x86/idctdsp: port {put,add}_pixels_clamped to yasm

Also add sse2 versions for both.
put_pixels_clamped port and sse2 version originally written by Timothy Gu.

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>

5 files changed, 112 insertions, 139 deletions

diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile
index 7bf0e82ac0..9f34abddde 100644
--- a/libavcodec/x86/Makefile
+++ b/libavcodec/x86/Makefile
@@ -66,8 +66,7 @@ OBJS-$(CONFIG_WEBP_DECODER)            += x86/vp8dsp_init.o
 # subsystems
 MMX-OBJS-$(CONFIG_DIRAC_DECODER)       += x86/dirac_dwt.o
 MMX-OBJS-$(CONFIG_FDCTDSP)             += x86/fdct.o
-MMX-OBJS-$(CONFIG_IDCTDSP)             += x86/idctdsp_mmx.o             \
-                                          x86/simple_idct.o
+MMX-OBJS-$(CONFIG_IDCTDSP)             += x86/simple_idct.o
 
 # decoders/encoders
 MMX-OBJS-$(CONFIG_MPEG4_DECODER)       += x86/xvididct_mmx.o            \
diff --git a/libavcodec/x86/idctdsp.asm b/libavcodec/x86/idctdsp.asm
index c3f37c4910..0aa73459e2 100644
--- a/libavcodec/x86/idctdsp.asm
+++ b/libavcodec/x86/idctdsp.asm
@@ -78,3 +78,106 @@ INIT_MMX mmx
 PUT_SIGNED_PIXELS_CLAMPED 0
 INIT_XMM sse2
 PUT_SIGNED_PIXELS_CLAMPED 3
+
+;--------------------------------------------------------------------------
+; void ff_put_pixels_clamped(const int16_t *block, uint8_t *pixels,
+;                            ptrdiff_t line_size);
+;--------------------------------------------------------------------------
+; %1 = block offset
+%macro PUT_PIXELS_CLAMPED_HALF 1
+    mova     m0, [blockq+mmsize*0+%1]
+    mova     m1, [blockq+mmsize*2+%1]
+%if mmsize == 8
+    mova     m2, [blockq+mmsize*4+%1]
+    mova     m3, [blockq+mmsize*6+%1]
+%endif
+    packuswb m0, [blockq+mmsize*1+%1]
+    packuswb m1, [blockq+mmsize*3+%1]
+%if mmsize == 8
+    packuswb m2, [blockq+mmsize*5+%1]
+    packuswb m3, [blockq+mmsize*7+%1]
+    movq           [pixelsq], m0
+    movq    [lsizeq+pixelsq], m1
+    movq  [2*lsizeq+pixelsq], m2
+    movq   [lsize3q+pixelsq], m3
+%else
+    movq           [pixelsq], m0
+    movhps  [lsizeq+pixelsq], m0
+    movq  [2*lsizeq+pixelsq], m1
+    movhps [lsize3q+pixelsq], m1
+%endif
+%endmacro
+
+%macro PUT_PIXELS_CLAMPED 0
+cglobal put_pixels_clamped, 3, 4, 2, block, pixels, lsize, lsize3
+    lea lsize3q, [lsizeq*3]
+    PUT_PIXELS_CLAMPED_HALF 0
+    lea pixelsq, [pixelsq+lsizeq*4]
+    PUT_PIXELS_CLAMPED_HALF 64
+    RET
+%endmacro
+
+INIT_MMX mmx
+PUT_PIXELS_CLAMPED
+INIT_XMM sse2
+PUT_PIXELS_CLAMPED
+
+;--------------------------------------------------------------------------
+; void ff_add_pixels_clamped(const int16_t *block, uint8_t *pixels,
+;                            ptrdiff_t line_size);
+;--------------------------------------------------------------------------
+; %1 = block offset
+%macro ADD_PIXELS_CLAMPED 1
+    mova       m0, [blockq+mmsize*0+%1]
+    mova       m1, [blockq+mmsize*1+%1]
+%if mmsize == 8
+    mova       m5, [blockq+mmsize*2+%1]
+    mova       m6, [blockq+mmsize*3+%1]
+%endif
+    movq       m2, [pixelsq]
+    movq       m3, [pixelsq+lsizeq]
+%if mmsize == 8
+    mova       m7, m2
+    punpcklbw  m2, m4
+    punpckhbw  m7, m4
+    paddsw     m0, m2
+    paddsw     m1, m7
+    mova       m7, m3
+    punpcklbw  m3, m4
+    punpckhbw  m7, m4
+    paddsw     m5, m3
+    paddsw     m6, m7
+%else
+    punpcklbw  m2, m4
+    punpcklbw  m3, m4
+    paddsw     m0, m2
+    paddsw     m1, m3
+%endif
+    packuswb   m0, m1
+%if mmsize == 8
+    packuswb   m5, m6
+    movq       [pixelsq], m0
+    movq       [pixelsq+lsizeq], m5
+%else
+    movq       [pixelsq], m0
+    movhps     [pixelsq+lsizeq], m0
+%endif
+%endmacro
+
+%macro ADD_PIXELS_CLAMPED 0
+cglobal add_pixels_clamped, 3, 3, 5, block, pixels, lsize
+    pxor       m4, m4
+    ADD_PIXELS_CLAMPED 0
+    lea        pixelsq, [pixelsq+lsizeq*2]
+    ADD_PIXELS_CLAMPED 32
+    lea        pixelsq, [pixelsq+lsizeq*2]
+    ADD_PIXELS_CLAMPED 64
+    lea        pixelsq, [pixelsq+lsizeq*2]
+    ADD_PIXELS_CLAMPED 96
+    RET
+%endmacro
+
+INIT_MMX mmx
+ADD_PIXELS_CLAMPED
+INIT_XMM sse2
+ADD_PIXELS_CLAMPED
diff --git a/libavcodec/x86/idctdsp.h b/libavcodec/x86/idctdsp.h
index 252e704338..6408df1e32 100644
--- a/libavcodec/x86/idctdsp.h
+++ b/libavcodec/x86/idctdsp.h
@@ -23,8 +23,12 @@
 
 void ff_add_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels,
                                ptrdiff_t line_size);
+void ff_add_pixels_clamped_sse2(const int16_t *block, uint8_t *pixels,
+                                ptrdiff_t line_size);
 void ff_put_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels,
                                ptrdiff_t line_size);
+void ff_put_pixels_clamped_sse2(const int16_t *block, uint8_t *pixels,
+                                ptrdiff_t line_size);
 void ff_put_signed_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels,
                                       ptrdiff_t line_size);
 void ff_put_signed_pixels_clamped_sse2(const int16_t *block, uint8_t *pixels,
diff --git a/libavcodec/x86/idctdsp_init.c b/libavcodec/x86/idctdsp_init.c
index 3f438f44db..2c26a98850 100644
--- a/libavcodec/x86/idctdsp_init.c
+++ b/libavcodec/x86/idctdsp_init.c
@@ -64,9 +64,6 @@ av_cold void ff_idctdsp_init_x86(IDCTDSPContext *c, AVCodecContext *avctx,
     int cpu_flags = av_get_cpu_flags();
 
     if (INLINE_MMX(cpu_flags)) {
-        c->put_pixels_clamped        = ff_put_pixels_clamped_mmx;
-        c->add_pixels_clamped        = ff_add_pixels_clamped_mmx;
-
         if (!high_bit_depth &&
             avctx->lowres == 0 &&
             (avctx->idct_algo == FF_IDCT_AUTO ||
@@ -80,8 +77,12 @@ av_cold void ff_idctdsp_init_x86(IDCTDSPContext *c, AVCodecContext *avctx,
     }
     if (EXTERNAL_MMX(cpu_flags)) {
         c->put_signed_pixels_clamped = ff_put_signed_pixels_clamped_mmx;
+        c->put_pixels_clamped        = ff_put_pixels_clamped_mmx;
+        c->add_pixels_clamped        = ff_add_pixels_clamped_mmx;
     }
     if (EXTERNAL_SSE2(cpu_flags)) {
         c->put_signed_pixels_clamped = ff_put_signed_pixels_clamped_sse2;
+        c->put_pixels_clamped        = ff_put_pixels_clamped_sse2;
+        c->add_pixels_clamped        = ff_add_pixels_clamped_sse2;
     }
 }
diff --git a/libavcodec/x86/idctdsp_mmx.c b/libavcodec/x86/idctdsp_mmx.c
deleted file mode 100644
index 0ce9a92390..0000000000
--- a/libavcodec/x86/idctdsp_mmx.c
+++ /dev/null
@@ -1,134 +0,0 @@
-/*
- * SIMD-optimized IDCT-related routines
- * Copyright (c) 2000, 2001 Fabrice Bellard
- * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
- *
- * MMX optimization by Nick Kurshev <nickols_k@mail.ru>
- *
- * This file is part of FFmpeg.
- *
- * FFmpeg is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * FFmpeg is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with FFmpeg; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#include "config.h"
-#include "libavutil/cpu.h"
-#include "libavutil/x86/asm.h"
-#include "libavcodec/idctdsp.h"
-#include "idctdsp.h"
-#include "inline_asm.h"
-
-#if HAVE_INLINE_ASM
-
-void ff_put_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels,
-                               ptrdiff_t line_size)
-{
-    const int16_t *p;
-    uint8_t *pix;
-
-    /* read the pixels */
-    p   = block;
-    pix = pixels;
-    /* unrolled loop */
-    __asm__ volatile (
-        "movq      (%3), %%mm0          \n\t"
-        "movq     8(%3), %%mm1          \n\t"
-        "movq    16(%3), %%mm2          \n\t"
-        "movq    24(%3), %%mm3          \n\t"
-        "movq    32(%3), %%mm4          \n\t"
-        "movq    40(%3), %%mm5          \n\t"
-        "movq    48(%3), %%mm6          \n\t"
-        "movq    56(%3), %%mm7          \n\t"
-        "packuswb %%mm1, %%mm0          \n\t"
-        "packuswb %%mm3, %%mm2          \n\t"
-        "packuswb %%mm5, %%mm4          \n\t"
-        "packuswb %%mm7, %%mm6          \n\t"
-        "movq     %%mm0, (%0)           \n\t"
-        "movq     %%mm2, (%0, %1)       \n\t"
-        "movq     %%mm4, (%0, %1, 2)    \n\t"
-        "movq     %%mm6, (%0, %2)       \n\t"
-        :: "r" (pix), "r" ((x86_reg) line_size), "r" ((x86_reg) line_size * 3),
-           "r" (p)
-        : "memory");
-    pix += line_size * 4;
-    p   += 32;
-
-    // if here would be an exact copy of the code above
-    // compiler would generate some very strange code
-    // thus using "r"
-    __asm__ volatile (
-        "movq       (%3), %%mm0         \n\t"
-        "movq      8(%3), %%mm1         \n\t"
-        "movq     16(%3), %%mm2         \n\t"
-        "movq     24(%3), %%mm3         \n\t"
-        "movq     32(%3), %%mm4         \n\t"
-        "movq     40(%3), %%mm5         \n\t"
-        "movq     48(%3), %%mm6         \n\t"
-        "movq     56(%3), %%mm7         \n\t"
-        "packuswb  %%mm1, %%mm0         \n\t"
-        "packuswb  %%mm3, %%mm2         \n\t"
-        "packuswb  %%mm5, %%mm4         \n\t"
-        "packuswb  %%mm7, %%mm6         \n\t"
-        "movq      %%mm0, (%0)          \n\t"
-        "movq      %%mm2, (%0, %1)      \n\t"
-        "movq      %%mm4, (%0, %1, 2)   \n\t"
-        "movq      %%mm6, (%0, %2)      \n\t"
-        :: "r" (pix), "r" ((x86_reg) line_size), "r" ((x86_reg) line_size * 3),
-           "r" (p)
-        : "memory");
-}
-
-void ff_add_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels,
-                               ptrdiff_t line_size)
-{
-    const int16_t *p;
-    uint8_t *pix;
-    int i;
-
-    /* read the pixels */
-    p   = block;
-    pix = pixels;
-    MOVQ_ZERO(mm7);
-    i = 4;
-    do {
-        __asm__ volatile (
-            "movq        (%2), %%mm0    \n\t"
-            "movq       8(%2), %%mm1    \n\t"
-            "movq      16(%2), %%mm2    \n\t"
-            "movq      24(%2), %%mm3    \n\t"
-            "movq          %0, %%mm4    \n\t"
-            "movq          %1, %%mm6    \n\t"
-            "movq       %%mm4, %%mm5    \n\t"
-            "punpcklbw  %%mm7, %%mm4    \n\t"
-            "punpckhbw  %%mm7, %%mm5    \n\t"
-            "paddsw     %%mm4, %%mm0    \n\t"
-            "paddsw     %%mm5, %%mm1    \n\t"
-            "movq       %%mm6, %%mm5    \n\t"
-            "punpcklbw  %%mm7, %%mm6    \n\t"
-            "punpckhbw  %%mm7, %%mm5    \n\t"
-            "paddsw     %%mm6, %%mm2    \n\t"
-            "paddsw     %%mm5, %%mm3    \n\t"
-            "packuswb   %%mm1, %%mm0    \n\t"
-            "packuswb   %%mm3, %%mm2    \n\t"
-            "movq       %%mm0, %0       \n\t"
-            "movq       %%mm2, %1       \n\t"
-            : "+m" (*pix), "+m" (*(pix + line_size))
-            : "r" (p)
-            : "memory");
-        pix += line_size * 2;
-        p   += 16;
-    } while (--i);
-}
-
-#endif /* HAVE_INLINE_ASM */