Add libgmp 6.2.0 to third_party
Don't build it yet. That will come in the next review.
Change-Id: Idf3266558165e5ab45f4a41c98cc8c838c8244d5
diff --git a/third_party/gmp/mpf/sqrt_ui.c b/third_party/gmp/mpf/sqrt_ui.c
new file mode 100644
index 0000000..9f91f99
--- /dev/null
+++ b/third_party/gmp/mpf/sqrt_ui.c
@@ -0,0 +1,108 @@
+/* mpf_sqrt_ui -- Compute the square root of an unsigned integer.
+
+Copyright 1993, 1994, 1996, 2000, 2001, 2004, 2005, 2015 Free Software
+Foundation, Inc.
+
+This file is part of the GNU MP Library.
+
+The GNU MP Library is free software; you can redistribute it and/or modify
+it under the terms of either:
+
+ * the GNU Lesser General Public License as published by the Free
+ Software Foundation; either version 3 of the License, or (at your
+ option) any later version.
+
+or
+
+ * the GNU General Public License as published by the Free Software
+ Foundation; either version 2 of the License, or (at your option) any
+ later version.
+
+or both in parallel, as here.
+
+The GNU MP Library is distributed in the hope that it will be useful, but
+WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
+or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+for more details.
+
+You should have received copies of the GNU General Public License and the
+GNU Lesser General Public License along with the GNU MP Library. If not,
+see https://www.gnu.org/licenses/. */
+
+#include <stdio.h> /* for NULL */
+#include "gmp-impl.h"
+
+
+/* As usual the aim is to produce PREC(r) limbs of result with the high limb
+ non-zero. That high limb will end up floor(sqrt(u)), and limbs below are
+ produced by padding the input with zeros, two for each desired result
+ limb, being 2*(prec-1) for a total 2*prec-1 limbs passed to mpn_sqrtrem.
+ The way mpn_sqrtrem calculates floor(sqrt(x)) ensures the root is correct
+ to the intended accuracy, ie. truncated to prec limbs.
+
+ With nails, u might be two limbs, in which case a total 2*prec limbs is
+ passed to mpn_sqrtrem (still giving a prec limb result). If uhigh is
+ zero we adjust back to 2*prec-1, since mpn_sqrtrem requires the high
+ non-zero. 2*prec limbs are always allocated, even when uhigh is zero, so
+ the store of uhigh can be done without a conditional.
+
+ u==0 is a special case so the rest of the code can assume the result is
+ non-zero (ie. will have a non-zero high limb on the result).
+
+ Not done:
+
+ No attempt is made to identify perfect squares. It's considered this can
+ be left to an application if it might occur with any frequency. As it
+ stands, mpn_sqrtrem does its normal amount of work on a perfect square
+ followed by zero limbs, though of course only an mpn_sqrtrem1 would be
+ actually needed. We also end up leaving our mpf result with lots of low
+ trailing zeros, slowing down subsequent operations.
+
+ We're not aware of any optimizations that can be made using the fact the
+ input has lots of trailing zeros (apart from the perfect square
+ case). */
+
+
+/* 1 if we (might) need two limbs for u */
+#define U2 (GMP_NUMB_BITS < BITS_PER_ULONG)
+
+void
+mpf_sqrt_ui (mpf_ptr r, unsigned long int u)
+{
+ mp_size_t rsize, zeros;
+ mp_ptr tp;
+ mp_size_t prec;
+ TMP_DECL;
+
+ if (UNLIKELY (u <= 1))
+ {
+ SIZ (r) = EXP (r) = u;
+ *PTR (r) = u;
+ return;
+ }
+
+ TMP_MARK;
+
+ prec = PREC (r);
+ zeros = 2 * prec - 2;
+ rsize = zeros + 1 + U2;
+
+ tp = TMP_ALLOC_LIMBS (rsize);
+
+ MPN_ZERO (tp, zeros);
+ tp[zeros] = u & GMP_NUMB_MASK;
+
+#if U2
+ {
+ mp_limb_t uhigh = u >> GMP_NUMB_BITS;
+ tp[zeros + 1] = uhigh;
+ rsize -= (uhigh == 0);
+ }
+#endif
+
+ mpn_sqrtrem (PTR (r), NULL, tp, rsize);
+
+ SIZ (r) = prec;
+ EXP (r) = 1;
+ TMP_FREE;
+}