Re: Could this become a race condition?
Matti Aarnio (matti.aarnio@tele.fi)
Thu, 17 Oct 1996 19:10:01 +0300 (EET DST)
Linus comments at my answer on the list:
> > .. is there a deadlock ? ..
> > ...
> > > Is there any way to avoid that?
> >
> > Yes. Yes:
> >
> > while (set_bit(1,&sb->s_lock) != 0)
> > __wait_on_super(sb);
> >
> > While the old value of the s_lock is non-zero, call
> > __wait_on_super(), else leave with s_lock set.
> > See <asm/bitops.h> for the routines.
>
> Actually, the old code is not a race condition, because "sb_lock" is only
> modified in a "process context" and never from interrupts. As such, the
> general non-re-entrancy of the kernel will guarantee that you can do
> "atomic" operations like the above without using the setbit() etc atomic
> stuff
Indeed, however I have understood that a long-term goal is to
get the SMP locks deeper/wider into the kernel from current
SINGLE spin-lock at system-call entrance. That way there can
be multiple processors running in kernel, and thus having a
need for this kind of atomic stuff.
> Linus
/Matti Aarnio <matti.aarnio@tele.fi>