kernel/linux.git/drivers/gpu/drm/msm, branch v4.20.5

drm/msm: Fix error return checking

2018-12-03T13:46:14+00:00

The error checks on ret for a negative error return always fails because the return value of iommu_map_sg() is unsigned and can never be negative. Detected with Coccinelle: drivers/gpu/drm/msm/msm_iommu.c:69:9-12: WARNING: Unsigned expression compared with zero: ret < 0 Signed-off-by: Wen Yang CC: Rob Clark CC: David Airlie CC: Julia Lawall CC: linux-arm-msm@vger.kernel.org CC: dri-devel@lists.freedesktop.org CC: freedreno@lists.freedesktop.org CC: linux-kernel@vger.kernel.org Signed-off-by: Sean Paul

drm/msm/dpu: Ignore alpha for XBGR8888 format

2018-12-03T13:46:14+00:00

Alpha enable in the pixel format will help in selecting the blend rule. By keeping alpha enable to true we are allowing foreground alpha to blend with the layer. If alpha is don't care, then we should not allow pixel alpha to be part of blend equation. Signed-off-by: Jayant Shekhar Signed-off-by: Sean Paul

drm/msm: dpu: Fix "WARNING: invalid free of devm_ allocated data"

2018-12-03T13:46:14+00:00

'dpu_enc' is a member of 'drm_enc' And 'drm_enc' got allocated with devm_kzalloc in dpu_encoder_init. This gives this error message: ./drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c:459:1-6: WARNING: invalid free of devm_ allocated data Signed-off-by: YueHaibing Signed-off-by: Rob Clark Signed-off-by: Sean Paul

drm/msm/hdmi: Drop pointless static qualifier in msm_hdmi_bind()

2018-12-03T13:46:14+00:00

There is no need to have the 'struct hdmi_platform_config *hdmi_cfg' variable static since new value always be assigned before use it. Signed-off-by: YueHaibing Signed-off-by: Rob Clark Signed-off-by: Sean Paul

drm/msm: Move fence put to where failure occurs

2018-12-03T13:46:13+00:00

If dma_fence_wait fails to wait for a supplied in-fence in msm_ioctl_gem_submit, make sure we release that in-fence. Also remove this dma_fence_put() from the 'out' label. Signed-off-by: Robert Foss Reviewed-by: Chris Wilson Cc: stable@vger.kernel.org Signed-off-by: Rob Clark Signed-off-by: Sean Paul

drm/msm: dpu: Don't set legacy plane->crtc pointer

2018-12-03T13:46:13+00:00

It causes a WARN in drm_atomic_get_plane_state(), and is not used by atomic (or dpu). Signed-off-by: Sean Paul Reviewed-by: Daniel Vetter Signed-off-by: Rob Clark

drm/msm/gpu: Don't map command buffers with nr_relocs equal to 0

2018-12-03T13:46:13+00:00

If a command buffer doesn't have any relocs assigned to it there then is no need to map it in the kernel address space. Signed-off-by: Jordan Crouse Signed-off-by: Rob Clark Signed-off-by: Sean Paul

drm/msm/hdmi: Enable HPD after HDMI IRQ is set up

2018-12-03T13:46:13+00:00

SoCs that contain MDP5 have a top level wrapper called MDSS that manages locks, power and irq for the sub-blocks within it. Irq for HDMI is also routed through the MDSS. Shortly after the Hot Plug Detection (HPD) is enabled in HDMI, HDMI interrupts are recieved by the MDSS interrupt handler. However at this moment the HDMI irq is still not mapped to the MDSS irq domain so the HDMI irq handler cannot be called to process the interrupts. This leads to a flood of HDMI interrupts on CPU 0. If we are lucky to have the HDMI initialization running on a different CPU, it will eventually map the HDMI irq to MDSS irq domain, the next HDMI interrupt will be handled by the HDMI irq handler, the interrupt flood will stop and we will recover. If the HDMI initialization is running on CPU 0, then it cannot complete and there is nothing to stop the interrupt flood on CPU 0. The system is stuck. Fix this by moving the HPD enablement after the HDMI irq is mapped to the MDSS irq domain. Signed-off-by: Todor Tomov Signed-off-by: Rob Clark Signed-off-by: Sean Paul

drm/msm: validate display and event threads

2018-12-03T13:46:07+00:00

While creating display and event threads per crtc, validate them before setting their priorities. changes in v2: - use dev_warn (Abhinav Kumar) changes in v3: - fix compilation error changes in v4: - Remove Change-Id (Sean Paul) - Keep logging within 80 char limit (Sean Paul) Reviewed-by: Sean Paul Signed-off-by: Jeykumar Sankaran Signed-off-by: Rob Clark Signed-off-by: Sean Paul

drm/msm: Grab a vblank reference when waiting for commit_done

2018-11-30T16:58:03+00:00

Similar to the atomic helpers, we should enable vblank while we're waiting for the commit to finish. DPU needs this, MDP5 seems to work fine without it. Reviewed-by: Abhinav Kumar Signed-off-by: Sean Paul Signed-off-by: Rob Clark