#include "util/u_inlines.h"
#include "nv50/nv50_context.h"
+#include "nv50/nv50_query_hw.h"
+
+#include "nv50/nv50_compute.xml.h"
void
nv50_constbufs_validate(struct nv50_context *nv50)
continue;
}
if (!nv50->state.uniform_buffer_bound[s]) {
- nv50->state.uniform_buffer_bound[s] = TRUE;
+ nv50->state.uniform_buffer_bound[s] = true;
BEGIN_NV04(push, NV50_3D(SET_PROGRAM_CB), 1);
PUSH_DATA (push, (b << 12) | (i << 8) | p | 1);
}
while (words) {
- unsigned nr;
-
- if (!PUSH_SPACE(push, 16))
- break;
- nr = PUSH_AVAIL(push);
- assert(nr >= 16);
- nr = MIN2(MIN2(nr - 3, words), NV04_PFIFO_MAX_PACKET_LEN);
+ unsigned nr = MIN2(words, NV04_PFIFO_MAX_PACKET_LEN);
+ PUSH_SPACE(push, nr + 3);
BEGIN_NV04(push, NV50_3D(CB_ADDR), 1);
PUSH_DATA (push, (start << 8) | b);
BEGIN_NI04(push, NV50_3D(CB_DATA(0)), nr);
BEGIN_NV04(push, NV50_3D(SET_PROGRAM_CB), 1);
PUSH_DATA (push, (b << 12) | (i << 8) | p | 1);
- BCTX_REFN(nv50->bufctx_3d, CB(s, i), res, RD);
+ BCTX_REFN(nv50->bufctx_3d, 3D_CB(s, i), res, RD);
+
+ nv50->cb_dirty = 1; /* Force cache flush for UBO. */
+ res->cb_bindings[s] |= 1 << i;
} else {
BEGIN_NV04(push, NV50_3D(SET_PROGRAM_CB), 1);
PUSH_DATA (push, (i << 8) | p | 0);
}
if (i == 0)
- nv50->state.uniform_buffer_bound[s] = FALSE;
+ nv50->state.uniform_buffer_bound[s] = false;
}
}
}
}
-static boolean
+static bool
nv50_program_validate(struct nv50_context *nv50, struct nv50_program *prog)
{
if (!prog->translated) {
prog->translated = nv50_program_translate(
- prog, nv50->screen->base.device->chipset);
+ prog, nv50->screen->base.device->chipset, &nv50->base.debug);
if (!prog->translated)
- return FALSE;
+ return false;
} else
if (prog->mem)
- return TRUE;
+ return true;
return nv50_program_upload_code(nv50, prog);
}
-static INLINE void
+static inline void
nv50_program_update_context_state(struct nv50_context *nv50,
struct nv50_program *prog, int stage)
{
if (prog && prog->tls_space) {
if (nv50->state.new_tls_space)
- nouveau_bufctx_reset(nv50->bufctx_3d, NV50_BIND_TLS);
+ nouveau_bufctx_reset(nv50->bufctx_3d, NV50_BIND_3D_TLS);
if (!nv50->state.tls_required || nv50->state.new_tls_space)
- BCTX_REFN_bo(nv50->bufctx_3d, TLS, flags, nv50->screen->tls_bo);
- nv50->state.new_tls_space = FALSE;
+ BCTX_REFN_bo(nv50->bufctx_3d, 3D_TLS, flags, nv50->screen->tls_bo);
+ nv50->state.new_tls_space = false;
nv50->state.tls_required |= 1 << stage;
} else {
if (nv50->state.tls_required == (1 << stage))
- nouveau_bufctx_reset(nv50->bufctx_3d, NV50_BIND_TLS);
+ nouveau_bufctx_reset(nv50->bufctx_3d, NV50_BIND_3D_TLS);
nv50->state.tls_required &= ~(1 << stage);
}
}
{
struct nouveau_pushbuf *push = nv50->base.pushbuf;
struct nv50_program *fp = nv50->fragprog;
+ struct pipe_rasterizer_state *rast = &nv50->rast->pipe;
+
+ if (nv50->zsa && nv50->zsa->pipe.alpha.enabled) {
+ struct pipe_framebuffer_state *fb = &nv50->framebuffer;
+ bool blendable = fb->nr_cbufs == 0 || !fb->cbufs[0] ||
+ nv50->screen->base.base.is_format_supported(
+ &nv50->screen->base.base,
+ fb->cbufs[0]->format,
+ fb->cbufs[0]->texture->target,
+ fb->cbufs[0]->texture->nr_samples,
+ PIPE_BIND_BLENDABLE);
+ /* If we already have alphatest code, we have to keep updating
+ * it. However we only have to have different code if the current RT0 is
+ * non-blendable. Otherwise we just set it to always pass and use the
+ * hardware alpha test.
+ */
+ if (fp->fp.alphatest || !blendable) {
+ uint8_t alphatest = PIPE_FUNC_ALWAYS + 1;
+ if (!blendable)
+ alphatest = nv50->zsa->pipe.alpha.func + 1;
+ if (!fp->fp.alphatest)
+ nv50_program_destroy(nv50, fp);
+ else if (fp->mem && fp->fp.alphatest != alphatest)
+ nouveau_heap_free(&fp->mem);
+
+ fp->fp.alphatest = alphatest;
+ }
+ } else if (fp->fp.alphatest && fp->fp.alphatest != PIPE_FUNC_ALWAYS + 1) {
+ /* Alpha test is disabled but we have a shader where it's filled
+ * in. Make sure to reset the function to 'always', otherwise it'll end
+ * up discarding fragments incorrectly.
+ */
+ if (fp->mem)
+ nouveau_heap_free(&fp->mem);
+
+ fp->fp.alphatest = PIPE_FUNC_ALWAYS + 1;
+ }
+
+ if (fp->fp.force_persample_interp != rast->force_persample_interp) {
+ /* Force the program to be reuploaded, which will trigger interp fixups
+ * to get applied
+ */
+ if (fp->mem)
+ nouveau_heap_free(&fp->mem);
+
+ fp->fp.force_persample_interp = rast->force_persample_interp;
+ }
+
+ if (fp->mem && !(nv50->dirty_3d & (NV50_NEW_3D_FRAGPROG | NV50_NEW_3D_MIN_SAMPLES)))
+ return;
if (!nv50_program_validate(nv50, fp))
- return;
+ return;
nv50_program_update_context_state(nv50, fp, 1);
BEGIN_NV04(push, NV50_3D(FP_REG_ALLOC_TEMP), 1);
PUSH_DATA (push, fp->fp.flags[1]);
BEGIN_NV04(push, NV50_3D(FP_START_ID), 1);
PUSH_DATA (push, fp->code_base);
+
+ if (nv50->screen->tesla->oclass >= NVA3_3D_CLASS) {
+ BEGIN_NV04(push, SUBC_3D(NVA3_3D_FP_MULTISAMPLE), 1);
+ if (nv50->min_samples > 1 || fp->fp.has_samplemask)
+ PUSH_DATA(push,
+ NVA3_3D_FP_MULTISAMPLE_FORCE_PER_SAMPLE |
+ (NVA3_3D_FP_MULTISAMPLE_EXPORT_SAMPLE_MASK *
+ fp->fp.has_samplemask));
+ else
+ PUSH_DATA(push, 0);
+ }
}
void
/* GP_ENABLE is updated in linkage validation */
}
+void
+nv50_compprog_validate(struct nv50_context *nv50)
+{
+ struct nouveau_pushbuf *push = nv50->base.pushbuf;
+ struct nv50_program *cp = nv50->compprog;
+
+ if (cp && !nv50_program_validate(nv50, cp))
+ return;
+
+ BEGIN_NV04(push, NV50_CP(CODE_CB_FLUSH), 1);
+ PUSH_DATA (push, 0);
+}
+
static void
nv50_sprite_coords_validate(struct nv50_context *nv50)
{
for (i = 0; i < 8; ++i)
PUSH_DATA(push, 0);
- nv50->state.point_sprite = FALSE;
+ nv50->state.point_sprite = false;
}
return;
} else {
- nv50->state.point_sprite = TRUE;
+ nv50->state.point_sprite = true;
}
memset(pntc, 0, sizeof(pntc));
PUSH_DATA (push, !nv50->rast->pipe.rasterizer_discard);
}
- if (nv50->dirty & NV50_NEW_FRAGPROG)
+ if (nv50->dirty_3d & NV50_NEW_3D_FRAGPROG)
return;
psize = nv50->state.semantic_psize & ~NV50_3D_SEMANTIC_PTSZ_PTSZ_EN__MASK;
color = nv50->state.semantic_color & ~NV50_3D_SEMANTIC_COLOR_CLMP_EN;
struct nv50_varying dummy;
int i, n, c, m;
uint32_t primid = 0;
- uint32_t layerid = vp->gp.layerid;
+ uint32_t layerid = 0;
+ uint32_t viewportid = 0;
uint32_t psiz = 0x000;
uint32_t interp = fp->fp.interp;
uint32_t colors = fp->fp.colors;
+ uint32_t clpd_nr = util_last_bit(vp->vp.clip_enable | vp->vp.cull_enable);
uint32_t lin[4];
uint8_t map[64];
uint8_t so_map[64];
- if (!(nv50->dirty & (NV50_NEW_VERTPROG |
- NV50_NEW_FRAGPROG |
- NV50_NEW_GMTYPROG))) {
+ if (!(nv50->dirty_3d & (NV50_NEW_3D_VERTPROG |
+ NV50_NEW_3D_FRAGPROG |
+ NV50_NEW_3D_GMTYPROG))) {
uint8_t bfc, ffc;
ffc = (nv50->state.semantic_color & NV50_3D_SEMANTIC_COLOR_FFC0_ID__MASK);
bfc = (nv50->state.semantic_color & NV50_3D_SEMANTIC_COLOR_BFC0_ID__MASK)
dummy.linear = 0;
m = nv50_vec4_map(map, 0, lin, &dummy, &vp->out[0]);
- for (c = 0; c < vp->vp.clpd_nr; ++c)
+ for (c = 0; c < clpd_nr; ++c)
map[m++] = vp->vp.clpd[c / 4] + (c % 4);
colors |= m << 8; /* adjust BFC0 id */
if (vp->out[n].sn == fp->in[i].sn &&
vp->out[n].si == fp->in[i].si)
break;
- if (i == fp->gp.primid) {
+ switch (fp->in[i].sn) {
+ case TGSI_SEMANTIC_PRIMID:
primid = m;
+ break;
+ case TGSI_SEMANTIC_LAYER:
+ layerid = m;
+ break;
+ case TGSI_SEMANTIC_VIEWPORT_INDEX:
+ viewportid = m;
+ break;
}
m = nv50_vec4_map(map, m, lin,
&fp->in[i], (n < vp->out_nr) ? &vp->out[n] : &dummy);
}
- if (vp->gp.has_layer) {
- // In GL4.x, layer can be an fp input, but not in 3.x. Make sure to add
- // it to the output map.
- map[m++] = layerid;
+ if (vp->gp.has_layer && !layerid) {
+ layerid = m;
+ map[m++] = vp->gp.layerid;
+ }
+
+ if (vp->gp.has_viewport && !viewportid) {
+ viewportid = m;
+ map[m++] = vp->gp.viewportid;
}
if (nv50->rast->pipe.point_size_per_vertex) {
BEGIN_NV04(push, NV50_3D(SEMANTIC_PRIM_ID), 1);
PUSH_DATA (push, primid);
+ assert(m > 0);
BEGIN_NV04(push, NV50_3D(VP_RESULT_MAP_SIZE), 1);
PUSH_DATA (push, m);
BEGIN_NV04(push, NV50_3D(VP_RESULT_MAP(0)), n);
PUSH_DATAp(push, map, n);
}
- BEGIN_NV04(push, NV50_3D(SEMANTIC_COLOR), 4);
+ BEGIN_NV04(push, NV50_3D(GP_VIEWPORT_ID_ENABLE), 5);
+ PUSH_DATA (push, vp->gp.has_viewport);
PUSH_DATA (push, colors);
- PUSH_DATA (push, (vp->vp.clpd_nr << 8) | 4);
+ PUSH_DATA (push, (clpd_nr << 8) | 4);
PUSH_DATA (push, layerid);
PUSH_DATA (push, psiz);
+ BEGIN_NV04(push, NV50_3D(SEMANTIC_VIEWPORT), 1);
+ PUSH_DATA (push, viewportid);
+
BEGIN_NV04(push, NV50_3D(LAYER), 1);
PUSH_DATA (push, vp->gp.has_layer << 16);
oid += mv & 1;
}
}
+ if (!m)
+ map[m++] = 0;
return m;
}
BEGIN_NV04(push, NV50_3D(VP_GP_BUILTIN_ATTR_EN), 1);
PUSH_DATA (push, vp->vp.attrs[2] | gp->vp.attrs[2]);
+ assert(m > 0);
BEGIN_NV04(push, NV50_3D(VP_RESULT_MAP_SIZE), 1);
PUSH_DATA (push, m);
BEGIN_NV04(push, NV50_3D(VP_RESULT_MAP(0)), n);
BEGIN_NV04(push, NV50_3D(STRMOUT_BUFFERS_CTRL), 1);
PUSH_DATA (push, ctrl);
- nouveau_bufctx_reset(nv50->bufctx_3d, NV50_BIND_SO);
-
for (i = 0; i < nv50->num_so_targets; ++i) {
struct nv50_so_target *targ = nv50_so_target(nv50->so_target[i]);
struct nv04_resource *buf = nv04_resource(targ->pipe.buffer);
const unsigned n = nv50->screen->base.class_3d >= NVA0_3D_CLASS ? 4 : 3;
if (n == 4 && !targ->clean)
- nv84_query_fifo_wait(push, targ->pq);
+ nv84_hw_query_fifo_wait(push, nv50_query(targ->pq));
BEGIN_NV04(push, NV50_3D(STRMOUT_ADDRESS_HIGH(i)), n);
PUSH_DATAh(push, buf->address + targ->pipe.buffer_offset);
PUSH_DATA (push, buf->address + targ->pipe.buffer_offset);
PUSH_DATA (push, so->num_attribs[i]);
if (n == 4) {
PUSH_DATA(push, targ->pipe.buffer_size);
-
- BEGIN_NV04(push, NVA0_3D(STRMOUT_OFFSET(i)), 1);
if (!targ->clean) {
assert(targ->pq);
- nv50_query_pushbuf_submit(push, targ->pq, 0x4);
+ nv50_hw_query_pushbuf_submit(push, NVA0_3D_STRMOUT_OFFSET(i),
+ nv50_query(targ->pq), 0x4);
} else {
+ BEGIN_NV04(push, NVA0_3D(STRMOUT_OFFSET(i)), 1);
PUSH_DATA(push, 0);
- targ->clean = FALSE;
+ targ->clean = false;
}
} else {
const unsigned limit = targ->pipe.buffer_size /
(so->stride[i] * nv50->state.prim_size);
prims = MIN2(prims, limit);
}
- BCTX_REFN(nv50->bufctx_3d, SO, buf, WR);
+ targ->stride = so->stride[i];
+ BCTX_REFN(nv50->bufctx_3d, 3D_SO, buf, WR);
}
if (prims != ~0) {
BEGIN_NV04(push, NV50_3D(STRMOUT_PRIMITIVE_LIMIT), 1);
projects / mesa.git / blobdiff