Re: [PATCH][ATM] assorted he driver cleanup

[chas williams]

In message <[e-mail]>,Arnaldo Carvalho de Melo w
rites:
>no, no, I was talking just about the need for HE_SPIN_LOCK wrapper, not the
>locking, i.e. couldn't it be just:
>
>spin_lock_irqsave(&dev->global_lock, flags)
>
>used so that it is clear that it is a irqsave variation, etc?

i suppose i could change it all back.  at one point, he_spin_lock()
was 'optmized' away on non smp machines (since a single cpu doesnt
tickle the particular h/w problem). 

i didnt want a ton of #ifdef CONFIG_SMP in the driver.