Re: [PATCH] arm64: dts: qcom: qcm6490-idp: Enable PCIe1

[Sushrut Shree Trivedi]

On 2/12/2026 5:49 PM, Konrad Dybcio wrote:
> On 2/12/26 1:06 PM, Sushrut Shree Trivedi wrote:
> > Remove PCIe1 clocks from protected-list and enable PCIe1 controller
> > and its corresponding PHY nodes on qcm6490-idp platform.
> >
> > PCIe1 is used to connect NVMe based SSD's on this platform.
> Is that a M.2 slot? What key (B/M etc.)?
The NVMe is actually soldered onto the board.
> [...]
>
> >   &pm7250b_gpios {
> >   	lcd_disp_bias_en: lcd-disp-bias-en-state {
> >   		pins = "gpio2";
> > @@ -920,6 +931,22 @@ &tlmm {
> >   	gpio-reserved-ranges = <32 2>, /* ADSP */
> >   			       <48 4>; /* NFC */
> >   
> > +	pcie1_reset_n: pcie1-reset-n-state {
> > +		pins = "gpio2";
> > +		function = "gpio";
> > +		drive-strength = <16>;
> > +		output-low;
> You're asserting the active state of a pin permanently this way, unless
> the driver takes over, please drop this line
If we de-assert the reset, it is an indication to the endpoint that it 
can participate
in link-training with the host. Hence, until the host driver is probed 
and necessary
resources enabled, we keep the perst asserted so endpoint doesn't try to
participate in link training.
> > +		bias-disable;
> > +		};
> Wrong indentation
Ack'd.
> > +
> > +	pcie1_wake_n: pcie1-wake-n-state {
> > +		pins = "gpio3";
> > +		function = "gpio";
> > +		drive-strength = <2>;
> > +		bias-pull-up;
> > +	};
> > +
> > +
> Double \n
Ack'd.
> >   	sd_cd: sd-cd-state {
> >   		pins = "gpio91";
> >   		function = "gpio";
> >
> > ---
> > base-commit: 4f938c7d3b25d87b356af4106c2682caf8c835a2
> > change-id: 20260212-qcm6490-idp-24f7b6a1812d
> >
> > Best regards,