Re: [patch] SMP alternatives: skip with UP kernels.

[Mikael Pettersson]

On 11 May 2006 20:28:09 +0200, Andi Kleen wrote:
>Mikael Pettersson <mikpe@xxxxxxxx> writes:
>
>> On Thu, 11 May 2006 14:13:22 +0200, Gerd Hoffmann wrote:
>> >+	if (0 == (__smp_alt_end - __smp_alt_begin))
>> >+		return; /* no tables, nothing to patch (UP kernel) */
>> 
>> That's an unnecessarily obscure way of stating the obvious:
>> 
>> 	if (__smp_alt_end == __smp_alt_begin)
>
>iirc ISO-C guarantees that symbols have different values and the
>optimizer could possibly make use of that fact. So you might actually
>need some RELOC_HIDE()s to make this safe.

OK, I didn't consider that (with "extern $type a[], b[];", "a == b"
is false by definition). However, the (0 == (_ - _)) version is no
safer, since a compiler can legally turn it back to a single "==".

Since you brought up the "rules of C" argument: the
(__smp_alt_end - __smp_alt_begin) expression is undefined because
pointer subtraction is only valid if both pointers point into
the same object (or just after it), which isn't the case here.

A plain "==" with RELOC_HIDE() on the operands would be best IMO.

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