Re: [RFA] sh-tdep.c: Fix float handling

[Elena Zannoni <ezannoni at redhat dot com>]

Corinna Vinschen writes:
 > On Fri, Oct 10, 2003 at 02:53:18PM -0400, Elena Zannoni wrote:
 > > Please add this expanation as a comment. I find cleared than your
 > > shorter comments. Just add it before the helper function.
 > > 
 > > >From here....
 > > 
 > >  > First, in contrast to non-FPU CPUs, arguments are never split between
 > >  > registers and stack.  If an argument doesn't fit in the remaining registers
 > >  > it's always pushed entirely on the stack.  My testing led me to the wrong
 > >  > assumption that just types > 16 bytes are pushed entirely on the stack.
 > 
 > I have a problem with this first part of the comment.  This one does
 > not fit well as a comment to the helper function.  I would suggest to
 > add this (carefully rephrased) to sh_push_dummy_call_fpu(), substituting
 > the very short
 > 
 >   /* No data is passed partially in registers.  */
 > 

Where is that? I think you can add it before the function
sh_push_dummy_call_fpu. Or where it makes more sense. I don't care, as
long as the whole thing is as clear as possible. It is wasy to
understand the code now, but I bet a few months from now we will all
have forgotten the details.

elena


 > comment instead.  Ok?
 > 
 > > Since this is a predicate, can you call it sh_treat_as_flt_p .
 > 
 > Ok.
 > 
 > > Ok otherwise.
 > 
 > Thanks, I'd just like to hear about the above comment.
 > 
 > > BTW, what multilibs are you testing?
 > 
 > Several, -m2, -m2e, m3e, m4, m4 -ml, -m4-single.
 > 
 > Corinna
 > 
 > -- 
 > Corinna Vinschen
 > Cygwin Developer
 > Red Hat, Inc.