diff --git a/configure b/configure
--- a/configure
+++ b/configure
@@ -6940,18 +6940,67 @@ else
   je_cv_function_ffsl=no
 fi
 rm -f core conftest.err conftest.$ac_objext \
     conftest$ac_exeext conftest.$ac_ext
 fi
 { $as_echo "$as_me:${as_lineno-$LINENO}: result: $je_cv_function_ffsl" >&5
 $as_echo "$je_cv_function_ffsl" >&6; }
 
-if test "x${je_cv_function_ffsl}" != "xyes" ; then
-   as_fn_error $? "Cannot build without ffsl(3)" "$LINENO" 5
+if test "x${je_cv_function_ffsl}" == "xyes" ; then
+  $as_echo "#define JEMALLOC_INTERNAL_FFSL ffsl" >>confdefs.h
+
+  $as_echo "#define JEMALLOC_INTERNAL_FFS ffs" >>confdefs.h
+
+else
+
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking whether a program using __builtin_ffsl is compilable" >&5
+$as_echo_n "checking whether a program using __builtin_ffsl is compilable... " >&6; }
+if ${je_cv_gcc_builtin_ffsl+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+
+  #include <stdio.h>
+  #include <strings.h>
+  #include <string.h>
+
+int
+main ()
+{
+
+	{
+		int rv = __builtin_ffsl(0x08);
+		printf("%d\n", rv);
+	}
+
+  ;
+  return 0;
+}
+_ACEOF
+if ac_fn_c_try_link "$LINENO"; then :
+  je_cv_gcc_builtin_ffsl=yes
+else
+  je_cv_gcc_builtin_ffsl=no
+fi
+rm -f core conftest.err conftest.$ac_objext \
+    conftest$ac_exeext conftest.$ac_ext
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $je_cv_gcc_builtin_ffsl" >&5
+$as_echo "$je_cv_gcc_builtin_ffsl" >&6; }
+
+  if test "x${je_cv_gcc_builtin_ffsl}" == "xyes" ; then
+    $as_echo "#define JEMALLOC_INTERNAL_FFSL __builtin_ffsl" >>confdefs.h
+
+    $as_echo "#define JEMALLOC_INTERNAL_FFS __builtin_ffs" >>confdefs.h
+
+  else
+    as_fn_error $? "Cannot build without ffsl(3) or __builtin_ffsl()" "$LINENO" 5
+  fi
 fi
 
 
 
 { $as_echo "$as_me:${as_lineno-$LINENO}: checking whether atomic(9) is compilable" >&5
 $as_echo_n "checking whether atomic(9) is compilable... " >&6; }
 if ${je_cv_atomic9+:} false; then :
   $as_echo_n "(cached) " >&6
diff --git a/configure.ac b/configure.ac
--- a/configure.ac
+++ b/configure.ac
@@ -1160,31 +1160,51 @@ fi
 AC_SUBST([enable_tls])
 if test "x${enable_tls}" = "x1" ; then
   AC_DEFINE_UNQUOTED([JEMALLOC_TLS], [ ])
 elif test "x${force_tls}" = "x1" ; then
   AC_MSG_ERROR([Failed to configure TLS, which is mandatory for correct function])
 fi
 
 dnl ============================================================================
-dnl Check for ffsl(3), and fail if not found.  This function exists on all
-dnl platforms that jemalloc currently has a chance of functioning on without
-dnl modification.
+dnl Check for ffsl(3), then __builtin_ffsl(), and fail if neither are found.
+dnl One of those two functions should (theoretically) exist on all platforms
+dnl that jemalloc currently has a chance of functioning on without modification.
+dnl We additionally assume ffs() or __builtin_ffs() are defined if
+dnl ffsl() or __builtin_ffsl() are defined, respectively.
 JE_COMPILABLE([a program using ffsl], [
 #include <stdio.h>
 #include <strings.h>
 #include <string.h>
 ], [
 	{
 		int rv = ffsl(0x08);
 		printf("%d\n", rv);
 	}
 ], [je_cv_function_ffsl])
-if test "x${je_cv_function_ffsl}" != "xyes" ; then
-   AC_MSG_ERROR([Cannot build without ffsl(3)])
+if test "x${je_cv_function_ffsl}" == "xyes" ; then
+  AC_DEFINE([JEMALLOC_INTERNAL_FFSL], [ffsl])
+  AC_DEFINE([JEMALLOC_INTERNAL_FFS], [ffs])
+else
+  JE_COMPILABLE([a program using __builtin_ffsl], [
+  #include <stdio.h>
+  #include <strings.h>
+  #include <string.h>
+  ], [
+	{
+		int rv = __builtin_ffsl(0x08);
+		printf("%d\n", rv);
+	}
+  ], [je_cv_gcc_builtin_ffsl])
+  if test "x${je_cv_gcc_builtin_ffsl}" == "xyes" ; then
+    AC_DEFINE([JEMALLOC_INTERNAL_FFSL], [__builtin_ffsl])
+    AC_DEFINE([JEMALLOC_INTERNAL_FFS], [__builtin_ffs])
+  else
+    AC_MSG_ERROR([Cannot build without ffsl(3) or __builtin_ffsl()])
+  fi
 fi
 
 dnl ============================================================================
 dnl Check for atomic(9) operations as provided on FreeBSD.
 
 JE_COMPILABLE([atomic(9)], [
 #include <sys/types.h>
 #include <machine/atomic.h>
diff --git a/include/jemalloc/internal/arena.h b/include/jemalloc/internal/arena.h
--- a/include/jemalloc/internal/arena.h
+++ b/include/jemalloc/internal/arena.h
@@ -810,17 +810,17 @@ arena_run_regind(arena_run_t *run, arena
 	 * Avoid doing division with a variable divisor if possible.  Using
 	 * actual division here can reduce allocator throughput by over 20%!
 	 */
 	diff = (unsigned)((uintptr_t)ptr - (uintptr_t)run -
 	    bin_info->reg0_offset);
 
 	/* Rescale (factor powers of 2 out of the numerator and denominator). */
 	interval = bin_info->reg_interval;
-	shift = ffs(interval) - 1;
+	shift = jemalloc_ffs(interval) - 1;
 	diff >>= shift;
 	interval >>= shift;
 
 	if (interval == 1) {
 		/* The divisor was a power of 2. */
 		regind = diff;
 	} else {
 		/*
diff --git a/include/jemalloc/internal/bitmap.h b/include/jemalloc/internal/bitmap.h
--- a/include/jemalloc/internal/bitmap.h
+++ b/include/jemalloc/internal/bitmap.h
@@ -125,21 +125,21 @@ bitmap_sfu(bitmap_t *bitmap, const bitma
 	size_t bit;
 	bitmap_t g;
 	unsigned i;
 
 	assert(bitmap_full(bitmap, binfo) == false);
 
 	i = binfo->nlevels - 1;
 	g = bitmap[binfo->levels[i].group_offset];
-	bit = ffsl(g) - 1;
+	bit = jemalloc_ffsl(g) - 1;
 	while (i > 0) {
 		i--;
 		g = bitmap[binfo->levels[i].group_offset + bit];
-		bit = (bit << LG_BITMAP_GROUP_NBITS) + (ffsl(g) - 1);
+		bit = (bit << LG_BITMAP_GROUP_NBITS) + (jemalloc_ffsl(g) - 1);
 	}
 
 	bitmap_set(bitmap, binfo, bit);
 	return (bit);
 }
 
 JEMALLOC_INLINE void
 bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
diff --git a/include/jemalloc/internal/jemalloc_internal.h.in b/include/jemalloc/internal/jemalloc_internal.h.in
--- a/include/jemalloc/internal/jemalloc_internal.h.in
+++ b/include/jemalloc/internal/jemalloc_internal.h.in
@@ -9,21 +9,23 @@
 #  define EPERM  ERROR_WRITE_FAULT
 #  define EFAULT ERROR_INVALID_ADDRESS
 #  define ENOMEM ERROR_NOT_ENOUGH_MEMORY
 #  undef ERANGE
 #  define ERANGE ERROR_INVALID_DATA
 #else
 #  include <sys/param.h>
 #  include <sys/mman.h>
-#  include <sys/syscall.h>
-#  if !defined(SYS_write) && defined(__NR_write)
-#    define SYS_write __NR_write
+#  if !defined(__pnacl__) && !defined(__native_client__)
+#    include <sys/syscall.h>
+#    if !defined(SYS_write) && defined(__NR_write)
+#      define SYS_write __NR_write
+#    endif
+#    include <sys/uio.h>
 #  endif
-#  include <sys/uio.h>
 #  include <pthread.h>
 #  include <errno.h>
 #endif
 #include <sys/types.h>
 
 #include <limits.h>
 #ifndef SIZE_T_MAX
 #  define SIZE_T_MAX	SIZE_MAX
@@ -275,16 +277,19 @@ static const bool config_ivsalloc =
 #    define LG_QUANTUM		4
 #  endif
 #  ifdef __SH4__
 #    define LG_QUANTUM		4
 #  endif
 #  ifdef __tile__
 #    define LG_QUANTUM		4
 #  endif
+#  ifdef __le32__
+#    define LG_QUANTUM		4
+#  endif
 #  ifndef LG_QUANTUM
 #    error "No LG_QUANTUM definition for architecture; specify via CPPFLAGS"
 #  endif
 #endif
 
 #define	QUANTUM			((size_t)(1U << LG_QUANTUM))
 #define	QUANTUM_MASK		(QUANTUM - 1)
 
diff --git a/include/jemalloc/internal/jemalloc_internal_defs.h.in b/include/jemalloc/internal/jemalloc_internal_defs.h.in
--- a/include/jemalloc/internal/jemalloc_internal_defs.h.in
+++ b/include/jemalloc/internal/jemalloc_internal_defs.h.in
@@ -153,16 +153,23 @@
  * memory map holes, much like munmap(2) does.
  */
 #undef JEMALLOC_MREMAP
 
 /* TLS is used to map arenas and magazine caches to threads. */
 #undef JEMALLOC_TLS
 
 /*
+ * ffs()/ffsl() functions to use for bitmapping.  Don't use these directly;
+ * instead, use jemalloc_ffs() or jemalloc_ffsl() from util.h.
+ */
+#undef JEMALLOC_INTERNAL_FFSL
+#undef JEMALLOC_INTERNAL_FFS
+
+/*
  * JEMALLOC_IVSALLOC enables ivsalloc(), which verifies that pointers reside
  * within jemalloc-owned chunks before dereferencing them.
  */
 #undef JEMALLOC_IVSALLOC
 
 /*
  * Darwin (OS X) uses zones to work around Mach-O symbol override shortcomings.
  */
diff --git a/include/jemalloc/internal/util.h b/include/jemalloc/internal/util.h
--- a/include/jemalloc/internal/util.h
+++ b/include/jemalloc/internal/util.h
@@ -104,22 +104,44 @@ void malloc_cprintf(void (*write)(void *
 void	malloc_printf(const char *format, ...)
     JEMALLOC_ATTR(format(printf, 1, 2));
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
 #ifdef JEMALLOC_H_INLINES
 
 #ifndef JEMALLOC_ENABLE_INLINE
+int	jemalloc_ffsl(long bitmap);
+int	jemalloc_ffs(int bitmap);
 size_t	pow2_ceil(size_t x);
 void	set_errno(int errnum);
 int	get_errno(void);
 #endif
 
 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_UTIL_C_))
+
+/* Sanity check: */
+#if !defined(JEMALLOC_INTERNAL_FFSL) || !defined(JEMALLOC_INTERNAL_FFS)
+#  error Both JEMALLOC_INTERNAL_FFSL && JEMALLOC_INTERNAL_FFS should have been defined by configure
+#endif
+
+JEMALLOC_ALWAYS_INLINE int
+jemalloc_ffsl(long bitmap)
+{
+
+        return (JEMALLOC_INTERNAL_FFSL(bitmap));
+}
+
+JEMALLOC_ALWAYS_INLINE int
+jemalloc_ffs(int bitmap)
+{
+
+        return (JEMALLOC_INTERNAL_FFS(bitmap));
+}
+
 /* Compute the smallest power of 2 that is >= x. */
 JEMALLOC_INLINE size_t
 pow2_ceil(size_t x)
 {
 
 	x--;
 	x |= x >> 1;
 	x |= x >> 2;
diff --git a/src/arena.c b/src/arena.c
--- a/src/arena.c
+++ b/src/arena.c
@@ -2378,17 +2378,17 @@ bin_info_run_size_calc(arena_bin_info_t 
 	/*
 	 * Determine redzone size based on minimum alignment and minimum
 	 * redzone size.  Add padding to the end of the run if it is needed to
 	 * align the regions.  The padding allows each redzone to be half the
 	 * minimum alignment; without the padding, each redzone would have to
 	 * be twice as large in order to maintain alignment.
 	 */
 	if (config_fill && opt_redzone) {
-		size_t align_min = ZU(1) << (ffs(bin_info->reg_size) - 1);
+		size_t align_min = ZU(1) << (jemalloc_ffs(bin_info->reg_size) - 1);
 		if (align_min <= REDZONE_MINSIZE) {
 			bin_info->redzone_size = REDZONE_MINSIZE;
 			pad_size = 0;
 		} else {
 			bin_info->redzone_size = align_min >> 1;
 			pad_size = bin_info->redzone_size;
 		}
 	} else {
diff --git a/src/rtree.c b/src/rtree.c
--- a/src/rtree.c
+++ b/src/rtree.c
@@ -4,18 +4,18 @@
 rtree_t *
 rtree_new(unsigned bits, rtree_alloc_t *alloc, rtree_dalloc_t *dalloc)
 {
 	rtree_t *ret;
 	unsigned bits_per_level, bits_in_leaf, height, i;
 
 	assert(bits > 0 && bits <= (sizeof(uintptr_t) << 3));
 
-	bits_per_level = ffs(pow2_ceil((RTREE_NODESIZE / sizeof(void *)))) - 1;
-	bits_in_leaf = ffs(pow2_ceil((RTREE_NODESIZE / sizeof(uint8_t)))) - 1;
+	bits_per_level = jemalloc_ffs(pow2_ceil((RTREE_NODESIZE / sizeof(void *)))) - 1;
+	bits_in_leaf = jemalloc_ffs(pow2_ceil((RTREE_NODESIZE / sizeof(uint8_t)))) - 1;
 	if (bits > bits_in_leaf) {
 		height = 1 + (bits - bits_in_leaf) / bits_per_level;
 		if ((height-1) * bits_per_level + bits_in_leaf != bits)
 			height++;
 	} else {
 		height = 1;
 	}
 	assert((height-1) * bits_per_level + bits_in_leaf >= bits);