[Intel-gfx] [PATCH] drm/vgem: Allow root to inject hanging fences onto dmabufs

[Chris Wilson]

On Fri, Nov 18, 2016 at 10:40:02AM +0100, Daniel Vetter wrote:
> On Fri, Nov 18, 2016 at 08:49:37AM +0000, Chris Wilson wrote:
> > On Wed, Jul 20, 2016 at 08:39:43PM +0100, Chris Wilson wrote:
> > > When performing driver testing, one factor we want to test is how we
> > > handle a foreign fence that is never signaled. We can wait on that fence
> > > indefinitely, in which case the driver appears hung, or we can take some
> > > remedial action (with risks regarding the state of any shared content).
> > > Whatever the action choosen by the driver, in order to perform the test
> > > we want to disable the safety feature of vgem fence (which is then used
> > > to test implicit dma-buf fencing). This is regarded as a highly
> > > dangerous feature and so hidden behind an expert config option and only
> > > available to root when enabled.
> > > 
> > > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > 
> > Ping. This is invaluable for my testing (some tests are impossible
> > without it).
> 
> Hm, I thought I voted for a module option (so that tests can frob at
> runtime) that auto-taints the kernel. And then maybe always allow it. I
> just don't like when Kconfig settings can result in testcase skips, makes
> it uncertain whether QA did QA or not.

We know when it is disabled, and so when it skips. We also control the
kconfig set by QA. There was very strong objections to having yet
another dangling fence that I thought making it available in default
builds was not desired.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre