[9fans] MIPS toolchain failure

[9fans] MIPS toolchain failure

[lucio]

This is on a conventional 386 system.  Compiling the linker (vl)
failed.  It could be caused by local changes, but I can't see how.
Can anyone confirm?

term% objtype=mips mk install
vc -FTVw asm.c
vc -FTVw list.c
vc -FTVw noop.c
vc -FTVw sched.c
vc -FTVw obj.c
vc 216994: suicide: sys: trap: fault read addr=0x50 pc=0x00006891
mk: vc -FTVw obj.c  : exit status=rc 216992: vc 216994: sys: trap: fault read addr=0x50 pc=0x00006891
term% acid 216994
/proc/216994/text:386 plan 9 executable
/sys/lib/acid/port
/sys/lib/acid/386
acid: lstk()
peep()+0x272 /sys/src/cmd/vc/peep.c:38
	t=0xe
	r=0xd51d0
	r1=0xfa400
	p=0x0
regopt(p=0x1a8ee8)+0xb26 /sys/src/cmd/vc/reg.c:395
	log5=0xc35
	val=0x7f
	r=0x0
	bit=0x2
	initpc=0x987
	npc=0x7f
	r1=0xd4098
	i=0xd
	vreg=0x80
codgen(nn=0x1a1a90,n=0x1a1d00)+0x195 /sys/src/cmd/cc/pgen.c:59
	n1=0x1a1808
	sp=0x1a4fa0
	nod1=0x0
	nod=0x0
yyparse()+0x3de /sys/src/cmd/cc/cc.y:1733
	save1=0x0
	save2=0x0
	save3=0x0
	save4=0x0
	yystate=0x2
	yychar=0xffffffff
	yys=0x0
	yyp=0xdfffd284
	yyn=0xfffffc18
	yypt=0xdfffd2cc
	w=0x0
	n=0x0
	s=0x0
compile(file=0xdfffefc2,ndef=0x0,defs=0xdfffee8c)+0x2ce /sys/src/cmd/cc/lex.c:284
	ofile=0x2e6a626f
	p=0x0
	incfile=0x70696d2f
	c=0x3
	fd=0x0
	av=0x0
	i=0x0
	opt=0x0
main(argv=0xdfffefac,argc=0x1)+0x199 /sys/src/cmd/cc/lex.c:144
	ndef=0x0
	_argc=0x77
	_args=0xdfffefc1
	defs=0x0
	nproc=0x0
	c=0x0
	nout=0x0
	status=0x0
_main+0x31 /sys/src/libc/386/main9.s:16
acid: \x04

Re: [9fans] MIPS toolchain failure

[erik quanstrom]

On Wed May 13 16:00:35 EDT 2009, lucio@proxima.alt.za wrote:
> This is on a conventional 386 system.  Compiling the linker (vl)
> failed.  It could be caused by local changes, but I can't see how.
> Can anyone confirm?
>

vc is able to compile vl here.  unless your source is different,
i don't see how you could fail there.  here's my source

>		switch(p->as) {
		default:
			r2 = rega();
			r->link = r2;
			r2->link = r1;

			r2->prog = p;
			r2->p1 = r;
			r->s1 = r2;
			r2->s1 = r1;
			r1->p1 = r2;

			r = r2;
			t++;

		case ADATA:
		case AGLOBL:
		case ANAME:
		case ASIGNAME:
>>			p = p->link;

i would think since p is not changed between the switch
and the reassignment of p, it must be that p is nil
at the switch statement, and therefore i would expect
the crash to be at > and not at >>.

- erik