linux-toradex.git/drivers/gpu/drm/xlnx, branch v5.16-rc5

drm/xlnx/zynqmp_disp: Fix incorrectly named enum 'zynqmp_disp_layer_id'

2021-08-08T23:28:05+00:00

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/xlnx/zynqmp_disp.c:101: warning: expecting prototype for enum zynqmp_disp_id. Prototype was for enum zynqmp_disp_layer_id instead

Signed-off-by: Lee Jones 
Reviewed-by: Laurent Pinchart 
Signed-off-by: Laurent Pinchart

drm: xlnx: zynqmp_dpsub: Expose plane ordering to userspace

2021-08-08T23:28:05+00:00

While the DPSUB has a fixed plane order, it still makes sense to expose
it to userspace to avoid hardcoding assumptions. Do so by adding an
immutable zpos property to planes.

Signed-off-by: Laurent Pinchart

drm: xlnx: zynqmp_dpsub: Add global alpha support

2021-08-08T23:28:04+00:00

The graphics plane has a global alpha setting. Expose it through the
plane's alpha property.

Signed-off-by: Laurent Pinchart

drm: xlnx: zynqmp_dpsub: Fix graphics layer blending

2021-08-08T23:28:04+00:00

To display the graphics layer, the global alpha needs to be enabled.
Enable it when the graphics plane is enabled (with full opacity), and
disable it otherwise.

Signed-off-by: Laurent Pinchart

drm: xlnx: zynqmp_dpsub: Pass disp structure to all internal functions

2021-08-08T23:28:04+00:00

The internal functions dealing with the audio/video buffer manager, the
blender, and the audio mixer, all receive pointers to the respective
objects. Those objects are embedded in the zynqmp_disp structure, and a
very small. Treating them as separate objects would require expanding
them with back-pointers to the zynqmp_disp in order to access fields
such as the device pointer for debug messages, and this isn't worth it.

Instead, merge those structures with the zynqmp_disp structure, and pass
the zynqmp_disp pointer to all internal functions.

Signed-off-by: Laurent Pinchart

drm: xlnx: zynqmp: Add zynqmp_disp_layer_is_video() to simplify the code

2021-08-08T23:28:04+00:00

Add a new function zynqmp_disp_layer_is_video() to simplify the code
that judges if a layer is the video layer.

Acked-by: Paul Cercueil 
Signed-off-by: Quanyang Wang 
Reviewed-by: Laurent Pinchart 
[Renamed is_layer_vid() to zynqmp_disp_layer_is_video()]]
Signed-off-by: Laurent Pinchart

drm: xlnx: zynqmp: release reset to DP controller before accessing DP registers

2021-08-08T23:28:04+00:00

When insmod zynqmp-dpsub.ko after rmmod it, system will hang with the
error log as below:

root@xilinx-zynqmp:~# insmod zynqmp-dpsub.ko
[   88.391289] [drm] Initialized zynqmp-dpsub 1.0.0 20130509 for fd4a0000.display on minor 0
[   88.529906] Console: switching to colour frame buffer device 128x48
[   88.549402] zynqmp-dpsub fd4a0000.display: [drm] fb0: zynqmp-dpsubdrm frame buffer device
[   88.571624] zynqmp-dpsub fd4a0000.display: ZynqMP DisplayPort Subsystem driver probed
root@xilinx-zynqmp:~# rmmod zynqmp_dpsub
[   94.023404] Console: switching to colour dummy device 80x25
root@xilinx-zynqmp:~# insmod zynqmp-dpsub.ko
	

This is because that in zynqmp_dp_probe it tries to access some DP
registers while the DP controller is still in the reset state. When
running "rmmod zynqmp_dpsub", zynqmp_dp_reset(dp, true) in
zynqmp_dp_phy_exit is called to force the DP controller into the reset
state. Then insmod will call zynqmp_dp_probe to program the DP registers,
but at this moment the DP controller hasn't been brought out of the reset
state yet since the function zynqmp_dp_reset(dp, false) is called later and
this will result the system hang.

Releasing the reset to DP controller before any read/write operation to it
will fix this issue. And for symmetry, move zynqmp_dp_reset() call from
zynqmp_dp_phy_exit() to zynqmp_dp_remove().

Signed-off-by: Quanyang Wang 
Reviewed-by: Laurent Pinchart 
Signed-off-by: Laurent Pinchart

drm: xlnx: zynqmp_dpsub: Update dependencies for ZynqMP DP

2021-08-08T23:28:04+00:00

ZynqMP DP requires the ZynqMP PHY and DPDMA to operate properly. So
depend on both the PHY and DPDMA.

Signed-off-by: Dylan Yip 
Signed-off-by: Laurent Pinchart

drm: xlnx: zynqmp_dpsub: Call pm_runtime_get_sync before setting pixel clock

2021-08-08T23:28:04+00:00

The Runtime PM subsystem will force the device "fd4a0000.zynqmp-display"
to enter suspend state while booting if the following conditions are met:
- the usage counter is zero (pm_runtime_get_sync hasn't been called yet)
- no 'active' children (no zynqmp-dp-snd-xx node under dpsub node)
- no other device in the same power domain (dpdma node has no
		"power-domains = <&zynqmp_firmware PD_DP>" property)

So there is a scenario as below:
1) DP device enters suspend state   <- call zynqmp_gpd_power_off
2) zynqmp_disp_crtc_setup_clock	    <- configurate register VPLL_FRAC_CFG
3) pm_runtime_get_sync		    <- call zynqmp_gpd_power_on and clear previous
				       VPLL_FRAC_CFG configuration
4) clk_prepare_enable(disp->pclk)   <- enable failed since VPLL_FRAC_CFG
				       configuration is corrupted

From above, we can see that pm_runtime_get_sync may clear register
VPLL_FRAC_CFG configuration and result the failure of clk enabling.
Putting pm_runtime_get_sync at the very beginning of the function
zynqmp_disp_crtc_atomic_enable can resolve this issue.

Signed-off-by: Quanyang Wang 
Reviewed-by: Laurent Pinchart 
Signed-off-by: Laurent Pinchart

drm/xlnx: Don't set struct drm_device.irq_enabled

2021-06-29T13:40:53+00:00

The field drm_device.irq_enabled is only used by legacy drivers
with userspace modesetting. Don't set it in xlnx.

Signed-off-by: Thomas Zimmermann 
Reviewed-by: Laurent Pinchart 
Acked-by: Daniel Vetter 
Link: https://patchwork.freedesktop.org/patch/msgid/20210625082222.3845-27-tzimmermann@suse.de