Move threadpool to extension

Summary:
1. Replace `executorch/backends/xnnpack/threadpool` with `executorch/extension/threadpool`
2. `hg mv xplat/executorch/backends/xnnpack/threadpool xplat/executorch/extension` and `hg mv fbcode/executorch/backends/xnnpack/threadpool fbcode/executorch/extension`

Differential Revision: D61938189

Author: kirklandsign

backends/xnnpack/runtime/XNNCompiler.cpp

@@ -9,7 +9,7 @@
 #include <executorch/backends/xnnpack/runtime/XNNCompiler.h>
 #include <executorch/backends/xnnpack/runtime/XNNHeader.h>
 #include <executorch/backends/xnnpack/serialization/schema_generated.h>
-#include <executorch/backends/xnnpack/threadpool/threadpool.h>
+#include <executorch/extension/threadpool/threadpool.h>
 #include <executorch/runtime/core/exec_aten/util/scalar_type_util.h>
 #include <unordered_map>
 

backends/xnnpack/targets.bzl

@@ -47,7 +47,7 @@ def define_common_targets():
         deps = [
             third_party_dep("XNNPACK"),
             "//executorch/backends/xnnpack/serialization:xnnpack_flatbuffer_header",
-            "//executorch/backends/xnnpack/threadpool:threadpool",
+            "//executorch/extension/threadpool:threadpool",
             "//executorch/runtime/core/exec_aten/util:tensor_util",
         ],
         # XnnpackBackend.cpp needs to compile with executor as whole

backends/xnnpack/threadpool/test/targets.bzl

@@ -15,6 +15,6 @@ def define_common_targets():
         name = "threadpool_test",
         srcs = _THREADPOOL_TESTS,
         deps = [
-            "//executorch/backends/xnnpack/threadpool:threadpool",
+            "//executorch/extension/threadpool:threadpool",
         ],
     )

backends/xnnpack/threadpool/test/threadpool_test.cpp

@@ -11,8 +11,8 @@
 #include <numeric>
 #include <random>
 
-#include <executorch/backends/xnnpack/threadpool/threadpool.h>
-#include <executorch/backends/xnnpack/threadpool/threadpool_guard.h>
+#include <executorch/extension/threadpool/threadpool.h>
+#include <executorch/extension/threadpool/threadpool_guard.h>
 
 using namespace ::testing;
 

backends/xnnpack/threadpool/threadpool.cpp

@@ -6,8 +6,8 @@
  * LICENSE file in the root directory of this source tree.
  */
 
-#include <executorch/backends/xnnpack/threadpool/threadpool.h>
-#include <executorch/backends/xnnpack/threadpool/threadpool_guard.h>
+#include <executorch/extension/threadpool/threadpool.h>
+#include <executorch/extension/threadpool/threadpool_guard.h>
 #include <executorch/runtime/platform/assert.h>
 #include <algorithm>
 

backends/xnnpack/threadpool/threadpool_guard.cpp

@@ -6,7 +6,7 @@
  * LICENSE file in the root directory of this source tree.
  */
 
-#include <executorch/backends/xnnpack/threadpool/threadpool_guard.h>
+#include <executorch/extension/threadpool/threadpool_guard.h>
 
 namespace torch {
 namespace executorch {

configurations/targets.bzl

@@ -20,7 +20,7 @@ def define_common_targets():
     runtime.cxx_library(
         name = "executor_cpu_optimized",
         exported_deps = [
-            "//executorch/backends/xnnpack/threadpool:threadpool",
+            "//executorch/extension/threadpool:threadpool",
         ] + get_all_cpu_backend_targets(),
         visibility = [
             "//executorch/test/...",

examples/models/llama2/main.cpp

@@ -11,8 +11,8 @@
 #include <executorch/examples/models/llama2/runner/runner.h>
 
 #if defined(ET_USE_THREADPOOL)
-#include <executorch/backends/xnnpack/threadpool/cpuinfo_utils.h>
-#include <executorch/backends/xnnpack/threadpool/threadpool.h>
+#include <executorch/extension/threadpool/cpuinfo_utils.h>
+#include <executorch/extension/threadpool/threadpool.h>
 #endif
 
 DEFINE_string(

examples/models/llama2/targets.bzl

@@ -17,8 +17,8 @@ def define_common_targets():
                 deps = [
                     "//executorch/examples/models/llama2/runner:runner" + aten_suffix,
                     "//executorch/extension/evalue_util:print_evalue",
-                    "//executorch/backends/xnnpack/threadpool:threadpool",
-                    "//executorch/backends/xnnpack/threadpool:cpuinfo_utils",
+                    "//executorch/extension/threadpool:threadpool",
+                    "//executorch/extension/threadpool:cpuinfo_utils",
                 ],
                 external_deps = [
                     "gflags",

examples/models/llava/main.cpp

@@ -11,8 +11,8 @@
 #include <torch/torch.h>
 
 #if defined(ET_USE_THREADPOOL)
-#include <executorch/backends/xnnpack/threadpool/cpuinfo_utils.h>
-#include <executorch/backends/xnnpack/threadpool/threadpool.h>
+#include <executorch/extension/threadpool/cpuinfo_utils.h>
+#include <executorch/extension/threadpool/threadpool.h>
 #endif
 
 DEFINE_string(

extension/android/jni/jni_layer_llama.cpp

@@ -23,8 +23,8 @@
 #include <executorch/runtime/platform/runtime.h>
 
 #if defined(ET_USE_THREADPOOL)
-#include <executorch/backends/xnnpack/threadpool/cpuinfo_utils.h>
-#include <executorch/backends/xnnpack/threadpool/threadpool.h>
+#include <executorch/extension/threadpool/cpuinfo_utils.h>
+#include <executorch/extension/threadpool/threadpool.h>
 #endif
 
 #include <fbjni/ByteBuffer.h>

extension/llm/custom_ops/op_sdpa.cpp

@@ -20,7 +20,7 @@
 #include <vector>
 
 #ifdef ET_USE_THREADPOOL
-#include <executorch/backends/xnnpack/threadpool/threadpool.h>
+#include <executorch/extension/threadpool/threadpool.h>
 #include <executorch/extension/parallel/thread_parallel.h>
 #endif
 #include <executorch/extension/kernel_util/make_boxed_from_unboxed_functor.h>

extension/llm/custom_ops/targets.bzl

@@ -17,7 +17,7 @@ def define_common_targets():
             "//executorch/kernels/optimized:libvec",
             "//executorch/extension/kernel_util:kernel_util",
             "//executorch/extension/parallel:thread_parallel",
-            "//executorch/backends/xnnpack/threadpool:threadpool",
+            "//executorch/extension/threadpool:threadpool",
         ],
         compiler_flags = ["-Wno-missing-prototypes", "-Wno-global-constructors"],
         visibility = [

extension/parallel/targets.bzl

@@ -23,7 +23,7 @@ def define_common_targets():
                 "@EXECUTORCH_CLIENTS",
             ],
             deps = [
-                "//executorch/backends/xnnpack/threadpool:threadpool",
+                "//executorch/extension/threadpool:threadpool",
                 "//executorch/runtime/core:core",
                 "//executorch/runtime/core/exec_aten/util:tensor_util" + aten_suffix,
             ],

extension/parallel/thread_parallel.cpp

@@ -8,7 +8,7 @@
 
 #include <tuple>
 
-#include <executorch/backends/xnnpack/threadpool/threadpool.h>
+#include <executorch/extension/threadpool/threadpool.h>
 #include <executorch/extension/parallel/thread_parallel.h>
 #include <executorch/runtime/core/exec_aten/util/tensor_util.h>
 #include <executorch/runtime/platform/assert.h>