machinst ABI: Pass fixed frame size to gen_clobber_restore

[uweigand]

The ABI common code currently passes the fixed frame size to
the gen_clobber_save back-end routine, which is required to
emit code to allocate the required stack space in the prologue.

Similarly, the back-end needs to emit code to de-allocate the
stack in the epilogue. However, at this point the back-end
does not have access to that fixed frame size value any more.
With targets that use a frame pointer, this does not matter,
since de-allocation can be done simply by assigning the frame
pointer back to the stack pointer. However, on targets that
do not use a frame pointer, the frame size is required.

To allow back-ends that option, this patch changes ABI common
code to pass the fixed frame size to get_clobber_restore as
well (computed identically to how it is done for get_clobber_save).

[uweigand]

@cfallin And this is the other change discussed here #2128 (comment) (required for the IBM Z back-end).

[bjorn3]

Suggested change

	let mut fixed_frame_storage_size = 0;
	if !self.call_conv.extends_baldrdash() {
	fixed_frame_storage_size += self.frame_size();
	}
	let fixed_frame_storage_size = if !self.call_conv.extends_baldrdash() {
	self.frame_size()
	} else {
	0
	};

Also why this condition?

[uweigand]

If you look at the code in gen_prologue above, it passes a fixed_frame_storage_size of 0 to gen_clobber_save with the baldrdash calling convention. For consistency, we need to do the same for gen_clobber_restore.

Likewise, I followed the same coding convention with the "mut fixed_frame_storage_size" as is currently in place in gen_prologue. Given that any future changes to the computation of the size will need to be kept in sync between gen_prologue and gen_epilogue, it seems preferable to me to also have the code itself as similar as possible.

[cfallin]

I think it would be better to simply store the fixed_frame_storage_size in self -- we get a mutable self in gen_prologue so we should be able to write it there. This avoids the duplication of logic and potential for mismatches...

[uweigand]

OK, updated the patch accordingly.

[cfallin]

Thanks! As noted below, I think there may be a better way to factor things, but it's a sensible change overall.

[cfallin]

I think it would be better to simply store the fixed_frame_storage_size in self -- we get a mutable self in gen_prologue so we should be able to write it there. This avoids the duplication of logic and potential for mismatches...

[cfallin]

Thanks!