[Intel-gfx] [PATCH] present: Avoid dereferencing vblank past a successful call to present_flip()

2014-05-31 Thread Chris Wilson
The driver may complete the flip instantaneously resulting in an immediate call to present_flip_notify(). This has the side-effect of destroying the vblank. We could add a reference count to the vblank structure to defer the destruction, or make the assume that further accesses are invalid and

[Intel-gfx] [PATCH 3/3] drm/i915: Use transcoder as index to MIPI regs

2014-05-31 Thread Shashank Sharma
Conceptually, the MIPI registers are addressed by the MIPI transcoder index, not the pipe. It doesn't matter right now, because there's a 1:1 relationship between pipes and MIPI transcoders, but that change allows us to break that link in the future V1: Created new patch to address Damien's

[Intel-gfx] [PATCH 2/3] drm/i915: Change Mipi register definitions

2014-05-31 Thread Shashank Sharma
Re-define MIPI register definitions in such a way that most of the existing DSI code can be re-used for future platforms. Register definitions are re-written using MMIO offset variable, so that without changing the existing sequence, same code can be generically applied. V4: Addressing review

Re: [Intel-gfx] [PATCH 2/3] drm/i915: Change Mipi register definitions

2014-05-31 Thread Damien Lespiau
On Sat, May 31, 2014 at 01:32:42PM +0530, Shashank Sharma wrote: Re-define MIPI register definitions in such a way that most of the existing DSI code can be re-used for future platforms. Register definitions are re-written using MMIO offset variable, so that without changing the existing

Re: [Intel-gfx] [PATCH 2/3] drm/i915: Change Mipi register definitions

2014-05-31 Thread Sharma, Shashank
Hi Damien, Please correct me if I am missing something, but the only reason we are seeing those extra formatting changes is, almost all of the old MIPI register definitions were beyond 80 char, and was checked in like that with warnings (How ?) So when I was checking for checkpatch errors, I