[23.0] Sync with upstream #633

zakkak · 2023-12-12T13:37:01Z

Closes #618

I went through the changes and don't see anything concerning.

(cherry picked from commit e27665f)

Some cleanups and additional sanity checks. (cherry picked from commit 6c0aa30)

(cherry picked from commit 1159814)

(cherry picked from commit 6bd08da)

(cherry picked from commit 6b05065)

(cherry picked from commit 65f4a3e)

…sue. PullRequest: graal/15912

… darwin-aarch64. PullRequest: graal/15910

…els. PullRequest: graal/15873

Disable the stripping of debug symbols from the generated native bits Closes oracle#6995 (cherry picked from commit 0162b22)

…f /json/list path. (GR-49664) (cherry picked from commit de48c84)

…ot reach here. PullRequest: graal/15886

…ead() in JFR.

… mx argument PullRequest: graal/15928

…8.2. PullRequest: js/2970

…XBean.

(cherry picked from commit 129fa26)

… called from a host call. (cherry picked from commit 0b14c1e) Adapt fix to 23.0

… entered context during close called from a host call PullRequest: graal/16015

(cherry picked from commit 514f292)

(cherry picked from commit 12e9a8b)

PullRequest: graal/15943

Some versions of `ld64` cannot deal with DIRECT8 relocations in the `.text` section. Approximately this is since version 820.1 (Xcode 14). Starting with Xcode15 beta3 the default linker has been replaced with "the new linker" (version 902.11) which does not suffer from this bug anymore. However, Xcode also ships a `ld-classic` which reassembles "the old linker" and is still affected by this bug (and it of course prints the same version number). See some more in-depth analysis of this ld64 bug in https://openradar.appspot.com/FB11942354 The workaround: Instead of emitting the address of a HostedMethod inlined, it will be put in the data section and thus requires a memory load to obtain it. (cherry picked from commit e30698b) resolve conflicts resolve errors backport fixup

(cherry picked from commit 6f02719) fix error add back method isAssociative() remove unused imports fix formatting issues

…nt range. PullRequest: graal/16151

PullRequest: graal/16231

…sThanNode PullRequest: graal/16170

…ge is queried via MemoryPoolMXBean. PullRequest: graal/16008

(cherry picked from commit 993a429)

…_info. PullRequest: graal/16227

PullRequest: graal/16264

…he header size. PullRequest: graal/16103

zakkak · 2023-12-12T14:00:07Z

Looking at build failure

zakkak · 2023-12-12T14:16:17Z

Looking at build failure

This was due to some conflicts with #575

Requested @roberttoyonaga's review in the corresponding lines.

zakkak · 2023-12-12T15:04:30Z

substratevm/src/com.oracle.svm.core/src/com/oracle/svm/core/jfr/JfrEvent.java

@@ -101,6 +102,6 @@ public boolean shouldEmit(long durationTicks) {

    @Uninterruptible(reason = "Prevent races with VM operations that start/stop recording.", callerMustBe = true)
    private boolean shouldEmit0() {
-        return SubstrateJVM.get().isRecording() && SubstrateJVM.get().isEnabled(this) && !SubstrateJVM.get().isCurrentThreadExcluded();
+        return SubstrateJVM.get().isRecording() && SubstrateJVM.get().isEnabled(this) && !JfrThreadLocal.isThreadExcluded(JavaThreads.getCurrentThreadOrNull());


@roberttoyonaga can you please check if this is correct?

Yes this is correct. But https://github.com/graalvm/mandrel/pull/633/files/f3864b7fbb9d325ae604fdb0ffab0e71506bf636#diff-5140a589f5128cadb659640373aafca28c66b801fef8c975d65c08fb9eded219R105 !SubstrateJVM.get().isCurrentThreadExcluded(); needs to be removed because now the checking is done in JfrThreadLocal

oh it looks like you already removed it

I remember there was some weird quirk that may have required isThreadExcluded to be checked before isRecording and isEnabled. Let me quickly investigate to be sure.

Ok this should be correct for 23.0.

jerboaa · 2023-12-12T16:16:46Z

The mandrel IT failure seems quarkusio/quarkus#37633

zakkak · 2023-12-12T22:25:39Z

The mandrel IT failure seems quarkusio/quarkus#37633

Indeed, the test no longer fails now that quarkusio/quarkus#37633 is merged, however with 23.0 we are getting:

com.oracle.svm.core.util.UserError$UserException: Incompatible change of initialization policy for io.netty.buffer.PooledByteBufAllocator: trying to change RUN_TIME from 'META-INF/native-image/io.netty/netty-buffer/native-image.properties' in 'file:///home/zakkak/code/mandrel-integration-tests/apps/jfr-native-image-performance/target/jfr-plaintext-native-image-source-jar/lib/io.netty.netty-buffer-4.1.100.Final.jar' with 'io.netty.buffer.PooledByteBufAllocator' and from feature io.quarkus.runner.Feature.beforeAnalysis with 'PooledByteBufAllocator.class' to RERUN Quarkus
	at org.graalvm.nativeimage.builder/com.oracle.svm.core.util.UserError.abort(UserError.java:73)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.classinitialization.ClassInitializationConfiguration.insertRec(ClassInitializationConfiguration.java:103)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.classinitialization.ClassInitializationConfiguration.insertRec(ClassInitializationConfiguration.java:117)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.classinitialization.ClassInitializationConfiguration.insertRec(ClassInitializationConfiguration.java:117)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.classinitialization.ClassInitializationConfiguration.insertRec(ClassInitializationConfiguration.java:117)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.classinitialization.ClassInitializationConfiguration.insertRec(ClassInitializationConfiguration.java:117)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.classinitialization.ClassInitializationConfiguration.insert(ClassInitializationConfiguration.java:64)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.classinitialization.ProvenSafeClassInitializationSupport.rerunInitialization(ProvenSafeClassInitializationSupport.java:162)
	at io.quarkus.runner.Feature.runtimeReinitializedClasses(Unknown Source)
	at io.quarkus.runner.Feature.beforeAnalysis(Unknown Source)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.NativeImageGenerator.lambda$runPointsToAnalysis$9(NativeImageGenerator.java:757)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.FeatureHandler.forEachFeature(FeatureHandler.java:89)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.NativeImageGenerator.runPointsToAnalysis(NativeImageGenerator.java:757)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.NativeImageGenerator.doRun(NativeImageGenerator.java:582)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.NativeImageGenerator.run(NativeImageGenerator.java:539)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.NativeImageGeneratorRunner.buildImage(NativeImageGeneratorRunner.java:408)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.NativeImageGeneratorRunner.build(NativeImageGeneratorRunner.java:612)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.NativeImageGeneratorRunner.start(NativeImageGeneratorRunner.java:134)
	at org.graalvm.nativeimage.builder/com.oracle.svm.hosted.NativeImageGeneratorRunner.main(NativeImageGeneratorRunner.java:94)

which is not the case with 23.1 and 24.0-dev (probably related to the new class initialization strategy adopted in 23.1 https://github.com/oracle/graal/blob/master/substratevm/CHANGELOG.md#graalvm-for-jdk-21-internal-version-2310)

We probably need to switch the corresponding netty classes to runtime initialize to make them compatible with 23.0. I will explore this following days.

zakkak · 2023-12-13T15:10:54Z

We probably need to switch the corresponding netty classes to runtime initialize to make them compatible with 23.0. I will explore this following days.

Fix in quarkusio/quarkus#37707

jerboaa

OK. Tests seem alright.

jerboaa · 2023-12-13T15:30:14Z

It would be good for @roberttoyonaga to approve the jfr changes.

roberttoyonaga

It would be good for @roberttoyonaga to approve the jfr changes.

It looks good to me. I tested it out on my machine as well. I just neglected to actually click approve.

christianhaeubl and others added 30 commits October 16, 2023 10:04

Add stack walking heuristic to Native Image crash log.

6f58dcb

[GS-5266] New dev cycle GraalVM 23.0.3

de1fe97

canon: expose dead phi cycle detection as feature that can be disabled

7434dd0

(cherry picked from commit e27665f)

Use some minimal alignment for unaligned heap chunks.

4ccf57e

Some cleanups and additional sanity checks. (cherry picked from commit 6c0aa30)

Adapt fix to version 23.0

9929dac

[GR-47839] Use fixed OS version for GraalVM builds on darwin-aarch64

ca1fc1b

(cherry picked from commit 1159814)

Workaround CPU features Xcode 15 miscomp issue (GR-49301)

4b3b001

(cherry picked from commit 6bd08da)

svm: fix placement of the NO_INLINE definition in cpuid.c

705bbf7

(cherry picked from commit 6b05065)

svm: add comment about the reason for NO_INLINE

aef9a21

(cherry picked from commit 65f4a3e)

[GR-49504] Backport 23.0: Workaround CPU features Xcode 15 miscomp is…

eecfaff

…sue. PullRequest: graal/15912

[GR-49649] Backport 23.0 : Use fixed OS version for GraalVM builds on…

f57a072

… darwin-aarch64. PullRequest: graal/15910

[GR-49508] [GR-49654] Backport to 23.0 : Refactor NI Optimization Lev…

5cecedb

…els. PullRequest: graal/15873

Introduce --disable-debuginfo-stripping mx argument

cdb0389

Disable the stripping of debug symbols from the generated native bits Closes oracle#6995 (cherry picked from commit 0162b22)

Add the mandatory experimental property 'debuggerId' and handle GET o…

ae28c57

…f /json/list path. (GR-49664) (cherry picked from commit de48c84)

[GR-48695] Backport to 23.0: Uncommitting of memory failed - should n…

d0322f8

…ot reach here. PullRequest: graal/15886

Adapt fix to 23.0

02b4f1e

Use JavaThreads.getCurrentThreadOrNull() instead of Thread.currentThr…

f16ae26

…ead() in JFR.

[GR-49726] Backport to 23.0 : Introduce --disable-debuginfo-stripping…

b66155c

… mx argument PullRequest: graal/15928

[GR-49798] Backport: Upgrading the underlying Node.js to version 18.1…

408429e

…8.2. PullRequest: js/2970

Report consistent values when memory usage is queried via MemoryPoolM…

d8c7571

…XBean.

Don't zap pre-zeroed unaligned chunks.

853f7fb

(cherry picked from commit 129fa26)

Prevent automatic leave of an explicitly entered context during close…

283d0f8

… called from a host call. (cherry picked from commit 0b14c1e) Adapt fix to 23.0

[GR-49914] Backport to 23.0: Prevent automatic leave of an explicitly…

e006df4

… entered context during close called from a host call PullRequest: graal/16015

Fix crash during isolate teardown.

70ea07c

(cherry picked from commit 514f292)

Set IsolateThread to null after teardown.

9f43ac7

(cherry picked from commit 12e9a8b)

Upgrade org.json to version 20231013. (GR-50058)

1b723bb

[GR-49794] Merge tag jdk-17.0.10+4.

20b47ba

[GR-49789] Backport to 23.0: Don't zap pre-zeroed unaligned chunks.

b24d7ba

PullRequest: graal/15943

Clarify semantics of binary arithmetic reassociation

6d973dd

(cherry picked from commit 6f02719) fix error add back method isAssociative() remove unused imports fix formatting issues

OracleLabsAutomation and others added 9 commits November 30, 2023 15:58

update to jvmci-23.0-b25

4903592

[GR-49946] Backport to 23.0: Loop unroll: handle stride overflow in i…

95bd3e3

…nt range. PullRequest: graal/16151

[GR-49909] Backport to 23.0: Allow UnlockExperimentalVMOptions.

e6a28d3

PullRequest: graal/16231

[GR-50041] Backport to 23.0: Check stamp before casting in IntegerLes…

2503381

…sThanNode PullRequest: graal/16170

[GR-49856] Backport to 23.0: Report consistent values when memory usa…

de57dd5

…ge is queried via MemoryPoolMXBean. PullRequest: graal/16008

Add an option to artificially increase the header size.

2abe819

(cherry picked from commit 993a429)

[GR-50356] Backport to 23.0: Use the vendor version for __svm_version…

217305b

…_info. PullRequest: graal/16227

[GR-50416] Update JVMCI to jvmci-23.0-b25

ed0e1de

PullRequest: graal/16264

[GR-49879] Backport to 23.0: Add an option to artificially increase t…

7edc294

…he header size. PullRequest: graal/16103

zakkak added affects/JDK17 JDK17 related affects/23.0 labels Dec 12, 2023

oracle-contributor-agreement bot added the OCA Verified All contributors have signed the Oracle Contributor Agreement. label Dec 12, 2023

zakkak requested a review from jerboaa December 12, 2023 13:37

zakkak linked an issue Dec 12, 2023 that may be closed by this pull request

[23.0] Backport: Synchronize accesses to reachability handlers #618

Closed

2 tasks

zakkak force-pushed the 2023-12-12-sync-23.0 branch from 15c098c to f3864b7 Compare December 12, 2023 14:15

zakkak added this to the 23.0.3.0-Final milestone Dec 12, 2023

Merge branch 'oracle/release/graal-vm/23.0' into 'mandrel/23.0'

7fbf185

zakkak force-pushed the 2023-12-12-sync-23.0 branch from f3864b7 to 7fbf185 Compare December 12, 2023 15:03

zakkak commented Dec 12, 2023

View reviewed changes

jerboaa approved these changes Dec 13, 2023

View reviewed changes

roberttoyonaga approved these changes Dec 13, 2023

View reviewed changes

zakkak merged commit a55af47 into graalvm:mandrel/23.0 Dec 13, 2023
66 of 67 checks passed

zakkak deleted the 2023-12-12-sync-23.0 branch December 13, 2023 15:51

jerboaa mentioned this pull request Dec 15, 2023

[23.0] Backport: Synchronize accesses to reachability handlers #618

Closed

2 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[23.0] Sync with upstream #633

[23.0] Sync with upstream #633

zakkak commented Dec 12, 2023

zakkak commented Dec 12, 2023

zakkak commented Dec 12, 2023 •

edited

Loading

zakkak Dec 12, 2023

roberttoyonaga Dec 12, 2023

roberttoyonaga Dec 12, 2023

roberttoyonaga Dec 12, 2023

roberttoyonaga Dec 12, 2023

jerboaa commented Dec 12, 2023

zakkak commented Dec 12, 2023

zakkak commented Dec 13, 2023

jerboaa left a comment

jerboaa commented Dec 13, 2023

roberttoyonaga left a comment

[23.0] Sync with upstream #633

[23.0] Sync with upstream #633

Conversation

zakkak commented Dec 12, 2023

zakkak commented Dec 12, 2023

zakkak commented Dec 12, 2023 • edited Loading

zakkak Dec 12, 2023

Choose a reason for hiding this comment

roberttoyonaga Dec 12, 2023

Choose a reason for hiding this comment

roberttoyonaga Dec 12, 2023

Choose a reason for hiding this comment

roberttoyonaga Dec 12, 2023

Choose a reason for hiding this comment

roberttoyonaga Dec 12, 2023

Choose a reason for hiding this comment

jerboaa commented Dec 12, 2023

zakkak commented Dec 12, 2023

zakkak commented Dec 13, 2023

jerboaa left a comment

Choose a reason for hiding this comment

jerboaa commented Dec 13, 2023

roberttoyonaga left a comment

Choose a reason for hiding this comment

zakkak commented Dec 12, 2023 •

edited

Loading