This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug middle-end/23954] [4.1 regression] FAIL: gcc.c-torture/execute/20040709-1.c execution, -Os
- From: "dave at hiauly1 dot hia dot nrc dot ca" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 18 Sep 2005 21:35:05 -0000
- Subject: [Bug middle-end/23954] [4.1 regression] FAIL: gcc.c-torture/execute/20040709-1.c execution, -Os
- References: <20050918173415.23954.danglin@gcc.gnu.org>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Additional Comments From dave at hiauly1 dot hia dot nrc dot ca 2005-09-18 21:35 -------
Subject: Re: [4.1 regression] FAIL: gcc.c-torture/execute/20040709-1.c execution, -Os
> Hmm, does this fail also on "hppa2.0w-hp-hpux11.11"? I want to say it does.
It's reproducible on hppa2.0w-hp-hpux11.11 with "-mschedule=7100LC".
It doesn't fail with "-mschedule=8000" where we have somewhat different
code:
L$0014:
.CALL
bl myrnd,%r2
nop
stb %r28,0(%r5)
ldo 1(%r5),%r5
ldo 8(%r8),%r28
comb,<> %r28,%r5,L$0014
addil LR'sB-$global$,%r27
ldw -8(%r5),%r28
Hmmm, I just realized that the reason that the 7100LC code fails is
that the "ldw -8(%r5),%r28" has been placed in the delay slot of the
"comb" causing it to be executed each trip around the loop.
Dave
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23954