Hi Geert,
On Monday 13 Feb 2017 11:13:31 Geert Uytterhoeven wrote:
> On Mon, Feb 13, 2017 at 11:05 AM, Geert Uytterhoeven wrote:
> > On Sun, Feb 12, 2017 at 2:13 PM, Laurent Pinchart wrote:
> >> The following changes since commit
5685cab84a6dc376a8d0e4534ae4b973546158b4:
> >> Merge remote-tracking branch 'drm-misc/drm-misc-next' into
> >> drm/next/base
> >>
> >> (2017-02-12 13:18:50 +0200)
> >>
> >> are available in the git repository at:
> >> git://linuxtv.org/pinchartl/media.git drm/next/du
> >>
> >> for you to fetch changes up to a194138cd82dff52d4c39895fd89dc6f26eafc97:
> >> drm: rcar-du: Use DRM core's atomic commit helper (2017-02-12 14:15:36
> >>
> >> +0200)
> >
> > Thank you, scheduled for inclusion in next renesas-drivers release
> > (probably renesas-drivers-2017-02-21-v4.10).
>
> Hit send too soon...
>
> Note that it already fails to merge cleanly into today's linux-next (holding
> wood, rabbit ears and tails, ... for next week).
Hardly a surprise, drm-misc evolves really fast, any patch series that touches
the DRM core has a high risk of conflicting. I hope that Maarten will resubmit
his patches soon and get them merged.
> My untested resolution:
Looks good to me.
> diff --cc drivers/gpu/drm/drm_atomic.c
> index a5673107db26c403,428743efc031e717..0000000000000000
> --- a/drivers/gpu/drm/drm_atomic.c
> +++ b/drivers/gpu/drm/drm_atomic.c
> @@@ -1938,8 -1949,8 +1949,8 @@@ static int prepare_crtc_signaling(struc
> if (arg->flags & DRM_MODE_ATOMIC_TEST_ONLY)
> return 0;
>
> - for_each_crtc_in_state(state, crtc, crtc_state, i) {
> + for_each_new_crtc_in_state(state, crtc, crtc_state, i) {
> - u64 __user *fence_ptr;
> + s32 __user *fence_ptr;
>
> fence_ptr = get_out_fence_for_crtc(crtc_state->state, crtc);
>
> @@@ -2018,17 -2029,14 +2029,17 @@@ static void complete_crtc_signaling(str
> return;
> }
>
> - for_each_crtc_in_state(state, crtc, crtc_state, i) {
> + for_each_new_crtc_in_state(state, crtc, crtc_state, i) {
> + struct drm_pending_vblank_event *event = crtc_state->event;
> /*
> - * TEST_ONLY and PAGE_FLIP_EVENT are mutually
> - * exclusive, if they weren't, this code should be
> - * called on success for TEST_ONLY too.
> + * Free the allocated event. drm_atomic_helper_setup_commit
> + * can allocate an event too, so only free it if it's ours
> + * to prevent a double free in drm_atomic_state_clear.
> */
> - if (crtc_state->event)
> - drm_event_cancel_free(dev,
> &crtc_state->event->base); + if (event && (event->base.fence
> || event->base.file_priv)) { +
> drm_event_cancel_free(dev, &event->base);
> + crtc_state->event = NULL;
> + }
> }
>
> if (!fence_state)
--
Regards,
Laurent Pinchart
