| 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
| ||
|
Message-ID: <8fba3bb0-cfb5-4427-97d0-0cf9493a5bcf@cjdns.fr>
Date: Sat, 22 Mar 2025 00:43:42 +0100
From: Caleb James DeLisle <cjd@...ns.fr>
To: Krzysztof Kozlowski <krzk@...nel.org>, linux-mips@...r.kernel.org
Cc: Thomas Gleixner <tglx@...utronix.de>, Rob Herring <robh@...nel.org>,
Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley
<conor+dt@...nel.org>, Thomas Bogendoerfer <tsbogend@...ha.franken.de>,
Daniel Lezcano <daniel.lezcano@...aro.org>, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org, benjamin.larsson@...exis.eu
Subject: Re: [PATCH v1 7/8] mips: Add EcoNet MIPS platform support
On 21/03/2025 22:00, Krzysztof Kozlowski wrote:
> On 21/03/2025 14:46, Caleb James DeLisle wrote:
>> Add platform support for EcoNet MIPS SoCs.
>>
>> Signed-off-by: Caleb James DeLisle <cjd@...ns.fr>
>> ---
>> As is typical of embedded platforms, it's not realistic to imagine building
>> a fully functioning system from kernel sources alone. In the interest of
>> providing something without external dependencies, I have included build and
>> device tree for a minimal testing / PoC image that can be booted from memory
>> on these devices.
>> ---
>> arch/mips/Kbuild.platforms | 1 +
>> arch/mips/Kconfig | 25 ++++++
>> arch/mips/boot/compressed/uart-16550.c | 5 ++
>> arch/mips/boot/dts/Makefile | 1 +
>> arch/mips/boot/dts/econet/Makefile | 2 +
>> arch/mips/boot/dts/econet/en751221.dtsi | 62 +++++++++++++++
>> .../boot/dts/econet/en751221_test_image.dts | 19 +++++
>> arch/mips/econet/Kconfig | 42 ++++++++++
>> arch/mips/econet/Makefile | 2 +
>> arch/mips/econet/Platform | 5 ++
>> arch/mips/econet/init.c | 78 +++++++++++++++++++
>> 11 files changed, 242 insertions(+)
>> create mode 100644 arch/mips/boot/dts/econet/Makefile
>> create mode 100644 arch/mips/boot/dts/econet/en751221.dtsi
>> create mode 100644 arch/mips/boot/dts/econet/en751221_test_image.dts
> DTS are always, always separate patches. See also DTS coding style and
> submitting patches in the bindings, which already covers this.
Ok got it, somehow I though arch/mips/* made it different.
>
>
> ...
>
>
>> new file mode 100644
>> index 000000000000..e4404aed5705
>> --- /dev/null
>> +++ b/arch/mips/boot/dts/econet/en751221.dtsi
>> @@ -0,0 +1,62 @@
>> +// SPDX-License-Identifier: GPL-2.0
>> +/dts-v1/;
>> +
>> +/ {
>> + compatible = "econet,en751221";
>> + #address-cells = <1>;
>> + #size-cells = <1>;
>> +
>> + hpt_clock: hpt_clock {
> No underscores in node names.
>
> Follow DTS coding style.
Right, I know this, my fault.
>
>> + compatible = "fixed-clock";
>> + #clock-cells = <0>;
>> + clock-frequency = <200000000>; /* 200 MHz */
>> + };
>> +
>> + cpus: cpus {
>> + #address-cells = <1>;
>> + #size-cells = <0>;
>> +
>> + cpu@0 {
>> + device_type = "cpu";
>> + compatible = "mips,mips24KEc";
>> + reg = <0>;
>> + };
>> + };
>> +
>> + cpuintc: interrupt-controller {
>> + #address-cells = <0>;
>> + #interrupt-cells = <1>;
>> + interrupt-controller;
>> + compatible = "mti,cpu-interrupt-controller";
>> + };
>> +
>> + intc: interrupt-controller@...40000 {
>> + compatible = "econet,en751221-intc";
>> + reg = <0x1fb40000 0x100>;
>> + interrupt-parent = <&cpuintc>;
>> + interrupts = <2>;
>> +
>> + interrupt-controller;
>> + #interrupt-cells = <1>;
>> + econet,shadow-interrupts = <7 2>, <8 3>, <13 12>, <30 29>;
>> + };
>> +
>> + uart: serial@...f0000 {
>> + compatible = "ns16550";
>> + reg = <0x1fbf0000 0x30>;
>> + reg-io-width = <4>;
>> + reg-shift = <2>;
>> + interrupt-parent = <&intc>;
>> + interrupts = <0>;
>> + clock-frequency = <1843200>;
>> + };
>> +
>> + timer_hpt: timer_hpt@...f0400 {
> Same problem as with binding.
Will fix.
>
>> + compatible = "econet,timer-hpt";
>> + reg = <0x1fbf0400 0x100>;
>> +
>> + interrupt-parent = <&intc>;
>> + interrupts = <30>;
>> + clocks = <&hpt_clock>;
>> + };
>> +};
>> diff --git a/arch/mips/boot/dts/econet/en751221_test_image.dts b/arch/mips/boot/dts/econet/en751221_test_image.dts
>> new file mode 100644
>> index 000000000000..bc140c4043b2
>> --- /dev/null
>> +++ b/arch/mips/boot/dts/econet/en751221_test_image.dts
> Does not look like a board. We do not take some testing/debugging
> thingies. Please upstream *real* board.
No problem, will swap this out for a cheap available modem that I like to dev on.
>
>> @@ -0,0 +1,19 @@
>> +// SPDX-License-Identifier: GPL-2.0
>> +/dts-v1/;
>> +
>> +#include "en751221.dtsi"
>> +
>> +/ {
>> + model = "Generic EN751221";
> Missing compatible.
>
>> +
>> + memory@0 {
>> + /* We hope at least 64MB will be available wherever we are run */
>> + device_type = "memory";
>> + reg = <0x00000000 0x4000000>;
>> + };
>> +
>> + chosen {
>> + bootargs = "console=ttyS0,115200";
> Drop bootargs and use standard property - stdout.
>
> See how all other platforms are doing it (and not some ancient MIPS, but
> the most recent arm64 or riscv).
Sure thing.
>
>
>> + linux,usable-memory-range = <0x00020000 0x3fe0000>;
>> + };
>> +};
>> diff --git a/arch/mips/econet/Kconfig b/arch/mips/econet/Kconfig
>> new file mode 100644
>> index 000000000000..12f85d638e47
>> --- /dev/null
>> +++ b/arch/mips/econet/Kconfig
>> @@ -0,0 +1,42 @@
>> +# SPDX-License-Identifier: GPL-2.0
>> +if ECONET
>> +
>> +config SOC_ECONET_EN751221
>> + bool
>> + select COMMON_CLK
>> + select ECONET_EN751221_INTC
>> + select IRQ_MIPS_CPU
>> + select SMP
>> + select SMP_UP
>> + select SYS_SUPPORTS_SMP
>> +
>> +choice
>> + prompt "EcoNet SoC selection"
>> + default SOC_ECONET_EN751221_FAMILY
>> + help
>> + Select EcoNet MIPS SoC type. Individual SoCs within a family are
>> + similar enough that is it enough to select the right family, and
>> + then customize to the specific SoC using the device tree only.
>> +
>> + config SOC_ECONET_EN751221_FAMILY
>> + bool "EN751221 family"
>> + select SOC_ECONET_EN751221
>> + help
>> + The EN751221 family includes EN7512, RN7513, EN7521, EN7526.
>> + They are based on single core MIPS 34Kc processors. To boot
>> + this kernel, you will need a device tree such as
>> + MIPS_RAW_APPENDED_DTB=y, and a root filesystem.
>> +
>> + config SOC_ECONET_EN751221_TEST_IMAGE
>> + bool "EN751221 test image"
>> + select SOC_ECONET_EN751221
>> + select BUILTIN_DTB
>> + help
>> + Build a minimal kernel that will boot on any EN751221 board
>> + with at least 64MB of memory. This has a builtin device tree
>> + so it can boot with nothing more than an appended initramfs.
>> + This is good for validating that a given SoC is EN751221
>> + compatible, or for regression testing.
>> +endchoice
>> +
>> +endif
>> diff --git a/arch/mips/econet/Makefile b/arch/mips/econet/Makefile
>> new file mode 100644
>> index 000000000000..7e4529e7d3d7
>> --- /dev/null
>> +++ b/arch/mips/econet/Makefile
>> @@ -0,0 +1,2 @@
>> +
>> +obj-y := init.o
>> diff --git a/arch/mips/econet/Platform b/arch/mips/econet/Platform
>> new file mode 100644
>> index 000000000000..bb659876d855
>> --- /dev/null
>> +++ b/arch/mips/econet/Platform
>> @@ -0,0 +1,5 @@
>> +# To address a 7.2MB kernel size limit in the EcoNet SDK bootloader,
>> +# we put the load address well above where the bootloader loads and then use
>> +# zboot. So please set CONFIG_ZBOOT_LOAD_ADDRESS to the address where your
>> +# bootloader actually places the kernel.
>> +load-$(CONFIG_ECONET) += 0xffffffff81000000
>> \ No newline at end of file
> You have patch warnings
"No newline at end of file" - that's embarrassing, I'll try to figure out how that got into the patch.
Thank you very much for your time.
Thanks,
Caleb
>
> Best regards,
> Krzysztof
Powered by blists - more mailing lists