[Mesa-dev] [PATCH 13/40] i965/blorp: Pipeline upload support for gen8
Pohjolainen, Topi
topi.pohjolainen at intel.com
Mon Apr 18 07:45:34 UTC 2016
On Mon, Apr 18, 2016 at 12:22:06AM -0700, Kenneth Graunke wrote:
> On Saturday, April 16, 2016 4:42:41 PM PDT Topi Pohjolainen wrote:
> > Signed-off-by: Topi Pohjolainen <topi.pohjolainen at intel.com>
> > ---
> > src/mesa/drivers/dri/i965/Makefile.sources | 1 +
> > src/mesa/drivers/dri/i965/brw_blorp.h | 3 +
> > src/mesa/drivers/dri/i965/gen8_blorp.cpp | 694 ++++++++++++++++++++++++++
> +++
> > 3 files changed, 698 insertions(+)
> > create mode 100644 src/mesa/drivers/dri/i965/gen8_blorp.cpp
> >
> > diff --git a/src/mesa/drivers/dri/i965/Makefile.sources b/src/mesa/drivers/
> dri/i965/Makefile.sources
> > index c314d74..0143c27 100644
> > --- a/src/mesa/drivers/dri/i965/Makefile.sources
> > +++ b/src/mesa/drivers/dri/i965/Makefile.sources
> > @@ -208,6 +208,7 @@ i965_FILES = \
> > gen7_wm_state.c \
> > gen7_wm_surface_state.c \
> > gen8_blend_state.c \
> > + gen8_blorp.cpp \
> > gen8_depth_state.c \
> > gen8_disable.c \
> > gen8_draw_upload.c \
> > diff --git a/src/mesa/drivers/dri/i965/brw_blorp.h b/src/mesa/drivers/dri/
> i965/brw_blorp.h
> > index 9a0fffc..8e30770 100644
> > --- a/src/mesa/drivers/dri/i965/brw_blorp.h
> > +++ b/src/mesa/drivers/dri/i965/brw_blorp.h
> > @@ -246,6 +246,9 @@ void
> > gen7_blorp_exec(struct brw_context *brw,
> > const brw_blorp_params *params);
> >
> > +void
> > +gen8_blorp_exec(struct brw_context *brw, const brw_blorp_params *params);
> > +
> > /**
> > * Parameters for a HiZ or depth resolve operation.
> > *
> > diff --git a/src/mesa/drivers/dri/i965/gen8_blorp.cpp b/src/mesa/drivers/
> dri/i965/gen8_blorp.cpp
> > new file mode 100644
> > index 0000000..0c15d36
> > --- /dev/null
> > +++ b/src/mesa/drivers/dri/i965/gen8_blorp.cpp
> > @@ -0,0 +1,694 @@
> > +/*
> > + * Copyright © 2016 Intel Corporation
> > + *
> > + * Permission is hereby granted, free of charge, to any person obtaining a
> > + * copy of this software and associated documentation files (the
> "Software"),
> > + * to deal in the Software without restriction, including without
> limitation
> > + * the rights to use, copy, modify, merge, publish, distribute, sublicense,
> > + * and/or sell copies of the Software, and to permit persons to whom the
> > + * Software is furnished to do so, subject to the following conditions:
> > + *
> > + * The above copyright notice and this permission notice (including the
> next
> > + * paragraph) shall be included in all copies or substantial portions of
> the
> > + * Software.
> > + *
> > + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
> OR
> > + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> > + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
> > + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
> OTHER
> > + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
> > + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
> DEALINGS
> > + * IN THE SOFTWARE.
> > + */
> > +
> > +#include <assert.h>
> > +
> > +#include "intel_batchbuffer.h"
> > +#include "intel_fbo.h"
> > +#include "intel_mipmap_tree.h"
> > +
> > +#include "brw_context.h"
> > +#include "brw_defines.h"
> > +#include "brw_state.h"
> > +
> > +#include "brw_blorp.h"
> > +
> > +
> > +/* SURFACE_STATE for renderbuffer or texture surface (see
> > + * brw_update_renderbuffer_surface and brw_update_texture_surface)
> > + */
> > +static uint32_t
> > +gen8_blorp_emit_surface_state(struct brw_context *brw,
> > + const brw_blorp_surface_info *surface,
> > + uint32_t read_domains, uint32_t write_domain,
> > + bool is_render_target)
> > +{
> > + uint32_t wm_surf_offset;
> > + const struct intel_mipmap_tree *mt = surface->mt;
> > + const uint32_t mocs_wb = is_render_target ?
> > + (brw->gen >= 9 ? SKL_MOCS_PTE :
> BDW_MOCS_PTE) :
> > + (brw->gen >= 9 ? SKL_MOCS_WB : BDW_MOCS_WB);
> > + const uint32_t tiling = surface->map_stencil_as_y_tiled
> > + ? I915_TILING_Y : mt->tiling;
> > + uint32_t tile_x, tile_y;
> > +
> > + uint32_t *surf = gen8_allocate_surface_state(brw, &wm_surf_offset, -1);
> > +
> > + surf[0] = BRW_SURFACE_2D << BRW_SURFACE_TYPE_SHIFT |
> > + surface->brw_surfaceformat << BRW_SURFACE_FORMAT_SHIFT |
> > + gen8_vertical_alignment(brw, mt, BRW_SURFACE_2D) |
> > + gen8_horizontal_alignment(brw, mt, BRW_SURFACE_2D) |
> > + gen8_surface_tiling_mode(tiling);
> > +
> > + surf[1] = SET_FIELD(mocs_wb, GEN8_SURFACE_MOCS) | mt->qpitch >> 2;
> > +
> > + surf[2] = SET_FIELD(surface->width - 1, GEN7_SURFACE_WIDTH) |
> > + SET_FIELD(surface->height - 1, GEN7_SURFACE_HEIGHT);
> > +
> > + uint32_t pitch_bytes = mt->pitch;
> > + if (surface->map_stencil_as_y_tiled)
> > + pitch_bytes *= 2;
> > + surf[3] = pitch_bytes - 1;
> > +
> > + surf[4] = gen7_surface_msaa_bits(surface->num_samples,
> > + surface->msaa_layout);
> > +
> > + if (surface->mt->mcs_mt) {
> > + surf[6] = SET_FIELD(surface->mt->qpitch / 4, GEN8_SURFACE_AUX_QPITCH)
> |
> > + SET_FIELD((surface->mt->mcs_mt->pitch / 128) - 1,
> > + GEN8_SURFACE_AUX_PITCH) |
> > + GEN8_SURFACE_AUX_MODE_MCS;
> > + } else {
> > + surf[6] = 0;
> > + }
> > +
> > + gen8_emit_fast_clear_color(brw, mt, surf);
> > + surf[7] |= SET_FIELD(HSW_SCS_RED, GEN7_SURFACE_SCS_R) |
> > + SET_FIELD(HSW_SCS_GREEN, GEN7_SURFACE_SCS_G) |
> > + SET_FIELD(HSW_SCS_BLUE, GEN7_SURFACE_SCS_B) |
> > + SET_FIELD(HSW_SCS_ALPHA, GEN7_SURFACE_SCS_A);
> > +
> > + /* reloc */
> > + *((uint64_t *)&surf[8]) =
> > + surface->compute_tile_offsets(&tile_x, &tile_y) + mt->bo->offset64;
> > +
> > + /* Note that the low bits of these fields are missing, so there's the
> > + * possibility of getting in trouble.
> > + */
> > + assert(tile_x % 4 == 0);
> > + assert(tile_y % 4 == 0);
> > + surf[5] = SET_FIELD(tile_x / 4, BRW_SURFACE_X_OFFSET) |
> > + SET_FIELD(tile_y / 4, GEN8_SURFACE_Y_OFFSET);
> > +
> > + if (brw->gen >= 9) {
> > + /* Disable Mip Tail by setting a large value. */
> > + surf[5] |= SET_FIELD(15, GEN9_SURFACE_MIP_TAIL_START_LOD);
> > + }
> > +
> > + if (surface->mt->mcs_mt) {
> > + *((uint64_t *) &surf[10]) = surface->mt->mcs_mt->bo->offset64;
> > + drm_intel_bo_emit_reloc(brw->batch.bo,
> > + wm_surf_offset + 10 * 4,
> > + surface->mt->mcs_mt->bo, 0,
> > + read_domains, write_domain);
> > + }
> > +
> > + /* Emit relocation to surface contents */
> > + drm_intel_bo_emit_reloc(brw->batch.bo,
> > + wm_surf_offset + 8 * 4,
> > + mt->bo,
> > + surf[8] - mt->bo->offset64,
> > + read_domains, write_domain);
> > +
> > + return wm_surf_offset;
> > +}
> > +
> > +static uint32_t
> > +gen8_blorp_emit_blend_state(struct brw_context *brw,
> > + const brw_blorp_params *params)
> > +{
> > + uint32_t blend_state_offset;
> > +
> > + assume(params->num_draw_buffers);
> > +
> > + const unsigned size = 4 + 8 * params->num_draw_buffers;
> > + uint32_t *blend = (uint32_t *)brw_state_batch(brw,
> AUB_TRACE_BLEND_STATE,
> > + size, 64,
> > + &blend_state_offset);
> > + memset(blend, 0, size);
> > +
> > + for (unsigned i = 0; i < params->num_draw_buffers; ++i) {
> > + blend[1 + 2 * i + 1] = GEN8_BLEND_PRE_BLEND_COLOR_CLAMP_ENABLE |
> > + GEN8_BLEND_POST_BLEND_COLOR_CLAMP_ENABLE |
> > + GEN8_BLEND_COLOR_CLAMP_RANGE_RTFORMAT;
> > + }
> > +
> > + return blend_state_offset;
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_disable_constant_state(struct brw_context *brw,
> > + unsigned opcode)
> > +{
> > + BEGIN_BATCH(11);
> > + OUT_BATCH(opcode << 16 | (11 - 2));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_disable_binding_table(struct brw_context *brw,
> > + unsigned opcode)
> > +{
> > +
> > + BEGIN_BATCH(2);
> > + OUT_BATCH(opcode << 16 | (2 - 2));
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +/* 3DSTATE_VS
> > + *
> > + * Disable vertex shader.
> > + */
> > +static void
> > +gen8_blorp_emit_vs_disable(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(9);
> > + OUT_BATCH(_3DSTATE_VS << 16 | (9 - 2));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +/* 3DSTATE_HS
> > + *
> > + * Disable the hull shader.
> > + */
> > +static void
> > +gen8_blorp_emit_hs_disable(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(9);
> > + OUT_BATCH(_3DSTATE_HS << 16 | (9 - 2));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +/* 3DSTATE_DS
> > + *
> > + * Disable the domain shader.
> > + */
> > +static void
> > +gen8_blorp_emit_ds_disable(struct brw_context *brw)
> > +{
> > + const int ds_pkt_len = brw->gen >= 9 ? 11 : 9;
> > + BEGIN_BATCH(ds_pkt_len);
> > + OUT_BATCH(_3DSTATE_DS << 16 | (ds_pkt_len - 2));
> > + for (int i = 0; i < ds_pkt_len - 1; i++)
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +/* 3DSTATE_GS
> > + *
> > + * Disable the geometry shader.
> > + */
> > +static void
> > +gen8_blorp_emit_gs_disable(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(10);
> > + OUT_BATCH(_3DSTATE_GS << 16 | (10 - 2));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +/* 3DSTATE_STREAMOUT
> > + *
> > + * Disable streamout.
> > + */
> > +static void
> > +gen8_blorp_emit_streamout_disable(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(5);
> > + OUT_BATCH(_3DSTATE_STREAMOUT << 16 | (5 - 2));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_raster_state(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(5);
> > + OUT_BATCH(_3DSTATE_RASTER << 16 | (5 - 2));
> > + OUT_BATCH(GEN8_RASTER_CULL_NONE | GEN8_RASTER_FRONT_WINDING_CCW);
>
> I don't think we should be setting GEN8_RASTER_FRONT_WINDING_CCW - we
> didn't set GEN6_SF_WINDING_CCW on Gen6-7.
I have no explanation for it either, I recall adding it when working on
meta+blorp. I'll drop it and re-check that nothing breaks.
>
> > + OUT_BATCH_F(0);
> > + OUT_BATCH_F(0);
> > + OUT_BATCH_F(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_sbe_state(struct brw_context *brw,
> > + const brw_blorp_params *params)
> > +{
> > + /* 3DSTATE_SBE */
> > + {
> > + const unsigned sbe_cmd_length = brw->gen == 8 ? 4 : 6;
> > + BEGIN_BATCH(sbe_cmd_length);
> > + OUT_BATCH(_3DSTATE_SBE << 16 | (sbe_cmd_length - 2));
> > + OUT_BATCH(GEN7_SBE_SWIZZLE_ENABLE |
> > + params->num_varyings << GEN7_SBE_NUM_OUTPUTS_SHIFT |
> > + 1 << GEN7_SBE_URB_ENTRY_READ_LENGTH_SHIFT |
> > + BRW_SF_URB_ENTRY_READ_OFFSET <<
> > + GEN8_SBE_URB_ENTRY_READ_OFFSET_SHIFT |
> > + GEN8_SBE_FORCE_URB_ENTRY_READ_LENGTH |
> > + GEN8_SBE_FORCE_URB_ENTRY_READ_OFFSET);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + if (sbe_cmd_length >= 6) {
> > + OUT_BATCH(GEN9_SBE_ACTIVE_COMPONENT_XYZW << (0 << 1));
> > + OUT_BATCH(0);
> > + }
> > + ADVANCE_BATCH();
> > + }
> > +
> > + {
> > + BEGIN_BATCH(11);
> > + OUT_BATCH(_3DSTATE_SBE_SWIZ << 16 | (11 - 2));
> > +
> > + /* Output DWords 1 through 8: */
> > + for (int i = 0; i < 8; i++) {
> > + OUT_BATCH(0);
> > + }
> > +
> > + OUT_BATCH(0); /* wrapshortest enables 0-7 */
> > + OUT_BATCH(0); /* wrapshortest enables 8-15 */
> > + ADVANCE_BATCH();
> > + }
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_sf_config(struct brw_context *brw)
> > +{
> > + /* See gen6_blorp_emit_sf_config() */
> > + BEGIN_BATCH(4);
> > + OUT_BATCH(_3DSTATE_SF << 16 | (4 - 2));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(GEN6_SF_LINE_AA_MODE_TRUE);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +/**
> > + * Disable thread dispatch (dw5.19) and enable the HiZ op.
> > + */
> > +static void
> > +gen8_blorp_emit_wm_state(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(2);
> > + OUT_BATCH(_3DSTATE_WM << 16 | (2 - 2));
> > + OUT_BATCH(GEN7_WM_LINE_AA_WIDTH_1_0 |
> > + GEN7_WM_LINE_END_CAP_AA_WIDTH_0_5 |
> > + GEN7_WM_POINT_RASTRULE_UPPER_RIGHT);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +/**
> > + * 3DSTATE_PS
> > + *
> > + * Pixel shader dispatch is disabled above in 3DSTATE_WM, dw1.29. Despite
> > + * that, thread dispatch info must still be specified.
> > + * - Maximum Number of Threads (dw4.24:31) must be nonzero, as the
> > + * valid range for this field is [0x3, 0x2f].
> > + * - A dispatch mode must be given; that is, at least one of the
> > + * "N Pixel Dispatch Enable" (N=8,16,32) fields must be set. This was
> > + * discovered through simulator error messages.
> > + */
> > +static void
> > +gen8_blorp_emit_ps_config(struct brw_context *brw,
> > + const brw_blorp_params *params,
> > + uint32_t prog_offset,
> > + brw_blorp_prog_data *prog_data)
> > +{
> > + uint32_t dw3, dw5, dw6, dw7;
> > +
> > + dw3 = dw5 = dw6 = dw7 = 0;
> > + dw3 |= GEN7_PS_VECTOR_MASK_ENABLE;
> > +
> > + dw6 |= GEN7_PS_16_DISPATCH_ENABLE;
> > +
> > + dw3 |= 1 << GEN7_PS_SAMPLER_COUNT_SHIFT; /* Up to 4 samplers */
> > + dw3 |= 2 << GEN7_PS_BINDING_TABLE_ENTRY_COUNT_SHIFT; /* Two surfaces */
> > + dw6 |= GEN7_PS_PUSH_CONSTANT_ENABLE;
> > + dw7 |= prog_data->first_curbe_grf << GEN7_PS_DISPATCH_START_GRF_SHIFT_0;
> > +
> > + /* 3DSTATE_PS expects the number of threads per PSD, which is always 64;
> > + * it implicitly scales for different GT levels (which have some # of
> PSDs).
> > + *
> > + * In Gen8 the format is U8-2 whereas in Gen9 it is U8-1.
> > + */
> > + if (brw->gen >= 9)
> > + dw6 |= (64 - 1) << HSW_PS_MAX_THREADS_SHIFT;
> > + else
> > + dw6 |= (64 - 2) << HSW_PS_MAX_THREADS_SHIFT;
> > +
> > + dw6 |= GEN7_PS_POSOFFSET_NONE;
> > +
> > + BEGIN_BATCH(12);
> > + OUT_BATCH(_3DSTATE_PS << 16 | (12 - 2));
> > + OUT_BATCH(prog_offset);
> > + OUT_BATCH(0);
> > + OUT_BATCH(dw3);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(dw6);
> > + OUT_BATCH(dw7);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_ps_blend(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(2);
> > + OUT_BATCH(_3DSTATE_PS_BLEND << 16 | (2 - 2));
> > + OUT_BATCH(GEN8_PS_BLEND_HAS_WRITEABLE_RT);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_ps_extra(struct brw_context *brw,
> > + const brw_blorp_params *params,
> > + const brw_blorp_prog_data *prog_data)
> > +{
> > + uint32_t dw1 = 0;
> > +
> > + dw1 |= GEN8_PSX_PIXEL_SHADER_VALID;
> > +
> > + if (params->src.mt) {
>
> This might be better as:
>
> if (params->use_wm_prog) {
>
> It should be equivalent, as I don't think we execute programs for
> clears, but...maybe a bit clearer? It's at least what Gen7 does.
We do execute them. We could skip for fast clear and resolves I guess but at
least normal rep-clear needs to run a shader.
>
> > + dw1 |= GEN8_PSX_KILL_ENABLE;
> > + dw1 |= GEN8_PSX_ATTRIBUTE_ENABLE;
> > + }
> > +
> > + if (params->dst.num_samples > 1 && prog_data &&
> > + prog_data->persample_msaa_dispatch)
> > + dw1 |= GEN8_PSX_SHADER_IS_PER_SAMPLE;
>
> Maybe put this block inside that the use_wm_prog condition, like Gen7...
Hmm, I wonder which packet are you referring to. In gen7 there is an odd
looking indentation in gen7_blorp_emit_wm_config() but it is really
controlled by the very same condition.
I dropped all the "use_wm_prog" comditions as gen8/9 blorp don't need them,
it is used to distuingish hiz execution. On gen8/9 there is entirely separate
path for hiz.
>
> > +
> > + BEGIN_BATCH(2);
> > + OUT_BATCH(_3DSTATE_PS_EXTRA << 16 | (2 - 2));
> > + OUT_BATCH(dw1);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_depth_disable(struct brw_context *brw)
> > +{
> > + /* Skip repeated NULL depth/stencil emits (think 2D rendering). */
> > + if (brw->no_depth_or_stencil)
> > + return;
> > +
> > + brw_emit_depth_stall_flushes(brw);
> > +
> > + BEGIN_BATCH(8);
> > + OUT_BATCH(GEN7_3DSTATE_DEPTH_BUFFER << 16 | (8 - 2));
> > + OUT_BATCH((BRW_DEPTHFORMAT_D32_FLOAT << 18) | (BRW_SURFACE_NULL << 29));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +
> > + BEGIN_BATCH(5);
> > + OUT_BATCH(GEN7_3DSTATE_HIER_DEPTH_BUFFER << 16 | (5 - 2));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +
> > + BEGIN_BATCH(5);
> > + OUT_BATCH(GEN7_3DSTATE_STENCIL_BUFFER << 16 | (5 - 2));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_vf_topology(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(2);
> > + OUT_BATCH(_3DSTATE_VF_TOPOLOGY << 16 | (2 - 2));
> > + OUT_BATCH(_3DPRIM_RECTLIST);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_vf_sys_gen_vals_state(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(2);
> > + OUT_BATCH(_3DSTATE_VF_SGVS << 16 | (2 - 2));
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_vf_instancing_state(struct brw_context *brw,
> > + unsigned num_elems)
> > +{
> > + for (unsigned i = 0; i < num_elems; ++i) {
> > + BEGIN_BATCH(3);
> > + OUT_BATCH(_3DSTATE_VF_INSTANCING << 16 | (3 - 2));
> > + OUT_BATCH(i);
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > + }
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_vf_state(struct brw_context *brw)
> > +{
> > + BEGIN_BATCH(2);
> > + OUT_BATCH(_3DSTATE_VF << 16 | (2 - 2));
> > + OUT_BATCH(0);
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_depth_stencil_state(struct brw_context *brw,
> > + const brw_blorp_params *params)
> > +{
> > + const unsigned pkt_len = brw->gen >= 9 ? 4 : 3;
> > +
> > + BEGIN_BATCH(pkt_len);
> > + OUT_BATCH(_3DSTATE_WM_DEPTH_STENCIL << 16 | (pkt_len - 2));
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + if (pkt_len > 3) {
> > + OUT_BATCH(0);
> > + }
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static void
> > +gen8_blorp_emit_constant_ps(struct brw_context *brw,
> > + uint32_t wm_push_const_offset)
> > +{
> > + const int dwords = brw->gen >= 8 ? 11 : 7;
> > + BEGIN_BATCH(dwords);
> > + OUT_BATCH(_3DSTATE_CONSTANT_PS << 16 | (dwords - 2));
> > +
> > + if (brw->gen >= 9) {
> > + OUT_BATCH(0);
> > + OUT_BATCH(BRW_BLORP_NUM_PUSH_CONST_REGS);
> > + } else {
> > + OUT_BATCH(BRW_BLORP_NUM_PUSH_CONST_REGS);
> > + OUT_BATCH(0);
> > + }
> > +
> > + if (brw->gen >= 9) {
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_RELOC64(brw->batch.bo, I915_GEM_DOMAIN_RENDER, 0,
> > + wm_push_const_offset);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + } else {
> > + OUT_BATCH(wm_push_const_offset);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + OUT_BATCH(0);
> > + }
> > +
> > + ADVANCE_BATCH();
> > +}
> > +
> > +static uint32_t
> > +gen8_blorp_emit_surface_states(struct brw_context *brw,
> > + const brw_blorp_params *params)
> > +{
> > + uint32_t wm_surf_offset_renderbuffer;
> > + uint32_t wm_surf_offset_texture = 0;
> > +
> > + intel_miptree_used_for_rendering(params->dst.mt);
> > +
> > + wm_surf_offset_renderbuffer =
> > + gen8_blorp_emit_surface_state(brw, ¶ms->dst,
> > + I915_GEM_DOMAIN_RENDER,
> > + I915_GEM_DOMAIN_RENDER,
> > + true /* is_render_target */);
> > + if (params->src.mt) {
> > + const brw_blorp_surface_info *surface = ¶ms->src;
> > + intel_mipmap_tree *mt = surface->mt;
> > +
> > + /* Textures are always sampled as 2D. */
> > + const bool is_cube = mt->target == GL_TEXTURE_CUBE_MAP_ARRAY ||
> > + mt->target == GL_TEXTURE_CUBE_MAP;
> > + const unsigned depth = (is_cube ? 6 : 1) * mt->logical_depth0;
> > + const GLenum target = is_cube ? GL_TEXTURE_2D_ARRAY : mt->target;
> > + const unsigned layer = surface->layer / MAX2(mt->num_samples, 1);
> > + const unsigned max_level = surface->level + mt->last_level + 1;
> > +
> > + brw->vtbl.emit_texture_surface_state(brw, mt, target,
> > + layer, layer + depth,
> > + surface->level, max_level,
> > + surface->brw_surfaceformat,
> > + SWIZZLE_XYZW,
> > + &wm_surf_offset_texture,
> > + -1, false, false);
> > + }
> > +
> > + return gen6_blorp_emit_binding_table(brw,
> > + wm_surf_offset_renderbuffer,
> > + wm_surf_offset_texture);
> > +}
> > +
> > +/**
> > + * \copydoc gen6_blorp_exec()
> > + */
> > +void
> > +gen8_blorp_exec(struct brw_context *brw, const brw_blorp_params *params)
> > +{
> > + brw_blorp_prog_data *prog_data = NULL;
> > + uint32_t wm_bind_bo_offset = 0;
> > +
> > + uint32_t prog_offset = params->get_wm_prog(brw, &prog_data);
> > +
> > + gen8_upload_state_base_address(brw);
> > + gen7_blorp_emit_cc_viewport(brw);
> > + gen7_l3_state.emit(brw);
> > +
> > + gen7_blorp_emit_urb_config(brw);
> > +
> > + const uint32_t cc_blend_state_offset =
> > + gen8_blorp_emit_blend_state(brw, params);
> > + gen7_blorp_emit_blend_state_pointer(brw, cc_blend_state_offset);
> > +
> > + const uint32_t cc_state_offset = gen6_blorp_emit_cc_state(brw);
> > + gen7_blorp_emit_cc_state_pointer(brw, cc_state_offset);
> > +
> > + gen8_blorp_emit_disable_constant_state(brw, _3DSTATE_CONSTANT_VS);
> > + gen8_blorp_emit_disable_constant_state(brw, _3DSTATE_CONSTANT_HS);
> > + gen8_blorp_emit_disable_constant_state(brw, _3DSTATE_CONSTANT_DS);
> > + gen8_blorp_emit_disable_constant_state(brw, _3DSTATE_CONSTANT_GS);
> > +
> > + const uint32_t wm_push_const_offset =
> > + gen6_blorp_emit_wm_constants(brw, params);
> > + gen8_blorp_emit_constant_ps(brw, wm_push_const_offset);
> > + wm_bind_bo_offset = gen8_blorp_emit_surface_states(brw, params);
> > +
> > + gen8_blorp_emit_disable_binding_table(brw,
> > +
> _3DSTATE_BINDING_TABLE_POINTERS_VS);
> > + gen8_blorp_emit_disable_binding_table(brw,
> > +
> _3DSTATE_BINDING_TABLE_POINTERS_HS);
> > + gen8_blorp_emit_disable_binding_table(brw,
> > +
> _3DSTATE_BINDING_TABLE_POINTERS_DS);
> > + gen8_blorp_emit_disable_binding_table(brw,
> > +
> _3DSTATE_BINDING_TABLE_POINTERS_GS);
> > +
> > + gen7_blorp_emit_binding_table_pointers_ps(brw, wm_bind_bo_offset);
> > + const uint32_t sampler_offset =
> > + gen6_blorp_emit_sampler_state(brw, BRW_MAPFILTER_LINEAR, 0, true);
> > + gen7_blorp_emit_sampler_state_pointers_ps(brw, sampler_offset);
> > +
> > + gen8_emit_3dstate_multisample(brw, params->dst.num_samples);
> > + gen6_emit_3dstate_sample_mask(brw,
> > + params->dst.num_samples > 1 ?
> > + (1 << params->dst.num_samples) - 1 :
> 1);
> > +
> > + gen8_disable_stages.emit(brw);
> > + gen8_blorp_emit_vs_disable(brw);
> > + gen8_blorp_emit_hs_disable(brw);
> > + gen7_blorp_emit_te_disable(brw);
> > + gen8_blorp_emit_ds_disable(brw);
> > + gen8_blorp_emit_gs_disable(brw);
> > +
> > + gen8_blorp_emit_streamout_disable(brw);
> > + gen6_blorp_emit_clip_disable(brw);
> > + gen8_blorp_emit_raster_state(brw);
> > + gen8_blorp_emit_sbe_state(brw, params);
> > + gen8_blorp_emit_sf_config(brw);
> > +
> > + gen8_blorp_emit_ps_blend(brw);
> > + gen8_blorp_emit_ps_extra(brw, params, prog_data);
> > +
> > + gen8_blorp_emit_ps_config(brw, params, prog_offset, prog_data);
> > +
> > + gen8_blorp_emit_depth_stencil_state(brw, params);
> > + gen8_blorp_emit_wm_state(brw);
> > +
> > + gen8_blorp_emit_depth_disable(brw);
> > + gen7_blorp_emit_clear_params(brw, params);
> > + gen6_blorp_emit_drawing_rectangle(brw, params);
> > + gen8_blorp_emit_vf_topology(brw);
> > + gen8_blorp_emit_vf_sys_gen_vals_state(brw);
> > + gen6_blorp_emit_vertices(brw, params);
> > + gen8_blorp_emit_vf_instancing_state(brw, 2);
> > + gen8_blorp_emit_vf_state(brw);
> > + gen7_blorp_emit_primitive(brw, params);
> > +}
> >
>
> I think we need to call write_pma_stall_bits(brw, 0) here. The GPU gets
> very upset if you leave the PMA stall bits configured incorrectly - and
> they might be enabled prior to a BLORP operation.
>
> Disabling it is always safe. It's also the right thing to do currently,
> since BLORP doesn't use the depth pipeline.
Right, I think considered it but I guess I was thinking too heavily in SKL
terms. There is the "brw->gen >= 9" condition in
gen8_emit_pma_stall_workaround(). I'll add it. Good catch.
More information about the mesa-dev
mailing list