Re: [PATCHv2] bitops: add _local bitops

[Avi Kivity <avi@redhat.com>]

On 05/13/2012 01:45 PM, Michael S. Tsirkin wrote:
> On Sun, May 13, 2012 at 01:13:21PM +0300, Avi Kivity wrote:
> > On 05/10/2012 10:04 PM, Christoph Lameter wrote:
> > > On Wed, 9 May 2012, Michael S. Tsirkin wrote:
> > >
> > > > kvm needs to update some hypervisor variables atomically
> > > > in a sense that the operation can't be interrupted
> > > > in the middle. However the hypervisor always runs
> > > > on the same CPU so it does not need any memory
> > > > barrier or lock prefix.
> > > >
> > > > Add _local bitops for this purpose: define them
> > > > as non-atomics for x86 and (for now) atomics for
> > > > everyone else.
> > >
> > > Have you tried to use the this_cpu_ops for that purpose? They create the
> > > per cpu atomic instructions that you want without a lock prefix and can
> > > also relocate the per cpu pointer to the correct processor via a
> > > segment register prefix.
> > >
> > > There are no bit operations provided right now but those can either be
> > > improvised using this_cpu_cmpxchg or added.
> > 
> > this_cpu_xchg() should be sufficient, since only bit zero has any
> > meaning in our use case (so xchg with zero is equivalent to
> > test_and_clear_bit).
>
> Yes it should work. No idea how it'd perform:
> arch/x86/include/asm/percpu.h implies it's expensive.  

It says a naive implementation is expensive, whereas the implementation
actually used (a cmpxchg loop) is better.

> My latest version
> simply documents what __test_and_clear does anyway.
> Which was indicated is acceptable ...
> Did you change your mind?

No.

-- 
error compiling committee.c: too many arguments to function