[PATCH weston v2 1/1] compositor-drm: Unref current & pending fb when deactivating the session

[Daniel Stone]

Hi Miguel,

On 11 August 2017 at 00:55, Miguel A. Vico <mvicomoya at nvidia.com> wrote:
> we stopped forcing a modeset when restoring the session. The motivation
> was that we would use a stale fb, so better to let the next repaint
> handle it.
>
> However, if drm_output::fb_current != NULL, we won't issue a modeset
> upon repaint.
>
> This change unreferences both drm_output::fb_current and
> drm_output::fb_pending when deactivating the current session. This
> ensures the very first repaint after restoring the session will issue a
> modeset.

Sorry for missing this the first time around. But I think this fix
from Pekka should be equivalent:

commit 6b65d8f12021d8fff0db37fd10b9a469769178b2
Refs: 2.99.92-4-g6b65d8f1
Author:     Pekka Paalanen <pekka.paalanen at collabora.co.uk>
AuthorDate: Thu Jul 27 13:44:32 2017 +0300

    compositor-drm: reset KMS state on VT-switch in

    Fix a regression with VT-switching away from Weston and then back
    causing drmModePageFlip() to fail with ENOSPC or EINVAL, leaving one or
    more outputs not updated. The regression appeared in
    47224cc9312fef05c1a523ea0da0a1aae66f100d:

            compositor-drm: Delete drm_backend_set_modes

    Fix it by forcing a drmModeSetCrtc() on all outputs both initially
    created and after VT-switch in.

Do you still see VT-switching issues?

Cheers,
Daniel