[llvm] [Offload] Implement the remaining initial Offload API (PR #122106)

[via llvm-commits]

================
@@ -0,0 +1,48 @@
+//===-- Memory.td - Memory definitions for Offload ---------*- tablegen -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file contains Offload API definitions related to memory allocations
+//
+//===----------------------------------------------------------------------===//
+
+def : Enum {
+  let name = "ol_alloc_type_t";
+  let desc = "Represents the type of allocation made with olMemAlloc.";
+  let etors = [
+    Etor<"HOST", "Host allocation">,
----------------
RaviNarayanaswamy wrote:

I don't think it should be called pinned.  The memory is allocated on the host, accessible by the device and does not migrate to device.  ie stays on host.  pinned in non offload term could mean pinned on the host and not paged out.
Also the "migratable" memory is  which is accessible by both host and device and the implementation may not migrate the memory but allocate in a location where both host and device can access.

https://github.com/llvm/llvm-project/pull/122106