optimization/3311: GCC-SH: gcc loses result of AND operation due to force_to_mode bug

[tm2 at best dot com]

>Number:         3311
>Category:       optimization
>Synopsis:       GCC-SH: gcc loses result of AND operation due to force_to_mode bug
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Jun 21 01:16:00 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator:     Toshiyasu Morita
>Release:        3.0
>Organization:
>Environment:
host: i686-linux (Red Hat 7.0)
target: sh-elf
>Description:
When this code:

int b = 5;

void main(void)

{
        int a = 0;

        if (b) {
                func();
        }

        printf("%d\n", (a + 23) & 0xfffffffc);
}

is compiled with these options

-O1 -m4-single-only -ml -S 


the following code is created:

_main:
        mov.l   r14,@-r15
        sts.l   pr,@-r15
        mov.l   .L3,r1
        mov.l   @r1,r1
        tst     r1,r1
        bt.s    .L2
        mov     r15,r14
        mov.l   .L4,r0
        jsr     @r0
        nop
.L2:
        mov.l   .L5,r4
        mov.l   .L6,r0
        jsr     @r0
        mov     #23,r5
        mov     r14,r15
        lds.l   @r15+,pr
        rts
        mov.l   @r15+,r14
.L7:
        .align 2
.L3:
        .long   _b
.L4:
        .long   _func
.L5:
        .long   .LC0
.L6:
        .long   _printf


Basically, the code calls printf with the value "23" in r5, which is wrong.
.It should be calling printf with the value "20" in r5.

My analysis of the bug and patch are at:

http://gcc.gnu.org/ml/gcc-patches/2001-04/msg01267.html

>How-To-Repeat:

>Fix:
See message http://gcc.gnu.org/ml/gcc-patches/2001-04/msg01267.html
>Release-Note:
>Audit-Trail:
>Unformatted: