mips gcc 2.95.2 and 2.91.66 bug

[Fabrice Bellard <bellard@email.enst.fr>]

Hi!

Maybe this bug can interest you: when using byte swaping in le16_to_cpu
for example, mips gcc 2.95.2 and 2.91.66 sometime do not generate correct
code : the u16 to u32 convertion is missing. I found this bug while
compiling drivers/mtd/ftl.c in build_maps(). Here is a sample source to
reproduce the bug:

typedef unsigned short __u16;

extern __inline__ __const__ __u16 le16_to_cpu(__u16 x)
{
    return ((__u16)( \
                (((__u16)(x) & (__u16)0x00ffU) << 8) | \
                (((__u16)(x) & (__u16)0xff00U) >> 8) ));
}

int test(int xtrans, int xvalid, __u16 *ptr)
{
    if ((xvalid+xtrans != le16_to_cpu(*ptr))) {
        return -1;
    }
    return 0;
}

The generated asm is :

test:
        .frame  $sp,0,$31               # vars= 0, regs= 0/0, args= 0,
extra= 0
        .mask   0x00000000,0
        .fmask  0x00000000,0
        .set    noreorder
        .cpload $25
        .set    reorder
        lhu     $3,0($7)
        addu    $5,$6,$5
        sll     $4,$3,8
        srl     $3,$3,8
        or      $4,$4,$3
        .set    noreorder
        .set    nomacro
        bne     $5,$4,$L7
        li      $2,-1                   # 0xffffffff
        .set    macro
        .set    reorder

        move    $2,$0
$L7:
        j       $31
        .end    test

The andi op is missing.

egcs-1.0.3 seems to be OK.

I have no experience in submitting gcc bugs, so if someone could forward
this mail to the relevant gcc mailing list...

Fabrice.