[Qemu-devel] Regression in QEMU 2.6.0 - Pentium III SSE fails

[Stefan Weil]

Hi Richard,

I got this error report from a Windows user of QEMU:

    QEMU 2.6.0 RC4 can no longer boot Slackware or Zenwalk.
    Boot fails with "Panic: Attempted to kill init."

    RC3 didn't work too (screenshot attached)
    My command line is:
    qemu-system-i386w.exe -m 512 -cpu pentium3 -vga cirrus -soundhw sb16
-net nic,model=ne2k_pci -net user -hda c.img -boot c
    I tried Slackware 9.0 and Zenwalk 2.6.

    The last working version of QEMU was 2.5.0

The error is not Windows related, but can be reproduced on Linux
when running latest QEMU with TCG. In my tests, I used a simplified
command line:

qemu-system-i386 -m 512 -net nic,model=ne2k_pci -net none -hda ~/c.img
-snapshot

Slackware uses a 2.4 kernel. This kernel fails while measuring
checksumming speed with pIII_sse:

md: linear personality registered as nr 1
md: raid0 personality registered as nr 2
md: raid1 personality registered as nr 3
md: raid5 personality registered as nr 4
raid5: measuring checksumming speed
   8regs     :   384.400 MB/sec
   32regs    :   259.200 MB/sec
invalid operand: 0000
CPU:    0
EIP:    0010:[<c0240b2a>]    Not tainted
EFLAGS: 00000246
eax: c15d8000   ebx: 00000000   ecx: 00000000   edx: c15d5000
esi: 8005003b   edi: 00000004   ebp: 00000000   esp: c15bdf50
ds: 0018   es: 0018   ss: 0018     
Process swapper (pid: 1, stackpage=c15bd000)
Stack: 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000
       00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000
       00000000 00000206 c0241c6c 00001000 c15d4000 c15d7000 c15d4000
c15d4000
Call Trace:    [<c0241c6c>] [<c0105000>] [<c0241db4>] [<c010503b>]
[<c0105000>]
  [<c0107416>] [<c0105030>]            

Code: 0f ae f8 0f 10 04 24 0f 10 4c 24 10 0f 10 54 24 20 0f 10 5c
 <0>Kernel panic: Attempted to kill init!

It should have been like this:

md: linear personality registered as nr 1
md: raid0 personality registered as nr 2
md: raid1 personality registered as nr 3
md: raid5 personality registered as nr 4
raid5: measuring checksumming speed
   8regs     :   372.800 MB/sec
   32regs    :   241.200 MB/sec
   pIII_sse  :   540.400 MB/sec
   pII_mmx   :   176.400 MB/sec
   p5_mmx    :   205.600 MB/sec
raid5: using function: pIII_sse (540.400 MB/sec)
md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
md: Autodetecting RAID arrays.
md: autorun ...

I think that the problem can be reproduced with other guests
running a 2.4 kernel, too.

Bisecting was difficult, but I could narrow down the range of
commits which caused the regression:

64dbaff09bb768dbbb13142862554f18ab642866 is good.
f4f1110e4b34797ddfa87bb28f9518b9256778be is bad.

Regards
Stefan

signature.asc
Description: OpenPGP digital signature