Adds crtc page-flip support by passing the new requested framebuffer to vbox_crtc_do_set_base().
There is no attempt to support vblank interrupts, so this page-flip implementation does not try to sync the page-flip to vertical blanking, so expect tearing effects. Is it possible to access the host vblank intervals in VBOX?
Steve Longerbeam (2): staging: vboxvideo: Pass a new framebuffer to vbox_crtc_do_set_base staging: vboxvideo: Add page-flip support
drivers/staging/vboxvideo/vbox_mode.c | 34 +++++++++++++++++++++++++++++++--- 1 file changed, 31 insertions(+), 3 deletions(-)
This modifies vbox_crtc_do_set_base() to take a new framebuffer to be activated, instead of the existing framebuffer attached to the crtc. This change allows the function to be given the new framebuffer from a page-flip request.
Signed-off-by: Steve Longerbeam steve_longerbeam@mentor.com --- drivers/staging/vboxvideo/vbox_mode.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/vboxvideo/vbox_mode.c b/drivers/staging/vboxvideo/vbox_mode.c index b265fe9..688e80d 100644 --- a/drivers/staging/vboxvideo/vbox_mode.c +++ b/drivers/staging/vboxvideo/vbox_mode.c @@ -222,7 +222,9 @@ static bool vbox_set_up_input_mapping(struct vbox_private *vbox) }
static int vbox_crtc_do_set_base(struct drm_crtc *crtc, - struct drm_framebuffer *old_fb, int x, int y) + struct drm_framebuffer *old_fb, + struct drm_framebuffer *new_fb, + int x, int y) { struct vbox_private *vbox = crtc->dev->dev_private; struct vbox_crtc *vbox_crtc = to_vbox_crtc(crtc); @@ -245,7 +247,7 @@ static int vbox_crtc_do_set_base(struct drm_crtc *crtc, vbox_bo_unreserve(bo); }
- vbox_fb = to_vbox_framebuffer(CRTC_FB(crtc)); + vbox_fb = to_vbox_framebuffer(new_fb); obj = vbox_fb->obj; bo = gem_to_vbox_bo(obj);
@@ -281,7 +283,7 @@ static int vbox_crtc_do_set_base(struct drm_crtc *crtc, static int vbox_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y, struct drm_framebuffer *old_fb) { - return vbox_crtc_do_set_base(crtc, old_fb, x, y); + return vbox_crtc_do_set_base(crtc, old_fb, CRTC_FB(crtc), x, y); }
static int vbox_crtc_mode_set(struct drm_crtc *crtc,
Adds crtc page-flip support by passing the new requested framebuffer to vbox_crtc_do_set_base().
Note there is no attempt to support vblank interrupts, it's not not known how to do this in VBOX or if it is even possible. Since this page-flip implementation does not try to sync the page-flip to vertical blanking, tearing effects are possible.
Signed-off-by: Steve Longerbeam steve_longerbeam@mentor.com --- drivers/staging/vboxvideo/vbox_mode.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+)
diff --git a/drivers/staging/vboxvideo/vbox_mode.c b/drivers/staging/vboxvideo/vbox_mode.c index 688e80d..285d8ad 100644 --- a/drivers/staging/vboxvideo/vbox_mode.c +++ b/drivers/staging/vboxvideo/vbox_mode.c @@ -308,6 +308,31 @@ static int vbox_crtc_mode_set(struct drm_crtc *crtc, return ret; }
+static int vbox_crtc_page_flip(struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_pending_vblank_event *event, + uint32_t page_flip_flags, + struct drm_modeset_acquire_ctx *ctx) +{ + struct vbox_private *vbox = crtc->dev->dev_private; + struct drm_device *drm = vbox->dev; + unsigned long flags; + int rc; + + rc = vbox_crtc_do_set_base(crtc, CRTC_FB(crtc), fb, 0, 0); + if (rc) + return rc; + + spin_lock_irqsave(&drm->event_lock, flags); + + if (event) + drm_crtc_send_vblank_event(crtc, event); + + spin_unlock_irqrestore(&drm->event_lock, flags); + + return 0; +} + static void vbox_crtc_disable(struct drm_crtc *crtc) { } @@ -346,6 +371,7 @@ static const struct drm_crtc_funcs vbox_crtc_funcs = { .reset = vbox_crtc_reset, .set_config = drm_crtc_helper_set_config, /* .gamma_set = vbox_crtc_gamma_set, */ + .page_flip = vbox_crtc_page_flip, .destroy = vbox_crtc_destroy, };
On Fri, Jul 20, 2018 at 10:17:29AM -0700, Steve Longerbeam wrote:
Adds crtc page-flip support by passing the new requested framebuffer to vbox_crtc_do_set_base().
There is no attempt to support vblank interrupts, so this page-flip implementation does not try to sync the page-flip to vertical blanking, so expect tearing effects. Is it possible to access the host vblank intervals in VBOX?
I think it'd be great to move vboxvideo over to atomic instead, which is required for destaging anyway, and would give you page flip support for free. -Daniel
Steve Longerbeam (2): staging: vboxvideo: Pass a new framebuffer to vbox_crtc_do_set_base staging: vboxvideo: Add page-flip support
drivers/staging/vboxvideo/vbox_mode.c | 34 +++++++++++++++++++++++++++++++--- 1 file changed, 31 insertions(+), 3 deletions(-)
-- 2.7.4
dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
On 08/06/2018 01:34 AM, Daniel Vetter wrote:
On Fri, Jul 20, 2018 at 10:17:29AM -0700, Steve Longerbeam wrote:
Adds crtc page-flip support by passing the new requested framebuffer to vbox_crtc_do_set_base().
There is no attempt to support vblank interrupts, so this page-flip implementation does not try to sync the page-flip to vertical blanking, so expect tearing effects. Is it possible to access the host vblank intervals in VBOX?
I think it'd be great to move vboxvideo over to atomic instead, which is required for destaging anyway, and would give you page flip support for free.
Hi Daniel, yes I agree vboxvideo needs to move to the DRM atomic framework.
Steve
Steve Longerbeam (2): staging: vboxvideo: Pass a new framebuffer to vbox_crtc_do_set_base staging: vboxvideo: Add page-flip support
drivers/staging/vboxvideo/vbox_mode.c | 34 +++++++++++++++++++++++++++++++--- 1 file changed, 31 insertions(+), 3 deletions(-)
-- 2.7.4
dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
Hi all,
On 08/07/2018 02:57 PM, Steve Longerbeam wrote:
On 08/06/2018 01:34 AM, Daniel Vetter wrote:
On Fri, Jul 20, 2018 at 10:17:29AM -0700, Steve Longerbeam wrote:
Adds crtc page-flip support by passing the new requested framebuffer to vbox_crtc_do_set_base().
There is no attempt to support vblank interrupts, so this page-flip implementation does not try to sync the page-flip to vertical blanking, so expect tearing effects. Is it possible to access the host vblank intervals in VBOX?
I think it'd be great to move vboxvideo over to atomic instead, which is required for destaging anyway, and would give you page flip support for free.
Hi Daniel, yes I agree vboxvideo needs to move to the DRM atomic framework.
Irrespective of moving vboxvideo to atomic framework, I still need to get an answer to the question, is it possible to access the host vblank intervals in order to sync page-flip to vblank, to prevent tearing. Is this possible somehow, perhaps by calling into the VBOX hypervisor?
Thanks, Steve
dri-devel@lists.freedesktop.org
-
Daniel Vetter -
Steve Longerbeam