Re: spinlock

David Wragg (dpw@doc.ic.ac.uk)
01 Dec 1999 01:01:11 +0000

Jamie Lokier <lkd@tantalophile.demon.co.uk> writes:
> David Wragg wrote:
> > Pavel Machek <pavel@suse.cz> writes:
> > > We should not do such tricks. gcc 2.95 may not be clever enough, but
> > > gcc 3.05? egcs people are doing subtle tricks, and we should be very
> > > carefull.
> >
> > I thought that the trick was only there to work around a bug in older
> > versions of gcc. The proper use of volatile should be enough to get
> > the compiler to do the right thing, but at one point gcc didn't.
>
> AFAIK volatile is not enough to get _asm_ to do the right thing.
> volatile guarantees visibility, but in this case it's atomicity that has
> to be guaranteed.

Indeed. I was just discussing compiler issues, not asm atomicity
issues.

> If volatile causes asm "m" constraints to always use the original memory
> location, that would be quite useful but it is not documented to do so.

As far as I know, it doesn't and isn't.

However, the volatile qualifier is defined in the ISO C standard. If
gcc breaks the restrictions on what it is allowed to do with volatile
objects implied by that definition, then it has a bug. If the same
restrictions don't also apply for the C expressions in inline asm
constraints, then that would also be a bug, since it would be
worthless as a feature. But I haven't noticed any problems in this
area with recent versions of gcc/egcs.

David Wragg

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