Re: [RFC PATCH] Add split_lock

[Thomas Gleixner]

On Thu, Apr 08 2021 at 12:18, Peter Zijlstra wrote:
> On Thu, Apr 08, 2021 at 06:23:38AM +0100, Matthew Wilcox (Oracle) wrote:
>> bit_spinlocks are horrible on RT because there's absolutely nowhere
>> to put the mutex to sleep on.  They also do not participate in lockdep
>> because there's nowhere to put the map.
>> 
>> Most (all?) bit spinlocks are actually a split lock; logically they
>> could be treated as a single spinlock, but for performance, we want to
>> split the lock over many objects.  Introduce the split_lock as somewhere
>> to store the lockdep map and as somewhere that the RT kernel can put
>> a mutex.  It may also let us store a ticket lock for better performance
>> on non-RT kernels in the future, but I have left the current cpu_relax()
>> implementation intact for now.
>
> I think I like it, but I'm not sure it'll work for RT as is. It's a bit
> like qrwlock in that it only uses the internal (split) lock for
> contention, but that doesn't work for PI.
>
> I've not recently looked at RT, but I think they simply used to bloat a
> number of the data structures with a real lock. Sebastian and Thomas
> will know better.

Correct, but it depends on the nature of the bit spinlock and the lock
held times. Some of them we just keep as is as it's not worth the
trouble.

Thanks,

        tglx