Status: CLOSED RAWHIDE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 15CC: berrange, ehabkost, extras-orphan, gansalmon, itamar, jonathan, kernel-maint, madhu.chinakonda, xen-maint
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-06-23 12:18:51 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
Driver on restore should ask xen for request-abs-pointer again if it's available none

Description Igor Mammedov 2011-04-14 10:45:05 UTC
Description of problem:
Mouse stuck after restore of PV guest but buttons are in working condition.

cat /proc/interrupts - shows that vkbd receives mouse motion events
but userspace test http://people.freedesktop.org/~whot/evtest.c
is blocked on reading from mouse event device. So events kind of lost
in PV kernel.

Version-Release number of selected component (if applicable):
kernel-2.6.38.2-9.fc15.x86_64

How reproducible:
reproducible always.

Steps to Reproduce:
1. boot PV guest to GDM login screen.
2. save the guest: virsh save guest guest.dump, then restore guest: virsh restore guest.dump.
3. put mouse pointer back to guest's vnc window and try to move mouse.
4. as result mouse doesn't move any more till the guest is rebooted
  
Actual results:
mouse doesn't move any more till the guest is rebooted

Expected results:
mouse move pointer after guest restore.

Additional info:

Comment 1 Igor Mammedov 2011-04-14 10:47:33 UTC
Attempted to run FC13 PV guest on RHEL5.
Mouse works as expected after resume.
So it looks like regression.

Comment 2 Igor Mammedov 2011-04-14 14:32:00 UTC
Created attachment 492126 [details]
Driver on restore should ask xen for request-abs-pointer again if it's available

    If driver has been configured for ABS coordinates at
    start it will get XENKBD_TYPE_POS events and then
    suddenly after restore it'll start getting
    XENKBD_TYPE_MOTION events, that will be dropped later
    and they won't get into user-space.

    Regression was introduced by hunk 5 and 6 of 5ea5254.

    Driver on restore should ask xen for request-abs-pointer
    again if it's available.

Comment 3 Igor Mammedov 2011-04-20 09:20:57 UTC
The upstream git commit is c36b58e8a9112017c2bcc322cc98e71241814303.

Comment 4 Igor Mammedov 2011-04-22 11:43:25 UTC
assigning to justin for integration

Comment 5 Chuck Ebbert 2011-05-03 07:48:25 UTC
This is fixed in 2.6.38.5