x86_64,riscv64: fix incorrect incoming_stack_alignment handling

The whole motivation behind this proposal in the first place was that
the LLVM backend disagrees with the self-hosted backends on what
`@setAlignStack` meant, so we can't just translate the old logic to the
new system! These backends can introduce support for overriding
`incoming_stack_alignment` later on.

src/Zcu.zig

@@ -3575,7 +3575,8 @@ pub fn callconvSupported(zcu: *Zcu, cc: std.builtin.CallingConvention) union(enu
             else => false,
         },
         .stage2_x86_64 => switch (cc) {
-            .x86_64_sysv, .x86_64_win, .naked => true, // stack alignment supported
+            .x86_64_sysv, .x86_64_win => |opts| opts.incoming_stack_alignment == null,
+            .naked => true,
             else => false,
         },
         .stage2_aarch64 => switch (cc) {

src/arch/riscv64/CodeGen.zig

@@ -820,10 +820,7 @@ pub fn generate(
     try function.frame_allocs.resize(gpa, FrameIndex.named_count);
     function.frame_allocs.set(
         @intFromEnum(FrameIndex.stack_frame),
-        FrameAlloc.init(.{
-            .size = 0,
-            .alignment = .fromByteUnits(target_util.stackAlignment(function.target.*, fn_type.fnCallingConvention(zcu))),
-        }),
+        FrameAlloc.init(.{ .size = 0, .aligment = .@"1" }),
     );
     function.frame_allocs.set(
         @intFromEnum(FrameIndex.call_frame),

src/arch/x86_64/CodeGen.zig

@@ -871,10 +871,7 @@ pub fn generate(
     try function.frame_allocs.resize(gpa, FrameIndex.named_count);
     function.frame_allocs.set(
         @intFromEnum(FrameIndex.stack_frame),
-        FrameAlloc.init(.{
-            .size = 0,
-            .alignment = .fromByteUnits(target_util.stackAlignment(function.target.*, fn_type.fnCallingConvention(zcu))),
-        }),
+        FrameAlloc.init(.{ .size = 0, .alignment = .@"1" }),
     );
     function.frame_allocs.set(
         @intFromEnum(FrameIndex.call_frame),