--- Additional Comments From hjl at lucon dot org 2005-05-15 15:36 ---
The change in ia64_expand_move
if (addend)
{
rtx subtarget = no_new_pseudos ? op0 : gen_reg_rtx (mode);
emit_insn (gen_rtx_SET (VOIDmode, subtarget, op1));
op1 = expand_si
--- Additional Comments From hjl at lucon dot org 2005-05-15 14:35 ---
It seems that "-fgcse -O" will trigger the bug.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21551
--- Additional Comments From hjl at lucon dot org 2005-05-15 14:12 ---
Created an attachment (id=8891)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8891&action=view)
A testcase
With the bad compiler, I got
[EMAIL PROTECTED] gcc]$ stage1/xgcc -Bstage1/ -O2 -S /tmp/foo.c
[EMAIL PRO
--- Additional Comments From hjl at lucon dot org 2005-05-15 01:03 ---
FYI, I was comparing the assembly outputs from the SAME source with and without
the patch mentioned. I will try to come up with a small testcase.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21551
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-15
00:59 ---
I would not doubt that CODE_FOR_insv changed also after his patch so that might
not be it.
--
What|Removed |Added
---
--- Additional Comments From hjl at lucon dot org 2005-05-15 00:12 ---
This patch
http://gcc.gnu.org/ml/gcc-patches/2005-05/msg01151.html
seems the cause. expmed.c compiled with the after compiler has
.mii
nop 0
(p6) addl r48 = @ltoffx(insn_data#+32768), r1
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-13
23:06 ---
The only change which could have caused this would be RTH but I know RTH
bootstrapped and tested
the patch.
Are you sure that you are not using a broken binutils?
--
http://gcc.gnu.org/bugzilla/show_b
