[PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Dave Airlie
On Mon, Sep 10, 2012 at 9:07 PM, Alan Cox wrote: > On Mon, 10 Sep 2012 14:31:52 +1000 > Dave Airlie wrote: > >> From: Dave Airlie >> >> For secondary GPUs in laptops, i.e. optimus or powerxpress, we have >> methods for powering down the GPU completely. This adds support >> to the drm core for po

[PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Dave Airlie
(oops' forgot reply to all) > > My midlayer-smell-o-meter just cranked up to 11 when reading this comment > ;-) > > I'd have expected: > - Drivers to check the power state and enable the gpu if it's off in their > cs ioctl (instead of the brute-force every ioctl there is approach in > the drm

[PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Dave Airlie
From: Dave Airlie For secondary GPUs in laptops, i.e. optimus or powerxpress, we have methods for powering down the GPU completely. This adds support to the drm core for powering back up the GPU on any access from ioctls or sysfs interfaces, and fires a 5s timer to test if we can power the GPU of

[PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Alan Cox
On Mon, 10 Sep 2012 14:31:52 +1000 Dave Airlie wrote: > From: Dave Airlie > > For secondary GPUs in laptops, i.e. optimus or powerxpress, we have > methods for powering down the GPU completely. This adds support > to the drm core for powering back up the GPU on any access from > ioctls or sysfs

[PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Alan Cox
> fine grained suspend/resume. Not sure if that concept helps here, but it > may be worth digging around to see how they went about waking up > individual devices. Badly. I don't believe the code ever worked properly. It certainly was full of races. I've reworked chunks of it in the GMA500 oaktrai

[PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Daniel Vetter
On Mon, Sep 10, 2012 at 10:23 AM, Dave Airlie wrote: > (oops' forgot reply to all) > >> >> My midlayer-smell-o-meter just cranked up to 11 when reading this comment >> ;-) >> >> I'd have expected: >> - Drivers to check the power state and enable the gpu if it's off in their >> cs ioctl (instead

[PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Chris Wilson
On Mon, 10 Sep 2012 18:23:05 +1000, Dave Airlie wrote: > (oops' forgot reply to all) > > > > > My midlayer-smell-o-meter just cranked up to 11 when reading this comment > > ;-) > > > > I'd have expected: > > - Drivers to check the power state and enable the gpu if it's off in their > > cs ioctl

[PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Daniel Vetter
On Mon, Sep 10, 2012 at 02:31:52PM +1000, Dave Airlie wrote: > From: Dave Airlie > > For secondary GPUs in laptops, i.e. optimus or powerxpress, we have > methods for powering down the GPU completely. This adds support > to the drm core for powering back up the GPU on any access from > ioctls or

Re: [PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Dave Airlie
On Mon, Sep 10, 2012 at 9:07 PM, Alan Cox wrote: > On Mon, 10 Sep 2012 14:31:52 +1000 > Dave Airlie wrote: > >> From: Dave Airlie >> >> For secondary GPUs in laptops, i.e. optimus or powerxpress, we have >> methods for powering down the GPU completely. This adds support >> to the drm core for po

Re: [PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Alan Cox
On Mon, 10 Sep 2012 14:31:52 +1000 Dave Airlie wrote: > From: Dave Airlie > > For secondary GPUs in laptops, i.e. optimus or powerxpress, we have > methods for powering down the GPU completely. This adds support > to the drm core for powering back up the GPU on any access from > ioctls or sysfs

Re: [PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Alan Cox
> fine grained suspend/resume. Not sure if that concept helps here, but it > may be worth digging around to see how they went about waking up > individual devices. Badly. I don't believe the code ever worked properly. It certainly was full of races. I've reworked chunks of it in the GMA500 oaktrai

Re: [PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Daniel Vetter
On Mon, Sep 10, 2012 at 10:23 AM, Dave Airlie wrote: > (oops' forgot reply to all) > >> >> My midlayer-smell-o-meter just cranked up to 11 when reading this comment >> ;-) >> >> I'd have expected: >> - Drivers to check the power state and enable the gpu if it's off in their >> cs ioctl (instead

Re: [PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Chris Wilson
On Mon, 10 Sep 2012 18:23:05 +1000, Dave Airlie wrote: > (oops' forgot reply to all) > > > > > My midlayer-smell-o-meter just cranked up to 11 when reading this comment > > ;-) > > > > I'd have expected: > > - Drivers to check the power state and enable the gpu if it's off in their > > cs ioctl

Re: [PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Dave Airlie
(oops' forgot reply to all) > > My midlayer-smell-o-meter just cranked up to 11 when reading this comment > ;-) > > I'd have expected: > - Drivers to check the power state and enable the gpu if it's off in their > cs ioctl (instead of the brute-force every ioctl there is approach in > the drm

Re: [PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-10 Thread Daniel Vetter
On Mon, Sep 10, 2012 at 02:31:52PM +1000, Dave Airlie wrote: > From: Dave Airlie > > For secondary GPUs in laptops, i.e. optimus or powerxpress, we have > methods for powering down the GPU completely. This adds support > to the drm core for powering back up the GPU on any access from > ioctls or

[PATCH 2/5] drm: Add initial dnyamic power off feature

2012-09-09 Thread Dave Airlie
From: Dave Airlie For secondary GPUs in laptops, i.e. optimus or powerxpress, we have methods for powering down the GPU completely. This adds support to the drm core for powering back up the GPU on any access from ioctls or sysfs interfaces, and fires a 5s timer to test if we can power the GPU of