From 0fb0fb1d8b5cdd915d049635df031575bcaa079e Mon Sep 17 00:00:00 2001
From: foreman <dl.constructicon@amd.com>
Date: Fri, 23 Sep 2016 18:18:01 -0400
Subject: [PATCH] P4 to Git Change 1318412 by lmoriche@lmoriche_opencl_dev on
 2016/09/23 17:57:01

	SWDEV-94610 - Remove the padding at the end of the kernargs (It was for the hidden arguments, but now, LC reports the correct size). Set the LLVM triple to amdgcn-amd-amdhsa-opencl when building the built-in library.

Affected files ...

... //depot/stg/opencl/drivers/opencl/opencldefs#186 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/roccompiler.cpp#20 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocprogram.cpp#37 edit
---
 rocclr/runtime/device/rocm/roccompiler.cpp | 2 +-
 rocclr/runtime/device/rocm/rocprogram.cpp  | 5 +----
 2 files changed, 2 insertions(+), 5 deletions(-)

diff --git a/rocclr/runtime/device/rocm/roccompiler.cpp b/rocclr/runtime/device/rocm/roccompiler.cpp
index 05fa1fdc20..ce1a62261e 100644
--- a/rocclr/runtime/device/rocm/roccompiler.cpp
+++ b/rocclr/runtime/device/rocm/roccompiler.cpp
@@ -168,7 +168,7 @@ HSAILProgram::compileImpl_LC(
         std::ofstream f(options->getDumpFileName(".cl").c_str(), std::ios::trunc);
         if(f.is_open()) {
             f << "/* Compiler options:\n" \
-                   "-c -emit-llvm -target amdgcn--amdhsa-opencl -x cl" \
+                   "-c -emit-llvm -target amdgcn-amd-amdhsa-opencl -x cl" \
                    " -include opencl-c.h " << driverOptions
                 << "\n*/\n\n" << sourceCode;
         } else {
diff --git a/rocclr/runtime/device/rocm/rocprogram.cpp b/rocclr/runtime/device/rocm/rocprogram.cpp
index d49f1c27de..37f54cb353 100644
--- a/rocclr/runtime/device/rocm/rocprogram.cpp
+++ b/rocclr/runtime/device/rocm/rocprogram.cpp
@@ -1119,10 +1119,7 @@ HSAILProgram::setKernels_LC(amd::option::Options *options, void* binary, size_t
             kernelCodeHandle,
             workgroupGroupSegmentByteSize,
             workitemPrivateSegmentByteSize,
-            // TODO: remove the workaround
-            //   add 24 bytes for global offsets as workaround for LC reporting
-            //   excluded the hidden arguments
-            amd::alignUp(kernargSegmentByteSize, sizeof(size_t)) + 3 * sizeof(size_t),
+            kernargSegmentByteSize,
             amd::alignUp(kernargSegmentAlignment,device().info().globalMemCacheLineSize_));
         if (!aKernel->init()) {
             return false;