[Intel-gfx] [PATCH] drm/i915/guc: Make sure vma containing firmware is GuC mappable

[Chris Wilson]

On Thu, Jan 12, 2017 at 12:44:52PM +0200, Tomi Sarvela wrote:
> On Thursday, 12 January 2017 10:33:47 EET Chris Wilson wrote:
> > On Wed, Jan 11, 2017 at 04:17:39PM +0100, Michał Winiarski wrote:
> > > Since commit 4741da925fa3 ("drm/i915/guc: Assert that all GGTT
> > > offsets used by the GuC are mappable"), we're asserting that GuC
> > > firmware is in the GuC mappable range.
> > > Except we're not pinning the object with bias, which means it's
> > > possible to trigger this assert. Let's add a proper bias.
> > > 
> > > Cc: Chris Wilson <chris at chris-wilson.co.uk>
> > > Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio at intel.com>
> > > Signed-off-by: Michał Winiarski <michal.winiarski at intel.com>
> > 
> > My spider sense is tingling....
> > 
> > CI has been reporting a suspend freeze with guc enabled. Tomi?
> 
> DRM-Tip kernels on Skylake from commit 8e6dfee and onwards
> (drm-tip: 2017y-01m-05d-16h-48m-32s UTC integration manifest)
> get into infinite suspend with following options turned on:
> i915.enable_guc_submission=1 i915.enable_guc_loading=1
> 
> With this patch the testhost i6700K survived dozen rounds of
> igt at gem_exec_suspend@basic-s3, without it the first run hangs the 
> machine.
> 
> Tested-by: Tomi Sarvela <tomi.p.sarvela at intel.com>

Thanks for the patch Michał, and for Tomi patiently proving how badly I
broke everything. Pushed,
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre