nouveau: Add support for SV_WORK_DIM

Add support for SV_WORK_DIM for nvc0 and nve4.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
This commit is contained in:
Hans de Goede 2016-04-25 11:42:12 +02:00
parent 3345f70f63
commit 2aa1197eee
8 changed files with 29 additions and 12 deletions

View file

@ -461,6 +461,7 @@ enum SVSemantic
SV_BASEVERTEX,
SV_BASEINSTANCE,
SV_DRAWID,
SV_WORK_DIM,
SV_UNDEFINED,
SV_LAST
};

View file

@ -408,6 +408,7 @@ static nv50_ir::SVSemantic translateSysVal(uint sysval)
case TGSI_SEMANTIC_BASEVERTEX: return nv50_ir::SV_BASEVERTEX;
case TGSI_SEMANTIC_BASEINSTANCE: return nv50_ir::SV_BASEINSTANCE;
case TGSI_SEMANTIC_DRAWID: return nv50_ir::SV_DRAWID;
case TGSI_SEMANTIC_WORK_DIM: return nv50_ir::SV_WORK_DIM;
default:
assert(0);
return nv50_ir::SV_CLOCK;

View file

@ -2388,6 +2388,8 @@ NVC0LoweringPass::handleRDSV(Instruction *i)
i->setSrc(0, bld.mkImm(sv == SV_GRIDID ? 0 : 1));
return true;
}
// Fallthrough
case SV_WORK_DIM:
addr += prog->driver->prop.cp.gridInfoBase;
bld.mkLoad(TYPE_U32, i->getDef(0),
bld.mkSymbol(FILE_MEMORY_CONST, prog->driver->io.auxCBSlot,

View file

@ -295,6 +295,7 @@ TargetNVC0::getSVAddress(DataFile shaderFile, const Symbol *sym) const
case SV_NTID: return kepler ? (0x00 + idx * 4) : ~0;
case SV_NCTAID: return kepler ? (0x0c + idx * 4) : ~0;
case SV_GRIDID: return kepler ? 0x18 : ~0;
case SV_WORK_DIM: return 0x1c;
case SV_SAMPLE_INDEX: return 0;
case SV_SAMPLE_POS: return 0;
case SV_SAMPLE_MASK: return 0;

View file

@ -380,7 +380,8 @@ nvc0_state_validate_cp(struct nvc0_context *nvc0, uint32_t mask)
}
static void
nvc0_compute_upload_input(struct nvc0_context *nvc0, const void *input)
nvc0_compute_upload_input(struct nvc0_context *nvc0,
const struct pipe_grid_info *info)
{
struct nouveau_pushbuf *push = nvc0->base.pushbuf;
struct nvc0_screen *screen = nvc0->screen;
@ -399,13 +400,23 @@ nvc0_compute_upload_input(struct nvc0_context *nvc0, const void *input)
/* NOTE: size is limited to 4 KiB, which is < NV04_PFIFO_MAX_PACKET_LEN */
BEGIN_1IC0(push, NVC0_CP(CB_POS), 1 + cp->parm_size / 4);
PUSH_DATA (push, 0);
PUSH_DATAp(push, input, cp->parm_size / 4);
PUSH_DATAp(push, info->input, cp->parm_size / 4);
nvc0_compute_invalidate_constbufs(nvc0);
BEGIN_NVC0(push, NVC0_CP(FLUSH), 1);
PUSH_DATA (push, NVC0_COMPUTE_FLUSH_CB);
}
BEGIN_NVC0(push, NVC0_CP(CB_SIZE), 3);
PUSH_DATA (push, 2048);
PUSH_DATAh(push, screen->uniform_bo->offset + NVC0_CB_AUX_INFO(5));
PUSH_DATA (push, screen->uniform_bo->offset + NVC0_CB_AUX_INFO(5));
BEGIN_1IC0(push, NVC0_CP(CB_POS), 1 + 1);
/* (7) as we only upload work_dim on nvc0, the rest uses special regs */
PUSH_DATA (push, NVC0_CB_AUX_GRID_INFO(7));
PUSH_DATA (push, info->work_dim);
BEGIN_NVC0(push, NVC0_CP(FLUSH), 1);
PUSH_DATA (push, NVC0_COMPUTE_FLUSH_CB);
}
void
@ -422,7 +433,7 @@ nvc0_launch_grid(struct pipe_context *pipe, const struct pipe_grid_info *info)
return;
}
nvc0_compute_upload_input(nvc0, info->input);
nvc0_compute_upload_input(nvc0, info);
BEGIN_NVC0(push, NVC0_CP(CP_START_ID), 1);
PUSH_DATA (push, nvc0_program_symbol_offset(cp, info->pc));

View file

@ -114,9 +114,9 @@
/* 8 sets of 32-bits coordinate offsets */
#define NVC0_CB_AUX_MS_INFO 0x0a0
#define NVC0_CB_AUX_MS_SIZE (8 * 2 * 4)
/* block/grid size, at 3 32-bits integers each and gridid */
/* block/grid size, at 3 32-bits integers each, gridid and work_dim */
#define NVC0_CB_AUX_GRID_INFO(i) 0x0e0 + (i) * 4 /* CP */
#define NVC0_CB_AUX_GRID_SIZE (7 * 4)
#define NVC0_CB_AUX_GRID_SIZE (8 * 4)
/* 8 user clip planes, at 4 32-bits floats each */
#define NVC0_CB_AUX_UCP_INFO 0x100
#define NVC0_CB_AUX_UCP_SIZE (PIPE_MAX_CLIP_PLANES * 4 * 4)

View file

@ -569,9 +569,9 @@ nvc0_program_translate(struct nvc0_program *prog, uint16_t chipset,
if (chipset >= NVISA_GK104_CHIPSET) {
info->io.auxCBSlot = 7;
info->io.msInfoCBSlot = 7;
info->prop.cp.gridInfoBase = NVC0_CB_AUX_GRID_INFO(0);
info->io.uboInfoBase = NVC0_CB_AUX_UBO_INFO(0);
}
info->prop.cp.gridInfoBase = NVC0_CB_AUX_GRID_INFO(0);
} else {
info->io.sampleInfoBase = NVC0_CB_AUX_SAMPLE_INFO;
}

View file

@ -434,7 +434,7 @@ nve4_compute_upload_input(struct nvc0_context *nvc0,
PUSH_DATAh(push, address + NVC0_CB_AUX_GRID_INFO(0));
PUSH_DATA (push, address + NVC0_CB_AUX_GRID_INFO(0));
BEGIN_NVC0(push, NVE4_CP(UPLOAD_LINE_LENGTH_IN), 2);
PUSH_DATA (push, 7 * 4);
PUSH_DATA (push, 8 * 4);
PUSH_DATA (push, 0x1);
if (unlikely(info->indirect)) {
@ -444,18 +444,19 @@ nve4_compute_upload_input(struct nvc0_context *nvc0,
nouveau_pushbuf_space(push, 16, 0, 1);
PUSH_REFN(push, res->bo, NOUVEAU_BO_RD | res->domain);
BEGIN_1IC0(push, NVE4_CP(UPLOAD_EXEC), 1 + 7);
BEGIN_1IC0(push, NVE4_CP(UPLOAD_EXEC), 1 + 8);
PUSH_DATA (push, NVE4_COMPUTE_UPLOAD_EXEC_LINEAR | (0x20 << 1));
PUSH_DATAp(push, info->block, 3);
nouveau_pushbuf_data(push, res->bo, offset,
NVC0_IB_ENTRY_1_NO_PREFETCH | 3 * 4);
} else {
BEGIN_1IC0(push, NVE4_CP(UPLOAD_EXEC), 1 + 7);
BEGIN_1IC0(push, NVE4_CP(UPLOAD_EXEC), 1 + 8);
PUSH_DATA (push, NVE4_COMPUTE_UPLOAD_EXEC_LINEAR | (0x20 << 1));
PUSH_DATAp(push, info->block, 3);
PUSH_DATAp(push, info->grid, 3);
}
PUSH_DATA (push, 0);
PUSH_DATA (push, info->work_dim);
BEGIN_NVC0(push, NVE4_CP(FLUSH), 1);
PUSH_DATA (push, NVE4_COMPUTE_FLUSH_CB);