linux-kernel - Re: [PATCH v1 3/3] arm64: dts: exynos990: define all PERIC USI nodes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAPLW+4mbJoLHJ7TDcn9z8-WOEAFSiH7wBMt4Rk3fsw0h1wSSMA@mail.gmail.com>
Date: Fri, 14 Feb 2025 09:39:52 -0600
From: Sam Protsenko <semen.protsenko@...aro.org>
To: Tudor Ambarus <tudor.ambarus@...aro.org>
Cc: Denzeel Oliva <wachiturroxd150@...il.com>, alim.akhtar@...sung.com, conor+dt@...nel.org, 
	devicetree@...r.kernel.org, gregkh@...uxfoundation.org, jirislaby@...nel.org, 
	krzk+dt@...nel.org, linux-arm-kernel@...ts.infradead.org, 
	linux-kernel@...r.kernel.org, linux-samsung-soc@...r.kernel.org, 
	linux-serial@...r.kernel.org, robh@...nel.org
Subject: Re: [PATCH v1 3/3] arm64: dts: exynos990: define all PERIC USI nodes

On Fri, Feb 14, 2025 at 12:49 AM Tudor Ambarus <tudor.ambarus@...aro.org> wrote:
>
>
>
> On 2/14/25 6:17 AM, Denzeel Oliva wrote:
> > On Thu, Feb 13, 2025 at 07:38:35AM +0000, Tudor Ambarus wrote:
>
> >> node properties shall be specified in a specific order. Follow similar
> >> nodes that are already accepted, gs101 is one.
> >
> > Not all Exynos SoCs will follow the same order
>
> you an fix them then. Please follow
> https://docs.kernel.org/devicetree/bindings/dts-coding-style.html#order-of-properties-in-device-node
>
> >
> >> <&cmu_peric0 CLK_GOUT_PERIC0_TOP0_IPCLK_4>;
> >
> > Is
> >
> > GATE(CLK_GOUT_PERIC0_TOP0_IPCLK_4, "gout_peric0_top0_ipclk_4",
> >      "dout_peric0_uart_dbg",
> >      CLK_CON_GAT_GOUT_BLK_PERIC0_UID_PERIC0_TOP0_IPCLKPORT_IPCLK_4,
> >      21, 0, 0), [Mainline CLK]
>
> I don't get this reasoning, sorry.
> >
> > You can find it in the cmucal-node.c driver downstream of the kernel. [0]
> >
> >>> +                   };
> >>> +           };
> >>> +
> >>> +           usi0: usi@...500c0 {
>
> cut
>
> >>> +                   serial_2: serial@...50000 {
> >>
> >> why not serial_0 since you're in USI0.
> >
> > Because it is simply displayed in the exynos9830-usi.dtsi [1]
>
> I don't think it matters what downstream specifies for labels. Use what
> common sense says.
>

Yeah, in upstream we don't care about downstream at all. Downstream
code is very often is quite low-quality and doesn't follow mainline
kernel conventions and best practices. Mentioning the downstream code
as an answer to questions like "why it was done this way?" would often
trigger people. In this case I think it should be consecutive
ordering, if there are no very good explanation why it shouldn't be
so. Then aliases can be used in the board dts if you need to keep a
fixed order in the system.

> >
> >>> +           };
> >>> +
> >>> +           usi_i2c_0: usi@...600c0 {
> >>> +                   compatible = "samsung,exynos990-usi", "samsung,exynos850-usi";
> >>> +                   reg = <0x105600c0 0x20>;
> >>> +                   samsung,sysreg = <&sysreg_peric0 0x1008>;
> >>> +                   samsung,mode = <USI_V2_I2C>;
> >>> +                   #address-cells = <1>;
> >>> +                   #size-cells = <1>;
> >>> +                   ranges;
> >>> +                   clocks = <&cmu_peric0 CLK_GOUT_PERIC0_TOP0_PCLK_6>,
> >>> +                            <&cmu_peric0 CLK_GOUT_PERIC0_TOP0_IPCLK_6>;
> >>> +                   clock-names = "pclk", "ipclk";
> >>> +                   status = "disabled";
> >>> +
> >>> +                   hsi2c_1: i2c@...60000 {
> >>> +                           compatible = "samsung,exynos990-hsi2c",
> >>> +                                        "samsung,exynosautov9-hsi2c";
> >>> +                           reg = <0x10560000 0xc0>;
> >>> +                           interrupts = <GIC_SPI 398 IRQ_TYPE_LEVEL_HIGH>;
> >>> +                           pinctrl-names = "default";
> >>> +                           pinctrl-0 = <&hsi2c1_bus>;
> >>> +                           clocks = <&cmu_peric0 CLK_GOUT_PERIC0_TOP0_IPCLK_6>,
> >>> +                                    <&cmu_peric0 CLK_GOUT_PERIC0_TOP0_PCLK_6>;
> >>> +                           clock-names = "hsi2c", "hsi2c_pclk";
> >>> +                           #address-cells = <1>;
> >>> +                           #size-cells = <0>;
> >>> +                           status = "disabled";
> >>> +                   };
> >>
> >> shouldn't you define serial and SPI too?
> >
> > As shown in the node it only uses i2c which
> > corresponds to the exynos9830-usi.dts. [2]
>
> If you can't specify all the protocol modes for all the USI nodes, then
> make it clear in the commit message.
>