Re: [linux-usb-devel] Re: [PATCH 00/10] usb-serial: Switches fromspin lock to atomic_t.

[Arjan van de Ven]

On Wed, 2005-12-07 at 11:01 -0500, Alan Stern wrote:
> On Wed, 7 Dec 2005, Arjan van de Ven wrote:
> 
> > > On the other hand, Oliver needs to be careful about claiming too much.  In 
> > > general atomic_t operations _are_ superior to the spinlock approach.
> > 
> > No they're not. Both are just about equally expensive cpu wise,
> > sometimes the atomic_t ones are a bit more expensive (like on parisc
> > architecture). But on x86 in either case it's a locked cycle, which is
> > just expensive no matter which side you flip the coin... 
> 
> You're overgeneralizing.

to some degree yes.

> 
> Sure, a locked cycle has a certain expense.  But it's a lot less than the 
> expense of a contested spinlock. 

the chances that *this* spinlock ends up being contested are near zero,
and.. in that scenario a locked cycle does the same thing, just in
hardware..... (eg the other cpu will busy wait until this locked cycle
is done)

-
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/