translate: add support for 8/16-bit indices

Currently, only 32-bit indices are supported, but some use cases
translate needs support for all types.

diff --git a/src/gallium/auxiliary/rtasm/rtasm_x86sse.c b/src/gallium/auxiliary/rtasm/rtasm_x86sse.c
index 9f70b73698aaa2735bf0fda24ae110d4b495605a..63007c1feb8650db2212aae980b27fd772ebf147 100644 (file)
--- a/src/gallium/auxiliary/rtasm/rtasm_x86sse.c
+++ b/src/gallium/auxiliary/rtasm/rtasm_x86sse.c
@@ -586,6 +586,20 @@ void x86_mov( struct x86_function *p,
    emit_op_modrm( p, 0x8b, 0x89, dst, src );
 }
 
+void x86_movzx8(struct x86_function *p, struct x86_reg dst, struct x86_reg src )
+{
+   DUMP_RR( dst, src );
+   emit_2ub(p, 0x0f, 0xb6);
+   emit_modrm(p, dst, src);
+}
+
+void x86_movzx16(struct x86_function *p, struct x86_reg dst, struct x86_reg src )
+{
+   DUMP_RR( dst, src );
+   emit_2ub(p, 0x0f, 0xb7);
+   emit_modrm(p, dst, src);
+}
+
 void x86_xor( struct x86_function *p,
              struct x86_reg dst,
              struct x86_reg src )

diff --git a/src/gallium/auxiliary/rtasm/rtasm_x86sse.h b/src/gallium/auxiliary/rtasm/rtasm_x86sse.h
index 6208e8f707f26a89f593d54d5f8ea530ae0f9eda..365dec109e779d3c53e771fdb926a34c638432a2 100644 (file)
--- a/src/gallium/auxiliary/rtasm/rtasm_x86sse.h
+++ b/src/gallium/auxiliary/rtasm/rtasm_x86sse.h
@@ -237,6 +237,8 @@ void x86_dec( struct x86_function *p, struct x86_reg reg );
 void x86_inc( struct x86_function *p, struct x86_reg reg );
 void x86_lea( struct x86_function *p, struct x86_reg dst, struct x86_reg src );
 void x86_mov( struct x86_function *p, struct x86_reg dst, struct x86_reg src );
+void x86_movzx8( struct x86_function *p, struct x86_reg dst, struct x86_reg src );
+void x86_movzx16( struct x86_function *p, struct x86_reg dst, struct x86_reg src );
 void x86_mul( struct x86_function *p, struct x86_reg src );
 void x86_imul( struct x86_function *p, struct x86_reg dst, struct x86_reg src );
 void x86_or( struct x86_function *p, struct x86_reg dst, struct x86_reg src );

diff --git a/src/gallium/auxiliary/translate/translate.h b/src/gallium/auxiliary/translate/translate.h
index eb6f2cc4862b0793dbaa4e3902438db6edf19692..a75380228b1dd21f116843e35d4b9c07a750ae91 100644 (file)
--- a/src/gallium/auxiliary/translate/translate.h
+++ b/src/gallium/auxiliary/translate/translate.h
@@ -85,6 +85,18 @@ struct translate {
                                 unsigned instance_id,
                                 void *output_buffer);
 
+   void (PIPE_CDECL *run_elts16)( struct translate *,
+                                const uint16_t *elts,
+                                unsigned count,
+                                unsigned instance_id,
+                                void *output_buffer);
+
+   void (PIPE_CDECL *run_elts8)( struct translate *,
+                                const uint8_t *elts,
+                                unsigned count,
+                                unsigned instance_id,
+                                void *output_buffer);
+
    void (PIPE_CDECL *run)( struct translate *,
                            unsigned start,
                            unsigned count,

diff --git a/src/gallium/auxiliary/translate/translate_generic.c b/src/gallium/auxiliary/translate/translate_generic.c
index 828b76dc77fee4da3baa52b0b99d5cf1d89c5f5a..975f23a6f4c8d7f6dcee3d6fedd6e83947f84384 100644 (file)
--- a/src/gallium/auxiliary/translate/translate_generic.c
+++ b/src/gallium/auxiliary/translate/translate_generic.c
@@ -441,6 +441,38 @@ static void PIPE_CDECL generic_run_elts( struct translate *translate,
    }
 }
 
+static void PIPE_CDECL generic_run_elts16( struct translate *translate,
+                                         const uint16_t *elts,
+                                         unsigned count,
+                                         unsigned instance_id,
+                                         void *output_buffer )
+{
+   struct translate_generic *tg = translate_generic(translate);
+   char *vert = output_buffer;
+   unsigned i;
+
+   for (i = 0; i < count; i++) {
+      generic_run_one(tg, *elts++, instance_id, vert);
+      vert += tg->translate.key.output_stride;
+   }
+}
+
+static void PIPE_CDECL generic_run_elts8( struct translate *translate,
+                                         const uint8_t *elts,
+                                         unsigned count,
+                                         unsigned instance_id,
+                                         void *output_buffer )
+{
+   struct translate_generic *tg = translate_generic(translate);
+   char *vert = output_buffer;
+   unsigned i;
+
+   for (i = 0; i < count; i++) {
+      generic_run_one(tg, *elts++, instance_id, vert);
+      vert += tg->translate.key.output_stride;
+   }
+}
+
 static void PIPE_CDECL generic_run( struct translate *translate,
                                     unsigned start,
                                     unsigned count,
@@ -498,6 +530,8 @@ struct translate *translate_generic_create( const struct translate_key *key )
    tg->translate.release = generic_release;
    tg->translate.set_buffer = generic_set_buffer;
    tg->translate.run_elts = generic_run_elts;
+   tg->translate.run_elts16 = generic_run_elts16;
+   tg->translate.run_elts8 = generic_run_elts8;
    tg->translate.run = generic_run;
 
    for (i = 0; i < key->nr_elements; i++) {

diff --git a/src/gallium/auxiliary/translate/translate_sse.c b/src/gallium/auxiliary/translate/translate_sse.c
index 68c71f42513357e5a7d3ac988114146f12604a4e..f9aab9232c551847cce4d7a9c0cd959a076e8e07 100644 (file)
--- a/src/gallium/auxiliary/translate/translate_sse.c
+++ b/src/gallium/auxiliary/translate/translate_sse.c
@@ -67,6 +67,8 @@ struct translate_sse {
 
    struct x86_function linear_func;
    struct x86_function elt_func;
+   struct x86_function elt16_func;
+   struct x86_function elt8_func;
    struct x86_function *func;
 
    boolean loaded_identity;
@@ -362,7 +364,7 @@ static boolean translate_attr( struct translate_sse *p,
 
 
 static boolean init_inputs( struct translate_sse *p,
-                            boolean linear )
+                            unsigned index_size )
 {
    unsigned i;
    struct x86_reg instance_id = x86_make_disp(p->machine_EDX,
@@ -372,7 +374,7 @@ static boolean init_inputs( struct translate_sse *p,
       struct translate_buffer_varient *varient = &p->buffer_varient[i];
       struct translate_buffer *buffer = &p->buffer[varient->buffer_index];
 
-      if (linear || varient->instance_divisor) {
+      if (!index_size || varient->instance_divisor) {
          struct x86_reg buf_stride   = x86_make_disp(p->machine_EDX,
                                                      get_offset(p, &buffer->stride));
          struct x86_reg buf_ptr      = x86_make_disp(p->machine_EDX,
@@ -421,7 +423,7 @@ static boolean init_inputs( struct translate_sse *p,
          /* In the linear case, keep the buffer pointer instead of the
           * index number.
           */
-         if (linear && p->nr_buffer_varients == 1)
+         if (!index_size && p->nr_buffer_varients == 1)
             x86_mov(p->func, elt, tmp_EAX);
          else
             x86_mov(p->func, buf_ptr, tmp_EAX);
@@ -433,7 +435,7 @@ static boolean init_inputs( struct translate_sse *p,
 
 
 static struct x86_reg get_buffer_ptr( struct translate_sse *p,
-                                      boolean linear,
+                                      unsigned index_size,
                                       unsigned var_idx,
                                       struct x86_reg elt )
 {
@@ -441,10 +443,10 @@ static struct x86_reg get_buffer_ptr( struct translate_sse *p,
       return x86_make_disp(p->machine_EDX,
                            get_offset(p, &p->instance_id));
    }
-   if (linear && p->nr_buffer_varients == 1) {
+   if (!index_size && p->nr_buffer_varients == 1) {
       return p->idx_EBX;
    }
-   else if (linear || p->buffer_varient[var_idx].instance_divisor) {
+   else if (!index_size || p->buffer_varient[var_idx].instance_divisor) {
       struct x86_reg ptr = p->tmp_EAX;
       struct x86_reg buf_ptr = 
          x86_make_disp(p->machine_EDX, 
@@ -469,8 +471,19 @@ static struct x86_reg get_buffer_ptr( struct translate_sse *p,
 
       /* Calculate pointer to current attrib:
        */
-      x86_mov(p->func, ptr, buf_stride);
-      x86_imul(p->func, ptr, elt);
+      switch(index_size)
+      {
+      case 1:
+         x86_movzx8(p->func, ptr, elt);
+         break;
+      case 2:
+         x86_movzx16(p->func, ptr, elt);
+         break;
+      case 4:
+         x86_mov(p->func, ptr, elt);
+         break;
+      }
+      x86_imul(p->func, ptr, buf_stride);
       x86_add(p->func, ptr, buf_base_ptr);
       return ptr;
    }
@@ -479,9 +492,9 @@ static struct x86_reg get_buffer_ptr( struct translate_sse *p,
 
 
 static boolean incr_inputs( struct translate_sse *p, 
-                            boolean linear )
+                            unsigned index_size )
 {
-   if (linear && p->nr_buffer_varients == 1) {
+   if (!index_size && p->nr_buffer_varients == 1) {
       struct x86_reg stride = x86_make_disp(p->machine_EDX,
                                             get_offset(p, &p->buffer[0].stride));
 
@@ -490,7 +503,7 @@ static boolean incr_inputs( struct translate_sse *p,
          sse_prefetchnta(p->func, x86_make_disp(p->idx_EBX, 192));
       }
    }
-   else if (linear) {
+   else if (!index_size) {
       unsigned i;
 
       /* Is this worthwhile??
@@ -511,7 +524,7 @@ static boolean incr_inputs( struct translate_sse *p,
       }
    } 
    else {
-      x86_lea(p->func, p->idx_EBX, x86_make_disp(p->idx_EBX, 4));
+      x86_lea(p->func, p->idx_EBX, x86_make_disp(p->idx_EBX, index_size));
    }
    
    return TRUE;
@@ -536,7 +549,7 @@ static boolean incr_inputs( struct translate_sse *p,
  */
 static boolean build_vertex_emit( struct translate_sse *p,
                                  struct x86_function *func,
-                                 boolean linear )
+                                 unsigned index_size )
 {
    int fixup, label;
    unsigned j;
@@ -585,13 +598,13 @@ static boolean build_vertex_emit( struct translate_sse *p,
 
    /* always load, needed or not:
     */
-   init_inputs(p, linear);
+   init_inputs(p, index_size);
 
    /* Note address for loop jump
     */
    label = x86_get_label(p->func);
    {
-      struct x86_reg elt = linear ? p->idx_EBX : x86_deref(p->idx_EBX);
+      struct x86_reg elt = !index_size ? p->idx_EBX : x86_deref(p->idx_EBX);
       int last_varient = -1;
       struct x86_reg vb;
 
@@ -603,7 +616,7 @@ static boolean build_vertex_emit( struct translate_sse *p,
           */
          if (varient != last_varient) {
             last_varient = varient;
-            vb = get_buffer_ptr(p, linear, varient, elt);
+            vb = get_buffer_ptr(p, index_size, varient, elt);
          }
          
          if (!translate_attr( p, a, 
@@ -621,7 +634,7 @@ static boolean build_vertex_emit( struct translate_sse *p,
 
       /* Incr index
        */ 
-      incr_inputs( p, linear );
+      incr_inputs( p, index_size );
    }
 
    /* decr count, loop if not zero
@@ -736,10 +749,16 @@ struct translate *translate_sse2_create( const struct translate_key *key )
 
    if (0) debug_printf("nr_buffers: %d\n", p->nr_buffers);
 
-   if (!build_vertex_emit(p, &p->linear_func, TRUE))
+   if (!build_vertex_emit(p, &p->linear_func, 0))
+      goto fail;
+
+   if (!build_vertex_emit(p, &p->elt_func, 4))
+      goto fail;
+
+   if (!build_vertex_emit(p, &p->elt16_func, 2))
       goto fail;
 
-   if (!build_vertex_emit(p, &p->elt_func, FALSE))
+   if (!build_vertex_emit(p, &p->elt8_func, 1))
       goto fail;
 
    p->translate.run = (void*)x86_get_func(&p->linear_func);
@@ -750,6 +769,14 @@ struct translate *translate_sse2_create( const struct translate_key *key )
    if (p->translate.run_elts == NULL)
       goto fail;
 
+   p->translate.run_elts16 = (void*)x86_get_func(&p->elt16_func);
+   if (p->translate.run_elts16 == NULL)
+      goto fail;
+
+   p->translate.run_elts8 = (void*)x86_get_func(&p->elt8_func);
+   if (p->translate.run_elts8 == NULL)
+      goto fail;
+
    return &p->translate;
 
  fail: