Mercurial > hg > graal-compiler
view src/share/vm/interpreter/interpreterRuntime.cpp @ 2149:7e37af9d69ef
7011379: G1: overly long concurrent marking cycles
Summary: This changeset introduces filtering of SATB buffers at the point when they are about to be enqueued. If this filtering clears enough entries on each buffer, the buffer can then be re-used and not enqueued. This cuts down the number of SATB buffers that need to be processed by the concurrent marking threads.
Reviewed-by: johnc, ysr
author | tonyp |
---|---|
date | Wed, 19 Jan 2011 09:35:17 -0500 |
parents | dd031b2226de |
children | 633a44a9fc45 |
line wrap: on
line source
/* * Copyright (c) 1997, 2011, 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 "classfile/systemDictionary.hpp" #include "classfile/vmSymbols.hpp" #include "compiler/compileBroker.hpp" #include "gc_interface/collectedHeap.hpp" #include "interpreter/interpreter.hpp" #include "interpreter/interpreterRuntime.hpp" #include "interpreter/linkResolver.hpp" #include "interpreter/templateTable.hpp" #include "memory/oopFactory.hpp" #include "memory/universe.inline.hpp" #include "oops/constantPoolOop.hpp" #include "oops/cpCacheOop.hpp" #include "oops/instanceKlass.hpp" #include "oops/methodDataOop.hpp" #include "oops/objArrayKlass.hpp" #include "oops/oop.inline.hpp" #include "oops/symbolOop.hpp" #include "prims/jvmtiExport.hpp" #include "prims/nativeLookup.hpp" #include "runtime/biasedLocking.hpp" #include "runtime/compilationPolicy.hpp" #include "runtime/deoptimization.hpp" #include "runtime/fieldDescriptor.hpp" #include "runtime/handles.inline.hpp" #include "runtime/interfaceSupport.hpp" #include "runtime/java.hpp" #include "runtime/jfieldIDWorkaround.hpp" #include "runtime/osThread.hpp" #include "runtime/sharedRuntime.hpp" #include "runtime/stubRoutines.hpp" #include "runtime/synchronizer.hpp" #include "runtime/threadCritical.hpp" #include "utilities/events.hpp" #ifdef TARGET_ARCH_x86 # include "vm_version_x86.hpp" #endif #ifdef TARGET_ARCH_sparc # include "vm_version_sparc.hpp" #endif #ifdef TARGET_ARCH_zero # include "vm_version_zero.hpp" #endif #ifdef COMPILER2 #include "opto/runtime.hpp" #endif class UnlockFlagSaver { private: JavaThread* _thread; bool _do_not_unlock; public: UnlockFlagSaver(JavaThread* t) { _thread = t; _do_not_unlock = t->do_not_unlock_if_synchronized(); t->set_do_not_unlock_if_synchronized(false); } ~UnlockFlagSaver() { _thread->set_do_not_unlock_if_synchronized(_do_not_unlock); } }; //------------------------------------------------------------------------------------------------------------------------ // State accessors void InterpreterRuntime::set_bcp_and_mdp(address bcp, JavaThread *thread) { last_frame(thread).interpreter_frame_set_bcp(bcp); if (ProfileInterpreter) { // ProfileTraps uses MDOs independently of ProfileInterpreter. // That is why we must check both ProfileInterpreter and mdo != NULL. methodDataOop mdo = last_frame(thread).interpreter_frame_method()->method_data(); if (mdo != NULL) { NEEDS_CLEANUP; last_frame(thread).interpreter_frame_set_mdp(mdo->bci_to_dp(last_frame(thread).interpreter_frame_bci())); } } } //------------------------------------------------------------------------------------------------------------------------ // Constants IRT_ENTRY(void, InterpreterRuntime::ldc(JavaThread* thread, bool wide)) // access constant pool constantPoolOop pool = method(thread)->constants(); int index = wide ? get_index_u2(thread, Bytecodes::_ldc_w) : get_index_u1(thread, Bytecodes::_ldc); constantTag tag = pool->tag_at(index); if (tag.is_unresolved_klass() || tag.is_klass()) { klassOop klass = pool->klass_at(index, CHECK); oop java_class = klass->klass_part()->java_mirror(); thread->set_vm_result(java_class); } else { #ifdef ASSERT // If we entered this runtime routine, we believed the tag contained // an unresolved string, an unresolved class or a resolved class. // However, another thread could have resolved the unresolved string // or class by the time we go there. assert(tag.is_unresolved_string()|| tag.is_string(), "expected string"); #endif oop s_oop = pool->string_at(index, CHECK); thread->set_vm_result(s_oop); } IRT_END IRT_ENTRY(void, InterpreterRuntime::resolve_ldc(JavaThread* thread, Bytecodes::Code bytecode)) { assert(bytecode == Bytecodes::_fast_aldc || bytecode == Bytecodes::_fast_aldc_w, "wrong bc"); ResourceMark rm(thread); methodHandle m (thread, method(thread)); Bytecode_loadconstant* ldc = Bytecode_loadconstant_at(m, bci(thread)); oop result = ldc->resolve_constant(THREAD); DEBUG_ONLY(ConstantPoolCacheEntry* cpce = m->constants()->cache()->entry_at(ldc->cache_index())); assert(result == cpce->f1(), "expected result for assembly code"); } IRT_END //------------------------------------------------------------------------------------------------------------------------ // Allocation IRT_ENTRY(void, InterpreterRuntime::_new(JavaThread* thread, constantPoolOopDesc* pool, int index)) klassOop k_oop = pool->klass_at(index, CHECK); instanceKlassHandle klass (THREAD, k_oop); // Make sure we are not instantiating an abstract klass klass->check_valid_for_instantiation(true, CHECK); // Make sure klass is initialized klass->initialize(CHECK); // At this point the class may not be fully initialized // because of recursive initialization. If it is fully // initialized & has_finalized is not set, we rewrite // it into its fast version (Note: no locking is needed // here since this is an atomic byte write and can be // done more than once). // // Note: In case of classes with has_finalized we don't // rewrite since that saves us an extra check in // the fast version which then would call the // slow version anyway (and do a call back into // Java). // If we have a breakpoint, then we don't rewrite // because the _breakpoint bytecode would be lost. oop obj = klass->allocate_instance(CHECK); thread->set_vm_result(obj); IRT_END IRT_ENTRY(void, InterpreterRuntime::newarray(JavaThread* thread, BasicType type, jint size)) oop obj = oopFactory::new_typeArray(type, size, CHECK); thread->set_vm_result(obj); IRT_END IRT_ENTRY(void, InterpreterRuntime::anewarray(JavaThread* thread, constantPoolOopDesc* pool, int index, jint size)) // Note: no oopHandle for pool & klass needed since they are not used // anymore after new_objArray() and no GC can happen before. // (This may have to change if this code changes!) klassOop klass = pool->klass_at(index, CHECK); objArrayOop obj = oopFactory::new_objArray(klass, size, CHECK); thread->set_vm_result(obj); IRT_END IRT_ENTRY(void, InterpreterRuntime::multianewarray(JavaThread* thread, jint* first_size_address)) // We may want to pass in more arguments - could make this slightly faster constantPoolOop constants = method(thread)->constants(); int i = get_index_u2(thread, Bytecodes::_multianewarray); klassOop klass = constants->klass_at(i, CHECK); int nof_dims = number_of_dimensions(thread); assert(oop(klass)->is_klass(), "not a class"); assert(nof_dims >= 1, "multianewarray rank must be nonzero"); // We must create an array of jints to pass to multi_allocate. ResourceMark rm(thread); const int small_dims = 10; jint dim_array[small_dims]; jint *dims = &dim_array[0]; if (nof_dims > small_dims) { dims = (jint*) NEW_RESOURCE_ARRAY(jint, nof_dims); } for (int index = 0; index < nof_dims; index++) { // offset from first_size_address is addressed as local[index] int n = Interpreter::local_offset_in_bytes(index)/jintSize; dims[index] = first_size_address[n]; } oop obj = arrayKlass::cast(klass)->multi_allocate(nof_dims, dims, CHECK); thread->set_vm_result(obj); IRT_END IRT_ENTRY(void, InterpreterRuntime::register_finalizer(JavaThread* thread, oopDesc* obj)) assert(obj->is_oop(), "must be a valid oop"); assert(obj->klass()->klass_part()->has_finalizer(), "shouldn't be here otherwise"); instanceKlass::register_finalizer(instanceOop(obj), CHECK); IRT_END // Quicken instance-of and check-cast bytecodes IRT_ENTRY(void, InterpreterRuntime::quicken_io_cc(JavaThread* thread)) // Force resolving; quicken the bytecode int which = get_index_u2(thread, Bytecodes::_checkcast); constantPoolOop cpool = method(thread)->constants(); // We'd expect to assert that we're only here to quicken bytecodes, but in a multithreaded // program we might have seen an unquick'd bytecode in the interpreter but have another // thread quicken the bytecode before we get here. // assert( cpool->tag_at(which).is_unresolved_klass(), "should only come here to quicken bytecodes" ); klassOop klass = cpool->klass_at(which, CHECK); thread->set_vm_result(klass); IRT_END //------------------------------------------------------------------------------------------------------------------------ // Exceptions // Assume the compiler is (or will be) interested in this event. // If necessary, create an MDO to hold the information, and record it. void InterpreterRuntime::note_trap(JavaThread* thread, int reason, TRAPS) { assert(ProfileTraps, "call me only if profiling"); methodHandle trap_method(thread, method(thread)); if (trap_method.not_null()) { methodDataHandle trap_mdo(thread, trap_method->method_data()); if (trap_mdo.is_null()) { methodOopDesc::build_interpreter_method_data(trap_method, THREAD); if (HAS_PENDING_EXCEPTION) { assert((PENDING_EXCEPTION->is_a(SystemDictionary::OutOfMemoryError_klass())), "we expect only an OOM error here"); CLEAR_PENDING_EXCEPTION; } trap_mdo = methodDataHandle(thread, trap_method->method_data()); // and fall through... } if (trap_mdo.not_null()) { // Update per-method count of trap events. The interpreter // is updating the MDO to simulate the effect of compiler traps. int trap_bci = trap_method->bci_from(bcp(thread)); Deoptimization::update_method_data_from_interpreter(trap_mdo, trap_bci, reason); } } } static Handle get_preinitialized_exception(klassOop k, TRAPS) { // get klass instanceKlass* klass = instanceKlass::cast(k); assert(klass->is_initialized(), "this klass should have been initialized during VM initialization"); // create instance - do not call constructor since we may have no // (java) stack space left (should assert constructor is empty) Handle exception; oop exception_oop = klass->allocate_instance(CHECK_(exception)); exception = Handle(THREAD, exception_oop); if (StackTraceInThrowable) { java_lang_Throwable::fill_in_stack_trace(exception); } return exception; } // Special handling for stack overflow: since we don't have any (java) stack // space left we use the pre-allocated & pre-initialized StackOverflowError // klass to create an stack overflow error instance. We do not call its // constructor for the same reason (it is empty, anyway). IRT_ENTRY(void, InterpreterRuntime::throw_StackOverflowError(JavaThread* thread)) Handle exception = get_preinitialized_exception( SystemDictionary::StackOverflowError_klass(), CHECK); THROW_HANDLE(exception); IRT_END IRT_ENTRY(void, InterpreterRuntime::create_exception(JavaThread* thread, char* name, char* message)) // lookup exception klass symbolHandle s = oopFactory::new_symbol_handle(name, CHECK); if (ProfileTraps) { if (s == vmSymbols::java_lang_ArithmeticException()) { note_trap(thread, Deoptimization::Reason_div0_check, CHECK); } else if (s == vmSymbols::java_lang_NullPointerException()) { note_trap(thread, Deoptimization::Reason_null_check, CHECK); } } // create exception Handle exception = Exceptions::new_exception(thread, s(), message); thread->set_vm_result(exception()); IRT_END IRT_ENTRY(void, InterpreterRuntime::create_klass_exception(JavaThread* thread, char* name, oopDesc* obj)) ResourceMark rm(thread); const char* klass_name = Klass::cast(obj->klass())->external_name(); // lookup exception klass symbolHandle s = oopFactory::new_symbol_handle(name, CHECK); if (ProfileTraps) { note_trap(thread, Deoptimization::Reason_class_check, CHECK); } // create exception, with klass name as detail message Handle exception = Exceptions::new_exception(thread, s(), klass_name); thread->set_vm_result(exception()); IRT_END IRT_ENTRY(void, InterpreterRuntime::throw_ArrayIndexOutOfBoundsException(JavaThread* thread, char* name, jint index)) char message[jintAsStringSize]; // lookup exception klass symbolHandle s = oopFactory::new_symbol_handle(name, CHECK); if (ProfileTraps) { note_trap(thread, Deoptimization::Reason_range_check, CHECK); } // create exception sprintf(message, "%d", index); THROW_MSG(s(), message); IRT_END IRT_ENTRY(void, InterpreterRuntime::throw_ClassCastException( JavaThread* thread, oopDesc* obj)) ResourceMark rm(thread); char* message = SharedRuntime::generate_class_cast_message( thread, Klass::cast(obj->klass())->external_name()); if (ProfileTraps) { note_trap(thread, Deoptimization::Reason_class_check, CHECK); } // create exception THROW_MSG(vmSymbols::java_lang_ClassCastException(), message); IRT_END // required can be either a MethodType, or a Class (for a single argument) // actual (if not null) can be either a MethodHandle, or an arbitrary value (for a single argument) IRT_ENTRY(void, InterpreterRuntime::throw_WrongMethodTypeException(JavaThread* thread, oopDesc* required, oopDesc* actual)) { ResourceMark rm(thread); char* message = SharedRuntime::generate_wrong_method_type_message(thread, required, actual); if (ProfileTraps) { note_trap(thread, Deoptimization::Reason_constraint, CHECK); } // create exception THROW_MSG(vmSymbols::java_dyn_WrongMethodTypeException(), message); } IRT_END // exception_handler_for_exception(...) returns the continuation address, // the exception oop (via TLS) and sets the bci/bcp for the continuation. // The exception oop is returned to make sure it is preserved over GC (it // is only on the stack if the exception was thrown explicitly via athrow). // During this operation, the expression stack contains the values for the // bci where the exception happened. If the exception was propagated back // from a call, the expression stack contains the values for the bci at the // invoke w/o arguments (i.e., as if one were inside the call). IRT_ENTRY(address, InterpreterRuntime::exception_handler_for_exception(JavaThread* thread, oopDesc* exception)) Handle h_exception(thread, exception); methodHandle h_method (thread, method(thread)); constantPoolHandle h_constants(thread, h_method->constants()); typeArrayHandle h_extable (thread, h_method->exception_table()); bool should_repeat; int handler_bci; int current_bci = bci(thread); // Need to do this check first since when _do_not_unlock_if_synchronized // is set, we don't want to trigger any classloading which may make calls // into java, or surprisingly find a matching exception handler for bci 0 // since at this moment the method hasn't been "officially" entered yet. if (thread->do_not_unlock_if_synchronized()) { ResourceMark rm; assert(current_bci == 0, "bci isn't zero for do_not_unlock_if_synchronized"); thread->set_vm_result(exception); #ifdef CC_INTERP return (address) -1; #else return Interpreter::remove_activation_entry(); #endif } do { should_repeat = false; // assertions #ifdef ASSERT assert(h_exception.not_null(), "NULL exceptions should be handled by athrow"); assert(h_exception->is_oop(), "just checking"); // Check that exception is a subclass of Throwable, otherwise we have a VerifyError if (!(h_exception->is_a(SystemDictionary::Throwable_klass()))) { if (ExitVMOnVerifyError) vm_exit(-1); ShouldNotReachHere(); } #endif // tracing if (TraceExceptions) { ttyLocker ttyl; ResourceMark rm(thread); tty->print_cr("Exception <%s> (" INTPTR_FORMAT ")", h_exception->print_value_string(), (address)h_exception()); tty->print_cr(" thrown in interpreter method <%s>", h_method->print_value_string()); tty->print_cr(" at bci %d for thread " INTPTR_FORMAT, current_bci, thread); } // Don't go paging in something which won't be used. // else if (h_extable->length() == 0) { // // disabled for now - interpreter is not using shortcut yet // // (shortcut is not to call runtime if we have no exception handlers) // // warning("performance bug: should not call runtime if method has no exception handlers"); // } // for AbortVMOnException flag NOT_PRODUCT(Exceptions::debug_check_abort(h_exception)); // exception handler lookup KlassHandle h_klass(THREAD, h_exception->klass()); handler_bci = h_method->fast_exception_handler_bci_for(h_klass, current_bci, THREAD); if (HAS_PENDING_EXCEPTION) { // We threw an exception while trying to find the exception handler. // Transfer the new exception to the exception handle which will // be set into thread local storage, and do another lookup for an // exception handler for this exception, this time starting at the // BCI of the exception handler which caused the exception to be // thrown (bug 4307310). h_exception = Handle(THREAD, PENDING_EXCEPTION); CLEAR_PENDING_EXCEPTION; if (handler_bci >= 0) { current_bci = handler_bci; should_repeat = true; } } } while (should_repeat == true); // notify JVMTI of an exception throw; JVMTI will detect if this is a first // time throw or a stack unwinding throw and accordingly notify the debugger if (JvmtiExport::can_post_on_exceptions()) { JvmtiExport::post_exception_throw(thread, h_method(), bcp(thread), h_exception()); } #ifdef CC_INTERP address continuation = (address)(intptr_t) handler_bci; #else address continuation = NULL; #endif address handler_pc = NULL; if (handler_bci < 0 || !thread->reguard_stack((address) &continuation)) { // Forward exception to callee (leaving bci/bcp untouched) because (a) no // handler in this method, or (b) after a stack overflow there is not yet // enough stack space available to reprotect the stack. #ifndef CC_INTERP continuation = Interpreter::remove_activation_entry(); #endif // Count this for compilation purposes h_method->interpreter_throwout_increment(); } else { // handler in this method => change bci/bcp to handler bci/bcp and continue there handler_pc = h_method->code_base() + handler_bci; #ifndef CC_INTERP set_bcp_and_mdp(handler_pc, thread); continuation = Interpreter::dispatch_table(vtos)[*handler_pc]; #endif } // notify debugger of an exception catch // (this is good for exceptions caught in native methods as well) if (JvmtiExport::can_post_on_exceptions()) { JvmtiExport::notice_unwind_due_to_exception(thread, h_method(), handler_pc, h_exception(), (handler_pc != NULL)); } thread->set_vm_result(h_exception()); return continuation; IRT_END IRT_ENTRY(void, InterpreterRuntime::throw_pending_exception(JavaThread* thread)) assert(thread->has_pending_exception(), "must only ne called if there's an exception pending"); // nothing to do - eventually we should remove this code entirely (see comments @ call sites) IRT_END IRT_ENTRY(void, InterpreterRuntime::throw_AbstractMethodError(JavaThread* thread)) THROW(vmSymbols::java_lang_AbstractMethodError()); IRT_END IRT_ENTRY(void, InterpreterRuntime::throw_IncompatibleClassChangeError(JavaThread* thread)) THROW(vmSymbols::java_lang_IncompatibleClassChangeError()); IRT_END //------------------------------------------------------------------------------------------------------------------------ // Fields // IRT_ENTRY(void, InterpreterRuntime::resolve_get_put(JavaThread* thread, Bytecodes::Code bytecode)) // resolve field FieldAccessInfo info; constantPoolHandle pool(thread, method(thread)->constants()); bool is_static = (bytecode == Bytecodes::_getstatic || bytecode == Bytecodes::_putstatic); { JvmtiHideSingleStepping jhss(thread); LinkResolver::resolve_field(info, pool, get_index_u2_cpcache(thread, bytecode), bytecode, false, CHECK); } // end JvmtiHideSingleStepping // check if link resolution caused cpCache to be updated if (already_resolved(thread)) return; // compute auxiliary field attributes TosState state = as_TosState(info.field_type()); // We need to delay resolving put instructions on final fields // until we actually invoke one. This is required so we throw // exceptions at the correct place. If we do not resolve completely // in the current pass, leaving the put_code set to zero will // cause the next put instruction to reresolve. bool is_put = (bytecode == Bytecodes::_putfield || bytecode == Bytecodes::_putstatic); Bytecodes::Code put_code = (Bytecodes::Code)0; // We also need to delay resolving getstatic instructions until the // class is intitialized. This is required so that access to the static // field will call the initialization function every time until the class // is completely initialized ala. in 2.17.5 in JVM Specification. instanceKlass *klass = instanceKlass::cast(info.klass()->as_klassOop()); bool uninitialized_static = ((bytecode == Bytecodes::_getstatic || bytecode == Bytecodes::_putstatic) && !klass->is_initialized()); Bytecodes::Code get_code = (Bytecodes::Code)0; if (!uninitialized_static) { get_code = ((is_static) ? Bytecodes::_getstatic : Bytecodes::_getfield); if (is_put || !info.access_flags().is_final()) { put_code = ((is_static) ? Bytecodes::_putstatic : Bytecodes::_putfield); } } cache_entry(thread)->set_field( get_code, put_code, info.klass(), info.field_index(), info.field_offset(), state, info.access_flags().is_final(), info.access_flags().is_volatile() ); IRT_END //------------------------------------------------------------------------------------------------------------------------ // Synchronization // // The interpreter's synchronization code is factored out so that it can // be shared by method invocation and synchronized blocks. //%note synchronization_3 static void trace_locking(Handle& h_locking_obj, bool is_locking) { ObjectSynchronizer::trace_locking(h_locking_obj, false, true, is_locking); } //%note monitor_1 IRT_ENTRY_NO_ASYNC(void, InterpreterRuntime::monitorenter(JavaThread* thread, BasicObjectLock* elem)) #ifdef ASSERT thread->last_frame().interpreter_frame_verify_monitor(elem); #endif if (PrintBiasedLockingStatistics) { Atomic::inc(BiasedLocking::slow_path_entry_count_addr()); } Handle h_obj(thread, elem->obj()); assert(Universe::heap()->is_in_reserved_or_null(h_obj()), "must be NULL or an object"); if (UseBiasedLocking) { // Retry fast entry if bias is revoked to avoid unnecessary inflation ObjectSynchronizer::fast_enter(h_obj, elem->lock(), true, CHECK); } else { ObjectSynchronizer::slow_enter(h_obj, elem->lock(), CHECK); } assert(Universe::heap()->is_in_reserved_or_null(elem->obj()), "must be NULL or an object"); #ifdef ASSERT thread->last_frame().interpreter_frame_verify_monitor(elem); #endif IRT_END //%note monitor_1 IRT_ENTRY_NO_ASYNC(void, InterpreterRuntime::monitorexit(JavaThread* thread, BasicObjectLock* elem)) #ifdef ASSERT thread->last_frame().interpreter_frame_verify_monitor(elem); #endif Handle h_obj(thread, elem->obj()); assert(Universe::heap()->is_in_reserved_or_null(h_obj()), "must be NULL or an object"); if (elem == NULL || h_obj()->is_unlocked()) { THROW(vmSymbols::java_lang_IllegalMonitorStateException()); } ObjectSynchronizer::slow_exit(h_obj(), elem->lock(), thread); // Free entry. This must be done here, since a pending exception might be installed on // exit. If it is not cleared, the exception handling code will try to unlock the monitor again. elem->set_obj(NULL); #ifdef ASSERT thread->last_frame().interpreter_frame_verify_monitor(elem); #endif IRT_END IRT_ENTRY(void, InterpreterRuntime::throw_illegal_monitor_state_exception(JavaThread* thread)) THROW(vmSymbols::java_lang_IllegalMonitorStateException()); IRT_END IRT_ENTRY(void, InterpreterRuntime::new_illegal_monitor_state_exception(JavaThread* thread)) // Returns an illegal exception to install into the current thread. The // pending_exception flag is cleared so normal exception handling does not // trigger. Any current installed exception will be overwritten. This // method will be called during an exception unwind. assert(!HAS_PENDING_EXCEPTION, "no pending exception"); Handle exception(thread, thread->vm_result()); assert(exception() != NULL, "vm result should be set"); thread->set_vm_result(NULL); // clear vm result before continuing (may cause memory leaks and assert failures) if (!exception->is_a(SystemDictionary::ThreadDeath_klass())) { exception = get_preinitialized_exception( SystemDictionary::IllegalMonitorStateException_klass(), CATCH); } thread->set_vm_result(exception()); IRT_END //------------------------------------------------------------------------------------------------------------------------ // Invokes IRT_ENTRY(Bytecodes::Code, InterpreterRuntime::get_original_bytecode_at(JavaThread* thread, methodOopDesc* method, address bcp)) return method->orig_bytecode_at(method->bci_from(bcp)); IRT_END IRT_ENTRY(void, InterpreterRuntime::set_original_bytecode_at(JavaThread* thread, methodOopDesc* method, address bcp, Bytecodes::Code new_code)) method->set_orig_bytecode_at(method->bci_from(bcp), new_code); IRT_END IRT_ENTRY(void, InterpreterRuntime::_breakpoint(JavaThread* thread, methodOopDesc* method, address bcp)) JvmtiExport::post_raw_breakpoint(thread, method, bcp); IRT_END IRT_ENTRY(void, InterpreterRuntime::resolve_invoke(JavaThread* thread, Bytecodes::Code bytecode)) // extract receiver from the outgoing argument list if necessary Handle receiver(thread, NULL); if (bytecode == Bytecodes::_invokevirtual || bytecode == Bytecodes::_invokeinterface) { ResourceMark rm(thread); methodHandle m (thread, method(thread)); Bytecode_invoke* call = Bytecode_invoke_at(m, bci(thread)); symbolHandle signature (thread, call->signature()); receiver = Handle(thread, thread->last_frame().interpreter_callee_receiver(signature)); assert(Universe::heap()->is_in_reserved_or_null(receiver()), "sanity check"); assert(receiver.is_null() || Universe::heap()->is_in_reserved(receiver->klass()), "sanity check"); } // resolve method CallInfo info; constantPoolHandle pool(thread, method(thread)->constants()); { JvmtiHideSingleStepping jhss(thread); LinkResolver::resolve_invoke(info, receiver, pool, get_index_u2_cpcache(thread, bytecode), bytecode, CHECK); if (JvmtiExport::can_hotswap_or_post_breakpoint()) { int retry_count = 0; while (info.resolved_method()->is_old()) { // It is very unlikely that method is redefined more than 100 times // in the middle of resolve. If it is looping here more than 100 times // means then there could be a bug here. guarantee((retry_count++ < 100), "Could not resolve to latest version of redefined method"); // method is redefined in the middle of resolve so re-try. LinkResolver::resolve_invoke(info, receiver, pool, get_index_u2_cpcache(thread, bytecode), bytecode, CHECK); } } } // end JvmtiHideSingleStepping // check if link resolution caused cpCache to be updated if (already_resolved(thread)) return; if (bytecode == Bytecodes::_invokeinterface) { if (TraceItables && Verbose) { ResourceMark rm(thread); tty->print_cr("Resolving: klass: %s to method: %s", info.resolved_klass()->name()->as_C_string(), info.resolved_method()->name()->as_C_string()); } if (info.resolved_method()->method_holder() == SystemDictionary::Object_klass()) { // NOTE: THIS IS A FIX FOR A CORNER CASE in the JVM spec // (see also cpCacheOop.cpp for details) methodHandle rm = info.resolved_method(); assert(rm->is_final() || info.has_vtable_index(), "should have been set already"); cache_entry(thread)->set_method(bytecode, rm, info.vtable_index()); } else { // Setup itable entry int index = klassItable::compute_itable_index(info.resolved_method()()); cache_entry(thread)->set_interface_call(info.resolved_method(), index); } } else { cache_entry(thread)->set_method( bytecode, info.resolved_method(), info.vtable_index()); } IRT_END // First time execution: Resolve symbols, create a permanent CallSite object. IRT_ENTRY(void, InterpreterRuntime::resolve_invokedynamic(JavaThread* thread)) { ResourceMark rm(thread); assert(EnableInvokeDynamic, ""); const Bytecodes::Code bytecode = Bytecodes::_invokedynamic; methodHandle caller_method(thread, method(thread)); constantPoolHandle pool(thread, caller_method->constants()); pool->set_invokedynamic(); // mark header to flag active call sites int caller_bci = 0; int site_index = 0; { address caller_bcp = bcp(thread); caller_bci = caller_method->bci_from(caller_bcp); site_index = Bytes::get_native_u4(caller_bcp+1); } assert(site_index == InterpreterRuntime::bytecode(thread)->get_index_u4(bytecode), ""); assert(constantPoolCacheOopDesc::is_secondary_index(site_index), "proper format"); // there is a second CPC entries that is of interest; it caches signature info: int main_index = pool->cache()->secondary_entry_at(site_index)->main_entry_index(); int pool_index = pool->cache()->entry_at(main_index)->constant_pool_index(); // first resolve the signature to a MH.invoke methodOop if (!pool->cache()->entry_at(main_index)->is_resolved(bytecode)) { JvmtiHideSingleStepping jhss(thread); CallInfo callinfo; LinkResolver::resolve_invoke(callinfo, Handle(), pool, site_index, bytecode, CHECK); // The main entry corresponds to a JVM_CONSTANT_InvokeDynamic, and serves // as a common reference point for all invokedynamic call sites with // that exact call descriptor. We will link it in the CP cache exactly // as if it were an invokevirtual of MethodHandle.invoke. pool->cache()->entry_at(main_index)->set_method( bytecode, callinfo.resolved_method(), callinfo.vtable_index()); } // The method (f2 entry) of the main entry is the MH.invoke for the // invokedynamic target call signature. oop f1_value = pool->cache()->entry_at(main_index)->f1(); methodHandle signature_invoker(THREAD, (methodOop) f1_value); assert(signature_invoker.not_null() && signature_invoker->is_method() && signature_invoker->is_method_handle_invoke(), "correct result from LinkResolver::resolve_invokedynamic"); Handle info; // optional argument(s) in JVM_CONSTANT_InvokeDynamic Handle bootm = SystemDictionary::find_bootstrap_method(caller_method, caller_bci, main_index, info, CHECK); if (!java_dyn_MethodHandle::is_instance(bootm())) { THROW_MSG(vmSymbols::java_lang_IllegalStateException(), "no bootstrap method found for invokedynamic"); } // Short circuit if CallSite has been bound already: if (!pool->cache()->secondary_entry_at(site_index)->is_f1_null()) return; symbolHandle call_site_name(THREAD, pool->name_ref_at(site_index)); Handle call_site = SystemDictionary::make_dynamic_call_site(bootm, // Callee information: call_site_name, signature_invoker, info, // Caller information: caller_method, caller_bci, CHECK); // In the secondary entry, the f1 field is the call site, and the f2 (index) // field is some data about the invoke site. Currently, it is just the BCI. // Later, it might be changed to help manage inlining dependencies. pool->cache()->secondary_entry_at(site_index)->set_dynamic_call(call_site, signature_invoker); } IRT_END //------------------------------------------------------------------------------------------------------------------------ // Miscellaneous nmethod* InterpreterRuntime::frequency_counter_overflow(JavaThread* thread, address branch_bcp) { nmethod* nm = frequency_counter_overflow_inner(thread, branch_bcp); assert(branch_bcp != NULL || nm == NULL, "always returns null for non OSR requests"); if (branch_bcp != NULL && nm != NULL) { // This was a successful request for an OSR nmethod. Because // frequency_counter_overflow_inner ends with a safepoint check, // nm could have been unloaded so look it up again. It's unsafe // to examine nm directly since it might have been freed and used // for something else. frame fr = thread->last_frame(); methodOop method = fr.interpreter_frame_method(); int bci = method->bci_from(fr.interpreter_frame_bcp()); nm = method->lookup_osr_nmethod_for(bci, CompLevel_none, false); } return nm; } IRT_ENTRY(nmethod*, InterpreterRuntime::frequency_counter_overflow_inner(JavaThread* thread, address branch_bcp)) // use UnlockFlagSaver to clear and restore the _do_not_unlock_if_synchronized // flag, in case this method triggers classloading which will call into Java. UnlockFlagSaver fs(thread); frame fr = thread->last_frame(); assert(fr.is_interpreted_frame(), "must come from interpreter"); methodHandle method(thread, fr.interpreter_frame_method()); const int branch_bci = branch_bcp != NULL ? method->bci_from(branch_bcp) : InvocationEntryBci; const int bci = branch_bcp != NULL ? method->bci_from(fr.interpreter_frame_bcp()) : InvocationEntryBci; nmethod* osr_nm = CompilationPolicy::policy()->event(method, method, branch_bci, bci, CompLevel_none, thread); if (osr_nm != NULL) { // We may need to do on-stack replacement which requires that no // monitors in the activation are biased because their // BasicObjectLocks will need to migrate during OSR. Force // unbiasing of all monitors in the activation now (even though // the OSR nmethod might be invalidated) because we don't have a // safepoint opportunity later once the migration begins. if (UseBiasedLocking) { ResourceMark rm; GrowableArray<Handle>* objects_to_revoke = new GrowableArray<Handle>(); for( BasicObjectLock *kptr = fr.interpreter_frame_monitor_end(); kptr < fr.interpreter_frame_monitor_begin(); kptr = fr.next_monitor_in_interpreter_frame(kptr) ) { if( kptr->obj() != NULL ) { objects_to_revoke->append(Handle(THREAD, kptr->obj())); } } BiasedLocking::revoke(objects_to_revoke); } } return osr_nm; IRT_END IRT_LEAF(jint, InterpreterRuntime::bcp_to_di(methodOopDesc* method, address cur_bcp)) assert(ProfileInterpreter, "must be profiling interpreter"); int bci = method->bci_from(cur_bcp); methodDataOop mdo = method->method_data(); if (mdo == NULL) return 0; return mdo->bci_to_di(bci); IRT_END IRT_ENTRY(void, InterpreterRuntime::profile_method(JavaThread* thread)) // use UnlockFlagSaver to clear and restore the _do_not_unlock_if_synchronized // flag, in case this method triggers classloading which will call into Java. UnlockFlagSaver fs(thread); assert(ProfileInterpreter, "must be profiling interpreter"); frame fr = thread->last_frame(); assert(fr.is_interpreted_frame(), "must come from interpreter"); methodHandle method(thread, fr.interpreter_frame_method()); methodOopDesc::build_interpreter_method_data(method, THREAD); if (HAS_PENDING_EXCEPTION) { assert((PENDING_EXCEPTION->is_a(SystemDictionary::OutOfMemoryError_klass())), "we expect only an OOM error here"); CLEAR_PENDING_EXCEPTION; // and fall through... } IRT_END #ifdef ASSERT IRT_LEAF(void, InterpreterRuntime::verify_mdp(methodOopDesc* method, address bcp, address mdp)) assert(ProfileInterpreter, "must be profiling interpreter"); methodDataOop mdo = method->method_data(); assert(mdo != NULL, "must not be null"); int bci = method->bci_from(bcp); address mdp2 = mdo->bci_to_dp(bci); if (mdp != mdp2) { ResourceMark rm; ResetNoHandleMark rnm; // In a LEAF entry. HandleMark hm; tty->print_cr("FAILED verify : actual mdp %p expected mdp %p @ bci %d", mdp, mdp2, bci); int current_di = mdo->dp_to_di(mdp); int expected_di = mdo->dp_to_di(mdp2); tty->print_cr(" actual di %d expected di %d", current_di, expected_di); int expected_approx_bci = mdo->data_at(expected_di)->bci(); int approx_bci = -1; if (current_di >= 0) { approx_bci = mdo->data_at(current_di)->bci(); } tty->print_cr(" actual bci is %d expected bci %d", approx_bci, expected_approx_bci); mdo->print_on(tty); method->print_codes(); } assert(mdp == mdp2, "wrong mdp"); IRT_END #endif // ASSERT IRT_ENTRY(void, InterpreterRuntime::update_mdp_for_ret(JavaThread* thread, int return_bci)) assert(ProfileInterpreter, "must be profiling interpreter"); ResourceMark rm(thread); HandleMark hm(thread); frame fr = thread->last_frame(); assert(fr.is_interpreted_frame(), "must come from interpreter"); methodDataHandle h_mdo(thread, fr.interpreter_frame_method()->method_data()); // Grab a lock to ensure atomic access to setting the return bci and // the displacement. This can block and GC, invalidating all naked oops. MutexLocker ml(RetData_lock); // ProfileData is essentially a wrapper around a derived oop, so we // need to take the lock before making any ProfileData structures. ProfileData* data = h_mdo->data_at(h_mdo->dp_to_di(fr.interpreter_frame_mdp())); RetData* rdata = data->as_RetData(); address new_mdp = rdata->fixup_ret(return_bci, h_mdo); fr.interpreter_frame_set_mdp(new_mdp); IRT_END IRT_ENTRY(void, InterpreterRuntime::at_safepoint(JavaThread* thread)) // We used to need an explict preserve_arguments here for invoke bytecodes. However, // stack traversal automatically takes care of preserving arguments for invoke, so // this is no longer needed. // IRT_END does an implicit safepoint check, hence we are guaranteed to block // if this is called during a safepoint if (JvmtiExport::should_post_single_step()) { // We are called during regular safepoints and when the VM is // single stepping. If any thread is marked for single stepping, // then we may have JVMTI work to do. JvmtiExport::at_single_stepping_point(thread, method(thread), bcp(thread)); } IRT_END IRT_ENTRY(void, InterpreterRuntime::post_field_access(JavaThread *thread, oopDesc* obj, ConstantPoolCacheEntry *cp_entry)) // check the access_flags for the field in the klass instanceKlass* ik = instanceKlass::cast((klassOop)cp_entry->f1()); typeArrayOop fields = ik->fields(); int index = cp_entry->field_index(); assert(index < fields->length(), "holders field index is out of range"); // bail out if field accesses are not watched if ((fields->ushort_at(index) & JVM_ACC_FIELD_ACCESS_WATCHED) == 0) return; switch(cp_entry->flag_state()) { case btos: // fall through case ctos: // fall through case stos: // fall through case itos: // fall through case ftos: // fall through case ltos: // fall through case dtos: // fall through case atos: break; default: ShouldNotReachHere(); return; } bool is_static = (obj == NULL); HandleMark hm(thread); Handle h_obj; if (!is_static) { // non-static field accessors have an object, but we need a handle h_obj = Handle(thread, obj); } instanceKlassHandle h_cp_entry_f1(thread, (klassOop)cp_entry->f1()); jfieldID fid = jfieldIDWorkaround::to_jfieldID(h_cp_entry_f1, cp_entry->f2(), is_static); JvmtiExport::post_field_access(thread, method(thread), bcp(thread), h_cp_entry_f1, h_obj, fid); IRT_END IRT_ENTRY(void, InterpreterRuntime::post_field_modification(JavaThread *thread, oopDesc* obj, ConstantPoolCacheEntry *cp_entry, jvalue *value)) klassOop k = (klassOop)cp_entry->f1(); // check the access_flags for the field in the klass instanceKlass* ik = instanceKlass::cast(k); typeArrayOop fields = ik->fields(); int index = cp_entry->field_index(); assert(index < fields->length(), "holders field index is out of range"); // bail out if field modifications are not watched if ((fields->ushort_at(index) & JVM_ACC_FIELD_MODIFICATION_WATCHED) == 0) return; char sig_type = '\0'; switch(cp_entry->flag_state()) { case btos: sig_type = 'Z'; break; case ctos: sig_type = 'C'; break; case stos: sig_type = 'S'; break; case itos: sig_type = 'I'; break; case ftos: sig_type = 'F'; break; case atos: sig_type = 'L'; break; case ltos: sig_type = 'J'; break; case dtos: sig_type = 'D'; break; default: ShouldNotReachHere(); return; } bool is_static = (obj == NULL); HandleMark hm(thread); instanceKlassHandle h_klass(thread, k); jfieldID fid = jfieldIDWorkaround::to_jfieldID(h_klass, cp_entry->f2(), is_static); jvalue fvalue; #ifdef _LP64 fvalue = *value; #else // Long/double values are stored unaligned and also noncontiguously with // tagged stacks. We can't just do a simple assignment even in the non- // J/D cases because a C++ compiler is allowed to assume that a jvalue is // 8-byte aligned, and interpreter stack slots are only 4-byte aligned. // We assume that the two halves of longs/doubles are stored in interpreter // stack slots in platform-endian order. jlong_accessor u; jint* newval = (jint*)value; u.words[0] = newval[0]; u.words[1] = newval[Interpreter::stackElementWords]; // skip if tag fvalue.j = u.long_value; #endif // _LP64 Handle h_obj; if (!is_static) { // non-static field accessors have an object, but we need a handle h_obj = Handle(thread, obj); } JvmtiExport::post_raw_field_modification(thread, method(thread), bcp(thread), h_klass, h_obj, fid, sig_type, &fvalue); IRT_END IRT_ENTRY(void, InterpreterRuntime::post_method_entry(JavaThread *thread)) JvmtiExport::post_method_entry(thread, InterpreterRuntime::method(thread), InterpreterRuntime::last_frame(thread)); IRT_END IRT_ENTRY(void, InterpreterRuntime::post_method_exit(JavaThread *thread)) JvmtiExport::post_method_exit(thread, InterpreterRuntime::method(thread), InterpreterRuntime::last_frame(thread)); IRT_END IRT_LEAF(int, InterpreterRuntime::interpreter_contains(address pc)) { return (Interpreter::contains(pc) ? 1 : 0); } IRT_END // Implementation of SignatureHandlerLibrary address SignatureHandlerLibrary::set_handler_blob() { BufferBlob* handler_blob = BufferBlob::create("native signature handlers", blob_size); if (handler_blob == NULL) { return NULL; } address handler = handler_blob->code_begin(); _handler_blob = handler_blob; _handler = handler; return handler; } void SignatureHandlerLibrary::initialize() { if (_fingerprints != NULL) { return; } if (set_handler_blob() == NULL) { vm_exit_out_of_memory(blob_size, "native signature handlers"); } BufferBlob* bb = BufferBlob::create("Signature Handler Temp Buffer", SignatureHandlerLibrary::buffer_size); _buffer = bb->code_begin(); _fingerprints = new(ResourceObj::C_HEAP)GrowableArray<uint64_t>(32, true); _handlers = new(ResourceObj::C_HEAP)GrowableArray<address>(32, true); } address SignatureHandlerLibrary::set_handler(CodeBuffer* buffer) { address handler = _handler; int insts_size = buffer->pure_insts_size(); if (handler + insts_size > _handler_blob->code_end()) { // get a new handler blob handler = set_handler_blob(); } if (handler != NULL) { memcpy(handler, buffer->insts_begin(), insts_size); pd_set_handler(handler); ICache::invalidate_range(handler, insts_size); _handler = handler + insts_size; } return handler; } void SignatureHandlerLibrary::add(methodHandle method) { if (method->signature_handler() == NULL) { // use slow signature handler if we can't do better int handler_index = -1; // check if we can use customized (fast) signature handler if (UseFastSignatureHandlers && method->size_of_parameters() <= Fingerprinter::max_size_of_parameters) { // use customized signature handler MutexLocker mu(SignatureHandlerLibrary_lock); // make sure data structure is initialized initialize(); // lookup method signature's fingerprint uint64_t fingerprint = Fingerprinter(method).fingerprint(); handler_index = _fingerprints->find(fingerprint); // create handler if necessary if (handler_index < 0) { ResourceMark rm; ptrdiff_t align_offset = (address) round_to((intptr_t)_buffer, CodeEntryAlignment) - (address)_buffer; CodeBuffer buffer((address)(_buffer + align_offset), SignatureHandlerLibrary::buffer_size - align_offset); InterpreterRuntime::SignatureHandlerGenerator(method, &buffer).generate(fingerprint); // copy into code heap address handler = set_handler(&buffer); if (handler == NULL) { // use slow signature handler } else { // debugging suppport if (PrintSignatureHandlers) { tty->cr(); tty->print_cr("argument handler #%d for: %s %s (fingerprint = " UINT64_FORMAT ", %d bytes generated)", _handlers->length(), (method->is_static() ? "static" : "receiver"), method->name_and_sig_as_C_string(), fingerprint, buffer.insts_size()); Disassembler::decode(handler, handler + buffer.insts_size()); #ifndef PRODUCT tty->print_cr(" --- associated result handler ---"); address rh_begin = Interpreter::result_handler(method()->result_type()); address rh_end = rh_begin; while (*(int*)rh_end != 0) { rh_end += sizeof(int); } Disassembler::decode(rh_begin, rh_end); #endif } // add handler to library _fingerprints->append(fingerprint); _handlers->append(handler); // set handler index assert(_fingerprints->length() == _handlers->length(), "sanity check"); handler_index = _fingerprints->length() - 1; } } } else { CHECK_UNHANDLED_OOPS_ONLY(Thread::current()->clear_unhandled_oops()); } if (handler_index < 0) { // use generic signature handler method->set_signature_handler(Interpreter::slow_signature_handler()); } else { // set handler method->set_signature_handler(_handlers->at(handler_index)); } } #ifdef ASSERT int handler_index, fingerprint_index; { // '_handlers' and '_fingerprints' are 'GrowableArray's and are NOT synchronized // in any way if accessed from multiple threads. To avoid races with another // thread which may change the arrays in the above, mutex protected block, we // have to protect this read access here with the same mutex as well! MutexLocker mu(SignatureHandlerLibrary_lock); handler_index = _handlers->find(method->signature_handler()); fingerprint_index = _fingerprints->find(Fingerprinter(method).fingerprint()); } assert(method->signature_handler() == Interpreter::slow_signature_handler() || handler_index == fingerprint_index, "sanity check"); #endif } BufferBlob* SignatureHandlerLibrary::_handler_blob = NULL; address SignatureHandlerLibrary::_handler = NULL; GrowableArray<uint64_t>* SignatureHandlerLibrary::_fingerprints = NULL; GrowableArray<address>* SignatureHandlerLibrary::_handlers = NULL; address SignatureHandlerLibrary::_buffer = NULL; IRT_ENTRY(void, InterpreterRuntime::prepare_native_call(JavaThread* thread, methodOopDesc* method)) methodHandle m(thread, method); assert(m->is_native(), "sanity check"); // lookup native function entry point if it doesn't exist bool in_base_library; if (!m->has_native_function()) { NativeLookup::lookup(m, in_base_library, CHECK); } // make sure signature handler is installed SignatureHandlerLibrary::add(m); // The interpreter entry point checks the signature handler first, // before trying to fetch the native entry point and klass mirror. // We must set the signature handler last, so that multiple processors // preparing the same method will be sure to see non-null entry & mirror. IRT_END #if defined(IA32) || defined(AMD64) IRT_LEAF(void, InterpreterRuntime::popframe_move_outgoing_args(JavaThread* thread, void* src_address, void* dest_address)) if (src_address == dest_address) { return; } ResetNoHandleMark rnm; // In a LEAF entry. HandleMark hm; ResourceMark rm; frame fr = thread->last_frame(); assert(fr.is_interpreted_frame(), ""); jint bci = fr.interpreter_frame_bci(); methodHandle mh(thread, fr.interpreter_frame_method()); Bytecode_invoke* invoke = Bytecode_invoke_at(mh, bci); ArgumentSizeComputer asc(invoke->signature()); int size_of_arguments = (asc.size() + (invoke->has_receiver() ? 1 : 0)); // receiver Copy::conjoint_jbytes(src_address, dest_address, size_of_arguments * Interpreter::stackElementSize); IRT_END #endif