Move vorbis_inverse_coupling from dsputil to vorbisdspcontext.

Conveniently (together with Justin's earlier patches), this makes
our vorbis decoder entirely independent of dsputil.

5 files changed, 122 insertions, 69 deletions

diff --git a/libavcodec/arm/Makefile b/libavcodec/arm/Makefile
index ac486f47dc..71048f9c4c 100644
--- a/libavcodec/arm/Makefile
+++ b/libavcodec/arm/Makefile
@@ -15,6 +15,7 @@ OBJS-$(CONFIG_MPEGAUDIODSP)            += arm/mpegaudiodsp_init_arm.o
 ARMV6-OBJS-$(CONFIG_MPEGAUDIODSP)      += arm/mpegaudiodsp_fixed_armv6.o
 
 OBJS-$(CONFIG_MPEGVIDEO)               += arm/mpegvideo_arm.o
+OBJS-$(CONFIG_VORBIS_DECODER)          += arm/vorbisdsp_init_arm.o
 OBJS-$(CONFIG_VP3DSP)                  += arm/vp3dsp_init_arm.o
 OBJS-$(CONFIG_VP5_DECODER)             += arm/vp56dsp_init_arm.o
 OBJS-$(CONFIG_VP6_DECODER)             += arm/vp56dsp_init_arm.o
@@ -86,6 +87,8 @@ NEON-OBJS-$(CONFIG_RV40_DECODER)       += arm/rv34dsp_neon.o            \
                                           arm/rv40dsp_neon.o            \
                                           arm/h264cmc_neon.o            \
 
+NEON-OBJS-$(CONFIG_VORBIS_DECODER)     += arm/vorbisdsp_neon.o
+
 NEON-OBJS-$(CONFIG_VP3DSP)             += arm/vp3dsp_neon.o
 
 NEON-OBJS-$(CONFIG_VP5_DECODER)        += arm/vp56dsp_neon.o            \
diff --git a/libavcodec/arm/dsputil_init_neon.c b/libavcodec/arm/dsputil_init_neon.c
index 34bb6191f1..ee0e9afa88 100644
--- a/libavcodec/arm/dsputil_init_neon.c
+++ b/libavcodec/arm/dsputil_init_neon.c
@@ -154,8 +154,6 @@ void ff_vector_clipf_neon(float *dst, const float *src, float min, float max,
 void ff_vector_clip_int32_neon(int32_t *dst, const int32_t *src, int32_t min,
                                int32_t max, unsigned int len);
 
-void ff_vorbis_inverse_coupling_neon(float *mag, float *ang, int blocksize);
-
 int32_t ff_scalarproduct_int16_neon(const int16_t *v1, const int16_t *v2, int len);
 int32_t ff_scalarproduct_and_madd_int16_neon(int16_t *v1, const int16_t *v2,
                                              const int16_t *v3, int len, int mul);
@@ -307,9 +305,6 @@ void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx)
     c->vector_clipf               = ff_vector_clipf_neon;
     c->vector_clip_int32          = ff_vector_clip_int32_neon;
 
-    if (CONFIG_VORBIS_DECODER)
-        c->vorbis_inverse_coupling = ff_vorbis_inverse_coupling_neon;
-
     c->scalarproduct_int16 = ff_scalarproduct_int16_neon;
     c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_neon;
 
diff --git a/libavcodec/arm/dsputil_neon.S b/libavcodec/arm/dsputil_neon.S
index a0d201cd85..ebc70ac109 100644
--- a/libavcodec/arm/dsputil_neon.S
+++ b/libavcodec/arm/dsputil_neon.S
@@ -19,7 +19,6 @@
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
-#include "config.h"
 #include "libavutil/arm/asm.S"
 
 function ff_clear_block_neon, export=1
@@ -532,69 +531,6 @@ function ff_add_pixels_clamped_neon, export=1
         bx              lr
 endfunc
 
-#if CONFIG_VORBIS_DECODER
-function ff_vorbis_inverse_coupling_neon, export=1
-        vmov.i32        q10, #1<<31
-        subs            r2,  r2,  #4
-        mov             r3,  r0
-        mov             r12, r1
-        beq             3f
-
-        vld1.32         {d24-d25},[r1,:128]!
-        vld1.32         {d22-d23},[r0,:128]!
-        vcle.s32        q8,  q12, #0
-        vand            q9,  q11, q10
-        veor            q12, q12, q9
-        vand            q2,  q12, q8
-        vbic            q3,  q12, q8
-        vadd.f32        q12, q11, q2
-        vsub.f32        q11, q11, q3
-1:      vld1.32         {d2-d3},  [r1,:128]!
-        vld1.32         {d0-d1},  [r0,:128]!
-        vcle.s32        q8,  q1,  #0
-        vand            q9,  q0,  q10
-        veor            q1,  q1,  q9
-        vst1.32         {d24-d25},[r3, :128]!
-        vst1.32         {d22-d23},[r12,:128]!
-        vand            q2,  q1,  q8
-        vbic            q3,  q1,  q8
-        vadd.f32        q1,  q0,  q2
-        vsub.f32        q0,  q0,  q3
-        subs            r2,  r2,  #8
-        ble             2f
-        vld1.32         {d24-d25},[r1,:128]!
-        vld1.32         {d22-d23},[r0,:128]!
-        vcle.s32        q8,  q12, #0
-        vand            q9,  q11, q10
-        veor            q12, q12, q9
-        vst1.32         {d2-d3},  [r3, :128]!
-        vst1.32         {d0-d1},  [r12,:128]!
-        vand            q2,  q12, q8
-        vbic            q3,  q12, q8
-        vadd.f32        q12, q11, q2
-        vsub.f32        q11, q11, q3
-        b               1b
-
-2:      vst1.32         {d2-d3},  [r3, :128]!
-        vst1.32         {d0-d1},  [r12,:128]!
-        it              lt
-        bxlt            lr
-
-3:      vld1.32         {d2-d3},  [r1,:128]
-        vld1.32         {d0-d1},  [r0,:128]
-        vcle.s32        q8,  q1,  #0
-        vand            q9,  q0,  q10
-        veor            q1,  q1,  q9
-        vand            q2,  q1,  q8
-        vbic            q3,  q1,  q8
-        vadd.f32        q1,  q0,  q2
-        vsub.f32        q0,  q0,  q3
-        vst1.32         {d2-d3},  [r0,:128]!
-        vst1.32         {d0-d1},  [r1,:128]!
-        bx              lr
-endfunc
-#endif
-
 function ff_butterflies_float_neon, export=1
 1:      vld1.32         {q0},[r0,:128]
         vld1.32         {q1},[r1,:128]
diff --git a/libavcodec/arm/vorbisdsp_init_arm.c b/libavcodec/arm/vorbisdsp_init_arm.c
new file mode 100644
index 0000000000..ec8fbd5bc9
--- /dev/null
+++ b/libavcodec/arm/vorbisdsp_init_arm.c
@@ -0,0 +1,36 @@
+/*
+ * ARM NEON optimised DSP functions
+ * Copyright (c) 2008 Mans Rullgard <mans@mansr.com>
+ *
+ * This file is part of Libav.
+ *
+ * Libav 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.
+ *
+ * Libav 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 Libav; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/attributes.h"
+#include "libavutil/cpu.h"
+#include "libavutil/arm/cpu.h"
+#include "libavcodec/vorbisdsp.h"
+
+void ff_vorbis_inverse_coupling_neon(float *mag, float *ang, int blocksize);
+
+void ff_vorbisdsp_init_arm(VorbisDSPContext *c)
+{
+    int cpu_flags = av_get_cpu_flags();
+
+    if (have_neon(cpu_flags)) {
+        c->vorbis_inverse_coupling = ff_vorbis_inverse_coupling_neon;
+    }
+}
diff --git a/libavcodec/arm/vorbisdsp_neon.S b/libavcodec/arm/vorbisdsp_neon.S
new file mode 100644
index 0000000000..7df876c2bc
--- /dev/null
+++ b/libavcodec/arm/vorbisdsp_neon.S
@@ -0,0 +1,83 @@
+/*
+ * ARM NEON optimised DSP functions
+ * Copyright (c) 2008 Mans Rullgard <mans@mansr.com>
+ *
+ * This file is part of Libav.
+ *
+ * Libav 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.
+ *
+ * Libav 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 Libav; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/arm/asm.S"
+
+function ff_vorbis_inverse_coupling_neon, export=1
+        vmov.i32        q10, #1<<31
+        subs            r2,  r2,  #4
+        mov             r3,  r0
+        mov             r12, r1
+        beq             3f
+
+        vld1.32         {d24-d25},[r1,:128]!
+        vld1.32         {d22-d23},[r0,:128]!
+        vcle.s32        q8,  q12, #0
+        vand            q9,  q11, q10
+        veor            q12, q12, q9
+        vand            q2,  q12, q8
+        vbic            q3,  q12, q8
+        vadd.f32        q12, q11, q2
+        vsub.f32        q11, q11, q3
+1:      vld1.32         {d2-d3},  [r1,:128]!
+        vld1.32         {d0-d1},  [r0,:128]!
+        vcle.s32        q8,  q1,  #0
+        vand            q9,  q0,  q10
+        veor            q1,  q1,  q9
+        vst1.32         {d24-d25},[r3, :128]!
+        vst1.32         {d22-d23},[r12,:128]!
+        vand            q2,  q1,  q8
+        vbic            q3,  q1,  q8
+        vadd.f32        q1,  q0,  q2
+        vsub.f32        q0,  q0,  q3
+        subs            r2,  r2,  #8
+        ble             2f
+        vld1.32         {d24-d25},[r1,:128]!
+        vld1.32         {d22-d23},[r0,:128]!
+        vcle.s32        q8,  q12, #0
+        vand            q9,  q11, q10
+        veor            q12, q12, q9
+        vst1.32         {d2-d3},  [r3, :128]!
+        vst1.32         {d0-d1},  [r12,:128]!
+        vand            q2,  q12, q8
+        vbic            q3,  q12, q8
+        vadd.f32        q12, q11, q2
+        vsub.f32        q11, q11, q3
+        b               1b
+
+2:      vst1.32         {d2-d3},  [r3, :128]!
+        vst1.32         {d0-d1},  [r12,:128]!
+        it              lt
+        bxlt            lr
+
+3:      vld1.32         {d2-d3},  [r1,:128]
+        vld1.32         {d0-d1},  [r0,:128]
+        vcle.s32        q8,  q1,  #0
+        vand            q9,  q0,  q10
+        veor            q1,  q1,  q9
+        vand            q2,  q1,  q8
+        vbic            q3,  q1,  q8
+        vadd.f32        q1,  q0,  q2
+        vsub.f32        q0,  q0,  q3
+        vst1.32         {d2-d3},  [r0,:128]!
+        vst1.32         {d0-d1},  [r1,:128]!
+        bx              lr
+endfunc