Merge branch 'arm' into unstable

src/Makefile

@@ -14,6 +14,7 @@
 
 release_hdr := $(shell sh -c './mkreleasehdr.sh')
 uname_S := $(shell sh -c 'uname -s 2>/dev/null || echo not')
+uname_M := $(shell sh -c 'uname -m 2>/dev/null || echo not')
 OPTIMIZATION?=-O2
 DEPENDENCY_TARGETS=hiredis linenoise lua
 NODEPS:=clean distclean
@@ -27,11 +28,14 @@ PREFIX?=/usr/local
 INSTALL_BIN=$(PREFIX)/bin
 INSTALL=install
 
-# Default allocator
+# Default allocator defaults to Jemalloc if it's not an ARM
+MALLOC=libc
+ifneq ($(uname_M),armv6l)
+ifneq ($(uname_M),armv7l)
 ifeq ($(uname_S),Linux)
 	MALLOC=jemalloc
-else
-	MALLOC=libc
+endif
+endif
 endif
 
 # Backwards compatibility for selecting an allocator

src/bitops.c

@@ -654,8 +654,11 @@ void bitopCommand(client *c) {
 
         /* Fast path: as far as we have data for all the input bitmaps we
          * can take a fast path that performs much better than the
-         * vanilla algorithm. */
+         * vanilla algorithm. On ARM we skip the fast path since it will
+         * result in GCC compiling the code using multiple-words load/store
+         * operations that are not supported even in ARM >= v6. */
         j = 0;
+        #ifndef __arm__
         if (minlen >= sizeof(unsigned long)*4 && numkeys <= 16) {
             unsigned long *lp[16];
             unsigned long *lres = (unsigned long*) res;
@@ -716,6 +719,7 @@ void bitopCommand(client *c) {
                 }
             }
         }
+        #endif
 
         /* j is set to the next byte to process by the previous loop. */
         for (; j < maxlen; j++) {

src/config.h

@@ -206,4 +206,13 @@ void setproctitle(const char *fmt, ...);
 #endif
 #endif
 
+/* Make sure we can test for ARM just checking for __arm__, since sometimes
+ * __arm is defined but __arm__ is not. */
+#if defined(__arm) && !defined(__arm__)
+#define __arm__
+#endif
+#if defined (__aarch64__) && !defined(__arm64__)
+#define __arm64__
+#endif
+
 #endif

src/hyperloglog.c

@@ -401,7 +401,11 @@ uint64_t MurmurHash64A (const void * key, int len, unsigned int seed) {
         uint64_t k;
 
 #if (BYTE_ORDER == LITTLE_ENDIAN)
+	#if defined(__arm__) && !defined(__arm64__)
+	memcpy(&k,data,sizeof(uint64_t));
+	#else
         k = *((uint64_t*)data);
+	#endif
 #else
         k = (uint64_t) data[0];
         k |= (uint64_t) data[1] << 8;