Mercurial > hg > graal-compiler
view src/gpu/hsail/vm/gpu_hsail.cpp @ 16865:8b0a6a9c3a4a
Added tag graal-0.4 for changeset 7d4f630172a1
author | Bernhard Urban <bernhard.urban@jku.at> |
---|---|
date | Tue, 19 Aug 2014 17:30:54 +0200 |
parents | a29e6e7b7a86 |
children | bbb987aec58d |
line wrap: on
line source
/* * Copyright (c) 2013, Oracle and/or its affiliates. All rights reserved. * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * * This code is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License version 2 only, as * published by the Free Software Foundation. * * This code is distributed in the hope that it will be useful, but WITHOUT * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License * version 2 for more details (a copy is included in the LICENSE file that * accompanied this code). * * You should have received a copy of the GNU General Public License version * 2 along with this work; if not, write to the Free Software Foundation, * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. * * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA * or visit www.oracle.com if you need additional information or have any * questions. * */ #include "precompiled.hpp" #include "runtime/javaCalls.hpp" #include "runtime/gpu.hpp" #include "runtime/deoptimization.hpp" #include "gpu_hsail.hpp" #include "utilities/debug.hpp" #include "utilities/exceptions.hpp" #include "hsail/vm/gpu_hsail.hpp" #include "utilities/globalDefinitions.hpp" #include "utilities/ostream.hpp" #include "graal/graalEnv.hpp" #include "graal/graalRuntime.hpp" #include "graal/graalJavaAccess.hpp" #include "hsailKernelArguments.hpp" #include "hsailJavaCallArguments.hpp" #include "code/pcDesc.hpp" #include "code/scopeDesc.hpp" #include "gpu_hsail_Frame.hpp" // Entry to GPU native method implementation that transitions current thread to '_thread_in_vm'. #define GPU_VMENTRY(result_type, name, signature) \ JNIEXPORT result_type JNICALL name signature { \ GRAAL_VM_ENTRY_MARK; \ // Entry to GPU native method implementation that calls a JNI function // and hence cannot transition current thread to '_thread_in_vm'. #define GPU_ENTRY(result_type, name, signature) \ JNIEXPORT result_type JNICALL name signature { \ #define GPU_END } #define CC (char*) /*cast a literal from (const char*)*/ #define FN_PTR(f) CAST_FROM_FN_PTR(void*, &(f)) #define OBJECT "Ljava/lang/Object;" #define STRING "Ljava/lang/String;" #define JLTHREAD "Ljava/lang/Thread;" #define HS_INSTALLED_CODE "Lcom/oracle/graal/hotspot/meta/HotSpotInstalledCode;" #define HS_COMPILED_NMETHOD "Lcom/oracle/graal/hotspot/HotSpotCompiledNmethod;" #define HS_NMETHOD "Lcom/oracle/graal/hotspot/meta/HotSpotNmethod;" JNINativeMethod Hsail::HSAIL_methods[] = { {CC"initialize", CC"()Z", FN_PTR(Hsail::initialize)}, {CC"generateKernel", CC"([B" STRING ")J", FN_PTR(Hsail::generate_kernel)}, {CC"executeKernel0", CC"("HS_INSTALLED_CODE"I["OBJECT"II[I)Z", FN_PTR(Hsail::execute_kernel_void_1d)}, }; void* Hsail::_device_context = NULL; jint Hsail::_notice_safepoints = false; Hsail::okra_get_context_func_t Hsail::_okra_get_context; Hsail::okra_create_kernel_func_t Hsail::_okra_create_kernel; Hsail::okra_push_pointer_func_t Hsail::_okra_push_pointer; Hsail::okra_push_boolean_func_t Hsail::_okra_push_boolean; Hsail::okra_push_byte_func_t Hsail::_okra_push_byte; Hsail::okra_push_double_func_t Hsail::_okra_push_double; Hsail::okra_push_float_func_t Hsail::_okra_push_float; Hsail::okra_push_int_func_t Hsail::_okra_push_int; Hsail::okra_push_long_func_t Hsail::_okra_push_long; Hsail::okra_execute_kernel_func_t Hsail::_okra_execute_kernel; Hsail::okra_clear_args_func_t Hsail::_okra_clear_args; Hsail::okra_dispose_kernel_func_t Hsail::_okra_dispose_kernel; Hsail::okra_dispose_context_func_t Hsail::_okra_dispose_context; //static jint in_kernel = 0; void Hsail::notice_safepoints() { _notice_safepoints = true; // if (TraceGPUInteraction) { // tty->print_cr("[HSAIL] Notice safepoint in_kernel=%d", in_kernel); // } } void Hsail::ignore_safepoints() { _notice_safepoints = false; } GPU_VMENTRY(jboolean, Hsail::execute_kernel_void_1d, (JNIEnv* env, jclass, jobject kernel_handle, jint dimX, jobject args, jint num_tlabs, jint allocBytesPerWorkitem, jobject oop_map_array)) ResourceMark rm; jlong nmethodValue = InstalledCode::address(kernel_handle); if (nmethodValue == 0) { SharedRuntime::throw_and_post_jvmti_exception(JavaThread::current(), vmSymbols::com_oracle_graal_api_code_InvalidInstalledCodeException(), NULL); } nmethod* nm = (nmethod*) (address) nmethodValue; methodHandle mh = nm->method(); Symbol* signature = mh->signature(); void* kernel = (void*) HotSpotInstalledCode::codeStart(kernel_handle); if (kernel == NULL) { SharedRuntime::throw_and_post_jvmti_exception(JavaThread::current(), vmSymbols::com_oracle_graal_api_code_InvalidInstalledCodeException(), NULL); } return execute_kernel_void_1d_internal((address) kernel, dimX, args, mh, nm, num_tlabs, allocBytesPerWorkitem, oop_map_array, CHECK_0); GPU_END static void showRanges(jboolean* a, int len) { // show ranges bool lookFor = true; for (int i = 0; i < len; i++) { if ((lookFor == true) && (a[i] != 0)) { tty->print("%d", i); lookFor = false; } else if ((lookFor == false) && (a[i] == 0)) { tty->print_cr("-%d", i-1); lookFor = true; } } if (lookFor == false) { tty->print_cr("-%d", len-1); } } jboolean Hsail::execute_kernel_void_1d_internal(address kernel, int dimX, jobject args, methodHandle& mh, nmethod* nm, jint num_tlabs, int allocBytesPerWorkitem, jobject oop_map_array, TRAPS) { ResourceMark rm(THREAD); objArrayOop argsArray = (objArrayOop) JNIHandles::resolve(args); assert(THREAD->is_Java_thread(), "must be a JavaThread"); // We avoid HSAILAllocationInfo logic if kernel does not allocate // in which case the num_tlabs passed in will be 0 HSAILAllocationInfo* allocInfo = (num_tlabs == 0 ? NULL : new HSAILAllocationInfo(num_tlabs, dimX, allocBytesPerWorkitem)); // Reset the kernel arguments _okra_clear_args(kernel); JavaThread* thread = (JavaThread*)THREAD; HSAILDeoptimizationInfo* e; if (UseHSAILDeoptimization) { // get how many bytes per deopt save area are required int saveAreaCounts = HSAILOopMapHelper::get_save_area_counts(oop_map_array); int numSRegs = saveAreaCounts & 0xff; int numDRegs = (saveAreaCounts >> 8) & 0xff; int numStackSlots = (saveAreaCounts >> 16); int bytesPerSaveArea = numSRegs * 4 + (numDRegs + numStackSlots) * 8; e = new (MAX_DEOPT_SLOTS, bytesPerSaveArea) HSAILDeoptimizationInfo(MAX_DEOPT_SLOTS, bytesPerSaveArea, dimX, allocInfo, oop_map_array); // copy cur_tlab_infos if (allocInfo != NULL) { e->set_cur_tlabInfos(allocInfo->getCurTlabInfos()); } // set deopt info in thread so gc oops_do processing can find it thread->set_gpu_hsail_deopt_info(e); } // This object sets up the kernel arguments HSAILKernelArguments hka((address) kernel, mh->signature(), argsArray, mh->is_static(), e); if (TraceGPUInteraction) { tty->print_cr("[HSAIL] range=%d", dimX); } // If any object passed was null, throw an exception here. Doing this // means the kernel code can avoid null checks on the object parameters. if (hka.getFirstNullParameterIndex() >= 0) { char buf[64]; sprintf(buf, "Null Kernel Parameter seen, Parameter Index: %d", hka.getFirstNullParameterIndex()); thread->set_gpu_exception_bci(0); thread->set_gpu_exception_method(mh()); THROW_MSG_0(vmSymbols::java_lang_NullPointerException(), buf); } // Run the kernel jboolean success = false; { TraceTime t("execute kernel", TraceGPUInteraction); graal_okra_range_t kernel_range = {0}; //in_kernel = 1; // Run the kernel kernel_range.dimension = 1; kernel_range.global_size[0] = dimX; success = _okra_execute_kernel(_device_context, kernel, &kernel_range); //in_kernel = 0; } // avoid HSAILAllocationInfo logic if kernel does not allocate if (allocInfo != NULL) { allocInfo->postKernelCleanup(); } if (UseHSAILDeoptimization) { // check if any workitem requested a deopt int deoptcode = e->deopt_occurred(); if (deoptcode != 1) { if (deoptcode == 0) { if (TraceGPUInteraction && _notice_safepoints != 0) { tty->print_cr("[HSAIL] observed safepoint during kernel"); } } else { // error condition detected in deopt code char msg[200]; sprintf(msg, "deopt error detected, slot for workitem %d was not empty", -1 * (deoptcode + 1)); guarantee(deoptcode == 1, msg); } } else { { TraceTime t3("handle deoptimizing workitems", TraceGPUInteraction); if (TraceGPUInteraction) { tty->print_cr("deopt happened."); HSAILKernelDeoptimization* pdeopt = e->get_deopt_save_state(0); tty->print_cr("first deopter was workitem %d", pdeopt->workitem()); } // Handle any deopting workitems. int count_deoptimized = 0; for (int k = 0; k < e->num_deopts(); k++) { HSAILKernelDeoptimization* pdeopt = e->get_deopt_save_state(k); jint workitem = pdeopt->workitem(); if (workitem != -1) { int deoptId = pdeopt->pc_offset(); HSAILFrame* hsailFrame = pdeopt->first_frame(); JavaValue result(T_VOID); JavaCallArguments javaArgs; javaArgs.set_alternative_target(nm); javaArgs.push_int(deoptId); javaArgs.push_long((jlong) hsailFrame); // Override the deoptimization action with Action_none until we decide // how to handle the other actions. int myActionReason = Deoptimization::make_trap_request(Deoptimization::trap_request_reason(pdeopt->reason()), Deoptimization::Action_none); javaArgs.push_int(myActionReason); javaArgs.push_oop((oop) NULL); if (TraceGPUInteraction) { tty->print_cr("[HSAIL] Deoptimizing to host for workitem=%d (slot=%d) with deoptId=%d, frame=" INTPTR_FORMAT ", actionAndReason=%d", workitem, k, deoptId, hsailFrame, myActionReason); // show the $d registers or stack slots containing references int maxOopBits = hsailFrame->num_d_regs() + hsailFrame->num_stack_slots(); HSAILOopMapHelper oopMapHelper(oop_map_array); int pc_offset = hsailFrame->pc_offset(); for (int bit = 0; bit < maxOopBits; bit++) { if (oopMapHelper.is_oop(pc_offset, bit)) { if (bit < hsailFrame->num_d_regs()) { // show $d reg oop tty->print_cr(" oop $d%d = %p", bit, hsailFrame->get_oop_for_bit(bit)); } else { // show stack slot oop int stackOffset = (bit - hsailFrame->num_d_regs()) * 8; // 8 bytes per stack slot tty->print_cr(" oop stk:%d = %p", stackOffset, hsailFrame->get_oop_for_bit(bit)); } } } } JavaCalls::call(&result, mh, &javaArgs, THREAD); count_deoptimized++; e->set_deopt_work_index(k + 1); } } if (TraceGPUInteraction) { tty->print_cr("[HSAIL] Deoptimizing to host completed for %d workitems", count_deoptimized); } } } // when we are done with the deopts, we don't need to oops_do anything // in the saved state anymore thread->set_gpu_hsail_deopt_info(NULL); // Handle any never_ran workitems if there were any { TraceTime t("handle never-rans ", TraceGPUInteraction); int count_never_ran = 0; bool handleNeverRansHere = true; // turn off verbose trace stuff for javacall arg setup bool savedTraceGPUInteraction = TraceGPUInteraction; TraceGPUInteraction = false; jboolean* never_ran_array = e->never_ran_array(); if (handleNeverRansHere) { for (int k = 0; k < dimX; k++) { if (never_ran_array[k]) { // run it as a javaCall KlassHandle methKlass = mh->method_holder(); Thread* THREAD = Thread::current(); JavaValue result(T_VOID); JavaCallArguments javaArgs; // re-resolve the args_handle here objArrayOop resolvedArgsArray = (objArrayOop) JNIHandles::resolve(args); // This object sets up the javaCall arguments. The way // argsArray is set up, this should work for instance // methods as well (the receiver will be the first oop pushed) HSAILJavaCallArguments hjca(&javaArgs, k, mh->signature(), resolvedArgsArray, mh->is_static()); if (mh->is_static()) { JavaCalls::call_static(&result, methKlass, mh->name(), mh->signature(), &javaArgs, THREAD); } else { JavaCalls::call_virtual(&result, methKlass, mh->name(), mh->signature(), &javaArgs, THREAD); } count_never_ran++; } } TraceGPUInteraction = savedTraceGPUInteraction; if (TraceGPUInteraction && (count_never_ran > 0)) { tty->print_cr("%d workitems never ran, have been run via JavaCall", count_never_ran); showRanges(never_ran_array, dimX); } } // end of never-ran handling } delete e; delete allocInfo; } return success; } GPU_ENTRY(jlong, Hsail::generate_kernel, (JNIEnv* env, jclass, jbyteArray code_handle, jstring name_handle)) guarantee(_okra_create_kernel != NULL, "[HSAIL] Okra not linked"); ResourceMark rm; jsize name_len = env->GetStringLength(name_handle); jsize code_len = env->GetArrayLength(code_handle); char* name = NEW_RESOURCE_ARRAY(char, name_len + 1); unsigned char* code = NEW_RESOURCE_ARRAY(unsigned char, code_len + 1); code[code_len] = 0; name[name_len] = 0; env->GetByteArrayRegion(code_handle, 0, code_len, (jbyte*) code); env->GetStringUTFRegion(name_handle, 0, name_len, name); // The kernel entrypoint is always run for the time being const char* entryPointName = "&run"; jlong okra_kernel; jint okra_status = _okra_create_kernel(_device_context, code, entryPointName, (void**)&okra_kernel); guarantee(okra_status==0, "_okra_create_kernel failed"); return (jlong) okra_kernel; GPU_END #if defined(LINUX) static const char* okra_library_name = "libokra_x86_64.so"; #elif defined(_WINDOWS) static char const* okra_library_name = "okra_x86_64.dll"; #else static char const* okra_library_name = NULL; #endif #define STRINGIFY(x) #x #define LOOKUP_OKRA_FUNCTION(name, alias) \ _##alias = \ CAST_TO_FN_PTR(alias##_func_t, os::dll_lookup(okra_lib_handle, STRINGIFY(name))); \ if (_##alias == NULL) { \ tty->print_cr("[HSAIL] ***** Error: Failed to lookup %s in %s, wrong version of OKRA?", STRINGIFY(name), okra_library_name); \ return false; \ } \ GPU_ENTRY(jboolean, Hsail::initialize, (JNIEnv* env, jclass)) if (okra_library_name == NULL) { if (TraceGPUInteraction) { tty->print_cr("Unsupported HSAIL platform"); } return false; } // here we know we have a valid okra_library_name to try to load char ebuf[O_BUFLEN]; char* okra_lib_name_from_env_var = getenv("_OKRA_SIM_LIB_PATH_"); if (okra_lib_name_from_env_var != NULL) { okra_library_name = okra_lib_name_from_env_var; } if (TraceGPUInteraction) { tty->print_cr("[HSAIL] library is %s", okra_library_name); } void* okra_lib_handle = NULL; #if defined(LINUX) // Check first if the Okra library is already loaded. // TODO: Figure out how to do this on other OSes. okra_lib_handle = ::dlopen(okra_library_name, RTLD_LAZY | RTLD_NOLOAD); #endif // If Okra library is not already loaded, load it here if (okra_lib_handle == NULL) { okra_lib_handle = os::dll_load(okra_library_name, ebuf, O_BUFLEN); } if (okra_lib_handle == NULL) { // Unable to open Okra library if (TraceGPUInteraction) { tty->print_cr("[HSAIL] library load failed."); } return false; } guarantee(_okra_get_context == NULL, "cannot repeat GPU initialization"); // At this point we know okra_lib_handle is valid whether we loaded // here or earlier. In either case, we can lookup the functions. LOOKUP_OKRA_FUNCTION(okra_get_context, okra_get_context); LOOKUP_OKRA_FUNCTION(okra_create_kernel, okra_create_kernel); LOOKUP_OKRA_FUNCTION(okra_push_pointer, okra_push_pointer); LOOKUP_OKRA_FUNCTION(okra_push_boolean, okra_push_boolean); LOOKUP_OKRA_FUNCTION(okra_push_byte, okra_push_byte); LOOKUP_OKRA_FUNCTION(okra_push_double, okra_push_double); LOOKUP_OKRA_FUNCTION(okra_push_float, okra_push_float); LOOKUP_OKRA_FUNCTION(okra_push_int, okra_push_int); LOOKUP_OKRA_FUNCTION(okra_push_long, okra_push_long); LOOKUP_OKRA_FUNCTION(okra_execute_kernel, okra_execute_kernel); LOOKUP_OKRA_FUNCTION(okra_clear_args, okra_clear_args); LOOKUP_OKRA_FUNCTION(okra_dispose_kernel, okra_dispose_kernel); LOOKUP_OKRA_FUNCTION(okra_dispose_context, okra_dispose_context); // if we made it this far, real success Gpu::initialized_gpu(new Hsail()); // There is 1 context per process jint result = _okra_get_context(&_device_context); guarantee(result==0, "get context failed"); return true; GPU_END bool Hsail::register_natives(JNIEnv* env) { jclass klass = env->FindClass("com/oracle/graal/hotspot/hsail/HSAILHotSpotBackend"); if (klass == NULL) { if (TraceGPUInteraction) { tty->print_cr("HSAILHotSpotBackend class not found"); } return false; } jint status = env->RegisterNatives(klass, HSAIL_methods, sizeof(HSAIL_methods) / sizeof(JNINativeMethod)); if (status != JNI_OK) { if (TraceGPUInteraction) { tty->print_cr("Error registering natives for HSAILHotSpotBackend: %d", status); } return false; } return true; } void Hsail::HSAILDeoptimizationInfo::oops_do(OopClosure* f) { int unprocessed_deopts = num_deopts() - deopt_work_index(); if (TraceGPUInteraction) { tty->print_cr("HSAILDeoptimizationInfo::oops_do deopt_occurred=%d, total_deopts=%d, unprocessed_deopts=%d, oop_map_array=%p", _deopt_occurred, num_deopts(), unprocessed_deopts, _oop_map_array); } if (num_deopts() == 0 || unprocessed_deopts <= 0) { return; // nothing to do } HSAILOopMapHelper oopMapHelper(_oop_map_array); oopMapHelper.resolve_arrays(); // resolve once before processing // go thru the unprocessed deopt frames, finding each oop and applying the closre for (int k = deopt_work_index(); k < num_deopts(); k++) { HSAILKernelDeoptimization* pdeopt = get_deopt_save_state(k); assert (pdeopt->workitem() >= 0, "bad workitem in deopt"); if (TraceGPUInteraction) { tty->print_cr(" deopt %d, workitem %d, pc %d", k, pdeopt->workitem(), pdeopt->pc_offset()); } HSAILFrame* hsailFrame = pdeopt->first_frame(); hsailFrame->oops_do(f, &oopMapHelper); } }