[PATCH weston v6 69/73] compositor-drm: head attach requires a modeset

[Pekka Paalanen]

On Wed, 18 Apr 2018 12:09:45 +0300
Pekka Paalanen <ppaalanen at gmail.com> wrote:

> On Thu, 12 Apr 2018 14:55:01 +0200
> Daniel Stone <daniel at fooishbar.org> wrote:
> 
> > On 16 February 2018 at 15:57, Pekka Paalanen <ppaalanen at gmail.com> wrote:  
> > > For the attach on an enabled output to have an effect, we need to go
> > > through drmModeSetCrtc or ATOMIC_ALLOW_MODESET.
> > >
> > > Signed-off-by: Pekka Paalanen <pekka.paalanen at collabora.co.uk>
> > > ---
> > >  libweston/compositor-drm.c | 15 +++++++++++++++
> > >  1 file changed, 15 insertions(+)
> > >
> > > diff --git a/libweston/compositor-drm.c b/libweston/compositor-drm.c
> > > index d1fc66ac..d0196715 100644
> > > --- a/libweston/compositor-drm.c
> > > +++ b/libweston/compositor-drm.c
> > > @@ -4404,9 +4404,24 @@ static int
> > >  drm_output_attach_head(struct weston_output *output_base,
> > >                        struct weston_head *head_base)
> > >  {
> > > +       struct drm_backend *b = to_drm_backend(output_base->compositor);
> > > +
> > >         if (wl_list_length(&output_base->head_list) >= MAX_CLONED_CONNECTORS)
> > >                 return -1;
> > >
> > > +       if (!output_base->enabled)
> > > +               return 0;
> > > +
> > > +       /* XXX: ensure the configuration will work.
> > > +        * This is actually impossible without major infrastructure
> > > +        * work. */
> > > +
> > > +       /* Need to go through modeset to add connectors. */
> > > +       /* XXX: Ideally we'd do this per-output, not globally. */
> > > +       b->state_invalid = true;    
> > 
> > How about just an output->heads_dirty flag or something, which would
> > ensure ALLOW_MODESET/SetCrtc are called?  
> 
> Sure, I'll see about adding that.

Hi Daniel,

looks like I would get into a bit of pickle in
drm_pending_state_apply_atomic() if struct drm_output has heads_dirty
flag.

Currently the loop over heads is inside state_invalid==true branch, and
I would need to take it out of there if I no longer set state_invalid.
But when I go through the heads, I would not know which output they
were removed from, so I cannot check the output->heads_dirty flag.
Seems like I would need to add pending_removal flag to drm_head or add
a list of removed connectors to the drm_output.

Doesn't seem worth it to me, so maybe I'll just remove the XXX comments
and let it be?


Thanks,
pq
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <https://lists.freedesktop.org/archives/wayland-devel/attachments/20180418/d8de635e/attachment.sig>