Sanitize MCJIT remote execution

MCJIT remote execution (ChildTarget+RemoteTargetExternal) protocol was in
dire need of refactoring. It was fail-prone, had no error reporting and
implemented the same message logic on every single function.

This patch rectifies it, and makes it work on ARM, where it was randomly
failing. Other architectures shall profit from this change as well, making
their buildbots and releases more reliable.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@199261 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 47 insertions, 7 deletions

diff --git a/tools/lli/RemoteTargetMessage.h b/tools/lli/RemoteTargetMessage.h
index 12cfa9a404..cb934a1066 100644
--- a/tools/lli/RemoteTargetMessage.h
+++ b/tools/lli/RemoteTargetMessage.h
@@ -26,20 +26,60 @@ namespace llvm {
 // only here for testing purposes and is therefore intended to be the simplest
 // implementation that will work.  It is assumed that the parent and child
 // share characteristics like endianness.
+//
+// Quick description of the protocol:
+//
+// { Header + Payload Size + Payload }
+//
+// The protocol message consist of a header, the payload size (which can be
+// zero), and the payload itself. The payload can contain any number of items,
+// and the size has to be the sum of them all. Each end is responsible for
+// reading/writing the correct number of items with the correct sizes.
+//
+// The current four known exchanges are:
+//
+//  * Allocate Space:
+//   Parent: { LLI_AllocateSpace, 8, Alignment, Size }
+//    Child: { LLI_AllocationResult, 8, Address }
+//
+//  * Load Data:
+//   Parent: { LLI_LoadDataSection, 8+Size, Address, Data }
+//    Child: { LLI_LoadComplete, 4, StatusCode }
+//
+//  * Load Code:
+//   Parent: { LLI_LoadCodeSection, 8+Size, Address, Code }
+//    Child: { LLI_LoadComplete, 4, StatusCode }
+//
+//  * Execute Code:
+//   Parent: { LLI_Execute, 8, Address }
+//    Child: { LLI_ExecutionResult, 4, Result }
+//
+// It is the responsibility of either side to check for correct headers,
+// sizes and payloads, since any inconsistency would misalign the pipe, and
+// result in data corruption.
 
 enum LLIMessageType {
   LLI_Error = -1,
   LLI_ChildActive = 0,        // Data = not used
-  LLI_AllocateSpace,          // Data = struct { uint_32t Align, uint_32t Size }
-  LLI_AllocationResult,       // Data = uint64_t AllocAddress (in Child memory space)
-  LLI_LoadCodeSection,        // Data = uint32_t Addr, followed by section contests
-  LLI_LoadDataSection,        // Data = uint32_t Addr, followed by section contents
-  LLI_LoadComplete,           // Data = not used
-  LLI_Execute,                // Data = Address of function to execute
-  LLI_ExecutionResult,        // Data = uint64_t Result
+  LLI_AllocateSpace,          // Data = struct { uint32_t Align, uint_32t Size }
+  LLI_AllocationResult,       // Data = uint64_t Address (child memory space)
+
+  LLI_LoadCodeSection,        // Data = uint64_t Address, void * SectionData
+  LLI_LoadDataSection,        // Data = uint64_t Address, void * SectionData
+  LLI_LoadResult,             // Data = uint32_t LLIMessageStatus
+
+  LLI_Execute,                // Data = uint64_t Address
+  LLI_ExecutionResult,        // Data = uint32_t Result
+
   LLI_Terminate               // Data = not used
 };
 
+enum LLIMessageStatus {
+  LLI_Status_Success = 0,     // Operation succeeded
+  LLI_Status_NotAllocated,    // Address+Size not allocated in child space
+  LLI_Status_IncompleteMsg    // Size received doesn't match request
+};
+
 } // end namespace llvm
 
 #endif