Bug 741139

Summary:	spice mode - display freezes
Product:	[Fedora] Fedora	Reporter:	Richard Keech <rkeech>
Component:	virt-manager	Assignee:	Cole Robinson <crobinso>
Status:	CLOSED WORKSFORME	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	high	Docs Contact:
Priority:	unspecified
Version:	15	CC:	berrange, cfeller, crobinso, dougsland, dpierce, erik-fedora, hbrock, jforbes, nobody, virt-maint
Target Milestone:	---
Target Release:	---
Hardware:	Unspecified
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2012-01-25 21:51:30 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:

Description Richard Keech 2011-09-26 00:30:45 UTC

Description of problem:

The virt-manager display window completely freezes frequently.  Re-starting virt-manager is required. Spicec does not have the same problem.  This is with a WinXP guest using QXL driver.  Host is 64-bit F15  (fully updated).  Guest is 32-bit WinXP.

Aside from the freeze problem, the graphics performance is excellent.


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

virt-manager 0.8.7-6


How reproducible:

Depending on the Windows app running, I generally cannot get more than ten minutes of operation without the display freezing.  Re-starting virt-manager (or using spicec) is able to re-establish desktop to a working state.


Steps to Reproduce:
1.  start windows guest.
2.  connect using virt-manager
3.  start using windows app
4.  wait a short while
  
Actual results:

Display within the virt-manager window freezes.  

Expected results:


Additional info:

Freeze only ever happens when app is being actively used.  I've never seen the freeze when running MS standard desktop components like file manager, ie it only ever affects 3rd-party apps.  Apps affected include 3d and non-3d apps.   

QXL driver version 1.4.1.0 setup with a 32-bit frame buffer.  Full hardware acceleration and write combining are selected.

WindowsXP is fully updated.

Problem occurs at multiple display resolutions.  I'm generally using 1600x900

Comment 1 Cole Robinson 2011-09-26 14:49:33 UTC

Thanks for the report. Please run the following commands:

yum install yum-utils gdb
debuginfo-install virt-manager
(run virt-manager, reproduce the freeze, keep the app running)
ps axwww | grep virt-manager | grep -v grep
(the first number that command shows is virt-manager's pid)
gdp -p $pid
thread apply all bt
(copy that output to this bug report)

Comment 2 Richard Keech 2011-09-27 11:32:15 UTC

Virt-manager doesn't have a debuginfo bc it's not the executable - it's running on top of python.  An initial backtrace shows that it's hung in poll() which is called from libvirt.  I've installed debuginfo for libvirt, python and glib2, and this is what I get:

Thread 2 (Thread 0x7fef21764700 (LWP 29778)):
#0  0x000000342ecd7423 in __GI___poll (fds=<optimized out>, nfds=<optimized out>, timeout=<optimized out>)
    at ../sysdeps/unix/sysv/linux/poll.c:87
#1  0x0000003430c42d24 in g_main_context_poll (n_fds=1, fds=0x7fef1c001150, priority=<optimized out>, timeout=-1, 
    context=0x2b2dce0) at gmain.c:3405
#2  g_main_context_iterate (context=0x2b2dce0, block=<optimized out>, dispatch=1, self=<optimized out>) at gmain.c:3087
#3  0x0000003430c4360d in g_main_loop_run (loop=0x23d0ef0) at gmain.c:3300
#4  0x00007fef28dc1564 in gdbus_shared_thread_func (data=<optimized out>) at gdbusprivate.c:276
#5  0x0000003430c683a6 in g_thread_create_proxy (data=0x2b2ddd0) at gthread.c:1955
#6  0x000000342f407b31 in start_thread (arg=0x7fef21764700) at pthread_create.c:305
#7  0x000000342ecdfd2d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:115

Thread 1 (Thread 0x7fef3235c720 (LWP 29777)):
#0  0x000000342ecd7423 in __GI___poll (fds=<optimized out>, nfds=<optimized out>, timeout=<optimized out>)
    at ../sysdeps/unix/sysv/linux/poll.c:87
#1  0x0000003430c42d24 in g_main_context_poll (n_fds=16, fds=0x376b780, priority=<optimized out>, timeout=537, context=
    0x239d570) at gmain.c:3405
#2  g_main_context_iterate (context=0x239d570, block=<optimized out>, dispatch=1, self=<optimized out>) at gmain.c:3087
#3  0x0000003430c4360d in g_main_loop_run (loop=0x2e63780) at gmain.c:3300
#4  0x000000343914c007 in gtk_main () from /usr/lib64/libgtk-x11-2.0.so.0
#5  0x00007fef291dffa6 in ?? () from /usr/lib64/python2.7/site-packages/gtk-2.0/gtk/_gtk.so
#6  0x00000034348df592 in call_function (oparg=<optimized out>, pp_stack=0x7ffffc08bc68)
    at /usr/src/debug/Python-2.7.1/Python/ceval.c:4056
#7  PyEval_EvalFrameEx (f=<optimized out>, throwflag=<optimized out>) at /usr/src/debug/Python-2.7.1/Python/ceval.c:2722
#8  0x00000034348e0098 in fast_function (nk=<optimized out>, na=0, n=<optimized out>, pp_stack=0x7ffffc08bda8, func=
    <function at remote 0x1fe6ed8>) at /usr/src/debug/Python-2.7.1/Python/ceval.c:4158
#9  call_function (oparg=<optimized out>, pp_stack=0x7ffffc08bda8) at /usr/src/debug/Python-2.7.1/Python/ceval.c:4093
#10 PyEval_EvalFrameEx (f=<optimized out>, throwflag=<optimized out>) at /usr/src/debug/Python-2.7.1/Python/ceval.c:2722

Comment 3 Richard Keech 2011-10-18 23:08:35 UTC

Note, I haven't seen this fault for a couple of weeks now.  Could it be that recent errata have fixed this problem?

Comment 4 Cole Robinson 2012-01-25 21:51:30 UTC

Probably, closing as WORKSFORME. Thanks for all the info though