[Openmp-commits] [openmp] 7086a1d - [libomptarget] [amdgpu] Hostcall offset check should consider implicit args

[Dhruva Chakrabarti via Openmp-commits]

Author: Dhruva Chakrabarti
Date: 2022-04-15T00:53:47Z
New Revision: 7086a1db80e1e09463744f6f6840eb771e3faeef

URL: https://github.com/llvm/llvm-project/commit/7086a1db80e1e09463744f6f6840eb771e3faeef
DIFF: https://github.com/llvm/llvm-project/commit/7086a1db80e1e09463744f6f6840eb771e3faeef.diff

LOG: [libomptarget] [amdgpu] Hostcall offset check should consider implicit args

Fixed hostcall offset check to compare against kernarg segment size
and implicit arguments. Improved the corresponding debug print.

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D123827

Added: 
    

Modified: 
    openmp/libomptarget/plugins/amdgpu/src/rtl.cpp

Removed: 
    


################################################################################
diff  --git a/openmp/libomptarget/plugins/amdgpu/src/rtl.cpp b/openmp/libomptarget/plugins/amdgpu/src/rtl.cpp
index fec8865b3c5e8..3086334f8ef63 100644
--- a/openmp/libomptarget/plugins/amdgpu/src/rtl.cpp
+++ b/openmp/libomptarget/plugins/amdgpu/src/rtl.cpp
@@ -1240,6 +1240,8 @@ int32_t runRegionLocked(int32_t device_id, void *tgt_entry_ptr, void **tgt_args,
           return OFFLOAD_FAIL;
         }
 
+        DP("Implicit argument count: %d\n",
+           KernelInfoEntry.implicit_argument_count);
         if (KernelInfoEntry.implicit_argument_count >= 4) {
           // Initialise pointer for implicit_argument_count != 0 ABI
           // Guess that the right implicit argument is at offset 24 after
@@ -1247,8 +1249,10 @@ int32_t runRegionLocked(int32_t device_id, void *tgt_entry_ptr, void **tgt_args,
           // the offset from msgpack. Clang is not annotating it at present.
           uint64_t Offset =
               sizeof(void *) * (KernelInfoEntry.explicit_argument_count + 3);
-          if ((Offset + 8) > (ArgPool->kernarg_segment_size)) {
-            DP("Bad offset of hostcall, exceeds kernarg segment size\n");
+          if ((Offset + 8) > ArgPool->kernarg_size_including_implicit()) {
+            DP("Bad offset of hostcall: %lu, exceeds kernarg size w/ implicit "
+               "args: %d\n",
+               Offset + 8, ArgPool->kernarg_size_including_implicit());
           } else {
             memcpy(static_cast<char *>(kernarg) + Offset, &buffer, 8);
           }