Re: [PATCH v2 6/7] arm64: dts: qcom: Add support for ECS LIVA QC710

[Nikita Travkin <nikita@trvn.ru>]

Konrad Dybcio писал(а) 21.01.2026 16:20:
> On 1/21/26 12:30 AM, Val Packett wrote:
>> Add a device tree for the ECS LIVA QC710 (Snapdragon 7c) mini PC/devkit.
>> 
>> Working:
>> - Wi-Fi (wcn3990 hw1.0)
>> - Bluetooth
>> - USB Type-A (USB3 and USB2)
>> - Ethernet (over USB2)
>> - HDMI Display
>> - eMMC
>> - SDHC (microSD slot)
>> 
>> Not included:
>> - HDMI Audio
>> - EC (IT8987)
>> 
>> Signed-off-by: Val Packett <val@packett.cool>
>> ---
> 
> [...]
>> +&tlmm {
>> +	/*
>> +	 * The TZ seem to protect those because some boards can have
>> +	 * fingerprint sensor connected to this range. Not connected
>> +	 * on this board
>> +	 */
>> +	gpio-reserved-ranges = <58 5>;
> 
> Would it still work with <59 4>? 58 seems to not be related to a QUP,
> rather as an audio MCLK output (would that be why +Nikita couldn't get
> some sort of audio working on his aspire1? That DT has it reserved too)
> 

FWIW according to aspire1 schematic GPIO 57 is codec mclk (which is
implemented and works perfectly, only DMIC connected to the soc
instead of said codec is broken); GPIO 58~62 are NC with a note
"Finger Printer" [sic] near them. There is no bracket annotation to
associate the note to specific pin range, but it's smack in the
middle of 58~62 range and from the fact that 63 (speaker amp bclk)
and pretty much every other pin later is also used, I guessed all 5
are actually reserved. It's possible that one of them is not
actually reserved by the firmware but probably not the source of
aspire1 issues (the dmic is connected to lpass tlmm anyway) or
perhaps they reserved 58 to be an interrupt input for the fp, or
something like that...

Nikita

> [...]
> 
>> +		data-pins {
>> +			pins = "sdc1_data";
>> +			drive-strength = <2>;
>> +			bias-pull-up;
> 
> here you do:
> 
> drive-strength
> bias-
> 
> [...]
> 
>> +	sdc2_default: sdc2-default-state {
>> +		clk-pins {
>> +			pins = "sdc2_clk";
>> +			bias-disable;
>> +			drive-strength = <16>;
>> +		};
> 
> here you do the inverse
> 
> (the former is preferred/more common)
> 
> otherwise
> 
> Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> Konrad