This is the mail archive of the gcc@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]
Other format: [Raw text]

reload and asms


Hi,
I've tripped over this code in reload_as_needed (reload1.c),
	      /* If this was an ASM, make sure that all the reload insns
		 we have generated are valid.  If not, give an error
		 and delete them.  */

	      if (asm_noperands (PATTERN (insn)) >= 0)
		for (p = NEXT_INSN (prev); p != next; p = NEXT_INSN (p))
		  if (p != insn && INSN_P (p)
		      && GET_CODE (PATTERN (p)) != USE
		      && (recog_memoized (p) < 0
			  || (extract_insn (p), ! constrain_operands (1))))
		    {
		      error_for_asm (insn,
				     "`asm' operand requires impossible reload");
		      delete_insn (p);
		    }
It is generating false positives, because at this point the source operands
of the P insns are still pseudos, hence constrain_operands (1), fails,
even though things will all work out ok.  A second problem is that
extract_insn makes the insn memoized, so merely removing the
constrain_operands call makes us fall over later, should the move's source
operand be a stack slot for instance, because we'd have remembered it
as a reg-reg move.

Removing this block of code has not caused any problems, but I'm
at a loss to understand what legitimate cases it would detect.  Anybody
want to clue me in?

nathan

--
Nathan Sidwell    ::   http://www.codesourcery.com   ::     CodeSourcery LLC
nathan@codesourcery.com    ::     http://www.planetfall.pwp.blueyonder.co.uk



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