[Intel-gfx] [PATCH 06/15] drm/i915: Assert no external observers when unwind a failed request alloc

[Chris Wilson]

On Fri, Nov 25, 2016 at 12:39:25PM +0200, Joonas Lahtinen wrote:
> On pe, 2016-11-25 at 09:30 +0000, Chris Wilson wrote:
> > Before we return the request back to the kmem_cache after a failed
> > i915_gem_request_alloc(), we should assert that it has not been added to
> > any global state tracking.
> > 
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > ---
> >  drivers/gpu/drm/i915/i915_gem_request.c | 5 +++++
> >  1 file changed, 5 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/i915/i915_gem_request.c b/drivers/gpu/drm/i915/i915_gem_request.c
> > index bd7b21f70283..ed6cead22a86 100644
> > --- a/drivers/gpu/drm/i915/i915_gem_request.c
> > +++ b/drivers/gpu/drm/i915/i915_gem_request.c
> > @@ -599,6 +599,11 @@ i915_gem_request_alloc(struct intel_engine_cs *engine,
> > >  	return req;
> >  
> >  err_ctx:
> > +	/* Make sure we didn't add ourselves to external state before freeing */
> > +	GEM_BUG_ON(!list_empty(&req->active_list));
> > +	GEM_BUG_ON(!list_empty(&req->priotree.signalers_list));
> > +	GEM_BUG_ON(!list_empty(&req->priotree.waiters_list));
> 
> This is some paranoia already, but if you managed to hit it;

Fortunately not. I scoured the init_context to make sure we weren't
triggering it - but it is a bug we have had in the past and the reason
why requests are so tricky to construct so I thought having some
paranoia was worth the documentation.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre