Add mold linker build

[egtvedt]

Allows building the #mold linker, which can then be used in the cross-toolchain by passing the -fuse-ld=mold to the gcc flags. It is much faster than ld or gold.

This requires a C++20 compiler and cmake.

Initially implemented by Arnaud, and HC added configure check for cmake.

Outstanding task to validate compiler is C++20 compatible.

[egtvedt]

Originally implemented in PR #2150, and I adjusted the check for cmake.

Hopefully kept original author as primary to keep credits.

[cpackham]

I've hit the thing to allow the build but as there's no sample that uses this it won't test much (other than checking that the additions don't break anything).

[egtvedt]

I've hit the thing to allow the build but as there's no sample that uses this it won't test much (other than checking that the additions don't break anything).

I ran a local build, and could produce a toolchain with mold available and appears to run as expected.

Can do a bit more testing tomorrow after revisiting your comment.

[egtvedt]

I've hit the thing to allow the build but as there's no sample that uses this it won't test much (other than checking that the additions don't break anything).

I ran a local build, and could produce a toolchain with mold available and appears to run as expected.

Can do a bit more testing tomorrow after revisiting your comment.

Quick hello world test program executes as expected on target.

Bonus is the executable size reduction from 76kB to 16kB.

[rawoul]

Sorry, I haven't had much time to follow the feedback on my PR, and I'm happy to see you are taking things forward ! I've tested toolchains with mold for i686 and aarch64 targets with no issues, building and executing a lot of linux packages, including wpewebkit, qt, etc... I had few issues in the past but with 2.31 I don't have a single problem.