This is the mail archive of the gcc-patches@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]

Re: [PATCH] Compute mulwiden costs exactly on i386


On Mon, Mar 08, 2004 at 02:49:30AM -0500, Bonzini wrote:
> ! 	  /* Compute costs correctly for widening multiplication.  */
> ! 	  if ((GET_CODE (op0) == SIGN_EXTEND || GET_CODE (op1) == ZERO_EXTEND)
> ! 	      && GET_CODE (op0) == GET_CODE (op1)
> ! 	      && GET_MODE_SIZE (GET_MODE (XEXP (op0, 0))) * 2
> ! 	         == GET_MODE_SIZE (mode))

Not correct for immediates, since (ze int) is invalid rtl.


r~


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