On Wed, Dec 20, 2017 at 12:17:27PM +0100, Martin Pieuchot wrote:
> On 15/12/17(Fri) 22:03, Mateusz Guzik wrote:
> > > +void
> > > +__mtx_enter(struct mutex *mtx)
> > > +{
> > > +#ifdef MP_LOCKDEBUG
> > > + int nticks = __mp_lock_spinout;
> > > +#endif
> > > +
> > > + while
On 15/12/17(Fri) 22:03, Mateusz Guzik wrote:
> [...]
> However, contended behaviour is a regression compared to the asm
> variant.
Now that I checked the files in could you generate a diff with your
suggestions?
> From what I gather this is a step towards unifying all mutex
> implementations,
> Date: Mon, 18 Dec 2017 10:08:23 +0100
> From: Martin Pieuchot
>
> On 14/12/17(Thu) 16:06, Martin Pieuchot wrote:
> > Diff below moves amd64 and i386 mutex to the common C implementation.
> >
> > The differences are:
> > - membar_enter_after_atomic(9) instead of
On 14/12/17(Thu) 16:06, Martin Pieuchot wrote:
> Diff below moves amd64 and i386 mutex to the common C implementation.
>
> The differences are:
> - membar_enter_after_atomic(9) instead of membar_enter(9), and
> - membar_exit_before_atomic(9) instead of membar_exit(9)
>
> I'd appreciate any
On Thu, Dec 14, 2017 at 04:06:41PM +0100, Martin Pieuchot wrote:
> Diff below moves amd64 and i386 mutex to the common C implementation.
>
> The differences are:
> - membar_enter_after_atomic(9) instead of membar_enter(9), and
> - membar_exit_before_atomic(9) instead of membar_exit(9)
>
> I'd