This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH: simplify_and_const_int - disable call to force_to_mode
- To: jsm28 at cam dot ac dot uk (Joseph S. Myers)
- Subject: Re: PATCH: simplify_and_const_int - disable call to force_to_mode
- From: Toshi Morita <tm2 at best dot com>
- Date: Mon, 26 Feb 2001 15:10:38 -0800 (PST)
- Cc: gcc-bugs at gcc dot gnu dot org, gcc-patches at gcc dot gnu dot org
> On Mon, 26 Feb 2001, Toshi Morita wrote:
>
> > Mon Feb 26 10:37:43 PST 2001 Toshiyasu Morita (tm2@best.com)
> >
> > * combine.c (simplify_and_const_int): Disable call to
> > force_to_mode.
>
> Which testcase does this fix, on what target?
>
> --
> Joseph S. Myers
> jsm28@cam.ac.uk
>
>
Okay, I managed to create a testcase which elicits the bug.
Build an sh-elf target compiler, and compile this sample:
int b = 5;
void main(void)
{
int a = 0;
if (b) {
func();
}
printf("%d\n", (a + 23) & 0xfffffffc);
}
using the options:
-O1 -S test.c
and you'll get:
_main:
mov.l r14,@-r15
sts.l pr,@-r15
mov.l .L4,r1
jsr @r1
mov r15,r14
mov.l .L5,r1
mov.l @r1,r1
tst r1,r1
bt .L3
mov.l .L6,r1
jsr @r1
nop
.L3:
mov.l .L7,r1
mov.l .L8,r4
jsr @r1
mov #23,r5 <- doh! is never ANDed against 0xfffffffc
...
With the previous patch, I get:
_main:
mov.l r14,@-r15
sts.l pr,@-r15
mov.l .L4,r1
jsr @r1
mov r15,r14
mov.l .L5,r1
mov.l @r1,r1
tst r1,r1
bt .L3
mov.l .L6,r1
jsr @r1
nop
.L3:
mov #23,r0
and #60,r0 <- correctly ANDed before passing to printf
mov.l .L7,r1
mov.l .L8,r4
jsr @r1
mov r0,r5
Toshi