Cross-compiling and testing sparc64 on Mac OS X

[Andreas Färber]

Hello guys,

It's awesome seeing you've made some progress towards emulating  
Solaris/sparc64!

However trying myself, compiling on Mac OS X (after making sure that  
NEED_INT128_T is defined) fails with the following error:

HOSTCC host/kernel/bootstrap.o
In file included from ../include/openbios/sysinclude.h:7,
                  from ../kernel/bootstrap.c:9:
/usr/include/stdlib.h:267: error: syntax error before ‘arc4random’
make[1]: *** [host/kernel/bootstrap.o] Error 1

The return type of arc4random is u_int32_t, which is defined (in my  
case) in <ppc/types.h>. Not sure what goes wrong there. I thought it  
could be a name clash with OpenBIOS' __TYPES_H (malc would be  
furious!), but - luckily - Apple uses __TYPES_H_ for their _types.h.  
Anyway, not having these functions declared by defining  
_POSIX_C_SOURCE made the compilation proceed with only one warning:

HOSTCC host/kernel/bootstrap.o
../kernel/bootstrap.c: In function ‘interpret_source’:
../kernel/bootstrap.c:529: warning: implicit declaration of function  
‘strcasecmp’

If I symlink .../share/qemu/openbios-sparc64 to .../obj-sparc64/ 
openbios-builtin.elf (or copy it over) I get this at runtime:

qemu: fatal: Trap 0x0010 while trap level (5) >= MAXTL (5), Error state
pc: 0000000000004200  npc: 0000000000004204
General Registers:
%g0: 0000000000000000	%g1: 0000000000000000	%g2: 0000000008000000	%g3:  
0000000000000000	
%g4: 0000000000000000	%g5: 0000000000000000	%g6: 0000000000000000	%g7:  
0000000000000000	
Current Register Window:
%o0: 00000000ffe00000	%o1: 00000000ffe01000	%o2: 000001fff0100000	%o3:  
000001fff0000000	
%o4: 0000000000000000	%o5: 0000000000000000	%o6: 0000000000000000	%o7:  
000001ff00000000	
%l0: 0000000007e80000	%l1: 000001ff00000000	%l2: 000001fff0080000	%l3:  
0000000000000000	
%l4: 0000000000000000	%l5: 0000000000000000	%l6: 0000000000000000	%l7:  
0000000000000000	
%i0: 0000000000000000	%i1: 0000000000000000	%i2: 0000000000000000	%i3:  
0000000000000000	
%i4: 0000000000000000	%i5: 0000000000000000	%i6: 0000000000000000	%i7:  
0000000000000000

Floating Point Registers:
%f00: 000000000.000000 000000000.000000 000000000.000000  
000000000.000000
%f04: 000000000.000000 000000000.000000 000000000.000000  
000000000.000000
%f08: 000000000.000000 000000000.000000 000000000.000000  
000000000.000000
%f12: 000000000.000000 000000000.000000 000000000.000000  
000000000.000000
%f16: 000000000.000000 000000000.000000 000000000.000000  
000000000.000000
%f20: 000000000.000000 000000000.000000 000000000.000000  
000000000.000000
%f24: 000000000.000000 000000000.000000 000000000.000000  
000000000.000000
%f28: 000000000.000000 000000000.000000 000000000.000000  
000000000.000000
pstate: 0x00000015 ccr: 0x99 asi: 0x00 tl: 5 fprs: 0
cansave: 6 canrestore: 0 otherwin: 0 wstate 0 cleanwin 6 cwp 7
fsr: 0x00000000
../qemu-solaris-sparc.command: line 3:  3289 Abort trap              / 
Users/andreas/Q/latest/bin/qemu-system-sparc64 -nographic -boot d - 
cdrom /Users/andreas/Q/sol-10-u3-ga-sparc-dvd.iso

This looks different from what you've posted and is identical with  
QEMU's version of openbios-sparc64. Same with or without -nographic.

Any ideas?

Regards,

Andreas

[Mark Cave-Ayland]

Andreas Färber wrote:
> Hello guys,
>
> It's awesome seeing you've made some progress towards emulating 
> Solaris/sparc64!
>
> However trying myself, compiling on Mac OS X (after making sure that 
> NEED_INT128_T is defined) fails with the following error:
>
> HOSTCC host/kernel/bootstrap.o
> In file included from ../include/openbios/sysinclude.h:7,
>                  from ../kernel/bootstrap.c:9:
> /usr/include/stdlib.h:267: error: syntax error before ‘arc4random’
> make[1]: *** [host/kernel/bootstrap.o] Error 1
>
> The return type of arc4random is u_int32_t, which is defined (in my 
> case) in <ppc/types.h>. Not sure what goes wrong there. I thought it 
> could be a name clash with OpenBIOS' __TYPES_H (malc would be furious!), 
> but - luckily - Apple uses __TYPES_H_ for their _types.h. Anyway, not 
> having these functions declared by defining _POSIX_C_SOURCE made the 
> compilation proceed with only one warning:
>
> HOSTCC host/kernel/bootstrap.o
> ../kernel/bootstrap.c: In function ‘interpret_source’:
> ../kernel/bootstrap.c:529: warning: implicit declaration of function 
> ‘strcasecmp’
>
> If I symlink .../share/qemu/openbios-sparc64 to 
> .../obj-sparc64/openbios-builtin.elf (or copy it over) I get this at 
> runtime:
>
> qemu: fatal: Trap 0x0010 while trap level (5) >= MAXTL (5), Error state
> pc: 0000000000004200  npc: 0000000000004204
> General Registers:
> %g0: 0000000000000000    %g1: 0000000000000000    %g2: 
> 0000000008000000    %g3: 0000000000000000   
> %g4: 0000000000000000    %g5: 0000000000000000    %g6: 
> 0000000000000000    %g7: 0000000000000000   
> Current Register Window:
> %o0: 00000000ffe00000    %o1: 00000000ffe01000    %o2: 
> 000001fff0100000    %o3: 000001fff0000000   
> %o4: 0000000000000000    %o5: 0000000000000000    %o6: 
> 0000000000000000    %o7: 000001ff00000000   
> %l0: 0000000007e80000    %l1: 000001ff00000000    %l2: 
> 000001fff0080000    %l3: 0000000000000000   
> %l4: 0000000000000000    %l5: 0000000000000000    %l6: 
> 0000000000000000    %l7: 0000000000000000   
> %i0: 0000000000000000    %i1: 0000000000000000    %i2: 
> 0000000000000000    %i3: 0000000000000000   
> %i4: 0000000000000000    %i5: 0000000000000000    %i6: 
> 0000000000000000    %i7: 0000000000000000
>
> Floating Point Registers:
> %f00: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f04: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f08: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f12: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f16: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f20: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f24: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f28: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> pstate: 0x00000015 ccr: 0x99 asi: 0x00 tl: 5 fprs: 0
> cansave: 6 canrestore: 0 otherwin: 0 wstate 0 cleanwin 6 cwp 7
> fsr: 0x00000000
> ../qemu-solaris-sparc.command: line 3:  3289 Abort trap              
> /Users/andreas/Q/latest/bin/qemu-system-sparc64 -nographic -boot d 
> -cdrom /Users/andreas/Q/sol-10-u3-ga-sparc-dvd.iso
>
> This looks different from what you've posted and is identical with 
> QEMU's version of openbios-sparc64. Same with or without -nographic.
>
> Any ideas?

Hmmm that's a tricky one. Does it not even get to the OpenBIOS prompt? 
You need to be using a recent git master for Qemu as well as OpenBIOS 
SVN, although I haven't updated my local copy for a couple of weeks.

Can you try launching a PPC Qemu instance with both the supplied 
openbios-ppc and then building your own to confirm that they both 
display the same behaviour?

ATB,

Mark.

[Andreas Färber]

Hello Mark,

Am 29.11.2009 um 14:43 schrieb Mark Cave-Ayland:
> Andreas Färber wrote:
> > qemu: fatal: Trap 0x0010 while trap level (5) >= MAXTL (5), Error  
> > state
> > pc: 0000000000004200  npc: 0000000000004204
> > General Registers:
> > %g0: 0000000000000000    %g1: 0000000000000000    %g2:  
> > 0000000008000000    %g3: 0000000000000000   %g4:  
> > 0000000000000000    %g5: 0000000000000000    %g6:  
> > 0000000000000000    %g7: 0000000000000000   Current Register Window:
> > %o0: 00000000ffe00000    %o1: 00000000ffe01000    %o2:  
> > 000001fff0100000    %o3: 000001fff0000000   %o4:  
> > 0000000000000000    %o5: 0000000000000000    %o6:  
> > 0000000000000000    %o7: 000001ff00000000   %l0:  
> > 0000000007e80000    %l1: 000001ff00000000    %l2:  
> > 000001fff0080000    %l3: 0000000000000000   %l4:  
> > 0000000000000000    %l5: 0000000000000000    %l6:  
> > 0000000000000000    %l7: 0000000000000000   %i0:  
> > 0000000000000000    %i1: 0000000000000000    %i2:  
> > 0000000000000000    %i3: 0000000000000000   %i4:  
> > 0000000000000000    %i5: 0000000000000000    %i6:  
> > 0000000000000000    %i7: 0000000000000000   Floating Point Registers:
> > %f00: 000000000.000000 000000000.000000 000000000.000000  
> > 000000000.000000
> > %f04: 000000000.000000 000000000.000000 000000000.000000  
> > 000000000.000000
> > %f08: 000000000.000000 000000000.000000 000000000.000000  
> > 000000000.000000
> > %f12: 000000000.000000 000000000.000000 000000000.000000  
> > 000000000.000000
> > %f16: 000000000.000000 000000000.000000 000000000.000000  
> > 000000000.000000
> > %f20: 000000000.000000 000000000.000000 000000000.000000  
> > 000000000.000000
> > %f24: 000000000.000000 000000000.000000 000000000.000000  
> > 000000000.000000
> > %f28: 000000000.000000 000000000.000000 000000000.000000  
> > 000000000.000000
> > pstate: 0x00000015 ccr: 0x99 asi: 0x00 tl: 5 fprs: 0
> > cansave: 6 canrestore: 0 otherwin: 0 wstate 0 cleanwin 6 cwp 7
> > fsr: 0x00000000
> > ../qemu-solaris-sparc.command: line 3:  3289 Abort  
> > trap              /Users/andreas/Q/latest/bin/qemu-system-sparc64 - 
> > nographic -boot d -cdrom /Users/andreas/Q/sol-10-u3-ga-sparc-dvd.iso
> > This looks different from what you've posted and is identical with  
> > QEMU's version of openbios-sparc64. Same with or without -nographic.
> > Any ideas?
>
> Hmmm that's a tricky one. Does it not even get to the OpenBIOS prompt?

Nope. Not ruling out a QEMU issue, but it's specific to sparc64 on Mac  
OS X host.
I never got further on sparc64, so the errors you are discussing now  
seemed like a leap forward. ;)

On Linux/amd64 I do get the OpenBIOS prompt with either 'open is not  
unique' or 'byte-load exception' (testing Solaris 10 U3, Solaris  
Nevada b127 and MilaX 0.3.2).
> You need to be using a recent git master for Qemu as well as  
> OpenBIOS SVN [...]

I'm at OpenBIOS trunk HEAD (r635) and was at QEMU master  
63e86b2e13396310feff1fe1a5988b5742487636.
> Can you try launching a PPC Qemu instance with both the supplied  
> openbios-ppc and then building your own to confirm that they both  
> display the same behaviour?

With ppc or sparc32 emulation I do not have this issue. Debian/sparc32  
boots fine 'til login and Debian/ppc has some graphic glitches  
(penguin miscolored on OSX/Cocoa, missing on Linux/SDL) but gets to  
the first installation screen.

Andreas

[Mark Cave-Ayland]

Andreas Färber wrote:
> Nope. Not ruling out a QEMU issue, but it's specific to sparc64 on Mac 
> OS X host.
> I never got further on sparc64, so the errors you are discussing now 
> seemed like a leap forward. ;)
>
> On Linux/amd64 I do get the OpenBIOS prompt with either 'open is not 
> unique' or 'byte-load exception' (testing Solaris 10 U3, Solaris Nevada 
> b127 and MilaX 0.3.2).

Right. So if you copy the openbios binary you built under Linux/amd64 to 
your OS X host, does it work? This will tell you whether it's an 
OpenBIOS build problem or a Qemu on OS X problem.

ATB,

Mark.

[Andreas Färber]

Am 29.11.2009 um 17:12 schrieb Mark Cave-Ayland:
> Andreas Färber wrote:
> > Nope. Not ruling out a QEMU issue, but it's specific to sparc64 on  
> > Mac OS X host.
> > I never got further on sparc64, so the errors you are discussing  
> > now seemed like a leap forward. ;)
> > On Linux/amd64 I do get the OpenBIOS prompt with either 'open is  
> > not unique' or 'byte-load exception' (testing Solaris 10 U3,  
> > Solaris Nevada b127 and MilaX 0.3.2).
>
> Right. So if you copy the openbios binary you built under Linux/ 
> amd64 to your OS X host, does it work? This will tell you whether  
> it's an OpenBIOS build problem or a Qemu on OS X problem.

It's probably not an OpenBIOS build problem. I was able to execute an  
OSX-built OpenBIOS on Linux and OpenSolaris. Executing a Linux-built  
OpenBIOS on OSX/ppc failed. Using same GCC 4.4.2 and binutils 2.20.
That still leaves OS differences, the endianness, the bitness, etc.  
etc. Maybe Blue has a hunch?

Andreas

[Andreas Färber]

Am 29.11.2009 um 19:32 schrieb Andreas Färber:
> Am 29.11.2009 um 17:12 schrieb Mark Cave-Ayland:
> > Andreas Färber wrote:
> > > Not ruling out a QEMU issue, but it's specific to sparc64 on Mac  
> > > OS X host.
> > > I never got further on sparc64, so the errors you are discussing  
> > > now seemed like a leap forward. ;)
> > > On Linux/amd64 I do get the OpenBIOS prompt with either 'open is  
> > > not unique' or 'byte-load exception' (testing Solaris 10 U3,  
> > > Solaris Nevada b127 and MilaX 0.3.2).
> >
> > Right. So if you copy the openbios binary you built under Linux/ 
> > amd64 to your OS X host, does it work? This will tell you whether  
> > it's an OpenBIOS build problem or a Qemu on OS X problem.
>
> It's probably not an OpenBIOS build problem. I was able to execute  
> an OSX-built OpenBIOS on Linux and OpenSolaris. Executing a Linux- 
> built OpenBIOS on OSX/ppc failed. Using same GCC 4.4.2 and binutils  
> 2.20.

I've just checked on OSX/i386, and both stock and custom openbios- 
sparc64 boot without the reported crash. So it's likely not a 32-bit  
issue and it's not Mac OS X per se...

This is what I got booting Debian GNU/Linux 4.0r3 with -nographic btw:

[...]
Remapping the kernel... done.
Booting Linux...
Inconsistent console: input 5, output 5
EXIT
11a>

Without -nographic it hangs after "Remapping the kernel... done.",  
making the fan spin louder and not dropping back to the OpenBIOS prompt.

Andreas

[Blue Swirl]

On Sun, Nov 29, 2009 at 8:32 PM, Andreas Färber  wrote:
> Am 29.11.2009 um 17:12 schrieb Mark Cave-Ayland:
> > Andreas Färber wrote:
> > > Nope. Not ruling out a QEMU issue, but it's specific to sparc64 on Mac OS
> > > X host.
> > > I never got further on sparc64, so the errors you are discussing now
> > > seemed like a leap forward. ;)
> > > On Linux/amd64 I do get the OpenBIOS prompt with either 'open is not
> > > unique' or 'byte-load exception' (testing Solaris 10 U3, Solaris Nevada b127
> > > and MilaX 0.3.2).
> >
> > Right. So if you copy the openbios binary you built under Linux/amd64 to
> > your OS X host, does it work? This will tell you whether it's an OpenBIOS
> > build problem or a Qemu on OS X problem.
>
> It's probably not an OpenBIOS build problem. I was able to execute an
> OSX-built OpenBIOS on Linux and OpenSolaris. Executing a Linux-built
> OpenBIOS on OSX/ppc failed. Using same GCC 4.4.2 and binutils 2.20.
> That still leaves OS differences, the endianness, the bitness, etc. etc.
> Maybe Blue has a hunch?

What about x86_64 target or other 64 bit targets, do you have problems
with those?

[Andreas Färber]

Hey,

Am 30.11.2009 um 22:03 schrieb Blue Swirl:
> What about x86_64 target or other 64 bit targets, do you have problems
> with those?

Hadn't checked yet. Booting Fedora 12 DVD results in immediate segfault!

(gdb) run -boot d -cdrom ../Fedora-12-x86_64-DVD.iso -nographic
Starting program: /Users/andreas/QEMU/latest/bin/qemu-system-x86_64 - 
boot d -cdrom ../Fedora-12-x86_64-DVD.iso -nographic
Reading symbols for shared libraries +++++++++++++++ 
+............................................................... done
Reading symbols for shared libraries . done
Reading symbols for shared libraries . done
Reading symbols for shared libraries . done
Reading symbols for shared libraries . done
Reading symbols for shared libraries . done
Reading symbols for shared libraries .. done

Program received signal EXC_BAD_ACCESS, Could not access memory.

Reason: KERN_PROTECTION_FAILURE at address: 0x00000000
                                       __stl_mmu (addr=0, val=0,  
mmu_idx=-1) at /Users/andreas/QEMU/qemu/target-i386/op_helper.c:259
                                                    259	    return  
p[0] | (p[1] << 8) | (p[2] << 16) | (p[3] << 24);
(gdb) bt
         #0  __stl_mmu (addr=0, val=0, mmu_idx=-1) at /Users/andreas/ 
QEMU/qemu/target-i386/op_helper.c:259
                                                                                                          #1 
   0x06000500 in ?? ()
                                                                                                                                 #2 
   0x000dbcd0 in cpu_x86_exec (env1=<value temporarily unavailable,  
due to optimizations>) at /Users/andreas/QEMU/qemu/target-i386/ 
op_helper.c:259
                                                                                                                            #3 
   0x0000a5b8 in qemu_main (argc=6, argv=0xbffff7e4, envp=<value  
temporarily unavailable, due to optimizations>) at /Users/andreas/QEMU/ 
qemu/target-i386/op_helper.c:259
                                                                                                                                             #4 
   0x000c9268 in -[QemuCocoaAppController  
startEmulationWithArgc:argv:] (self=<value temporarily unavailable,  
due to optimizations>, _cmd=<value temporarily unavailable, due to  
optimizations>, argc=<value temporarily unavailable, due to  
optimizations>, argv=<value temporarily unavailable, due to  
optimizations>) at /Users/andreas/QEMU/qemu/target-i386/op_helper.c:259
                                                      #5  0x93765688  
in _nsnote_callback ()
                                                                                           #6 
   0x94c0643c in _CFXNotificationPostNotification ()
                                                                                                                                                #7 
   0x93762dc4 in -[NSNotificationCenter  
postNotificationName:object:userInfo:] ()
                                                                          #8 
   0x940f1858 in -[NSApplication _postDidFinishNotification] ()
                                                                                                                                          #9 
   0x940f1770 in -[NSApplication _sendFinishLaunchingNotification] ()
                                                        #10 0x94079114  
in -[NSApplication(NSAppleEventHandling) _handleAEOpen:] ()
                                                                                                                                  #11 
  0x9407894c in -[NSApplication(NSAppleEventHandling)  
_handleCoreEvent:withReplyEvent:] ()

#12 0x93786008 in -[NSAppleEventManager  
dispatchRawAppleEvent:withRawReply:handlerRefCon:] ()
           #13 0x93785ddc in _NSAppleEventManagerGenericHandler ()
                                                                  #14  
0x93498ce4 in aeDispatchAppleEvent ()
                                                                                                           #15 
  0x93498bec in dispatchEventAndSendReply ()
                                                                                                                                                        #16 
  0x934989f0 in aeProcessAppleEvent ()
                                        #17 0x95579410 in  
AEProcessAppleEvent ()
                                                                                #18 
  0x94076478 in _DPSNextEvent ()
                                                                                                                  #19 
  0x94075c00 in -[NSApplication  
nextEventMatchingMask:untilDate:inMode:dequeue:] ()
                                               #20 0x9406f8a0 in - 
[NSApplication run] ()
                                                                                        #21 
  0x000c9a8c in main (argc=<value temporarily unavailable, due to  
optimizations>, argv=<value temporarily unavailable, due to  
optimizations>) at /Users/andreas/QEMU/qemu/target-i386/op_helper.c:259
                                                                                                                                       Current 
  language:  auto; currently c
(gdb)

Andreas

[Blue Swirl]

On Fri, Nov 27, 2009 at 1:49 AM, Andreas Färber  wrote:
> Hello guys,
>
> It's awesome seeing you've made some progress towards emulating
> Solaris/sparc64!
>
> However trying myself, compiling on Mac OS X (after making sure that
> NEED_INT128_T is defined) fails with the following error:
>
> HOSTCC host/kernel/bootstrap.o
> In file included from ../include/openbios/sysinclude.h:7,
>                 from ../kernel/bootstrap.c:9:
> /usr/include/stdlib.h:267: error: syntax error before ‘arc4random’
> make[1]: *** [host/kernel/bootstrap.o] Error 1
>
> The return type of arc4random is u_int32_t, which is defined (in my case) in
> <ppc/types.h>. Not sure what goes wrong there. I thought it could be a name
> clash with OpenBIOS' __TYPES_H (malc would be furious!), but - luckily -
> Apple uses __TYPES_H_ for their _types.h. Anyway, not having these functions
> declared by defining _POSIX_C_SOURCE made the compilation proceed with only
> one warning:
>
> HOSTCC host/kernel/bootstrap.o
> ../kernel/bootstrap.c: In function ‘interpret_source’:
> ../kernel/bootstrap.c:529: warning: implicit declaration of function
> ‘strcasecmp’
>
> If I symlink .../share/qemu/openbios-sparc64 to
> .../obj-sparc64/openbios-builtin.elf (or copy it over) I get this at
> runtime:
>
> qemu: fatal: Trap 0x0010 while trap level (5) >= MAXTL (5), Error state
> pc: 0000000000004200  npc: 0000000000004204
> General Registers:
> %g0: 0000000000000000   %g1: 0000000000000000   %g2: 0000000008000000   %g3:
> 0000000000000000
> %g4: 0000000000000000   %g5: 0000000000000000   %g6: 0000000000000000   %g7:
> 0000000000000000
> Current Register Window:
> %o0: 00000000ffe00000   %o1: 00000000ffe01000   %o2: 000001fff0100000   %o3:
> 000001fff0000000
> %o4: 0000000000000000   %o5: 0000000000000000   %o6: 0000000000000000   %o7:
> 000001ff00000000
> %l0: 0000000007e80000   %l1: 000001ff00000000   %l2: 000001fff0080000   %l3:
> 0000000000000000
> %l4: 0000000000000000   %l5: 0000000000000000   %l6: 0000000000000000   %l7:
> 0000000000000000
> %i0: 0000000000000000   %i1: 0000000000000000   %i2: 0000000000000000   %i3:
> 0000000000000000
> %i4: 0000000000000000   %i5: 0000000000000000   %i6: 0000000000000000   %i7:
> 0000000000000000
>
> Floating Point Registers:
> %f00: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f04: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f08: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f12: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f16: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f20: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f24: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> %f28: 000000000.000000 000000000.000000 000000000.000000 000000000.000000
> pstate: 0x00000015 ccr: 0x99 asi: 0x00 tl: 5 fprs: 0
> cansave: 6 canrestore: 0 otherwin: 0 wstate 0 cleanwin 6 cwp 7
> fsr: 0x00000000
> ../qemu-solaris-sparc.command: line 3:  3289 Abort trap
>  /Users/andreas/Q/latest/bin/qemu-system-sparc64 -nographic -boot d -cdrom
> /Users/andreas/Q/sol-10-u3-ga-sparc-dvd.iso
>
> This looks different from what you've posted and is identical with QEMU's
> version of openbios-sparc64. Same with or without -nographic.
>
> Any ideas?

Please start QEMU with -d in_asm,out_asm and post the last 100 lines
(or so) of /tmp/qemu.log.

I'd suspect a bug in Sparc64 emulation, or PPC code generator.

[Andreas Färber]

Am 30.11.2009 um 22:01 schrieb Blue Swirl:
> Please start QEMU with -d in_asm,out_asm and post the last 100 lines
> (or so) of /tmp/qemu.log.
>
> I'd suspect a bug in Sparc64 emulation, or PPC code generator.

Attached is the output of `tail -n 100 < /tmp/qemu.log` for MilaX 0.3.2.

In light of x86_64 and ppc64 bus errors I'd suspect the ppc TCG target?

Thanks,
Andreas

[Andreas Färber]

Hi,

Am 30.11.2009 um 23:23 schrieb Andreas Färber:
> Am 30.11.2009 um 22:01 schrieb Blue Swirl:
> > Please start QEMU with -d in_asm,out_asm and post the last 100 lines
> > (or so) of /tmp/qemu.log.
> >
> > I'd suspect a bug in Sparc64 emulation, or PPC code generator.

I've ported the ppc64 TCG backend to OSX and found that both sparc64  
and x64 work there.

It's still odd, but thanks for the TCG hint!

Andreas