qemu discuss - Oct 2006 - qemu 0.8.2 and snv_49

After I did a fresh install with snv_49 qemu from the openSolaris  website does
not work anymore:

After I start qemu it shows the monitor with status "stopped" and
after a few seconds qemu core dumps.

The error occurs with and without the kqemu module (0.2)  loaded; the error is
the same in 64 bit and in 32 bit Solaris.

qemu 0.7.2 from blastwave and a self-compiled version (0.8.2 sources from the
qemu website) work fine

Is this a known bug?

regards 

Bernd

sample output (with kqemu NOT loaded):

(bash):root at ferrari:/export/install/config/cardbus # uname -a
SunOS ferrari 5.11 snv_49 i86pc i386 i86pc

(bash):root at ferrari:/export/install/config/cardbus # isainfo -kv
32-bit i386 kernel modules


truss /opt/tools/qemu/SUNWqemu/32/bin/qemu -L
/opt/tools/qemu/SUNWqemu/32/bin/bios -boot d  -m 128  -hda
/export/qemu/hdimg//dsl.hdd -cdrom /export/qemu/cdrom/dsl.iso

....

open64("/opt/tools/qemu/SUNWqemu/32/bin/bios/bios.bin", O_RDONLY) = 8
llseek(8, 0, SEEK_END)                          = 65536
llseek(8, 0, SEEK_SET)                          = 0
read(8, " U89E5 P Q06 W8B N\n83F9".., 65536)    = 65536
close(8)                                        = 0
open64("/opt/tools/qemu/SUNWqemu/32/bin/bios/vgabios-cirrus.bin",
O_RDONLY) = 8
llseek(8, 0, SEEK_END)                          = 35328
llseek(8, 0, SEEK_SET)                          = 0
read(8, " UAA EE9ED\0 P l e x 8 6".., 35328)    = 35328
close(8)                                        = 0
brk(0x1246FE60)                                 = 0
brk(0x12471E60)                                 = 0
brk(0x12471E60)                                 = 0
brk(0x12483E60)                                 = 0
    Received signal #14, SIGALRM [caught]
lwp_sigmask(SIG_SETMASK, 0xFFBFFEFF, 0x0000FFF7) = 0xFFBFFEFF [0x0000FFFF]
setcontext(0x080436A0)
brk(0x12483E60)                                 = 0
brk(0x1248FE60)                                 = 0
brk(0x1248FE60)                                 = 0
brk(0x12499E60)                                 = 0
llseek(5, 0, SEEK_SET)                          = 0
read(5, "\0\0\0\0\0\0\0\0\0\0\0\0".., 512)      = 512
time()                                          = 1160676739
brk(0x12499E60)                                 = 0
brk(0x1249BE60)                                 = 0
brk(0x1249BE60)                                 = 0
brk(0x1249DE60)                                 = 0
    Received signal #14, SIGALRM [caught]
setcontext(0x0804373C)
getcontext(0x080436A0)
lwp_sigmask(SIG_SETMASK, 0xFFBFFEFF, 0x0000FFF7) = 0xFFBFFEFF [0x0000FFFF]
setcontext(0x08043430)
brk(0x1249DE60)                                 = 0
brk(0x1249FE60)                                 = 0
brk(0x1249FE60)                                 = 0
brk(0x124A1E60)                                 = 0
brk(0x124A1E60)                                 = 0
brk(0x124A3E60)                                 = 0
    Incurred fault #6, FLTBOUNDS  %pc = 0x08107B87
      siginfo: SIGSEGV SEGV_MAPERR addr=0x08128000
    Received signal #11, SIGSEGV [default]
      siginfo: SIGSEGV SEGV_MAPERR addr=0x08128000
(bash):root at ferrari:/export/install/config/cardbus #
--
This message posted from opensolaris.org

Forgot to mention that the qemu binary worked without problems in snv_46 (with
and without kqemu)

Bernd
--
This message posted from opensolaris.org

Hi,

Bernd Schemmer wrote:
>After I did a fresh install with snv_49 qemu from the openSolaris  website
does not work anymore:
>
>After I start qemu it shows the monitor with status "stopped" and
after a few seconds qemu core dumps.
>
>The error occurs with and without the kqemu module (0.2)  loaded; the error
is the same in 64 bit and in 32 bit Solaris.
>
>qemu 0.7.2 from blastwave and a self-compiled version (0.8.2 sources from
the qemu website) work fine
>
>Is this a known bug?
>  
>

No.
But now it is known.

Thanks for reporting this.
I''m still running snv41 on x64, but I can upgrade my U20 next week.
I need to play with register usage setting myself on exactly snv_49.
But I''m uploading a new global SUNWqemu package update in a couple of
hours.
Built under Solaris 8 (and 10_FCS for amd64).
Built in the same way, as CSWqemu had been.
This might work around the problem before I can find the precise reason
for the scenario.


Regards,
Martin