basilisk-devel

[B2-devel] SheepShaver on x86_64 continued

[Tomek J. <to...@je...>]

I know what is causing the following errors:

I/usr/include/glib-1.2 -I/usr/lib64/glib/include -I/usr/X11R6/include -
fpermissive -c ../kpx_cpu/src/cpu/ppc/ppc-dyngen.cpp -o obj/ppc-dyngen.=
o
/tmp/ccP0F64j.s: Assembler messages:
/tmp/ccP0F64j.s:54225: Error: suffix or operands invalid for `push'
/tmp/ccP0F64j.s:54225: Error: suffix or operands invalid for `push'
/tmp/ccP0F64j.s:54225: Error: suffix or operands invalid for `pop'
/tmp/ccP0F64j.s:54225: Error: suffix or operands invalid for `pop'
make: *** [obj/ppc-dyngen.o] B=C5=82=C4=85d 1


In SheepShaver/src/cpu/kpx_cpu/src/cpu/ppc/ppc-dyngen.cpp I commented
out the following lines (starting frrom line 67):

/*=20
	__asm__ ("push %%ecx ; push %%ebx ; cpuid ; pop %%ebx ; pop %%ecx"
			 : "=3Da" (fl1) : "0" (0) : "edx", "cc");
	if (fl1 =3D=3D 0)
*/

Now it compiles fine.

Jerzu

Re: [B2-devel] SheepShaver on x86_64 continued

[Gwenole B. <gb...@di...>]

mercredi 10 novembre 2004, =E0 07:51  pm, Tomek Jerzykowski a =E9crit :

> In SheepShaver/src/cpu/kpx_cpu/src/cpu/ppc/ppc-dyngen.cpp I commented
> out the following lines (starting frrom line 67):

Yes, but you will get lower AltiVec performance since you wouldn't use=20=

MMX/SSE in that case.

I have fixed it in CVS. The cpuid test was from GCC and got fixed later=20=

too.