#include "util/u_inlines.h"
#include "util/u_pack_color.h"
-#include "util/u_format.h"
+#include "util/format/u_format.h"
#include "util/u_surface.h"
#include "os/os_thread.h"
#include "nvc0/nvc0_context.h"
#include "nvc0/nvc0_resource.h"
-#include "nv50/nv50_defs.xml.h"
-#include "nv50/nv50_texture.xml.h"
+#include "nv50/g80_defs.xml.h"
+#include "nv50/g80_texture.xml.h"
/* these are used in nv50_blit.h */
#define NV50_ENG2D_SUPPORTED_FORMATS 0xff9ccfe1cce3ccc9ULL
/* A8_UNORM is treated as I8_UNORM as far as the 2D engine is concerned. */
if (!dst && unlikely(format == PIPE_FORMAT_I8_UNORM) && !dst_src_equal)
- return NV50_SURFACE_FORMAT_A8_UNORM;
+ return G80_SURFACE_FORMAT_A8_UNORM;
/* Hardware values for color formats range from 0xc0 to 0xff,
* but the 2D engine doesn't support all of them.
switch (util_format_get_blocksize(format)) {
case 1:
- return NV50_SURFACE_FORMAT_R8_UNORM;
+ return G80_SURFACE_FORMAT_R8_UNORM;
case 2:
- return NV50_SURFACE_FORMAT_RG8_UNORM;
+ return G80_SURFACE_FORMAT_RG8_UNORM;
case 4:
- return NV50_SURFACE_FORMAT_BGRA8_UNORM;
+ return G80_SURFACE_FORMAT_BGRA8_UNORM;
case 8:
- return NV50_SURFACE_FORMAT_RGBA16_UNORM;
+ return G80_SURFACE_FORMAT_RGBA16_UNORM;
case 16:
- return NV50_SURFACE_FORMAT_RGBA32_FLOAT;
+ return G80_SURFACE_FORMAT_RGBA32_FLOAT;
default:
assert(0);
return 0;
struct pipe_surface *dst,
const union pipe_color_union *color,
unsigned dstx, unsigned dsty,
- unsigned width, unsigned height)
+ unsigned width, unsigned height,
+ bool render_condition_enabled)
{
struct nvc0_context *nvc0 = nvc0_context(pipe);
struct nouveau_pushbuf *push = nvc0->base.pushbuf;
struct nv04_resource *res = nv04_resource(sf->base.texture);
unsigned z;
+ assert(dst->texture->target != PIPE_BUFFER);
+
if (!PUSH_SPACE(push, 32 + sf->depth))
return;
nvc0_resource_fence(res, NOUVEAU_BO_WR);
}
+ if (!render_condition_enabled)
+ IMMED_NVC0(push, NVC0_3D(COND_MODE), NVC0_3D_COND_MODE_ALWAYS);
+
BEGIN_NIC0(push, NVC0_3D(CLEAR_BUFFERS), sf->depth);
for (z = 0; z < sf->depth; ++z) {
PUSH_DATA (push, 0x3c |
(z << NVC0_3D_CLEAR_BUFFERS_LAYER__SHIFT));
}
- nvc0->dirty |= NVC0_NEW_FRAMEBUFFER;
+ if (!render_condition_enabled)
+ IMMED_NVC0(push, NVC0_3D(COND_MODE), nvc0->cond_condmode);
+
+ nvc0->dirty_3d |= NVC0_NEW_3D_FRAMEBUFFER;
}
static void
-nvc0_clear_buffer_cpu(struct pipe_context *pipe,
- struct pipe_resource *res,
- unsigned offset, unsigned size,
- const void *data, int data_size)
+nvc0_clear_buffer_push_nvc0(struct pipe_context *pipe,
+ struct pipe_resource *res,
+ unsigned offset, unsigned size,
+ const void *data, int data_size)
{
+ struct nvc0_context *nvc0 = nvc0_context(pipe);
+ struct nouveau_pushbuf *push = nvc0->base.pushbuf;
struct nv04_resource *buf = nv04_resource(res);
- struct pipe_transfer *pt;
- struct pipe_box box;
- unsigned elements, i;
+ unsigned i;
- elements = size / data_size;
+ nouveau_bufctx_refn(nvc0->bufctx, 0, buf->bo, buf->domain | NOUVEAU_BO_WR);
+ nouveau_pushbuf_bufctx(push, nvc0->bufctx);
+ nouveau_pushbuf_validate(push);
+
+ unsigned count = (size + 3) / 4;
+ unsigned data_words = data_size / 4;
- u_box_1d(offset, size, &box);
+ while (count) {
+ unsigned nr_data = MIN2(count, NV04_PFIFO_MAX_PACKET_LEN) / data_words;
+ unsigned nr = nr_data * data_words;
+
+ if (!PUSH_SPACE(push, nr + 9))
+ break;
- uint8_t *map = buf->vtbl->transfer_map(pipe, res, 0, PIPE_TRANSFER_WRITE,
- &box, &pt);
+ BEGIN_NVC0(push, NVC0_M2MF(OFFSET_OUT_HIGH), 2);
+ PUSH_DATAh(push, buf->address + offset);
+ PUSH_DATA (push, buf->address + offset);
+ BEGIN_NVC0(push, NVC0_M2MF(LINE_LENGTH_IN), 2);
+ PUSH_DATA (push, MIN2(size, nr * 4));
+ PUSH_DATA (push, 1);
+ BEGIN_NVC0(push, NVC0_M2MF(EXEC), 1);
+ PUSH_DATA (push, 0x100111);
+
+ /* must not be interrupted (trap on QUERY fence, 0x50 works however) */
+ BEGIN_NIC0(push, NVC0_M2MF(DATA), nr);
+ for (i = 0; i < nr_data; i++)
+ PUSH_DATAp(push, data, data_words);
+
+ count -= nr;
+ offset += nr * 4;
+ size -= nr * 4;
+ }
+
+ nvc0_resource_validate(buf, NOUVEAU_BO_WR);
+
+ nouveau_bufctx_reset(nvc0->bufctx, 0);
+}
+
+static void
+nvc0_clear_buffer_push_nve4(struct pipe_context *pipe,
+ struct pipe_resource *res,
+ unsigned offset, unsigned size,
+ const void *data, int data_size)
+{
+ struct nvc0_context *nvc0 = nvc0_context(pipe);
+ struct nouveau_pushbuf *push = nvc0->base.pushbuf;
+ struct nv04_resource *buf = nv04_resource(res);
+ unsigned i;
+
+ nouveau_bufctx_refn(nvc0->bufctx, 0, buf->bo, buf->domain | NOUVEAU_BO_WR);
+ nouveau_pushbuf_bufctx(push, nvc0->bufctx);
+ nouveau_pushbuf_validate(push);
+
+ unsigned count = (size + 3) / 4;
+ unsigned data_words = data_size / 4;
+
+ while (count) {
+ unsigned nr_data = MIN2(count, NV04_PFIFO_MAX_PACKET_LEN) / data_words;
+ unsigned nr = nr_data * data_words;
+
+ if (!PUSH_SPACE(push, nr + 10))
+ break;
+
+ BEGIN_NVC0(push, NVE4_P2MF(UPLOAD_DST_ADDRESS_HIGH), 2);
+ PUSH_DATAh(push, buf->address + offset);
+ PUSH_DATA (push, buf->address + offset);
+ BEGIN_NVC0(push, NVE4_P2MF(UPLOAD_LINE_LENGTH_IN), 2);
+ PUSH_DATA (push, MIN2(size, nr * 4));
+ PUSH_DATA (push, 1);
+ /* must not be interrupted (trap on QUERY fence, 0x50 works however) */
+ BEGIN_1IC0(push, NVE4_P2MF(UPLOAD_EXEC), nr + 1);
+ PUSH_DATA (push, 0x1001);
+ for (i = 0; i < nr_data; i++)
+ PUSH_DATAp(push, data, data_words);
+
+ count -= nr;
+ offset += nr * 4;
+ size -= nr * 4;
+ }
+
+ nvc0_resource_validate(buf, NOUVEAU_BO_WR);
+
+ nouveau_bufctx_reset(nvc0->bufctx, 0);
+}
- for (i = 0; i < elements; ++i)
- memcpy(&map[i*data_size], data, data_size);
+static void
+nvc0_clear_buffer_push(struct pipe_context *pipe,
+ struct pipe_resource *res,
+ unsigned offset, unsigned size,
+ const void *data, int data_size)
+{
+ struct nvc0_context *nvc0 = nvc0_context(pipe);
+ unsigned tmp;
+
+ if (data_size == 1) {
+ tmp = *(unsigned char *)data;
+ tmp = (tmp << 24) | (tmp << 16) | (tmp << 8) | tmp;
+ data = &tmp;
+ data_size = 4;
+ } else if (data_size == 2) {
+ tmp = *(unsigned short *)data;
+ tmp = (tmp << 16) | tmp;
+ data = &tmp;
+ data_size = 4;
+ }
- buf->vtbl->transfer_unmap(pipe, pt);
+ if (nvc0->screen->base.class_3d < NVE4_3D_CLASS)
+ nvc0_clear_buffer_push_nvc0(pipe, res, offset, size, data, data_size);
+ else
+ nvc0_clear_buffer_push_nve4(pipe, res, offset, size, data, data_size);
}
static void
memcpy(&color.ui, data, 16);
break;
case 12:
- /* This doesn't work, RGB32 is not a valid RT format.
- * dst_fmt = PIPE_FORMAT_R32G32B32_UINT;
- * memcpy(&color.ui, data, 12);
- * memset(&color.ui[3], 0, 4);
+ /* RGB32 is not a valid RT format. This will be handled by the pushbuf
+ * uploader.
*/
break;
case 8:
return;
}
+ util_range_add(&buf->base, &buf->valid_buffer_range, offset, offset + size);
+
assert(size % data_size == 0);
if (data_size == 12) {
- /* TODO: Find a way to do this with the GPU! */
- nvc0_clear_buffer_cpu(pipe, res, offset, size, data, data_size);
+ nvc0_clear_buffer_push(pipe, res, offset, size, data, data_size);
return;
}
+ if (offset & 0xff) {
+ unsigned fixup_size = MIN2(size, align(offset, 0x100) - offset);
+ assert(fixup_size % data_size == 0);
+ nvc0_clear_buffer_push(pipe, res, offset, fixup_size, data, data_size);
+ offset += fixup_size;
+ size -= fixup_size;
+ if (!size)
+ return;
+ }
+
elements = size / data_size;
height = (elements + 16383) / 16384;
width = elements / height;
+ if (height > 1)
+ width &= ~0xff;
+ assert(width > 0);
if (!PUSH_SPACE(push, 40))
return;
PUSH_REFN (push, buf->bo, buf->domain | NOUVEAU_BO_WR);
BEGIN_NVC0(push, NVC0_3D(CLEAR_COLOR(0)), 4);
- PUSH_DATAf(push, color.f[0]);
- PUSH_DATAf(push, color.f[1]);
- PUSH_DATAf(push, color.f[2]);
- PUSH_DATAf(push, color.f[3]);
+ PUSH_DATA (push, color.ui[0]);
+ PUSH_DATA (push, color.ui[1]);
+ PUSH_DATA (push, color.ui[2]);
+ PUSH_DATA (push, color.ui[3]);
BEGIN_NVC0(push, NVC0_3D(SCREEN_SCISSOR_HORIZ), 2);
PUSH_DATA (push, width << 16);
PUSH_DATA (push, height << 16);
BEGIN_NVC0(push, NVC0_3D(RT_ADDRESS_HIGH(0)), 9);
PUSH_DATAh(push, buf->address + offset);
PUSH_DATA (push, buf->address + offset);
- PUSH_DATA (push, width * data_size);
+ PUSH_DATA (push, align(width * data_size, 0x100));
PUSH_DATA (push, height);
PUSH_DATA (push, nvc0_format_table[dst_fmt].rt);
PUSH_DATA (push, NVC0_3D_RT_TILE_MODE_LINEAR);
IMMED_NVC0(push, NVC0_3D(ZETA_ENABLE), 0);
IMMED_NVC0(push, NVC0_3D(MULTISAMPLE_MODE), 0);
+ IMMED_NVC0(push, NVC0_3D(COND_MODE), NVC0_3D_COND_MODE_ALWAYS);
+
IMMED_NVC0(push, NVC0_3D(CLEAR_BUFFERS), 0x3c);
+ IMMED_NVC0(push, NVC0_3D(COND_MODE), nvc0->cond_condmode);
+
+ nvc0_resource_validate(buf, NOUVEAU_BO_WR);
+
if (width * height != elements) {
offset += width * height * data_size;
width = elements - width * height;
- height = 1;
-
- BEGIN_NVC0(push, NVC0_3D(RT_ADDRESS_HIGH(0)), 4);
- PUSH_DATAh(push, buf->address + offset);
- PUSH_DATA (push, buf->address + offset);
- PUSH_DATA (push, width * data_size);
- PUSH_DATA (push, height);
-
- IMMED_NVC0(push, NVC0_3D(CLEAR_BUFFERS), 0x3c);
+ nvc0_clear_buffer_push(pipe, res, offset, width * data_size,
+ data, data_size);
}
- nouveau_fence_ref(nvc0->screen->base.fence.current, &buf->fence);
- nouveau_fence_ref(nvc0->screen->base.fence.current, &buf->fence_wr);
- nvc0->dirty |= NVC0_NEW_FRAMEBUFFER;
+ nvc0->dirty_3d |= NVC0_NEW_3D_FRAMEBUFFER;
}
static void
double depth,
unsigned stencil,
unsigned dstx, unsigned dsty,
- unsigned width, unsigned height)
+ unsigned width, unsigned height,
+ bool render_condition_enabled)
{
struct nvc0_context *nvc0 = nvc0_context(pipe);
struct nouveau_pushbuf *push = nvc0->base.pushbuf;
int unk = mt->base.base.target == PIPE_TEXTURE_2D;
unsigned z;
+ assert(dst->texture->target != PIPE_BUFFER);
+
if (!PUSH_SPACE(push, 32 + sf->depth))
return;
PUSH_DATA (push, dst->u.tex.first_layer);
IMMED_NVC0(push, NVC0_3D(MULTISAMPLE_MODE), mt->ms_mode);
+ if (!render_condition_enabled)
+ IMMED_NVC0(push, NVC0_3D(COND_MODE), NVC0_3D_COND_MODE_ALWAYS);
+
BEGIN_NIC0(push, NVC0_3D(CLEAR_BUFFERS), sf->depth);
for (z = 0; z < sf->depth; ++z) {
PUSH_DATA (push, mode |
(z << NVC0_3D_CLEAR_BUFFERS_LAYER__SHIFT));
}
- nvc0->dirty |= NVC0_NEW_FRAMEBUFFER;
+ if (!render_condition_enabled)
+ IMMED_NVC0(push, NVC0_3D(COND_MODE), nvc0->cond_condmode);
+
+ nvc0->dirty_3d |= NVC0_NEW_3D_FRAMEBUFFER;
}
void
nvc0_clear(struct pipe_context *pipe, unsigned buffers,
+ const struct pipe_scissor_state *scissor_state,
const union pipe_color_union *color,
double depth, unsigned stencil)
{
uint32_t mode = 0;
/* don't need NEW_BLEND, COLOR_MASK doesn't affect CLEAR_BUFFERS */
- if (!nvc0_state_validate(nvc0, NVC0_NEW_FRAMEBUFFER, 9 + (fb->nr_cbufs * 2)))
+ if (!nvc0_state_validate_3d(nvc0, NVC0_NEW_3D_FRAMEBUFFER))
return;
if (buffers & PIPE_CLEAR_COLOR && fb->nr_cbufs) {
}
}
+static void
+gm200_evaluate_depth_buffer(struct pipe_context *pipe)
+{
+ struct nvc0_context *nvc0 = nvc0_context(pipe);
+ struct nouveau_pushbuf *push = nvc0->base.pushbuf;
+
+ nvc0_state_validate_3d(nvc0, NVC0_NEW_3D_FRAMEBUFFER);
+ IMMED_NVC0(push, SUBC_3D(0x11fc), 1);
+}
+
/* =============================== BLIT CODE ===================================
*/
struct nv50_tsc_entry sampler[2]; /* nearest, bilinear */
- pipe_mutex mutex;
+ mtx_t mutex;
struct nvc0_screen *screen;
};
enum pipe_texture_target target;
struct {
struct pipe_framebuffer_state fb;
+ struct nvc0_window_rect_stateobj window_rect;
struct nvc0_rasterizer_stateobj *rast;
struct nvc0_program *vp;
struct nvc0_program *tcp;
struct pipe_sampler_view *texture[2];
struct nv50_tsc_entry *sampler[2];
unsigned min_samples;
- uint32_t dirty;
+ uint32_t dirty_3d;
} saved;
struct nvc0_rasterizer_stateobj rast;
};
};
static const uint32_t code_gm107[] =
{
- 0xfc0007e0, 0x001f8000, /* sched 0x7e0 0x7e0 0x7e0 */
- 0x0807ff04, 0xefd8ff80, /* ld b64 $r4 a[0x80] 0x0 */
- 0x0907ff00, 0xefd97f80, /* ld b96 $r0 a[0x90] 0x0 */
- 0x0707ff04, 0xeff0ff80, /* st b64 a[0x70] $r4 0x0 */
- 0xfc0007e0, 0x00000000, /* sched 0x7e0 0x7e0 0x0 */
- 0x0807ff00, 0xeff17f80, /* st b96 a[0x80] $r0 0x0 */
+ 0xe4200701, 0x001d0400, /* sched (st 0x1 wr 0x0) (st 0x1 wr 0x1) (st 0x1 wr 0x2) */
+ 0x0807ff00, 0xefd87f80, /* ld b32 $r0 a[0x80] 0x0 */
+ 0x0847ff01, 0xefd87f80, /* ld b32 $r1 a[0x84] 0x0 */
+ 0x0907ff02, 0xefd87f80, /* ld b32 $r2 a[0x90] 0x0 */
+ 0xf0200761, 0x003f8400, /* sched (st 0x1 wr 0x3) (st 0x1 wr 0x4) (st 0x1 wt 0x1) */
+ 0x0947ff03, 0xefd87f80, /* ld b32 $r3 a[0x94] 0x0 */
+ 0x0987ff04, 0xefd87f80, /* ld b32 $r4 a[0x98] 0x0 */
+ 0x0707ff00, 0xeff07f80, /* st b32 a[0x70] $r0 0x0 */
+ 0xfc2017e1, 0x011f8404, /* sched (st 0x1 wt 0x2) (st 0x1 wt 0x4) (st 0x1 wt 0x8) */
+ 0x0747ff01, 0xeff07f80, /* st b32 a[0x74] $r1 0x0 */
+ 0x0807ff02, 0xeff07f80, /* st b32 a[0x80] $r2 0x0 */
+ 0x0847ff03, 0xeff07f80, /* st b32 a[0x84] $r3 0x0 */
+ 0xfde087e1, 0x001f8000, /* sched (st 0x1 wt 0x10) (st 0xf) (st 0x0) */
+ 0x0887ff04, 0xeff07f80, /* st b32 a[0x88] $r4 0x0 */
0x0007000f, 0xe3000000, /* exit */
};
blit->sampler[0].id = -1;
- blit->sampler[0].tsc[0] = NV50_TSC_0_SRGB_CONVERSION_ALLOWED |
- (NV50_TSC_WRAP_CLAMP_TO_EDGE << NV50_TSC_0_WRAPS__SHIFT) |
- (NV50_TSC_WRAP_CLAMP_TO_EDGE << NV50_TSC_0_WRAPT__SHIFT) |
- (NV50_TSC_WRAP_CLAMP_TO_EDGE << NV50_TSC_0_WRAPR__SHIFT);
+ blit->sampler[0].tsc[0] = G80_TSC_0_SRGB_CONVERSION |
+ (G80_TSC_WRAP_CLAMP_TO_EDGE << G80_TSC_0_ADDRESS_U__SHIFT) |
+ (G80_TSC_WRAP_CLAMP_TO_EDGE << G80_TSC_0_ADDRESS_V__SHIFT) |
+ (G80_TSC_WRAP_CLAMP_TO_EDGE << G80_TSC_0_ADDRESS_P__SHIFT);
blit->sampler[0].tsc[1] =
- NV50_TSC_1_MAGF_NEAREST | NV50_TSC_1_MINF_NEAREST | NV50_TSC_1_MIPF_NONE;
+ G80_TSC_1_MAG_FILTER_NEAREST |
+ G80_TSC_1_MIN_FILTER_NEAREST |
+ G80_TSC_1_MIP_FILTER_NONE;
/* clamp to edge, min/max lod = 0, bilinear filtering */
blit->sampler[1].tsc[0] = blit->sampler[0].tsc[0];
blit->sampler[1].tsc[1] =
- NV50_TSC_1_MAGF_LINEAR | NV50_TSC_1_MINF_LINEAR | NV50_TSC_1_MIPF_NONE;
+ G80_TSC_1_MAG_FILTER_LINEAR |
+ G80_TSC_1_MIN_FILTER_LINEAR |
+ G80_TSC_1_MIP_FILTER_NONE;
}
static void
const unsigned mode = ctx->mode;
if (!blitter->fp[targ][mode]) {
- pipe_mutex_lock(blitter->mutex);
+ mtx_lock(&blitter->mutex);
if (!blitter->fp[targ][mode])
blitter->fp[targ][mode] =
nv50_blitter_make_fp(&ctx->nvc0->base.pipe, mode, ptarg);
- pipe_mutex_unlock(blitter->mutex);
+ mtx_unlock(&blitter->mutex);
}
ctx->fp = blitter->fp[targ][mode];
}
templ.format = format;
templ.u.tex.first_layer = templ.u.tex.last_layer = layer;
templ.u.tex.first_level = templ.u.tex.last_level = level;
- templ.swizzle_r = PIPE_SWIZZLE_RED;
- templ.swizzle_g = PIPE_SWIZZLE_GREEN;
- templ.swizzle_b = PIPE_SWIZZLE_BLUE;
- templ.swizzle_a = PIPE_SWIZZLE_ALPHA;
+ templ.swizzle_r = PIPE_SWIZZLE_X;
+ templ.swizzle_g = PIPE_SWIZZLE_Y;
+ templ.swizzle_b = PIPE_SWIZZLE_Z;
+ templ.swizzle_a = PIPE_SWIZZLE_W;
if (layer == -1) {
templ.u.tex.first_layer = 0;
}
static void
-nvc0_blitctx_pre_blit(struct nvc0_blitctx *ctx)
+nvc0_blitctx_pre_blit(struct nvc0_blitctx *ctx,
+ const struct pipe_blit_info *info)
{
struct nvc0_context *nvc0 = ctx->nvc0;
struct nvc0_blitter *blitter = nvc0->screen->blitter;
ctx->saved.fb.width = nvc0->framebuffer.width;
ctx->saved.fb.height = nvc0->framebuffer.height;
+ ctx->saved.fb.samples = nvc0->framebuffer.samples;
+ ctx->saved.fb.layers = nvc0->framebuffer.layers;
ctx->saved.fb.nr_cbufs = nvc0->framebuffer.nr_cbufs;
ctx->saved.fb.cbufs[0] = nvc0->framebuffer.cbufs[0];
ctx->saved.fb.zsbuf = nvc0->framebuffer.zsbuf;
ctx->saved.fp = nvc0->fragprog;
ctx->saved.min_samples = nvc0->min_samples;
+ ctx->saved.window_rect = nvc0->window_rect;
nvc0->rast = &ctx->rast;
nvc0->gmtyprog = NULL;
nvc0->fragprog = ctx->fp;
+ nvc0->window_rect.rects =
+ MIN2(info->num_window_rectangles, NVC0_MAX_WINDOW_RECTANGLES);
+ nvc0->window_rect.inclusive = info->window_rectangle_include;
+ if (nvc0->window_rect.rects)
+ memcpy(nvc0->window_rect.rect, info->window_rectangles,
+ sizeof(struct pipe_scissor_state) * nvc0->window_rect.rects);
+
for (s = 0; s <= 4; ++s) {
ctx->saved.num_textures[s] = nvc0->num_textures[s];
ctx->saved.num_samplers[s] = nvc0->num_samplers[s];
nvc0->min_samples = 1;
- ctx->saved.dirty = nvc0->dirty;
+ ctx->saved.dirty_3d = nvc0->dirty_3d;
nvc0->textures_dirty[4] |= 3;
nvc0->samplers_dirty[4] |= 3;
- nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_FB);
- nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_TEX(4, 0));
- nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_TEX(4, 1));
+ nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_3D_FB);
+ nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_3D_TEX(4, 0));
+ nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_3D_TEX(4, 1));
- nvc0->dirty = NVC0_NEW_FRAMEBUFFER | NVC0_NEW_MIN_SAMPLES |
- NVC0_NEW_VERTPROG | NVC0_NEW_FRAGPROG |
- NVC0_NEW_TCTLPROG | NVC0_NEW_TEVLPROG | NVC0_NEW_GMTYPROG |
- NVC0_NEW_TEXTURES | NVC0_NEW_SAMPLERS;
+ nvc0->dirty_3d = NVC0_NEW_3D_FRAMEBUFFER | NVC0_NEW_3D_MIN_SAMPLES |
+ NVC0_NEW_3D_VERTPROG | NVC0_NEW_3D_FRAGPROG |
+ NVC0_NEW_3D_TCTLPROG | NVC0_NEW_3D_TEVLPROG | NVC0_NEW_3D_GMTYPROG |
+ NVC0_NEW_3D_TEXTURES | NVC0_NEW_3D_SAMPLERS | NVC0_NEW_3D_WINDOW_RECTS;
}
static void
nvc0->framebuffer.width = blit->saved.fb.width;
nvc0->framebuffer.height = blit->saved.fb.height;
+ nvc0->framebuffer.samples = blit->saved.fb.samples;
+ nvc0->framebuffer.layers = blit->saved.fb.layers;
nvc0->framebuffer.nr_cbufs = blit->saved.fb.nr_cbufs;
nvc0->framebuffer.cbufs[0] = blit->saved.fb.cbufs[0];
nvc0->framebuffer.zsbuf = blit->saved.fb.zsbuf;
nvc0->fragprog = blit->saved.fp;
nvc0->min_samples = blit->saved.min_samples;
+ nvc0->window_rect = blit->saved.window_rect;
pipe_sampler_view_reference(&nvc0->textures[4][0], NULL);
pipe_sampler_view_reference(&nvc0->textures[4][1], NULL);
nvc0->base.pipe.render_condition(&nvc0->base.pipe, nvc0->cond_query,
nvc0->cond_cond, nvc0->cond_mode);
- nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_FB);
- nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_TEX(4, 0));
- nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_TEX(4, 1));
-
- nvc0->dirty = blit->saved.dirty |
- (NVC0_NEW_FRAMEBUFFER | NVC0_NEW_SCISSOR | NVC0_NEW_SAMPLE_MASK |
- NVC0_NEW_RASTERIZER | NVC0_NEW_ZSA | NVC0_NEW_BLEND |
- NVC0_NEW_VIEWPORT |
- NVC0_NEW_TEXTURES | NVC0_NEW_SAMPLERS |
- NVC0_NEW_VERTPROG | NVC0_NEW_FRAGPROG |
- NVC0_NEW_TCTLPROG | NVC0_NEW_TEVLPROG | NVC0_NEW_GMTYPROG |
- NVC0_NEW_TFB_TARGETS | NVC0_NEW_VERTEX | NVC0_NEW_ARRAYS);
+ nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_3D_VTX_TMP);
+ nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_3D_TEXT);
+ nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_3D_FB);
+ nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_3D_TEX(4, 0));
+ nouveau_bufctx_reset(nvc0->bufctx_3d, NVC0_BIND_3D_TEX(4, 1));
+ nouveau_scratch_done(&nvc0->base);
+
+ nvc0->dirty_3d = blit->saved.dirty_3d |
+ (NVC0_NEW_3D_FRAMEBUFFER | NVC0_NEW_3D_SCISSOR | NVC0_NEW_3D_SAMPLE_MASK |
+ NVC0_NEW_3D_RASTERIZER | NVC0_NEW_3D_ZSA | NVC0_NEW_3D_BLEND |
+ NVC0_NEW_3D_VIEWPORT | NVC0_NEW_3D_WINDOW_RECTS |
+ NVC0_NEW_3D_TEXTURES | NVC0_NEW_3D_SAMPLERS |
+ NVC0_NEW_3D_VERTPROG | NVC0_NEW_3D_FRAGPROG |
+ NVC0_NEW_3D_TCTLPROG | NVC0_NEW_3D_TEVLPROG | NVC0_NEW_3D_GMTYPROG |
+ NVC0_NEW_3D_TFB_TARGETS | NVC0_NEW_3D_VERTEX | NVC0_NEW_3D_ARRAYS);
nvc0->scissors_dirty |= 1;
nvc0->viewports_dirty |= 1;
static void
nvc0_blit_3d(struct nvc0_context *nvc0, const struct pipe_blit_info *info)
{
+ struct nvc0_screen *screen = nvc0->screen;
struct nvc0_blitctx *blit = nvc0->blit;
struct nouveau_pushbuf *push = nvc0->base.pushbuf;
struct pipe_resource *src = info->src.resource;
blit->render_condition_enable = info->render_condition_enable;
nvc0_blit_select_fp(blit, info);
- nvc0_blitctx_pre_blit(blit);
+ nvc0_blitctx_pre_blit(blit, info);
nvc0_blit_set_dst(blit, dst, info->dst.level, -1, info->dst.format);
nvc0_blit_set_src(blit, src, info->src.level, -1, info->src.format,
nvc0_blitctx_prepare_state(blit);
- nvc0_state_validate(nvc0, ~0, 48);
+ nvc0_state_validate_3d(nvc0, ~0);
x_range = (float)info->src.box.width / (float)info->dst.box.width;
y_range = (float)info->src.box.height / (float)info->dst.box.height;
x0 = (float)info->src.box.x - x_range * (float)info->dst.box.x;
y0 = (float)info->src.box.y - y_range * (float)info->dst.box.y;
- x1 = x0 + 16384.0f * x_range;
- y1 = y0 + 16384.0f * y_range;
+ x1 = x0 + 32768.0f * x_range;
+ y1 = y0 + 32768.0f * y_range;
x0 *= (float)(1 << nv50_miptree(src)->ms_x);
x1 *= (float)(1 << nv50_miptree(src)->ms_x);
* render target, with scissors defining the destination region.
* The vertex is supplied with non-normalized texture coordinates
* arranged in a way to yield the desired offset and scale.
+ *
+ * Note that while the source texture is presented to the sampler as
+ * non-MSAA (even if it is), the destination texture is treated as MSAA for
+ * rendering. This means that
+ * - destination coordinates shouldn't be scaled
+ * - without per-sample rendering, the target will be a solid-fill for all
+ * of the samples
+ *
+ * The last point implies that this process is very bad for 1:1 blits, as
+ * well as scaled blits between MSAA surfaces. This works fine for
+ * upscaling and downscaling though. The 1:1 blits should ideally be
+ * handled by the 2d engine, which can do it perfectly.
*/
minx = info->dst.box.x;
return;
}
- BCTX_REFN_bo(nvc0->bufctx_3d, VTX_TMP, NOUVEAU_BO_GART | NOUVEAU_BO_RD, vtxbuf_bo);
+ BCTX_REFN_bo(nvc0->bufctx_3d, 3D_VTX_TMP,
+ NOUVEAU_BO_GART | NOUVEAU_BO_RD, vtxbuf_bo);
+ BCTX_REFN_bo(nvc0->bufctx_3d, 3D_TEXT,
+ NV_VRAM_DOMAIN(&screen->base) | NOUVEAU_BO_RD, screen->text);
nouveau_pushbuf_validate(push);
BEGIN_NVC0(push, NVC0_3D(VERTEX_ARRAY_FETCH(0)), 4);
NVC0_3D_VERTEX_ATTRIB_FORMAT_SIZE_32 |
NVC0_3D_VERTEX_ATTRIB_FORMAT_CONST);
}
+ for (i = 1; i < n; ++i)
+ IMMED_NVC0(push, NVC0_3D(VERTEX_ARRAY_FETCH(i)), 0);
if (nvc0->state.instance_elts) {
nvc0->state.instance_elts = 0;
BEGIN_NVC0(push, NVC0_3D(MACRO_VERTEX_ARRAY_PER_INSTANCE), 2);
}
nvc0->state.num_vtxelts = 2;
+ if (nvc0->state.prim_restart) {
+ IMMED_NVC0(push, NVC0_3D(PRIM_RESTART_ENABLE), 0);
+ nvc0->state.prim_restart = 0;
+ }
+
+ if (nvc0->state.index_bias) {
+ IMMED_NVC0(push, NVC0_3D(VB_ELEMENT_BASE), 0);
+ IMMED_NVC0(push, NVC0_3D(VERTEX_ID_BASE), 0);
+ nvc0->state.index_bias = 0;
+ }
+
for (i = 0; i < info->dst.box.depth; ++i, z += dz) {
if (info->dst.box.z + i) {
BEGIN_NVC0(push, NVC0_3D(LAYER), 1);
*(vbuf++) = fui(y0);
*(vbuf++) = fui(z);
- *(vbuf++) = fui(16384 << nv50_miptree(dst)->ms_x);
+ *(vbuf++) = fui(32768.0f);
*(vbuf++) = fui(0.0f);
*(vbuf++) = fui(x1);
*(vbuf++) = fui(y0);
*(vbuf++) = fui(z);
*(vbuf++) = fui(0.0f);
- *(vbuf++) = fui(16384 << nv50_miptree(dst)->ms_y);
+ *(vbuf++) = fui(32768.0f);
*(vbuf++) = fui(x0);
*(vbuf++) = fui(y1);
*(vbuf++) = fui(z);
struct nouveau_pushbuf *push = nvc0->base.pushbuf;
bool eng3d = false;
+ if (info->src.box.width == 0 || info->src.box.height == 0 ||
+ info->dst.box.width == 0 || info->dst.box.height == 0) {
+ pipe_debug_message(&nvc0->base.debug, ERROR,
+ "Blit with zero-size src or dst box");
+ return;
+ }
+
if (util_format_is_depth_or_stencil(info->dst.resource->format)) {
if (!(info->mask & PIPE_MASK_ZS))
return;
else
if (util_format_is_alpha(info->src.format))
eng3d = info->src.format != PIPE_FORMAT_A8_UNORM;
+ else
+ if (util_format_is_srgb(info->dst.format) &&
+ util_format_get_nr_components(info->src.format) == 1)
+ eng3d = true;
else
eng3d = !nv50_2d_format_supported(info->src.format);
}
info->src.box.height != -info->dst.box.height))
eng3d = true;
+ if (info->num_window_rectangles > 0 || info->window_rectangle_include)
+ eng3d = true;
+
if (nvc0->screen->num_occlusion_queries_active)
IMMED_NVC0(push, NVC0_3D(SAMPLECNT_ENABLE), 0);
}
screen->blitter->screen = screen;
- pipe_mutex_init(screen->blitter->mutex);
+ (void) mtx_init(&screen->blitter->mutex, mtx_plain);
nvc0_blitter_make_vp(screen->blitter);
nvc0_blitter_make_sampler(screen->blitter);
}
}
+ mtx_destroy(&blitter->mutex);
FREE(blitter);
}
pipe->clear_depth_stencil = nvc0_clear_depth_stencil;
pipe->clear_texture = nv50_clear_texture;
pipe->clear_buffer = nvc0_clear_buffer;
+ if (nvc0->screen->base.class_3d >= GM200_3D_CLASS)
+ pipe->evaluate_depth_buffer = gm200_evaluate_depth_buffer;
}
projects / mesa.git / blobdiff