[PATCH] arm64: dts: zynqmp: Add power domain for the DisplayPort DMA controller

[Michal Simek]

Hi,

On 3/8/21 11:58 AM, Rajan Vaja wrote:
> Hi,
> 
>> -----Original Message-----
>> From: Michal Simek <michal.simek at xilinx.com>
>> Sent: Monday, March 8, 2021 2:09 AM
>> To: Laurent Pinchart <laurent.pinchart at ideasonboard.com>; linux-arm-
>> kernel at lists.infradead.org; Rajan Vaja <RAJANV at xilinx.com>
>> Cc: Michal Simek <michals at xilinx.com>
>> Subject: Re: [PATCH] arm64: dts: zynqmp: Add power domain for the DisplayPort
>> DMA controller
>>
>> Hi,
>>
>> On 3/7/21 12:09 AM, Laurent Pinchart wrote:
>>> The DisplayPort DMA controller (DPDMA) is located in the same power
>>> domain as the DisplayPort Subsystem (DPSUB). Specify the power domain in
>>> the device tree.
>>>
>>> Signed-off-by: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
>>> ---
>>>
>>> I'm actually not entirely sure about the truthfulness of the commit
>>> message, as I haven't been able to located this information in the
>>> reference manual. Nonetheless, without this patch, I get a clock enable
>>> failure when starting the display:
>>>
>>> [   70.517104] zynqmp_pll_enable() clock enable failed for vpll_int, ret = -22
>>> [   70.527552] zynqmp-dpsub fd4a0000.display: failed to enable a pixel clock
>>>
>>> I'm thus quite confident the patch is correct. If anyone has more
>>> information and would like to reword the commit message, please feel
>>> free to do so.
>>>
>>>  arch/arm64/boot/dts/xilinx/zynqmp.dtsi | 1 +
>>>  1 file changed, 1 insertion(+)
>>>
>>> diff --git a/arch/arm64/boot/dts/xilinx/zynqmp.dtsi
>> b/arch/arm64/boot/dts/xilinx/zynqmp.dtsi
>>> index c113c76f360d..dc5c3a09c77e 100644
>>> --- a/arch/arm64/boot/dts/xilinx/zynqmp.dtsi
>>> +++ b/arch/arm64/boot/dts/xilinx/zynqmp.dtsi
>>> @@ -749,6 +749,7 @@ zynqmp_dpdma: dma-controller at fd4c0000 {
>>>  			interrupts = <0 122 4>;
>>>  			interrupt-parent = <&gic>;
>>>  			clock-names = "axi_clk";
>>> +			power-domains = <&zynqmp_firmware PD_DP>;
>>>  			#dma-cells = <1>;
>>>  		};
>>>
>>>
>>
>> Rajan: Can you please comment this?
> [Rajan] Change looks fine to me. If power-domain entry is not present in device tree node, that device would not be requested and hence clock control would fail.

The key point here is if DPDMA is in also in PD_DP or there is separate
one. And can you please send any tag? acked-by, etc?

Thanks,
Michal