Re: [PATCH 3.13.y-ckt 019/103] drm/i915: Swap primary planes on gen2 for FBC

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Thu, Feb 19, 2015 at 10:55:42AM -0800, Kamal Mostafa wrote:
> On Thu, 2015-02-19 at 11:31 +0200, Ville Syrjälä wrote:
> > On Wed, Feb 18, 2015 at 04:31:46PM -0800, Kamal Mostafa wrote:
> > > 3.13.11-ckt16 -stable review patch.  If anyone has any objections, please let me know.
> > > 
> > > ------------------
> > > 
> > > From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
> > > 
> > > commit 1f1c2e2468f937cefd6bcb645c959c7b5d9821df upstream.
> > > 
> > > Only plane A is FBC capable on gen2 (like gen3), but the panel fitter
> > > is hooked up to pipe B, so we want to prefer pipe B + plane A.
> > 
> > FBC doesn't even work on gen2 (still missing my stolen mem patches) so
> > this patch doesn't really seem like stable material to me. But I take
> > it you backported it to get the prepare_page_flip() patch to apply cleanly?
> 
> Yes, exactly (and I forgot to note that in my commit message).
> 
> > If you really want to do that, then you should also backport
> > 
> > commit 38af6096781e2ed3d83b970218dedfe98e2f8f9c
> > Author: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > Date:   Tue Jan 7 16:15:37 2014 +0200
> > 
> >     drm/i915: Don't swap planes on 830M
> 
> I will pick up that one too.  So this 3.13-stable release will include
> all three of these, barring any further objections:
> 
> 7d47559 drm/i915: Don't call intel_prepare_page_flip() multiple times on gen2-4
> 38af609 drm/i915: Don't swap planes on 830M
> 1f1c2e2 drm/i915: Swap primary planes on gen2 for FBC
> 
> > othwerwise you may have some angry 830 users knocking at the door,
> > assuming there are more than two of us :) Although 830 support has
> > been more or less broken since forever until 3.18 so maybe no one
> > would notice anyway.
> 
> Thanks very much for the guidance, Ville.

No problem.

When fixing a regression we do try to identify the bad commit in the
commit message of the fix. So when considering backporting a commit, I
think it would be a good idea to grep more recent commit messages for
its hash, just in case it has caused some regression.

> 
>  -Kamal
> 
> 
> > 
> > > 
> > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
> > > [danvet: Add the code comment Chris requested in his review.]
> > > Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> > > 
> > > Signed-off-by: Kamal Mostafa <kamal@canonical.com>
> > > ---
> > >  drivers/gpu/drm/i915/i915_irq.c      | 14 +++++++++-----
> > >  drivers/gpu/drm/i915/intel_display.c |  7 +++++--
> > >  2 files changed, 14 insertions(+), 7 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> > > index 3a10da2..5e0b5e4 100644
> > > --- a/drivers/gpu/drm/i915/i915_irq.c
> > > +++ b/drivers/gpu/drm/i915/i915_irq.c
> > > @@ -3155,10 +3155,10 @@ static int i8xx_irq_postinstall(struct drm_device *dev)
> > >   * Returns true when a page flip has completed.
> > >   */
> > >  static bool i8xx_handle_vblank(struct drm_device *dev,
> > > -			       int pipe, u16 iir)
> > > +			       int plane, int pipe, u32 iir)
> > >  {
> > >  	drm_i915_private_t *dev_priv = dev->dev_private;
> > > -	u16 flip_pending = DISPLAY_PLANE_FLIP_PENDING(pipe);
> > > +	u16 flip_pending = DISPLAY_PLANE_FLIP_PENDING(plane);
> > >  
> > >  	if (!drm_handle_vblank(dev, pipe))
> > >  		return false;
> > > @@ -3166,7 +3166,7 @@ static bool i8xx_handle_vblank(struct drm_device *dev,
> > >  	if ((iir & flip_pending) == 0)
> > >  		return false;
> > >  
> > > -	intel_prepare_page_flip(dev, pipe);
> > > +	intel_prepare_page_flip(dev, plane);
> > >  
> > >  	/* We detect FlipDone by looking for the change in PendingFlip from '1'
> > >  	 * to '0' on the following vblank, i.e. IIR has the Pendingflip
> > > @@ -3235,9 +3235,13 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg)
> > >  			notify_ring(dev, &dev_priv->ring[RCS]);
> > >  
> > >  		for_each_pipe(pipe) {
> > > +			int plane = pipe;
> > > +			if (IS_MOBILE(dev))
> > > +				plane = !plane;
> > > +
> > >  			if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS &&
> > > -			    i8xx_handle_vblank(dev, pipe, iir))
> > > -				flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(pipe);
> > > +			    i8xx_handle_vblank(dev, plane, pipe, iir))
> > > +				flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(plane);
> > >  
> > >  			if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
> > >  				i9xx_pipe_crc_irq_handler(dev, pipe);
> > > diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> > > index 2d2c1e1..ae3bc77 100644
> > > --- a/drivers/gpu/drm/i915/intel_display.c
> > > +++ b/drivers/gpu/drm/i915/intel_display.c
> > > @@ -9949,10 +9949,13 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
> > >  		intel_crtc->lut_b[i] = i;
> > >  	}
> > >  
> > > -	/* Swap pipes & planes for FBC on pre-965 */
> > > +	/*
> > > +	 * On gen2/3 only plane A can do fbc, but the panel fitter and lvds port
> > > +	 * is hooked to plane B. Hence we want plane A feeding pipe B.
> > > +	 */
> > >  	intel_crtc->pipe = pipe;
> > >  	intel_crtc->plane = pipe;
> > > -	if (IS_MOBILE(dev) && IS_GEN3(dev)) {
> > > +	if (IS_MOBILE(dev) && INTEL_INFO(dev)->gen < 4) {
> > >  		DRM_DEBUG_KMS("swapping pipes & planes for FBC\n");
> > >  		intel_crtc->plane = !pipe;
> > >  	}
> > > -- 
> > > 1.9.1
> > 
> 

-- 
Ville Syrjälä
Intel OTC