R: disable stack protector on aarch64-darwin

[collares]

Motivation for this change

An attempt at fixing #158730, copied from #128606. I don't have an aarch64-darwin machine to test it on, but ofBorg seems happy.

cc @ConnorBaker

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandbox = true set in nix.conf? (See Nix manual)
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 22.05 Release Notes (or backporting 21.11 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
  • (Release notes changes) Ran nixos/doc/manual/md-to-db.sh to update generated release notes
• Fits CONTRIBUTING.md.

[collares]

The correct fix for this is #151983, since it is a gfortran problem. The present PR should be an acceptable stopgap, though, since the workaround is already used in a few places in nixpkgs.

[ConnorBaker]

@collares this does fixes the issue I was running into -- what are the possible repercussions of disabling stack protector hardening?

[collares]

Disabling the stack protector is a bad idea If you run R on malicious inputs that could cause buffer/stack overflows through exploiting R security bugs. If you're not worried about R security bugs (because you only run code you trust, say), you should be good. See https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html for more details.

[7c6f434c]

Isn't malicious R code a problem on its own without sack smashing? I guess there could be about maliciously crafted input data to attack the standard library functions or something like that.

[smaret]

This builds and runs fine on x86_darwin.