anv,iris: Fix input vertex max for tcs on gen12

gen12 does away with the single patch dispatch mode for tcs, and
increases some limits so that 8_patch mode can always work. Make the
necessary changes so we don't try to fall back to single patch mode.

Fixes KHR-GL46.tessellation_shader.single.max_patch_vertices and others

Fixes: 44754279ace7 ("intel/fs/gen12: Use TCS 8_PATCH mode.")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4843>

diff --git a/src/gallium/drivers/iris/iris_state.c b/src/gallium/drivers/iris/iris_state.c
index 0d434124aaf81e9b4be62721d03d39bbdc0abdf3..a5b825f1bafaeaad71cdd2b03278653e77bc7390 100644 (file)
--- a/src/gallium/drivers/iris/iris_state.c
+++ b/src/gallium/drivers/iris/iris_state.c
@@ -4178,6 +4178,8 @@ iris_store_tcs_state(struct iris_context *ice,
        *    more than 2 times the number of instance count.
        */
       assert((devinfo->max_tcs_threads / 2) > tcs_prog_data->instances);
+      hs.DispatchGRFStartRegisterForURBData = prog_data->dispatch_grf_start_reg & 0x1f;
+      hs.DispatchGRFStartRegisterForURBData5 = prog_data->dispatch_grf_start_reg >> 5;
 #endif
 
       hs.InstanceCount = tcs_prog_data->instances - 1;

diff --git a/src/intel/compiler/brw_vec4_tcs.cpp b/src/intel/compiler/brw_vec4_tcs.cpp
index 852c5a9865cd3425ae6365406a777b2f4d575aec..e9a6880cfbe46f21cb4c5a4cb080b75a5b3d698d 100644 (file)
--- a/src/intel/compiler/brw_vec4_tcs.cpp
+++ b/src/intel/compiler/brw_vec4_tcs.cpp
@@ -394,7 +394,7 @@ brw_compile_tcs(const struct brw_compiler *compiler,
 
    if (compiler->use_tcs_8_patch &&
        nir->info.tess.tcs_vertices_out <= (devinfo->gen >= 12 ? 32 : 16) &&
-       2 + has_primitive_id + key->input_vertices <= 31) {
+       2 + has_primitive_id + key->input_vertices <= (devinfo->gen >= 12 ? 63 : 31)) {
       /* 3DSTATE_HS imposes two constraints on using 8_PATCH mode. First, the
        * "Instance" field limits the number of output vertices to [1, 16] on
        * gen11 and below, or [1, 32] on gen12 and above. Secondly, the

diff --git a/src/intel/genxml/gen12.xml b/src/intel/genxml/gen12.xml
index 9be40f66003f888cf0156b17c840f86ef37b749b..3eebc30f4bef4f80423c74acab00da1e7bdffe7f 100644 (file)
--- a/src/intel/genxml/gen12.xml
+++ b/src/intel/genxml/gen12.xml
@@ -1964,7 +1964,7 @@
       <value name="9-12 Samplers" value="3"/>
       <value name="13-16 Samplers" value="4"/>
     </field>
-    <field name="Instance Count" start="64" end="67" type="uint"/>
+    <field name="Instance Count" start="64" end="68" type="uint"/>
     <field name="Maximum Number of Threads" start="72" end="80" type="uint"/>
     <field name="Statistics Enable" start="93" end="93" type="bool"/>
     <field name="Enable" start="95" end="95" type="bool"/>

diff --git a/src/intel/vulkan/genX_pipeline.c b/src/intel/vulkan/genX_pipeline.c
index c6f479168b6a5c0c11ef5e34d8646eec044cf160..5277d22489ef2b9af85740047199955998dd2616 100644 (file)
--- a/src/intel/vulkan/genX_pipeline.c
+++ b/src/intel/vulkan/genX_pipeline.c
@@ -1624,7 +1624,12 @@ emit_3dstate_hs_te_ds(struct anv_graphics_pipeline *pipeline,
       hs.VertexURBEntryReadLength = 0;
       hs.VertexURBEntryReadOffset = 0;
       hs.DispatchGRFStartRegisterForURBData =
-         tcs_prog_data->base.base.dispatch_grf_start_reg;
+         tcs_prog_data->base.base.dispatch_grf_start_reg & 0x1f;
+#if GEN_GEN >= 12
+      hs.DispatchGRFStartRegisterForURBData5 =
+         tcs_prog_data->base.base.dispatch_grf_start_reg >> 5;
+#endif
+
 
       hs.PerThreadScratchSpace = get_scratch_space(tcs_bin);
       hs.ScratchSpaceBasePointer =