This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: GCC build failed with your patch.
> On Mon, 31 Jul 2000, Jan Hubicka wrote:
> > I've just commited my version of the patch - I just forgot to fill
> > cvs.commit file so it was waiting till now, since I've made the fix
> > yesterdat (or today?) before I went sleep. Hope that gcc is back in
> > bootstrapland now.
>
> I am afraid it is not! I am still seeing the bootstrap failure reported
> in http://gcc.gnu.org/ml/gcc-bugs/2000-07/msg00767.html
Duh!
I see that I've cheked in the different version of patch that I've got
approved and bootstrapped!
I not have to bootstrap on different machine that I use for CVS, since
that one lacks some tools required now and I didn't had time to install
it.
Going to check in the obvious bugfix.
Interesting bit is, that this problem don't show in the bootstrap for me
anyway (I've already bootstrapped from fresh CVS). Purpose is that with
ppro tunning the insn is just never constructed. The patch just synchronizes
the broken patch I've installed with one I got approved.
Really embrasling.
BTW also interesting bit is that I didn't received the second failure notice
Thanks
Honza
Mon Jul 31 12:24:00 MET DST 2000 Jan Hubicka <jh@suse.cz>
* i386.md (addsi to lea splitter, ashlqi3_1_lea): Fix bugs
in my last checkin.
Index: egcs/gcc/config/i386//i386.md
===================================================================
RCS file: /cvs/gcc/egcs/gcc/config/i386/i386.md,v
retrieving revision 1.168
diff -c -3 -p -r1.168 i386.md
*** i386.md 2000/07/31 00:06:48 1.168
--- i386.md 2000/07/31 10:23:17
***************
*** 3902,3908 ****
"
{
rtx pat;
- operands[0] = gen_lowpart (SImode, operands[0]);
/* In -fPIC mode the constructs like (const (unspec [symbol_ref]))
may confuse gen_lowpart. */
if (GET_MODE (operands[0]) != Pmode)
--- 3902,3907 ----
***************
*** 3910,3915 ****
--- 3909,3915 ----
operands[1] = gen_lowpart (Pmode, operands[1]);
operands[2] = gen_lowpart (Pmode, operands[2]);
}
+ operands[0] = gen_lowpart (SImode, operands[0]);
pat = gen_rtx_PLUS (Pmode, operands[1], operands[2]);
if (Pmode != SImode)
pat = gen_rtx_SUBREG (SImode, pat, 0);
***************
*** 6691,6697 ****
(define_insn "*ashlqi3_1_lea"
[(set (match_operand:QI 0 "nonimmediate_operand" "=qm,r,r")
(ashift:QI (match_operand:QI 1 "nonimmediate_operand" "0,0,r")
! (match_operand:QI 2 "nonmemory_operand" "cI,cI,r")))
(clobber (reg:CC 17))]
"!TARGET_PARTIAL_REG_STALL
&& ix86_binary_operator_ok (ASHIFT, QImode, operands)"
--- 6691,6697 ----
(define_insn "*ashlqi3_1_lea"
[(set (match_operand:QI 0 "nonimmediate_operand" "=qm,r,r")
(ashift:QI (match_operand:QI 1 "nonimmediate_operand" "0,0,r")
! (match_operand:QI 2 "nonmemory_operand" "cI,cI,M")))
(clobber (reg:CC 17))]
"!TARGET_PARTIAL_REG_STALL
&& ix86_binary_operator_ok (ASHIFT, QImode, operands)"