Re: [LinuxBIOS] [PATCH][v3] Spinlock fixes

[Stefan Reinauer]

* Uwe Hermann  [070706 02:05]:
> Various spinlock-related cleanups:
> Revert back to spinlock_t (instead of struct spinlock). This is fine
> in this special case, as the contents of spinlock_t are not meant to
> ever be accessed directly (only by "accessor" functions).

NACK.

Why would that be required? The code in svn needs no fixing. Let's not
invent typedefs without any need for them
> Drop the spin_lock_string and spin_unlock_string macros, they're pretty
> useless as they're only used in a single place and a macro doesn't make
> this code any more readable, IMO.

NACK.

They are not useless at all as they are _the_ (one and only) spinlock
implementation on x86. The "only single place" is the declaration of the
x86 specific spin_lock and spin_unlock functions.