Re: [PATCH for 2.5] preemptible kernel

[george anzinger]

"David S. Miller" wrote:
> 
> Keith Owens writes:
>  > Or have I missed something?
> 
> Nope, it is a fundamental problem with such kernel pre-emption
> schemes.  As a result, it would also break our big-reader locks
> (see include/linux/brlock.h).

He has this one covered.  The patch puts preemption locks around
read_locks.

By the by, if a preemption lock is all that is needed the patch defines
it and it is rather fast (an inc going in and a dec & test comming
out).  A lot faster than a spin lock with its "LOCK" access.  A preempt
lock does not need to be "LOCK"ed because the only contender is the same
cpu.

George

> 
> Basically, anything which uses smp_processor_id() would need to
> be holding some lock so as to not get pre-empted.
> 
> Later,
> David S. Miller
> [e-mail]