Enable shim v2 build

Summary:
Follow up of pytorch#125087

This diff will create shim v2 header and cpp file and corresponding build

Differential Revision: D56617546

buckbuild.bzl

@@ -385,6 +385,8 @@ def get_aten_generated_files(enabled_backends):
         "core/TensorMethods.cpp",
         "core/aten_interned_strings.h",
         "core/enum_tag.h",
+        "torch/csrc/inductor/aoti_torch/generated/c_shim_cpu.cpp",
+        "torch/csrc/inductor/aoti_torch/generated/c_shim_cpu.h",
     ] + get_aten_derived_type_srcs(enabled_backends)
 
     # This is tiresome.  A better strategy would be to unconditionally
@@ -469,6 +471,7 @@ def gen_aten_files(
         cmd = "$(exe {}torchgen:gen) ".format(ROOT_PATH) + " ".join([
             "--source-path $(location {}:aten_src_path)/aten/src/ATen".format(ROOT),
             "--install_dir $OUT",
+            "--aoti_install_dir $OUT/torch/csrc/inductor/aoti_torch/generated"
         ] + extra_params),
         visibility = visibility,
         compatible_with = compatible_with,
@@ -1008,6 +1011,16 @@ def define_buck_targets(
         visibility = ["PUBLIC"],
     )
 
+    fb_xplat_cxx_library(
+        name = "generated-aoti-cpu-headers",
+        header_namespace = "",
+        exported_headers = {
+            "torch/csrc/inductor/aoti_torch/generated/c_shim_cpu.h": ":gen_aten[torch/csrc/inductor/aoti_torch/generated/c_shim_cpu.h]",
+        },
+        labels = labels,
+        visibility = ["PUBLIC"],
+    )
+
     fb_xplat_cxx_library(
         name = "generated-version-header",
         header_namespace = "torch",

build.bzl

@@ -73,6 +73,7 @@ def define_targets(rules):
         "$(execpath //torchgen:gen)",
         "--install_dir=$(RULEDIR)",
         "--source-path aten/src/ATen",
+        "--aoti_install_dir=$(RULEDIR)/torch/csrc/inductor/aoti_torch/generated"
     ] + (["--static_dispatch_backend CPU"] if rules.is_cpu_static_dispatch_build() else []))
 
     gen_aten_outs_cuda = (
@@ -83,6 +84,7 @@ def define_targets(rules):
     gen_aten_outs = (
         GENERATED_H + GENERATED_H_CORE +
         GENERATED_CPP + GENERATED_CPP_CORE +
+        GENERATED_AOTI_H + GENERATED_AOTI_CPP +
         aten_ufunc_generated_cpu_sources() +
         aten_ufunc_generated_cpu_kernel_sources() + [
             "Declarations.yaml",
@@ -316,3 +318,13 @@ GENERATED_AUTOGRAD_CPP = [
     "torch/csrc/lazy/generated/RegisterAutogradLazy.cpp",
     "torch/csrc/lazy/generated/RegisterLazy.cpp",
 ] + _GENERATED_AUTOGRAD_CPP_HEADERS + GENERATED_LAZY_H
+
+GENERATED_AOTI_H = [
+    "torch/csrc/inductor/aoti_torch/generated/c_shim_cpu.h",
+    "torch/csrc/inductor/aoti_torch/generated/c_shim_cuda.h",
+]
+
+GENERATED_AOTI_CPP = [
+    "torch/csrc/inductor/aoti_torch/generated/c_shim_cpu.cpp",
+    "torch/csrc/inductor/aoti_torch/generated/c_shim_cuda.cpp",
+]

build_variables.bzl

@@ -471,6 +471,7 @@ inductor_core_resources = [
     "torch/csrc/inductor/aoti_torch/shim_common.cpp",
     "torch/csrc/inductor/aoti_torch/tensor_converter.cpp",
     "torch/csrc/inductor/inductor_ops.cpp",
+    ":gen_aten[torch/csrc/inductor/aoti_torch/generated/c_shim_cpu.cpp]",
 ]
 
 libtorch_core_sources = sorted(