feat(linked): support TVM FFI providers

docs/linked-operators.md

@@ -1,9 +1,9 @@
 # Linked Operators
 
 The linked backend calls operators provided by an installed third-party shared
-library. It supports exact exported C++ symbols and registered PyTorch
-Dispatcher operators when a platform package does not provide source code or a
-stable C API.
+library. It supports exact exported symbols, TVM FFI entry points, and
+registered PyTorch Dispatcher operators when a platform package does not
+provide source code or a stable C API.
 
 ## Source Layout
 
@@ -15,16 +15,25 @@ src/linked/<transport>/<device>/
   ops/<operator>/
     <implementation>.yaml
     <implementation>.h
-    <implementation>.cc
+    <implementation>.{cc,cu}
 ```
 
+CUDA providers may use `<implementation>.cu` instead of `.cc`.
+
 The platform library file contains DSO discovery information:
 
 ```yaml
 python_distribution_package: vllm
 library_glob: vllm/_C*.so
 ```
 
+A library may also provide `include_glob` when its transport needs installed
+headers. Each glob must resolve to exactly one path in the Python distribution.
+A library may set `python_distribution_version` to a PEP 440 specifier. The
+resolver verifies the installed distribution version before looking up its DSO.
+A DSO that depends on another declared platform library lists that dependency
+under the implementation's optional `link_libraries` key.
+
 Files for an operator implementation use the provider name as their common
 stem. Multiple implementations for the same operator and device use distinct
 file stems and implementation slots.
@@ -54,15 +63,19 @@ partial Dispatcher contract or a binding that mixes both forms.
 
 ## Adapter Boundary
 
-Keep ABI behavior in `<implementation>.cc`, not in YAML. Shared operator
-templates own reusable tensor conversion, stream guards, layout staging, and
-copy-back behavior. Provider sources own exact typed function declarations,
+Keep ABI behavior in `<implementation>.cc` or `.cu`, not in YAML. Shared
+operator templates own reusable tensor conversion, stream guards, layout
+staging, and copy-back behavior. Provider sources own exact typed declarations,
 synthesized arguments, and provider-specific return handling.
 
 For the `torch` transport, an implementation backend inherits its device `C10`
 specialization for device identity and external-stream handling, then defines
 its provider-specific `Call` ABI.
 
+For the `tvm_ffi` transport, provider sources call exported TVM FFI entry
+points directly. The resolver supplies installed TVM FFI headers and links the
+provider DSO together with every library named in `link_libraries`.
+
 ## Configuration
 
 At configure time, `scripts/resolve_linked_ops.py` locates the installed Python
@@ -73,8 +86,10 @@ loading the DSO, comparing the registered schema exactly, and checking the
 requested dispatch key. The resolver writes a CMake manifest and diagnostic
 JSON under `generated/linked/`. Generated files are not committed.
 
-DSOs that provide Dispatcher registrations are force-loaded only for their own
-link item so that the linker cannot discard their static registration code.
+DSOs that provide Dispatcher registrations and DSOs named in
+`link_libraries` are retained with `--no-as-needed` only for their own link
+items. This prevents the linker from discarding registration code or an
+explicit dependency while preserving its state for unrelated libraries.
 
 Enable the backend independently of generated ATen implementations:
 
@@ -86,6 +101,11 @@ cmake -S . -B build \
   -DINFINI_OPS_OPS=silu_and_mul
 ```
 
+`WITH_TORCH=OFF` disables the generated ATen backend, but `WITH_LINKED=ON`
+currently still requires an installed `torch` package. Linked configuration
+shares its Python interpreter and C++ ABI setup with the existing `torch`
+transport.
+
 To resolve only selected linked implementation slots, pass an `ops.json` file
 through `INFINI_OPS_OPS`. The resolver reads each linked provider's slot from
 its sibling C++ header before locating external libraries, so an unselected
@@ -98,10 +118,11 @@ Provider and PyTorch C++ ABIs must match. Configuration fails before compilation
 when the distribution, shared library, or an exact required symbol is missing.
 
 InfiniOps does not bundle the provider library. Its resolved directory and the
-PyTorch runtime directories are recorded in the installed binary's RPATH, so a
-linked build is tied to that Python environment. Reconfigure and rebuild after
-moving or replacing the provider environment. In-place changes to a resolved
-provider DSO are tracked as CMake configure and link dependencies.
+directories of its linked dependencies are recorded in the installed binary's
+RPATH, so a linked build is tied to that Python environment. PyTorch runtime
+directories are recorded for the `torch` transport as well. Reconfigure and
+rebuild after moving or replacing the provider environment. In-place changes
+to a resolved provider DSO are tracked as CMake configure and link dependencies.
 
 ## Implementation Slots
 

pyproject.toml

@@ -1,5 +1,5 @@
 [build-system]
-requires = ["scikit-build-core", "pybind11", "libclang", "pyyaml"]
+requires = ["scikit-build-core", "pybind11", "libclang", "packaging", "pyyaml"]
 build-backend = "scikit_build_core.build"
 
 [project]