swr/rast: Convert C Sampler intrinsics

Convert portions of the C sampler to the rasty SIMD lib.

Also fix SRL call with a non-immediate.  Don't count on the compiler
automagically converting an srli call to srl if the shift count isn't
an immediate.

Reviewed-by: Bruce Cherniak <[email protected]>

2 files changed, 19 insertions, 0 deletions

diff --git a/src/gallium/drivers/swr/rasterizer/common/intrin.h b/src/gallium/drivers/swr/rasterizer/common/intrin.h
index 33d37e3cece..59d66bc60a8 100644
--- a/src/gallium/drivers/swr/rasterizer/common/intrin.h
+++ b/src/gallium/drivers/swr/rasterizer/common/intrin.h
@@ -26,7 +26,10 @@
 
 #include "os.h"
 
+#if !defined(SIMD_ARCH)
 #define SIMD_ARCH KNOB_ARCH
+#endif 
+
 #include "simdlib_types.hpp"
 
 typedef SIMDImpl::SIMD128Impl::Float                      simd4scalar;
diff --git a/src/gallium/drivers/swr/rasterizer/common/simdlib_128_avx.inl b/src/gallium/drivers/swr/rasterizer/common/simdlib_128_avx.inl
index 72327918937..b1511c6c0e2 100644
--- a/src/gallium/drivers/swr/rasterizer/common/simdlib_128_avx.inl
+++ b/src/gallium/drivers/swr/rasterizer/common/simdlib_128_avx.inl
@@ -174,6 +174,7 @@ SIMD_IWRAPPER_2_(xor_si, _mm_xor_si128);        // return a ^ b       (int)
 // Shift operations
 //-----------------------------------------------------------------------
 SIMD_IWRAPPER_1I(slli_epi32);               // return a << ImmT
+SIMD_IWRAPPER_1I(slli_epi64);               // return a << ImmT
 
 static SIMDINLINE Integer SIMDCALL sllv_epi32(Integer vA, Integer vB) // return a << b      (uint32)
 {
@@ -205,6 +206,11 @@ SIMD_IWRAPPER_1I(srai_epi32);               // return a >> ImmT   (int32)
 SIMD_IWRAPPER_1I(srli_epi32);               // return a >> ImmT   (uint32)
 SIMD_IWRAPPER_1I_(srli_si, _mm_srli_si128); // return a >> (ImmT*8) (uint)
 
+static SIMDINLINE Integer SIMDCALL srl_epi64(Integer a, Integer n)
+{
+    return _mm_srl_epi64(a, n);
+}
+
 template<int ImmT>                              // same as srli_si, but with Float cast to int
 static SIMDINLINE Float SIMDCALL srlisi_ps(Float a)
 {
@@ -272,6 +278,16 @@ static SIMDINLINE Float SIMDCALL cvtepi32_ps(Integer a) // return (float)a    (i
     return _mm_cvtepi32_ps(a);
 }
 
+static SIMDINLINE int32_t SIMDCALL cvtsi128_si32(Integer a) // return a.v[0]
+{
+    return _mm_cvtsi128_si32(a);
+}
+
+static SIMDINLINE Integer SIMDCALL cvtsi32_si128(int32_t n) // return a[0] = n, a[1]...a[3] = 0
+{
+    return _mm_cvtsi32_si128(n);
+}
+
 SIMD_IWRAPPER_1(cvtepu8_epi16);     // return (int16)a    (uint8 --> int16)
 SIMD_IWRAPPER_1(cvtepu8_epi32);     // return (int32)a    (uint8 --> int32)
 SIMD_IWRAPPER_1(cvtepu16_epi32);    // return (int32)a    (uint16 --> int32)