Compiling with BTI enforcement

Help
,
1

I wanted to compile Zig code with BTI enabled. I tried to add CPU feature:

    const aarch64_features = std.Target.aarch64.Feature;

    var module_target_query: std.Target.Query = .{
        .cpu_arch = .aarch64,
        .os_tag = .freestanding,
    };
    module_target_query.cpu_features_add.addFeature(
        @intFromEnum(aarch64_features.bti),
    );

    const obj = b.addObject(.{
        .name = "obj",
        .root_module = b.createModule(.{
            .root_source_file = b.path("main.zig"),
            .target = module_target,
            .optimize = optimize,
            .pic = false,
        }),
        .use_llvm = true,
    });

but it only tells LLVM that CPU supports it, which makes this flag useless.

I have to emit LLVM IR file, manually add “branch-target-enforcement”=“true” to attributes, and compile the IR.
So is there any more convenient way to add BTI enforcement? Can I somehow pass args to LLVM from build? Is it planned to have such arch-specific build options?