Merge v6.13 into drm-next
authorSimona Vetter <simona.vetter@ffwll.ch>
Thu, 23 Jan 2025 13:39:49 +0000 (14:39 +0100)
committerSimona Vetter <simona.vetter@ffwll.ch>
Thu, 23 Jan 2025 13:42:21 +0000 (14:42 +0100)
A regression was caused by commit e4b5ccd392b9 ("drm/v3d: Ensure job
pointer is set to NULL after job completion"), but this commit is not
yet in next-fixes, fast-forward it.

Note that this recreates Linus merge in 96c84703f1cf ("Merge tag
'drm-next-2025-01-17' of https://gitlab.freedesktop.org/drm/kernel")
because I didn't want to backmerge a random point in the merge window.

Signed-off-by: Simona Vetter <simona.vetter@ffwll.ch>
16 files changed:
1  2 
MAINTAINERS
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
drivers/gpu/drm/bridge/ite-it6263.c
drivers/gpu/drm/display/drm_bridge_connector.c
drivers/gpu/drm/drm_connector.c
drivers/gpu/drm/mediatek/mtk_dp.c
drivers/gpu/drm/mediatek/mtk_drm_drv.c
drivers/gpu/drm/tests/drm_connector_test.c
drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
drivers/gpu/drm/xe/tests/xe_bo.c
drivers/gpu/drm/xe/tests/xe_mocs.c
drivers/gpu/drm/xe/xe_gt.h
drivers/gpu/drm/xe/xe_hw_engine.c
drivers/gpu/drm/xe/xe_oa.c

diff --cc MAINTAINERS
Simple merge
index 6d5d81f0dc4e7bffa6f7d0e8f2aa0daf60d06199,1d155463d044b2ad739cca256a9fa8c515b39de8..784b03abb3a4398b4c54436004bee9c3eede5a12
@@@ -2072,20 -2008,14 +2073,23 @@@ void amdgpu_gfx_enforce_isolation_ring_
        mutex_lock(&adev->enforce_isolation_mutex);
        if (adev->enforce_isolation[idx]) {
                if (adev->kfd.init_complete)
-                       amdgpu_gfx_kfd_sch_ctrl(adev, idx, false);
+                       sched_work = true;
        }
        mutex_unlock(&adev->enforce_isolation_mutex);
+       if (sched_work)
+               amdgpu_gfx_kfd_sch_ctrl(adev, idx, false);
  }
  
 +/**
 + * amdgpu_gfx_enforce_isolation_ring_end_use - End use of a ring with enforced isolation
 + * @ring: Pointer to the amdgpu_ring structure
 + *
 + * Ring end_use helper implementation for gfx which serializes access to the
 + * gfx IP between kernel submission IOCTLs and KFD user queues when isolation
 + * enforcement is enabled. The kernel submission IOCTLs and KFD user queues
 + * each get a time slice when both are active.
 + */
  void amdgpu_gfx_enforce_isolation_ring_end_use(struct amdgpu_ring *ring)
  {
        struct amdgpu_device *adev = ring->adev;
Simple merge
index 32108307de66560029ddf319169ac7d7e2af40d2,3a5a4f92c979accaa2a8f79ca0f15396dd579429..56f977bbe62d612f95e723122afece486e854fd4
@@@ -596,8 -460,9 +596,11 @@@ struct drm_connector *drm_bridge_connec
                return ERR_PTR(-EINVAL);
  
        if (bridge_connector->bridge_hdmi) {
+               if (!connector->ycbcr_420_allowed)
+                       supported_formats &= ~BIT(HDMI_COLORSPACE_YUV420);
 +              bridge = bridge_connector->bridge_hdmi;
 +
                ret = drmm_connector_hdmi_init(drm, connector,
                                               bridge_connector->bridge_hdmi->vendor,
                                               bridge_connector->bridge_hdmi->product,
Simple merge
Simple merge
Simple merge
Simple merge
Simple merge
Simple merge
Simple merge
Simple merge