This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: Code gen error on PowerPC
- To: law at cygnus dot com,Jeffrey A Law <law at hurl dot cygnus dot com>
- Subject: Re: Code gen error on PowerPC
- From: Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>
- Date: Wed, 10 Mar 1999 22:54:56 +0100
- Cc: Richard Henderson <rth at cygnus dot com>,Gary Thomas <gdt at linuxppc dot org>,egcs-ppc at egcs dot cygnus dot com,egcs-bugs at egcs dot cygnus dot com
- References: <4723.920359818@hurl.cygnus.com>
Am Tue, 02 Mar 1999 schrieb Jeffrey A Law:
>In message <4.2.0.25.19990302025929.0370ae50@mail.lauterbach.com>you write:
> > Hmm, I think this giv is setup by this hunk of code in find_mem_givs:
> >
> > case MEM:
> > {
> > rtx src_reg;
> > rtx add_val;
> > rtx mult_val;
> > int benefit;
> >
> > benefit = general_induction_var (XEXP (x, 0),
> > &src_reg, &add_val, &mult_val);
> >
> > /* Don't make a DEST_ADDR giv with mult_val == 1 && add_val == 0.
> > Such a giv isn't useful. */
> > if (benefit > 0 && (mult_val != const1_rtx || add_val != const0_rtx
> > ))
> > {
> > /* Found one; record it. */
> > struct induction *v
> > = (struct induction *) oballoc (sizeof (struct induction));
> >
> > record_giv (v, insn, src_reg, addr_placeholder, mult_val,
> > add_val, benefit, DEST_ADDR, not_every_iteration,
> > &XEXP (x, 0), loop_start, loop_end);
> >
> > v->mem_mode = GET_MODE (x);
> > }
> > }
> > return;
> >
> > But it's too late now and I'm not sure about anything :-) I'll continue
> > debugging tomorrow or wednesday.
>Could be. There's a call to general_induction_var, so it is a likely
>candidate.
It took me a little bit longer to get back to this problem, and I have a little
question now. I looked at the .loop file and noticed the following:
Insn 71: dest address src reg 81 benefit 7 used 1 lifetime 1 replaceable mult
8 add (plus:SI (plus:SI (reg:SI 88) (const_int -8))
(const_int 4))
Is this valid at this point? Shouldn't there be a single (plus:SI (reg:SI 88)
(const_int -4))?
This bug turns out to be quite hard to debug for me, as I can't see what's
going wrong yet.
Franz.