Re: Raid-1 woes

[Herbert Xu <herbert@gondor.apana.org.au>]

On Fri, Jun 08, 2001 at 09:04:44AM +1000, Craig Small wrote:
> 
>      e94:       d8 03 f0 a0     ldl     t6,984(a0)
>      e98:       25 95 a4 40     subq    t4,0x24,t4
>      e9c:       26 95 c4 40     subq    t5,0x24,t5
>      ea0:       00 00 25 a0     ldl     t0,0(t4)

This is where the fault is.  Compiling the same file here with gcc 2.95
reveals a compiler bug.  A few lines up from here, you'll see something
similar to

        s8addq $25,$25,$1
        mov $23,$28
        addq $16,32,$25
        s4addq $1,0,$23
        addl $2,$31,$24
        addq $22,$25,$4
        addq $22,$28,$3

where $25 is never initialised when the s8addq happens.  Perhaps you
should give gcc 3.0 a spin.
-- 
Debian GNU/Linux 2.2 is out! ( http://www.debian.org/ )
Email:  Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt