Avoid ldbl-96 stack corruption from range reduction of pseudo-zero (bug 25487).

Bug 25487 reports stack corruption in ldbl-96 sinl on a pseudo-zero
argument (an representation where all the significand bits, including
the explicit high bit, are zero, but the exponent is not zero, which
is not a valid representation for the long double type).

Although this is not a valid long double representation, existing
practice in this area (see bug 4586, originally marked invalid but
subsequently fixed) is that we still seek to avoid invalid memory
accesses as a result, in case of programs that treat arbitrary binary
data as long double representations, although the invalid
representations of the ldbl-96 format do not need to be consistently
handled the same as any particular valid representation.

This patch makes the range reduction detect pseudo-zero and unnormal
representations that would otherwise go to __kernel_rem_pio2, and
returns a NaN for them instead of continuing with the range reduction
process.  (Pseudo-zero and unnormal representations whose unbiased
exponent is less than -1 have already been safely returned from the
function before this point without going through the rest of range
reduction.)  Pseudo-zero representations would previously result in
the value passed to __kernel_rem_pio2 being all-zero, which is
definitely unsafe; unnormal representations would previously result in
a value passed whose high bit is zero, which might well be unsafe
since that is not a form of input expected by __kernel_rem_pio2.

Tested for x86_64.

sysdeps/ieee754/ldbl-96/Makefile

@@ -17,5 +17,6 @@
 # <https://www.gnu.org/licenses/>.
 
 ifeq ($(subdir),math)
-tests += test-canonical-ldbl-96 test-totalorderl-ldbl-96
+tests += test-canonical-ldbl-96 test-totalorderl-ldbl-96 test-sinl-pseudo
+CFLAGS-test-sinl-pseudo.c += -fstack-protector-all
 endif

sysdeps/ieee754/ldbl-96/e_rem_pio2l.c

@@ -210,6 +210,18 @@ __ieee754_rem_pio2l (long double x, long double *y)
       return 0;
     }
 
+  if ((i0 & 0x80000000) == 0)
+    {
+      /* Pseudo-zero and unnormal representations are not valid
+	 representations of long double.  We need to avoid stack
+	 corruption in __kernel_rem_pio2, which expects input in a
+	 particular normal form, but those representations do not need
+	 to be consistently handled like any particular floating-point
+	 value.  */
+      y[1] = y[0] = __builtin_nanl ("");
+      return 0;
+    }
+
   /* Split the 64 bits of the mantissa into three 24-bit integers
      stored in a double array.  */
   exp = j0 - 23;

sysdeps/ieee754/ldbl-96/test-sinl-pseudo.c

@@ -0,0 +1,41 @@
+/* Test sinl for pseudo-zeros and unnormals for ldbl-96 (bug 25487).
+   Copyright (C) 2020 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C 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
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <math.h>
+#include <math_ldbl.h>
+#include <stdint.h>
+
+static int
+do_test (void)
+{
+  for (int i = 0; i < 64; i++)
+    {
+      uint64_t sig = i == 63 ? 0 : 1ULL << i;
+      long double ld;
+      SET_LDOUBLE_WORDS (ld, 0x4141,
+			 sig >> 32, sig & 0xffffffffULL);
+      /* The requirement is that no stack overflow occurs when the
+	 pseudo-zero or unnormal goes through range reduction.  */
+      volatile long double ldr;
+      ldr = sinl (ld);
+      (void) ldr;
+    }
+  return 0;
+}
+
+#include <support/test-driver.c>