[Xen-bugs] [Bug 1612] x86-64 : Can't start VM when vif set

[bugzilla-daemon@xxxxxxxxxxxxxxxxxxx]

http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1612





------- Comment #7 from joseph.deming@xxxxxxxxxxxxxx  2010-05-21 11:18 -------
yes, i do get that error.  here's a bit more from my end:

[    0.026832] CPU: Physical Processor ID: 0
[    0.026884] CPU: Processor Core ID: 0
[    0.026938] Performance Events: AMD PMU driver.
[    0.027023] ------------[ cut here ]------------
[    0.027085] WARNING: at
/build/buildd-linux-2.6_2.6.32-12-amd64-tnQeYz/linux-2.6-2.6.32/debian/build/source_amd64_xen/arch/x86/xen/enlighten.c:721
ini
t_hw_perf_events+0x32d/0x3cd()
[    0.027189] Hardware name: PowerEdge R805
[    0.027241] Modules linked in:
[    0.027325] Pid: 0, comm: swapper Not tainted 2.6.32-5-xen-amd64 #1
[    0.027382] Call Trace:
[    0.027436]  [<ffffffff81500a0f>] ? init_hw_perf_events+0x32d/0x3cd
[    0.027494]  [<ffffffff81500a0f>] ? init_hw_perf_events+0x32d/0x3cd
[    0.027554]  [<ffffffff8104fc68>] ? warn_slowpath_common+0x77/0xa3
[    0.027612]  [<ffffffff81500a0f>] ? init_hw_perf_events+0x32d/0x3cd
[    0.027673]  [<ffffffff812fd38f>] ? identify_cpu+0x2ff/0x308
[    0.027732]  [<ffffffff8100eb4f>] ? xen_restore_fl_direct_end+0x0/0x1
[    0.027792]  [<ffffffff810e85f5>] ? kmem_cache_alloc+0x8c/0xf0
[    0.027850]  [<ffffffff81500636>] ? identify_boot_cpu+0x15/0x3e
[    0.027907]  [<ffffffff815006bd>] ? check_bugs+0x9/0x2e
[    0.027974]  [<ffffffff814f9cc2>] ? start_kernel+0x3cd/0x3e8
[    0.028001]  [<ffffffff814fbad7>] ? xen_start_kernel+0x649/0x64f
[    0.028001] ---[ end trace a7919e7f17c0a725 ]---
[    0.028001] ... version:                0
[    0.028001] ... bit width:              48
[    0.028001] ... generic registers:      4
[    0.028001] ... value mask:             0000ffffffffffff
[    0.028001] ... max period:             00007fffffffffff
[    0.028001] ... fixed-purpose events:   0
[    0.028001] ... event mask:             000000000000000f

[  155.228843] Event-channel device installed.
[  155.658856]   alloc irq_desc for 3238 on node 0
[  155.658860]   alloc kstat_irqs on node 0
[  155.658927]   alloc irq_desc for 3237 on node 0
[  155.658930]   alloc kstat_irqs on node 0
[  155.659332] XENBUS: Unable to read cpu state
[  155.659456] XENBUS: Unable to read cpu state
[  155.659624] XENBUS: Unable to read cpu state
[  155.659778] XENBUS: Unable to read cpu state
[  155.659922] XENBUS: Unable to read cpu state
[  155.660857] XENBUS: Unable to read cpu state
[  155.661027] XENBUS: Unable to read cpu state
[  155.661183] XENBUS: Unable to read cpu state
[  155.661340] XENBUS: Unable to read cpu state
[  155.661606] XENBUS: Unable to read cpu state
[  155.661757] XENBUS: Unable to read cpu state
[  155.661906] XENBUS: Unable to read cpu state
[  157.156279] Bridge firewalling registered

May 21 13:31:16 kraid udevd-work[522]: device 0x17303e0 has devpath
'/devices/virtual/misc/evtchn'
May 21 13:31:16 kraid udevd-work[522]: LINK 'char/10:57'
/lib/udev/rules.d/50-udev-default.rules:2
May 21 13:31:16 kraid udevd-work[522]: NAME 'xen/evtchn'
/lib/udev/rules.d/50-udev-default.rules:48
May 21 13:31:16 kraid udevd-work[522]: LINK 'xen/evtchn'
/lib/udev/rules.d/xen-backend.rules:8
May 21 13:31:16 kraid udevd-work[522]: kernel-provided name 'evtchn' and NAME=
'xen/evtchn' disagree, please use SYMLINK+= or change the kernel to provid
e the proper name
May 21 13:31:16 kraid udevd-work[522]: created db link (xen/evtchn char/10:57
xen/evtchn)
May 21 13:31:16 kraid udevd-work[522]: creating device node '/dev/xen/evtchn',
devnum=10:57, mode=0660, uid=0, gid=0
May 21 13:31:16 kraid udevd-work[522]: mknod(/dev/xen/evtchn, 020660, (10,57))
May 21 13:31:16 kraid udevd-work[522]: set permissions /dev/xen/evtchn, 020660,
uid=0, gid=0
May 21 13:31:16 kraid udevd-work[522]: creating symlink '/dev/char/10:57' to
'../xen/evtchn'
May 21 13:31:16 kraid udevd-work[522]: creating link '/dev/xen/evtchn' to
'/dev/xen/evtchn'
May 21 13:31:16 kraid udevd-work[522]: found existing node instead of symlink
'/dev/xen/evtchn'
May 21 13:31:16 kraid udevd-work[522]: device node '/dev/xen/evtchn' already
exists, link to '/dev/xen/evtchn' will not overwrite it
May 21 13:31:16 kraid udevd-work[522]: passed -1 bytes to netlink monitor
0x173cb20

 - one line like you mentioned for each of core in my machine (2x6-core amd)
 - the first 'crash' appears immediately after boot, don't have reason to
suspect it's related to this issue directly
 - and, i included the NAME vs. SYMLINK msg that keeps coming out for the
xen/evtchn driver, i also don't have reason to suspect it's related, but i was
looking into udev issues for a bit
 - there's also some apparently outdated code that triggers the 'physdev match:
using --physdev-out in the OUTPUT, FORWARD and POSTROUTING chains for
non-bridged traffic is not supported anymore' warnings almost postively from
the file /etc/xen/scripts/vif-common.sh:76-78

but i feel most of this is superfluous.  i noticed i can trigger the crash
explicitly by calling 'xm network-attach wario vifname=vif1.99' where wario is
a currently running vm that has been created and already caused the crash
during creation for it's existing interfaces, and vif1.99 is a made up name
that is not defined elsewhere in my configs.

i suspect you could create it more cleanly by bringing up a domain with no
vif's in the config file, then calling 'xm network-attach <domain_name>
<vifname=XXX> etc'


-- 
Configure bugmail: 
http://bugzilla.xensource.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

_______________________________________________
Xen-bugs mailing list
Xen-bugs@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-bugs