Tim Waugh (firstname.lastname@example.org)
Wed, 2 Sep 1998 19:07:33 +0100 (BST)
On Mon, 31 Aug 1998, Philip Blundell wrote:
> >I think that with the current code it' s impossible that a deadlock
> >happnes, since currently every spinlock is used with the
> >spin_[un]lock_irqsave version and there is an unlock after every lock. It
> >works as it' s true that 2.0.x works in SMP machines.
> Since, in practice, the waitlist is the only area that's touched frequently, I
> don't think the lock is going to come under contention very much even with
> only a single shared lock. Maybe your patch is an improvement but I don't
> think this is the right time for it.
My main concern when I put the spinlock code in was to make sure that the
code was still safe. I was converting it from using cli and was still a
bit unsure about how to use spinlocks.
I think Andrea's changes are theoretically good, by making the locking
more finely grained, but I can't see any bugs being fixed.
-- To unsubscribe, send mail to: email@example.com --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:18:13 EST