diff --git a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ArithInterfaceImpl.cpp b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ArithInterfaceImpl.cpp
--- a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ArithInterfaceImpl.cpp
+++ b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ArithInterfaceImpl.cpp
@@ -49,7 +49,6 @@
     auto globalMemref = globalCreator.getGlobalFor(constantOp);
     Value memref = b.create<memref::GetGlobalOp>(
         constantOp.getLoc(), globalMemref.type(), globalMemref.getName());
-    state.aliasInfo.insertNewBufferEquivalence(memref, constantOp.getResult());
     state.mapBuffer(constantOp, memref);
 
     return success();
diff --git a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.cpp b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.cpp
--- a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.cpp
+++ b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.cpp
@@ -586,7 +586,6 @@
   Value casted = allocated.getValue();
   if (memRefType && memRefType != allocMemRefType) {
     casted = b.create<memref::CastOp>(loc, memRefType, allocated.getValue());
-    aliasInfo.insertNewBufferEquivalence(casted, allocated.getValue());
   }
 
   // 2. Create memory deallocation.
diff --git a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/LinalgInterfaceImpl.cpp b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/LinalgInterfaceImpl.cpp
--- a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/LinalgInterfaceImpl.cpp
+++ b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/LinalgInterfaceImpl.cpp
@@ -271,8 +271,6 @@
         return failure();
 
       // Insert mapping and aliasing info.
-      state.aliasInfo.createAliasInfoEntry(resultBuffer);
-      state.aliasInfo.insertNewBufferEquivalence(opResult, resultBuffer);
       state.mapBuffer(opResult, resultBuffer);
 
       // Insert new operand and bbArg.
@@ -281,9 +279,6 @@
           body->insertArgument(nextOutputBBArgIndex, resultBuffer.getType());
       BlockArgument oldTensorBBArg = body->getArgument(oldOutputBBArgIndex);
       // Insert mapping and aliasing info.
-      state.aliasInfo.createAliasInfoEntry(newBufferBBArg);
-      state.aliasInfo.insertNewBufferEquivalence(oldTensorBBArg,
-                                                 newBufferBBArg);
       state.mapBuffer(oldTensorBBArg, newBufferBBArg);
 
       // Set operand of `linalg.yield` to the bbArg so it just canonicalizes
@@ -321,9 +316,6 @@
       BlockArgument oldTensorBBArg = body->getArgument(oldInputBBArgIndex);
 
       // Insert mapping and aliasing info.
-      state.aliasInfo.createAliasInfoEntry(newBufferBBArg);
-      state.aliasInfo.insertNewBufferEquivalence(oldTensorBBArg,
-                                                 newBufferBBArg);
       state.mapBuffer(oldTensorBBArg, newBufferBBArg);
 
       // Increment indices.
diff --git a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ModuleBufferization.cpp b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ModuleBufferization.cpp
--- a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ModuleBufferization.cpp
+++ b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ModuleBufferization.cpp
@@ -174,7 +174,6 @@
                                              BufferizationState &state) {
   LLVM_DEBUG(DBGS() << "Begin bufferizeFuncOpBoundary:\n" << funcOp << "\n");
   ModuleBufferizationState &moduleState = getModuleBufferizationState(state);
-  BufferizationAliasInfo &aliasInfo = state.aliasInfo;
 
   // If nothing to do then we are done.
   if (!llvm::any_of(funcOp.getType().getInputs(), isaTensor) &&
@@ -271,15 +270,12 @@
         auto castOp = b.create<memref::CastOp>(
             funcOp.getLoc(), toMemrefOp.memref().getType(), memref);
         toMemrefOp.memref().replaceAllUsesWith(castOp);
-        aliasInfo.insertNewBufferEquivalence(castOp.dest(),
-                                             toMemrefOp.memref());
       }
     }
     // Replace all remaining uses by a to_tensor.
     if (!bbArg.use_empty()) {
       auto toTensorOp =
           b.create<bufferization::ToTensorOp>(funcOp.getLoc(), memref);
-      aliasInfo.insertNewBufferEquivalence(toTensorOp, bbArg);
       bbArg.replaceAllUsesWith(toTensorOp);
     }
     frontBlock.eraseArgument(0);
@@ -521,7 +517,6 @@
           Value buffer = state.lookupBuffer(callOp->getOperand(idx));
           // Add CallOp operand/result equivalence: this is interprocedural
           // info.
-          state.aliasInfo.insertNewBufferEquivalence(oldRes, buffer);
           state.mapBuffer(oldRes, buffer);
           // Add a ToTensorOp to kill all uses of the CallOp return.
           // Replace all uses of the CallOp results so we can erase the CallOp.
@@ -531,7 +526,6 @@
               b.create<bufferization::ToTensorOp>(callOp.getLoc(), buffer);
           oldRes.replaceAllUsesWith(toTensorOp);
           // Add new op equivalence info.
-          state.aliasInfo.insertNewBufferEquivalence(toTensorOp, buffer);
           state.mapBuffer(toTensorOp, buffer);
           continue;
         }
@@ -574,7 +568,6 @@
         Value castBuffer =
             b.create<memref::CastOp>(callOp.getLoc(), memRefType, buffer);
         // Add new op equivalence info.
-        state.aliasInfo.insertNewBufferEquivalence(castBuffer, buffer);
         state.mapBuffer(tensorOperand, castBuffer);
         buffer = castBuffer;
       }
@@ -627,7 +620,6 @@
       Value returnTensor = b.create<bufferization::ToTensorOp>(
           returnOp.getLoc(), v);
       operand.set(returnTensor);
-      state.aliasInfo.insertNewBufferEquivalence(returnTensor, v);
       state.mapBuffer(returnTensor, v);
     }
     return success();
@@ -657,7 +649,6 @@
                             : getContiguousOrUnrankedMemRefType(tensorType);
       Value bufferCast = b.create<bufferization::ToMemrefOp>(funcOp.getLoc(),
                                                              memRefType, bbArg);
-      state.aliasInfo.insertNewBufferEquivalence(bufferCast, bbArg);
       state.mapBuffer(bbArg, bufferCast);
     }
 
diff --git a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/SCFInterfaceImpl.cpp b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/SCFInterfaceImpl.cpp
--- a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/SCFInterfaceImpl.cpp
+++ b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/SCFInterfaceImpl.cpp
@@ -149,7 +149,6 @@
       if (!resultBuffer)
         return failure();
 
-      state.aliasInfo.createAliasInfoEntry(resultBuffer);
       state.mapBuffer(opResult, resultBuffer);
     }
 
diff --git a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/TensorInterfaceImpl.cpp b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/TensorInterfaceImpl.cpp
--- a/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/TensorInterfaceImpl.cpp
+++ b/mlir/lib/Dialect/Linalg/ComprehensiveBufferize/TensorInterfaceImpl.cpp
@@ -89,7 +89,6 @@
         castOp.getResult().getType(), layout, memorySpace);
     Value res =
         b.create<memref::CastOp>(castOp.getLoc(), memRefType, resultBuffer);
-    state.aliasInfo.insertNewBufferEquivalence(res, castOp.getResult());
     state.mapBuffer(castOp.getResult(), res);
     return success();
   }
@@ -267,7 +266,6 @@
     b.create<memref::StoreOp>(loc, insertOp.scalar(), destMemref,
                               insertOp.indices());
     state.mapBuffer(insertOp, destMemref);
-    state.aliasInfo.insertNewBufferAlias(insertOp, destMemref);
     return success();
   }
 
@@ -464,8 +462,6 @@
       Value subView = b.create<memref::SubViewOp>(
           loc, subviewMemRefType, dstMemref, insertSliceOp.getMixedOffsets(),
           insertSliceOp.getMixedSizes(), insertSliceOp.getMixedStrides());
-      // Insert new alias.
-      state.aliasInfo.insertNewBufferAlias(subView, dstMemref);
       // Copy tensor.
       Value srcMemref = state.lookupBuffer(insertSliceOp.source());
       state.options.allocationFns->memCpyFn(b, insertSliceOp.getLoc(),