Re: fs/locks.c BKL removal

Matthew Wilcox (willy@debian.org)
Sat, 11 May 2002 20:48:20 +0100


On Fri, May 10, 2002 at 04:17:29PM -0700, Andrew Morton wrote:
> It was I who put the BKL back into locks.c, much to
> Matthew's disgust...

The disgust was targetted more at removing the abstraction of
locking scheme which I'd put in and having explicit lock_kernel() /
unlock_kernel() calls. I'd used (iirc) acquire_lock() / release_lock()
macros which could have just been redefined.

> The problem was that replacing the BKL with a semaphore
> seriously damaged Apache thoughput on 8-way. Apache
> was using flock()-based synchronisation and replacing
> a spin with a schedule just killed it.

Which says that our semaphores suck, because they don't try to spin for a
bit before scheduling. Of course, your change back was the right thing
to do in the 2.3.late timeframe.

-- 
Revolutions do not require corporate support.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/