Re: [PATCH 2/5] locking/lockdep: Eliminate redundant irqs check in __lock_acquire()

[Ingo Molnar]

* Waiman Long <longman@xxxxxxxxxx> wrote:

> The static __lock_acquire() function has only two callers:
> 
>  1) lock_acquire()
>  2) reacquire_held_locks()
> 
> In lock_acquire(), raw_local_irq_save() is called before hand. So
> irqs must have been disabled. So the check

s/before hand/
 /beforehand

s/irqs
 /IRQs

> 
> 	DEBUG_LOCKS_WARN_ON(!irqs_disabled())
> 
> is kind of redundant in thise case. So move the above check
> to reacquire_held_locks() to eliminate redundant code in the
> lock_acquire path.

s/thise
 /this

s/lock_acquire path
 /lock_acquire() path
 
> Signed-off-by: Waiman Long <longman@xxxxxxxxxx>
> ---
>  kernel/locking/lockdep.c | 15 +++++++--------
>  1 file changed, 7 insertions(+), 8 deletions(-)
> 
> diff --git a/kernel/locking/lockdep.c b/kernel/locking/lockdep.c
> index 8f9de7c..add0468 100644
> --- a/kernel/locking/lockdep.c
> +++ b/kernel/locking/lockdep.c
> @@ -3192,6 +3192,10 @@ void lockdep_init_map(struct lockdep_map *lock, const char *name,
>  /*
>   * This gets called for every mutex_lock*()/spin_lock*() operation.
>   * We maintain the dependency maps and validate the locking attempt:
> + *
> + * The callers must make sure that IRQs are disabled before calling it.
> + * otherwise we could get an interrupt which would want to take locks,
> + * which would end up in lockdep again.

Spelling nit: a comma after the first line, like it was in the original version:

> -	/*
> -	 * Lockdep should run with IRQs disabled, otherwise we could
> -	 * get an interrupt which would want to take locks, which would
> -	 * end up in lockdep and have you got a head-ache already?
> -	 */

Thanks,

	Ingo