[MLIR] Create memref dialect and move several dialect-specific ops from std.

Create the memref dialect and move several dialect-specific ops without
dependencies to other ops from std dialect to this dialect.

Moved ops:
AllocOp -> MemRef_AllocOp
AllocaOp -> MemRef_AllocaOp
DeallocOp -> MemRef_DeallocOp
MemRefCastOp -> MemRef_CastOp
GetGlobalMemRefOp -> MemRef_GetGlobalOp
GlobalMemRefOp -> MemRef_GlobalOp
PrefetchOp -> MemRef_PrefetchOp
ReshapeOp -> MemRef_ReshapeOp
StoreOp -> MemRef_StoreOp
TransposeOp -> MemRef_TransposeOp
ViewOp -> MemRef_ViewOp

The roadmap to split the memref dialect from std is discussed here:
https://llvm.discourse.group/t/rfc-split-the-memref-dialect-from-std/2667

Differential Revision: https://reviews.llvm.org/D96425

1 files changed, 6 insertions, 4 deletions

diff --git a/mlir/lib/Transforms/Utils/LoopUtils.cpp b/mlir/lib/Transforms/Utils/LoopUtils.cpp
index a8c32c84..e4cbba3 100644
--- a/mlir/lib/Transforms/Utils/LoopUtils.cpp
+++ b/mlir/lib/Transforms/Utils/LoopUtils.cpp
@@ -18,6 +18,7 @@
 #include "mlir/Analysis/Utils.h"
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/Affine/IR/AffineValueMap.h"
+#include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/IR/AffineMap.h"
 #include "mlir/IR/BlockAndValueMapping.h"
@@ -2470,7 +2471,8 @@ static LogicalResult generateCopy(
 
     // Create the fast memory space buffer just before the 'affine.for'
     // operation.
-    fastMemRef = prologue.create<AllocOp>(loc, fastMemRefType).getResult();
+    fastMemRef =
+        prologue.create<memref::AllocOp>(loc, fastMemRefType).getResult();
     // Record it.
     fastBufferMap[memref] = fastMemRef;
     // fastMemRefType is a constant shaped memref.
@@ -2540,7 +2542,7 @@ static LogicalResult generateCopy(
     // Create a tag (single element 1-d memref) for the DMA.
     auto tagMemRefType = MemRefType::get({1}, top.getIntegerType(32), {},
                                          copyOptions.tagMemorySpace);
-    auto tagMemRef = prologue.create<AllocOp>(loc, tagMemRefType);
+    auto tagMemRef = prologue.create<memref::AllocOp>(loc, tagMemRefType);
 
     SmallVector<Value, 4> tagIndices({zeroIndex});
     auto tagAffineMap = b.getMultiDimIdentityMap(tagIndices.size());
@@ -2568,7 +2570,7 @@ static LogicalResult generateCopy(
                               numElementsSSA);
 
     // Generate dealloc for the tag.
-    auto tagDeallocOp = epilogue.create<DeallocOp>(loc, tagMemRef);
+    auto tagDeallocOp = epilogue.create<memref::DeallocOp>(loc, tagMemRef);
     if (*nEnd == end && isCopyOutAtEndOfBlock)
       // Since new ops are being appended (for outgoing DMAs), adjust the end to
       // mark end of range of the original.
@@ -2577,7 +2579,7 @@ static LogicalResult generateCopy(
 
   // Generate dealloc for the buffer.
   if (!existingBuf) {
-    auto bufDeallocOp = epilogue.create<DeallocOp>(loc, fastMemRef);
+    auto bufDeallocOp = epilogue.create<memref::DeallocOp>(loc, fastMemRef);
     // When generating pointwise copies, `nEnd' has to be set to deallocOp on
     // the fast buffer (since it marks the new end insertion point).
     if (!copyOptions.generateDma && *nEnd == end && isCopyOutAtEndOfBlock)