[Intel-gfx] [PATCH 0/5] Add power feature debugfs disabling

[Daniel Vetter]

On Sat, Feb 01, 2014 at 05:14:22PM +0000, Chris Wilson wrote:
> On Fri, Jan 31, 2014 at 03:42:47PM -0600, jeff.mcgee at intel.com wrote:
> > From: Jeff McGee <jeff.mcgee at intel.com>
> > 
> > This series has recently been accepted into the Haswell Android kernel and
> > helps with debugging and profiling these power features. I would like it
> > to be considered for upstream incorporation. The patches here have been
> > rebased (minimal changes required) and compile-tested only.
> > 
> > Broad device support is provided, accept for RPS and RC6 with Broadwell
> > and Valleyview. Both of these were somewhat of a moving target and I
> > didn't have devices to work with. Support can of course be added with
> > help from appropriate folks.
> > 
> > The hooks introduce some amount of overhead as an additional check is
> > often needed to determine whether the feature is on or off - similar to
> > the module parameters that already exist. I felt that the overhead was
> > minimal enough and didn't want to ugly up the code with CONFIG_DEBUG_FS
> > compile conditionals. But I'm open to the list's thoughts on this.
> > 
> > IGT tests of these new interfaces can certainly be added. I wanted to
> > make sure there was sufficient interest in having these interfaces before
> > starting on the tests. So please provide feedback.
> 
> I can see the value of adding this for power testing and the code looks
> quite neat and self-contained. (The one bikeshed I have is that I would
> like the parameter check and debug.disable check combined into a single
> function call, similar to intel_enable_rc6() so that all the similar
> logic is together, well commented and easy to verify, and hard for
> callers to get wrong.) Longer term, should we not consider this for
> our /sys/drm/card0/power API? i.e. do we see value beyond debugging and
> testing?

For fbc/psr/ips I'd even use a connector/crtc property if we really need
to frob this. Would be a good excuse for some better kms cmdline
utilities ...
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch