Re: [PATCH v5 05/24] ARM: tegra: Add device-tree for ASUS Transformer EeePad TF101

From: Dmitry Osipenko <digetx@gmail.com>
To: Thierry Reding <thierry.reding@gmail.com>
Cc: "Jonathan Hunter" <jonathanh@nvidia.com>,
	"David Heidelberg" <david@ixit.cz>,
	"Svyatoslav Ryhel" <clamor95@gmail.com>,
	"Anton Bambura" <jenneron@protonmail.com>,
	"Antoni Aloy Torrens" <aaloytorrens@gmail.com>,
	"Nikola Milosavljevic" <mnidza@outlook.com>,
	"Ion Agorria" <ion@agorria.com>,
	"Michał Mirosław" <mirq-linux@rere.qmqm.pl>,
	"Ihor Didenko" <tailormoon@rambler.ru>,
	"Andreas Westman Dorcsak" <hedmoo@yahoo.com>,
	"Maxim Schwalm" <maxim.schwalm@gmail.com>,
	"Raffaele Tranquillini" <raffaele.tranquillini@gmail.com>,
	"Jasper Korten" <jja2000@gmail.com>,
	"Thomas Graichen" <thomas.graichen@gmail.com>,
	"Stefan Eichenberger" <stefan.eichenberger@toradex.com>,
	devicetree@vger.kernel.org, linux-tegra@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v5 05/24] ARM: tegra: Add device-tree for ASUS Transformer EeePad TF101
Date: Fri, 10 Dec 2021 19:42:18 +0300	[thread overview]
Message-ID: <de357dd7-ad4b-758c-f041-b3385d164b98@gmail.com> (raw)
In-Reply-To: <YbN2T5guOfIRLXg1@orome>

10.12.2021 18:46, Thierry Reding пишет:
...
>> +	gpio@6000d000 {
>> +		charging-enable-hog {
>> +			gpio-hog;
>> +			gpios = <TEGRA_GPIO(R, 6) GPIO_ACTIVE_HIGH>;
>> +			output-low;
>> +		};
>> +	};
> 
> Isn't this something that we may want to change at some point? My
> understanding is that GPIO hogs are permanent, so it won't be possible
> to grab GPIO R.6 and change this.
> 
> Are there any plans to allow setting this at runtime?

I'm not aware about plans to change that charging hog.

It's not a problem to remove the hog. I don't understand why you're
saying that it's permanent. We have such hogs in Nexus7 DT for the 3G
modem pins. If we'll ever have a driver for that modem, then we will
remove those hogs, not a problem.

> [...]
>> +	i2c2: i2c@7000c400 {
>> +		status = "okay";
>> +		clock-frequency = <100000>;
>> +	};
>> +
>> +	i2cmux {
> 
> This doesn't belong here. The ordering is by unit-address and everything
> without unit-address needs to move after the nodes with unit-addresses
> and be sorted alphabetically.

Logically it belongs here, since mux uses i2c2. All other DTs do the same.

...
>> +	memory-controller@7000f400 {
>> +		nvidia,use-ram-code;
>> +
>> +		emc-tables@3 {
>> +			reg = <0x3>;
>> +
>> +			#address-cells = <1>;
>> +			#size-cells = <0>;
>> +
>> +			lpddr2 {
>> +				compatible = "elpida,B8132B2PB-6D-F", "jedec,lpddr2-s4";
>> +				revision-id1 = <1>;
>> +				density = <2048>;
>> +				io-width = <16>;
>> +			};
>> +
>> +			emc-table@25000 {
> 
> Ugh... looking at the bindings for this the naming here is rather
> unfortunate. emc-tables@3 and emc-table@*... the top-level emc-tables
> are really tables, but the emc-table@* are really entries or rows in
> that table, not tables themselves. It's also rather unfortunate that we
> duplicate the frequency in both the "reg" and "clock-frequency"
> properties. One of them would've been enough.

It's emc-tables@3 here because initially there were other tables that I
removed during preparation of the DT for upstream. Those tables were
untested and looked questionable to me.

> Anyway, looks like this has basically been like this since forever, so
> not much that can be done about it.
> 
> Again, the memory-controller node needs to be sorted differently. There
> are other occurrences of this throughout the file.

Please feel free to reorder it.