qemu problem

Mark Picone mark.picone at deakin.edu.au
Tue Sep 9 01:13:29 UTC 2008


I am using qemu without any problems on a 7.0-p4 host, here are a few things I 
learnt a lot the way:

1. qemu (with kqemu kernel module loaded) will cause the problems you are having 
(extreme instability).
2. qemu-devel (with kqemu-devel kernel module loaded) will work fine until you 
start swapping memory, then it seems some sort of memory corruption will occur and 
crash the VM.

If you have plenty of RAM give the -devel versions a go, otherwise do not use 
kqemu as it seems to be unstable.

At the moment I am running a WindowsXP-SP3 VM with qemu-devel (albeit very slowly) 
without kqemu or stability problems.

Thanks,


Mark Picone, Trainee Unix Administrator
Information Technology Services Division
Phone:   03 5227 8602   International: +61 3 5227 0806
Fax:     03 5227 8799   International: +61 3 5227 8799
Email:   mark.picone at deakin.edu.au
Website: http://www.deakin.edu.au


> -----Original Message-----
> From: owner-freebsd-ports at freebsd.org [mailto:owner-freebsd-
> ports at freebsd.org] On Behalf Of Ganbold
> Sent: Sunday, 7 September 2008 3:54 PM
> To: Carlos A. M. dos Santos
> Cc: freebsd-ports
> Subject: Re: qemu problem
>
> Carlos A. M. dos Santos wrote:
> > On Sat, Sep 6, 2008 at 9:52 AM, Ganbold <ganbold at micom.mng.net> wrote:
> >
> >> Hi,
> >>
> >> I have problem installing FreeBSD-7.0 using qemu in RELENG_7.
> >> It starts installing FreeBSD, but it crashes and dumps core in different
> >> places.
> >>
> >
> > It would be important to know what "different places" means. It is
> > *during* installation or *after*  it?
> >
> >
>
> Both. It happens when it tries to copy something, or when it tries to
> compile something.
> Yesterday it hanged and crashed when I tried to upgrade 7.0 to CURRENT
> (buildworld).
>
> >> Did somebody experience this before?
> >>
> >> devil# uname -an
> >> FreeBSD devil.micom.mng.net 7.0-STABLE FreeBSD 7.0-STABLE #9: Tue Aug 19
> >> 18:35:02 ULAT 2008
> tsgan at devil.micom.mng.net:/usr/obj/usr/src/sys/DEVIL
> >>  i386
> >> devil#
> >>
> >> devil# qemu -boot d -hda freebsd.img -m 256 -cdrom
> >> ~tsgan/7.0-RELEASE-i386-disc1.iso -localtime -net nic -net tap smp 2
> >> qemu: fatal: triple fault
> >> EAX=c0bfe67c EBX=0000000c ECX=f001003f EDX=f001003f
> >> ESI=c0bfe67c EDI=c24f7c60 EBP=c0bfe670 ESP=c0bfe5e0
> >> EIP=c0a49004 EFL=00000002 [-------] CPL=0 II=0 A20=1 SMM=0 HLT=0
> >> ES =0028 00000000 ffffffff 00cf9300
> >> CS =0020 00000000 ffffffff 00cf9b00
> >> SS =0028 00000000 ffffffff 00cf9300
> >> DS =0028 00000000 ffffffff 00cf9300
> >> FS =0008 ffc00000 ffffffff ffcf93c0
> >> GS =0028 00000000 ffffffff 00cf9300
> >> LDT=0050 c0bfef20 00000087 c000e2bf
> >> TR =0060 c0bff1c0 00000067 c00089bf
> >> GDT=     c0bfe5a0 00000097
> >> IDT=     c0c00240 000007ff
> >> CR0=e005003b CR2=f0010043 CR3=0141e000 CR4=00000690
> >> CCS=c0bfe67c CCD=c0bfe6e8 CCO=ADDB   FCW=127f FSW=0020 [ST=0] FTW=00
> >> MXCSR=00001f80
> >> FPR0=0000000000000000 0000 FPR1=0000000000000000 0000
> >> FPR2=0000000000000000 0000 FPR3=0000000000000000 0000
> >> FPR4=ccccccccccc40000 3ffe FPR5=8000000000000000 3ffe
> >> FPR6=e670d1fa33376800 3ffe FPR7=8e670d1fa3337800 4002
> >> XMM00=00000000000000000000000000000000
> >> XMM01=00000000000000000000000000000000
> >> XMM02=00000000000000000000000000000000
> >> XMM03=00000000000000000000000000000000
> >> XMM04=00000000000000000000000000000000
> >> XMM05=00000000000000000000000000000000
> >> XMM06=00000000000000000000000000000000
> >> XMM07=00000000000000000000000000000000
> >> Abort (core dumped)
> >>
> >
> > QEMU treats triple faults generated by the guest OS as fatal errors,
> > so it aborts execution and dumps core. In my opinion this is a too
> > self-punishing behavior that chould be replaced by a less harmful VM
> > restart. Triple faults are in fact fatal errors, so QEMU is correct,
> > in theory. In practice, however, some operating systems generate
> > triple faults on purpose in order to force a system reboot. The Linux
> > kernel used to do this. It appears that the FreeBSD boot loader does
> > it as well, so if you start FreeBSD and choose option 7 in the boot
> > prompt you will ever crash QEMU.
> >
> >
> >> %pkg_info|grep qemu
> >> kqemu-kmod-1.3.0.p11_9 Kernel Accelerator for QEMU CPU Emulator
> >> qemu-0.9.1_9        QEMU CPU Emulator
> >> %kldstat
> >> Id Refs Address    Size     Name
> >> 1   22 0xc0400000 701ae4   kernel
> >> 2    1 0xc0b02000 5844     if_tap.ko
> >> 3    1 0xc0b08000 15524    snd_hda.ko
> >> 4    2 0xc0b1e000 52a08    sound.ko
> >> 5    2 0xc0b71000 10ebc    drm.ko
> >> 6    1 0xc0b82000 71c4     i915.ko
> >> 7    1 0xc0b8a000 1ff24    kqemu.ko
> >> 8    1 0xc0baa000 b8c8     aio.ko
> >> 9    1 0xc0bb6000 6b3d0    acpi.ko
> >> 10    1 0xc4340000 9000     if_bridge.ko
> >> 11    1 0xc4349000 6000     bridgestp.ko
> >> 12    2 0xc44ac000 d000     ipfw.ko
> >> 13    1 0xc4500000 4000     ipdivert.ko
> >> 14    1 0xc4526000 22000    linux.ko
> >> 15    1 0xc45a0000 e000     fuse.ko
> >>
> >
> >
>
> _______________________________________________
> freebsd-ports at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> To unsubscribe, send any mail to "freebsd-ports-unsubscribe at freebsd.org"
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3110 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-ports/attachments/20080909/48e99d77/smime-0001.bin


More information about the freebsd-ports mailing list