Merge branch 'master' of ssh://git.freedesktop.org/git/mesa/mesa into pipe-video

Conflicts:
	src/gallium/include/pipe/p_format.h

9 files changed, 320 insertions, 6 deletions

diff --git a/src/gallium/auxiliary/util/u_dl.c b/src/gallium/auxiliary/util/u_dl.c
index 220860ebf4b..aca435d6cad 100644
--- a/src/gallium/auxiliary/util/u_dl.c
+++ b/src/gallium/auxiliary/util/u_dl.c
@@ -38,6 +38,7 @@
 #endif
 
 #include "u_dl.h"
+#include "u_pointer.h"
 
 
 struct util_dl_library *
@@ -58,7 +59,7 @@ util_dl_get_proc_address(struct util_dl_library *library,
                          const char *procname)
 {
 #if defined(PIPE_OS_UNIX)
-   return (util_dl_proc)dlsym((void *)library, procname);
+   return (util_dl_proc) pointer_to_func(dlsym((void *)library, procname));
 #elif defined(PIPE_OS_WINDOWS)
    return (util_dl_proc)GetProcAddress((HMODULE)library, procname);
 #else
diff --git a/src/gallium/auxiliary/util/u_format.csv b/src/gallium/auxiliary/util/u_format.csv
index 0811280b97b..8e5d4487a67 100644
--- a/src/gallium/auxiliary/util/u_format.csv
+++ b/src/gallium/auxiliary/util/u_format.csv
@@ -109,9 +109,12 @@ PIPE_FORMAT_Z32_UNORM               , plain, 1, 1, un32,     ,     ,     , x___,
 PIPE_FORMAT_Z32_FLOAT               , plain, 1, 1, f32 ,     ,     ,     , x___, zs
 PIPE_FORMAT_Z24_UNORM_S8_USCALED    , plain, 1, 1, un24, u8  ,     ,     , xy__, zs
 PIPE_FORMAT_S8_USCALED_Z24_UNORM    , plain, 1, 1, u8 ,  un24,     ,     , yx__, zs
+PIPE_FORMAT_X24S8_USCALED           , plain, 1, 1, x24,  u8  ,     ,     , _y__, zs
+PIPE_FORMAT_S8X24_USCALED           , plain, 1, 1, u8  , x24 ,     ,     , _x__, zs
 PIPE_FORMAT_Z24X8_UNORM             , plain, 1, 1, un24, x8  ,     ,     , x___, zs
 PIPE_FORMAT_X8Z24_UNORM             , plain, 1, 1, x8  , un24,     ,     , y___, zs
 PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED , plain, 1, 1, f32,  u8  , x24 ,     , xy__, zs
+PIPE_FORMAT_X32_S8X24_USCALED       , plain, 1, 1, x32,  u8  , x24 ,     , _y__, zs
 
 # YUV formats
 # http://www.fourcc.org/yuv.php#UYVY
diff --git a/src/gallium/auxiliary/util/u_format_srgb.py b/src/gallium/auxiliary/util/u_format_srgb.py
index 3e8000f3687..cd63ae78919 100644
--- a/src/gallium/auxiliary/util/u_format_srgb.py
+++ b/src/gallium/auxiliary/util/u_format_srgb.py
@@ -1,6 +1,6 @@
 #!/usr/bin/env python
 
-'''
+CopyRight = '''
 /**************************************************************************
  *
  * Copyright 2010 VMware, Inc.
@@ -89,7 +89,7 @@ def main():
     print '/* This file is autogenerated by u_format_srgb.py. Do not edit directly. */'
     print
     # This will print the copyright message on the top of this file
-    print __doc__.strip()
+    print CopyRight.strip()
     print
     print '#include "u_format_srgb.h"'
     print
diff --git a/src/gallium/auxiliary/util/u_format_table.py b/src/gallium/auxiliary/util/u_format_table.py
index f0b407b8b8e..8cc22a56371 100755
--- a/src/gallium/auxiliary/util/u_format_table.py
+++ b/src/gallium/auxiliary/util/u_format_table.py
@@ -1,6 +1,6 @@
 #!/usr/bin/env python
 
-'''
+CopyRight = '''
 /**************************************************************************
  *
  * Copyright 2010 VMware, Inc.
@@ -83,7 +83,7 @@ def write_format_table(formats):
     print '/* This file is autogenerated by u_format_table.py from u_format.csv. Do not edit directly. */'
     print
     # This will print the copyright message on the top of this file
-    print __doc__.strip()
+    print CopyRight.strip()
     print
     print '#include "u_format.h"'
     print '#include "u_format_s3tc.h"'
diff --git a/src/gallium/auxiliary/util/u_format_zs.c b/src/gallium/auxiliary/util/u_format_zs.c
index 792d69c214c..80081e22f7c 100644
--- a/src/gallium/auxiliary/util/u_format_zs.c
+++ b/src/gallium/auxiliary/util/u_format_zs.c
@@ -918,3 +918,56 @@ util_format_z32_float_s8x24_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned d
    }
 }
 
+
+void
+util_format_x24s8_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+   util_format_z24_unorm_s8_uscaled_unpack_s_8uscaled(dst_row, dst_stride,
+						      src_row, src_stride,
+						      width, height);
+}
+
+void
+util_format_x24s8_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+   util_format_z24_unorm_s8_uscaled_pack_s_8uscaled(dst_row, dst_stride,
+						    src_row, src_stride,
+						    width, height);
+}
+
+void
+util_format_s8x24_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+   util_format_s8_uscaled_z24_unorm_unpack_s_8uscaled(dst_row, dst_stride,
+						      src_row, src_stride,
+						      width, height);
+}
+
+void
+util_format_s8x24_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+   util_format_s8_uscaled_z24_unorm_pack_s_8uscaled(dst_row, dst_stride,
+						      src_row, src_stride,
+						      width, height);
+}
+
+void
+util_format_x32_s8x24_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+						const uint8_t *src_row, unsigned src_stride,
+						unsigned width, unsigned height)
+{
+   util_format_z32_float_s8x24_uscaled_unpack_s_8uscaled(dst_row, dst_stride,
+							 src_row, src_stride,
+							 width, height);
+
+}
+
+void
+util_format_x32_s8x24_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+					      const uint8_t *src_row, unsigned src_stride,
+					      unsigned width, unsigned height)
+{
+   util_format_z32_float_s8x24_uscaled_pack_s_8uscaled(dst_row, dst_stride,
+                                                       src_row, src_stride,
+						       width, height);
+}
diff --git a/src/gallium/auxiliary/util/u_format_zs.h b/src/gallium/auxiliary/util/u_format_zs.h
index 650db4b95fd..1604cc3eee2 100644
--- a/src/gallium/auxiliary/util/u_format_zs.h
+++ b/src/gallium/auxiliary/util/u_format_zs.h
@@ -192,5 +192,21 @@ util_format_z32_float_s8x24_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned
 void
 util_format_z32_float_s8x24_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
+void
+util_format_x24s8_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+void
+util_format_x24s8_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+void
+util_format_s8x24_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
+void
+util_format_s8x24_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+void
+util_format_x32_s8x24_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+void
+util_format_x32_s8x24_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_sride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 #endif /* U_FORMAT_ZS_H_ */
diff --git a/src/gallium/auxiliary/util/u_math.h b/src/gallium/auxiliary/util/u_math.h
index 69a76814945..37294b7203f 100644
--- a/src/gallium/auxiliary/util/u_math.h
+++ b/src/gallium/auxiliary/util/u_math.h
@@ -118,6 +118,11 @@ __inline double __cdecl atan2(double val)
 #endif
 
 
+#ifndef M_SQRT2
+#define M_SQRT2 1.41421356237309504880
+#endif
+
+
 #if defined(_MSC_VER) 
 
 #if _MSC_VER < 1400 && !defined(__cplusplus) || defined(PIPE_SUBSYSTEM_WINDOWS_CE)
diff --git a/src/gallium/auxiliary/util/u_sse.h b/src/gallium/auxiliary/util/u_sse.h
index 03198c91da4..1df6c872677 100644
--- a/src/gallium/auxiliary/util/u_sse.h
+++ b/src/gallium/auxiliary/util/u_sse.h
@@ -71,6 +71,96 @@ _mm_castps_si128(__m128 a)
 
 #endif /* defined(_MSC_VER) && _MSC_VER < 1500 */
 
+union m128i {
+   __m128i m;
+   ubyte ub[16];
+   ushort us[8];
+   uint ui[4];
+};
+
+static INLINE void u_print_epi8(const char *name, __m128i r)
+{
+   union { __m128i m; ubyte ub[16]; } u;
+   u.m = r;
+
+   debug_printf("%s: "
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x/"
+                "%02x\n",
+                name,
+                u.ub[0],  u.ub[1],  u.ub[2],  u.ub[3],
+                u.ub[4],  u.ub[5],  u.ub[6],  u.ub[7],
+                u.ub[8],  u.ub[9],  u.ub[10], u.ub[11],
+                u.ub[12], u.ub[13], u.ub[14], u.ub[15]);
+}
+
+static INLINE void u_print_epi16(const char *name, __m128i r)
+{
+   union { __m128i m; ushort us[8]; } u;
+   u.m = r;
+
+   debug_printf("%s: "
+                "%04x/"
+                "%04x/"
+                "%04x/"
+                "%04x/"
+                "%04x/"
+                "%04x/"
+                "%04x/"
+                "%04x\n",
+                name,
+                u.us[0],  u.us[1],  u.us[2],  u.us[3],
+                u.us[4],  u.us[5],  u.us[6],  u.us[7]);
+}
+
+static INLINE void u_print_epi32(const char *name, __m128i r)
+{
+   union { __m128i m; uint ui[4]; } u;
+   u.m = r;
+
+   debug_printf("%s: "
+                "%08x/"
+                "%08x/"
+                "%08x/"
+                "%08x\n",
+                name,
+                u.ui[0],  u.ui[1],  u.ui[2],  u.ui[3]);
+}
+
+static INLINE void u_print_ps(const char *name, __m128 r)
+{
+   union { __m128 m; float f[4]; } u;
+   u.m = r;
+
+   debug_printf("%s: "
+                "%f/"
+                "%f/"
+                "%f/"
+                "%f\n",
+                name,
+                u.f[0],  u.f[1],  u.f[2],  u.f[3]);
+}
+
+
+#define U_DUMP_EPI32(a) u_print_epi32(#a, a)
+#define U_DUMP_EPI16(a) u_print_epi16(#a, a)
+#define U_DUMP_EPI8(a)  u_print_epi8(#a, a)
+#define U_DUMP_PS(a)    u_print_ps(#a, a)
+
+
 
 #if defined(PIPE_ARCH_SSSE3)
 
@@ -98,6 +188,68 @@ _mm_shuffle_epi8(__m128i a, __m128i mask)
 #endif /* !PIPE_ARCH_SSSE3 */
 
 
-#endif /* PIPE_ARCH_X86 || PIPE_ARCH_X86_64 */
+
+
+/* Provide an SSE2 implementation of _mm_mullo_epi32() in terms of
+ * _mm_mul_epu32().
+ *
+ * I suspect this works fine for us because one of our operands is
+ * always positive, but not sure that this can be used for general
+ * signed integer multiplication.
+ *
+ * This seems close enough to the speed of SSE4 and the real
+ * _mm_mullo_epi32() intrinsic as to not justify adding an sse4
+ * dependency at this point.
+ */
+static INLINE __m128i mm_mullo_epi32(const __m128i a, const __m128i b)
+{
+   __m128i a4   = _mm_srli_epi64(a, 32);  /* shift by one dword */
+   __m128i b4   = _mm_srli_epi64(b, 32);  /* shift by one dword */
+   __m128i ba   = _mm_mul_epu32(b, a);   /* multply dwords 0, 2 */
+   __m128i b4a4 = _mm_mul_epu32(b4, a4); /* multiply dwords 1, 3 */
+
+   /* Interleave the results, either with shuffles or (slightly
+    * faster) direct bit operations:
+    */
+#if 0
+   __m128i ba8             = _mm_shuffle_epi32(ba, 8);
+   __m128i b4a48           = _mm_shuffle_epi32(b4a4, 8);
+   __m128i result          = _mm_unpacklo_epi32(ba8, b4a48);
+#else
+   __m128i mask            = _mm_setr_epi32(~0,0,~0,0);
+   __m128i ba_mask         = _mm_and_si128(ba, mask);
+   __m128i b4a4_mask_shift = _mm_slli_epi64(b4a4, 32);
+   __m128i result          = _mm_or_si128(ba_mask, b4a4_mask_shift);
+#endif
+
+   return result;
+}
+
+
+static INLINE void
+transpose4_epi32(const __m128i * restrict a,
+                 const __m128i * restrict b,
+                 const __m128i * restrict c,
+                 const __m128i * restrict d,
+                 __m128i * restrict o,
+                 __m128i * restrict p,
+                 __m128i * restrict q,
+                 __m128i * restrict r)
+{
+  __m128i t0 = _mm_unpacklo_epi32(*a, *b);
+  __m128i t1 = _mm_unpacklo_epi32(*c, *d);
+  __m128i t2 = _mm_unpackhi_epi32(*a, *b);
+  __m128i t3 = _mm_unpackhi_epi32(*c, *d);
+
+  *o = _mm_unpacklo_epi64(t0, t1);
+  *p = _mm_unpackhi_epi64(t0, t1);
+  *q = _mm_unpacklo_epi64(t2, t3);
+  *r = _mm_unpackhi_epi64(t2, t3);
+}
+
+#define SCALAR_EPI32(m, i) _mm_shuffle_epi32((m), _MM_SHUFFLE(i,i,i,i))
+
+
+#endif /* PIPE_ARCH_SSE */
 
 #endif /* U_SSE_H_ */
diff --git a/src/gallium/auxiliary/util/u_tile.c b/src/gallium/auxiliary/util/u_tile.c
index f7aa1403d08..44cadbfcdd0 100644
--- a/src/gallium/auxiliary/util/u_tile.c
+++ b/src/gallium/auxiliary/util/u_tile.c
@@ -217,6 +217,81 @@ z24s8_get_tile_rgba(const unsigned *src,
    }
 }
 
+/*** PIPE_FORMAT_S8X24_USCALED ***/
+
+/**
+ * Return S component as four uint32_t in [0..255].  Z part ignored.
+ */
+static void
+s8x24_get_tile_rgba(const unsigned *src,
+                    unsigned w, unsigned h,
+                    float *p,
+                    unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (float)((*src++ >> 24) & 0xff);
+      }
+
+      p += dst_stride;
+   }
+}
+
+/*** PIPE_FORMAT_X24S8_USCALED ***/
+
+/**
+ * Return S component as four uint32_t in [0..255].  Z part ignored.
+ */
+static void
+x24s8_get_tile_rgba(const unsigned *src,
+                    unsigned w, unsigned h,
+                    float *p,
+                    unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (float)(*src++ & 0xff);
+      }
+      p += dst_stride;
+   }
+}
+
+
+/**
+ * Return S component as four uint32_t in [0..255].  Z part ignored.
+ */
+static void
+s8_get_tile_rgba(const unsigned char *src,
+		 unsigned w, unsigned h,
+		 float *p,
+		 unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (float)(*src++ & 0xff);
+      }
+      p += dst_stride;
+   }
+}
 
 /*** PIPE_FORMAT_Z32_FLOAT ***/
 
@@ -261,10 +336,19 @@ pipe_tile_raw_to_rgba(enum pipe_format format,
    case PIPE_FORMAT_Z24X8_UNORM:
       s8z24_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
       break;
+   case PIPE_FORMAT_S8_USCALED:
+      s8_get_tile_rgba((unsigned char *) src, w, h, dst, dst_stride);
+      break;
+   case PIPE_FORMAT_X24S8_USCALED:
+      s8x24_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
+      break;
    case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
    case PIPE_FORMAT_X8Z24_UNORM:
       z24s8_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
       break;
+   case PIPE_FORMAT_S8X24_USCALED:
+      x24s8_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
+      break;
    case PIPE_FORMAT_Z32_FLOAT:
       z32f_get_tile_rgba((float *) src, w, h, dst, dst_stride);
       break;