[Codegen][Tuner] Clarify tuning spec linking order. NFC. (iree-org#19370

)

This clarification was suggested in
iree-org#19337 (comment).

compiler/src/iree/compiler/Codegen/Common/Passes.h

@@ -56,7 +56,8 @@ void addEncodingToNopPasses(FunctionLikeNest &passManager);
 /// Links nested transform dialect tuning specs named sequences into a single
 /// entry point. Returns the new named sequence op (inserted into the `module`)
 /// that includes the nested tuning specs, or a null op when no nested named
-/// sequences were found.
+/// sequences were found. The order of inclusion is the same as the order in
+/// which these nested tuning specs appear in the IR.
 FailureOr<transform::NamedSequenceOp> linkTuningSpecs(ModuleOp module);
 
 //------------------------------------------------------------------------------

compiler/src/iree/compiler/Codegen/Common/Passes.td

@@ -412,7 +412,7 @@ def LinkTuningSpecsPass : Pass<"iree-codegen-link-tuning-specs", "ModuleOp"> {
   let description = [{
     Given a module with multiple nested tuning specs, introduce a new named sequence
     that includes all the other tuning spec entry points. The order of inclusion is the same
-    as the in which these nested tuning specs appear in the IR.
+    as the order in which these nested tuning specs appear in the IR.
 
     A tuning spec entry point is a `transform.named_sequence` op annotated with the
     `iree_codegen.tuning_spec` unit attribute. We require it to perform in-place op