Re: Your Jan 15 change broke the x86

[Jeffrey A Law <law at cygnus dot com>]

  In message <19990131131438.A17484@cygnus.com>you write:
  > On Sun, Jan 31, 1999 at 01:46:55PM -0700, Jeffrey A Law wrote:
  > > I'm a little suprised "bm" was shoved into memory.
  > 
  > This is -O0, remember?  Of course it lives in memory.
Oh yea.  Duh.


  > Yes, but they allow memory only of a very particular form.  And
  > at rtl generation time that form _only_ exists in push instructions
  > created by expand_call.  So there's no point in matching them.
Ah.  Yes, the effected function is only called during rtl generation.

  > Or is there?  Come to think of it, how can you possibly usefully
  > use "<" by itself as an input constraint from an asm?  There does
  > not appear to be any code whatsoever to create an address of the
  > proper form, so the only way it could be used is in conjunction
  > with another memory specifier.  Even then I can't see how specifying
  > "o<>" wins over "m" -- in an asm you can't select different code
  > sequences based on alternative or whatnot.  
Are autoinc addresses always offsettable?  A reloaded autoinc is offsettable,
but I'm not sure a raw autoinc is offsettable.

Also some targets only allow limited autoinc addressing -- so an autoinc may
not match a 'm'.  Consider a target which can do postinc loads and preinc
stores only.  For such a port 'm' would not match an autoinc address.

jeff