static void r200FireEB(r200ContextPtr rmesa, int vertex_count, int type)
{
BATCH_LOCALS(&rmesa->radeon);
if (vertex_count > 0) {
BEGIN_BATCH(8+2);
OUT_BATCH_PACKET3_CLIP(R200_CP_CMD_3D_DRAW_INDX_2, 0);
OUT_BATCH(R200_VF_PRIM_WALK_IND |
R200_VF_COLOR_ORDER_RGBA |
((vertex_count + 0) << 16) |
type);
if (!rmesa->radeon.radeonScreen->kernel_mm) {
OUT_BATCH_PACKET3(R200_CP_CMD_INDX_BUFFER, 2);
OUT_BATCH((0x80 << 24) | (0 << 16) | 0x810);
OUT_BATCH_RELOC(rmesa->radeon.tcl.elt_dma_offset,
rmesa->radeon.tcl.elt_dma_bo,
rmesa->radeon.tcl.elt_dma_offset,
RADEON_GEM_DOMAIN_GTT, 0, 0);
OUT_BATCH((vertex_count + 1)/2);
} else {
OUT_BATCH_PACKET3(R200_CP_CMD_INDX_BUFFER, 2);
OUT_BATCH((0x80 << 24) | (0 << 16) | 0x810);
OUT_BATCH(rmesa->radeon.tcl.elt_dma_offset);
OUT_BATCH((vertex_count + 1)/2);
radeon_cs_write_reloc(rmesa->radeon.cmdbuf.cs,
rmesa->radeon.tcl.elt_dma_bo,
RADEON_GEM_DOMAIN_GTT, 0, 0);
}
END_BATCH();
}
}
static void r300EmitVbufPrim(r300ContextPtr rmesa, GLuint primitive, GLuint vertex_nr)
{
BATCH_LOCALS(&rmesa->radeon);
int type, num_verts;
if (RADEON_DEBUG & RADEON_VERTS)
fprintf(stderr, "%s\n", __func__);
type = r300PrimitiveType(rmesa, primitive);
num_verts = r300NumVerts(rmesa, vertex_nr, primitive);
BEGIN_BATCH(3);
OUT_BATCH_PACKET3(R300_PACKET3_3D_DRAW_VBUF_2, 0);
OUT_BATCH(R300_VAP_VF_CNTL__PRIM_WALK_VERTEX_LIST | (num_verts << 16) | type);
END_BATCH();
}
static void r300EmitVertexAOS(r300ContextPtr rmesa, GLuint vertex_size, struct radeon_bo *bo, GLuint offset)
{
BATCH_LOCALS(&rmesa->radeon);
radeon_print(RADEON_SWRENDER, RADEON_TRACE,
"%s: vertex_size %d, offset 0x%x \n",
__FUNCTION__, vertex_size, offset);
BEGIN_BATCH(7);
OUT_BATCH_PACKET3(R300_PACKET3_3D_LOAD_VBPNTR, 2);
OUT_BATCH(1);
OUT_BATCH(vertex_size | (vertex_size << 8));
OUT_BATCH_RELOC(offset, bo, offset, RADEON_GEM_DOMAIN_GTT, 0, 0);
END_BATCH();
}
void radeonEmitVertexAOS( r100ContextPtr rmesa,
GLuint vertex_size,
struct radeon_bo *bo,
GLuint offset )
{
#if RADEON_OLD_PACKETS
rmesa->ioctl.vertex_offset = offset;
rmesa->ioctl.bo = bo;
#else
BATCH_LOCALS(&rmesa->radeon);
if (RADEON_DEBUG & (RADEON_PRIMS|DEBUG_IOCTL))
fprintf(stderr, "%s: vertex_size 0x%x offset 0x%x \n",
__FUNCTION__, vertex_size, offset);
BEGIN_BATCH(7);
OUT_BATCH_PACKET3(RADEON_CP_PACKET3_3D_LOAD_VBPNTR, 2);
OUT_BATCH(1);
OUT_BATCH(vertex_size | (vertex_size << 8));
OUT_BATCH_RELOC(offset, bo, offset, RADEON_GEM_DOMAIN_GTT, 0, 0);
END_BATCH();
#endif
}
void radeonEmitAOS( r100ContextPtr rmesa,
GLuint nr,
GLuint offset )
{
#if RADEON_OLD_PACKETS
assert( nr == 1 );
rmesa->ioctl.bo = rmesa->radeon.tcl.aos[0].bo;
rmesa->ioctl.vertex_offset =
(rmesa->radeon.tcl.aos[0].offset + offset * rmesa->radeon.tcl.aos[0].stride * 4);
rmesa->ioctl.vertex_max = rmesa->radeon.tcl.aos[0].count;
#else
BATCH_LOCALS(&rmesa->radeon);
uint32_t voffset;
// int sz = AOS_BUFSZ(nr);
int sz = 1 + (nr >> 1) * 3 + (nr & 1) * 2;
int i;
if (RADEON_DEBUG & RADEON_IOCTL)
fprintf(stderr, "%s\n", __FUNCTION__);
BEGIN_BATCH(sz+2+(nr * 2));
OUT_BATCH_PACKET3(RADEON_CP_PACKET3_3D_LOAD_VBPNTR, sz - 1);
OUT_BATCH(nr);
{
for (i = 0; i + 1 < nr; i += 2) {
OUT_BATCH((rmesa->radeon.tcl.aos[i].components << 0) |
(rmesa->radeon.tcl.aos[i].stride << 8) |
(rmesa->radeon.tcl.aos[i + 1].components << 16) |
(rmesa->radeon.tcl.aos[i + 1].stride << 24));
voffset = rmesa->radeon.tcl.aos[i + 0].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 0].stride;
OUT_BATCH(voffset);
voffset = rmesa->radeon.tcl.aos[i + 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 1].stride;
OUT_BATCH(voffset);
}
if (nr & 1) {
OUT_BATCH((rmesa->radeon.tcl.aos[nr - 1].components << 0) |
(rmesa->radeon.tcl.aos[nr - 1].stride << 8));
voffset = rmesa->radeon.tcl.aos[nr - 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[nr - 1].stride;
OUT_BATCH(voffset);
}
for (i = 0; i + 1 < nr; i += 2) {
voffset = rmesa->radeon.tcl.aos[i + 0].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 0].stride;
radeon_cs_write_reloc(rmesa->radeon.cmdbuf.cs,
rmesa->radeon.tcl.aos[i+0].bo,
RADEON_GEM_DOMAIN_GTT,
0, 0);
voffset = rmesa->radeon.tcl.aos[i + 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 1].stride;
radeon_cs_write_reloc(rmesa->radeon.cmdbuf.cs,
rmesa->radeon.tcl.aos[i+1].bo,
RADEON_GEM_DOMAIN_GTT,
0, 0);
}
if (nr & 1) {
voffset = rmesa->radeon.tcl.aos[nr - 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[nr - 1].stride;
radeon_cs_write_reloc(rmesa->radeon.cmdbuf.cs,
rmesa->radeon.tcl.aos[nr-1].bo,
RADEON_GEM_DOMAIN_GTT,
0, 0);
}
}
END_BATCH();
#endif
}
void r200EmitAOS(r200ContextPtr rmesa, GLuint nr, GLuint offset)
{
BATCH_LOCALS(&rmesa->radeon);
uint32_t voffset;
int sz = 1 + (nr >> 1) * 3 + (nr & 1) * 2;
int i;
radeon_print(RADEON_RENDER, RADEON_VERBOSE,
"%s: nr=%d, ofs=0x%08x\n",
__FUNCTION__, nr, offset);
BEGIN_BATCH(sz+2+ (nr*2));
OUT_BATCH_PACKET3(R200_CP_CMD_3D_LOAD_VBPNTR, sz - 1);
OUT_BATCH(nr);
if (!rmesa->radeon.radeonScreen->kernel_mm) {
for (i = 0; i + 1 < nr; i += 2) {
OUT_BATCH((rmesa->radeon.tcl.aos[i].components << 0) |
(rmesa->radeon.tcl.aos[i].stride << 8) |
(rmesa->radeon.tcl.aos[i + 1].components << 16) |
(rmesa->radeon.tcl.aos[i + 1].stride << 24));
voffset = rmesa->radeon.tcl.aos[i + 0].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 0].stride;
OUT_BATCH_RELOC(voffset,
rmesa->radeon.tcl.aos[i].bo,
voffset,
RADEON_GEM_DOMAIN_GTT,
0, 0);
voffset = rmesa->radeon.tcl.aos[i + 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 1].stride;
OUT_BATCH_RELOC(voffset,
rmesa->radeon.tcl.aos[i+1].bo,
voffset,
RADEON_GEM_DOMAIN_GTT,
0, 0);
}
if (nr & 1) {
OUT_BATCH((rmesa->radeon.tcl.aos[nr - 1].components << 0) |
(rmesa->radeon.tcl.aos[nr - 1].stride << 8));
voffset = rmesa->radeon.tcl.aos[nr - 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[nr - 1].stride;
OUT_BATCH_RELOC(voffset,
rmesa->radeon.tcl.aos[nr - 1].bo,
voffset,
RADEON_GEM_DOMAIN_GTT,
0, 0);
}
} else {
for (i = 0; i + 1 < nr; i += 2) {
OUT_BATCH((rmesa->radeon.tcl.aos[i].components << 0) |
(rmesa->radeon.tcl.aos[i].stride << 8) |
(rmesa->radeon.tcl.aos[i + 1].components << 16) |
(rmesa->radeon.tcl.aos[i + 1].stride << 24));
voffset = rmesa->radeon.tcl.aos[i + 0].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 0].stride;
OUT_BATCH(voffset);
voffset = rmesa->radeon.tcl.aos[i + 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 1].stride;
OUT_BATCH(voffset);
}
if (nr & 1) {
OUT_BATCH((rmesa->radeon.tcl.aos[nr - 1].components << 0) |
(rmesa->radeon.tcl.aos[nr - 1].stride << 8));
voffset = rmesa->radeon.tcl.aos[nr - 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[nr - 1].stride;
OUT_BATCH(voffset);
}
for (i = 0; i + 1 < nr; i += 2) {
voffset = rmesa->radeon.tcl.aos[i + 0].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 0].stride;
radeon_cs_write_reloc(rmesa->radeon.cmdbuf.cs,
rmesa->radeon.tcl.aos[i+0].bo,
RADEON_GEM_DOMAIN_GTT,
0, 0);
voffset = rmesa->radeon.tcl.aos[i + 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[i + 1].stride;
radeon_cs_write_reloc(rmesa->radeon.cmdbuf.cs,
rmesa->radeon.tcl.aos[i+1].bo,
RADEON_GEM_DOMAIN_GTT,
0, 0);
}
if (nr & 1) {
voffset = rmesa->radeon.tcl.aos[nr - 1].offset +
offset * 4 * rmesa->radeon.tcl.aos[nr - 1].stride;
radeon_cs_write_reloc(rmesa->radeon.cmdbuf.cs,
rmesa->radeon.tcl.aos[nr-1].bo,
RADEON_GEM_DOMAIN_GTT,
0, 0);
}
}
END_BATCH();
}