VkDevice _device,
struct anv_pipeline_cache * cache,
const VkGraphicsPipelineCreateInfo* pCreateInfo,
- const struct anv_graphics_pipeline_create_info *extra,
const VkAllocationCallbacks* pAllocator,
VkPipeline* pPipeline)
{
ANV_FROM_HANDLE(anv_device, device, _device);
ANV_FROM_HANDLE(anv_render_pass, pass, pCreateInfo->renderPass);
+ const struct anv_physical_device *physical_device =
+ &device->instance->physicalDevice;
+ const struct gen_device_info *devinfo = &physical_device->info;
struct anv_subpass *subpass = &pass->subpasses[pCreateInfo->subpass];
struct anv_pipeline *pipeline;
VkResult result;
return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
result = anv_pipeline_init(pipeline, device, cache,
- pCreateInfo, extra, pAllocator);
+ pCreateInfo, pAllocator);
if (result != VK_SUCCESS) {
anv_free2(&device->alloc, pAllocator, pipeline);
return result;
}
assert(pCreateInfo->pVertexInputState);
- emit_vertex_input(pipeline, pCreateInfo->pVertexInputState, extra);
+ emit_vertex_input(pipeline, pCreateInfo->pVertexInputState);
assert(pCreateInfo->pRasterizationState);
emit_rs_state(pipeline, pCreateInfo->pRasterizationState,
- pass, subpass, extra);
+ pCreateInfo->pMultisampleState, pass, subpass);
emit_ds_state(pipeline, pCreateInfo->pDepthStencilState, pass, subpass);
emit_urb_setup(pipeline);
emit_3dstate_clip(pipeline, pCreateInfo->pViewportState,
- pCreateInfo->pRasterizationState, extra);
+ pCreateInfo->pRasterizationState);
emit_3dstate_streamout(pipeline, pCreateInfo->pRasterizationState);
- if (pCreateInfo->pMultisampleState &&
- pCreateInfo->pMultisampleState->rasterizationSamples > 1)
- anv_finishme("VK_STRUCTURE_TYPE_PIPELINE_MULTISAMPLE_STATE_CREATE_INFO");
-
- uint32_t samples = 1;
- uint32_t log2_samples = __builtin_ffs(samples) - 1;
-
- anv_batch_emit(&pipeline->batch, GENX(3DSTATE_MULTISAMPLE), ms) {
- ms.PixelLocation = PIXLOC_CENTER;
- ms.NumberofMultisamples = log2_samples;
- }
-
- anv_batch_emit(&pipeline->batch, GENX(3DSTATE_SAMPLE_MASK), sm) {
- sm.SampleMask = 0xff;
- }
+ emit_ms_state(pipeline, pCreateInfo->pMultisampleState);
const struct brw_vs_prog_data *vs_prog_data = get_vs_prog_data(pipeline);
gen7_emit_vs_workaround_flush(brw);
#endif
- if (pipeline->vs_vec4 == NO_KERNEL || (extra && extra->disable_vs))
+ if (pipeline->vs_vec4 == NO_KERNEL)
anv_batch_emit(&pipeline->batch, GENX(3DSTATE_VS), vs);
else
anv_batch_emit(&pipeline->batch, GENX(3DSTATE_VS), vs) {
vs.VertexURBEntryReadLength = vs_prog_data->base.urb_read_length;
vs.VertexURBEntryReadOffset = 0;
- vs.MaximumNumberofThreads = device->info.max_vs_threads - 1;
+ vs.MaximumNumberofThreads = devinfo->max_vs_threads - 1;
vs.StatisticsEnable = true;
vs.VSFunctionEnable = true;
}
const struct brw_gs_prog_data *gs_prog_data = get_gs_prog_data(pipeline);
- if (pipeline->gs_kernel == NO_KERNEL || (extra && extra->disable_vs)) {
+ if (pipeline->gs_kernel == NO_KERNEL) {
anv_batch_emit(&pipeline->batch, GENX(3DSTATE_GS), gs);
} else {
anv_batch_emit(&pipeline->batch, GENX(3DSTATE_GS), gs) {
gs.DispatchGRFStartRegisterforURBData =
gs_prog_data->base.base.dispatch_grf_start_reg;
- gs.MaximumNumberofThreads = device->info.max_gs_threads - 1;
+ gs.MaximumNumberofThreads = devinfo->max_gs_threads - 1;
/* This in the next dword on HSW. */
gs.ControlDataFormat = gs_prog_data->control_data_format;
gs.ControlDataHeaderSize = gs_prog_data->control_data_header_size_hwords;
* don't at least set the maximum number of threads.
*/
anv_batch_emit(&pipeline->batch, GENX(3DSTATE_PS), ps) {
- ps.MaximumNumberofThreads = device->info.max_wm_threads - 1;
+ ps.MaximumNumberofThreads = devinfo->max_wm_threads - 1;
}
} else {
const struct brw_wm_prog_data *wm_prog_data = get_wm_prog_data(pipeline);
anv_batch_emit(&pipeline->batch, GENX(3DSTATE_PS), ps) {
ps.KernelStartPointer0 = pipeline->ps_ksp0;
+ ps.KernelStartPointer1 = 0;
+ ps.KernelStartPointer2 = pipeline->ps_ksp0 + wm_prog_data->prog_offset_2;
ps.ScratchSpaceBasePointer = (struct anv_address) {
.bo = anv_scratch_pool_alloc(device, &device->scratch_pool,
.offset = 0,
};
ps.PerThreadScratchSpace = scratch_space(&wm_prog_data->base);
- ps.MaximumNumberofThreads = device->info.max_wm_threads - 1;
+ ps.MaximumNumberofThreads = devinfo->max_wm_threads - 1;
ps.PushConstantEnable = wm_prog_data->base.nr_params > 0;
ps.AttributeEnable = wm_prog_data->num_varying_inputs > 0;
ps.oMaskPresenttoRenderTarget = wm_prog_data->uses_omask;
wm_prog_data->base.dispatch_grf_start_reg,
ps.DispatchGRFStartRegisterforConstantSetupData1 = 0,
ps.DispatchGRFStartRegisterforConstantSetupData2 =
- wm_prog_data->dispatch_grf_start_reg_2,
+ wm_prog_data->dispatch_grf_start_reg_2;
/* Haswell requires the sample mask to be set in this packet as well as
* in 3DSTATE_SAMPLE_MASK; the values should match. */
/* _NEW_BUFFERS, _NEW_MULTISAMPLE */
-
- ps.KernelStartPointer1 = 0;
- ps.KernelStartPointer2 = pipeline->ps_ksp0 + wm_prog_data->prog_offset_2;
+#if GEN_IS_HASWELL
+ ps.SampleMask = 0xff;
+#endif
}
+ uint32_t samples = pCreateInfo->pMultisampleState ?
+ pCreateInfo->pMultisampleState->rasterizationSamples : 1;
+
/* FIXME-GEN7: This needs a lot more work, cf gen7 upload_wm_state(). */
anv_batch_emit(&pipeline->batch, GENX(3DSTATE_WM), wm) {
wm.StatisticsEnable = true;
}
wm.BarycentricInterpolationMode = wm_prog_data->barycentric_interp_modes;
+
+ wm.MultisampleRasterizationMode = samples > 1 ?
+ MSRASTMODE_ON_PATTERN : MSRASTMODE_OFF_PIXEL;
+ wm.MultisampleDispatchMode = ((samples == 1) ||
+ (samples > 1 && wm_prog_data->persample_dispatch)) ?
+ MSDISPMODE_PERSAMPLE : MSDISPMODE_PERPIXEL;
}
}