[Docs] Update stale links

docs/dev/pass_infra.rst

@@ -344,7 +344,7 @@ We've covered the concept of different level of passes and the context used for
 compilation. It would be interesting to see how easily users can register
 a pass.  Let's take const folding as an example. This pass has already been
 implemented to fold constants in a Relay function (found in
-`src/relay/pass/fold_constant.cc`_).
+`src/relay/transforms/fold_constant.cc`_).
 
 An API was provided to perform the ``Expr`` to ``Expr`` transformation.
 
@@ -536,7 +536,7 @@ optimization pipeline and debug Relay and tir passes, please refer to the
 
 .. _src/ir/transform.cc: https://github.com/apache/tvm/blob/main/src/ir/transform.cc
 
-.. _src/relay/pass/fold_constant.cc: https://github.com/apache/tvm/blob/main/src/relay/pass/fold_constant.cc
+.. _src/relay/transforms/fold_constant.cc: https://github.com/apache/tvm/blob/main/src/relay/transforms/fold_constant.cc
 
 .. _python/tvm/relay/transform/transform.py: https://github.com/apache/tvm/blob/main/python/tvm/relay/transform/transform.py
 

docs/dev/relay_add_op.rst

@@ -469,7 +469,7 @@ Adding a Gradient in C++
 Adding a gradient in C++ is similar to adding one in Python, but the
 interface for registering is slightly different.
 
-First, make sure ``src/relay/pass/pattern_utils.h`` is included. It provides
+First, make sure ``src/relay/transforms/pattern_utils.h`` is included. It provides
 helper functions for creating nodes in the Relay AST. Then, define the
 gradient in a similar fashion as in the Python example:
 

docs/dev/relay_add_pass.rst

@@ -397,10 +397,10 @@ the below code applies both the ``FoldConstant`` and ``ToANormalForm`` passes
 More detail about registration can be found in :ref:`tvm-runtime-system` and more
 information about the pass manager interface can be found in :ref:`pass-infra`.
 Relay's standard passes are listed in `include/tvm/relay/transform.h`_ and implemented
-in `src/relay/pass/`_.
+in `src/relay/transforms/`_.
 
 .. _include/tvm/relay/transform.h: https://github.com/apache/tvm/blob/main/include/tvm/relay/transform.h
 
-.. _src/relay/pass/: https://github.com/apache/tvm/tree/main/src/relay/pass
+.. _src/relay/transforms/: https://github.com/apache/tvm/tree/main/src/relay/transforms
 
 .. _src/relay/transforms/fold_constant.cc: https://github.com/apache/tvm/blob/main/src/relay/transforms/fold_constant.cc

docs/langref/relay_type.rst

@@ -43,7 +43,7 @@ Relay to be ahead-of-time compiled and provides much more information about
 tensors for optimizations further in the compilation pipeline. Such optimizations
 can be implemented as passes, which are Relay-to-Relay AST transformations, and
 may use the inferred types (e.g., shape information) for making decisions about
-program transformations. For instance, :code:`src/relay/pass/fuse_ops.cc` gives
+program transformations. For instance, :code:`src/relay/transforms/fuse_ops.cc` gives
 an implementation of a pass that uses inferred tensor shapes to replace invocations
 of operators in a Relay program with fused operator implementations.
 

src/relay/analysis/annotated_region_set.h

@@ -18,7 +18,7 @@
  */
 
 /*!
- * \file tvm/relay/pass/annotated_region_set.h
+ * \file tvm/relay/transforms/annotated_region_set.h
  * \brief Define data structures to extract and manipulate regions from
  * a relay function. Regions are denoted by region_begin and region_end
  * annotations that exist on all the input and output edges of the region.