[Intel-gfx] [PATCH] drm/gem: support BO freeing without dev->struct_mutex

[Chris Wilson]

On Wed, Apr 27, 2016 at 02:12:46PM +0200, Daniel Vetter wrote:
> On Wed, Apr 27, 2016 at 12:58:51PM +0100, Chris Wilson wrote:
> > On Wed, Apr 27, 2016 at 01:50:00PM +0200, Daniel Vetter wrote:
> > > Finally all the core gem and a lot of drivers are entirely free of
> > > dev->struct_mutex depencies, and we can start to have an entirely
> > > lockless unref path.
> > > 
> > > To make sure that no one who touches the core code accidentally breaks
> > > existing drivers which still require dev->struct_mutex I've made the
> > > might_lock check unconditional.
> > > 
> > > While at it de-inline the ref/unref functions, they've become a bit
> > > too big.
> > > 
> > > v2: Make it not leak like a sieve.
> > > 
> > > v3: Review from Lucas:
> > > - drop != NULL in pointer checks.
> > > - fixup copypasted kerneldoc to actually match the functions.
> > > 
> > > Cc: Chris Wilson <chris at chris-wilson.co.uk>
> > > Cc: Alex Deucher <alexdeucher at gmail.com>
> > > Cc: Lucas Stach <l.stach at pengutronix.de>
> > > Reviewed-by: Lucas Stach <l.stach at pengutronix.de>
> > > Signed-off-by: Daniel Vetter <daniel.vetter at intel.com>
> > > ---
> > >  drivers/gpu/drm/drm_gem.c | 64 ++++++++++++++++++++++++++++++++++++++---------
> > >  include/drm/drmP.h        | 15 ++++++++---
> > >  include/drm/drm_gem.h     | 45 ++-------------------------------
> > >  3 files changed, 66 insertions(+), 58 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
> > > index 25dac31eef37..a4684a306c48 100644
> > > --- a/drivers/gpu/drm/drm_gem.c
> > > +++ b/drivers/gpu/drm/drm_gem.c
> > > @@ -788,16 +788,7 @@ drm_gem_object_release(struct drm_gem_object *obj)
> > >  }
> > >  EXPORT_SYMBOL(drm_gem_object_release);
> > >  
> > > -/**
> > > - * drm_gem_object_free - free a GEM object
> > > - * @kref: kref of the object to free
> > > - *
> > > - * Called after the last reference to the object has been lost.
> > > - * Must be called holding struct_ mutex
> > > - *
> > > - * Frees the object
> > > - */
> > > -void
> > > +static void
> > >  drm_gem_object_free(struct kref *kref)
> > >  {
> > >  	struct drm_gem_object *obj =
> > > @@ -806,10 +797,59 @@ drm_gem_object_free(struct kref *kref)
> > >  
> > >  	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
> > >  
> > > -	if (dev->driver->gem_free_object != NULL)
> > > +	if (dev->driver->gem_free_object_unlocked)
> > > +		dev->driver->gem_free_object_unlocked(obj);
> > > +	else if (dev->driver->gem_free_object)
> > >  		dev->driver->gem_free_object(obj);
> > >  }
> > > -EXPORT_SYMBOL(drm_gem_object_free);
> > 
> > With my plan, I want to keep drm_gem_object_free exported.
> > The plan being a __drm_gem_object_free_lockless fastpath.
> 
> Yeah, but I think we can reexport again in the patch that adds
> __drm_gem_object_free_lockless. Just not much personally a friend of
> sprawling EXPORT_SYMBOL when not needed. But I can drop that part too if
> you feel strongly.

The alternative is to add the function now and then I can just drop the
support into my tree and start testing :)
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre