Add umul_ppmm to gmp-arch.hdoc

To enable “longlong.h” removal, the umul_ppmm is moved to a gmp-arch.h.
The generic implementation now uses a static inline, which provides
better type checking than the GNU extension to cast the asm constraint
(and it works better with clang).

Most of the architecture uses the generic implementation, which is
expanded from a macro, except for alpha, arm, hppa, x86, m68k, mips,
powerpc, and sparc.  The 32 bit architectures the compiler generates
good enough code using uint64_t types, where for 64 bit architecture
the patch leverages the math_u128.h definitions that uses 128-bit
integers when available (all 64 bit architectures on gcc 15).

Reviewed-by: Wilco Dijkstra  <Wilco.Dijkstra@arm.com>

diff --git a/sysdeps/generic/gmp-arch.h b/sysdeps/generic/gmp-arch.h
index 3adab107329afe6ee7f2b12449c1b74df7a480e8..f1dba1384bc1fa8682fdf8be617233d6274135eb 100644 (file)
--- a/sysdeps/generic/gmp-arch.h
+++ b/sysdeps/generic/gmp-arch.h
@@ -37,6 +37,32 @@ ll_highpart (mp_limb_t t)
   return t >> (BITS_PER_MP_LIMB / 2);
 }
 
+/* umul_ppmm(high_prod, low_prod, multiplier, multiplicand) multiplies two
+   UWtype integers MULTIPLIER and MULTIPLICAND, and generates a two UWtype
+   word product in HIGH_PROD and LOW_PROD.  */
+static __always_inline void
+umul_ppmm_generic (mp_limb_t *w1, mp_limb_t *w0, mp_limb_t u, mp_limb_t v)
+{
+#if __WORDSIZE == 32
+  uint64_t t0 = (uint64_t)u * v;
+  *w1 = t0 >> 32;
+  *w0 = t0;
+#else
+  u128 r = u128_mul(u128_from_u64 (u), u128_from_u64 (v));
+  *w1 = u128_high (r);
+  *w0 = u128_low (r);
+#endif
+}
+#undef umul_ppmm
+#define umul_ppmm(__w1, __w0, __u, __v)                        \
+  ({                                                   \
+    __typeof (__w0) __w0t;                             \
+    __typeof (__w1) __w1t;                             \
+    umul_ppmm_generic (&__w1t, &__w0t, __u, __v);      \
+    __w1 = __w1t;                                      \
+    __w0 = __w0t;                                      \
+  })
+
 /* udiv_qrnnd(quotient, remainder, high_numerator, low_numerator,
    denominator) divides a UDWtype, composed by the UWtype integers
    HIGH_NUMERATOR and LOW_NUMERATOR, by DENOMINATOR and places the quotient