Merge svn changes up to 27824

Conflicts:
	cfg-common-opts.h
	libmpcodecs/dec_video.c
	libmpcodecs/vd.c
	libvo/x11_common.h
	mplayer.c
	stream/cache2.c

31 files changed, 95 insertions, 120 deletions

diff --git a/libmpcodecs/ad_ffmpeg.c b/libmpcodecs/ad_ffmpeg.c
index a8a8a18120..560dd90c0a 100644
--- a/libmpcodecs/ad_ffmpeg.c
+++ b/libmpcodecs/ad_ffmpeg.c
@@ -177,6 +177,8 @@ static int decode_audio(sh_audio_t *sh_audio,unsigned char *buf,int minlen,int m
               src_ch_layout = AF_CHANNEL_LAYOUT_LAVC_LIBA52_DEFAULT;
             else if (!strcasecmp(codec, "vorbis"))
               src_ch_layout = AF_CHANNEL_LAYOUT_VORBIS_DEFAULT;
+            else if (!strcasecmp(codec, "flac"))
+              src_ch_layout = AF_CHANNEL_LAYOUT_FLAC_DEFAULT;
             else
               src_ch_layout = AF_CHANNEL_LAYOUT_MPLAYER_DEFAULT;
             reorder_channel_nch(buf, src_ch_layout,
diff --git a/libmpcodecs/ad_twin.c b/libmpcodecs/ad_twin.c
index 3915101ae4..51b260d061 100644
--- a/libmpcodecs/ad_twin.c
+++ b/libmpcodecs/ad_twin.c
@@ -138,7 +138,7 @@ static int init_vqf_audio_codec(sh_audio_t *sh_audio){
     priv->framesize=TvqGetFrameSize();
     sh_audio->audio_in_minsize=priv->framesize*in_fmt->nChannels;
     sh_audio->a_in_buffer_size=4*sh_audio->audio_in_minsize;
-    sh_audio->a_in_buffer=malloc(sh_audio->a_in_buffer_size);
+    sh_audio->a_in_buffer=av_malloc(sh_audio->a_in_buffer_size);
     sh_audio->a_in_buffer_len=0;
 
 
diff --git a/libmpcodecs/dec_audio.c b/libmpcodecs/dec_audio.c
index 23f19ac405..947b47bbfc 100644
--- a/libmpcodecs/dec_audio.c
+++ b/libmpcodecs/dec_audio.c
@@ -74,8 +74,7 @@ static int init_audio_codec(sh_audio_t *sh_audio)
 	sh_audio->a_in_buffer_size = sh_audio->audio_in_minsize;
 	mp_msg(MSGT_DECAUDIO, MSGL_V, MSGTR_AllocatingBytesForInputBuffer,
 	       sh_audio->a_in_buffer_size);
-	sh_audio->a_in_buffer = memalign(16, sh_audio->a_in_buffer_size);
-	memset(sh_audio->a_in_buffer, 0, sh_audio->a_in_buffer_size);
+	sh_audio->a_in_buffer = av_mallocz(sh_audio->a_in_buffer_size);
 	sh_audio->a_in_buffer_len = 0;
     }
 
@@ -84,12 +83,11 @@ static int init_audio_codec(sh_audio_t *sh_audio)
     mp_msg(MSGT_DECAUDIO, MSGL_V, MSGTR_AllocatingBytesForOutputBuffer,
 	   sh_audio->audio_out_minsize, MAX_OUTBURST, sh_audio->a_buffer_size);
 
-    sh_audio->a_buffer = memalign(16, sh_audio->a_buffer_size);
+    sh_audio->a_buffer = av_mallocz(sh_audio->a_buffer_size);
     if (!sh_audio->a_buffer) {
 	mp_msg(MSGT_DECAUDIO, MSGL_ERR, MSGTR_CantAllocAudioBuf);
 	return 0;
     }
-    memset(sh_audio->a_buffer, 0, sh_audio->a_buffer_size);
     sh_audio->a_buffer_len = 0;
 
     if (!sh_audio->ad_driver->init(sh_audio)) {
@@ -306,12 +304,8 @@ void uninit_audio(sh_audio_t *sh_audio)
     free(sh_audio->a_out_buffer);
     sh_audio->a_out_buffer = NULL;
     sh_audio->a_out_buffer_size = 0;
-    if (sh_audio->a_buffer)
-	free(sh_audio->a_buffer);
-    sh_audio->a_buffer = NULL;
-    if (sh_audio->a_in_buffer)
-	free(sh_audio->a_in_buffer);
-    sh_audio->a_in_buffer = NULL;
+    av_freep(&sh_audio->a_buffer);
+    av_freep(&sh_audio->a_in_buffer);
 }
 
 
diff --git a/libmpcodecs/dec_video.c b/libmpcodecs/dec_video.c
index 7a3e826e45..4259377342 100644
--- a/libmpcodecs/dec_video.c
+++ b/libmpcodecs/dec_video.c
@@ -414,9 +414,9 @@ void *decode_video(sh_video_t *sh_video, unsigned char *start, int in_size,
     // some codecs are broken, and doesn't restore MMX state :(
     // it happens usually with broken/damaged files.
     if (gCpuCaps.has3DNow) {
-        __asm __volatile("femms\n\t":::"memory");
+        __asm__ volatile("femms\n\t":::"memory");
     } else if (gCpuCaps.hasMMX) {
-        __asm __volatile("emms\n\t":::"memory");
+        __asm__ volatile("emms\n\t":::"memory");
     }
 #endif
 
diff --git a/libmpcodecs/native/mmx.h b/libmpcodecs/native/mmx.h
index d9a6fb65cb..1661b318a4 100644
--- a/libmpcodecs/native/mmx.h
+++ b/libmpcodecs/native/mmx.h
@@ -57,24 +57,24 @@ typedef union {
 
 
 #define mmx_i2r(op,imm,reg) \
-        __asm__ __volatile__ (#op " %0, %%" #reg \
+        __asm__ volatile (#op " %0, %%" #reg \
                               : /* nothing */ \
                               : "i" (imm) )
 
 #define mmx_m2r(op, mem, reg) \
-        __asm__ __volatile__ (#op " %0, %%" #reg \
+        __asm__ volatile (#op " %0, %%" #reg \
                               : /* nothing */ \
                               : "m" (mem))
 
 #define mmx_r2m(op, reg, mem) \
-        __asm__ __volatile__ (#op " %%" #reg ", %0" \
+        __asm__ volatile (#op " %%" #reg ", %0" \
                               : "=m" (mem) \
                               : /* nothing */ )
 
 #define mmx_r2r(op, regs, regd) \
-        __asm__ __volatile__ (#op " %" #regs ", %" #regd)
+        __asm__ volatile (#op " %" #regs ", %" #regd)
 
 
-#define emms() __asm__ __volatile__ ("emms")
+#define emms() __asm__ volatile ("emms")
 
 #endif /* MPLAYER_MMX_H */
diff --git a/libmpcodecs/pullup.c b/libmpcodecs/pullup.c
index b5229052a4..f71bf790ce 100644
--- a/libmpcodecs/pullup.c
+++ b/libmpcodecs/pullup.c
@@ -14,7 +14,7 @@
 static int diff_y_mmx(unsigned char *a, unsigned char *b, int s)
 {
 	int ret;
-	asm volatile (
+	__asm__ volatile (
 		"movl $4, %%ecx \n\t"
 		"pxor %%mm4, %%mm4 \n\t"
 		"pxor %%mm7, %%mm7 \n\t"
@@ -61,7 +61,7 @@ static int diff_y_mmx(unsigned char *a, unsigned char *b, int s)
 static int licomb_y_mmx(unsigned char *a, unsigned char *b, int s)
 {
 	int ret;
-	asm volatile (
+	__asm__ volatile (
 		"movl $4, %%ecx \n\t"
 		"pxor %%mm6, %%mm6 \n\t"
 		"pxor %%mm7, %%mm7 \n\t"
@@ -150,7 +150,7 @@ static int licomb_y_mmx(unsigned char *a, unsigned char *b, int s)
 static int var_y_mmx(unsigned char *a, unsigned char *b, int s)
 {
 	int ret;
-	asm volatile (
+	__asm__ volatile (
 		"movl $3, %%ecx \n\t"
 		"pxor %%mm4, %%mm4 \n\t"
 		"pxor %%mm7, %%mm7 \n\t"
diff --git a/libmpcodecs/vd.c b/libmpcodecs/vd.c
index 1e23d2eced..fe9d1d7861 100644
--- a/libmpcodecs/vd.c
+++ b/libmpcodecs/vd.c
@@ -103,6 +103,9 @@ const vd_functions_t * const mpcodecs_vd_drivers[] = {
 
 #include "libvo/video_out.h"
 
+#define SCREEN_SIZE_X 1
+#define SCREEN_SIZE_Y 1
+
 int mpcodecs_config_vo(sh_video_t *sh, int w, int h,
                        unsigned int preferred_outfmt)
 {
diff --git a/libmpcodecs/vd_libmpeg2.c b/libmpcodecs/vd_libmpeg2.c
index eeab4b23be..7505768dee 100644
--- a/libmpcodecs/vd_libmpeg2.c
+++ b/libmpcodecs/vd_libmpeg2.c
@@ -83,9 +83,7 @@ static int init(sh_video_t *sh){
     #elif ARCH_ARM
        accel |= MPEG2_ACCEL_ARM;
     #endif
-    #ifdef HAVE_IWMMXT
-       accel |= MPEG2_ACCEL_ARM_IWMMXT;
-    #elif HAVE_MVI
+    #ifdef HAVE_MVI
        accel |= MPEG2_ACCEL_ALPHA_MVI;
     #elif HAVE_VIS
        accel |= MPEG2_ACCEL_SPARC_VIS;
diff --git a/libmpcodecs/vd_zrmjpeg.c b/libmpcodecs/vd_zrmjpeg.c
index a2d1d21e8b..2d35ec1420 100644
--- a/libmpcodecs/vd_zrmjpeg.c
+++ b/libmpcodecs/vd_zrmjpeg.c
@@ -109,7 +109,7 @@ static int get_int2(unsigned char *buf) {
 static unsigned int guess_mjpeg_type(unsigned char *data, unsigned int size,
 		int d_height) {
 	unsigned int p;
-	int marker, length, height, ncomps, i, hf[3], vf[3];
+	int marker, length, height, i, hf[3], vf[3];
 	unsigned int app0 = 0, header = 0;
 
 	/* The initial marker must be SIO */
diff --git a/libmpcodecs/vf_decimate.c b/libmpcodecs/vf_decimate.c
index 5fb633d3c5..81b58cccb7 100644
--- a/libmpcodecs/vf_decimate.c
+++ b/libmpcodecs/vf_decimate.c
@@ -23,7 +23,7 @@ struct vf_priv_s {
 static int diff_MMX(unsigned char *old, unsigned char *new, int os, int ns)
 {
 	volatile short out[4];
-	asm (
+	__asm__ (
 		"movl $8, %%ecx \n\t"
 		"pxor %%mm4, %%mm4 \n\t"
 		"pxor %%mm7, %%mm7 \n\t"
diff --git a/libmpcodecs/vf_divtc.c b/libmpcodecs/vf_divtc.c
index 4dade860f1..307a14890d 100644
--- a/libmpcodecs/vf_divtc.c
+++ b/libmpcodecs/vf_divtc.c
@@ -37,7 +37,7 @@ struct vf_priv_s
 static int diff_MMX(unsigned char *old, unsigned char *new, int os, int ns)
    {
    volatile short out[4];
-   asm (
+   __asm__ (
 	"movl $8, %%ecx \n\t"
 	"pxor %%mm4, %%mm4 \n\t"
 	"pxor %%mm7, %%mm7 \n\t"
@@ -128,7 +128,7 @@ static unsigned int checksum_plane(unsigned char *p, unsigned char *z,
    unsigned int shift;
    uint32_t sum, t;
    unsigned char *e, *e2;
-#if MP_WORDSIZE==64
+#if __WORDSIZE==64
    typedef uint64_t wsum_t;
 #else
    typedef uint32_t wsum_t;
@@ -143,7 +143,7 @@ static unsigned int checksum_plane(unsigned char *p, unsigned char *z,
       for(wsum=0, e2=e-sizeof(wsum_t)+1; p<e2; p+=sizeof(wsum_t))
 	 wsum^=*(wsum_t *)p;
 
-#if MP_WORDSIZE==64
+#if __WORDSIZE==64
       t=be2me_32((uint32_t)(wsum>>32^wsum));
 #else
       t=be2me_32(wsum);
diff --git a/libmpcodecs/vf_eq.c b/libmpcodecs/vf_eq.c
index 0b65720849..a0dc04ae39 100644
--- a/libmpcodecs/vf_eq.c
+++ b/libmpcodecs/vf_eq.c
@@ -44,7 +44,7 @@ static void process_MMX(unsigned char *dest, int dstride, unsigned char *src, in
 	contvec[0] = contvec[1] = contvec[2] = contvec[3] = contrast;
 		
 	while (h--) {
-		asm volatile (
+		__asm__ volatile (
 			"movq (%5), %%mm3 \n\t"
 			"movq (%6), %%mm4 \n\t"
 			"pxor %%mm0, %%mm0 \n\t"
@@ -82,7 +82,7 @@ static void process_MMX(unsigned char *dest, int dstride, unsigned char *src, in
 		src += sstep;
 		dest += dstep;
 	}
-	asm volatile ( "emms \n\t" ::: "memory" );
+	__asm__ volatile ( "emms \n\t" ::: "memory" );
 }
 #endif
 
diff --git a/libmpcodecs/vf_eq2.c b/libmpcodecs/vf_eq2.c
index dbe751fbca..b3a501719f 100644
--- a/libmpcodecs/vf_eq2.c
+++ b/libmpcodecs/vf_eq2.c
@@ -130,7 +130,7 @@ void affine_1d_MMX (eq2_param_t *par, unsigned char *dst, unsigned char *src,
   dstep = dstride - w;
 
   while (h-- > 0) {
-    asm volatile (
+    __asm__ volatile (
       "movq (%5), %%mm3 \n\t"
       "movq (%6), %%mm4 \n\t"
       "pxor %%mm0, %%mm0 \n\t"
@@ -170,7 +170,7 @@ void affine_1d_MMX (eq2_param_t *par, unsigned char *dst, unsigned char *src,
     dst += dstep;
   }
 
-  asm volatile ( "emms \n\t" ::: "memory" );
+  __asm__ volatile ( "emms \n\t" ::: "memory" );
 }
 #endif
 
diff --git a/libmpcodecs/vf_filmdint.c b/libmpcodecs/vf_filmdint.c
index f479569536..5c022e5920 100644
--- a/libmpcodecs/vf_filmdint.c
+++ b/libmpcodecs/vf_filmdint.c
@@ -366,13 +366,13 @@ block_metrics_faster_c(unsigned char *a, unsigned char *b, int as, int bs,
 #define MEQ(X,Y) ((X).even == (Y).even && (X).odd == (Y).odd && (X).temp == (Y).temp && (X).noise == (Y).noise)
 
 #define BLOCK_METRICS_TEMPLATE() \
-    asm volatile("pxor %mm7, %mm7\n\t"   /* The result is colleted in mm7 */ \
+    __asm__ volatile("pxor %mm7, %mm7\n\t"   /* The result is colleted in mm7 */ \
 		 "pxor %mm6, %mm6\n\t"   /* Temp to stay at 0 */	     \
 	);								     \
     a -= as;								     \
     b -= bs;								     \
     do {								     \
-	asm volatile(							     \
+	__asm__ volatile(						     \
 	    "movq (%0,%2), %%mm0\n\t"					     \
 	    "movq (%1,%3), %%mm1\n\t"   /* mm1 = even */		     \
 	    PSADBW(%%mm1, %%mm0, %%mm4, %%mm6)				     \
@@ -437,7 +437,7 @@ block_metrics_3dnow(unsigned char *a, unsigned char *b, int as, int bs,
     static const unsigned long long ones = 0x0101010101010101ull;
 
     BLOCK_METRICS_TEMPLATE();
-    asm volatile("movq %%mm7, %0\n\temms" : "=m" (tm));
+    __asm__ volatile("movq %%mm7, %0\n\temms" : "=m" (tm));
     get_block_stats(&tm, p, s);
 #endif
     return tm;
@@ -469,7 +469,7 @@ block_metrics_mmx2(unsigned char *a, unsigned char *b, int as, int bs,
 #ifdef DEBUG
     struct frame_stats ts = *s;
 #endif
-    asm volatile("prefetcht0 (%0,%2)\n\t"
+    __asm__ volatile("prefetcht0 (%0,%2)\n\t"
 		 "prefetcht0 (%1,%3)\n\t" :
 		 : "r" (a), "r" (b),
 		 "r" (prefetch_line * as), "r" (prefetch_line * bs));
@@ -477,7 +477,7 @@ block_metrics_mmx2(unsigned char *a, unsigned char *b, int as, int bs,
     BLOCK_METRICS_TEMPLATE();
 
     s->num_blocks++;
-    asm volatile(
+    __asm__ volatile(
 	"movq %3, %%mm0\n\t"
 	"movq %%mm7, %%mm1\n\t"
 	"psubusw %%mm0, %%mm1\n\t"
@@ -523,7 +523,7 @@ block_metrics_mmx2(unsigned char *a, unsigned char *b, int as, int bs,
 	s->interlaced_high += interlaced >> 16;
 	s->interlaced_low += interlaced;
     } else {
-	asm volatile(
+	__asm__ volatile(
 	    "pcmpeqw %%mm0, %%mm0\n\t" /* -1 */
 	    "psubw 	%%mm0, %%mm4\n\t"
 	    "psubw 	%%mm0, %%mm5\n\t"
@@ -537,7 +537,7 @@ block_metrics_mmx2(unsigned char *a, unsigned char *b, int as, int bs,
 	    : "=m" (s->tiny), "=m" (s->low), "=m" (s->high)
 	    );
 
-	asm volatile(
+	__asm__ volatile(
 	    "pshufw $0, %2, %%mm0\n\t"
 	    "psubusw %%mm7, %%mm0\n\t"
 	    "pcmpeqw %%mm6, %%mm0\n\t"   /* 0 if below sad_thres */
@@ -554,7 +554,7 @@ block_metrics_mmx2(unsigned char *a, unsigned char *b, int as, int bs,
 	    );
     }
 
-    asm volatile(
+    __asm__ volatile(
 	"movq %%mm7, (%1)\n\t"
 	PMAXUW((%0), %%mm7)
 	"movq %%mm7, (%0)\n\t"
@@ -595,7 +595,7 @@ dint_copy_line_mmx2(unsigned char *dst, unsigned char *a, long bos,
 #else
     unsigned long len = (w+7) >> 3;
     int ret;
-    asm volatile (
+    __asm__ volatile (
 	"pxor %%mm6, %%mm6 \n\t"       /* deinterlaced pixel counter */
 	"movd %0, %%mm7 \n\t"
 	"punpcklbw %%mm7, %%mm7 \n\t"
@@ -605,7 +605,7 @@ dint_copy_line_mmx2(unsigned char *dst, unsigned char *a, long bos,
 	: "rm" (t)
 	);
     do {
-	asm volatile (
+	__asm__ volatile (
 	    "movq (%0), %%mm0\n\t"
 	    "movq (%0,%3,2), %%mm1\n\t"
 	    "movq %%mm0, (%2)\n\t"
@@ -637,7 +637,7 @@ dint_copy_line_mmx2(unsigned char *dst, unsigned char *a, long bos,
 	dst += 8;
     } while (--len);
 
-    asm volatile ("pxor %%mm7, %%mm7 \n\t"
+    __asm__ volatile ("pxor %%mm7, %%mm7 \n\t"
 		  "psadbw %%mm6, %%mm7 \n\t"
 		  "movd %%mm7, %0 \n\t"
 		  "emms \n\t"
diff --git a/libmpcodecs/vf_fspp.c b/libmpcodecs/vf_fspp.c
index a6c2c58a3b..46c8468745 100644
--- a/libmpcodecs/vf_fspp.c
+++ b/libmpcodecs/vf_fspp.c
@@ -187,7 +187,7 @@ static void store_slice_mmx(uint8_t *dst, int16_t *src, long dst_stride, long sr
     width = (width+7)&~7;
     dst_stride-=width;
     //src_stride=(src_stride-width)*2;
-    asm volatile(
+    __asm__ volatile(
 	"mov %5, %%"REG_d"                \n\t"
 	"mov %6, %%"REG_S"                \n\t"
 	"mov %7, %%"REG_D"                \n\t"
@@ -255,7 +255,7 @@ static void store_slice2_mmx(uint8_t *dst, int16_t *src, long dst_stride, long s
     width = (width+7)&~7;
     dst_stride-=width;
     //src_stride=(src_stride-width)*2;
-    asm volatile(
+    __asm__ volatile(
 	"mov %5, %%"REG_d"                \n\t"
 	"mov %6, %%"REG_S"                \n\t"
 	"mov %7, %%"REG_D"                \n\t"
@@ -318,7 +318,7 @@ static void store_slice2_mmx(uint8_t *dst, int16_t *src, long dst_stride, long s
 static void mul_thrmat_mmx(struct vf_priv_s *p, int q)
 {
     uint64_t *adr=&p->threshold_mtx_noq[0];
-    asm volatile(
+    __asm__ volatile(
 	"movd %0, %%mm7                \n\t"
 	"add $8*8*2, %%"REG_D"            \n\t"
 	"movq 0*8(%%"REG_S"), %%mm0        \n\t"
@@ -558,10 +558,10 @@ static int put_image(struct vf_instance* vf, mp_image_t *mpi, double pts)
     }
 
 #ifdef HAVE_MMX
-    if(gCpuCaps.hasMMX) asm volatile ("emms\n\t");
+    if(gCpuCaps.hasMMX) __asm__ volatile ("emms\n\t");
 #endif
 #ifdef HAVE_MMX2
-    if(gCpuCaps.hasMMX2) asm volatile ("sfence\n\t");
+    if(gCpuCaps.hasMMX2) __asm__ volatile ("sfence\n\t");
 #endif
     return vf_next_put_image(vf,dmpi, pts);
 }
@@ -868,7 +868,7 @@ static void column_fidct_c(int16_t* thr_adr, DCTELEM *data, DCTELEM *output, int
 static void column_fidct_mmx(int16_t* thr_adr,  DCTELEM *data,  DCTELEM *output,  int cnt)
 {
     uint64_t __attribute__((aligned(8))) temps[4];
-    asm volatile(
+    __asm__ volatile(
 	ASMALIGN(4)
 	"1:                   \n\t"
 	"movq "DCTSIZE_S"*0*2(%%"REG_S"), %%mm1 \n\t"
@@ -1669,7 +1669,7 @@ static void row_idct_mmx (DCTELEM* workspace,
 			  int16_t* output_adr,  int output_stride,  int cnt)
 {
     uint64_t __attribute__((aligned(8))) temps[4];
-    asm volatile(
+    __asm__ volatile(
 	"lea (%%"REG_a",%%"REG_a",2), %%"REG_d"    \n\t"
 	"1:                     \n\t"
 	"movq "DCTSIZE_S"*0*2(%%"REG_S"), %%mm0 \n\t"
@@ -1935,7 +1935,7 @@ static void row_fdct_c(DCTELEM *data, const uint8_t *pixels, int line_size, int
 static void row_fdct_mmx(DCTELEM *data,  const uint8_t *pixels,  int line_size,  int cnt)
 {
     uint64_t __attribute__((aligned(8))) temps[4];
-    asm volatile(
+    __asm__ volatile(
 	"lea (%%"REG_a",%%"REG_a",2), %%"REG_d"    \n\t"
 	"6:                     \n\t"
 	"movd (%%"REG_S"), %%mm0           \n\t"
diff --git a/libmpcodecs/vf_geq.c b/libmpcodecs/vf_geq.c
index e8f16776d7..d2cbe6554b 100644
--- a/libmpcodecs/vf_geq.c
+++ b/libmpcodecs/vf_geq.c
@@ -33,7 +33,6 @@
 #include "mp_image.h"
 #include "vf.h"
 
-#define HAVE_AV_CONFIG_H
 #include "libavcodec/avcodec.h"
 #include "libavcodec/eval.h"
 
diff --git a/libmpcodecs/vf_halfpack.c b/libmpcodecs/vf_halfpack.c
index 39ba69bb81..f15a9a439b 100644
--- a/libmpcodecs/vf_halfpack.c
+++ b/libmpcodecs/vf_halfpack.c
@@ -37,7 +37,7 @@ static void halfpack_MMX(unsigned char *dst, unsigned char *src[3],
 	vinc = srcstride[2] - w/2;
 
 	for (h/=2; h; h--) {
-		asm (
+		__asm__ (
 			"pxor %%mm0, %%mm0 \n\t"
 			ASMALIGN(4)
 			"1: \n\t"
@@ -99,7 +99,7 @@ static void halfpack_MMX(unsigned char *dst, unsigned char *src[3],
 		v += vinc;
 		dst += dstinc;
 	}
-	asm volatile ( "emms \n\t" ::: "memory" );
+	__asm__ volatile ( "emms \n\t" ::: "memory" );
 }
 #endif
 
diff --git a/libmpcodecs/vf_ilpack.c b/libmpcodecs/vf_ilpack.c
index 0ffca173da..4b5612643d 100644
--- a/libmpcodecs/vf_ilpack.c
+++ b/libmpcodecs/vf_ilpack.c
@@ -61,7 +61,7 @@ static void pack_li_1_C(unsigned char *dst, unsigned char *y,
 static void pack_nn_MMX(unsigned char *dst, unsigned char *y,
 	unsigned char *u, unsigned char *v, int w)
 {
-	asm volatile (""
+	__asm__ volatile (""
 		ASMALIGN(4)
 		"1: \n\t"
 		"movq (%0), %%mm1 \n\t"
@@ -91,7 +91,7 @@ static void pack_nn_MMX(unsigned char *dst, unsigned char *y,
 static void pack_li_0_MMX(unsigned char *dst, unsigned char *y,
 	unsigned char *u, unsigned char *v, int w, int us, int vs)
 {
-	asm volatile (""
+	__asm__ volatile (""
 		"push %%"REG_BP" \n\t"
 #ifdef ARCH_X86_64
 		"mov %6, %%"REG_BP" \n\t"
@@ -199,7 +199,7 @@ static void pack_li_0_MMX(unsigned char *dst, unsigned char *y,
 static void pack_li_1_MMX(unsigned char *dst, unsigned char *y,
 	unsigned char *u, unsigned char *v, int w, int us, int vs)
 {
-	asm volatile (""
+	__asm__ volatile (""
 		"push %%"REG_BP" \n\t"
 #ifdef ARCH_X86_64
 		"mov %6, %%"REG_BP" \n\t"
diff --git a/libmpcodecs/vf_ivtc.c b/libmpcodecs/vf_ivtc.c
index c54405a4eb..450ccb0a3c 100644
--- a/libmpcodecs/vf_ivtc.c
+++ b/libmpcodecs/vf_ivtc.c
@@ -46,7 +46,7 @@ static void block_diffs_MMX(struct metrics *m, unsigned char *old, unsigned char
 	int i;
 	short out[24]; // output buffer for the partial metrics from the mmx code
 	
-	asm (
+	__asm__ (
 		"movl $4, %%ecx \n\t"
 		"pxor %%mm4, %%mm4 \n\t" // 4 even difference sums
 		"pxor %%mm5, %%mm5 \n\t" // 4 odd difference sums
@@ -105,7 +105,7 @@ static void block_diffs_MMX(struct metrics *m, unsigned char *old, unsigned char
 	m->o = out[4]+out[5]+out[6]+out[7];
 	m->d = m->e + m->o;
 
-	asm (
+	__asm__ (
 		// First loop to measure first four columns
 		"movl $4, %%ecx \n\t"
 		"pxor %%mm4, %%mm4 \n\t" // Past spacial noise
diff --git a/libmpcodecs/vf_noise.c b/libmpcodecs/vf_noise.c
index ab88cb5964..02fe9d9e16 100644
--- a/libmpcodecs/vf_noise.c
+++ b/libmpcodecs/vf_noise.c
@@ -150,7 +150,7 @@ static inline void lineNoise_MMX(uint8_t *dst, uint8_t *src, int8_t *noise, int
 	long mmx_len= len&(~7);
 	noise+=shift;
 
-	asm volatile(
+	__asm__ volatile(
 		"mov %3, %%"REG_a"		\n\t"
 		"pcmpeqb %%mm7, %%mm7		\n\t"
 		"psllw $15, %%mm7		\n\t"
@@ -179,7 +179,7 @@ static inline void lineNoise_MMX2(uint8_t *dst, uint8_t *src, int8_t *noise, int
 	long mmx_len= len&(~7);
 	noise+=shift;
 
-	asm volatile(
+	__asm__ volatile(
 		"mov %3, %%"REG_a"		\n\t"
 		"pcmpeqb %%mm7, %%mm7		\n\t"
 		"psllw $15, %%mm7		\n\t"
@@ -220,7 +220,7 @@ static inline void lineNoise_C(uint8_t *dst, uint8_t *src, int8_t *noise, int le
 static inline void lineNoiseAvg_MMX(uint8_t *dst, uint8_t *src, int len, int8_t **shift){
 	long mmx_len= len&(~7);
 
-	asm volatile(
+	__asm__ volatile(
 		"mov %5, %%"REG_a"		\n\t"
 		ASMALIGN(4)
 		"1:				\n\t"
@@ -357,10 +357,10 @@ static int put_image(struct vf_instance* vf, mp_image_t *mpi, double pts){
         vf_clone_mpi_attributes(dmpi, mpi);
 
 #ifdef HAVE_MMX
-	if(gCpuCaps.hasMMX) asm volatile ("emms\n\t");
+	if(gCpuCaps.hasMMX) __asm__ volatile ("emms\n\t");
 #endif
 #ifdef HAVE_MMX2
-	if(gCpuCaps.hasMMX2) asm volatile ("sfence\n\t");
+	if(gCpuCaps.hasMMX2) __asm__ volatile ("sfence\n\t");
 #endif
 
 	return vf_next_put_image(vf,dmpi, pts);
diff --git a/libmpcodecs/vf_palette.c b/libmpcodecs/vf_palette.c
index 449516f758..b2105f9487 100644
--- a/libmpcodecs/vf_palette.c
+++ b/libmpcodecs/vf_palette.c
@@ -108,15 +108,15 @@ static int put_image(struct vf_instance* vf, mp_image_t *mpi, double pts){
 	    break;
 	case 24:
 	    if (IMGFMT_IS_BGR(dmpi->imgfmt))
-		palette8tobgr24(mpi->planes[0],dmpi->planes[0],mpi->h*mpi->w,mpi->planes[1]);
+		palette8topacked24(mpi->planes[0],dmpi->planes[0],mpi->h*mpi->w,mpi->planes[1]);
 	    else
-		palette8torgb24(mpi->planes[0],dmpi->planes[0],mpi->h*mpi->w,mpi->planes[1]);
+		palette8topacked24(mpi->planes[0],dmpi->planes[0],mpi->h*mpi->w,mpi->planes[1]);
 	    break;
 	case 32:
 	    if (IMGFMT_IS_BGR(dmpi->imgfmt))
-		palette8tobgr32(mpi->planes[0],dmpi->planes[0],mpi->h*mpi->w,mpi->planes[1]);
+		palette8topacked32(mpi->planes[0],dmpi->planes[0],mpi->h*mpi->w,mpi->planes[1]);
 	    else
-		palette8torgb32(mpi->planes[0],dmpi->planes[0],mpi->h*mpi->w,mpi->planes[1]);
+		palette8topacked32(mpi->planes[0],dmpi->planes[0],mpi->h*mpi->w,mpi->planes[1]);
 	    break;
 	}
     } else {
@@ -139,15 +139,15 @@ static int put_image(struct vf_instance* vf, mp_image_t *mpi, double pts){
 		break;
 	    case 24:
 		if (IMGFMT_IS_BGR(dmpi->imgfmt))
-		    palette8tobgr24(src,dst,mpi->w,mpi->planes[1]);
+		    palette8topacked24(src,dst,mpi->w,mpi->planes[1]);
 		else
-		    palette8torgb24(src,dst,mpi->w,mpi->planes[1]);
+		    palette8topacked24(src,dst,mpi->w,mpi->planes[1]);
 		break;
 	    case 32:
 		if (IMGFMT_IS_BGR(dmpi->imgfmt))
-		    palette8tobgr32(src,dst,mpi->w,mpi->planes[1]);
+		    palette8topacked32(src,dst,mpi->w,mpi->planes[1]);
 		else
-		    palette8torgb32(src,dst,mpi->w,mpi->planes[1]);
+		    palette8topacked32(src,dst,mpi->w,mpi->planes[1]);
 		break;
 	    }
 	}
diff --git a/libmpcodecs/vf_pp7.c b/libmpcodecs/vf_pp7.c
index 2fc76b1752..4221625abf 100644
--- a/libmpcodecs/vf_pp7.c
+++ b/libmpcodecs/vf_pp7.c
@@ -157,7 +157,7 @@ static void dctB_c(DCTELEM *dst, DCTELEM *src){
 
 #ifdef HAVE_MMX
 static void dctB_mmx(DCTELEM *dst, DCTELEM *src){
-    asm volatile (
+    __asm__ volatile (
         "movq  (%0), %%mm0      \n\t"
         "movq  1*4*2(%0), %%mm1 \n\t"
         "paddw 6*4*2(%0), %%mm0 \n\t"
@@ -398,10 +398,10 @@ static int put_image(struct vf_instance* vf, mp_image_t *mpi, double pts){
     }
 
 #ifdef HAVE_MMX
-    if(gCpuCaps.hasMMX) asm volatile ("emms\n\t");
+    if(gCpuCaps.hasMMX) __asm__ volatile ("emms\n\t");
 #endif
 #ifdef HAVE_MMX2
-    if(gCpuCaps.hasMMX2) asm volatile ("sfence\n\t");
+    if(gCpuCaps.hasMMX2) __asm__ volatile ("sfence\n\t");
 #endif
 
     return vf_next_put_image(vf,dmpi, pts);
diff --git a/libmpcodecs/vf_qp.c b/libmpcodecs/vf_qp.c
index b08a0d5523..98aeab7535 100644
--- a/libmpcodecs/vf_qp.c
+++ b/libmpcodecs/vf_qp.c
@@ -33,8 +33,6 @@
 #include "vf.h"
 #include "libvo/fastmemcpy.h"
 
-// Needed to bring in lrintf.
-#define HAVE_AV_CONFIG_H
 #include "libavcodec/avcodec.h"
 #include "libavcodec/eval.h"
 
diff --git a/libmpcodecs/vf_sab.c b/libmpcodecs/vf_sab.c
index 1f7bdbccc1..61a5a7bc32 100644
--- a/libmpcodecs/vf_sab.c
+++ b/libmpcodecs/vf_sab.c
@@ -140,7 +140,7 @@ static int config(struct vf_instance* vf,
 	unsigned int flags, unsigned int outfmt){
 	
 	int sw, sh;
-//asm volatile("emms\n\t");
+//__asm__ volatile("emms\n\t");
 	allocStuff(&vf->priv->luma, width, height);
 	
 	getSubSampleFactors(&sw, &sh, outfmt);
diff --git a/libmpcodecs/vf_scale.c b/libmpcodecs/vf_scale.c
index 0daa23c605..02cd1ac919 100644
--- a/libmpcodecs/vf_scale.c
+++ b/libmpcodecs/vf_scale.c
@@ -557,7 +557,7 @@ void sws_getFlagsAndFilterFromCmdLine(int *flags, SwsFilter **srcFilterParam, Sw
 
 #ifdef ARCH_X86
 	if(gCpuCaps.hasMMX)
-		asm volatile("emms\n\t"::: "memory"); //FIXME this shouldnt be required but it IS (even for non mmx versions)
+		__asm__ volatile("emms\n\t"::: "memory"); //FIXME this should not be required but it IS (even for non-MMX versions)
 #endif
 	if(firstTime)
 	{
diff --git a/libmpcodecs/vf_spp.c b/libmpcodecs/vf_spp.c
index b94948d549..11a7cae1ff 100644
--- a/libmpcodecs/vf_spp.c
+++ b/libmpcodecs/vf_spp.c
@@ -153,7 +153,7 @@ static void hardthresh_mmx(DCTELEM dst[64], DCTELEM src[64], int qp, uint8_t *pe
 	
 	threshold1= qp*((1<<4) - bias) - 1;
 	
-        asm volatile(
+        __asm__ volatile(
 #define REQUANT_CORE(dst0, dst1, dst2, dst3, src0, src1, src2, src3) \
 		"movq " #src0 ", %%mm0	\n\t"\
 		"movq " #src1 ", %%mm1	\n\t"\
@@ -221,7 +221,7 @@ static void softthresh_mmx(DCTELEM dst[64], DCTELEM src[64], int qp, uint8_t *pe
 
 	threshold1= qp*((1<<4) - bias) - 1;
 	
-        asm volatile(
+        __asm__ volatile(
 #undef REQUANT_CORE
 #define REQUANT_CORE(