Re: [PATCH v4 2/4] arm64: dts: freescale: add initial device tree for TQMa8MPQS with i.MX8MP

[Alexander Stein <alexander.stein@ew.tq-group.com>]

Am Mittwoch, 3. Juni 2026, 19:41:47 CEST schrieb Frank Li:
> On Wed, Jun 03, 2026 at 10:05:59AM +0000, sashiko-bot@kernel.org wrote:
> > Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
> > - [High] Incorrect `reset-gpios` polarity on `dp_bridge` will permanently disable the device.
> > - [Medium] Missing `interrupts` property for the DisplayPort bridge despite hardware routing.
> > --
> >
> > commit ec4a92e3f2baae5d65ed787df020c8a40c851231
> > Author: Paul Gerber <paul.gerber@tq-group.com>
> >
> > arm64: dts: freescale: add initial device tree for TQMa8MPQS with i.MX8MP
> >
> > This adds support for TQMa8MPQS module on MB-SMARC-2 board.
> >
> > > diff --git a/arch/arm64/boot/dts/freescale/imx8mp-tqma8mpqs.dtsi b/arch/arm64/boot/dts/freescale/imx8mp-tqma8mpqs.dtsi
> > > new file mode 100644
> > > --- /dev/null
> > > +++ b/arch/arm64/boot/dts/freescale/imx8mp-tqma8mpqs.dtsi
> >
> > [ ... ]
> >
> > > +&i2c3 {
> > > +	clock-frequency = <384000>;
> > > +	pinctrl-names = "default", "gpio";
> > > +	pinctrl-0 = <&pinctrl_i2c3>;
> > > +	pinctrl-1 = <&pinctrl_i2c3_gpio>;
> > > +	scl-gpios = <&gpio5 18 (GPIO_ACTIVE_HIGH | GPIO_OPEN_DRAIN)>;
> > > +	sda-gpios = <&gpio5 19 (GPIO_ACTIVE_HIGH | GPIO_OPEN_DRAIN)>;
> > > +	status = "okay";
> > > +
> > > +	dp_bridge: dp-bridge@f {
> > > +		compatible = "toshiba,tc9595", "toshiba,tc358767";
> > > +		reg = <0xf>;
> > > +		pinctrl-names = "default";
> > > +		pinctrl-0 = <&pinctrl_tc9595>;
> > > +		clock-names = "ref";
> > > +		clocks = <&clk_dp>;
> >
> > [Severity: Medium]
> > Is this node missing the interrupt properties for the Hot Plug Detect signal?
> >
> > The pinctrl node pinctrl_tc9595 configures GPIO5_IO21 (DP_HPD_INT), but
> > without the interrupt-parent and interrupts properties here, the tc358767
> > driver won't be able to request the hardware IRQ.
> >
> > Will this cause the driver to silently fall back to 10-second software
> > polling for HPD and link training events instead of using the hardware
> > interrupt?
> >
> > > +		reset-gpios = <&expander0 4 GPIO_ACTIVE_HIGH>;
> >
> > [Severity: High]
> > Will this reset-gpios polarity leave the device permanently held in reset?
> >
> > The tc358767 driver initializes the reset GPIO to logical 0, pulses it to
> > logical 1 to assert, and returns it to logical 0 to deassert. Since this
> > is specified as GPIO_ACTIVE_HIGH, the final logical 0 maps to a physical
> > LOW signal.
> >
> > If this pin controls the bridge's Enable or active-low Reset, won't holding
> > it physically LOW leave the bridge permanently disabled? Does this need to be
> > GPIO_ACTIVE_LOW so the driver's logical 0 maps to a physical HIGH instead?
> 
> Does AI said correctly?

Currently in next there is only one position where the reset-gpio is touched in
tc358767:

/* Reset GPIO is optional */
tc->reset_gpio = devm_gpiod_get_optional(dev, "reset", GPIOD_OUT_LOW);
if (IS_ERR(tc->reset_gpio))
	return PTR_ERR(tc->reset_gpio);

if (tc->reset_gpio) {
	gpiod_set_value_cansleep(tc->reset_gpio, 1);
	usleep_range(5000, 10000);
}

So yes, the reset GPIO is set to logical 0 and to 1 afterwards. The RESX input
(reset) is indeed active-low. So AFAICS the driver is assuming
GPIO_ACTIVE_HIGH anyway.
Nevertheless right now the bridge is not functioning at all without
additional patches [1].

Without the polarity cleared/fixed in tc358767, there is not much benefit
in addressing the polarity in DT.

Best regards,
Alexander

[1] https://lore.kernel.org/all/20240625122824.148163-1-marex@denx.de/

> Frank
> 
> >
> > > +		toshiba,hpd-pin = <0>;
> > > +		status = "disabled";
> > > +
> > > +		ports {
> >
> > [ ... ]
> >
> > --
> > Sashiko AI review · https://sashiko.dev/#/patchset/20260603093621.2504490-1-alexander.stein@ew.tq-group.com?part=2
> 


-- 
TQ-Systems GmbH | Mühlstraße 2, Gut Delling | 82229 Seefeld, Germany
Amtsgericht München, HRB 105018
Geschäftsführer: Detlef Schneider, Rüdiger Stahl, Stefan Schneider
http://www.tq-group.com/