[Intel-gfx] [PATCH] drm/i915/dsi: Silence atomic update failure with DSI panel

[Daniel Vetter]

On Wed, Sep 06, 2017 at 07:48:33PM +0300, Martin Peres wrote:
> On 06/09/17 13:09, Mika Kahola wrote:
> > On Tue, 2017-09-05 at 18:11 +0200, Daniel Vetter wrote:
> > > On Tue, Sep 05, 2017 at 04:35:04PM +0300, Mika Kahola wrote:
> > > > 
> > > > It appears that we cannot trust scanline counters when MIPI/DSI
> > > > display is
> > > > connected. In CI system this appears as flickering errors that
> > > > randomly
> > > > appear in test cases. To avoid this flickering, let's just silence
> > > > atomic
> > > > update failure in case with DSI panel.
> > > > 
> > > > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102403
> > > > Signed-off-by: Mika Kahola <mika.kahola at intel.com>
> > > This just changes a loud atomic failure to a silent atomic failure.
> > > What
> > > we instead need to do is actually fix the bug, not hide it.
> > > 
> > BSpec has a notion that (PIPE_SCANLINE) that is "Not supported with
> > MIPI DSI."
> > 
> > That's why I thought it might be ok to silence the error as the
> > computation that we try to accomplish wouldn't work anyway. Maybe this
> > way we could remove DSI from being blackllisted.
> 
> I agree. If the HW can't do it, so what can we do here?
> 
> As long as this is well documented and the userspace knows about this issue
> (if anything relies on this feature), then what else can we do?
> 
> With the relevant BSpec quotes added above the changes, I can give my:
> Acked-by: Martin Peres <martin.peres at linux.intel.com>
> 
> I would however like to know if this breaks any feature the userspace relies
> on.

Catching up on this, for the record:

Lack of this breaks atomic. So yeah, not really an optional feature, and
definitely not a failure mode we should flush out to hide it.

And if the scanline trick doesn't work, then we need a different way to
achieve the same. Just because the one way we use everywhere else doesn't
work, doesn't mean atomic on dsi is going to be impossible. It would be
real bad if that's the case.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch