[Qemu-devel] two repeatable kqemu-related crashes

[Qemu-devel] two repeatable kqemu-related crashes

[Don Kitchen]

Hi all

I've found two repeatable (possibly related) ways to crash kqemu with 0.9.0
and several earlier versions also I think. It's under linux 2.6.9 fully
updated CentOS 4.4 (clone of RH enterprise linux 4.4)

First, I cannot use kqemu 1.3.0pre11 (either prebuilt or compiled from
source, they're bitwise identical) because as soon as qemu starts, the
window reports guest stopped, and within 3 seconds the host machine
freezes completely flashing caps lock and scroll lock lights. Hard reboot
necessary.

Second problem, so I am using a previous kqemu, pre9 I think. The following
only occurs with kqemu loaded...
 
I'm hosting win2k SP4: under administrative tools I run Computer Management,
and select Disk Management. I either get a qemu crash immediately or after
clicking the first disk. On crash gives output like this:

EAX=00010282 EBX=00c22008 ECX=63010101 EDX=bf42584c
ESI=00c22008 EDI=00c24610 EBP=0006f5d0 ESP=0006f5a8
EIP=77e149a7 EFL=00010246 [---Z-P-] CPL=3 II=0 A20=1 SMM=0 HLT=0
ES =0023 00000000 ffffffff 00cff300
CS =001b 00000000 ffffffff 00cffb00
SS =0023 00000000 ffffffff 00cff300
DS =0023 00000000 ffffffff 00cff300
FS =003b 7ffde000 00000fff 7f40f3fd
GS =0000 00000000 00000000 00000000
LDT=0000 00000000 00000000 00008000
TR =0028 803eb000 000020ab 8000893e
GDT=     80036000 000003ff
IDT=     80036400 000007ff
CR0=e001003b CR2=77f8dfd0 CR3=07a3f000 CR4=00000690
Unsupported return value: 0xffffffff

I hope this information allows for whatever bugs are causing this to be
identified and fixed.

Thanks

Re: [Qemu-devel] two repeatable kqemu-related crashes

[John Morris]

On Fri, 2007-04-13 at 02:31, Don Kitchen wrote:
> Hi all
> 
> I've found two repeatable (possibly related) ways to crash kqemu with 0.9.0
> and several earlier versions also I think. It's under linux 2.6.9 fully
> updated CentOS 4.4 (clone of RH enterprise linux 4.4)

I'm running in a similar setting without a problem.  Kqemu 1.3.0pre11 +
a locally cooked rpm of qemu 0.9.0 based an earlier dag package of
qemu.  I running WBEL 4 instead of CentOS 4, but I really doubt that is
where your problem is.  I'm running on x86_64 and have to admit I
haven't bothered going to the very latest kernel package, still running
kernel-2.6.9-34.EL.  The guest I have been playing with this past week
is Debian 4.0 i386.  Installed and have been running since with
-kernel-kqemu.

The only reliable way I have to nuke the machine is boot FC5 with
-kernel-kqemu and start OO.o.  Without -kernel-kqemu it only kills the
VM.  It is something with FC5 and GCJ, been reported on the list
before.  Btw, OO.o as supplied with Debian 4.0 starts up without a
lockup, will be throwing some documents at it later to see if it holds
up.

-- 
John M.      http://www.beau.org/~jmorris     This post is 100% M$Free!
Geekcode 3.1:GCS C+++ UL++++$ P++ L+++ W++ w--- Y++ b++ 5+++ R tv- e* r