Re: [PATCH 6/8] i386: bitops: Don't mark memory as clobbered unnecessarily

[Jeremy Fitzhardinge <jeremy@goop.org>]

Linus Torvalds wrote:
> Sure, that's *one* thing that "volatile" guarantees (it guarantees that 
> gcc won't optimize away things where the end result isn't actually visibly 
> used).
>
> But gcc docs also talk about the other things volatile means, including 
> "not significantly moved".
>   

Actually, it doesn't.  In fact it goes out of its way to say that "asm
volatile" statements can be moved quite a bit, with respect to other
asms, other code, jumps, basic blocks, etc.  The only reliable way to
control the placement of an asm is have the right dependencies.

    The `volatile' keyword indicates that the instruction has important
    side-effects.  GCC will not delete a volatile `asm' if it is reachable.
    (The instruction can still be deleted if GCC can prove that
    control-flow will never reach the location of the instruction.)  Note
    that even a volatile `asm' instruction can be moved relative to other
    code, including across jump instructions.

also:

    An `asm' instruction without any output operands will be treated
    identically to a volatile `asm' instruction.

So there isn't anything very special about "asm volatile".  It's purely
to stop apparently useless code from being removed.

    J