Merge remote-tracking branch 'origin/released'

8 files changed, 889 insertions, 722 deletions

diff --git a/pixman/configure.ac b/pixman/configure.ac
index dc523df0f..481d0bb00 100644
--- a/pixman/configure.ac
+++ b/pixman/configure.ac
@@ -303,7 +303,7 @@ CFLAGS=$xserver_save_CFLAGS
 
 AC_ARG_ENABLE(mmx,
    [AC_HELP_STRING([--disable-mmx],
-                   [disable MMX fast paths])],
+                   [disable x86 MMX fast paths])],
    [enable_mmx=$enableval], [enable_mmx=auto])
 
 if test $enable_mmx = no ; then
@@ -311,17 +311,17 @@ if test $enable_mmx = no ; then
 fi
 
 if test $have_mmx_intrinsics = yes ; then
-   AC_DEFINE(USE_MMX, 1, [use MMX compiler intrinsics])
+   AC_DEFINE(USE_X86_MMX, 1, [use x86 MMX compiler intrinsics])
 else
    MMX_CFLAGS=
 fi
 
 AC_MSG_RESULT($have_mmx_intrinsics)
 if test $enable_mmx = yes && test $have_mmx_intrinsics = no ; then
-   AC_MSG_ERROR([MMX intrinsics not detected])
+   AC_MSG_ERROR([x86 MMX intrinsics not detected])
 fi
 
-AM_CONDITIONAL(USE_MMX, test $have_mmx_intrinsics = yes)
+AM_CONDITIONAL(USE_X86_MMX, test $have_mmx_intrinsics = yes)
 
 dnl ===========================================================================
 dnl Check for SSE2
@@ -540,6 +540,54 @@ if test $enable_arm_neon = yes && test $have_arm_neon = no ; then
    AC_MSG_ERROR([ARM NEON intrinsics not detected])
 fi
 
+dnl ===========================================================================
+dnl Check for IWMMXT
+
+if test "x$IWMMXT_CFLAGS" = "x" ; then
+   IWMMXT_CFLAGS="-march=iwmmxt -flax-vector-conversions -Winline"
+fi
+
+have_iwmmxt_intrinsics=no
+AC_MSG_CHECKING(whether to use ARM IWMMXT intrinsics)
+xserver_save_CFLAGS=$CFLAGS
+CFLAGS="$IWMMXT_CFLAGS $CFLAGS"
+AC_COMPILE_IFELSE([
+#if defined(__GNUC__) && (__GNUC__ < 4 || (__GNUC__ == 3 && __GNUC_MINOR__ < 6))
+#error "Need GCC >= 4.6 for IWMMXT intrinsics"
+#endif
+#include <mmintrin.h>
+int main () {
+	union {
+		__m64 v;
+		[char c[8];]
+	} a = { .c = {1, 2, 3, 4, 5, 6, 7, 8} };
+	int b = 4;
+	__m64 c = _mm_srli_si64 (a.v, b);
+}], have_iwmmxt_intrinsics=yes)
+CFLAGS=$xserver_save_CFLAGS
+
+AC_ARG_ENABLE(arm-iwmmxt,
+   [AC_HELP_STRING([--disable-arm-iwmmxt],
+                   [disable ARM IWMMXT fast paths])],
+   [enable_iwmmxt=$enableval], [enable_iwmmxt=auto])
+
+if test $enable_iwmmxt = no ; then
+   have_iwmmxt_intrinsics=disabled
+fi
+
+if test $have_iwmmxt_intrinsics = yes ; then
+   AC_DEFINE(USE_ARM_IWMMXT, 1, [use ARM IWMMXT compiler intrinsics])
+else
+   IWMMXT_CFLAGS=
+fi
+
+AC_MSG_RESULT($have_iwmmxt_intrinsics)
+if test $enable_iwmmxt = yes && test $have_iwmmxt_intrinsics = no ; then
+   AC_MSG_ERROR([IWMMXT intrinsics not detected])
+fi
+
+AM_CONDITIONAL(USE_ARM_IWMMXT, test $have_iwmmxt_intrinsics = yes)
+
 dnl =========================================================================================
 dnl Check for GNU-style inline assembly support
 
diff --git a/pixman/demos/Makefile.am b/pixman/demos/Makefile.am
index 171f8f419..070c2d748 100644
--- a/pixman/demos/Makefile.am
+++ b/pixman/demos/Makefile.am
@@ -1,36 +1,36 @@
-if HAVE_GTK

-

-AM_CFLAGS = @OPENMP_CFLAGS@

-AM_LDFLAGS = @OPENMP_CFLAGS@

-

-LDADD = $(top_builddir)/pixman/libpixman-1.la -lm $(GTK_LIBS)

-INCLUDES = -I$(top_srcdir)/pixman -I$(top_builddir)/pixman $(GTK_CFLAGS)

-

-GTK_UTILS = gtk-utils.c gtk-utils.h

-

-DEMOS =				\

-	clip-test		\

-	clip-in			\

-	composite-test		\

-	gradient-test		\

-	radial-test		\

-	alpha-test		\

-	screen-test		\

-	convolution-test	\

-	trap-test		\

-	tri-test

-

-gradient_test_SOURCES = gradient-test.c $(GTK_UTILS)

-alpha_test_SOURCES = alpha-test.c $(GTK_UTILS)

-composite_test_SOURCES = composite-test.c $(GTK_UTILS)

-clip_test_SOURCES = clip-test.c $(GTK_UTILS)

-clip_in_SOURCES = clip-in.c $(GTK_UTILS)

-trap_test_SOURCES = trap-test.c $(GTK_UTILS)

-screen_test_SOURCES = screen-test.c $(GTK_UTILS)

-convolution_test_SOURCES = convolution-test.c $(GTK_UTILS)

-radial_test_SOURCES = radial-test.c ../test/utils.c ../test/utils.h $(GTK_UTILS)

-tri_test_SOURCES = tri-test.c ../test/utils.c ../test/utils.h $(GTK_UTILS)

-

-noinst_PROGRAMS = $(DEMOS)

-

-endif

+if HAVE_GTK
+
+AM_CFLAGS = $(OPENMP_CFLAGS)
+AM_LDFLAGS = $(OPENMP_CFLAGS)
+
+LDADD = $(top_builddir)/pixman/libpixman-1.la -lm $(GTK_LIBS) $(PNG_LIBS)
+INCLUDES = -I$(top_srcdir)/pixman -I$(top_builddir)/pixman $(GTK_CFLAGS) $(PNG_CFLAGS)
+
+GTK_UTILS = gtk-utils.c gtk-utils.h
+
+DEMOS =				\
+	clip-test		\
+	clip-in			\
+	composite-test		\
+	gradient-test		\
+	radial-test		\
+	alpha-test		\
+	screen-test		\
+	convolution-test	\
+	trap-test		\
+	tri-test
+
+gradient_test_SOURCES = gradient-test.c $(GTK_UTILS)
+alpha_test_SOURCES = alpha-test.c $(GTK_UTILS)
+composite_test_SOURCES = composite-test.c $(GTK_UTILS)
+clip_test_SOURCES = clip-test.c $(GTK_UTILS)
+clip_in_SOURCES = clip-in.c $(GTK_UTILS)
+trap_test_SOURCES = trap-test.c $(GTK_UTILS)
+screen_test_SOURCES = screen-test.c $(GTK_UTILS)
+convolution_test_SOURCES = convolution-test.c $(GTK_UTILS)
+radial_test_SOURCES = radial-test.c ../test/utils.c ../test/utils.h $(GTK_UTILS)
+tri_test_SOURCES = tri-test.c ../test/utils.c ../test/utils.h $(GTK_UTILS)
+
+noinst_PROGRAMS = $(DEMOS)
+
+endif
diff --git a/pixman/pixman/Makefile.am b/pixman/pixman/Makefile.am
index 2421a4f9e..286b7cf36 100644
--- a/pixman/pixman/Makefile.am
+++ b/pixman/pixman/Makefile.am
@@ -22,7 +22,7 @@ EXTRA_DIST =				\
 DISTCLEANFILES = $(BUILT_SOURCES)
 
 # mmx code
-if USE_MMX
+if USE_X86_MMX
 noinst_LTLIBRARIES += libpixman-mmx.la
 libpixman_mmx_la_SOURCES = \
 	pixman-mmx.c
@@ -90,5 +90,17 @@ libpixman_1_la_LIBADD += libpixman-arm-neon.la
 ASM_CFLAGS_arm_neon=
 endif
 
+# iwmmxt code
+if USE_ARM_IWMMXT
+noinst_LTLIBRARIES += libpixman-iwmmxt.la
+libpixman_iwmmxt_la_SOURCES = pixman-mmx.c
+libpixman_iwmmxt_la_CFLAGS = $(DEP_CFLAGS) $(IWMMXT_CFLAGS)
+libpixman_iwmmxt_la_LIBADD = $(DEP_LIBS)
+libpixman_1_la_LDFLAGS += $(IWMMXT_LDFLAGS)
+libpixman_1_la_LIBADD += libpixman-iwmmxt.la
+
+ASM_CFLAGS_IWMMXT=$(IWMMXT_CFLAGS)
+endif
+
 .c.s : $(libpixmaninclude_HEADERS) $(BUILT_SOURCES)
 	$(CC) $(CFLAGS) $(ASM_CFLAGS_$(@:pixman-%.s=%)) $(ASM_CFLAGS_$(@:pixman-arm-%.s=arm_%)) -DHAVE_CONFIG_H -I$(srcdir) -I$(builddir) -I$(top_builddir) -S -o $@ $<
diff --git a/pixman/pixman/Makefile.win32 b/pixman/pixman/Makefile.win32
index beff4a068..381f2cd9d 100644
--- a/pixman/pixman/Makefile.win32
+++ b/pixman/pixman/Makefile.win32
@@ -14,7 +14,7 @@ ifeq ($(SSE2_VAR),)
 SSE2_VAR=on
 endif
 
-MMX_CFLAGS = -DUSE_MMX -w14710 -w14714
+MMX_CFLAGS = -DUSE_X86_MMX -w14710 -w14714
 SSE2_CFLAGS = -DUSE_SSE2
 
 # MMX compilation flags
diff --git a/pixman/pixman/pixman-cpu.c b/pixman/pixman/pixman-cpu.c
index 77e5b7f7a..dff27d1ea 100644
--- a/pixman/pixman/pixman-cpu.c
+++ b/pixman/pixman/pixman-cpu.c
@@ -1,613 +1,631 @@
-/*

- * Copyright © 2000 SuSE, Inc.

- * Copyright © 2007 Red Hat, Inc.

- *

- * Permission to use, copy, modify, distribute, and sell this software and its

- * documentation for any purpose is hereby granted without fee, provided that

- * the above copyright notice appear in all copies and that both that

- * copyright notice and this permission notice appear in supporting

- * documentation, and that the name of SuSE not be used in advertising or

- * publicity pertaining to distribution of the software without specific,

- * written prior permission.  SuSE makes no representations about the

- * suitability of this software for any purpose.  It is provided "as is"

- * without express or implied warranty.

- *

- * SuSE DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL

- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL SuSE

- * BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES

- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION

- * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN

- * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.

- */

-#ifdef HAVE_CONFIG_H

-#include <config.h>

-#endif

-

-#include <string.h>

-

-#if defined(USE_ARM_SIMD) && defined(_MSC_VER)

-/* Needed for EXCEPTION_ILLEGAL_INSTRUCTION */

-#include <windows.h>

-#endif

-

-#include "pixman-private.h"

-

-#ifdef USE_VMX

-

-/* The CPU detection code needs to be in a file not compiled with

- * "-maltivec -mabi=altivec", as gcc would try to save vector register

- * across function calls causing SIGILL on cpus without Altivec/vmx.

- */

-static pixman_bool_t initialized = FALSE;

-static volatile pixman_bool_t have_vmx = TRUE;

-

-#ifdef __APPLE__

-#include <sys/sysctl.h>

-

-static pixman_bool_t

-pixman_have_vmx (void)

-{

-    if (!initialized)

-    {

-	size_t length = sizeof(have_vmx);

-	int error =

-	    sysctlbyname ("hw.optional.altivec", &have_vmx, &length, NULL, 0);

-

-	if (error)

-	    have_vmx = FALSE;

-

-	initialized = TRUE;

-    }

-    return have_vmx;

-}

-

-#elif defined (__OpenBSD__)

-#include <sys/param.h>

-#include <sys/sysctl.h>

-#include <machine/cpu.h>

-

-static pixman_bool_t

-pixman_have_vmx (void)

-{

-    if (!initialized)

-    {

-	int mib[2] = { CTL_MACHDEP, CPU_ALTIVEC };

-	size_t length = sizeof(have_vmx);

-	int error =

-	    sysctl (mib, 2, &have_vmx, &length, NULL, 0);

-

-	if (error != 0)

-	    have_vmx = FALSE;

-

-	initialized = TRUE;

-    }

-    return have_vmx;

-}

-

-#elif defined (__linux__)

-#include <sys/types.h>

-#include <sys/stat.h>

-#include <fcntl.h>

-#include <unistd.h>

-#include <stdio.h>

-#include <linux/auxvec.h>

-#include <asm/cputable.h>

-

-static pixman_bool_t

-pixman_have_vmx (void)

-{

-    if (!initialized)

-    {

-	char fname[64];

-	unsigned long buf[64];

-	ssize_t count = 0;

-	pid_t pid;

-	int fd, i;

-

-	pid = getpid ();

-	snprintf (fname, sizeof(fname) - 1, "/proc/%d/auxv", pid);

-

-	fd = open (fname, O_RDONLY);

-	if (fd >= 0)

-	{

-	    for (i = 0; i <= (count / sizeof(unsigned long)); i += 2)

-	    {

-		/* Read more if buf is empty... */

-		if (i == (count / sizeof(unsigned long)))

-		{

-		    count = read (fd, buf, sizeof(buf));

-		    if (count <= 0)

-			break;

-		    i = 0;

-		}

-

-		if (buf[i] == AT_HWCAP)

-		{

-		    have_vmx = !!(buf[i + 1] & PPC_FEATURE_HAS_ALTIVEC);

-		    initialized = TRUE;

-		    break;

-		}

-		else if (buf[i] == AT_NULL)

-		{

-		    break;

-		}

-	    }

-	    close (fd);

-	}

-    }

-    if (!initialized)

-    {

-	/* Something went wrong. Assume 'no' rather than playing

-	   fragile tricks with catching SIGILL. */

-	have_vmx = FALSE;

-	initialized = TRUE;

-    }

-

-    return have_vmx;

-}

-

-#else /* !__APPLE__ && !__OpenBSD__ && !__linux__ */

-#include <signal.h>

-#include <setjmp.h>

-

-static jmp_buf jump_env;

-

-static void

-vmx_test (int        sig,

-	  siginfo_t *si,

-	  void *     unused)

-{

-    longjmp (jump_env, 1);

-}

-

-static pixman_bool_t

-pixman_have_vmx (void)

-{

-    struct sigaction sa, osa;

-    int jmp_result;

-

-    if (!initialized)

-    {

-	sa.sa_flags = SA_SIGINFO;

-	sigemptyset (&sa.sa_mask);

-	sa.sa_sigaction = vmx_test;

-	sigaction (SIGILL, &sa, &osa);

-	jmp_result = setjmp (jump_env);

-	if (jmp_result == 0)

-	{

-	    asm volatile ( "vor 0, 0, 0" );

-	}

-	sigaction (SIGILL, &osa, NULL);

-	have_vmx = (jmp_result == 0);

-	initialized = TRUE;

-    }

-    return have_vmx;

-}

-

-#endif /* __APPLE__ */

-#endif /* USE_VMX */

-

-#if defined(USE_ARM_SIMD) || defined(USE_ARM_NEON)

-

-#if defined(_MSC_VER)

-

-#if defined(USE_ARM_SIMD)

-extern int pixman_msvc_try_arm_simd_op ();

-

-pixman_bool_t

-pixman_have_arm_simd (void)

-{

-    static pixman_bool_t initialized = FALSE;

-    static pixman_bool_t have_arm_simd = FALSE;

-

-    if (!initialized)

-    {

-	__try {

-	    pixman_msvc_try_arm_simd_op ();

-	    have_arm_simd = TRUE;

-	} __except (GetExceptionCode () == EXCEPTION_ILLEGAL_INSTRUCTION) {

-	    have_arm_simd = FALSE;

-	}

-	initialized = TRUE;

-    }

-

-    return have_arm_simd;

-}

-

-#endif /* USE_ARM_SIMD */

-

-#if defined(USE_ARM_NEON)

-extern int pixman_msvc_try_arm_neon_op ();

-

-pixman_bool_t

-pixman_have_arm_neon (void)

-{

-    static pixman_bool_t initialized = FALSE;

-    static pixman_bool_t have_arm_neon = FALSE;

-

-    if (!initialized)

-    {

-	__try

-	{

-	    pixman_msvc_try_arm_neon_op ();

-	    have_arm_neon = TRUE;

-	}

-	__except (GetExceptionCode () == EXCEPTION_ILLEGAL_INSTRUCTION)

-	{

-	    have_arm_neon = FALSE;

-	}

-	initialized = TRUE;

-    }

-

-    return have_arm_neon;

-}

-

-#endif /* USE_ARM_NEON */

-

-#elif defined (__linux__) /* linux ELF */

-

-#include <stdlib.h>

-#include <unistd.h>

-#include <sys/types.h>

-#include <sys/stat.h>

-#include <sys/mman.h>

-#include <fcntl.h>

-#include <string.h>

-#include <elf.h>

-

-static pixman_bool_t arm_has_v7 = FALSE;

-static pixman_bool_t arm_has_v6 = FALSE;

-static pixman_bool_t arm_has_vfp = FALSE;

-static pixman_bool_t arm_has_neon = FALSE;

-static pixman_bool_t arm_has_iwmmxt = FALSE;

-static pixman_bool_t arm_tests_initialized = FALSE;

-

-static void

-pixman_arm_read_auxv ()

-{

-    int fd;

-    Elf32_auxv_t aux;

-

-    fd = open ("/proc/self/auxv", O_RDONLY);

-    if (fd >= 0)

-    {

-	while (read (fd, &aux, sizeof(Elf32_auxv_t)) == sizeof(Elf32_auxv_t))

-	{

-	    if (aux.a_type == AT_HWCAP)

-	    {

-		uint32_t hwcap = aux.a_un.a_val;

-		/* hardcode these values to avoid depending on specific

-		 * versions of the hwcap header, e.g. HWCAP_NEON

-		 */

-		arm_has_vfp = (hwcap & 64) != 0;

-		arm_has_iwmmxt = (hwcap & 512) != 0;

-		/* this flag is only present on kernel 2.6.29 */

-		arm_has_neon = (hwcap & 4096) != 0;

-	    }

-	    else if (aux.a_type == AT_PLATFORM)

-	    {

-		const char *plat = (const char*) aux.a_un.a_val;

-		if (strncmp (plat, "v7l", 3) == 0)

-		{

-		    arm_has_v7 = TRUE;

-		    arm_has_v6 = TRUE;

-		}

-		else if (strncmp (plat, "v6l", 3) == 0)

-		{

-		    arm_has_v6 = TRUE;

-		}

-	    }

-	}

-	close (fd);

-    }

-

-    arm_tests_initialized = TRUE;

-}

-

-#if defined(USE_ARM_SIMD)

-pixman_bool_t

-pixman_have_arm_simd (void)

-{

-    if (!arm_tests_initialized)

-	pixman_arm_read_auxv ();

-

-    return arm_has_v6;

-}

-

-#endif /* USE_ARM_SIMD */

-

-#if defined(USE_ARM_NEON)

-pixman_bool_t

-pixman_have_arm_neon (void)

-{

-    if (!arm_tests_initialized)

-	pixman_arm_read_auxv ();

-

-    return arm_has_neon;

-}

-

-#endif /* USE_ARM_NEON */

-

-#else /* linux ELF */

-

-#define pixman_have_arm_simd() FALSE

-#define pixman_have_arm_neon() FALSE

-

-#endif

-

-#endif /* USE_ARM_SIMD || USE_ARM_NEON */

-

-#if defined(USE_MMX) || defined(USE_SSE2)

-/* The CPU detection code needs to be in a file not compiled with

- * "-mmmx -msse", as gcc would generate CMOV instructions otherwise

- * that would lead to SIGILL instructions on old CPUs that don't have

- * it.

- */

-#if !defined(__amd64__) && !defined(__x86_64__) && !defined(_M_AMD64)

-

-#ifdef HAVE_GETISAX

-#include <sys/auxv.h>

-#endif

-

-typedef enum

-{

-    NO_FEATURES = 0,

-    MMX = 0x1,

-    MMX_EXTENSIONS = 0x2,

-    SSE = 0x6,

-    SSE2 = 0x8,

-    CMOV = 0x10

-} cpu_features_t;

-

-

-static unsigned int

-detect_cpu_features (void)

-{

-    unsigned int features = 0;

-    unsigned int result = 0;

-

-#ifdef HAVE_GETISAX

-    if (getisax (&result, 1))

-    {

-	if (result & AV_386_CMOV)

-	    features |= CMOV;

-	if (result & AV_386_MMX)

-	    features |= MMX;

-	if (result & AV_386_AMD_MMX)

-	    features |= MMX_EXTENSIONS;

-	if (result & AV_386_SSE)

-	    features |= SSE;

-	if (result & AV_386_SSE2)

-	    features |= SSE2;

-    }

-#else

-    char vendor[13];

-#ifdef _MSC_VER

-    int vendor0 = 0, vendor1, vendor2;

-#endif

-    vendor[0] = 0;

-    vendor[12] = 0;

-

-#ifdef __GNUC__

-    /* see p. 118 of amd64 instruction set manual Vol3 */

-    /* We need to be careful about the handling of %ebx and

-     * %esp here. We can't declare either one as clobbered

-     * since they are special registers (%ebx is the "PIC

-     * register" holding an offset to global data, %esp the

-     * stack pointer), so we need to make sure they have their

-     * original values when we access the output operands.

-     */

-    __asm__ (

-        "pushf\n"

-        "pop %%eax\n"

-        "mov %%eax, %%ecx\n"

-        "xor $0x00200000, %%eax\n"

-        "push %%eax\n"

-        "popf\n"

-        "pushf\n"

-        "pop %%eax\n"

-        "mov $0x0, %%edx\n"

-        "xor %%ecx, %%eax\n"

-        "jz 1f\n"

-

-        "mov $0x00000000, %%eax\n"

-        "push %%ebx\n"

-        "cpuid\n"

-        "mov %%ebx, %%eax\n"

-        "pop %%ebx\n"

-        "mov %%eax, %1\n"

-        "mov %%edx, %2\n"

-        "mov %%ecx, %3\n"

-        "mov $0x00000001, %%eax\n"

-        "push %%ebx\n"

-        "cpuid\n"

-        "pop %%ebx\n"

-        "1:\n"

-        "mov %%edx, %0\n"

-	: "=r" (result),

-        "=m" (vendor[0]),

-        "=m" (vendor[4]),

-        "=m" (vendor[8])

-	:

-	: "%eax", "%ecx", "%edx"

-        );

-

-#elif defined (_MSC_VER)

-

-    _asm {

-	pushfd

-	pop eax

-	mov ecx, eax

-	xor eax, 00200000h

-	push eax

-	popfd

-	pushfd

-	pop eax

-	mov edx, 0

-	xor eax, ecx

-	jz nocpuid

-

-	mov eax, 0

-	push ebx

-	cpuid

-	mov eax, ebx

-	pop ebx

-	mov vendor0, eax

-	mov vendor1, edx

-	mov vendor2, ecx

-	mov eax, 1

-	push ebx

-	cpuid

-	pop ebx

-    nocpuid:

-	mov result, edx

-    }

-    memmove (vendor + 0, &vendor0, 4);

-    memmove (vendor + 4, &vendor1, 4);

-    memmove (vendor + 8, &vendor2, 4);

-

-#else

-#   error unsupported compiler

-#endif

-

-    features = 0;

-    if (result)

-    {

-	/* result now contains the standard feature bits */

-	if (result & (1 << 15))

-	    features |= CMOV;

-	if (result & (1 << 23))

-	    features |= MMX;

-	if (result & (1 << 25))

-	    features |= SSE;

-	if (result & (1 << 26))

-	    features |= SSE2;

-	if ((features & MMX) && !(features & SSE) &&

-	    (strcmp (vendor, "AuthenticAMD") == 0 ||

-	     strcmp (vendor, "Geode by NSC") == 0))

-	{

-	    /* check for AMD MMX extensions */

-#ifdef __GNUC__

-	    __asm__ (

-	        "	push %%ebx\n"

-	        "	mov $0x80000000, %%eax\n"

-	        "	cpuid\n"

-	        "	xor %%edx, %%edx\n"

-	        "	cmp $0x1, %%eax\n"

-	        "	jge 2f\n"

-	        "	mov $0x80000001, %%eax\n"

-	        "	cpuid\n"

-	        "2:\n"

-	        "	pop %%ebx\n"

-	        "	mov %%edx, %0\n"

-		: "=r" (result)

-		:

-		: "%eax", "%ecx", "%edx"

-	        );

-#elif defined _MSC_VER

-	    _asm {

-		push ebx

-		mov eax, 80000000h

-		cpuid

-		xor edx, edx

-		cmp eax, 1

-		jge notamd

-		mov eax, 80000001h

-		cpuid

-	    notamd:

-		pop ebx

-		mov result, edx

-	    }

-#endif

-	    if (result & (1 << 22))

-		features |= MMX_EXTENSIONS;

-	}

-    }

-#endif /* HAVE_GETISAX */

-

-    return features;

-}

-

-static pixman_bool_t

-pixman_have_mmx (void)

-{

-    static pixman_bool_t initialized = FALSE;

-    static pixman_bool_t mmx_present;

-

-    if (!initialized)

-    {

-	unsigned int features = detect_cpu_features ();

-	mmx_present = (features & (MMX | MMX_EXTENSIONS)) == (MMX | MMX_EXTENSIONS);

-	initialized = TRUE;

-    }

-

-    return mmx_present;

-}

-

-#ifdef USE_SSE2

-static pixman_bool_t

-pixman_have_sse2 (void)

-{

-    static pixman_bool_t initialized = FALSE;

-    static pixman_bool_t sse2_present;

-

-    if (!initialized)

-    {

-	unsigned int features = detect_cpu_features ();

-	sse2_present = (features & (MMX | MMX_EXTENSIONS | SSE | SSE2)) == (MMX | MMX_EXTENSIONS | SSE | SSE2);

-	initialized = TRUE;

-    }

-

-    return sse2_present;

-}

-

-#endif

-

-#else /* __amd64__ */

-#ifdef USE_MMX

-#define pixman_have_mmx() TRUE

-#endif

-#ifdef USE_SSE2

-#define pixman_have_sse2() TRUE

-#endif

-#endif /* __amd64__ */

-#endif

-

-pixman_implementation_t *

-_pixman_choose_implementation (void)

-{

-    pixman_implementation_t *imp;

-

-    imp = _pixman_implementation_create_general();

-    imp = _pixman_implementation_create_fast_path (imp);

-    

-#ifdef USE_MMX

-    if (pixman_have_mmx ())

-	imp = _pixman_implementation_create_mmx (imp);

-#endif

-

-#ifdef USE_SSE2

-    if (pixman_have_sse2 ())

-	imp = _pixman_implementation_create_sse2 (imp);

-#endif

-

-#ifdef USE_ARM_SIMD

-    if (pixman_have_arm_simd ())

-	imp = _pixman_implementation_create_arm_simd (imp);

-#endif

-

-#ifdef USE_ARM_NEON

-    if (pixman_have_arm_neon ())

-	imp = _pixman_implementation_create_arm_neon (imp);

-#endif

-    

-#ifdef USE_VMX

-    if (pixman_have_vmx ())

-	imp = _pixman_implementation_create_vmx (imp);

-#endif

-

-    imp = _pixman_implementation_create_noop (imp);

-    

-    return imp;

-}

-

+/*
+ * Copyright © 2000 SuSE, Inc.
+ * Copyright © 2007 Red Hat, Inc.
+ *
+ * Permission to use, copy, modify, distribute, and sell this software and its
+ * documentation for any purpose is hereby granted without fee, provided that
+ * the above copyright notice appear in all copies and that both that
+ * copyright notice and this permission notice appear in supporting
+ * documentation, and that the name of SuSE not be used in advertising or
+ * publicity pertaining to distribution of the software without specific,
+ * written prior permission.  SuSE makes no representations about the
+ * suitability of this software for any purpose.  It is provided "as is"
+ * without express or implied warranty.
+ *
+ * SuSE DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL SuSE
+ * BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
+ * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
+ * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+#ifdef HAVE_CONFIG_H
+#include <config.h>
+#endif
+
+#include <string.h>
+
+#if defined(USE_ARM_SIMD) && defined(_MSC_VER)
+/* Needed for EXCEPTION_ILLEGAL_INSTRUCTION */
+#include <windows.h>
+#endif
+
+#include "pixman-private.h"
+
+#ifdef USE_VMX
+
+/* The CPU detection code needs to be in a file not compiled with
+ * "-maltivec -mabi=altivec", as gcc would try to save vector register
+ * across function calls causing SIGILL on cpus without Altivec/vmx.
+ */
+static pixman_bool_t initialized = FALSE;
+static volatile pixman_bool_t have_vmx = TRUE;
+
+#ifdef __APPLE__
+#include <sys/sysctl.h>
+
+static pixman_bool_t
+pixman_have_vmx (void)
+{
+    if (!initialized)
+    {
+	size_t length = sizeof(have_vmx);
+	int error =
+	    sysctlbyname ("hw.optional.altivec", &have_vmx, &length, NULL, 0);
+
+	if (error)
+	    have_vmx = FALSE;
+
+	initialized = TRUE;
+    }
+    return have_vmx;
+}
+
+#elif defined (__OpenBSD__)
+#include <sys/param.h>
+#include <sys/sysctl.h>
+#include <machine/cpu.h>
+
+static pixman_bool_t
+pixman_have_vmx (void)
+{
+    if (!initialized)
+    {
+	int mib[2] = { CTL_MACHDEP, CPU_ALTIVEC };
+	size_t length = sizeof(have_vmx);
+	int error =
+	    sysctl (mib, 2, &have_vmx, &length, NULL, 0);
+
+	if (error != 0)
+	    have_vmx = FALSE;
+
+	initialized = TRUE;
+    }
+    return have_vmx;
+}
+
+#elif defined (__linux__)
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <linux/auxvec.h>
+#include <asm/cputable.h>
+
+static pixman_bool_t
+pixman_have_vmx (void)
+{
+    if (!initialized)
+    {
+	char fname[64];
+	unsigned long buf[64];
+	ssize_t count = 0;
+	pid_t pid;
+	int fd, i;
+
+	pid = getpid ();
+	snprintf (fname, sizeof(fname) - 1, "/proc/%d/auxv", pid);
+
+	fd = open (fname, O_RDONLY);
+	if (fd >= 0)
+	{
+	    for (i = 0; i <= (count / sizeof(unsigned long)); i += 2)
+	    {
+		/* Read more if buf is empty... */
+		if (i == (count / sizeof(unsigned long)))
+		{
+		    count = read (fd, buf, sizeof(buf));
+		    if (count <= 0)
+			break;
+		    i = 0;
+		}
+
+		if (buf[i] == AT_HWCAP)
+		{
+		    have_vmx = !!(buf[i + 1] & PPC_FEATURE_HAS_ALTIVEC);
+		    initialized = TRUE;
+		    break;
+		}
+		else if (buf[i] == AT_NULL)
+		{
+		    break;
+		}
+	    }
+	    close (fd);
+	}
+    }
+    if (!initialized)
+    {
+	/* Something went wrong. Assume 'no' rather than playing
+	   fragile tricks with catching SIGILL. */
+	have_vmx = FALSE;
+	initialized = TRUE;
+    }
+
+    return have_vmx;
+}
+
+#else /* !__APPLE__ && !__OpenBSD__ && !__linux__ */
+#include <signal.h>
+#include <setjmp.h>
+
+static jmp_buf jump_env;
+
+static void
+vmx_test (int        sig,
+	  siginfo_t *si,
+	  void *     unused)
+{
+    longjmp (jump_env, 1);
+}
+
+static pixman_bool_t
+pixman_have_vmx (void)
+{
+    struct sigaction sa, osa;
+    int jmp_result;
+
+    if (!initialized)
+    {
+	sa.sa_flags = SA_SIGINFO;
+	sigemptyset (&sa.sa_mask);
+	sa.sa_sigaction = vmx_test;
+	sigaction (SIGILL, &sa, &osa);
+	jmp_result = setjmp (jump_env);
+	if (jmp_result == 0)
+	{
+	    asm volatile ( "vor 0, 0, 0" );
+	}
+	sigaction (SIGILL, &osa, NULL);
+	have_vmx = (jmp_result == 0);
+	initialized = TRUE;
+    }
+    return have_vmx;
+}
+
+#endif /* __APPLE__ */
+#endif /* USE_VMX */
+
+#if defined(USE_ARM_SIMD) || defined(USE_ARM_NEON) || defined(USE_ARM_IWMMXT)
+
+#if defined(_MSC_VER)
+
+#if defined(USE_ARM_SIMD)
+extern int pixman_msvc_try_arm_simd_op ();
+
+pixman_bool_t
+pixman_have_arm_simd (void)
+{
+    static pixman_bool_t initialized = FALSE;
+    static pixman_bool_t have_arm_simd = FALSE;
+
+    if (!initialized)
+    {
+	__try {
+	    pixman_msvc_try_arm_simd_op ();
+	    have_arm_simd = TRUE;
+	} __except (GetExceptionCode () == EXCEPTION_ILLEGAL_INSTRUCTION) {
+	    have_arm_simd = FALSE;
+	}
+	initialized = TRUE;
+    }
+
+    return have_arm_simd;
+}
+
+#endif /* USE_ARM_SIMD */
+
+#if defined(USE_ARM_NEON)
+extern int pixman_msvc_try_arm_neon_op ();
+
+pixman_bool_t
+pixman_have_arm_neon (void)
+{
+    static pixman_bool_t initialized = FALSE;
+    static pixman_bool_t have_arm_neon = FALSE;
+
+    if (!initialized)
+    {
+	__try
+	{
+	    pixman_msvc_try_arm_neon_op ();
+	    have_arm_neon = TRUE;
+	}
+	__except (GetExceptionCode () == EXCEPTION_ILLEGAL_INSTRUCTION)
+	{
+	    have_arm_neon = FALSE;
+	}
+	initialized = TRUE;
+    }
+
+    return have_arm_neon;
+}
+
+#endif /* USE_ARM_NEON */
+
+#elif defined (__linux__) /* linux ELF */
+
+#include <stdlib.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/mman.h>
+#include <fcntl.h>
+#include <string.h>
+#include <elf.h>
+
+static pixman_bool_t arm_has_v7 = FALSE;
+static pixman_bool_t arm_has_v6 = FALSE;
+static pixman_bool_t arm_has_vfp = FALSE;
+static pixman_bool_t arm_has_neon = FALSE;
+static pixman_bool_t arm_has_iwmmxt = FALSE;
+static pixman_bool_t arm_tests_initialized = FALSE;
+
+static void
+pixman_arm_read_auxv ()
+{
+    int fd;
+    Elf32_auxv_t aux;
+
+    fd = open ("/proc/self/auxv", O_RDONLY);
+    if (fd >= 0)
+    {
+	while (read (fd, &aux, sizeof(Elf32_auxv_t)) == sizeof(Elf32_auxv_t))
+	{
+	    if (aux.a_type == AT_HWCAP)
+	    {
+		uint32_t hwcap = aux.a_un.a_val;
+		/* hardcode these values to avoid depending on specific
+		 * versions of the hwcap header, e.g. HWCAP_NEON
+		 */
+		arm_has_vfp = (hwcap & 64) != 0;
+		arm_has_iwmmxt = (hwcap & 512) != 0;
+		/* this flag is only present on kernel 2.6.29 */
+		arm_has_neon = (hwcap & 4096) != 0;
+	    }
+	    else if (aux.a_type == AT_PLATFORM)
+	    {
+		const char *plat = (const char*) aux.a_un.a_val;
+		if (strncmp (plat, "v7l", 3) == 0)
+		{
+		    arm_has_v7 = TRUE;
+		    arm_has_v6 = TRUE;
+		}
+		else if (strncmp (plat, "v6l", 3) == 0)
+		{
+		    arm_has_v6 = TRUE;
+		}
+	    }
+	}
+	close (fd);
+    }
+
+    arm_tests_initialized = TRUE;
+}
+
+#if defined(USE_ARM_SIMD)
+pixman_bool_t
+pixman_have_arm_simd (void)
+{
+    if (!arm_tests_initialized)
+	pixman_arm_read_auxv ();
+
+    return arm_has_v6;
+}
+
+#endif /* USE_ARM_SIMD */
+
+#if defined(USE_ARM_NEON)
+pixman_bool_t
+pixman_have_arm_neon (void)
+{
+    if (!arm_tests_initialized)
+	pixman_arm_read_auxv ();
+
+    return arm_has_neon;
+}
+
+#endif /* USE_ARM_NEON */
+
+#if defined(USE_ARM_IWMMXT)
+pixman_bool_t
+pixman_have_arm_iwmmxt (void)
+{
+    if (!arm_tests_initialized)
+	pixman_arm_read_auxv ();
+
+    return arm_has_iwmmxt;
+}
+
+#endif /* USE_ARM_IWMMXT */
+
+#else /* linux ELF */
+
+#define pixman_have_arm_simd() FALSE
+#define pixman_have_arm_neon() FALSE
+#define pixman_have_arm_iwmmxt() FALSE
+
+#endif
+
+#endif /* USE_ARM_SIMD || USE_ARM_NEON || USE_ARM_IWMMXT */
+
+#if defined(USE_X86_MMX) || defined(USE_SSE2)
+/* The CPU detection code needs to be in a file not compiled with
+ * "-mmmx -msse", as gcc would generate CMOV instructions otherwise
+ * that would lead to SIGILL instructions on old CPUs that don't have
+ * it.
+ */
+#if !defined(__amd64__) && !defined(__x86_64__) && !defined(_M_AMD64)
+
+#ifdef HAVE_GETISAX
+#include <sys/auxv.h>
+#endif
+
+typedef enum
+{
+    NO_FEATURES = 0,
+    MMX = 0x1,
+    MMX_EXTENSIONS = 0x2,
+    SSE = 0x6,
+    SSE2 = 0x8,
+    CMOV = 0x10
+} cpu_features_t;
+
+
+static unsigned int
+detect_cpu_features (void)
+{
+    unsigned int features = 0;
+    unsigned int result = 0;
+
+#ifdef HAVE_GETISAX
+    if (getisax (&result, 1))
+    {
+	if (result & AV_386_CMOV)
+	    features |= CMOV;
+	if (result & AV_386_MMX)
+	    features |= MMX;
+	if (result & AV_386_AMD_MMX)
+	    features |= MMX_EXTENSIONS;
+	if (result & AV_386_SSE)
+	    features |= SSE;
+	if (result & AV_386_SSE2)
+	    features |= SSE2;
+    }
+#else
+    char vendor[13];
+#ifdef _MSC_VER
+    int vendor0 = 0, vendor1, vendor2;
+#endif
+    vendor[0] = 0;
+    vendor[12] = 0;
+
+#ifdef __GNUC__
+    /* see p. 118 of amd64 instruction set manual Vol3 */
+    /* We need to be careful about the handling of %ebx and
+     * %esp here. We can't declare either one as clobbered
+     * since they are special registers (%ebx is the "PIC
+     * register" holding an offset to global data, %esp the
+     * stack pointer), so we need to make sure they have their
+     * original values when we access the output operands.
+     */
+    __asm__ (
+        "pushf\n"
+        "pop %%eax\n"
+        "mov %%eax, %%ecx\n"
+        "xor $0x00200000, %%eax\n"
+        "push %%eax\n"
+        "popf\n"
+        "pushf\n"
+        "pop %%eax\n"
+        "mov $0x0, %%edx\n"
+        "xor %%ecx, %%eax\n"
+        "jz 1f\n"
+
+        "mov $0x00000000, %%eax\n"
+        "push %%ebx\n"
+        "cpuid\n"
+        "mov %%ebx, %%eax\n"
+        "pop %%ebx\n"
+        "mov %%eax, %1\n"
+        "mov %%edx, %2\n"
+        "mov %%ecx, %3\n"
+        "mov $0x00000001, %%eax\n"
+        "push %%ebx\n"
+        "cpuid\n"
+        "pop %%ebx\n"
+        "1:\n"
+        "mov %%edx, %0\n"
+	: "=r" (result),
+        "=m" (vendor[0]),
+        "=m" (vendor[4]),
+        "=m" (vendor[8])
+	:
+	: "%eax", "%ecx", "%edx"
+        );
+
+#elif defined (_MSC_VER)
+
+    _asm {
+	pushfd
+	pop eax
+	mov ecx, eax
+	xor eax, 00200000h
+	push eax
+	popfd
+	pushfd
+	pop eax
+	mov edx, 0
+	xor eax, ecx
+	jz nocpuid
+
+	mov eax, 0
+	push ebx
+	cpuid
+	mov eax, ebx
+	pop ebx
+	mov vendor0, eax
+	mov vendor1, edx
+	mov vendor2, ecx
+	mov eax, 1
+	push ebx
+	cpuid
+	pop ebx
+    nocpuid:
+	mov result, edx
+    }
+    memmove (vendor + 0, &vendor0, 4);
+    memmove (vendor + 4, &vendor1, 4);
+    memmove (vendor + 8, &vendor2, 4);
+
+#else
+#   error unsupported compiler
+#endif
+
+    features = 0;
+    if (result)
+    {
+	/* result now contains the standard feature bits */
+	if (result & (1 << 15))
+	    features |= CMOV;
+	if (result & (1 << 23))
+	    features |= MMX;
+	if (result & (1 << 25))
+	    features |= SSE;
+	if (result & (1 << 26))
+	    features |= SSE2;
+	if ((features & MMX) && !(features & SSE) &&
+	    (strcmp (vendor, "AuthenticAMD") == 0 ||
+	     strcmp (vendor, "Geode by NSC") == 0))
+	{
+	    /* check for AMD MMX extensions */
+#ifdef __GNUC__
+	    __asm__ (
+	        "	push %%ebx\n"
+	        "	mov $0x80000000, %%eax\n"
+	        "	cpuid\n"
+	        "	xor %%edx, %%edx\n"
+	        "	cmp $0x1, %%eax\n"
+	        "	jge 2f\n"
+	        "	mov $0x80000001, %%eax\n"
+	        "	cpuid\n"
+	        "2:\n"
+	        "	pop %%ebx\n"
+	        "	mov %%edx, %0\n"
+		: "=r" (result)
+		:
+		: "%eax", "%ecx", "%edx"
+	        );
+#elif defined _MSC_VER
+	    _asm {
+		push ebx
+		mov eax, 80000000h
+		cpuid
+		xor edx, edx
+		cmp eax, 1
+		jge notamd
+		mov eax, 80000001h
+		cpuid
+	    notamd:
+		pop ebx
+		mov result, edx
+	    }
+#endif
+	    if (result & (1 << 22))
+		features |= MMX_EXTENSIONS;
+	}
+    }
+#endif /* HAVE_GETISAX */
+
+    return features;
+}
+
+static pixman_bool_t
+pixman_have_mmx (void)
+{
+    static pixman_bool_t initialized = FALSE;
+    static pixman_bool_t mmx_present;
+
+    if (!initialized)
+    {
+	unsigned int features = detect_cpu_features ();
+	mmx_present = (features & (MMX | MMX_EXTENSIONS)) == (MMX | MMX_EXTENSIONS);
+	initialized = TRUE;
+    }
+
+    return mmx_present;
+}
+
+#ifdef USE_SSE2
+static pixman_bool_t
+pixman_have_sse2 (void)
+{
+    static pixman_bool_t initialized = FALSE;
+    static pixman_bool_t sse2_present;
+
+    if (!initialized)
+    {
+	unsigned int features = detect_cpu_features ();
+	sse2_present = (features & (MMX | MMX_EXTENSIONS | SSE | SSE2)) == (MMX | MMX_EXTENSIONS | SSE | SSE2);
+	initialized = TRUE;
+    }
+
+    return sse2_present;
+}
+
+#endif
+
+#else /* __amd64__ */
+#ifdef USE_X86_MMX
+#define pixman_have_mmx() TRUE
+#endif
+#ifdef USE_SSE2
+#define pixman_have_sse2() TRUE
+#endif
+#endif /* __amd64__ */
+#endif
+
+pixman_implementation_t *
+_pixman_choose_implementation (void)
+{
+    pixman_implementation_t *imp;
+
+    imp = _pixman_implementation_create_general();
+    imp = _pixman_implementation_create_fast_path (imp);
+    
+#ifdef USE_X86_MMX
+    if (pixman_have_mmx ())
+	imp = _pixman_implementation_create_mmx (imp);
+#endif
+
+#ifdef USE_SSE2
+    if (pixman_have_sse2 ())
+	imp = _pixman_implementation_create_sse2 (imp);
+#endif
+
+#ifdef USE_ARM_SIMD
+    if (pixman_have_arm_simd ())
+	imp = _pixman_implementation_create_arm_simd (imp);
+#endif
+
+#ifdef USE_ARM_IWMMXT
+    if (pixman_have_arm_iwmmxt ())
+	imp = _pixman_implementation_create_mmx (imp);
+#endif
+
+#ifdef USE_ARM_NEON
+    if (pixman_have_arm_neon ())
+	imp = _pixman_implementation_create_arm_neon (imp);
+#endif
+
+#ifdef USE_VMX
+    if (pixman_have_vmx ())
+	imp = _pixman_implementation_create_vmx (imp);
+#endif
+
+    imp = _pixman_implementation_create_noop (imp);
+    
+    return imp;
+}
+
diff --git a/pixman/pixman/pixman-mmx.c b/pixman/pixman/pixman-mmx.c
index 11644e87b..9c428a3de 100644
--- a/pixman/pixman/pixman-mmx.c
+++ b/pixman/pixman/pixman-mmx.c
@@ -33,7 +33,7 @@
 #include <config.h>
 #endif
 
-#ifdef USE_MMX
+#if defined USE_X86_MMX || defined USE_ARM_IWMMXT
 
 #include <mmintrin.h>
 #include "pixman-private.h"
@@ -47,6 +47,15 @@
 #define CHECKPOINT()
 #endif
 
+#ifdef USE_ARM_IWMMXT
+/* Empty the multimedia state. For some reason, ARM's mmintrin.h doesn't provide this.  */
+extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_empty (void)
+{
+
+}
+#endif
+
 /* Notes about writing mmx code
  *
  * give memory operands as the second operand. If you give it as the
@@ -302,6 +311,39 @@ in_over (__m64 src, __m64 srca, __m64 mask, __m64 dest)
 
 #endif
 
+/* Elemental unaligned loads */
+
+static __inline__ uint64_t ldq_u(uint64_t *p)
+{
+#ifdef USE_X86_MMX
+    /* x86's alignment restrictions are very relaxed. */
+    return *p;
+#elif defined USE_ARM_IWMMXT
+    int align = (uintptr_t)p & 7;
+    __m64 *aligned_p;
+    if (align == 0)
+	return *p;
+    aligned_p = (__m64 *)((uintptr_t)p & ~7);
+    return _mm_align_si64 (aligned_p[0], aligned_p[1], align);
+#else
+    struct __una_u64 { uint64_t x __attribute__((packed)); };
+    const struct __una_u64 *ptr = (const struct __una_u64 *) p;
+    return ptr->x;
+#endif
+}
+
+static __inline__ uint32_t ldl_u(uint32_t *p)
+{
+#ifdef USE_X86_MMX
+    /* x86's alignment restrictions are very relaxed. */
+    return *p;
+#else
+    struct __una_u32 { uint32_t x __attribute__((packed)); };
+    const struct __una_u32 *ptr = (const struct __una_u32 *) p;
+    return ptr->x;
+#endif
+}
+
 static force_inline __m64
 load8888 (uint32_t v)
 {
@@ -1382,7 +1424,7 @@ mmx_composite_over_8888_n_8888 (pixman_implementation_t *imp,
 
 	while (w >= 2)
 	{
-	    __m64 vs = *(__m64 *)src;
+	    __m64 vs = (__m64)ldq_u((uint64_t *)src);
 	    __m64 vd = *(__m64 *)dst;
 	    __m64 vsrc0 = expand8888 (vs, 0);
 	    __m64 vsrc1 = expand8888 (vs, 1);
@@ -1463,14 +1505,14 @@ mmx_composite_over_x888_n_8888 (pixman_implementation_t *imp,
 	    __m64 vd6 = *(__m64 *)(dst + 12);
 	    __m64 vd7 = *(__m64 *)(dst + 14);
 
-	    __m64 vs0 = *(__m64 *)(src + 0);
-	    __m64 vs1 = *(__m64 *)(src + 2);
-	    __m64 vs2 = *(__m64 *)(src + 4);
-	    __m64 vs3 = *(__m64 *)(src + 6);
-	    __m64 vs4 = *(__m64 *)(src + 8);
-	    __m64 vs5 = *(__m64 *)(src + 10);
-	    __m64 vs6 = *(__m64 *)(src + 12);
-	    __m64 vs7 = *(__m64 *)(src + 14);
+	    __m64 vs0 = (__m64)ldq_u((uint64_t *)(src + 0));
+	    __m64 vs1 = (__m64)ldq_u((uint64_t *)(src + 2));
+	    __m64 vs2 = (__m64)ldq_u((uint64_t *)(src + 4));
+	    __m64 vs3 = (__m64)ldq_u((uint64_t *)(src + 6));
+	    __m64 vs4 = (__m64)ldq_u((uint64_t *)(src + 8));
+	    __m64 vs5 = (__m64)ldq_u((uint64_t *)(src + 10));
+	    __m64 vs6 = (__m64)ldq_u((uint64_t *)(src + 12));
+	    __m64 vs7 = (__m64)ldq_u((uint64_t *)(src + 14));
 
 	    vd0 = pack8888 (
 	        in_over (expandx888 (vs0, 0), srca, vmask, expand8888 (vd0, 0)),
@@ -1800,7 +1842,7 @@ pixman_fill_mmx (uint32_t *bits,
     uint32_t byte_width;
     uint8_t     *byte_line;
 
-#ifdef __GNUC__
+#if defined __GNUC__ && defined USE_X86_MMX
     __m64 v1, v2, v3, v4, v5, v6, v7;
 #endif
 
@@ -1834,7 +1876,7 @@ pixman_fill_mmx (uint32_t *bits,
     fill = ((uint64_t)xor << 32) | xor;
     vfill = to_m64 (fill);
 
-#ifdef __GNUC__
+#if defined __GNUC__ && defined USE_X86_MMX
     __asm__ (
         "movq		%7,	%0\n"
         "movq		%7,	%1\n"
@@ -1880,7 +1922,7 @@ pixman_fill_mmx (uint32_t *bits,
 
 	while (w >= 64)
 	{
-#ifdef __GNUC__
+#if defined __GNUC__ && defined USE_X86_MMX
 	    __asm__ (
 	        "movq	%1,	  (%0)\n"
 	        "movq	%2,	 8(%0)\n"
@@ -2507,23 +2549,35 @@ mmx_composite_in_n_8_8 (pixman_implementation_t *imp,
 	mask_line += mask_stride;
 	w = width;
 
-	if ((((unsigned long)dest_image & 3) == 0) &&
-	    (((unsigned long)src_image & 3) == 0))
+	while (w && (unsigned long)dst & 7)
 	{
-	    while (w >= 4)
-	    {
-		__m64 vmask;
-		__m64 vdest;
+	    uint16_t tmp;
+	    uint8_t a;
+	    uint32_t m, d;
 
-		vmask = load8888 (*(uint32_t *)mask);
-		vdest = load8888 (*(uint32_t *)dst);
+	    a = *mask++;
+	    d = *dst;
 
-		*(uint32_t *)dst = store8888 (in (in (vsrca, vmask), vdest));
+	    m = MUL_UN8 (sa, a, tmp);
+	    d = MUL_UN8 (m, d, tmp);
 
-		dst += 4;
-		mask += 4;
-		w -= 4;
-	    }
+	    *dst++ = d;
+	    w--;
+	}
+
+	while (w >= 4)
+	{
+	    __m64 vmask;
+	    __m64 vdest;
+
+	    vmask = load8888 (ldl_u((uint32_t *)mask));
+	    vdest = load8888 (*(uint32_t *)dst);
+
+	    *(uint32_t *)dst = store8888 (in (in (vsrca, vmask), vdest));
+
+	    dst += 4;
+	    mask += 4;
+	    w -= 4;
 	}
 
 	while (w--)
@@ -2566,20 +2620,31 @@ mmx_composite_in_8_8 (pixman_implementation_t *imp,
 	src_line += src_stride;
 	w = width;
 
-	if ((((unsigned long)dest_image & 3) == 0) &&
-	    (((unsigned long)src_image & 3) == 0))
+	while (w && (unsigned long)dst & 3)
 	{
-	    while (w >= 4)
-	    {
-		uint32_t *s = (uint32_t *)src;
-		uint32_t *d = (uint32_t *)dst;
+	    uint8_t s, d;
+	    uint16_t tmp;
 
-		*d = store8888 (in (load8888 (*s), load8888 (*d)));
+	    s = *src;
+	    d = *dst;
 
-		w -= 4;
-		dst += 4;
-		src += 4;
-	    }
+	    *dst = MUL_UN8 (s, d, tmp);
+
+	    src++;
+	    dst++;
+	    w--;
+	}
+
+	while (w >= 4)
+	{
+	    uint32_t *s = (uint32_t *)src;
+	    uint32_t *d = (uint32_t *)dst;
+
+	    *d = store8888 (in (load8888 (ldl_u((uint32_t *)s)), load8888 (*d)));
+
+	    w -= 4;
+	    dst += 4;
+	    src += 4;
 	}
 
 	while (w--)
@@ -2634,20 +2699,36 @@ mmx_composite_add_n_8_8 (pixman_implementation_t *imp,
 	mask_line += mask_stride;
 	w = width;
 
-	if ((((unsigned long)mask_image & 3) == 0) &&
-	    (((unsigned long)dest_image  & 3) == 0))
+	while (w && (unsigned long)dst & 3)
 	{
-	    while (w >= 4)
-	    {
-		__m64 vmask = load8888 (*(uint32_t *)mask);
-		__m64 vdest = load8888 (*(uint32_t *)dst);
+	    uint16_t tmp;
+	    uint16_t a;
+	    uint32_t m, d;
+	    uint32_t r;
 
-		*(uint32_t *)dst = store8888 (_mm_adds_pu8 (in (vsrca, vmask), vdest));
+	    a = *mask++;
+	    d = *dst;
 
-		w -= 4;
-		dst += 4;
-		mask += 4;
-	    }
+	    m = MUL_UN8 (sa, a, tmp);
+	    r = ADD_UN8 (m, d, tmp);
+
+	    *dst++ = r;
+	    w--;
+	}
+
+	while (w >= 4)
+	{
+	    __m64 vmask;
+	    __m64 vdest;
+
+	    vmask = load8888 (ldl_u((uint32_t *)mask));
+	    vdest = load8888 (*(uint32_t *)dst);
+
+	    *(uint32_t *)dst = store8888 (_mm_adds_pu8 (in (vsrca, vmask), vdest));
+
+	    dst += 4;
+	    mask += 4;
+	    w -= 4;
 	}
 
 	while (w--)
@@ -2710,7 +2791,7 @@ mmx_composite_add_8_8 (pixman_implementation_t *imp,
 
 	while (w >= 8)
 	{
-	    *(__m64*)dst = _mm_adds_pu8 (*(__m64*)src, *(__m64*)dst);
+	    *(__m64*)dst = _mm_adds_pu8 ((__m64)ldq_u((uint64_t *)src), *(__m64*)dst);
 	    dst += 8;
 	    src += 8;
 	    w -= 8;
@@ -2768,7 +2849,7 @@ mmx_composite_add_8888_8888 (pixman_implementation_t *imp,
 
 	while (w >= 2)
 	{
-	    dst64 = _mm_adds_pu8 (*(__m64*)src, *(__m64*)dst);
+	    dst64 = _mm_adds_pu8 ((__m64)ldq_u((uint64_t *)src), *(__m64*)dst);
 	    *(uint64_t*)dst = to_uint64 (dst64);
 	    dst += 2;
 	    src += 2;
@@ -2841,6 +2922,14 @@ pixman_blt_mmx (uint32_t *src_bits,
 	dst_bytes += dst_stride;
 	w = byte_width;
 
+	while (w >= 1 && ((unsigned long)d & 1))
+	{
+	    *(uint8_t *)d = *(uint8_t *)s;
+	    w -= 1;
+	    s += 1;
+	    d += 1;
+	}
+
 	while (w >= 2 && ((unsigned long)d & 3))
 	{
 	    *(uint16_t *)d = *(uint16_t *)s;
@@ -2851,7 +2940,7 @@ pixman_blt_mmx (uint32_t *src_bits,
 
 	while (w >= 4 && ((unsigned long)d & 7))
 	{
-	    *(uint32_t *)d = *(uint32_t *)s;
+	    *(uint32_t *)d = ldl_u((uint32_t *)s);
 
 	    w -= 4;
 	    s += 4;
@@ -2860,7 +2949,7 @@ pixman_blt_mmx (uint32_t *src_bits,
 
 	while (w >= 64)
 	{
-#if defined (__GNUC__) || (defined(__SUNPRO_C) && (__SUNPRO_C >= 0x590))
+#if (defined (__GNUC__) || (defined(__SUNPRO_C) && (__SUNPRO_C >= 0x590))) && defined USE_X86_MMX
 	    __asm__ (
 	        "movq	  (%1),	  %%mm0\n"
 	        "movq	 8(%1),	  %%mm1\n"
@@ -2885,14 +2974,14 @@ pixman_blt_mmx (uint32_t *src_bits,
 		  "%mm0", "%mm1", "%mm2", "%mm3",
 		  "%mm4", "%mm5", "%mm6", "%mm7");
 #else
-	    __m64 v0 = *(__m64 *)(s + 0);
-	    __m64 v1 = *(__m64 *)(s + 8);
-	    __m64 v2 = *(__m64 *)(s + 16);
-	    __m64 v3 = *(__m64 *)(s + 24);
-	    __m64 v4 = *(__m64 *)(s + 32);
-	    __m64 v5 = *(__m64 *)(s + 40);
-	    __m64 v6 = *(__m64 *)(s + 48);
-	    __m64 v7 = *(__m64 *)(s + 56);
+	    __m64 v0 = ldq_u((uint64_t *)(s + 0));
+	    __m64 v1 = ldq_u((uint64_t *)(s + 8));
+	    __m64 v2 = ldq_u((uint64_t *)(s + 16));
+	    __m64 v3 = ldq_u((uint64_t *)(s + 24));
+	    __m64 v4 = ldq_u((uint64_t *)(s + 32));
+	    __m64 v5 = ldq_u((uint64_t *)(s + 40));
+	    __m64 v6 = ldq_u((uint64_t *)(s + 48));
+	    __m64 v7 = ldq_u((uint64_t *)(s + 56));
 	    *(__m64 *)(d + 0)  = v0;
 	    *(__m64 *)(d + 8)  = v1;
 	    *(__m64 *)(d + 16) = v2;
@@ -2909,7 +2998,7 @@ pixman_blt_mmx (uint32_t *src_bits,
 	}
 	while (w >= 4)
 	{
-	    *(uint32_t *)d = *(uint32_t *)s;
+	    *(uint32_t *)d = ldl_u((uint32_t *)s);
 
 	    w -= 4;
 	    s += 4;
@@ -3161,4 +3250,4 @@ _pixman_implementation_create_mmx (pixman_implementation_t *fallback)
     return imp;
 }
 
-#endif /* USE_MMX */
+#endif /* USE_X86_MMX || USE_ARM_IWMMXT */
diff --git a/pixman/pixman/pixman-private.h b/pixman/pixman/pixman-private.h
index 6e716c6e6..4d645fe4b 100644
--- a/pixman/pixman/pixman-private.h
+++ b/pixman/pixman/pixman-private.h
@@ -545,7 +545,7 @@ _pixman_implementation_create_fast_path (pixman_implementation_t *fallback);
 pixman_implementation_t *
 _pixman_implementation_create_noop (pixman_implementation_t *fallback);
 
-#ifdef USE_MMX
+#if defined USE_X86_MMX || defined USE_ARM_IWMMXT
 pixman_implementation_t *
 _pixman_implementation_create_mmx (pixman_implementation_t *fallback);
 #endif
diff --git a/pixman/test/Makefile.am b/pixman/test/Makefile.am
index 802d3f4fc..eeb3679f0 100644
--- a/pixman/test/Makefile.am
+++ b/pixman/test/Makefile.am
@@ -1,9 +1,9 @@
 include $(top_srcdir)/test/Makefile.sources
 
-AM_CFLAGS = @OPENMP_CFLAGS@
-AM_LDFLAGS = @OPENMP_CFLAGS@ @TESTPROGS_EXTRA_LDFLAGS@
-LDADD = libutils.la $(top_builddir)/pixman/libpixman-1.la -lm  @PNG_LIBS@
-INCLUDES = -I$(top_srcdir)/pixman -I$(top_builddir)/pixman @PNG_CFLAGS@
+AM_CFLAGS = $(OPENMP_CFLAGS)
+AM_LDFLAGS = $(OPENMP_CFLAGS) $(TESTPROGS_EXTRA_LDFLAGS)
+LDADD = libutils.la $(top_builddir)/pixman/libpixman-1.la -lm  $(PNG_LIBS)
+INCLUDES = -I$(top_srcdir)/pixman -I$(top_builddir)/pixman $(PNG_CFLAGS)
 
 libutils_la_SOURCES = $(libutils_sources) $(libutils_headers)