[Intel-gfx] [PATCH] drm/i915/fbdev: Enable late fbdev initial configuration

[Chris Wilson]

Quoting Souza, Jose (2018-04-18 17:42:47)
> On Wed, 2018-04-18 at 09:26 +0100, Chris Wilson wrote:
> > Quoting Souza, Jose (2018-04-18 01:07:16)
> > > On Wed, 2018-04-18 at 00:44 +0100, Chris Wilson wrote:
> > > > Quoting José Roberto de Souza (2018-04-17 23:34:18)
> > > > > If the initial fbdev
> > > > > configuration(intel_fbdev_initial_config())
> > > > > runs and
> > > > > there still no sink connected it will cause
> > > > > drm_fb_helper_initial_config() to return 0 as no error
> > > > > happened(but
> > > > > internally the return is -EAGAIN).
> > > > > Because no framebuffer was allocated, when a sink is connected
> > > > > intel_fbdev_output_poll_changed() will not execute
> > > > > drm_fb_helper_hotplug_event() that would trigger another try to
> > > > > do
> > > > > the
> > > > > initial fbdev configuration.
> > > > > 
> > > > > So here creating a dummy framebuffer of 800x600, so
> > > > > drm_fb_helper_hotplug_event() will be executed and fbdev can be
> > > > > properly
> > > > > setup when a sink is connected, if needed the dummy framebuffer
> > > > > will be
> > > > > freed and a new with the proper size will be allocated.
> > > > > 
> > > > > This issue also happens when a MST DP sink is connected since
> > > > > boot,
> > > > > as
> > > > > the MST topology is discovered in parallel if
> > > > > intel_fbdev_initial_config()
> > > > > is executed before the first sink MST is discovered it will
> > > > > cause
> > > > > this
> > > > > same issue.
> > > > > 
> > > > > This is a follow up patch of
> > > > > https://patchwork.freedesktop.org/patch/196089/
> > > > > 
> > > > > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104158
> > > > > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104425
> > > > > Cc: Chris Wilson <chris at chris-wilson.co.uk>
> > > > > Cc: Rodrigo Vivi <rodrigo.vivi at intel.com>
> > > > > Signed-off-by: José Roberto de Souza <jose.souza at intel.com>
> > > > > ---
> > > > >  drivers/gpu/drm/i915/intel_fbdev.c | 24
> > > > > +++++++++++++++++++++++-
> > > > >  1 file changed, 23 insertions(+), 1 deletion(-)
> > > > > 
> > > > > diff --git a/drivers/gpu/drm/i915/intel_fbdev.c
> > > > > b/drivers/gpu/drm/i915/intel_fbdev.c
> > > > > index 7d41d139341b..773577d39782 100644
> > > > > --- a/drivers/gpu/drm/i915/intel_fbdev.c
> > > > > +++ b/drivers/gpu/drm/i915/intel_fbdev.c
> > > > > @@ -696,14 +696,36 @@ int intel_fbdev_init(struct drm_device
> > > > > *dev)
> > > > >         return 0;
> > > > >  }
> > > > >  
> > > > > +static void intel_fbdev_dummy_fb_create(struct intel_fbdev
> > > > > *ifbdev)
> > > > > +{
> > > > > +       struct drm_fb_helper_surface_size sizes;
> > > > > +
> > > > > +       sizes.fb_width = 800;
> > > > > +       sizes.fb_height = 600;
> > > > > +       sizes.surface_width = sizes.fb_width;
> > > > > +       sizes.surface_height = sizes.fb_height;
> > > > > +       sizes.surface_bpp = 32;
> > > > > +       sizes.surface_depth = 24;
> > > > > +
> > > > > +       if (intelfb_create(&ifbdev->helper, &sizes))
> > > > > +               DRM_ERROR("Unable to create dummy framebufer");
> > > > > +}
> > > > > +
> > > > >  static void intel_fbdev_initial_config(void *data,
> > > > > async_cookie_t
> > > > > cookie)
> > > > >  {
> > > > >         struct intel_fbdev *ifbdev = data;
> > > > >  
> > > > >         /* Due to peculiar init order wrt to hpd handling this
> > > > > is
> > > > > separate. */
> > > > >         if (drm_fb_helper_initial_config(&ifbdev->helper,
> > > > > -                                        ifbdev-
> > > > > >preferred_bpp))
> > > > > +                                        ifbdev-
> > > > > >preferred_bpp)) {
> > > > >                 intel_fbdev_unregister(to_i915(ifbdev-
> > > > > > helper.dev));
> > > > > 
> > > > > +               return;
> > > > > +       }
> > > > > +
> > > > > +       mutex_lock(&ifbdev->helper.lock);
> > > > > +       if (!ifbdev->vma)
> > > > > +               intel_fbdev_dummy_fb_create(ifbdev);
> > > > > +       mutex_unlock(&ifbdev->helper.lock);
> > > > >  }
> > > > 
> > > > Did you try
> > > > 
> > > > diff --git a/drivers/gpu/drm/i915/intel_fbdev.c
> > > > b/drivers/gpu/drm/i915/intel_fbdev.c
> > > > index 65a3313723c9..4120c635742d 100644
> > > > --- a/drivers/gpu/drm/i915/intel_fbdev.c
> > > > +++ b/drivers/gpu/drm/i915/intel_fbdev.c
> > > > @@ -493,6 +493,7 @@ static bool intel_fb_initial_config(struct
> > > > drm_fb_helper *fb_helper,
> > > >  bail:
> > > >                 DRM_DEBUG_KMS("Not using firmware
> > > > configuration\n");
> > > >                 memcpy(enabled, save_enabled, count);
> > > > +               fb_helper->deferred_setup = true;
> > > 
> > > This is too earlier to set deferred_setup when
> > > intel_fb_initial_config() fails,
> > > __drm_fb_helper_initial_config_and_unlock() will then try to
> > > probe/create a fb but as there is no sink connected it will fail
> > > too.
> > > 
> > > Also __drm_fb_helper_initial_config_and_unlock() is already setting
> > > deferred_setup if both attempts fails.
> > 
> > So...
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_fbdev.c
> > b/drivers/gpu/drm/i915/intel_fbdev.c
> > index 7d41d139341b..5f138a03dd2a 100644
> > --- a/drivers/gpu/drm/i915/intel_fbdev.c
> > +++ b/drivers/gpu/drm/i915/intel_fbdev.c
> > @@ -807,7 +807,7 @@ void intel_fbdev_output_poll_changed(struct
> > drm_device *dev)
> >                 return;
> >  
> >         intel_fbdev_sync(ifbdev);
> > -       if (ifbdev->vma)
> > +       if (ifbdev->vma || ifbdev->helper->deferred_setup)
> >                 drm_fb_helper_hotplug_event(&ifbdev->helper);
> >  }
> 
> That would work but according to your comment here: https://patchwork.f
> reedesktop.org/patch/196089/ that would lead to a crash in some case.

In the deferred_setup case, won't it go to intelfb_create() and after
that point if we still have no ifdev->vma we won't hit any other pathway?

Just because it's been historically buggy (at least 2 instances now of
having to apply protection here), it may have been fixed... Unlikely,
but maybe it all finally works.
-Chris