summaryrefslogtreecommitdiff
path: root/src/gallium/drivers/nv40
diff options
context:
space:
mode:
Diffstat (limited to 'src/gallium/drivers/nv40')
-rw-r--r--src/gallium/drivers/nv40/nv40_context.c9
-rw-r--r--src/gallium/drivers/nv40/nv40_context.h2
-rw-r--r--src/gallium/drivers/nv40/nv40_draw.c5
-rw-r--r--src/gallium/drivers/nv40/nv40_fragprog.c130
-rw-r--r--src/gallium/drivers/nv40/nv40_fragtex.c6
-rw-r--r--src/gallium/drivers/nv40/nv40_miptree.c65
-rw-r--r--src/gallium/drivers/nv40/nv40_screen.c7
-rw-r--r--src/gallium/drivers/nv40/nv40_state.c15
-rw-r--r--src/gallium/drivers/nv40/nv40_state.h1
-rw-r--r--src/gallium/drivers/nv40/nv40_state_emit.c11
-rw-r--r--src/gallium/drivers/nv40/nv40_transfer.c39
-rw-r--r--src/gallium/drivers/nv40/nv40_vbo.c5
-rw-r--r--src/gallium/drivers/nv40/nv40_vertprog.c92
13 files changed, 220 insertions, 167 deletions
diff --git a/src/gallium/drivers/nv40/nv40_context.c b/src/gallium/drivers/nv40/nv40_context.c
index 7f008274a4..d56c7a6b49 100644
--- a/src/gallium/drivers/nv40/nv40_context.c
+++ b/src/gallium/drivers/nv40/nv40_context.c
@@ -25,6 +25,12 @@ static void
nv40_destroy(struct pipe_context *pipe)
{
struct nv40_context *nv40 = nv40_context(pipe);
+ unsigned i;
+
+ for (i = 0; i < NV40_STATE_MAX; i++) {
+ if (nv40->state.hw[i])
+ so_ref(NULL, &nv40->state.hw[i]);
+ }
if (nv40->draw)
draw_destroy(nv40->draw);
@@ -58,6 +64,9 @@ nv40_create(struct pipe_screen *pscreen, unsigned pctx_id)
nv40->pipe.is_texture_referenced = nouveau_is_texture_referenced;
nv40->pipe.is_buffer_referenced = nouveau_is_buffer_referenced;
+ screen->base.channel->user_private = nv40;
+ screen->base.channel->flush_notify = nv40_state_flush_notify;
+
nv40_init_query_functions(nv40);
nv40_init_surface_functions(nv40);
nv40_init_state_functions(nv40);
diff --git a/src/gallium/drivers/nv40/nv40_context.h b/src/gallium/drivers/nv40/nv40_context.h
index a3d594167a..83fcf1785d 100644
--- a/src/gallium/drivers/nv40/nv40_context.h
+++ b/src/gallium/drivers/nv40/nv40_context.h
@@ -159,7 +159,6 @@ struct nv40_context {
unsigned vtxbuf_nr;
struct pipe_vertex_element vtxelt[PIPE_MAX_ATTRIBS];
unsigned vtxelt_nr;
- const unsigned *edgeflags;
};
static INLINE struct nv40_context *
@@ -204,6 +203,7 @@ extern void nv40_fragtex_bind(struct nv40_context *);
extern boolean nv40_state_validate(struct nv40_context *nv40);
extern boolean nv40_state_validate_swtnl(struct nv40_context *nv40);
extern void nv40_state_emit(struct nv40_context *nv40);
+extern void nv40_state_flush_notify(struct nouveau_channel *chan);
extern struct nv40_state_entry nv40_state_rasterizer;
extern struct nv40_state_entry nv40_state_scissor;
extern struct nv40_state_entry nv40_state_stipple;
diff --git a/src/gallium/drivers/nv40/nv40_draw.c b/src/gallium/drivers/nv40/nv40_draw.c
index b2f19ecb69..3875bc3545 100644
--- a/src/gallium/drivers/nv40/nv40_draw.c
+++ b/src/gallium/drivers/nv40/nv40_draw.c
@@ -261,7 +261,8 @@ nv40_draw_elements_swtnl(struct pipe_context *pipe,
map = pipe_buffer_map(pscreen,
nv40->constbuf[PIPE_SHADER_VERTEX],
PIPE_BUFFER_USAGE_CPU_READ);
- draw_set_mapped_constant_buffer(nv40->draw, map, nr);
+ draw_set_mapped_constant_buffer(nv40->draw, PIPE_SHADER_VERTEX,
+ map, nr);
}
draw_arrays(nv40->draw, mode, start, count);
@@ -285,7 +286,7 @@ static INLINE void
emit_attrib(struct nv40_context *nv40, unsigned hw, unsigned emit,
unsigned semantic, unsigned index)
{
- unsigned draw_out = draw_find_vs_output(nv40->draw, semantic, index);
+ unsigned draw_out = draw_find_shader_output(nv40->draw, semantic, index);
unsigned a = nv40->swtnl.nr_attribs++;
nv40->swtnl.hw[a] = hw;
diff --git a/src/gallium/drivers/nv40/nv40_fragprog.c b/src/gallium/drivers/nv40/nv40_fragprog.c
index 99277506fc..bb9c85cc43 100644
--- a/src/gallium/drivers/nv40/nv40_fragprog.c
+++ b/src/gallium/drivers/nv40/nv40_fragprog.c
@@ -149,7 +149,7 @@ emit_src(struct nv40_fpc *fpc, int pos, struct nv40_sreg src)
sizeof(uint32_t) * 4);
}
- sr |= (NV40_FP_REG_TYPE_CONST << NV40_FP_REG_TYPE_SHIFT);
+ sr |= (NV40_FP_REG_TYPE_CONST << NV40_FP_REG_TYPE_SHIFT);
break;
case NV40SR_NONE:
sr |= (NV40_FP_REG_TYPE_INPUT << NV40_FP_REG_TYPE_SHIFT);
@@ -255,50 +255,50 @@ tgsi_src(struct nv40_fpc *fpc, const struct tgsi_full_src_register *fsrc)
{
struct nv40_sreg src;
- switch (fsrc->SrcRegister.File) {
+ switch (fsrc->Register.File) {
case TGSI_FILE_INPUT:
src = nv40_sr(NV40SR_INPUT,
- fpc->attrib_map[fsrc->SrcRegister.Index]);
+ fpc->attrib_map[fsrc->Register.Index]);
break;
case TGSI_FILE_CONSTANT:
- src = constant(fpc, fsrc->SrcRegister.Index, NULL);
+ src = constant(fpc, fsrc->Register.Index, NULL);
break;
case TGSI_FILE_IMMEDIATE:
- assert(fsrc->SrcRegister.Index < fpc->nr_imm);
- src = fpc->imm[fsrc->SrcRegister.Index];
+ assert(fsrc->Register.Index < fpc->nr_imm);
+ src = fpc->imm[fsrc->Register.Index];
break;
case TGSI_FILE_TEMPORARY:
- src = fpc->r_temp[fsrc->SrcRegister.Index];
+ src = fpc->r_temp[fsrc->Register.Index];
break;
/* NV40 fragprog result regs are just temps, so this is simple */
case TGSI_FILE_OUTPUT:
- src = fpc->r_result[fsrc->SrcRegister.Index];
+ src = fpc->r_result[fsrc->Register.Index];
break;
default:
NOUVEAU_ERR("bad src file\n");
break;
}
- src.abs = fsrc->SrcRegisterExtMod.Absolute;
- src.negate = fsrc->SrcRegister.Negate;
- src.swz[0] = fsrc->SrcRegister.SwizzleX;
- src.swz[1] = fsrc->SrcRegister.SwizzleY;
- src.swz[2] = fsrc->SrcRegister.SwizzleZ;
- src.swz[3] = fsrc->SrcRegister.SwizzleW;
+ src.abs = fsrc->Register.Absolute;
+ src.negate = fsrc->Register.Negate;
+ src.swz[0] = fsrc->Register.SwizzleX;
+ src.swz[1] = fsrc->Register.SwizzleY;
+ src.swz[2] = fsrc->Register.SwizzleZ;
+ src.swz[3] = fsrc->Register.SwizzleW;
return src;
}
static INLINE struct nv40_sreg
tgsi_dst(struct nv40_fpc *fpc, const struct tgsi_full_dst_register *fdst) {
- switch (fdst->DstRegister.File) {
+ switch (fdst->Register.File) {
case TGSI_FILE_OUTPUT:
- return fpc->r_result[fdst->DstRegister.Index];
+ return fpc->r_result[fdst->Register.Index];
case TGSI_FILE_TEMPORARY:
- return fpc->r_temp[fdst->DstRegister.Index];
+ return fpc->r_temp[fdst->Register.Index];
case TGSI_FILE_NULL:
return nv40_sr(NV40SR_NONE, 0);
default:
- NOUVEAU_ERR("bad dst file %d\n", fdst->DstRegister.File);
+ NOUVEAU_ERR("bad dst file %d\n", fdst->Register.File);
return nv40_sr(NV40SR_NONE, 0);
}
}
@@ -364,8 +364,8 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
const struct tgsi_full_src_register *fsrc;
- fsrc = &finst->FullSrcRegisters[i];
- if (fsrc->SrcRegister.File == TGSI_FILE_TEMPORARY) {
+ fsrc = &finst->Src[i];
+ if (fsrc->Register.File == TGSI_FILE_TEMPORARY) {
src[i] = tgsi_src(fpc, fsrc);
}
}
@@ -373,9 +373,9 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
const struct tgsi_full_src_register *fsrc;
- fsrc = &finst->FullSrcRegisters[i];
+ fsrc = &finst->Src[i];
- switch (fsrc->SrcRegister.File) {
+ switch (fsrc->Register.File) {
case TGSI_FILE_INPUT:
case TGSI_FILE_CONSTANT:
case TGSI_FILE_TEMPORARY:
@@ -386,10 +386,10 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
break;
}
- switch (fsrc->SrcRegister.File) {
+ switch (fsrc->Register.File) {
case TGSI_FILE_INPUT:
- if (ai == -1 || ai == fsrc->SrcRegister.Index) {
- ai = fsrc->SrcRegister.Index;
+ if (ai == -1 || ai == fsrc->Register.Index) {
+ ai = fsrc->Register.Index;
src[i] = tgsi_src(fpc, fsrc);
} else {
src[i] = temp(fpc);
@@ -399,8 +399,8 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
break;
case TGSI_FILE_CONSTANT:
if ((ci == -1 && ii == -1) ||
- ci == fsrc->SrcRegister.Index) {
- ci = fsrc->SrcRegister.Index;
+ ci == fsrc->Register.Index) {
+ ci = fsrc->Register.Index;
src[i] = tgsi_src(fpc, fsrc);
} else {
src[i] = temp(fpc);
@@ -410,8 +410,8 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
break;
case TGSI_FILE_IMMEDIATE:
if ((ci == -1 && ii == -1) ||
- ii == fsrc->SrcRegister.Index) {
- ii = fsrc->SrcRegister.Index;
+ ii == fsrc->Register.Index) {
+ ii = fsrc->Register.Index;
src[i] = tgsi_src(fpc, fsrc);
} else {
src[i] = temp(fpc);
@@ -423,7 +423,7 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
/* handled above */
break;
case TGSI_FILE_SAMPLER:
- unit = fsrc->SrcRegister.Index;
+ unit = fsrc->Register.Index;
break;
case TGSI_FILE_OUTPUT:
break;
@@ -433,8 +433,8 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
}
}
- dst = tgsi_dst(fpc, &finst->FullDstRegisters[0]);
- mask = tgsi_mask(finst->FullDstRegisters[0].DstRegister.WriteMask);
+ dst = tgsi_dst(fpc, &finst->Dst[0]);
+ mask = tgsi_mask(finst->Dst[0].Register.WriteMask);
sat = (finst->Instruction.Saturate == TGSI_SAT_ZERO_ONE);
switch (finst->Instruction.Opcode) {
@@ -445,10 +445,11 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
arith(fpc, sat, ADD, dst, mask, src[0], src[1], none);
break;
case TGSI_OPCODE_CMP:
- tmp = temp(fpc);
- arith(fpc, sat, MOV, dst, mask, src[2], none, none);
+ tmp = nv40_sr(NV40SR_NONE, 0);
tmp.cc_update = 1;
arith(fpc, 0, MOV, tmp, 0xf, src[0], none, none);
+ dst.cc_test = NV40_VP_INST_COND_GE;
+ arith(fpc, sat, MOV, dst, mask, src[2], none, none);
dst.cc_test = NV40_VP_INST_COND_LT;
arith(fpc, sat, MOV, dst, mask, src[1], none, none);
break;
@@ -573,13 +574,28 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
neg(swz(tmp, X, X, X, X)), none, none);
break;
case TGSI_OPCODE_SCS:
- if (mask & MASK_X) {
- arith(fpc, sat, COS, dst, MASK_X,
- swz(src[0], X, X, X, X), none, none);
+ /* avoid overwriting the source */
+ if(src[0].swz[SWZ_X] != SWZ_X)
+ {
+ if (mask & MASK_X) {
+ arith(fpc, sat, COS, dst, MASK_X,
+ swz(src[0], X, X, X, X), none, none);
+ }
+ if (mask & MASK_Y) {
+ arith(fpc, sat, SIN, dst, MASK_Y,
+ swz(src[0], X, X, X, X), none, none);
+ }
}
- if (mask & MASK_Y) {
- arith(fpc, sat, SIN, dst, MASK_Y,
- swz(src[0], X, X, X, X), none, none);
+ else
+ {
+ if (mask & MASK_Y) {
+ arith(fpc, sat, SIN, dst, MASK_Y,
+ swz(src[0], X, X, X, X), none, none);
+ }
+ if (mask & MASK_X) {
+ arith(fpc, sat, COS, dst, MASK_X,
+ swz(src[0], X, X, X, X), none, none);
+ }
}
break;
case TGSI_OPCODE_SEQ:
@@ -644,15 +660,15 @@ nv40_fragprog_parse_decl_attrib(struct nv40_fpc *fpc,
{
int hw;
- switch (fdec->Semantic.SemanticName) {
+ switch (fdec->Semantic.Name) {
case TGSI_SEMANTIC_POSITION:
hw = NV40_FP_OP_INPUT_SRC_POSITION;
break;
case TGSI_SEMANTIC_COLOR:
- if (fdec->Semantic.SemanticIndex == 0) {
+ if (fdec->Semantic.Index == 0) {
hw = NV40_FP_OP_INPUT_SRC_COL0;
} else
- if (fdec->Semantic.SemanticIndex == 1) {
+ if (fdec->Semantic.Index == 1) {
hw = NV40_FP_OP_INPUT_SRC_COL1;
} else {
NOUVEAU_ERR("bad colour semantic index\n");
@@ -663,9 +679,9 @@ nv40_fragprog_parse_decl_attrib(struct nv40_fpc *fpc,
hw = NV40_FP_OP_INPUT_SRC_FOGC;
break;
case TGSI_SEMANTIC_GENERIC:
- if (fdec->Semantic.SemanticIndex <= 7) {
+ if (fdec->Semantic.Index <= 7) {
hw = NV40_FP_OP_INPUT_SRC_TC(fdec->Semantic.
- SemanticIndex);
+ Index);
} else {
NOUVEAU_ERR("bad generic semantic index\n");
return FALSE;
@@ -676,7 +692,7 @@ nv40_fragprog_parse_decl_attrib(struct nv40_fpc *fpc,
return FALSE;
}
- fpc->attrib_map[fdec->DeclarationRange.First] = hw;
+ fpc->attrib_map[fdec->Range.First] = hw;
return TRUE;
}
@@ -684,15 +700,15 @@ static boolean
nv40_fragprog_parse_decl_output(struct nv40_fpc *fpc,
const struct tgsi_full_declaration *fdec)
{
- unsigned idx = fdec->DeclarationRange.First;
+ unsigned idx = fdec->Range.First;
unsigned hw;
- switch (fdec->Semantic.SemanticName) {
+ switch (fdec->Semantic.Name) {
case TGSI_SEMANTIC_POSITION:
hw = 1;
break;
case TGSI_SEMANTIC_COLOR:
- switch (fdec->Semantic.SemanticIndex) {
+ switch (fdec->Semantic.Index) {
case 0: hw = 0; break;
case 1: hw = 2; break;
case 2: hw = 3; break;
@@ -738,9 +754,9 @@ nv40_fragprog_prepare(struct nv40_fpc *fpc)
goto out_err;
break;
case TGSI_FILE_TEMPORARY:
- if (fdec->DeclarationRange.Last > high_temp) {
+ if (fdec->Range.Last > high_temp) {
high_temp =
- fdec->DeclarationRange.Last;
+ fdec->Range.Last;
}
break;
default:
@@ -752,7 +768,7 @@ nv40_fragprog_prepare(struct nv40_fpc *fpc)
{
struct tgsi_full_immediate *imm;
float vals[4];
-
+
imm = &p.FullToken.FullImmediate;
assert(imm->Immediate.DataType == TGSI_IMM_FLOAT32);
assert(fpc->nr_imm < MAX_IMM);
@@ -836,7 +852,7 @@ nv40_fragprog_translate(struct nv40_context *nv40,
fp->insn[fpc->inst_offset + 1] = 0x00000000;
fp->insn[fpc->inst_offset + 2] = 0x00000000;
fp->insn[fpc->inst_offset + 3] = 0x00000000;
-
+
fp->translated = TRUE;
out_err:
tgsi_parse_free(&parse);
@@ -917,7 +933,7 @@ nv40_fragprog_validate(struct nv40_context *nv40)
update_constants:
if (fp->nr_consts) {
float *map;
-
+
map = pipe_buffer_map(pscreen, constbuf,
PIPE_BUFFER_USAGE_CPU_READ);
for (i = 0; i < fp->nr_consts; i++) {
@@ -948,6 +964,12 @@ void
nv40_fragprog_destroy(struct nv40_context *nv40,
struct nv40_fragment_program *fp)
{
+ if (fp->buffer)
+ pipe_buffer_reference(&fp->buffer, NULL);
+
+ if (fp->so)
+ so_ref(NULL, &fp->so);
+
if (fp->insn_len)
FREE(fp->insn);
}
diff --git a/src/gallium/drivers/nv40/nv40_fragtex.c b/src/gallium/drivers/nv40/nv40_fragtex.c
index e2ec57564d..44abc84596 100644
--- a/src/gallium/drivers/nv40/nv40_fragtex.c
+++ b/src/gallium/drivers/nv40/nv40_fragtex.c
@@ -117,11 +117,11 @@ nv40_fragtex_build(struct nv40_context *nv40, int unit)
so_data (so, NV40TCL_TEX_ENABLE_ENABLE | ps->en);
so_data (so, txs);
so_data (so, ps->filt | tf->sign | 0x2000 /*voodoo*/);
- so_data (so, (pt->width[0] << NV40TCL_TEX_SIZE0_W_SHIFT) |
- pt->height[0]);
+ so_data (so, (pt->width0 << NV40TCL_TEX_SIZE0_W_SHIFT) |
+ pt->height0);
so_data (so, ps->bcol);
so_method(so, nv40->screen->curie, NV40TCL_TEX_SIZE1(unit), 1);
- so_data (so, (pt->depth[0] << NV40TCL_TEX_SIZE1_DEPTH_SHIFT) | txp);
+ so_data (so, (pt->depth0 << NV40TCL_TEX_SIZE1_DEPTH_SHIFT) | txp);
return so;
}
diff --git a/src/gallium/drivers/nv40/nv40_miptree.c b/src/gallium/drivers/nv40/nv40_miptree.c
index 465dd3b069..89bd155ff4 100644
--- a/src/gallium/drivers/nv40/nv40_miptree.c
+++ b/src/gallium/drivers/nv40/nv40_miptree.c
@@ -1,14 +1,19 @@
#include "pipe/p_state.h"
#include "pipe/p_defines.h"
#include "pipe/p_inlines.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
#include "nv40_context.h"
+#include "../nv04/nv04_surface_2d.h"
+
+
static void
nv40_miptree_layout(struct nv40_miptree *mt)
{
struct pipe_texture *pt = &mt->base;
- uint width = pt->width[0], height = pt->height[0], depth = pt->depth[0];
+ uint width = pt->width0;
uint offset = 0;
int nr_faces, l, f;
uint wide_pitch = pt->tex_usage & (PIPE_TEXTURE_USAGE_SAMPLER |
@@ -21,29 +26,21 @@ nv40_miptree_layout(struct nv40_miptree *mt)
nr_faces = 6;
} else
if (pt->target == PIPE_TEXTURE_3D) {
- nr_faces = pt->depth[0];
+ nr_faces = pt->depth0;
} else {
nr_faces = 1;
}
for (l = 0; l <= pt->last_level; l++) {
- pt->width[l] = width;
- pt->height[l] = height;
- pt->depth[l] = depth;
- pt->nblocksx[l] = pf_get_nblocksx(&pt->block, width);
- pt->nblocksy[l] = pf_get_nblocksy(&pt->block, height);
-
if (wide_pitch && (pt->tex_usage & NOUVEAU_TEXTURE_USAGE_LINEAR))
- mt->level[l].pitch = align(pt->width[0] * pt->block.size, 64);
+ mt->level[l].pitch = align(util_format_get_stride(pt->format, pt->width0), 64);
else
- mt->level[l].pitch = pt->width[l] * pt->block.size;
+ mt->level[l].pitch = util_format_get_stride(pt->format, width);
mt->level[l].image_offset =
CALLOC(nr_faces, sizeof(unsigned));
- width = MAX2(1, width >> 1);
- height = MAX2(1, height >> 1);
- depth = MAX2(1, depth >> 1);
+ width = u_minify(width, 1);
}
for (f = 0; f < nr_faces; f++) {
@@ -51,14 +48,14 @@ nv40_miptree_layout(struct nv40_miptree *mt)
mt->level[l].image_offset[f] = offset;
if (!(pt->tex_usage & NOUVEAU_TEXTURE_USAGE_LINEAR) &&
- pt->width[l + 1] > 1 && pt->height[l + 1] > 1)
- offset += align(mt->level[l].pitch * pt->height[l], 64);
+ u_minify(pt->width0, l + 1) > 1 && u_minify(pt->height0, l + 1) > 1)
+ offset += align(mt->level[l].pitch * u_minify(pt->height0, l), 64);
else
- offset += mt->level[l].pitch * pt->height[l];
+ offset += mt->level[l].pitch * u_minify(pt->height0, l);
}
mt->level[l].image_offset[f] = offset;
- offset += mt->level[l].pitch * pt->height[l];
+ offset += mt->level[l].pitch * u_minify(pt->height0, l);
}
mt->total_size = offset;
@@ -79,8 +76,8 @@ nv40_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *pt)
mt->base.screen = pscreen;
/* Swizzled textures must be POT */
- if (pt->width[0] & (pt->width[0] - 1) ||
- pt->height[0] & (pt->height[0] - 1))
+ if (pt->width0 & (pt->width0 - 1) ||
+ pt->height0 & (pt->height0 - 1))
mt->base.tex_usage |= NOUVEAU_TEXTURE_USAGE_LINEAR;
else
if (pt->tex_usage & (PIPE_TEXTURE_USAGE_PRIMARY |
@@ -109,6 +106,12 @@ nv40_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *pt)
if (pt->tex_usage & PIPE_TEXTURE_USAGE_DYNAMIC)
buf_usage |= PIPE_BUFFER_USAGE_CPU_READ_WRITE;
+ /* apparently we can't render to swizzled surfaces smaller than 64 bytes, so make them linear.
+ * If the user did not ask for a render target, they can still render to it, but it will cost them an extra copy.
+ * This also happens for small mipmaps of large textures. */
+ if (pt->tex_usage & PIPE_TEXTURE_USAGE_RENDER_TARGET && util_format_get_stride(pt->format, pt->width0) < 64)
+ mt->base.tex_usage |= NOUVEAU_TEXTURE_USAGE_LINEAR;
+
nv40_miptree_layout(mt);
mt->buffer = pscreen->buffer_create(pscreen, 256, buf_usage, mt->total_size);
@@ -116,7 +119,7 @@ nv40_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *pt)
FREE(mt);
return NULL;
}
-
+ mt->bo = nouveau_bo(mt->buffer);
return &mt->base;
}
@@ -128,7 +131,7 @@ nv40_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
/* Only supports 2D, non-mipmapped textures for the moment */
if (pt->target != PIPE_TEXTURE_2D || pt->last_level != 0 ||
- pt->depth[0] != 1)
+ pt->depth0 != 1)
return NULL;
mt = CALLOC_STRUCT(nv40_miptree);
@@ -145,6 +148,7 @@ nv40_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
mt->base.tex_usage |= NOUVEAU_TEXTURE_USAGE_LINEAR;
pipe_buffer_reference(&mt->buffer, pb);
+ mt->bo = nouveau_bo(mt->buffer);
return &mt->base;
}
@@ -176,8 +180,8 @@ nv40_miptree_surface_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
return NULL;
pipe_texture_reference(&ns->base.texture, pt);
ns->base.format = pt->format;
- ns->base.width = pt->width[level];
- ns->base.height = pt->height[level];
+ ns->base.width = u_minify(pt->width0, level);
+ ns->base.height = u_minify(pt->height0, level);
ns->base.usage = flags;
pipe_reference_init(&ns->base.reference, 1);
ns->base.face = face;
@@ -194,12 +198,27 @@ nv40_miptree_surface_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
ns->base.offset = mt->level[level].image_offset[0];
}
+ /* create a linear temporary that we can render into if necessary.
+ * Note that ns->pitch is always a multiple of 64 for linear surfaces and swizzled surfaces are POT, so
+ * ns->pitch & 63 is equivalent to (ns->pitch < 64 && swizzled)*/
+ if((ns->pitch & 63) && (ns->base.usage & (PIPE_BUFFER_USAGE_GPU_WRITE | NOUVEAU_BUFFER_USAGE_NO_RENDER)) == PIPE_BUFFER_USAGE_GPU_WRITE)
+ return &nv04_surface_wrap_for_render(pscreen, ((struct nv40_screen*)pscreen)->eng2d, ns)->base;
+
return &ns->base;
}
static void
nv40_miptree_surface_del(struct pipe_surface *ps)
{
+ struct nv04_surface* ns = (struct nv04_surface*)ps;
+ if(ns->backing)
+ {
+ struct nv40_screen* screen = (struct nv40_screen*)ps->texture->screen;
+ if(ns->backing->base.usage & PIPE_BUFFER_USAGE_GPU_WRITE)
+ screen->eng2d->copy(screen->eng2d, &ns->backing->base, 0, 0, ps, 0, 0, ns->base.width, ns->base.height);
+ nv40_miptree_surface_del(&ns->backing->base);
+ }
+
pipe_texture_reference(&ps->texture, NULL);
FREE(ps);
}
diff --git a/src/gallium/drivers/nv40/nv40_screen.c b/src/gallium/drivers/nv40/nv40_screen.c
index bd13dfddd1..d01e712805 100644
--- a/src/gallium/drivers/nv40/nv40_screen.c
+++ b/src/gallium/drivers/nv40/nv40_screen.c
@@ -140,6 +140,12 @@ static void
nv40_screen_destroy(struct pipe_screen *pscreen)
{
struct nv40_screen *screen = nv40_screen(pscreen);
+ unsigned i;
+
+ for (i = 0; i < NV40_STATE_MAX; i++) {
+ if (screen->state[i])
+ so_ref(NULL, &screen->state[i]);
+ }
nouveau_resource_free(&screen->vp_exec_heap);
nouveau_resource_free(&screen->vp_data_heap);
@@ -147,6 +153,7 @@ nv40_screen_destroy(struct pipe_screen *pscreen)
nouveau_notifier_free(&screen->query);
nouveau_notifier_free(&screen->sync);
nouveau_grobj_free(&screen->curie);
+ nv04_surface_2d_takedown(&screen->eng2d);
nouveau_screen_fini(&screen->base);
diff --git a/src/gallium/drivers/nv40/nv40_state.c b/src/gallium/drivers/nv40/nv40_state.c
index c3ee4d2345..ed55d29aff 100644
--- a/src/gallium/drivers/nv40/nv40_state.c
+++ b/src/gallium/drivers/nv40/nv40_state.c
@@ -687,16 +687,6 @@ nv40_set_vertex_elements(struct pipe_context *pipe, unsigned count,
nv40->draw_dirty |= NV40_NEW_ARRAYS;
}
-static void
-nv40_set_edgeflags(struct pipe_context *pipe, const unsigned *bitfield)
-{
- struct nv40_context *nv40 = nv40_context(pipe);
-
- nv40->edgeflags = bitfield;
- nv40->dirty |= NV40_NEW_ARRAYS;
- nv40->draw_dirty |= NV40_NEW_ARRAYS;
-}
-
void
nv40_init_state_functions(struct nv40_context *nv40)
{
@@ -705,9 +695,9 @@ nv40_init_state_functions(struct nv40_context *nv40)
nv40->pipe.delete_blend_state = nv40_blend_state_delete;
nv40->pipe.create_sampler_state = nv40_sampler_state_create;
- nv40->pipe.bind_sampler_states = nv40_sampler_state_bind;
+ nv40->pipe.bind_fragment_sampler_states = nv40_sampler_state_bind;
nv40->pipe.delete_sampler_state = nv40_sampler_state_delete;
- nv40->pipe.set_sampler_textures = nv40_set_sampler_texture;
+ nv40->pipe.set_fragment_sampler_textures = nv40_set_sampler_texture;
nv40->pipe.create_rasterizer_state = nv40_rasterizer_state_create;
nv40->pipe.bind_rasterizer_state = nv40_rasterizer_state_bind;
@@ -736,7 +726,6 @@ nv40_init_state_functions(struct nv40_context *nv40)
nv40->pipe.set_scissor_state = nv40_set_scissor_state;
nv40->pipe.set_viewport_state = nv40_set_viewport_state;
- nv40->pipe.set_edgeflags = nv40_set_edgeflags;
nv40->pipe.set_vertex_buffers = nv40_set_vertex_buffers;
nv40->pipe.set_vertex_elements = nv40_set_vertex_elements;
}
diff --git a/src/gallium/drivers/nv40/nv40_state.h b/src/gallium/drivers/nv40/nv40_state.h
index 8a9d8c8fdf..192074e747 100644
--- a/src/gallium/drivers/nv40/nv40_state.h
+++ b/src/gallium/drivers/nv40/nv40_state.h
@@ -75,6 +75,7 @@ struct nv40_fragment_program {
struct nv40_miptree {
struct pipe_texture base;
+ struct nouveau_bo *bo;
struct pipe_buffer *buffer;
uint total_size;
diff --git a/src/gallium/drivers/nv40/nv40_state_emit.c b/src/gallium/drivers/nv40/nv40_state_emit.c
index 198692965d..789ed16126 100644
--- a/src/gallium/drivers/nv40/nv40_state_emit.c
+++ b/src/gallium/drivers/nv40/nv40_state_emit.c
@@ -57,7 +57,7 @@ nv40_state_emit(struct nv40_context *nv40)
struct nouveau_channel *chan = nv40->screen->base.channel;
struct nv40_state *state = &nv40->state;
struct nv40_screen *screen = nv40->screen;
- unsigned i, samplers;
+ unsigned i;
uint64_t states;
if (nv40->pctx_id != screen->cur_pctx) {
@@ -87,6 +87,14 @@ nv40_state_emit(struct nv40_context *nv40)
}
state->dirty = 0;
+}
+
+void
+nv40_state_flush_notify(struct nouveau_channel *chan)
+{
+ struct nv40_context *nv40 = chan->user_private;
+ struct nv40_state *state = &nv40->state;
+ unsigned i, samplers;
so_emit_reloc_markers(chan, state->hw[NV40_STATE_FB]);
for (i = 0, samplers = state->fp_samplers; i < 16 && samplers; i++) {
@@ -160,7 +168,6 @@ nv40_state_validate_swtnl(struct nv40_context *nv40)
draw_set_viewport_state(draw, &nv40->viewport);
if (nv40->draw_dirty & NV40_NEW_ARRAYS) {
- draw_set_edgeflags(draw, nv40->edgeflags);
draw_set_vertex_buffers(draw, nv40->vtxbuf_nr, nv40->vtxbuf);
draw_set_vertex_elements(draw, nv40->vtxelt_nr, nv40->vtxelt);
}
diff --git a/src/gallium/drivers/nv40/nv40_transfer.c b/src/gallium/drivers/nv40/nv40_transfer.c
index 92caee6f38..791ee6823d 100644
--- a/src/gallium/drivers/nv40/nv40_transfer.c
+++ b/src/gallium/drivers/nv40/nv40_transfer.c
@@ -1,7 +1,9 @@
#include <pipe/p_state.h>
#include <pipe/p_defines.h>
#include <pipe/p_inlines.h>
+#include <util/u_format.h>
#include <util/u_memory.h>
+#include <util/u_math.h>
#include <nouveau/nouveau_winsys.h>
#include "nv40_context.h"
#include "nv40_screen.h"
@@ -10,22 +12,19 @@
struct nv40_transfer {
struct pipe_transfer base;
struct pipe_surface *surface;
- bool direct;
+ boolean direct;
};
static void
-nv40_compatible_transfer_tex(struct pipe_texture *pt, unsigned level,
+nv40_compatible_transfer_tex(struct pipe_texture *pt, unsigned width, unsigned height,
struct pipe_texture *template)
{
memset(template, 0, sizeof(struct pipe_texture));
template->target = pt->target;
template->format = pt->format;
- template->width[0] = pt->width[level];
- template->height[0] = pt->height[level];
- template->depth[0] = 1;
- template->block = pt->block;
- template->nblocksx[0] = pt->nblocksx[level];
- template->nblocksy[0] = pt->nblocksx[level];
+ template->width0 = width;
+ template->height0 = height;
+ template->depth0 = 1;
template->last_level = 0;
template->nr_samples = pt->nr_samples;
@@ -48,14 +47,10 @@ nv40_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
return NULL;
pipe_texture_reference(&tx->base.texture, pt);
- tx->base.format = pt->format;
tx->base.x = x;
tx->base.y = y;
tx->base.width = w;
tx->base.height = h;
- tx->base.block = pt->block;
- tx->base.nblocksx = pt->nblocksx[level];
- tx->base.nblocksy = pt->nblocksy[level];
tx->base.stride = mt->level[level].pitch;
tx->base.usage = usage;
tx->base.face = face;
@@ -76,7 +71,7 @@ nv40_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
tx->direct = false;
- nv40_compatible_transfer_tex(pt, level, &tx_tex_template);
+ nv40_compatible_transfer_tex(pt, w, h, &tx_tex_template);
tx_tex = pscreen->texture_create(pscreen, &tx_tex_template);
if (!tx_tex)
@@ -85,6 +80,8 @@ nv40_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
return NULL;
}
+ tx->base.stride = ((struct nv40_miptree*)tx_tex)->level[0].pitch;
+
tx->surface = pscreen->get_tex_surface(pscreen, tx_tex,
0, 0, 0,
pipe_transfer_buffer_flags(&tx->base));
@@ -110,8 +107,8 @@ nv40_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
/* TODO: Check if SIFM can un-swizzle */
nvscreen->eng2d->copy(nvscreen->eng2d,
tx->surface, 0, 0,
- src, 0, 0,
- src->width, src->height);
+ src, x, y,
+ w, h);
pipe_surface_reference(&src, NULL);
}
@@ -131,13 +128,13 @@ nv40_transfer_del(struct pipe_transfer *ptx)
dst = pscreen->get_tex_surface(pscreen, ptx->texture,
ptx->face, ptx->level, ptx->zslice,
- PIPE_BUFFER_USAGE_GPU_WRITE);
+ PIPE_BUFFER_USAGE_GPU_WRITE | NOUVEAU_BUFFER_USAGE_NO_RENDER);
/* TODO: Check if SIFM can deal with x,y,w,h when swizzling */
nvscreen->eng2d->copy(nvscreen->eng2d,
- dst, 0, 0,
+ dst, tx->base.x, tx->base.y,
tx->surface, 0, 0,
- dst->width, dst->height);
+ tx->base.width, tx->base.height);
pipe_surface_reference(&dst, NULL);
}
@@ -156,8 +153,10 @@ nv40_transfer_map(struct pipe_screen *pscreen, struct pipe_transfer *ptx)
void *map = pipe_buffer_map(pscreen, mt->buffer,
pipe_transfer_buffer_flags(ptx));
- return map + ns->base.offset +
- ptx->y * ns->pitch + ptx->x * ptx->block.size;
+ if(!tx->direct)
+ return map + ns->base.offset;
+ else
+ return map + ns->base.offset + ptx->y * ns->pitch + ptx->x * util_format_get_blocksize(ptx->texture->format);
}
static void
diff --git a/src/gallium/drivers/nv40/nv40_vbo.c b/src/gallium/drivers/nv40/nv40_vbo.c
index b2753b8e2e..af3fcf6a34 100644
--- a/src/gallium/drivers/nv40/nv40_vbo.c
+++ b/src/gallium/drivers/nv40/nv40_vbo.c
@@ -484,11 +484,6 @@ nv40_vbo_validate(struct nv40_context *nv40)
unsigned vb_flags = NOUVEAU_BO_VRAM | NOUVEAU_BO_GART | NOUVEAU_BO_RD;
int hw;
- if (nv40->edgeflags) {
- nv40->fallback_swtnl |= NV40_NEW_ARRAYS;
- return FALSE;
- }
-
vtxbuf = so_new(20, 18);
so_method(vtxbuf, curie, NV40TCL_VTXBUF_ADDRESS(0), nv40->vtxelt_nr);
vtxfmt = so_new(17, 0);
diff --git a/src/gallium/drivers/nv40/nv40_vertprog.c b/src/gallium/drivers/nv40/nv40_vertprog.c
index 31dae2457f..d9fc31006f 100644
--- a/src/gallium/drivers/nv40/nv40_vertprog.c
+++ b/src/gallium/drivers/nv40/nv40_vertprog.c
@@ -295,30 +295,30 @@ static INLINE struct nv40_sreg
tgsi_src(struct nv40_vpc *vpc, const struct tgsi_full_src_register *fsrc) {
struct nv40_sreg src;
- switch (fsrc->SrcRegister.File) {
+ switch (fsrc->Register.File) {
case TGSI_FILE_INPUT:
- src = nv40_sr(NV40SR_INPUT, fsrc->SrcRegister.Index);
+ src = nv40_sr(NV40SR_INPUT, fsrc->Register.Index);
break;
case TGSI_FILE_CONSTANT:
- src = constant(vpc, fsrc->SrcRegister.Index, 0, 0, 0, 0);
+ src = constant(vpc, fsrc->Register.Index, 0, 0, 0, 0);
break;
case TGSI_FILE_IMMEDIATE:
- src = vpc->imm[fsrc->SrcRegister.Index];
+ src = vpc->imm[fsrc->Register.Index];
break;
case TGSI_FILE_TEMPORARY:
- src = vpc->r_temp[fsrc->SrcRegister.Index];
+ src = vpc->r_temp[fsrc->Register.Index];
break;
default:
NOUVEAU_ERR("bad src file\n");
break;
}
- src.abs = fsrc->SrcRegisterExtMod.Absolute;
- src.negate = fsrc->SrcRegister.Negate;
- src.swz[0] = fsrc->SrcRegister.SwizzleX;
- src.swz[1] = fsrc->SrcRegister.SwizzleY;
- src.swz[2] = fsrc->SrcRegister.SwizzleZ;
- src.swz[3] = fsrc->SrcRegister.SwizzleW;
+ src.abs = fsrc->Register.Absolute;
+ src.negate = fsrc->Register.Negate;
+ src.swz[0] = fsrc->Register.SwizzleX;
+ src.swz[1] = fsrc->Register.SwizzleY;
+ src.swz[2] = fsrc->Register.SwizzleZ;
+ src.swz[3] = fsrc->Register.SwizzleW;
return src;
}
@@ -326,15 +326,15 @@ static INLINE struct nv40_sreg
tgsi_dst(struct nv40_vpc *vpc, const struct tgsi_full_dst_register *fdst) {
struct nv40_sreg dst;
- switch (fdst->DstRegister.File) {
+ switch (fdst->Register.File) {
case TGSI_FILE_OUTPUT:
- dst = vpc->r_result[fdst->DstRegister.Index];
+ dst = vpc->r_result[fdst->Register.Index];
break;
case TGSI_FILE_TEMPORARY:
- dst = vpc->r_temp[fdst->DstRegister.Index];
+ dst = vpc->r_temp[fdst->Register.Index];
break;
case TGSI_FILE_ADDRESS:
- dst = vpc->r_address[fdst->DstRegister.Index];
+ dst = vpc->r_address[fdst->Register.Index];
break;
default:
NOUVEAU_ERR("bad dst file\n");
@@ -405,8 +405,8 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
const struct tgsi_full_src_register *fsrc;
- fsrc = &finst->FullSrcRegisters[i];
- if (fsrc->SrcRegister.File == TGSI_FILE_TEMPORARY) {
+ fsrc = &finst->Src[i];
+ if (fsrc->Register.File == TGSI_FILE_TEMPORARY) {
src[i] = tgsi_src(vpc, fsrc);
}
}
@@ -414,9 +414,9 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
const struct tgsi_full_src_register *fsrc;
- fsrc = &finst->FullSrcRegisters[i];
+ fsrc = &finst->Src[i];
- switch (fsrc->SrcRegister.File) {
+ switch (fsrc->Register.File) {
case TGSI_FILE_INPUT:
case TGSI_FILE_CONSTANT:
case TGSI_FILE_TEMPORARY:
@@ -427,10 +427,10 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
break;
}
- switch (fsrc->SrcRegister.File) {
+ switch (fsrc->Register.File) {
case TGSI_FILE_INPUT:
- if (ai == -1 || ai == fsrc->SrcRegister.Index) {
- ai = fsrc->SrcRegister.Index;
+ if (ai == -1 || ai == fsrc->Register.Index) {
+ ai = fsrc->Register.Index;
src[i] = tgsi_src(vpc, fsrc);
} else {
src[i] = temp(vpc);
@@ -440,8 +440,8 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
break;
case TGSI_FILE_CONSTANT:
if ((ci == -1 && ii == -1) ||
- ci == fsrc->SrcRegister.Index) {
- ci = fsrc->SrcRegister.Index;
+ ci == fsrc->Register.Index) {
+ ci = fsrc->Register.Index;
src[i] = tgsi_src(vpc, fsrc);
} else {
src[i] = temp(vpc);
@@ -451,8 +451,8 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
break;
case TGSI_FILE_IMMEDIATE:
if ((ci == -1 && ii == -1) ||
- ii == fsrc->SrcRegister.Index) {
- ii = fsrc->SrcRegister.Index;
+ ii == fsrc->Register.Index) {
+ ii = fsrc->Register.Index;
src[i] = tgsi_src(vpc, fsrc);
} else {
src[i] = temp(vpc);
@@ -469,8 +469,8 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
}
}
- dst = tgsi_dst(vpc, &finst->FullDstRegisters[0]);
- mask = tgsi_mask(finst->FullDstRegisters[0].DstRegister.WriteMask);
+ dst = tgsi_dst(vpc, &finst->Dst[0]);
+ mask = tgsi_mask(finst->Dst[0].Register.WriteMask);
switch (finst->Instruction.Opcode) {
case TGSI_OPCODE_ABS:
@@ -577,19 +577,19 @@ static boolean
nv40_vertprog_parse_decl_output(struct nv40_vpc *vpc,
const struct tgsi_full_declaration *fdec)
{
- unsigned idx = fdec->DeclarationRange.First;
+ unsigned idx = fdec->Range.First;
int hw;
- switch (fdec->Semantic.SemanticName) {
+ switch (fdec->Semantic.Name) {
case TGSI_SEMANTIC_POSITION:
hw = NV40_VP_INST_DEST_POS;
vpc->hpos_idx = idx;
break;
case TGSI_SEMANTIC_COLOR:
- if (fdec->Semantic.SemanticIndex == 0) {
+ if (fdec->Semantic.Index == 0) {
hw = NV40_VP_INST_DEST_COL0;
} else
- if (fdec->Semantic.SemanticIndex == 1) {
+ if (fdec->Semantic.Index == 1) {
hw = NV40_VP_INST_DEST_COL1;
} else {
NOUVEAU_ERR("bad colour semantic index\n");
@@ -597,10 +597,10 @@ nv40_vertprog_parse_decl_output(struct nv40_vpc *vpc,
}
break;
case TGSI_SEMANTIC_BCOLOR:
- if (fdec->Semantic.SemanticIndex == 0) {
+ if (fdec->Semantic.Index == 0) {
hw = NV40_VP_INST_DEST_BFC0;
} else
- if (fdec->Semantic.SemanticIndex == 1) {
+ if (fdec->Semantic.Index == 1) {
hw = NV40_VP_INST_DEST_BFC1;
} else {
NOUVEAU_ERR("bad bcolour semantic index\n");
@@ -614,13 +614,17 @@ nv40_vertprog_parse_decl_output(struct nv40_vpc *vpc,
hw = NV40_VP_INST_DEST_PSZ;
break;
case TGSI_SEMANTIC_GENERIC:
- if (fdec->Semantic.SemanticIndex <= 7) {
- hw = NV40_VP_INST_DEST_TC(fdec->Semantic.SemanticIndex);
+ if (fdec->Semantic.Index <= 7) {
+ hw = NV40_VP_INST_DEST_TC(fdec->Semantic.Index);
} else {
NOUVEAU_ERR("bad generic semantic index\n");
return FALSE;
}
break;
+ case TGSI_SEMANTIC_EDGEFLAG:
+ /* not really an error just a fallback */
+ NOUVEAU_ERR("cannot handle edgeflag output\n");
+ return FALSE;
default:
NOUVEAU_ERR("bad output semantic\n");
return FALSE;
@@ -652,16 +656,16 @@ nv40_vertprog_prepare(struct nv40_vpc *vpc)
fdec = &p.FullToken.FullDeclaration;
switch (fdec->Declaration.File) {
case TGSI_FILE_TEMPORARY:
- if (fdec->DeclarationRange.Last > high_temp) {
+ if (fdec->Range.Last > high_temp) {
high_temp =
- fdec->DeclarationRange.Last;
+ fdec->Range.Last;
}
break;
#if 0 /* this would be nice.. except gallium doesn't track it */
case TGSI_FILE_ADDRESS:
- if (fdec->DeclarationRange.Last > high_addr) {
+ if (fdec->Range.Last > high_addr) {
high_addr =
- fdec->DeclarationRange.Last;
+ fdec->Range.Last;
}
break;
#endif
@@ -681,11 +685,11 @@ nv40_vertprog_prepare(struct nv40_vpc *vpc)
const struct tgsi_full_dst_register *fdst;
finst = &p.FullToken.FullInstruction;
- fdst = &finst->FullDstRegisters[0];
+ fdst = &finst->Dst[0];
- if (fdst->DstRegister.File == TGSI_FILE_ADDRESS) {
- if (fdst->DstRegister.Index > high_addr)
- high_addr = fdst->DstRegister.Index;
+ if (fdst->Register.File == TGSI_FILE_ADDRESS) {
+ if (fdst->Register.Index > high_addr)
+ high_addr = fdst->Register.Index;
}
}