Merge remote branch 'origin/master' into lp-binning

Conflicts:
	src/gallium/drivers/llvmpipe/lp_quad.h
	src/gallium/drivers/llvmpipe/lp_setup.c

8 files changed, 34 insertions, 27 deletions

diff --git a/src/gallium/drivers/cell/spu/spu_command.c b/src/gallium/drivers/cell/spu/spu_command.c
index 12b855a3db..55bd85bde2 100644
--- a/src/gallium/drivers/cell/spu/spu_command.c
+++ b/src/gallium/drivers/cell/spu/spu_command.c
@@ -53,8 +53,7 @@ struct spu_vs_context draw;
 /**
  * Buffers containing dynamically generated SPU code:
  */
-static unsigned char attribute_fetch_code_buffer[136 * PIPE_MAX_ATTRIBS]
-    ALIGN16_ATTRIB;
+PIPE_ALIGN_VAR(16) static unsigned char attribute_fetch_code_buffer[136 * PIPE_MAX_ATTRIBS];
 
 
 
@@ -543,7 +542,7 @@ cmd_batch(uint opcode)
 {
    const uint buf = (opcode >> 8) & 0xff;
    uint size = (opcode >> 16);
-   qword buffer[CELL_BUFFER_SIZE / 16] ALIGN16_ATTRIB;
+   PIPE_ALIGN_VAR(16) qword buffer[CELL_BUFFER_SIZE / 16];
    const unsigned usize = ROUNDUP16(size) / sizeof(buffer[0]);
    uint pos;
 
diff --git a/src/gallium/drivers/cell/spu/spu_exec.c b/src/gallium/drivers/cell/spu/spu_exec.c
index d86d8e09a5..d2166a4901 100644
--- a/src/gallium/drivers/cell/spu/spu_exec.c
+++ b/src/gallium/drivers/cell/spu/spu_exec.c
@@ -1839,10 +1839,11 @@ spu_exec_machine_run( struct spu_exec_machine *mach )
    /* execute declarations (interpolants) */
    if( mach->Processor == TGSI_PROCESSOR_FRAGMENT ) {
       for (i = 0; i < mach->NumDeclarations; i++) {
+         PIPE_ALIGN_VAR(16)
          union {
             struct tgsi_full_declaration decl;
             qword buffer[ROUNDUP16(sizeof(struct tgsi_full_declaration)) / 16];
-         } d ALIGN16_ATTRIB;
+         } d;
          unsigned ea = (unsigned) (mach->Declarations + pc);
 
          spu_dcache_fetch_unaligned(d.buffer, ea, sizeof(d.decl));
@@ -1853,10 +1854,11 @@ spu_exec_machine_run( struct spu_exec_machine *mach )
 
    /* execute instructions, until pc is set to -1 */
    while (pc != -1) {
+      PIPE_ALIGN_VAR(16)
       union {
          struct tgsi_full_instruction inst;
          qword buffer[ROUNDUP16(sizeof(struct tgsi_full_instruction)) / 16];
-      } i ALIGN16_ATTRIB;
+      } i;
       unsigned ea = (unsigned) (mach->Instructions + pc);
 
       spu_dcache_fetch_unaligned(i.buffer, ea, sizeof(i.inst));
diff --git a/src/gallium/drivers/cell/spu/spu_exec.h b/src/gallium/drivers/cell/spu/spu_exec.h
index 8605679940..0ca92af248 100644
--- a/src/gallium/drivers/cell/spu/spu_exec.h
+++ b/src/gallium/drivers/cell/spu/spu_exec.h
@@ -98,9 +98,9 @@ struct spu_exec_machine
     * 4  internal temporaries
     * 1  address
     */
+   PIPE_ALIGN_VAR(16)
    struct spu_exec_vector       Temps[TGSI_EXEC_NUM_TEMPS 
-                                      + TGSI_EXEC_NUM_TEMP_EXTRAS + 1]
-       ALIGN16_ATTRIB;
+                                      + TGSI_EXEC_NUM_TEMP_EXTRAS + 1];
 
    struct spu_exec_vector       *Addrs;
 
diff --git a/src/gallium/drivers/cell/spu/spu_funcs.c b/src/gallium/drivers/cell/spu/spu_funcs.c
index ff3d609d25..98919c43ff 100644
--- a/src/gallium/drivers/cell/spu/spu_funcs.c
+++ b/src/gallium/drivers/cell/spu/spu_funcs.c
@@ -144,7 +144,7 @@ export_func(struct cell_spu_function_info *spu_functions,
 void
 return_function_info(void)
 {
-   struct cell_spu_function_info funcs ALIGN16_ATTRIB;
+   PIPE_ALIGN_VAR(16) struct cell_spu_function_info funcs;
    int tag = TAG_MISC;
 
    ASSERT(sizeof(funcs) == 256); /* must be multiple of 16 bytes */
diff --git a/src/gallium/drivers/cell/spu/spu_main.h b/src/gallium/drivers/cell/spu/spu_main.h
index 33767e7c51..b18f4c22ef 100644
--- a/src/gallium/drivers/cell/spu/spu_main.h
+++ b/src/gallium/drivers/cell/spu/spu_main.h
@@ -93,6 +93,7 @@ typedef vector unsigned int (*spu_fragment_program_func)(vector float *inputs,
                                                          vector float *constants);
 
 
+PIPE_ALIGN_TYPE(16,
 struct spu_framebuffer
 {
    void *color_start;              /**< addr of color surface in main memory */
@@ -107,10 +108,11 @@ struct spu_framebuffer
 
    uint zsize;                     /**< 0, 2 or 4 bytes per Z */
    float zscale;                   /**< 65535.0, 2^24-1 or 2^32-1 */
-} ALIGN16_ATTRIB;
+});
 
 
 /** per-texture level info */
+PIPE_ALIGN_TYPE(16,
 struct spu_texture_level
 {
    void *start;
@@ -123,20 +125,22 @@ struct spu_texture_level
    vector signed int mask_s, mask_t, mask_r;
    /** texcoord clamp limits */
    vector signed int max_s, max_t, max_r;
-} ALIGN16_ATTRIB;
+});
 
 
+PIPE_ALIGN_TYPE(16,
 struct spu_texture
 {
    struct spu_texture_level level[CELL_MAX_TEXTURE_LEVELS];
    uint max_level;
    uint target;  /**< PIPE_TEXTURE_x */
-} ALIGN16_ATTRIB;
+});
 
 
 /**
  * All SPU global/context state will be in a singleton object of this type:
  */
+PIPE_ALIGN_TYPE(16,
 struct spu_global
 {
    /** One-time init/constant info */
@@ -155,8 +159,8 @@ struct spu_global
    struct vertex_info vertex_info;
 
    /** Current color and Z tiles */
-   tile_t ctile ALIGN16_ATTRIB;
-   tile_t ztile ALIGN16_ATTRIB;
+   PIPE_ALIGN_VAR(16) tile_t ctile;
+   PIPE_ALIGN_VAR(16) tile_t ztile;
 
    /** Read depth/stencil tiles? */
    boolean read_depth_stencil;
@@ -165,8 +169,8 @@ struct spu_global
    ubyte cur_ctile_status, cur_ztile_status;
 
    /** Status of all tiles in framebuffer */
-   ubyte ctile_status[CELL_MAX_HEIGHT/TILE_SIZE][CELL_MAX_WIDTH/TILE_SIZE] ALIGN16_ATTRIB;
-   ubyte ztile_status[CELL_MAX_HEIGHT/TILE_SIZE][CELL_MAX_WIDTH/TILE_SIZE] ALIGN16_ATTRIB;
+   PIPE_ALIGN_VAR(16) ubyte ctile_status[CELL_MAX_HEIGHT/TILE_SIZE][CELL_MAX_WIDTH/TILE_SIZE];
+   PIPE_ALIGN_VAR(16) ubyte ztile_status[CELL_MAX_HEIGHT/TILE_SIZE][CELL_MAX_WIDTH/TILE_SIZE];
 
    /** Current fragment ops machine code, at 8-byte boundary */
    uint *fragment_ops_code;
@@ -175,7 +179,7 @@ struct spu_global
    spu_fragment_ops_func fragment_ops[2];
 
    /** Current fragment program machine code, at 8-byte boundary */
-   uint fragment_program_code[SPU_MAX_FRAGMENT_PROGRAM_INSTS] ALIGN8_ATTRIB;
+   PIPE_ALIGN_VAR(8) uint fragment_program_code[SPU_MAX_FRAGMENT_PROGRAM_INSTS];
    /** Current fragment ops function */
    spu_fragment_program_func fragment_program;
 
@@ -187,7 +191,7 @@ struct spu_global
    /** Fragment program constants */
    vector float constants[4 * CELL_MAX_CONSTANTS];
 
-} ALIGN16_ATTRIB;
+});
 
 
 extern struct spu_global spu;
diff --git a/src/gallium/drivers/cell/spu/spu_render.c b/src/gallium/drivers/cell/spu/spu_render.c
index 5ffb7073ab..14987e3c3a 100644
--- a/src/gallium/drivers/cell/spu/spu_render.c
+++ b/src/gallium/drivers/cell/spu/spu_render.c
@@ -169,7 +169,7 @@ void
 cmd_render(const struct cell_command_render *render, uint *pos_incr)
 {
    /* we'll DMA into these buffers */
-   ubyte vertex_data[CELL_BUFFER_SIZE] ALIGN16_ATTRIB;
+   PIPE_ALIGN_VAR(16) ubyte vertex_data[CELL_BUFFER_SIZE];
    const uint vertex_size = render->vertex_size; /* in bytes */
    /*const*/ uint total_vertex_bytes = render->num_verts * vertex_size;
    uint index_bytes;
diff --git a/src/gallium/drivers/cell/spu/spu_vertex_fetch.c b/src/gallium/drivers/cell/spu/spu_vertex_fetch.c
index 03375d84a5..087963960d 100644
--- a/src/gallium/drivers/cell/spu/spu_vertex_fetch.c
+++ b/src/gallium/drivers/cell/spu/spu_vertex_fetch.c
@@ -43,7 +43,8 @@ typedef void (*spu_fetch_func)(qword *out, const qword *in,
 			       const qword *shuffle_data);
 
 
-static const qword fetch_shuffle_data[5] ALIGN16_ATTRIB = {
+PIPE_ALIGN_VAR(16) static const qword
+fetch_shuffle_data[5] = {
    /* Shuffle used by CVT_64_FLOAT
     */
    {
@@ -110,7 +111,7 @@ static void generic_vertex_fetch(struct spu_vs_context *draw,
       unsigned idx;
       const unsigned bytes_per_entry = draw->vertex_fetch.size[attr];
       const unsigned quads_per_entry = (bytes_per_entry + 15) / 16;
-      qword in[2 * 4] ALIGN16_ATTRIB;
+      PIPE_ALIGN_VAR(16) qword in[2 * 4];
 
 
       /* Fetch four attributes for four vertices.  
diff --git a/src/gallium/drivers/cell/spu/spu_vertex_shader.c b/src/gallium/drivers/cell/spu/spu_vertex_shader.c
index fbe5b34d39..3e9804bf8e 100644
--- a/src/gallium/drivers/cell/spu/spu_vertex_shader.c
+++ b/src/gallium/drivers/cell/spu/spu_vertex_shader.c
@@ -107,8 +107,8 @@ run_vertex_program(struct spu_vs_context *draw,
    struct spu_exec_machine *machine = &draw->machine;
    unsigned int j;
 
-   ALIGN16_DECL(struct spu_exec_vector, inputs, PIPE_MAX_ATTRIBS);
-   ALIGN16_DECL(struct spu_exec_vector, outputs, PIPE_MAX_ATTRIBS);
+   PIPE_ALIGN_VAR(16) struct spu_exec_vector inputs[PIPE_MAX_ATTRIBS];
+   PIPE_ALIGN_VAR(16) struct spu_exec_vector outputs[PIPE_MAX_ATTRIBS];
    const float *scale = draw->viewport.scale;
    const float *trans = draw->viewport.translate;
 
@@ -119,8 +119,8 @@ run_vertex_program(struct spu_vs_context *draw,
    ASSERT_ALIGN16(draw->constants);
    machine->Consts = (float (*)[4]) draw->constants;
 
-   machine->Inputs = ALIGN16_ASSIGN(inputs);
-   machine->Outputs = ALIGN16_ASSIGN(outputs);
+   machine->Inputs = inputs;
+   machine->Outputs = outputs;
 
    spu_vertex_fetch( draw, machine, elts, count );
 
@@ -132,8 +132,9 @@ run_vertex_program(struct spu_vs_context *draw,
    for (j = 0; j < count; j++) {
       unsigned slot;
       float x, y, z, w;
+      PIPE_ALIGN_VAR(16)
       unsigned char buffer[sizeof(struct vertex_header)
-          + MAX_VERTEX_SIZE] ALIGN16_ATTRIB;
+          + MAX_VERTEX_SIZE];
       struct vertex_header *const tmpOut =
           (struct vertex_header *) buffer;
       const unsigned vert_size = ROUNDUP16(sizeof(struct vertex_header)
@@ -186,8 +187,8 @@ run_vertex_program(struct spu_vs_context *draw,
 }
 
 
-unsigned char immediates[(sizeof(float) * 4 * TGSI_EXEC_NUM_IMMEDIATES) + 32]
-    ALIGN16_ATTRIB;
+PIPE_ALIGN_VAR(16) unsigned char
+immediates[(sizeof(float) * 4 * TGSI_EXEC_NUM_IMMEDIATES) + 32]);
 
 
 void