intel: Add driver support for hardware generated local invocation IDs

This adds a few new fields in the brw_cs_prog_data struct and then
uses them to fill in the relevant COMPUTE_WALKER fields.

Although the Tile Layout field theoretically has different settings for
32/64/128bpe, it appears that the recommended programming is to always
pick either TileY 32bpe or Linear.  It's not very practical to look at
the surface formats involved, anyway.

Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27167>

src/gallium/drivers/iris/iris_state.c

@@ -8677,6 +8677,14 @@ iris_upload_compute_walker(struct iris_context *ice,
          cw.PostSync.MOCS                  = iris_mocs(NULL, &screen->isl_dev, 0);
          cw.InterfaceDescriptor            = idd;
 
+#if GFX_VERx10 >= 125
+         cw.GenerateLocalID = cs_prog_data->generate_local_id != 0;
+         cw.EmitLocal       = cs_prog_data->generate_local_id;
+         cw.WalkOrder       = cs_prog_data->walk_order;
+         cw.TileLayout = cs_prog_data->walk_order == BRW_WALK_ORDER_YXZ ?
+                         TileY32bpe : Linear;
+#endif
+
          assert(brw_cs_push_const_total_size(cs_prog_data, dispatch.threads) == 0);
       }
    }

src/intel/blorp/blorp_genX_exec.h

@@ -2191,6 +2191,14 @@ blorp_exec_compute(struct blorp_batch *batch, const struct blorp_params *params)
       cw.IndirectDataStartAddress       = push_const_offset;
       cw.IndirectDataLength             = push_const_size;
 
+#if GFX_VERx10 >= 125
+      cw.GenerateLocalID                = cs_prog_data->generate_local_id != 0;
+      cw.EmitLocal                      = cs_prog_data->generate_local_id;
+      cw.WalkOrder                      = cs_prog_data->walk_order;
+      cw.TileLayout = cs_prog_data->walk_order == BRW_WALK_ORDER_YXZ ?
+                      TileY32bpe : Linear;
+#endif
+
       cw.InterfaceDescriptor = (struct GENX(INTERFACE_DESCRIPTOR_DATA)) {
          .KernelStartPointer = params->cs_prog_kernel,
          .SamplerStatePointer = samplers_offset,

src/intel/compiler/brw_compiler.h

@@ -1322,6 +1322,15 @@ struct brw_push_const_block {
    unsigned size;       /* Bytes, register aligned */
 };
 
+enum PACKED brw_compute_walk_order {
+   BRW_WALK_ORDER_XYZ = 0,
+   BRW_WALK_ORDER_XZY = 1,
+   BRW_WALK_ORDER_YXZ = 2,
+   BRW_WALK_ORDER_YZX = 3,
+   BRW_WALK_ORDER_ZXY = 4,
+   BRW_WALK_ORDER_ZYX = 5,
+};
+
 struct brw_cs_prog_data {
    struct brw_stage_prog_data base;
 
@@ -1344,6 +1353,8 @@ struct brw_cs_prog_data {
    bool uses_inline_data;
    bool uses_btd_stack_ids;
    bool uses_systolic;
+   uint8_t generate_local_id;
+   enum brw_compute_walk_order walk_order;
 
    struct {
       struct brw_push_const_block cross_thread;

src/intel/vulkan/genX_cmd_compute.c

@@ -359,6 +359,11 @@ emit_compute_walker(struct anv_cmd_buffer *cmd_buffer,
 #if GFX_VERx10 == 125
          .SystolicModeEnable             = prog_data->uses_systolic,
 #endif
+         .GenerateLocalID                = prog_data->generate_local_id != 0,
+         .EmitLocal                      = prog_data->generate_local_id,
+         .WalkOrder                      = prog_data->walk_order,
+         .TileLayout = prog_data->walk_order == BRW_WALK_ORDER_YXZ ?
+                       TileY32bpe : Linear,
          .LocalXMaximum                  = prog_data->local_size[0] - 1,
          .LocalYMaximum                  = prog_data->local_size[1] - 1,
          .LocalZMaximum                  = prog_data->local_size[2] - 1,

src/intel/vulkan/genX_simple_shader.c

@@ -557,6 +557,14 @@ genX(emit_simple_shader_dispatch)(struct anv_simple_shader *state,
          cw.ExecutionMask                  = dispatch.right_mask;
          cw.PostSync.MOCS                  = anv_mocs(device, NULL, 0);
 
+#if GFX_VERx10 >= 125
+         cw.GenerateLocalID                = prog_data->generate_local_id != 0;
+         cw.EmitLocal                      = prog_data->generate_local_id;
+         cw.WalkOrder                      = prog_data->walk_order;
+         cw.TileLayout = prog_data->walk_order == BRW_WALK_ORDER_YXZ ?
+                         TileY32bpe : Linear;
+#endif
+
          cw.InterfaceDescriptor = (struct GENX(INTERFACE_DESCRIPTOR_DATA)) {
             .KernelStartPointer                = state->kernel->kernel.offset +
                                                  brw_cs_prog_data_prog_offset(prog_data,