Two problems with current_function_pretend_args_size

[Jim Wilson]

On Fri, 2003-09-12 at 02:45, Richard Sandiford wrote:
> Yes, the testcase fails because the caller & callee think the arguments
> are in different places.

OK.  That is a good enough reason for an ABI change.

> locate_and_pad_parm isn't the only issue though.  init_emit will set
> REGNO_POINTER_ALIGN (ARG_POINTER_REGNUM) to STACK_BOUNDARY. 

That is a good point.  This is sufficient reason to use STACK_BOUNDARY.

> Yes, either += or = should be OK.  I'm happy to use = if you think
> that's stylistically better.

I don't really care.  Either way is fine with me.  I was mainly trying
to make sure I understood the patch correctly.

The patch is OK.
-- 
Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com