Decoupling AOT from graph memory planner (apache#8096)

* Fix an issue with storage-rewrite pass and packed functions

Change-Id: I13888471d4b8927a4012d6a8e749fb7a8935dd77

* Rebasing

Change-Id: I7aa12e0217b8a2e1ff2a97a7c5fdda6b7597ae64

* Addressing comments

Change-Id: If9f1ee190690f9a810fe41eb1933d736f1eb4ec3

* Add a pass to legalize packed calls

Change-Id: I8aa43d3a1b837b03a5cf3c6b32fc760bd78d3436

* Add a unit test for the legalization pass

Change-Id: I5b0d75380ff660dd5a0acf5b14fa84bb992fbec4

* rebasing

Change-Id: I52ceab5cf6e9b54390cb36c18dbb8e22505d8e18

* Use common StorageInfo

Change-Id: Ia8b7de1373f167ca7d0d69a99846d417405bbe48

Author: Giuseppe Rossini

include/tvm/tir/transform.h

@@ -418,6 +418,11 @@ TVM_DLL Pass ConvertBlocksToOpaque();
  */
 TVM_DLL Pass CompactBufferAllocation();
 
+/*!
+ * This pass legalizes packed calls by wrapping their arguments into TVMValues
+ */
+TVM_DLL Pass LegalizePackedCalls();
+
 /*!
  * \brief Flatten the multi-dimensional BufferLoad and BufferStore
  *        to single dimensional Load/Store. Also remove Block to

python/tvm/tir/transform/transform.py

@@ -451,6 +451,17 @@ def LowerTVMBuiltin():
     return _ffi_api.LowerTVMBuiltin()
 
 
+def LegalizePackedCalls():
+    """Legalize packed calls to have its arguments wrapped in TVMValues
+
+    Returns
+    -------
+    fpass : tvm.transform.Pass
+        The result pass
+    """
+    return _ffi_api.LegalizePackedCalls()
+
+
 def LowerIntrin():
     """Lower target specific intrinsic calls.