KonstantinSeurer
/
mesa
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

anv: Add support for sample locations 

Allowing the user to set custom sample locations, by filling the
extension structs and chaining them to the pipeline structs according
to the Vulkan specification section [26.5. Custom Sample Locations]
for the following structures:

'VkPipelineSampleLocationsStateCreateInfoEXT'
'VkSampleLocationsInfoEXT'
'VkSampleLocationEXT'

Once custom locations are used, the default locations are lost and need
to be re-emitted again in the next pipeline creation. For that, we emit
the 3DSTATE_SAMPLE_PATTERN at every pipeline creation.

v2: In v1, we used the custom anv_sample struct to store the location
    and the distance from the pixel center because we would then use
    this distance to sort the locations and send them in increasing
    monotonical order to the GPU. That was because the Skylake PRM Vol.
    2a "3DSTATE_SAMPLE_PATTERN" says that the samples must have
    monotonically increasing distance from the pixel center to get the
    correct centroid computation in the device. However, the Vulkan
    spec seems to require that the samples occur in the order provided
    through the API and this requirement is only for the standard
    locations. As long as this only affects centroid calculations as
    the docs say, we should be ok because OpenGL and Vulkan only
    require that the centroid be some lit sample and that it's the same
    for all samples in a pixel; they have no requirement that it be the
    one closest to center. (Jason Ekstrand)
    For that we made the following changes:
    1- We removed the custom structs and functions from anv_private.h
       and anv_sample_locations.h and anv_sample_locations.c (the last
       two files were removed). (Jason Ekstrand)
    2- We modified the macros used to take also the array as parameter
       and we renamed them to start by GEN_. (Jason Ekstrand)
    3- We don't sort the samples anymore. (Jason Ekstrand)

v3 (Jason Ekstrand):
    Break the refactoring out into multiple commits

v4: Merge dynamic/non-dynamic changes into a single commit (Lionel)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/1887>
This commit is contained in:
Eleni Maria Stea 2019-03-14 21:52:02 +02:00 committed by Marge Bot
parent 43acc10bd0
commit 27ee40f4c9
6 changed files with 110 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
src/intel/vulkan/anv_cmd_buffer.c
View File

@ -188,6 +188,14 @@ anv_dynamic_state_copy(struct anv_dynamic_state *dest,
ANV_CMP_COPY(dyn_vbo_stride, ANV_CMD_DIRTY_DYNAMIC_VERTEX_INPUT_BINDING_STRIDE);
ANV_CMP_COPY(dyn_vbo_size, ANV_CMD_DIRTY_DYNAMIC_VERTEX_INPUT_BINDING_STRIDE);
if (copy_mask & ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS) {
dest->sample_locations.samples = src->sample_locations.samples;
typed_memcpy(dest->sample_locations.locations,
src->sample_locations.locations,
dest->sample_locations.samples);
changed |= ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS;
}
#undef ANV_CMP_COPY
return changed;
@ -687,6 +695,22 @@ void anv_CmdSetStencilReference(
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_STENCIL_REFERENCE;
}
void anv_CmdSetSampleLocationsEXT(
VkCommandBuffer commandBuffer,
const VkSampleLocationsInfoEXT* pSampleLocationsInfo)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
struct anv_dynamic_state *dyn_state = &cmd_buffer->state.gfx.dynamic;
uint32_t samples = pSampleLocationsInfo->sampleLocationsPerPixel;
dyn_state->sample_locations.samples = samples;
typed_memcpy(dyn_state->sample_locations.locations,
pSampleLocationsInfo->pSampleLocations, samples);
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS;
}
void anv_CmdSetLineStippleEXT(
VkCommandBuffer commandBuffer,
uint32_t lineStippleFactor,

29
src/intel/vulkan/anv_pipeline.c
View File

@ -31,6 +31,7 @@
#include "util/os_time.h"
#include "common/gen_l3_config.h"
#include "common/gen_disasm.h"
#include "common/gen_sample_positions.h"
#include "anv_private.h"
#include "compiler/brw_nir.h"
#include "anv_nir.h"
@ -2044,6 +2045,34 @@ copy_non_dynamic_state(struct anv_graphics_pipeline *pipeline,
}
}
if (states & ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS) {
const VkPipelineMultisampleStateCreateInfo *ms_info =
pCreateInfo->pMultisampleState;
const VkPipelineSampleLocationsStateCreateInfoEXT *sl_info = ms_info ?
vk_find_struct_const(ms_info, PIPELINE_SAMPLE_LOCATIONS_STATE_CREATE_INFO_EXT) : NULL;
if (sl_info) {
dynamic->sample_locations.samples =
sl_info->sampleLocationsInfo.sampleLocationsCount;
const VkSampleLocationEXT *positions =
sl_info->sampleLocationsInfo.pSampleLocations;
for (uint32_t i = 0; i < dynamic->sample_locations.samples; i++) {
dynamic->sample_locations.locations[i].x = positions[i].x;
dynamic->sample_locations.locations[i].y = positions[i].y;
}
} else {
dynamic->sample_locations.samples =
ms_info ? ms_info->rasterizationSamples : 1;
const struct gen_sample_position *positions =
gen_get_sample_positions(dynamic->sample_locations.samples);
for (uint32_t i = 0; i < dynamic->sample_locations.samples; i++) {
dynamic->sample_locations.locations[i].x = positions[i].x;
dynamic->sample_locations.locations[i].y = positions[i].y;
}
}
}
pipeline->dynamic_state_mask = states;
}

12
src/intel/vulkan/anv_private.h
View File

@ -189,6 +189,7 @@ struct gen_perf_query_result;
#define ANV_SSBO_ALIGNMENT 4
#define ANV_SSBO_BOUNDS_CHECK_ALIGNMENT 4
#define MAX_VIEWS_FOR_PRIMITIVE_REPLICATION 16
#define MAX_SAMPLE_LOCATIONS 16
/* From the Skylake PRM Vol. 7 "Binding Table Surface State Model":
*
@ -2372,6 +2373,7 @@ enum anv_cmd_dirty_bits {
ANV_CMD_DIRTY_DYNAMIC_DEPTH_BOUNDS_TEST_ENABLE = 1 << 21, /* VK_DYNAMIC_STATE_DEPTH_BOUNDS_TEST_ENABLE_EXT */
ANV_CMD_DIRTY_DYNAMIC_STENCIL_TEST_ENABLE = 1 << 22, /* VK_DYNAMIC_STATE_STENCIL_TEST_ENABLE_EXT */
ANV_CMD_DIRTY_DYNAMIC_STENCIL_OP = 1 << 23, /* VK_DYNAMIC_STATE_STENCIL_OP_EXT */
ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS = 1 << 24, /* VK_DYNAMIC_STATE_SAMPLE_LOCATIONS_EXT */
};
typedef uint32_t anv_cmd_dirty_mask_t;
@ -2395,7 +2397,8 @@ typedef uint32_t anv_cmd_dirty_mask_t;
ANV_CMD_DIRTY_DYNAMIC_DEPTH_COMPARE_OP | \
ANV_CMD_DIRTY_DYNAMIC_DEPTH_BOUNDS_TEST_ENABLE | \
ANV_CMD_DIRTY_DYNAMIC_STENCIL_TEST_ENABLE | \
ANV_CMD_DIRTY_DYNAMIC_STENCIL_OP)
ANV_CMD_DIRTY_DYNAMIC_STENCIL_OP | \
ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS)
static inline enum anv_cmd_dirty_bits
anv_cmd_dirty_bit_for_vk_dynamic_state(VkDynamicState vk_state)
@ -2443,6 +2446,8 @@ anv_cmd_dirty_bit_for_vk_dynamic_state(VkDynamicState vk_state)
return ANV_CMD_DIRTY_DYNAMIC_STENCIL_TEST_ENABLE;
case VK_DYNAMIC_STATE_STENCIL_OP_EXT:
return ANV_CMD_DIRTY_DYNAMIC_STENCIL_OP;
case VK_DYNAMIC_STATE_SAMPLE_LOCATIONS_EXT:
return ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS;
default:
assert(!"Unsupported dynamic state");
return 0;
@ -2769,6 +2774,11 @@ struct anv_dynamic_state {
uint16_t pattern;
} line_stipple;
struct {
uint32_t samples;
VkSampleLocationEXT locations[MAX_SAMPLE_LOCATIONS];
} sample_locations;
VkCullModeFlags cull_mode;
VkFrontFace front_face;
VkPrimitiveTopology primitive_topology;

7
src/intel/vulkan/gen7_cmd_buffer.c
View File

@ -345,6 +345,13 @@ genX(cmd_buffer_flush_dynamic_state)(struct anv_cmd_buffer *cmd_buffer)
cmd_buffer->state.gfx.primitive_topology = topology;
}
if (cmd_buffer->device->enabled_extensions.EXT_sample_locations &&
cmd_buffer->state.gfx.dirty & ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS) {
genX(emit_multisample)(&cmd_buffer->batch,
cmd_buffer->state.gfx.dynamic.sample_locations.samples,
cmd_buffer->state.gfx.dynamic.sample_locations.locations);
}
cmd_buffer->state.gfx.dirty = 0;
}

7
src/intel/vulkan/gen8_cmd_buffer.c
View File

@ -648,6 +648,13 @@ genX(cmd_buffer_flush_dynamic_state)(struct anv_cmd_buffer *cmd_buffer)
}
}
if (cmd_buffer->device->enabled_extensions.EXT_sample_locations &&
cmd_buffer->state.gfx.dirty & ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS) {
genX(emit_sample_pattern)(&cmd_buffer->batch,
cmd_buffer->state.gfx.dynamic.sample_locations.samples,
cmd_buffer->state.gfx.dynamic.sample_locations.locations);
}
cmd_buffer->state.gfx.dirty = 0;
}

36
src/intel/vulkan/genX_pipeline.c
View File

@ -734,11 +734,39 @@ emit_rs_state(struct anv_graphics_pipeline *pipeline,
static void
emit_ms_state(struct anv_graphics_pipeline *pipeline,
const VkPipelineMultisampleStateCreateInfo *info)
const VkPipelineMultisampleStateCreateInfo *info,
uint32_t dynamic_states)
{
uint32_t samples = info ? info->rasterizationSamples : 1;
/* If the sample locations are dynamic, 3DSTATE_MULTISAMPLE on Gen7/7.5
* will be emitted dynamically, so skip it here. On Gen8+
* 3DSTATE_SAMPLE_PATTERN will be emitted dynamically, so skip it here.
*/
if (!(dynamic_states & ANV_CMD_DIRTY_DYNAMIC_SAMPLE_LOCATIONS)) {
/* Only lookup locations if the extensions is active, otherwise the
* default ones will be used either at device initialization time or
* through 3DSTATE_MULTISAMPLE on Gen7/7.5 by passing NULL locations.
*/
if (pipeline->base.device->enabled_extensions.EXT_sample_locations) {
#if GEN_GEN >= 8
genX(emit_sample_pattern)(&pipeline->base.batch,
pipeline->dynamic_state.sample_locations.samples,
pipeline->dynamic_state.sample_locations.locations);
#endif
}
genX(emit_multisample)(&pipeline->base.batch, samples, NULL);
genX(emit_multisample)(&pipeline->base.batch,
pipeline->dynamic_state.sample_locations.samples,
pipeline->dynamic_state.sample_locations.locations);
} else {
/* On Gen8+ 3DSTATE_MULTISAMPLE does not hold anything we need to modify
* for sample locations, so we don't have to emit it dynamically.
*/
#if GEN_GEN >= 8
genX(emit_multisample)(&pipeline->base.batch,
info ? info->rasterizationSamples : 1,
NULL);
#endif
}
/* From the Vulkan 1.0 spec:
* If pSampleMask is NULL, it is treated as if the mask has all bits
@ -2264,7 +2292,7 @@ genX(graphics_pipeline_create)(
pCreateInfo->pRasterizationState,
ms_info, line_info, dynamic_states, pass, subpass,
urb_deref_block_size);
emit_ms_state(pipeline, ms_info);
emit_ms_state(pipeline, ms_info, dynamic_states);
emit_ds_state(pipeline, ds_info, dynamic_states, pass, subpass);
emit_cb_state(pipeline, cb_info, ms_info);
compute_kill_pixel(pipeline, ms_info, subpass);