[Natty-SRU] Lenovo x420s panel display corruption when mode-setting

[Andy Whitcroft]

On Mon, Oct 03, 2011 at 01:19:55PM +0100, Andy Whitcroft wrote:

> Well its a vast and epic patch.  I am left wondering whether this is
> a backport or a cherry-pick in this context, below its reported as a
> backport which implies it was modified, but you say here "as-is" implying
> the opposite.
[...]
> > Subject: [PATCH] drm/i915: add pipe/plane enable/disable functions
> > 
> > Add plane enable/disable functions to prevent duplicated code and allow
> > us to easily check for plane enable/disable requirements (such as pipe
> > enable, plane status, pll status etc).
> 
> It is clear that this description is inadequate if we are actually
> changing pipe init order.  (Not that that is anything you should change
> as it _is_ the upstream commit description.)  Bad Jesse.
> 
> > 
> > Signed-off-by: Jesse Barnes <jbarnes at virtuousgeek.org>
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > 
> > (Backport to 2.6.38.y)
> 
> BugLink: http://bugs.launchpad.net/bugs/812638
> (backported from commit b24e71798871089da1a4ab049db2800afc1aac0c)
> 
> "backported from" if its modified, "cherry-picked from" if its not ... ?

Ok, having talked to Jose a bit on IRC I am now able to answer my own
question.  Actually this patch is the merge of _two_ upstream commits
only retaining one of the descriptions.

It is a merge of the two commits below:

  commit 65993d64a31844ad444694efb2d159eb9c883e49
  Author: Jesse Barnes <jbarnes at virtuousgeek.org>
  Date:   Tue Jan 4 15:09:29 2011 -0800

    drm/i915: don't enable plane, pipe and PLL prematurely
    
    On Ironlake+ we need to enable these in a specific order.
    
    Signed-off-by: Jesse Barnes <jbarnes at virtuousgeek.org>
    Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>

and:

  commit b24e71798871089da1a4ab049db2800afc1aac0c
  Author: Jesse Barnes <jbarnes at virtuousgeek.org>
  Date:   Tue Jan 4 15:09:30 2011 -0800

    drm/i915: add pipe/plane enable/disable functions
    
    Add plane enable/disable functions to prevent duplicated code and allow
    us to easily check for plane enable/disable requirements (such as pipe
    enable, plane status, pll status etc).
    
    Signed-off-by: Jesse Barnes <jbarnes at virtuousgeek.org>
    Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>

If we try applying these onto the current natty head the first is a clean
cherry-pick, the second collised in context _only_ with the patch below,
keeping both sides gives us the combined commit on this thread:

  commit bbeaf8811ba070fd186dfcabc957044c3a1149ac
  Author: Keng-Yu Lin <keng-yu.lin at canonical.com>
  Date:   Wed Jun 22 09:49:34 2011 +0800

    drm/i915: disable PCH ports if needed when disabling a CRTC

For me this should have been submitted as two patched as clearly the
change in the first one is a key component.  From these discusssions it
seems that this improves things from 100% failure to ~10% failure on
boot.  It looks a lot like the changes in the second patch, such as
waiting for vblank on enabling the pipes fixes that final 10%.

-apw