Re: [PATCH] udf: uint32_t can't be less than zero

[Denis Vlasenko <vda@ilport.com.ua>]

On Saturday 23 April 2005 21:48, Jesper Juhl wrote:
> Here's a patch that removes a few bits from fs/udf/balloc.c that 
> test uint32_t values for being less than zero, which is impossible.
> 
> I know not everyone agree with this sort of cleanup, but I figured I'd do 
> the patch in any case, then leave it up to the maintainer to apply it or 
> drop it.
> 
> Please keep me on CC: when replying.
> -	if (bloc.logicalBlockNum < 0 ||
> -		(bloc.logicalBlockNum + count) > UDF_SB_PARTLEN(sb, bloc.partitionReferenceNum))
> +	if ((bloc.logicalBlockNum + count) > UDF_SB_PARTLEN(sb, bloc.partitionReferenceNum))

It is not immediately visible here that bloc.logicalBlockNum is unsigned.
One needs to check that by looking at the definition.

Also if later someone changes bloc.logicalBlockNum into signed entity, code
becomes buggy. Not good.

gcc already optimizes out such checks:

# gcc -O2 t.c -S -fomit-frame-pointer
# cat t.c t.s
extern unsigned v;

int f() {
    return v<0 || v>100 || v==50;
}
        .file   "t.c"
        .text
        .p2align 2,,3
.globl f
        .type   f, @function
f:
        movl    v, %edx
        xorl    %eax, %eax
        cmpl    $100, %edx
        ja      .L3
        cmpl    $50, %edx
        je      .L3
        ret
        .p2align 2,,3
.L3:
        movl    $1, %eax
        ret
        .size   f, .-f
        .section        .note.GNU-stack,"",@progbits
        .ident  "GCC: (GNU) 3.4.3"
--
vda

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/