Merge remote-tracking branch 'qatar/master'

* qatar/master:
  qdm2: remove broken and disabled dump_context() debug function
  x86: h264_intrapred: use newly introduced SPLAT* and PSHUFLW macros
  x86inc: add SPLATB_LOAD, SPLATB_REG, PSHUFLW macros
  x86inc: modify ALIGN to not generate long nops on i586
  x86: h264_intrapred: port to cpuflag macros
  avplay: update input filter pointer when the filtergraph is reset.
  avconv: fix parsing of -force_key_frames option.
  h264: use templates to avoid excessive inlining
  xtea: Make the count parameter match the documentation
  blowfish: Make the count parameter match the documentation
  mpegvideo: Don't use ff_mspel_motion() for vc1
  xtea: invert branch and loop precedence
  blowfish: invert branch and loop precedence
  flvdec: optionally trust the metadata
  avconv: Set audio filter time base to the sample rate
  vp8: Add ifdef guards around the sse2 loopfilter in the sse2slow branch too

Conflicts:
	ffmpeg.c
	ffplay.c
	libavcodec/h264.c
	libavcodec/mpegvideo_common.h

Merged-by: Michael Niedermayer <michaelni@gmx.at>

4 files changed, 62 insertions, 18 deletions

diff --git a/libavutil/blowfish.c b/libavutil/blowfish.c
index 2ef037cf8e..fd894cdb80 100644
--- a/libavutil/blowfish.c
+++ b/libavutil/blowfish.c
@@ -381,8 +381,8 @@ void av_blowfish_crypt(AVBlowfish *ctx, uint8_t *dst, const uint8_t *src,
     uint32_t v0, v1;
     int i;
 
-    while (count > 0) {
-        if (decrypt) {
+    if (decrypt) {
+        while (count--) {
             v0 = AV_RB32(src);
             v1 = AV_RB32(src + 4);
 
@@ -396,7 +396,12 @@ void av_blowfish_crypt(AVBlowfish *ctx, uint8_t *dst, const uint8_t *src,
                     dst[i] = dst[i] ^ iv[i];
                 memcpy(iv, src, 8);
             }
-        } else {
+
+            src   += 8;
+            dst   += 8;
+        }
+    } else {
+        while (count--) {
             if (iv) {
                 for (i = 0; i < 8; i++)
                     dst[i] = src[i] ^ iv[i];
@@ -414,11 +419,10 @@ void av_blowfish_crypt(AVBlowfish *ctx, uint8_t *dst, const uint8_t *src,
 
             if (iv)
                 memcpy(iv, dst, 8);
-        }
 
-        src   += 8;
-        dst   += 8;
-        count -= 8;
+            src   += 8;
+            dst   += 8;
+        }
     }
 }
 
diff --git a/libavutil/x86/x86inc.asm b/libavutil/x86/x86inc.asm
index 971d210848..7d637a6353 100644
--- a/libavutil/x86/x86inc.asm
+++ b/libavutil/x86/x86inc.asm
@@ -604,6 +604,7 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
 ; All subsequent functions (up to the next INIT_CPUFLAGS) is built for the specified cpu.
 ; You shouldn't need to invoke this macro directly, it's a subroutine for INIT_MMX &co.
 %macro INIT_CPUFLAGS 0-2
+    CPU amdnop
     %if %0 >= 1
         %xdefine cpuname %1
         %assign cpuflags cpuflags_%1
@@ -625,6 +626,9 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
         %elifidn %1, sse3
             %define movu lddqu
         %endif
+        %if notcpuflag(mmx2)
+            CPU basicnop
+        %endif
     %else
         %xdefine SUFFIX
         %undef cpuname
diff --git a/libavutil/x86/x86util.asm b/libavutil/x86/x86util.asm
index 34959f2c94..c5c72e3d4b 100644
--- a/libavutil/x86/x86util.asm
+++ b/libavutil/x86/x86util.asm
@@ -256,15 +256,26 @@
 %define ABSB ABSB_MMX
 %define ABSB2 ABSB2_MMX
 
-%macro SPLATB_MMX 3
+%macro SPLATB_LOAD 3
+%if cpuflag(ssse3)
+    movd      %1, [%2-3]
+    pshufb    %1, %3
+%else
     movd      %1, [%2-3] ;to avoid crossing a cacheline
     punpcklbw %1, %1
     SPLATW    %1, %1, 3
+%endif
 %endmacro
 
-%macro SPLATB_SSSE3 3
-    movd      %1, [%2-3]
+%macro SPLATB_REG 3
+%if cpuflag(ssse3)
+    movd      %1, %2d
     pshufb    %1, %3
+%else
+    movd      %1, %2d
+    punpcklbw %1, %1
+    SPLATW    %1, %1, 0
+%endif
 %endmacro
 
 %macro PALIGNR_MMX 4-5 ; [dst,] src1, src2, imm, tmp
@@ -296,6 +307,14 @@
 %endif
 %endmacro
 
+%macro PSHUFLW 1+
+    %if mmsize == 8
+        pshufw %1
+    %else
+        pshuflw %1
+    %endif
+%endmacro
+
 %macro DEINTB 5 ; mask, reg1, mask, reg2, optional src to fill masks from
 %ifnum %5
     pand   m%3, m%5, m%4 ; src .. y6 .. y4
@@ -521,8 +540,22 @@
 %if mmsize == 16
     pshuflw    %1, %2, (%3)*0x55
     punpcklqdq %1, %1
-%else
+%elif cpuflag(mmx2)
     pshufw     %1, %2, (%3)*0x55
+%else
+    %ifnidn %1, %2
+        mova       %1, %2
+    %endif
+    %if %3 & 2
+        punpckhwd  %1, %1
+    %else
+        punpcklwd  %1, %1
+    %endif
+    %if %3 & 1
+        punpckhwd  %1, %1
+    %else
+        punpcklwd  %1, %1
+    %endif
 %endif
 %endmacro
 
diff --git a/libavutil/xtea.c b/libavutil/xtea.c
index 983cae90a5..0a5df1cec9 100644
--- a/libavutil/xtea.c
+++ b/libavutil/xtea.c
@@ -71,8 +71,8 @@ void av_xtea_crypt(AVXTEA *ctx, uint8_t *dst, const uint8_t *src, int count,
 {
     int i;
 
-    while (count > 0) {
-        if (decrypt) {
+    if (decrypt) {
+        while (count--) {
             xtea_crypt_ecb(ctx, dst, src, decrypt);
 
             if (iv) {
@@ -80,7 +80,12 @@ void av_xtea_crypt(AVXTEA *ctx, uint8_t *dst, const uint8_t *src, int count,
                     dst[i] = dst[i] ^ iv[i];
                 memcpy(iv, src, 8);
             }
-        } else {
+
+            src   += 8;
+            dst   += 8;
+        }
+    } else {
+        while (count--) {
             if (iv) {
                 for (i = 0; i < 8; i++)
                     dst[i] = src[i] ^ iv[i];
@@ -89,11 +94,9 @@ void av_xtea_crypt(AVXTEA *ctx, uint8_t *dst, const uint8_t *src, int count,
             } else {
                 xtea_crypt_ecb(ctx, dst, src, decrypt);
             }
+            src   += 8;
+            dst   += 8;
         }
-
-        src   += 8;
-        dst   += 8;
-        count -= 8;
     }
 }