In case of a modeset where a mode gets split across mutiple CRTCs in the driver specific implementation (bigjoiner in i915) we wrongly count the affected CRTCs based on the drm_crtc_mask and indicate the stolen CRTC as an affected CRTC in atomic_check_only(). This triggers a warning since affected CRTCs doent match requested CRTC.
To fix this in such bigjoiner configurations, we should only increment affected crtcs if that CRTC is enabled in UAPI not if it is just used internally in the driver to split the mode.
v3: Add the same uapi crtc_state->enable check in requested crtc calc (Ville)
Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Simon Ser contact@emersion.fr Cc: Pekka Paalanen pekka.paalanen@collabora.co.uk Cc: Daniel Stone daniels@collabora.com Cc: Daniel Vetter daniel.vetter@intel.com Cc: dri-devel@lists.freedesktop.org Signed-off-by: Manasi Navare manasi.d.navare@intel.com --- drivers/gpu/drm/drm_atomic.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index ff1416cd609a..a1e4c7905ebb 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1310,8 +1310,10 @@ int drm_atomic_check_only(struct drm_atomic_state *state)
DRM_DEBUG_ATOMIC("checking %p\n", state);
- for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) - requested_crtc |= drm_crtc_mask(crtc); + for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) { + if (new_crtc_state->enable) + requested_crtc |= drm_crtc_mask(crtc); + }
for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, i) { ret = drm_atomic_plane_check(old_plane_state, new_plane_state); @@ -1360,8 +1362,10 @@ int drm_atomic_check_only(struct drm_atomic_state *state) } }
- for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) - affected_crtc |= drm_crtc_mask(crtc); + for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) { + if (new_crtc_state->enable) + affected_crtc |= drm_crtc_mask(crtc); + }
/* * For commits that allow modesets drivers can add other CRTCs to the
Hi Ville,
Could you take a look at this, this addresses teh review comments from prev version
Manasi
On Mon, Oct 04, 2021 at 04:59:13AM -0700, Manasi Navare wrote:
In case of a modeset where a mode gets split across mutiple CRTCs in the driver specific implementation (bigjoiner in i915) we wrongly count the affected CRTCs based on the drm_crtc_mask and indicate the stolen CRTC as an affected CRTC in atomic_check_only(). This triggers a warning since affected CRTCs doent match requested CRTC.
To fix this in such bigjoiner configurations, we should only increment affected crtcs if that CRTC is enabled in UAPI not if it is just used internally in the driver to split the mode.
v3: Add the same uapi crtc_state->enable check in requested crtc calc (Ville)
Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Simon Ser contact@emersion.fr Cc: Pekka Paalanen pekka.paalanen@collabora.co.uk Cc: Daniel Stone daniels@collabora.com Cc: Daniel Vetter daniel.vetter@intel.com Cc: dri-devel@lists.freedesktop.org Signed-off-by: Manasi Navare manasi.d.navare@intel.com
drivers/gpu/drm/drm_atomic.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index ff1416cd609a..a1e4c7905ebb 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1310,8 +1310,10 @@ int drm_atomic_check_only(struct drm_atomic_state *state)
DRM_DEBUG_ATOMIC("checking %p\n", state);
- for_each_new_crtc_in_state(state, crtc, new_crtc_state, i)
requested_crtc |= drm_crtc_mask(crtc);
for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
if (new_crtc_state->enable)requested_crtc |= drm_crtc_mask(crtc);}
for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, i) { ret = drm_atomic_plane_check(old_plane_state, new_plane_state);
@@ -1360,8 +1362,10 @@ int drm_atomic_check_only(struct drm_atomic_state *state) } }
- for_each_new_crtc_in_state(state, crtc, new_crtc_state, i)
affected_crtc |= drm_crtc_mask(crtc);
for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
if (new_crtc_state->enable)affected_crtc |= drm_crtc_mask(crtc);}
/*
- For commits that allow modesets drivers can add other CRTCs to the
-- 2.19.1
On Fri, Oct 22, 2021 at 12:51:12PM -0700, Navare, Manasi wrote:
Hi Ville,
Could you take a look at this, this addresses teh review comments from prev version
I don't think I ever got an answer to my question as to whether this was tested with all the interesting scenarios: 1) just with the master crtc added by userspace into the commit 2) just with the slave crtc added by userspace into the commit 3) both crtcs added by userspace into the commit
I guess 1) has been tested since that happens all the time, but the other two scanarios would likely need to be done with a synthetic test to make sure we're actually hitting them.
I think it *should* work, but I'd like to have real proof of that. Reviewed-by: Ville Syrjälä ville.syrjala@linux.intel.com
Manasi
On Mon, Oct 04, 2021 at 04:59:13AM -0700, Manasi Navare wrote:
In case of a modeset where a mode gets split across mutiple CRTCs in the driver specific implementation (bigjoiner in i915) we wrongly count the affected CRTCs based on the drm_crtc_mask and indicate the stolen CRTC as an affected CRTC in atomic_check_only(). This triggers a warning since affected CRTCs doent match requested CRTC.
To fix this in such bigjoiner configurations, we should only increment affected crtcs if that CRTC is enabled in UAPI not if it is just used internally in the driver to split the mode.
v3: Add the same uapi crtc_state->enable check in requested crtc calc (Ville)
Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Simon Ser contact@emersion.fr Cc: Pekka Paalanen pekka.paalanen@collabora.co.uk Cc: Daniel Stone daniels@collabora.com Cc: Daniel Vetter daniel.vetter@intel.com Cc: dri-devel@lists.freedesktop.org Signed-off-by: Manasi Navare manasi.d.navare@intel.com
drivers/gpu/drm/drm_atomic.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index ff1416cd609a..a1e4c7905ebb 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1310,8 +1310,10 @@ int drm_atomic_check_only(struct drm_atomic_state *state)
DRM_DEBUG_ATOMIC("checking %p\n", state);
- for_each_new_crtc_in_state(state, crtc, new_crtc_state, i)
requested_crtc |= drm_crtc_mask(crtc);
for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
if (new_crtc_state->enable)requested_crtc |= drm_crtc_mask(crtc);}
for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, i) { ret = drm_atomic_plane_check(old_plane_state, new_plane_state);
@@ -1360,8 +1362,10 @@ int drm_atomic_check_only(struct drm_atomic_state *state) } }
- for_each_new_crtc_in_state(state, crtc, new_crtc_state, i)
affected_crtc |= drm_crtc_mask(crtc);
for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
if (new_crtc_state->enable)affected_crtc |= drm_crtc_mask(crtc);}
/*
- For commits that allow modesets drivers can add other CRTCs to the
-- 2.19.1
On Tue, Jan 18, 2022 at 06:34:20PM +0200, Ville Syrjälä wrote:
On Fri, Oct 22, 2021 at 12:51:12PM -0700, Navare, Manasi wrote:
Hi Ville,
Could you take a look at this, this addresses teh review comments from prev version
I don't think I ever got an answer to my question as to whether this was tested with all the interesting scenarios:
- just with the master crtc added by userspace into the commit
- just with the slave crtc added by userspace into the commit
- both crtcs added by userspace into the commit
I guess 1) has been tested since that happens all the time, but the other two scanarios would likely need to be done with a synthetic test to make sure we're actually hitting them.
I think it *should* work, but I'd like to have real proof of that. Reviewed-by: Ville Syrjälä ville.syrjala@linux.intel.com
Thank you for your review here Ville. I have triggered a separate email thread to understand all the above testing scenarios and get them tested with bigjoiner IGT.
Manasi
Manasi
On Mon, Oct 04, 2021 at 04:59:13AM -0700, Manasi Navare wrote:
In case of a modeset where a mode gets split across mutiple CRTCs in the driver specific implementation (bigjoiner in i915) we wrongly count the affected CRTCs based on the drm_crtc_mask and indicate the stolen CRTC as an affected CRTC in atomic_check_only(). This triggers a warning since affected CRTCs doent match requested CRTC.
To fix this in such bigjoiner configurations, we should only increment affected crtcs if that CRTC is enabled in UAPI not if it is just used internally in the driver to split the mode.
v3: Add the same uapi crtc_state->enable check in requested crtc calc (Ville)
Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Simon Ser contact@emersion.fr Cc: Pekka Paalanen pekka.paalanen@collabora.co.uk Cc: Daniel Stone daniels@collabora.com Cc: Daniel Vetter daniel.vetter@intel.com Cc: dri-devel@lists.freedesktop.org Signed-off-by: Manasi Navare manasi.d.navare@intel.com
drivers/gpu/drm/drm_atomic.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index ff1416cd609a..a1e4c7905ebb 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1310,8 +1310,10 @@ int drm_atomic_check_only(struct drm_atomic_state *state)
DRM_DEBUG_ATOMIC("checking %p\n", state);
- for_each_new_crtc_in_state(state, crtc, new_crtc_state, i)
requested_crtc |= drm_crtc_mask(crtc);
for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
if (new_crtc_state->enable)requested_crtc |= drm_crtc_mask(crtc);}
for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, i) { ret = drm_atomic_plane_check(old_plane_state, new_plane_state);
@@ -1360,8 +1362,10 @@ int drm_atomic_check_only(struct drm_atomic_state *state) } }
- for_each_new_crtc_in_state(state, crtc, new_crtc_state, i)
affected_crtc |= drm_crtc_mask(crtc);
for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
if (new_crtc_state->enable)affected_crtc |= drm_crtc_mask(crtc);}
/*
- For commits that allow modesets drivers can add other CRTCs to the
-- 2.19.1
-- Ville Syrjälä Intel
dri-devel@lists.freedesktop.org
-
Manasi Navare -
Navare, Manasi
-
Ville Syrjälä