gallium: Use util_cpu_to_le{16,32} in many more places.

... and util_le{16,32}_to_cpu. I think I've used the right ones for
describing the actual operation performed (even though they're both just
"byte-swap this if I'm on big-endian").

The Linux Kernel has typedefs __le32/__be32 and friends that static
analysis tools can use to check that byte-orderings are correct. It
might be interesting to apply that here as well.

Reviewed-by: Eric Anholt <[email protected]>
Reviewed-by: Jose Fonseca <[email protected]>

3 files changed, 79 insertions, 353 deletions

diff --git a/src/gallium/auxiliary/util/u_format_other.c b/src/gallium/auxiliary/util/u_format_other.c
index 85001c14f22..20752d469a0 100644
--- a/src/gallium/auxiliary/util/u_format_other.c
+++ b/src/gallium/auxiliary/util/u_format_other.c
@@ -42,10 +42,7 @@ util_format_r9g9b9e5_float_unpack_rgba_float(float *dst_row, unsigned dst_stride
       float *dst = dst_row;
       const uint8_t *src = src_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
          rgb9e5_to_float3(value, dst);
          dst[3] = 1; /* a */
          src += 4;
@@ -66,10 +63,7 @@ util_format_r9g9b9e5_float_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride
       const float *src = src_row;
       uint8_t *dst = dst_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = float3_to_rgb9e5(src);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(float3_to_rgb9e5(src));
          *(uint32_t *)dst = value;
          src += 4;
          dst += 4;
@@ -83,10 +77,7 @@ void
 util_format_r9g9b9e5_float_fetch_rgba_float(float *dst, const uint8_t *src,
                                        unsigned i, unsigned j)
 {
-   uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-   value = util_bswap32(value);
-#endif
+   uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
    rgb9e5_to_float3(value, dst);
    dst[3] = 1; /* a */
 }
@@ -103,10 +94,7 @@ util_format_r9g9b9e5_float_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_str
       uint8_t *dst = dst_row;
       const uint8_t *src = src_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
          rgb9e5_to_float3(value, p);
          dst[0] = float_to_ubyte(p[0]); /* r */
          dst[1] = float_to_ubyte(p[1]); /* g */
@@ -136,10 +124,7 @@ util_format_r9g9b9e5_float_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_strid
          p[0] = ubyte_to_float(src[0]);
          p[1] = ubyte_to_float(src[1]);
          p[2] = ubyte_to_float(src[2]);
-         value = float3_to_rgb9e5(p);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(float3_to_rgb9e5(p));
          *(uint32_t *)dst = value;
          src += 4;
          dst += 4;
@@ -160,10 +145,7 @@ util_format_r11g11b10_float_unpack_rgba_float(float *dst_row, unsigned dst_strid
       float *dst = dst_row;
       const uint8_t *src = src_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
          r11g11b10f_to_float3(value, dst);
          dst[3] = 1; /* a */
          src += 4;
@@ -184,10 +166,7 @@ util_format_r11g11b10_float_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
       const float *src = src_row;
       uint8_t *dst = dst_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = float3_to_r11g11b10f(src);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(float3_to_r11g11b10f(src));
          *(uint32_t *)dst = value;
          src += 4;
          dst += 4;
@@ -201,10 +180,7 @@ void
 util_format_r11g11b10_float_fetch_rgba_float(float *dst, const uint8_t *src,
                                        unsigned i, unsigned j)
 {
-   uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-   value = util_bswap32(value);
-#endif
+   uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
    r11g11b10f_to_float3(value, dst);
    dst[3] = 1; /* a */
 }
@@ -221,10 +197,7 @@ util_format_r11g11b10_float_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       uint8_t *dst = dst_row;
       const uint8_t *src = src_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
          r11g11b10f_to_float3(value, p);
          dst[0] = float_to_ubyte(p[0]); /* r */
          dst[1] = float_to_ubyte(p[1]); /* g */
@@ -254,10 +227,7 @@ util_format_r11g11b10_float_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          p[0] = ubyte_to_float(src[0]);
          p[1] = ubyte_to_float(src[1]);
          p[2] = ubyte_to_float(src[2]);
-         value = float3_to_r11g11b10f(p);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(float3_to_r11g11b10f(p));
          *(uint32_t *)dst = value;
          src += 4;
          dst += 4;
@@ -338,13 +308,9 @@ util_format_r8g8bx_snorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint16_t *src = (const uint16_t *)src_row;
       for(x = 0; x < width; x += 1) {
-         uint16_t value = *src++;
+         uint16_t value = util_cpu_to_le16(*src++);
          int16_t r, g;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-
          r = ((int16_t)(value << 8)) >> 8;
          g = ((int16_t)(value << 0)) >> 8;
 
@@ -370,13 +336,9 @@ util_format_r8g8bx_snorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_strid
       uint8_t *dst = dst_row;
       const uint16_t *src = (const uint16_t *)src_row;
       for(x = 0; x < width; x += 1) {
-         uint16_t value = *src++;
+         uint16_t value = util_cpu_to_le16(*src++);
          int16_t r, g;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-
          r = ((int16_t)(value << 8)) >> 8;
          g = ((int16_t)(value << 0)) >> 8;
 
@@ -407,11 +369,7 @@ util_format_r8g8bx_snorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= (uint16_t)(((int8_t)(CLAMP(src[0], -1, 1) * 0x7f)) & 0xff) ;
          value |= (uint16_t)((((int8_t)(CLAMP(src[1], -1, 1) * 0x7f)) & 0xff) << 8) ;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le16_to_cpu(value);
 
          src += 4;
       }
@@ -437,11 +395,7 @@ util_format_r8g8bx_snorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= src[0] >> 1;
          value |= (src[1] >> 1) << 8;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le16_to_cpu(value);
 
          src += 4;
       }
@@ -455,13 +409,9 @@ void
 util_format_r8g8bx_snorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                      unsigned i, unsigned j)
 {
-   uint16_t value = *(const uint16_t *)src;
+   uint16_t value = util_cpu_to_le16(*(const uint16_t *)src);
    int16_t r, g;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-   value = util_bswap16(value);
-#endif
-
    r = ((int16_t)(value << 8)) >> 8;
    g = ((int16_t)(value << 0)) >> 8;
 
diff --git a/src/gallium/auxiliary/util/u_format_yuv.c b/src/gallium/auxiliary/util/u_format_yuv.c
index 891d99cf673..16ed8af7ba6 100644
--- a/src/gallium/auxiliary/util/u_format_yuv.c
+++ b/src/gallium/auxiliary/util/u_format_yuv.c
@@ -52,11 +52,7 @@ util_format_r8g8_b8g8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_strid
       float r, g0, g1, b;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          r  = ubyte_to_float((value >>  0) & 0xff);
          g0 = ubyte_to_float((value >>  8) & 0xff);
@@ -77,11 +73,7 @@ util_format_r8g8_b8g8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_strid
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          r  = ubyte_to_float((value >>  0) & 0xff);
          g0 = ubyte_to_float((value >>  8) & 0xff);
@@ -114,11 +106,7 @@ util_format_r8g8_b8g8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       uint8_t r, g0, g1, b;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          r  = (value >>  0) & 0xff;
          g0 = (value >>  8) & 0xff;
@@ -139,11 +127,7 @@ util_format_r8g8_b8g8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          r  = (value >>  0) & 0xff;
          g0 = (value >>  8) & 0xff;
@@ -186,11 +170,7 @@ util_format_r8g8_b8g8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
          value |= float_to_ubyte(b)  << 16;
          value |= float_to_ubyte(g1) << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -206,11 +186,7 @@ util_format_r8g8_b8g8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
          value |= float_to_ubyte(b)  << 16;
          value |= float_to_ubyte(g1) << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -243,11 +219,7 @@ util_format_r8g8_b8g8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          value |= b  << 16;
          value |= g1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -263,11 +235,7 @@ util_format_r8g8_b8g8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          value |= b  << 16;
          value |= g1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -304,11 +272,7 @@ util_format_g8r8_g8b8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_strid
       float r, g0, g1, b;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          g0 = ubyte_to_float((value >>  0) & 0xff);
          r  = ubyte_to_float((value >>  8) & 0xff);
@@ -329,11 +293,7 @@ util_format_g8r8_g8b8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_strid
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          g0 = ubyte_to_float((value >>  0) & 0xff);
          r  = ubyte_to_float((value >>  8) & 0xff);
@@ -366,11 +326,7 @@ util_format_g8r8_g8b8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       uint8_t r, g0, g1, b;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          g0 = (value >>  0) & 0xff;
          r  = (value >>  8) & 0xff;
@@ -391,11 +347,7 @@ util_format_g8r8_g8b8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          g0 = (value >>  0) & 0xff;
          r  = (value >>  8) & 0xff;
@@ -438,11 +390,7 @@ util_format_g8r8_g8b8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
          value |= float_to_ubyte(g1) << 16;
          value |= float_to_ubyte(b)  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -458,11 +406,7 @@ util_format_g8r8_g8b8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
          value |= float_to_ubyte(g1) << 16;
          value |= float_to_ubyte(b)  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -495,11 +439,7 @@ util_format_g8r8_g8b8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          value |= g1 << 16;
          value |= b  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -515,11 +455,7 @@ util_format_g8r8_g8b8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          value |= g1 << 16;
          value |= b  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -556,11 +492,7 @@ util_format_uyvy_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       uint8_t y0, y1, u, v;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          u  = (value >>  0) & 0xff;
          y0 = (value >>  8) & 0xff;
@@ -577,11 +509,7 @@ util_format_uyvy_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          u  = (value >>  0) & 0xff;
          y0 = (value >>  8) & 0xff;
@@ -612,11 +540,7 @@ util_format_uyvy_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
       uint8_t y0, y1, u, v;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          u  = (value >>  0) & 0xff;
          y0 = (value >>  8) & 0xff;
@@ -633,11 +557,7 @@ util_format_uyvy_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          u  = (value >>  0) & 0xff;
          y0 = (value >>  8) & 0xff;
@@ -683,11 +603,7 @@ util_format_uyvy_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= v  << 16;
          value |= y1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -702,11 +618,7 @@ util_format_uyvy_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= v  << 16;
          value |= y1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -744,11 +656,7 @@ util_format_uyvy_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= v  << 16;
          value |= y1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -763,11 +671,7 @@ util_format_uyvy_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= v  << 16;
          value |= y1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -809,11 +713,7 @@ util_format_yuyv_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       uint8_t y0, y1, u, v;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          y0 = (value >>  0) & 0xff;
          u  = (value >>  8) & 0xff;
@@ -830,11 +730,7 @@ util_format_yuyv_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          y0 = (value >>  0) & 0xff;
          u  = (value >>  8) & 0xff;
@@ -865,11 +761,7 @@ util_format_yuyv_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
       uint8_t y0, y1, u, v;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          y0 = (value >>  0) & 0xff;
          u  = (value >>  8) & 0xff;
@@ -886,11 +778,7 @@ util_format_yuyv_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          y0 = (value >>  0) & 0xff;
          u  = (value >>  8) & 0xff;
@@ -936,11 +824,7 @@ util_format_yuyv_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= y1 << 16;
          value |= v  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -955,11 +839,7 @@ util_format_yuyv_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= y1 << 16;
          value |= v  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -997,11 +877,7 @@ util_format_yuyv_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= y1 << 16;
          value |= v  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -1016,11 +892,7 @@ util_format_yuyv_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= y1 << 16;
          value |= v  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
diff --git a/src/gallium/auxiliary/util/u_format_zs.c b/src/gallium/auxiliary/util/u_format_zs.c
index ed45c52f9c9..f1ed32f1d5c 100644
--- a/src/gallium/auxiliary/util/u_format_zs.c
+++ b/src/gallium/auxiliary/util/u_format_zs.c
@@ -147,10 +147,7 @@ util_format_z16_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint16_t *src = (const uint16_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint16_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
+         uint16_t value = util_cpu_to_le16(*src++);
          *dst++ = z16_unorm_to_z32_float(value);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -170,10 +167,7 @@ util_format_z16_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint16_t value;
          value = z32_float_to_z16_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_le16_to_cpu(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -190,10 +184,7 @@ util_format_z16_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
       uint32_t *dst = dst_row;
       const uint16_t *src = (const uint16_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint16_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
+         uint16_t value = util_cpu_to_le16(*src++);
          *dst++ = z16_unorm_to_z32_unorm(value);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -213,10 +204,7 @@ util_format_z16_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint16_t value;
          value = z32_unorm_to_z16_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_le16_to_cpu(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -233,10 +221,7 @@ util_format_z32_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z32_unorm_to_z32_float(value);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -256,10 +241,7 @@ util_format_z32_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_float_to_z32_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -362,10 +344,7 @@ util_format_z24_unorm_s8_uint_unpack_z_float(float *dst_row, unsigned dst_stride
       float *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value =  util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_float(value & 0xffffff);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -383,16 +362,10 @@ util_format_z24_unorm_s8_uint_pack_z_float(uint8_t *dst_row, unsigned dst_stride
       const float *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0xff000000;
          value |= z32_float_to_z24_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -409,10 +382,7 @@ util_format_z24_unorm_s8_uint_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_s
       uint32_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_unorm(value & 0xffffff);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -430,16 +400,10 @@ util_format_z24_unorm_s8_uint_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stri
       const uint32_t *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value= *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0xff000000;
          value |= z32_unorm_to_z24_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -456,10 +420,7 @@ util_format_z24_unorm_s8_uint_unpack_s_8uint(uint8_t *dst_row, unsigned dst_stri
       uint8_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = value >> 24;
       }
       src_row += src_stride/sizeof(*src_row);
@@ -477,16 +438,10 @@ util_format_z24_unorm_s8_uint_pack_s_8uint(uint8_t *dst_row, unsigned dst_stride
       const uint8_t *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0x00ffffff;
          value |= *src++ << 24;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -503,10 +458,7 @@ util_format_s8_uint_z24_unorm_unpack_z_float(float *dst_row, unsigned dst_stride
       float *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_float(value >> 8);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -524,16 +476,10 @@ util_format_s8_uint_z24_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride
       const float *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0x000000ff;
          value |= z32_float_to_z24_unorm(*src++) << 8;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -550,10 +496,7 @@ util_format_s8_uint_z24_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_s
       uint32_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_unorm(value >> 8);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -571,16 +514,10 @@ util_format_s8_uint_z24_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stri
       const uint32_t *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0x000000ff;
          value |= *src++ & 0xffffff00;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -597,10 +534,7 @@ util_format_s8_uint_z24_unorm_unpack_s_8uint(uint8_t *dst_row, unsigned dst_stri
       uint8_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = value & 0xff;
       }
       src_row += src_stride/sizeof(*src_row);
@@ -618,16 +552,10 @@ util_format_s8_uint_z24_unorm_pack_s_8uint(uint8_t *dst_row, unsigned dst_stride
       const uint8_t *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0xffffff00;
          value |= *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -644,10 +572,7 @@ util_format_z24x8_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_float(value & 0xffffff);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -667,10 +592,7 @@ util_format_z24x8_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_float_to_z24_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -687,10 +609,7 @@ util_format_z24x8_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
       uint32_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_unorm(value & 0xffffff);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -710,10 +629,7 @@ util_format_z24x8_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_unorm_to_z24_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -730,10 +646,7 @@ util_format_x8z24_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint32_t *src = (uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_float(value >> 8);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -753,10 +666,7 @@ util_format_x8z24_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_float_to_z24_unorm(*src++) << 8;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -773,10 +683,7 @@ util_format_x8z24_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
       uint32_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_unorm(value >> 8);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -796,10 +703,7 @@ util_format_x8z24_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_unorm_to_z24_unorm(*src++) << 8;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);