Update docs to use top-level CMake targets

ghstack-source-id: bd444bf
ghstack-comment-id: 3111967018
Pull-Request: #12807

Author: GregoryComer

backends/vulkan/README.md

@@ -193,12 +193,12 @@ GPU!
 
 ```shell
 # Build a model runner binary linked with the Vulkan delegate libs
-cmake --build cmake-android-out --target vulkan_executor_runner -j32
+cmake --build cmake-android-out --target executor_runner -j32
 
 # Push model to device
 adb push vk_add.pte /data/local/tmp/vk_add.pte
 # Push binary to device
-adb push cmake-android-out/backends/vulkan/vulkan_executor_runner /data/local/tmp/runner_bin
+adb push cmake-android-out/executor_runner /data/local/tmp/runner_bin
 
 # Run the model
 adb shell /data/local/tmp/runner_bin --model_path /data/local/tmp/vk_add.pte

docs/source/kernel-library-selective-build.md

@@ -34,9 +34,25 @@ The basic flow looks like this:
 3. A _kernel resolver _takes in the linked kernel libraries as well as the merged op info yaml file, then makes a decision on which kernels to be registered into ExecuTorch runtime.
 
 
+## Selective Build CMake Options
+
+To enable selective build when building the executorch kernel libraries as part of a CMake build, the following CMake options are exposed. These options affect the `executorch_kernels` CMake target. Make sure to link this target when using selective build.
+
+ * `EXECUTORCH_SELECT_OPS_YAML`: A path to a YAML file specifying the operators to include.
+ * `EXECUTORCH_SELECT_OPS_LIST`: A string containing the operators to include.
+ * `EXECUTORCH_SELECT_OPS_MODEL`: A path to a PTE file. Only operators used in this model will be included.
+ * `EXECUTORCH_ENABLE_DTYPE_SELECTIVE_BUILD`: If enabled, operators will be further specialized to only operator on the data types specified in the operator selection.
+
+Note that `EXECUTORCH_SELECT_OPS_YAML`, `EXECUTORCH_SELECT_OPS_LIST`, and `EXECUTORCH_SELECT_OPS_MODEL` are mutually exclusive. Only one operator specifier directive is allowed.
+
+As an example, to build with only operators used in mv2_xnnpack_fp32.pte, the CMake build can be configured as follows.
+```
+cmake .. -DEXECUTORCH_SELECT_OPS_MODEL=mv2_xnnpack_fp32.pte
+```
+
 ## APIs
 
-We expose a CMake macro [gen_selected_ops](https://github.com/pytorch/executorch/blob/main/tools/cmake/Codegen.cmake#L12), to allow users specifying op info:
+For fine-grained control, we expose a CMake macro [gen_selected_ops](https://github.com/pytorch/executorch/blob/main/tools/cmake/Codegen.cmake#L12) to allow users to specify op info:
 
 ```
 gen_selected_ops(
@@ -75,7 +91,7 @@ Beyond pruning the binary to remove unused operators, the binary size can furthe
 
 ## Example Walkthrough
 
-In [CMakeLists.txt](https://github.com/BujSet/executorch/blob/main/examples/selective_build/CMakeLists.txt#L48-L72), we have the following cmake config options:
+In [examples/selective_build/CMakeLists.txt](https://github.com/BujSet/executorch/blob/main/examples/selective_build/CMakeLists.txt#L48-L72), we have the following cmake config options:
 
 1. `EXECUTORCH_SELECT_OPS_YAML`
 2. `EXECUTORCH_SELECT_OPS_LIST`

docs/source/using-executorch-building-from-source.md

@@ -252,6 +252,17 @@ I 00:00:00.000764 executorch:executor_runner.cpp:180] Model executed successfull
 I 00:00:00.000770 executorch:executor_runner.cpp:184] 1 outputs:
 Output 0: tensor(sizes=[1], [2.])
 ```
+
+### CMake Targets
+
+To link against the ExecuTorch framework from CMake, the following top-level targets are exposed:
+
+ * `executorch::backends`: Contains all configured backends.
+ * `executorch::extensions`: Contains all configured extensions.
+ * `executorch::kernels`: Contains all configured kernel libraries.
+
+The backends, extensions, and kernels included in these targets are controlled by the various `EXECUTORCH_` CMake options specified by the build.
+
 ## Build ExecuTorch for Windows
 
 This document outlines the current known working build instructions for building and validating ExecuTorch on a Windows machine.

docs/source/using-executorch-cpp.md

@@ -40,7 +40,7 @@ Running a model using the low-level runtime APIs allows for a high-degree of con
 
 ## Building with CMake
 
-ExecuTorch uses CMake as the primary build system. Inclusion of the module and tensor APIs are controlled by the `EXECUTORCH_BUILD_EXTENSION_MODULE` and `EXECUTORCH_BUILD_EXTENSION_TENSOR` CMake options. As these APIs may not be supported on embedded systems, they are disabled by default when building from source. The low-level API surface is always included. To link, add the `executorch` target as a CMake dependency, along with `extension_module_static` and `extension_tensor`, if desired.
+ExecuTorch uses CMake as the primary build system. Inclusion of the module and tensor APIs are controlled by the `EXECUTORCH_BUILD_EXTENSION_MODULE` and `EXECUTORCH_BUILD_EXTENSION_TENSOR` CMake options. As these APIs may not be supported on embedded systems, they are disabled by default when building from source. The low-level API surface is always included. To link, add the `executorch` target as a CMake dependency, along with `executorch_backends`, `executorch_extensions`, and `extension_kernels`, to link all configured backends, extensions, and kernels.
 
 ```
 # CMakeLists.txt
@@ -49,10 +49,9 @@ add_subdirectory("executorch")
 target_link_libraries(
     my_target
     PRIVATE executorch
-    extension_module_static
-    extension_tensor
-    optimized_native_cpu_ops_lib
-    xnnpack_backend)
+    executorch::backends
+    executorch::extensions
+    executorch::kernels)
 ```
 
 See [Building from Source](using-executorch-building-from-source.md) for more information on the CMake build process.

tools/cmake/preset/default.cmake

@@ -228,12 +228,14 @@ define_overridable_option(
 define_overridable_option(
   EXECUTORCH_SELECT_OPS_MODEL
   "Build the executorch_kernels target with only operators from the given model .pte file."
-  STRING ""
+  STRING
+  ""
 )
 define_overridable_option(
   EXECUTORCH_ENABLE_DTYPE_SELECTIVE_BUILD
   "Build the executorch_kernels target with only operator implementations for selected data types."
-  BOOL FALSE
+  BOOL
+  FALSE
 )
 
 # ------------------------------------------------------------------------------
@@ -248,8 +250,7 @@ check_required_options_on(
 
 check_required_options_on(
   IF_ON EXECUTORCH_BUILD_EXECUTOR_RUNNER REQUIRES
-  EXECUTORCH_BUILD_EXTENSION_EVALUE_UTIL
-  EXECUTORCH_BUILD_EXTENSION_RUNNER_UTIL
+  EXECUTORCH_BUILD_EXTENSION_EVALUE_UTIL EXECUTORCH_BUILD_EXTENSION_RUNNER_UTIL
 )
 check_required_options_on(
   IF_ON EXECUTORCH_BUILD_EXTENSION_FLAT_TENSOR REQUIRES
@@ -295,26 +296,29 @@ check_required_options_on(
   IF_ON EXECUTORCH_BUILD_TESTS REQUIRES EXECUTORCH_BUILD_EXTENSION_FLAT_TENSOR
 )
 
+check_required_options_on(
+  IF_ON EXECUTORCH_ENABLE_DTYPE_SELECTIVE_BUILD REQUIRES
+  EXECUTORCH_SELECT_OPS_MODEL
+)
+
 check_required_options_on(
   IF_ON EXECUTORCH_BUILD_XNNPACK REQUIRES EXECUTORCH_BUILD_CPUINFO
   EXECUTORCH_BUILD_PTHREADPOOL
 )
 
 check_conflicting_options_on(
   IF_ON EXECUTORCH_BUILD_ARM_BAREMETAL CONFLICTS_WITH
-  EXECUTORCH_BUILD_PTHREADPOOL
-  EXECUTORCH_BUILD_CPUINFO
+  EXECUTORCH_BUILD_PTHREADPOOL EXECUTORCH_BUILD_CPUINFO
 )
 
 # Selective build specifiers are mutually exclusive.
 check_conflicting_options_on(
-  IF_ON EXECUTORCH_SELECT_OPS_YAML CONFLICTS_WITH
-  EXECUTORCH_SELECT_OPS_LIST EXECUTORCH_SELECT_OPS_MODEL
+  IF_ON EXECUTORCH_SELECT_OPS_YAML CONFLICTS_WITH EXECUTORCH_SELECT_OPS_LIST
+  EXECUTORCH_SELECT_OPS_MODEL
 )
 
 check_conflicting_options_on(
-  IF_ON EXECUTORCH_SELECT_OPS_LIST CONFLICTS_WITH
-  EXECUTORCH_SELECT_OPS_MODEL
+  IF_ON EXECUTORCH_SELECT_OPS_LIST CONFLICTS_WITH EXECUTORCH_SELECT_OPS_MODEL
 )
 
 if(NOT EXISTS ${EXECUTORCH_PAL_DEFAULT_FILE_PATH})