This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: RFC: Jump to const_int


> On Sun, Jan 14, 2001 at 04:46:48PM -0500, John David Anglin wrote:
> > I think that address_operand may be a slightly better choice
> > for indirect_jump.  
> 
> Depends on whether you believe the non-informative comment
> preceeding that pattern.  As it is, reload will simply force
> the operand into a register anyway.

I was thinking of using the same pattern as the m68k.  More is changed
than just the predicate.

> > Regarding situations with more than 255 arguments for a call, it
> > looks like this will now cause an abort.
> 
> No it will not.  Examine the companion change to RETURN_POPS_ARGS.

I saw that.  n_popped will be 0 when the the number of words is greater
than 255.  This will cause call/call_value insns to be generated but
rounded_stack_size_rtx won't match `const_int 0'.  The call/call_value
patterns only match the special case when there are no operands.  I
think the predicate needs to be const_int_operand for the 2nd/3rd
operand, respectively.

Dave
-- 
J. David Anglin                                  dave.anglin@nrc.ca
National Research Council of Canada              (613) 990-0752 (FAX: 952-6605)

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]