Make wasi-nn backends as separated shared libraries (#3509)

- All files under *core/iwasm/libraries/wasi-nn* are compiled as shared libraries
- *wasi-nn.c* is shared between backends
- Every backend has a separated shared library
- If wasi-nn feature is enabled, iwasm will depend on shared library libiwasm.so
  instead of linking static library libvmlib.a

core/iwasm/libraries/wasi-nn/README.md

@@ -2,15 +2,28 @@
 
 ## How to use
 
+### Host
+
 Enable WASI-NN in the WAMR by spefiying it in the cmake building configuration as follows,
 
-```
+```cmake
 set (WAMR_BUILD_WASI_NN  1)
 ```
 
-The definition of the functions provided by WASI-NN is in the header file `core/iwasm/libraries/wasi-nn/wasi_nn.h`.
+or in command line
 
-By only including this file in your WASM application you will bind WASI-NN into your module.
+```bash
+$ cmake -DWAMR_BUILD_WASI_NN=1 <other options> ...
+```
+
+> ![Caution]
+> If enable `WAMR_BUID_WASI_NN`, iwasm will link a shared WAMR library instead of a static one. Wasi-nn backends will be loaded dynamically at runtime. Users shall specify the path of the backend library and register it to the iwasm runtime with `--native-lib=<path of backend library>`. All shared libraries should be placed in the `LD_LIBRARY_PATH`.
+
+### Wasm
+
+The definition of functions provided by WASI-NN (Wasm imports) is in the header file _core/iwasm/libraries/wasi-nn/wasi_nn.h_.
+
+By only including this file in a WASM application you will bind WASI-NN into your module.
 
 ## Tests
 
@@ -27,9 +40,8 @@ Build the runtime image for your execution target type.
 - `vx-delegate`
 - `tpu`
 
-```
-EXECUTION_TYPE=cpu
-docker build -t wasi-nn-${EXECUTION_TYPE} -f core/iwasm/libraries/wasi-nn/test/Dockerfile.${EXECUTION_TYPE} .
+```bash
+EXECUTION_TYPE=cpu docker build -t wasi-nn-${EXECUTION_TYPE} -f core/iwasm/libraries/wasi-nn/test/Dockerfile.${EXECUTION_TYPE} .
 ```
 
 ### Build wasm app
@@ -50,15 +62,19 @@ If all the tests have run properly you will the the following message in the ter
 Tests: passed!
 ```
 
+> [!TIP]
+> Use _libwasi-nn-tflite.so_ as an example. You shall use whatever you have built.
+
 - CPU
 
-```
+```bash
 docker run \
     -v $PWD/core/iwasm/libraries/wasi-nn/test:/assets \
     -v $PWD/core/iwasm/libraries/wasi-nn/test/models:/models \
     wasi-nn-cpu \
     --dir=/ \
     --env="TARGET=cpu" \
+    --native-lib=/lib/libwasi-nn-tflite.so \
     /assets/test_tensorflow.wasm
 ```
 
@@ -66,7 +82,7 @@ docker run \
   - Requirements:
     - [NVIDIA docker](https://github.com/NVIDIA/nvidia-docker).
 
-```
+```bash
 docker run \
     --runtime=nvidia \
     -v $PWD/core/iwasm/libraries/wasi-nn/test:/assets \
@@ -74,17 +90,19 @@ docker run \
     wasi-nn-nvidia-gpu \
     --dir=/ \
     --env="TARGET=gpu" \
+    --native-lib=/lib/libwasi-nn-tflite.so \
     /assets/test_tensorflow.wasm
 ```
 
 - vx-delegate for NPU (x86 simulator)
 
-```
+```bash
 docker run \
     -v $PWD/core/iwasm/libraries/wasi-nn/test:/assets \
     wasi-nn-vx-delegate \
     --dir=/ \
     --env="TARGET=gpu" \
+    --native-lib=/lib/libwasi-nn-tflite.so \
     /assets/test_tensorflow_quantized.wasm
 ```
 
@@ -92,7 +110,7 @@ docker run \
   - Requirements:
     - [Coral USB](https://coral.ai/products/accelerator/).
 
-```
+```bash
 docker run \
     --privileged \
     --device=/dev/bus/usb:/dev/bus/usb \
@@ -100,6 +118,7 @@ docker run \
     wasi-nn-tpu \
     --dir=/ \
     --env="TARGET=tpu" \
+    --native-lib=/lib/libwasi-nn-tflite.so \
     /assets/test_tensorflow_quantized.wasm
 ```
 
@@ -120,20 +139,20 @@ Use [classification-example](https://github.com/bytecodealliance/wasi-nn/tree/ma
 
 ### Prepare the model and the wasm
 
-``` bash
+```bash
 $ pwd
 /workspaces/wasm-micro-runtime/core/iwasm/libraries/wasi-nn/test
 
 $ docker build -t wasi-nn-example:v1.0 -f Dockerfile.wasi-nn-example .
 ```
 
-There are model files(*mobilenet\**) and wasm files(*wasi-nn-example.wasm*) in the directory */workspaces/wasi-nn/rust/examples/classification-example/build* in the image of wasi-nn-example:v1.0.
+There are model files(\*mobilenet\**) and wasm files(*wasi-nn-example.wasm*) in the directory */workspaces/wasi-nn/rust/examples/classification-example/build\* in the image of wasi-nn-example:v1.0.
 
 ### build iwasm and test
 
-*TODO: May need alternative steps to build the iwasm and test in the container of wasi-nn-example:v1.0*
+_TODO: May need alternative steps to build the iwasm and test in the container of wasi-nn-example:v1.0_
 
-``` bash
+```bash
 $ pwd
 /workspaces/wasm-micro-runtime
 
@@ -143,9 +162,9 @@ $ docker run --rm -it -v $(pwd):/workspaces/wasm-micro-runtime wasi-nn-example:v
 > [!Caution]
 > The following steps are executed in the container of wasi-nn-example:v1.0.
 
-``` bash
+```bash
 $ cd /workspaces/wasm-micro-runtime/product-mini/platforms/linux
 $ cmake -S . -B build -DWAMR_BUILD_WASI_NN=1 -DWAMR_BUILD_WASI_EPHEMERAL_NN=1
 $ cmake --build build
 $ ./build/iwasm -v=5 --map-dir=/workspaces/wasi-nn/rust/examples/classification-example/build/::fixture /workspaces/wasi-nn/rust/examples/classification-example/build/wasi-nn-example.wasm
-```
+```