[Mesa-dev] [PATCH 1/8] vbo: Correctly handle attribute offsets in dlist draw.

Brian Paul brianp at vmware.com
Wed Jan 31 16:54:55 UTC 2018


Hi Matthias,

Nice work!  It's nice to get rid of some of those attribute loops.
The series looks good to me, just assorted nit-picks and comments on a 
few patches.

-Brian


On 01/31/2018 12:55 AM, Mathias.Froehlich at gmx.net wrote:
> From: Mathias Fröhlich <mathias.froehlich at web.de>
> 
> When executing a display list draw, for the offset
> list to be correct, the offset computation needs to
> accumulate all attribute size values in order.
> Specifically, if we are shuffling around the position
> and generic0 attributes, we may violate the order or
> if we do not walk the generic vbo attributes we may
> skip some of the attributes.
> Even if this is an unlikely usecase we can fix this

"use case"

> by precomputing the offsets on the full attribute list
> and store the full offset list in the display list node.
> 
> v2: Formatting fix
> v3: Rebase
> 
> Signed-off-by: Mathias Fröhlich <Mathias.Froehlich at web.de>
> ---
>   src/mesa/vbo/vbo_save.h      |  1 +
>   src/mesa/vbo/vbo_save_api.c  | 19 ++++++++++++++++++
>   src/mesa/vbo/vbo_save_draw.c | 47 +++++++++++++++-----------------------------
>   3 files changed, 36 insertions(+), 31 deletions(-)
> 
> diff --git a/src/mesa/vbo/vbo_save.h b/src/mesa/vbo/vbo_save.h
> index 1dc66a598e..cb0bff2341 100644
> --- a/src/mesa/vbo/vbo_save.h
> +++ b/src/mesa/vbo/vbo_save.h
> @@ -64,6 +64,7 @@ struct vbo_save_vertex_list {
>      GLbitfield64 enabled; /**< mask of enabled vbo arrays. */
>      GLubyte attrsz[VBO_ATTRIB_MAX];
>      GLenum16 attrtype[VBO_ATTRIB_MAX];
> +   GLuint offsets[VBO_ATTRIB_MAX];
>      GLuint vertex_size;  /**< size in GLfloats */
>   
>      /* Copy of the final vertex from node->vertex_store->bufferobj.
> diff --git a/src/mesa/vbo/vbo_save_api.c b/src/mesa/vbo/vbo_save_api.c
> index a0fb62d814..fb51bdb84e 100644
> --- a/src/mesa/vbo/vbo_save_api.c
> +++ b/src/mesa/vbo/vbo_save_api.c
> @@ -420,6 +420,8 @@ compile_vertex_list(struct gl_context *ctx)
>   {
>      struct vbo_save_context *save = &vbo_context(ctx)->save;
>      struct vbo_save_vertex_list *node;
> +   GLuint offset;
> +   unsigned i;
>   
>      /* Allocate space for this structure in the display list currently
>       * being compiled.
> @@ -443,6 +445,23 @@ compile_vertex_list(struct gl_context *ctx)
>      node->vertex_size = save->vertex_size;
>      node->buffer_offset =
>         (save->buffer_map - save->vertex_store->buffer_map) * sizeof(GLfloat);
> +   if (aligned_vertex_buffer_offset(node)) {
> +      /* The vertex size is an exact multiple of the buffer offset.
> +       * This means that we can use zero-based vertex attribute pointers
> +       * and specify the start of the primitive with the _mesa_prim::start
> +       * field.  This results in issuing several draw calls with identical
> +       * vertex attribute information.  This can result in fewer state
> +       * changes in drivers.  In particular, the Gallium CSO module will
> +       * filter out redundant vertex buffer changes.
> +       */
> +      offset = 0;
> +   } else {
> +      offset = node->buffer_offset;
> +   }
> +   for (i = 0; i < VBO_ATTRIB_MAX; ++i) {
> +      node->offsets[i] = offset;
> +      offset += node->attrsz[i] * sizeof(GLfloat);
> +   }
>      node->vertex_count = save->vert_count;
>      node->wrap_count = save->copied.nr;
>      node->dangling_attr_ref = save->dangling_attr_ref;
> diff --git a/src/mesa/vbo/vbo_save_draw.c b/src/mesa/vbo/vbo_save_draw.c
> index fd0ccc1230..291d99ed9a 100644
> --- a/src/mesa/vbo/vbo_save_draw.c
> +++ b/src/mesa/vbo/vbo_save_draw.c
> @@ -26,6 +26,7 @@
>    *    Keith Whitwell <keithw at vmware.com>
>    */
>   
> +#include <stdbool.h>
>   #include "main/glheader.h"
>   #include "main/bufferobj.h"
>   #include "main/context.h"
> @@ -137,29 +138,10 @@ bind_vertex_list(struct gl_context *ctx,
>      struct vbo_context *vbo = vbo_context(ctx);
>      struct vbo_save_context *save = &vbo->save;
>      struct gl_vertex_array *arrays = save->arrays;
> -   GLuint buffer_offset = node->buffer_offset;
>      const GLubyte *map;
>      GLuint attr;
> -   GLubyte node_attrsz[VBO_ATTRIB_MAX];  /* copy of node->attrsz[] */
> -   GLenum16 node_attrtype[VBO_ATTRIB_MAX];  /* copy of node->attrtype[] */
>      GLbitfield varying_inputs = 0x0;
> -
> -   STATIC_ASSERT(sizeof(node_attrsz) == sizeof(node->attrsz));
> -   memcpy(node_attrsz, node->attrsz, sizeof(node->attrsz));
> -   STATIC_ASSERT(sizeof(node_attrtype) == sizeof(node->attrtype));
> -   memcpy(node_attrtype, node->attrtype, sizeof(node->attrtype));
> -
> -   if (aligned_vertex_buffer_offset(node)) {
> -      /* The vertex size is an exact multiple of the buffer offset.
> -       * This means that we can use zero-based vertex attribute pointers
> -       * and specify the start of the primitive with the _mesa_prim::start
> -       * field.  This results in issuing several draw calls with identical
> -       * vertex attribute information.  This can result in fewer state
> -       * changes in drivers.  In particular, the Gallium CSO module will
> -       * filter out redundant vertex buffer changes.
> -       */
> -      buffer_offset = 0;
> -   }
> +   bool generic_from_pos = false;
>   
>      /* Install the default (ie Current) attributes first */
>      for (attr = 0; attr < VERT_ATTRIB_FF_MAX; attr++) {
> @@ -191,10 +173,7 @@ bind_vertex_list(struct gl_context *ctx,
>            ctx->VertexProgram._Current->info.inputs_read;
>         if ((inputs_read & VERT_BIT_POS) == 0 &&
>             (inputs_read & VERT_BIT_GENERIC0)) {
> -         save->inputs[VERT_ATTRIB_GENERIC0] = save->inputs[0];
> -         node_attrsz[VERT_ATTRIB_GENERIC0] = node_attrsz[0];
> -         node_attrtype[VERT_ATTRIB_GENERIC0] = node_attrtype[0];
> -         node_attrsz[0] = 0;
> +         generic_from_pos = true;
>         }
>         break;
>      default:
> @@ -203,31 +182,37 @@ bind_vertex_list(struct gl_context *ctx,
>   
>      for (attr = 0; attr < VERT_ATTRIB_MAX; attr++) {
>         const GLuint src = map[attr];
> +      const GLubyte size = node->attrsz[src];
>   
> -      if (node_attrsz[src]) {
> +      if (size) {
>            struct gl_vertex_array *array = &arrays[attr];
> +         const GLenum16 type = node->attrtype[src];
>   
>            /* override the default array set above */

That comment is no longer quite correct.

>            save->inputs[attr] = array;
>   
> -         array->Ptr = (const GLubyte *) NULL + buffer_offset;
> -         array->Size = node_attrsz[src];
> +         array->Ptr = (const GLubyte *) NULL + node->offsets[src];
> +         array->Size = size;
>            array->StrideB = node->vertex_size * sizeof(GLfloat);
> -         array->Type = node_attrtype[src];
> -         array->Integer = vbo_attrtype_to_integer_flag(node_attrtype[src]);
> +         array->Type = type;
> +         array->Integer = vbo_attrtype_to_integer_flag(type);
>            array->Format = GL_RGBA;
> -         array->_ElementSize = array->Size * sizeof(GLfloat);
> +         array->_ElementSize = size * sizeof(GLfloat);
>            _mesa_reference_buffer_object(ctx,
>                                          &array->BufferObj,
>                                          node->vertex_store->bufferobj);
>   
>            assert(array->BufferObj->Name);
>   
> -         buffer_offset += node_attrsz[src] * sizeof(GLfloat);
>            varying_inputs |= VERT_BIT(attr);
>         }
>      }
>   
> +   if (generic_from_pos) {
> +      varying_inputs |= (varying_inputs & VERT_BIT_POS) << VERT_ATTRIB_GENERIC0;
> +      save->inputs[VERT_ATTRIB_GENERIC0] = save->inputs[VERT_ATTRIB_POS];
> +   }
> +
>      _mesa_set_varying_vp_inputs(ctx, varying_inputs);
>      ctx->NewDriverState |= ctx->DriverFlags.NewArray;
>   }
> 



More information about the mesa-dev mailing list