Currently only the 3-element targets are available w/musl,
which is especially unfortunate since this means when
target'ing raspberryPi w/musl ("muslpi" in our tree)
GHC sees "armv6l-unknown-linux-musleabihf" and drops
the last bit, which is bad for two reasons:

• GHC's llvm-targets file doesn't have an entry for armv6l-unknown-linux,
  preventing compilation/use in this configuration.
• Information about float mode to use (hard vs soft) is lost,
  at least in this style of target string.

I think the second reason is actually the explanation for the first,
and is why GHC special-cases "-gnueabi" which this PR extends
to similarly treat "-musleabi".

This only impacts GHC builds using musl, to avoid the rebuild (and following what was done for Android).
If later this is expanded I think only makes sense/has impact when LLVM is used,
so same configurations as in #43303.

I was "on the fence" with preserving "-musl*" (not just "-musleabi*") to ensure in a mixed
environment there's less confusion about how to extend/interpret "x86_64-unknown-linux",
but in the end decided to just match the "-gnueabi*" behavior and revisit if there's
motivation from experiences "in the wild".

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS)
• 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 nox --run "nox-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)
• Fits CONTRIBUTING.md.

Tested a number of configurations building basic packages (earlier today),
still rebuilding-all-the-haskell after a small rebase.

Inspected output of ${targetPrefix}ghc --info, although I'm not a GHC expert
but things seem as they should be before/after.