OpenHarmony/third_party_mesa3d
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

radv: no need to do gs_alloc_req for newer chips in ngg vs/tes

Browse Source 
Copy from radeonsi.

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17130>
This commit is contained in:
Qiang Yu
2022-06-17 15:36:08 +08:00
committed by Marge Bot
parent 74e596a5f0
commit 3aa70d92ce
5 changed files with 23 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
src/amd/vulkan/radv_nir_to_llvm.c
View File

@@ -1278,8 +1278,10 @@ handle_ngg_outputs_post_2(struct radv_shader_context *ctx)
/* TODO: primitive culling */
ac_build_sendmsg_gs_alloc_req(&ctx->ac, get_wave_id_in_tg(ctx), ngg_get_vtx_cnt(ctx),
ngg_get_prim_cnt(ctx));
/* Newer chips can use PRIMGEN_PASSTHRU_NO_MSG to skip gs_alloc_req for NGG passthrough. */
if (!(ctx->shader_info->is_ngg_passthrough && ctx->ac.family >= CHIP_NAVI23))
ac_build_sendmsg_gs_alloc_req(&ctx->ac, get_wave_id_in_tg(ctx), ngg_get_vtx_cnt(ctx),
ngg_get_prim_cnt(ctx));
/* TODO: streamout queries */
/* Export primitive data to the index buffer.