Skip to content

[GR-63694] Thread termination note in isolate tear down. #10965

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

[GR-63694] Thread termination note in isolate tear down. #10965

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 18 additions & 0 deletions espresso/src/com.oracle.truffle.espresso.mokapot/include/graal_isolate_dynamic.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -127,6 +127,15 @@ typedef int (*graal_detach_thread_fn_t)(graal_isolatethread_t* thread);
* waiting for any attached threads to detach from it, then discards its objects,
* threads, and any other state or context that is associated with it.
* Returns 0 on success, or a non-zero value on failure.
*
*
* If this call blocks indefinitely, this means there are still Java threads running
* that do not terminate after receiving a Thread.interrupt() event. To prevent indefinite blocking,
* these threads should be cooperatively terminated within Java before invoking this call.
* To diagnose such issues, use the option '-R:TearDownWarningSeconds=<secs>' at image build time
* to detect the threads that are still running. This will print the stack traces of all threads that block tear-down.
* To resolve blocking issues, utilize any available method to terminate the offending threads. This may include
* calling shutdown API functions, adjusting the application configuration, or leveraging reflection.
*/
typedef int (*graal_tear_down_isolate_fn_t)(graal_isolatethread_t* isolateThread);

Expand All @@ -141,6 +150,15 @@ typedef int (*graal_tear_down_isolate_fn_t)(graal_isolatethread_t* isolateThread
* Java code at the time when this function is called or at any point in the future
* or this will cause entirely undefined (and likely fatal) behavior.
* Returns 0 on success, or a non-zero value on (non-fatal) failure.
*
*
* If this call blocks indefinitely, this means there are still Java threads running
* that do not terminate after receiving a Thread.interrupt() event. To prevent indefinite blocking,
* these threads should be cooperatively terminated within Java before invoking this call.
* To diagnose such issues, use the option '-R:TearDownWarningSeconds=<secs>' at image build time
* to detect the threads that are still running. This will print the stack traces of all threads that block tear-down.
* To resolve blocking issues, utilize any available method to terminate the offending threads. This may include
* calling shutdown API functions, adjusting the application configuration, or leveraging reflection.
*/
typedef int (*graal_detach_all_threads_and_tear_down_isolate_fn_t)(graal_isolatethread_t* isolateThread);

Expand Down
39 changes: 33 additions & 6 deletions substratevm/src/com.oracle.svm.core/src/com/oracle/svm/core/SubstrateOptions.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,6 +24,8 @@
*/
package com.oracle.svm.core;

import static com.oracle.svm.core.SubstrateOptions.DeprecatedOptions.TearDownFailureNanos;
import static com.oracle.svm.core.SubstrateOptions.DeprecatedOptions.TearDownWarningNanos;
import static com.oracle.svm.core.option.RuntimeOptionKey.RuntimeOptionKeyFlag.Immutable;
import static com.oracle.svm.core.option.RuntimeOptionKey.RuntimeOptionKeyFlag.IsolateCreationOnly;
import static com.oracle.svm.core.option.RuntimeOptionKey.RuntimeOptionKeyFlag.RelevantForCompilationIsolates;
Expand Down Expand Up @@ -63,6 +65,7 @@
import com.oracle.svm.core.option.RuntimeOptionKey;
import com.oracle.svm.core.option.SubstrateOptionsParser;
import com.oracle.svm.core.thread.VMOperationControl;
import com.oracle.svm.core.util.TimeUtils;
import com.oracle.svm.core.util.UserError;
import com.oracle.svm.util.LogUtils;
import com.oracle.svm.util.ModuleSupport;
Expand Down Expand Up @@ -576,6 +579,13 @@ protected void onValueUpdate(EconomicMap<OptionKey<?>, Object> values, Boolean o
}
}
};

@Option(help = "The number of nanoseconds before and between which tearing down an isolate gives a warning message. 0 implies no warning.", deprecated = true, deprecationMessage = "Use TearDownWarningSeconds instead")//
public static final RuntimeOptionKey<Long> TearDownWarningNanos = new RuntimeOptionKey<>(0L, RelevantForCompilationIsolates);

@Option(help = "The number of nanoseconds before tearing down an isolate gives a failure message. 0 implies no message.", deprecated = true, deprecationMessage = "Use TearDownFailureSeconds instead")//
public static final RuntimeOptionKey<Long> TearDownFailureNanos = new RuntimeOptionKey<>(0L, RelevantForCompilationIsolates);

}

@Option(help = "Enable detection and runtime container configuration support.")//
Expand Down Expand Up @@ -818,18 +828,35 @@ private static void validateZapNativeMemory(HostedOptionKey<Boolean> optionKey)
* Isolate tear down options.
*/

@Option(help = "The number of nanoseconds before and between which tearing down an isolate gives a warning message. 0 implies no warning.")//
public static final RuntimeOptionKey<Long> TearDownWarningNanos = new RuntimeOptionKey<>(0L, RelevantForCompilationIsolates);
@Option(help = "The number of seconds before and between which tearing down an isolate gives a warning message. 0 implies no warning.")//
public static final RuntimeOptionKey<Long> TearDownWarningSeconds = new RuntimeOptionKey<>(0L, key -> validateTearDownTimeoutOptions(), RelevantForCompilationIsolates);

@Option(help = "The number of nanoseconds before tearing down an isolate gives a failure message. 0 implies no message.")//
public static final RuntimeOptionKey<Long> TearDownFailureNanos = new RuntimeOptionKey<>(0L, RelevantForCompilationIsolates);
@Option(help = "The number of seconds before tearing down an isolate gives a failure message. 0 implies no message.")//
public static final RuntimeOptionKey<Long> TearDownFailureSeconds = new RuntimeOptionKey<>(0L, key -> validateTearDownTimeoutOptions(), RelevantForCompilationIsolates);

private static void validateTearDownTimeoutOptions() {
if (TearDownWarningSeconds.hasBeenSet() && TearDownWarningNanos.hasBeenSet()) {
throw new IllegalArgumentException("Can't set both TearDownWarningSeconds and TearDownWarningNanos at the same time. Use TearDownWarningSeconds.");
}
if (TearDownFailureSeconds.hasBeenSet() && TearDownFailureNanos.hasBeenSet()) {
throw new IllegalArgumentException("Can't set both TearDownFailureSeconds and TearDownFailureNanos at the same time. Use TearDownFailureSeconds.");
}
}

public static long getTearDownWarningNanos() {
return TearDownWarningNanos.getValue();
validateTearDownTimeoutOptions();
if (TearDownWarningNanos.hasBeenSet()) {
return TearDownWarningNanos.getValue();
}
return TearDownWarningSeconds.getValue() * TimeUtils.nanosPerSecond;
}

public static long getTearDownFailureNanos() {
return TearDownFailureNanos.getValue();
validateTearDownTimeoutOptions();
if (TearDownFailureNanos.hasBeenSet()) {
return TearDownFailureNanos.getValue();
}
return TearDownFailureSeconds.getValue() * TimeUtils.nanosPerSecond;
}

@Option(help = "Define the maximum number of stores for which the loop that zeroes out objects is unrolled.")//
Expand Down
52 changes: 35 additions & 17 deletions ...rc/com.oracle.svm.core/src/com/oracle/svm/core/c/function/CEntryPointNativeFunctions.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,6 @@

import java.util.function.Function;

import jdk.graal.compiler.word.Word;
import org.graalvm.nativeimage.CurrentIsolate;
import org.graalvm.nativeimage.Isolate;
import org.graalvm.nativeimage.IsolateThread;
Expand All @@ -41,6 +40,8 @@
import com.oracle.svm.core.c.function.CEntryPointOptions.NoPrologue;
import com.oracle.svm.core.thread.VMThreads;

import jdk.graal.compiler.word.Word;

@CHeader(value = GraalIsolateHeader.class)
public final class CEntryPointNativeFunctions {

Expand Down Expand Up @@ -158,12 +159,26 @@ public static int detachThread(IsolateThread thread) {
return CEntryPointActions.leaveDetachThread();
}

static final String THREAD_TERMINATION_NOTE = """

If this call blocks indefinitely, this means there are still Java threads running
that do not terminate after receiving a Thread.interrupt() event. To prevent indefinite blocking,
these threads should be cooperatively terminated within Java before invoking this call.
To diagnose such issues, use the option '-R:TearDownWarningSeconds=<secs>' at image build time
to detect the threads that are still running. This will print the stack traces of all threads that block tear-down.
To resolve blocking issues, utilize any available method to terminate the offending threads. This may include
calling shutdown API functions, adjusting the application configuration, or leveraging reflection.
""";

@Uninterruptible(reason = UNINTERRUPTIBLE_REASON)
@CEntryPoint(name = "tear_down_isolate", documentation = {
"Tears down the isolate of the passed (and still attached) isolate thread,",
"waiting for any attached threads to detach from it, then discards its objects,",
"threads, and any other state or context that is associated with it.",
"Returns 0 on success, or a non-zero value on failure."})
@CEntryPoint(name = "tear_down_isolate", documentation = {"""
Tears down the isolate of the passed (and still attached) isolate thread,
waiting for any attached threads to detach from it, then discards its objects,
threads, and any other state or context that is associated with it.
Returns 0 on success, or a non-zero value on failure.
""",
THREAD_TERMINATION_NOTE,
})
@CEntryPointOptions(prologue = NoPrologue.class, epilogue = NoEpilogue.class, nameTransformation = NameTransformation.class)
public static int tearDownIsolate(IsolateThread isolateThread) {
int result = CEntryPointActions.enter(isolateThread);
Expand All @@ -174,17 +189,20 @@ public static int tearDownIsolate(IsolateThread isolateThread) {
}

@Uninterruptible(reason = UNINTERRUPTIBLE_REASON)
@CEntryPoint(name = "detach_all_threads_and_tear_down_isolate", documentation = {
"In the isolate of the passed isolate thread, detach all those threads that were",
"externally started (not within Java, which includes the \"main thread\") and were",
"attached to the isolate afterwards. Afterwards, all threads that were started",
"within Java undergo a regular shutdown process, followed by the tear-down of the",
"entire isolate, which detaches the current thread and discards the objects,",
"threads, and any other state or context associated with the isolate.",
"None of the manually attached threads targeted by this function may be executing",
"Java code at the time when this function is called or at any point in the future",
"or this will cause entirely undefined (and likely fatal) behavior.",
"Returns 0 on success, or a non-zero value on (non-fatal) failure."})
@CEntryPoint(name = "detach_all_threads_and_tear_down_isolate", documentation = {"""
In the isolate of the passed isolate thread, detach all those threads that were
externally started (not within Java, which includes the "main thread") and were
attached to the isolate afterwards. Afterwards, all threads that were started
within Java undergo a regular shutdown process, followed by the tear-down of the
entire isolate, which detaches the current thread and discards the objects,
threads, and any other state or context associated with the isolate.
None of the manually attached threads targeted by this function may be executing
Java code at the time when this function is called or at any point in the future
or this will cause entirely undefined (and likely fatal) behavior.
Returns 0 on success, or a non-zero value on (non-fatal) failure.
""",
THREAD_TERMINATION_NOTE,
})
@CEntryPointOptions(prologue = NoPrologue.class, epilogue = NoEpilogue.class, nameTransformation = NameTransformation.class)
public static int detachAllThreadsAndTearDownIsolate(IsolateThread isolateThread) {
int result = CEntryPointActions.enter(isolateThread);
Expand Down
Loading