From f670ae2f8d44d0cd95d1c2349c0499a771ee4c04 Mon Sep 17 00:00:00 2001
From: foreman <dl.constructicon@amd.com>
Date: Thu, 1 Oct 2015 08:37:19 -0400
Subject: [PATCH] P4 to Git Change 1196211 by nhaustov@nhaustov_hsa on
 2015/10/01 08:27:23

	SWDEV-76911 - Bug 11070: ViennaCL fails in amd::hsa::loader::Executable::Destroy

	Destructors of global variables run in undefined order, so loader cannot
	be global. Encapsulate loader functionality in Loader class and link its lifecycle
	to Runtime.
	In HSA Runtime, create Loader as part of Runtime singleton.
	In ORCA Runtime, create Loader as part of HSAILProgram class.

	Testing: smoke, pre-checkin

Affected files ...

... //depot/stg/opencl/drivers/opencl/compiler/sc/HSAIL/ext/amdhsacod/amdhsacod.cpp#10 integrate
... //depot/stg/opencl/drivers/opencl/compiler/sc/HSAIL/ext/loader/executable.cpp#12 integrate
... //depot/stg/opencl/drivers/opencl/compiler/sc/HSAIL/ext/loader/executable.hpp#8 integrate
... //depot/stg/opencl/drivers/opencl/compiler/sc/HSAIL/include/amd_hsa_loader.hpp#5 integrate
... //depot/stg/opencl/drivers/opencl/compiler/tools/aoc2/aoc2.cpp#78 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuprogram.cpp#210 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuprogram.hpp#61 edit
---
 rocclr/runtime/device/gpu/gpuprogram.cpp | 8 ++++++--
 rocclr/runtime/device/gpu/gpuprogram.hpp | 2 ++
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/rocclr/runtime/device/gpu/gpuprogram.cpp b/rocclr/runtime/device/gpu/gpuprogram.cpp
index 014de53f12..431cca6e3d 100644
--- a/rocclr/runtime/device/gpu/gpuprogram.cpp
+++ b/rocclr/runtime/device/gpu/gpuprogram.cpp
@@ -18,6 +18,7 @@
 #include "utils/options.hpp"
 #include "hsa.h"
 #include "hsa_ext_image.h"
+#include "amd_hsa_loader.hpp"
 
 namespace gpu {
 
@@ -1670,6 +1671,7 @@ HSAILProgram::HSAILProgram(Device& device)
     binOpts_.bitness = ELFDATA2LSB;
     binOpts_.alloc = &::malloc;
     binOpts_.dealloc = &::free;
+    loader_ = amd::hsa::loader::Loader::Create(&loaderContext_);
 }
 
 HSAILProgram::HSAILProgram(NullDevice& device)
@@ -1689,6 +1691,7 @@ HSAILProgram::HSAILProgram(NullDevice& device)
     binOpts_.bitness = ELFDATA2LSB;
     binOpts_.alloc = &::malloc;
     binOpts_.dealloc = &::free;
+    loader_ = amd::hsa::loader::Loader::Create(&loaderContext_);
 }
 
 HSAILProgram::~HSAILProgram()
@@ -1710,9 +1713,10 @@ HSAILProgram::~HSAILProgram()
     }
     releaseClBinary();
     if (executable_ != NULL) {
-        Executable::Destroy(executable_);
+        loader_->DestroyExecutable(executable_);
     }
     delete kernels_;
+    amd::hsa::loader::Loader::Destroy(loader_);
 }
 
 bool
@@ -2132,7 +2136,7 @@ HSAILProgram::linkImpl(amd::option::Options* options)
     hsa_agent_t agent;
     agent.handle = 1;
     if (!isNull() && hsaLoad) {
-        executable_ = Executable::Create(HSA_PROFILE_BASE, &loaderContext_, NULL);
+        executable_ = loader_->CreateExecutable(HSA_PROFILE_BASE, NULL);
         if (executable_ == NULL) {
             buildLog_ += "Error: Executable for AMD HSA Code Object isn't created.\n";
             return false;
diff --git a/rocclr/runtime/device/gpu/gpuprogram.hpp b/rocclr/runtime/device/gpu/gpuprogram.hpp
index 01b61f1489..9f87328f35 100644
--- a/rocclr/runtime/device/gpu/gpuprogram.hpp
+++ b/rocclr/runtime/device/gpu/gpuprogram.hpp
@@ -15,6 +15,7 @@ class Options;
 } // option
 namespace hsa {
 namespace loader {
+class Loader;
 class Executable;
 class Context;
 } // loader
@@ -624,6 +625,7 @@ private:
     uint    maxScratchRegs_;    //!< Maximum number of scratch regs used in the program by individual kernel
     std::list<Sampler*>   staticSamplers_;    //!< List od internal static samplers
     bool            isNull_;        //!< Null program no memory allocations
+    amd::hsa::loader::Loader* loader_; //!< Loader object
     amd::hsa::loader::Executable* executable_;    //!< Executable for HSA Loader
     ORCAHSALoaderContext loaderContext_;    //!< Context for HSA Loader
 };