compiler-rt: fix build on armv7l

[thefloweringash]

Motivation for this change

Fixes #87266 - llvmPackages_7.compiler-rt fails to produce output path on armv7l

Still testing the builds, so opening as draft.

See also same change for x86 #85945

cc @matthewbauer @Atemu @TravisWhitaker

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Determined the impact on package closure size (by running nix path-info -S before and after)
• Ensured that relevant documentation is up to date
• Fits CONTRIBUTING.md.

[matthewbauer]

Nice work! We probably should consolidate this with set(X86 i386 i486 i586 i686) and try to get it accepted upstream.

[thefloweringash]

Successfully built for llvmPackages_{5,6,7,8,9,10}.compiler-rt. For llvmPackages_{8,10} I had to set doCheck = false in llvm due to test failures.

We probably should consolidate this with set(X86 i386 i486 i586 i686) and try to get it accepted upstream.

I'm following the pattern in #85945 with enough changes to make it work. I didn't look deeply into why this is an appropriate fix. Someone else would need to write the cover letter for upstreaming.

Incidentally, has this been tested in the i486 and i586 cases? I only see a definition of i386_sources and i686_sources in the CMakeLists.txt. I wouldn't be surprised if it fails with the same error as in #87266 (comment) for i486 and i586

[Atemu]

Hm, is it actually still causing a rebuild on Darwin or is ofBorg bugged?

[doronbehar]

Related: #87528