Re: Negative scalability by removal of lock_kernel()?(Was: Strange performance behavior of 2.4.0-test9)

From: Andrew Morton (andrewm@uow.edu.au)
Date: Fri Oct 27 2000 - 05:23:56 EST

Andi Kleen wrote:
>
> When you have two CPUs contending on common paths it is better to do:
> [ spinlock stuff ]

Andi, if the lock_kernel() is removed then the first time the CPUs will butt heads is on a semaphore. This is much more expensive.

I bet if acquire_fl_sem() and release_fl_sem() are turned into lock_kernel()/unlock_kernel() then the scalability will come back.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

This archive was generated by hypermail 2b29 : Tue Oct 31 2000 - 21:00:20 EST