C implementation of mod_1_1

[Torbjorn Granlund]

Torbjorn Granlund <tg at gmplib.org> writes:

    /* FIXME: Needs review and/or testing. */
    #if 0 && defined (__sparc__) && W_TYPE_SIZE == 64
    #define add_mssaaaa(m, sh, sl, ah, al, bh, bl)				\
      __asm__ (								\
          "addcc	%r5,%6,%2\n"						\
          "	addccc	%r7,%8,%%g0\n"						\
          "	addc	%r3,%4,%1\n"						\
          "	clr	%0\n"							\
          "	movcs	%%xcc, -1, %0\n"					\
           : "=r" (m),"=r" (sh), "=&r" (sl)					\
           : "rJ" (ah), "rI" (bh), "%rJ" (al), "rI" (bl),			\
    	 "rJ" ((al) >> 32), "rI" ((bl) >> 32),				\
    	 __CLOBBER_CC)
    #endif
    
  I cannot see how carry-out from ah.bh works.
  
Sorry, I am confused.  If addc means eat carry from icc and generate
carry to xcc, themn the code is probably correct.

I find it best to test these types of things before staring too much at
them.

    /* FIXME: Needs review and/or testing. I don't understand why
       constraints says s0 (%2) and a0 (%6) must share a register. */
    #if 0 && HAVE_HOST_CPU_FAMILY_powerpc && W_TYPE_SIZE == 64
    #define add_mssaaaa(m, s1, s0, a1, a0, b1, b0)                         \
      __asm__ (								\
           "add%I6c	%2,%5,%6\n"						\
          "	adde	%1,%3,%4\n"						\
          "	subfe	%0,%0,%0\n"						\
          "	neg	%0, %0"							\
             : "=r" (m), "=r" (s1), "=&r" (s0)                          \
             : "r"  ((UDItype)(a1)), "r" ((UDItype)(b1)),                 \
               "%2" ((UDItype)(a0)), "rI" ((UDItype)(b0)))
    #endif

  This cannot be right.  Replace the neg with a subfi %0,%0,-1.

Clarification: subtract-with-carry on ppc processors complement the
subtrahend, and add the carry flag and minuend to that.  I.e., CF=1
means no carry.

The binding 5->2 looks bogus.  The comment about 6->2 is wrong...

-- 
Torbjörn