[Platform] Does the Chip-tool support 32 bit Linux?

[qq8512852]

Reproduction steps / Feature

Hello Matter teams,
I saw the description of the building instruction of Chip-tool which as follow:
Before you can use the CHIP Tool, you must compile it from source on Linux (amd64/aarch64) or macOS. If you want to run it on Raspberry Pi, it must use a 64-bit OS.

So if the 32-bit OS is, how can I support the Chip Tool?
Background:
I have a 32-bit Linux device. Now I want to integrate Docker(build OpenThread border router) and chip tool. Then I can directly control the node device on my border router device with Chip Tool.

Platform

other

Platform Version(s)

No response

Type

Platform validated

(Optional) If manually tested please explain why this is only manually tested

No response

Anything else?

No response

[stlevkov]

You can, with a proper toolchain. Check the GN buld files on how to set a toolchain using the sysroot of your linux device.

[qq8512852]

You can, with a proper toolchain. Check the GN buld files on how to set a toolchain using the sysroot of your linux device.

Hi stlevkov, I tried to command gn gen out/arm --args='target_cpu="arm"' under the folder connectectedhomeip/examples/chip-tool. And then execute "ninja -C out/arm" ,the result as follow:

_/usr/include/glib-2.0/glib/gtypes.h: In function ‘gboolean GLIB_CHECKED_ADD_U64(guint64*, guint64, guint64)’: /usr/include/glib-2.0/glib/gmacros.h:742:31: error: static assertion failed: Expression evaluates to false #define G_STATIC_ASSERT(expr) static_assert (expr, "Expression evaluates to false") ^ /usr/include/glib-2.0/glib/gtypes.h:463:3: note: in expansion of macro ‘G_STATIC_ASSERT’ G_STATIC_ASSERT(sizeof (unsigned long long) == sizeof (guint64));

Please give some tips. Thanks.

[stlevkov]

Hi, There are two options to compile it for 32 bit:

• Compile directly on the target machine (you would need all of the libraries like avahi, glib, OpenSSL installed)
• Compile on the host machine (Linux amd64) using cross-compiler. You will also need all of the required libraries compiled as static or link them dynamically. You need to set the toolchain sysroot and the cross-compiler in the corresponding GN Build file for the chip-tool. I would recommend to read the GN Build documentation on how to use cross-compilation setup.

[qq8512852]

Hi, There are two options to compile it for 32 bit:

• Compile directly on the target machine (you would need all of the libraries like avahi, glib, OpenSSL installed)
• Compile on the host machine (Linux amd64) using cross-compiler. You will also need all of the required libraries compiled as static or link them dynamically. You need to set the toolchain sysroot and the cross-compiler in the corresponding GN Build file for the chip-tool. I would recommend to read the GN Build documentation on how to use cross-compilation setup.

Thanks for your help. Already done with 32 bit toolchain and work on my arm chip smoothly.

[taigila]

@qq8512852 hi sir, currently I am trying to cross-compile chip-tool on a custom toolchain for arm64 machine (with my amd64 based ubuntu 22.04). I already built some libraries (openssl, glib2.0) with my toolchain and put those in a directory, and try to direct gn and ninja to build with the libraries that i built. i have used --system_libdir and when i wanna compile using ninja - C, they still look for the libraries on the default ubuntu 22.04 libraries. Do you know can i change that? thank you