8316746: Top of lock-stack does not match the unlocked object #16406

TheRealMDoerr · 2023-10-27T19:26:54Z

It took me quite a long journey to figure out that C2 OSR goes wrong only in the test "vmTestbase/nsk/jdi/StepEvent" because the interpreter fills the slots in a different order as expected by C2. (Interpreter and C1 don't care about it.)
I've reimplemented the search loop in monitorenter, improved comments and cleaned up the related code a bit. The test is passing with this change.

Progress

Change must be properly reviewed (1 review required, with at least 1 Reviewer)
Change must not contain extraneous whitespace
Commit message must refer to an issue

Issue

JDK-8316746: Top of lock-stack does not match the unlocked object (Bug - P2)

Reviewers

Richard Reingruber (@reinrich - Reviewer) ⚠️ Review applies to 5011761f
Lutz Schmidt (@RealLucy - Reviewer) ⚠️ Review applies to 56209f27

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/16406/head:pull/16406
$ git checkout pull/16406

Update a local copy of the PR:
$ git checkout pull/16406
$ git pull https://git.openjdk.org/jdk.git pull/16406/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 16406

View PR using the GUI difftool:
$ git pr show -t 16406

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/16406.diff

Webrev

Link to Webrev Comment

bridgekeeper · 2023-10-27T19:27:49Z

👋 Welcome back mdoerr! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

openjdk · 2023-10-27T19:29:26Z

@TheRealMDoerr The following label will be automatically applied to this pull request:

hotspot-compiler

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing list. If you would like to change these labels, use the /label pull request command.

mlbridge · 2023-10-27T19:32:38Z

Webrevs

dean-long · 2023-10-27T22:32:51Z

Mystery solved!

…age.

TheRealMDoerr · 2023-10-29T16:39:59Z

Mystery solved!

Yeah! Thanks again for your help! I hope I didn't steal too much of your time and that it was at least interesting for everybody.

TobiHartmann · 2023-10-31T08:26:59Z

/label add hotspot-runtime

openjdk · 2023-10-31T08:28:49Z

@TobiHartmann
The hotspot-runtime label was successfully added.

reinrich

Hi Martin,

the order of BasicObjectLocks in the interpreter is specified if the locking operations are well-formed.

the comment on the PPC implementation of frame::interpreter_frame_monitor_end is wrong. It should be removed.

TheRealMDoerr · 2023-11-02T16:03:47Z

You're right! Removed the comment from all platforms which still had it.

reinrich · 2023-11-02T16:06:22Z

Thanks. I wasn't even aware that other implementations have it as well :)

reinrich

The code in this PR is a lot easier to follow than the baseline version and it looks correct to me.
I tried to understand what's actually wrong with the baseline but gave up. Do you know where precisely the bug is?

Anyway, thanks for fixing this!
Richard.

reinrich · 2023-11-02T15:40:15Z

src/hotspot/cpu/ppc/templateTable_ppc_64.cpp

-  ConditionRegister found_free_slot = CCR0,
-                    found_same_obj  = CCR1,
-                    reached_limit   = CCR6;
+  // Note: The order of the monitors is important for C2 OSR which derives the


You should refer to the decl. of the interpreter_frame_monitor_* where the expected order is specified.

openjdk · 2023-11-02T16:44:31Z

@TheRealMDoerr This change now passes all automated pre-integration checks.

ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details.

After integration, the commit message for the final commit will be:

8316746: Top of lock-stack does not match the unlocked object

Reviewed-by: rrich, lucy

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 136 new commits pushed to the master branch:

7d25f1c: 8319423: Improve Year.isLeap by checking divisibility by 16
59e9981: 8319376: ParallelGC: Forwarded objects found during heap inspection
7c7f8ea: 8319456: jdk/jfr/event/gc/collection/TestGCCauseWith[Serial|Parallel].java : GC cause 'GCLocker Initiated GC' not in the valid causes
7bc8e4c: 8294980: test/jdk/java/lang/invoke 15 test classes use experimental bytecode library
e841897: 8319374: JFR: Remove instrumentation for exception events
cd9719b: 8319306: Serial: Remove TenuredSpace::verify
1e687b4: 8316719: C2 compilation still fails with "bad AD file"
8555e0f: 8319318: bufferedStream fixed case can be removed
73c5f60: 8319556: Harmonize interface formatting in the FFM API
cc4b0d9: 8319378: Spec for j.util.Timer::purge and j.util.Timer::cancel could be improved
... and 126 more: https://git.openjdk.org/jdk/compare/2915d74a10c63cdca22d0055ebde67ef316a341b...master

As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

➡️ To integrate this PR with the above commit message to the master branch, type /integrate in a new comment.

TheRealMDoerr · 2023-11-02T16:51:40Z

Thanks for the review! The old code simply reuses the first empty slot it finds (from top to bottom). That causes the order to be reversed if e.g. 2 slots are free and we lock 2 objects. The new code uses the free slot which is closer to the bottom first.

TheRealMDoerr · 2023-11-02T22:04:02Z

@TobiHartmann: You were also interested in a test. I have included one which reproduces the issue on PPC64 without this fix.

RealLucy

Changes look good to me.
I added one minor optimization suggestion you may want to consider. It is not relevant for approval.

RealLucy · 2023-11-08T11:09:29Z

src/hotspot/cpu/ppc/templateTable_ppc_64.cpp

-  __ addi(Rcurrent_obj_addr, Rcurrent_obj_addr, - frame::interpreter_frame_monitor_size_in_bytes());
+  __ cmpdi(CCR0, Rfree_slot, 0);
+  __ beq(CCR0, Lallocate_new);
+  __ mr(Rcurrent_monitor, Rfree_slot);
  __ b(Lfound);


Wouldn't it be nice to use Rfree_slot from here on and save the move to Rcurrent_monitor and the unconditional branch?

__ cmpdi(CCR0, Rfree_slot, 0); __ bne(CCR0, Lfound);

Good idea. Changed.

RealLucy · 2023-11-08T11:10:52Z

src/hotspot/cpu/ppc/templateTable_ppc_64.cpp

-  __ addi(Rcurrent_obj_addr, Rcurrent_obj_addr, - frame::interpreter_frame_monitor_size_in_bytes());
+  __ cmpdi(CCR0, Rfree_slot, 0);
+  __ beq(CCR0, Lallocate_new);
+  __ mr(Rcurrent_monitor, Rfree_slot);
  __ b(Lfound);

  // We didn't find a free BasicObjLock => allocate one.
  __ align(32, 12);
  __ bind(Lallocate_new);
  __ add_monitor_to_stack(false, Rscratch1, Rscratch2);
  __ mr(Rcurrent_monitor, R26_monitor);


If you do the above optimization, this instruction would need to change to
__ mr(Rfree_slot, R26_monitor);

RealLucy · 2023-11-08T11:14:14Z

src/hotspot/cpu/ppc/templateTable_ppc_64.cpp

@@ -4241,7 +4232,7 @@ void TemplateTable::monitorenter() {
  // The object has already been popped from the stack, so the expression stack looks correct.
  __ addi(R14_bcp, R14_bcp, 1);

-  __ std(Robj_to_lock, 0, Rcurrent_obj_addr);
+  __ std(Robj_to_lock, in_bytes(BasicObjectLock::obj_offset()), Rcurrent_monitor);
  __ lock_object(Rcurrent_monitor, Robj_to_lock);


Rcurrent_monitor would need to be replaced by Rfree_slot in the above two lines as well.

TheRealMDoerr · 2023-11-08T16:47:09Z

Thanks for the review! I'm rerunning tests and will probably integrate tomorrow if everything is fine and there are no further comments.

TheRealMDoerr · 2023-11-09T10:12:22Z

Tests are green.
/integrate

openjdk · 2023-11-09T10:14:06Z

Going to push as commit 7d8adfa.
Since your change was applied there have been 147 commits pushed to the master branch:

dd9eab1: 8310886: C2 SuperWord: Two nodes should be isomorphic if they are loop invariant but pinned at different nodes outside the loop
7e4cb2f: 8318962: Update ProcessTools javadoc with suggestions in 8315097
0b58bde: 8319104: GtestWrapper crashes with SIGILL in AsyncLogTest::test_asynclog_raw on AIX opt
b0fc808: 8319725: G1: Subtracting virtual time from wall time after JDK-8319204
a26f7c0: 6928542: Chinese characters in RTF are not decoded
4e8c036: 8319665: [macOS] Obsolete imports of <Carbon/Carbon.h> in java.desktop
f4ab952: 8319253: [BACKOUT] Change LockingMode default from LM_LEGACY to LM_LIGHTWEIGHT
c137b17: 8187655: jdk.lambda.vm.InterfaceAccessFlagsTest.testPrivateMethodCall needs update after nestmates support
fa331d7: 8319184: RISC-V: improve MD5 intrinsic
4c0f6c6: 8311961: Update Manual Test Groups for ATR JDK22
... and 137 more: https://git.openjdk.org/jdk/compare/2915d74a10c63cdca22d0055ebde67ef316a341b...master

Your commit was automatically rebased without conflicts.

openjdk · 2023-11-09T10:14:13Z

@TheRealMDoerr Pushed as commit 7d8adfa.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.

8316746: Top of lock-stack does not match the unlocked object

c2d3668

openjdk bot added the rfr Pull request is ready for review label Oct 27, 2023

openjdk bot added the hotspot-compiler hotspot-compiler-dev@openjdk.org label Oct 27, 2023

Handle recursive locking like on other platforms. Improve register us…

1e678fd

…age.

Add regression test.

5011761

openjdk bot added the hotspot-runtime hotspot-runtime-dev@openjdk.org label Oct 31, 2023

reinrich reviewed Nov 2, 2023

View reviewed changes

Remove wrong comment.

d5977ed

reinrich approved these changes Nov 2, 2023

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Nov 2, 2023

Improve comment.

56209f2

RealLucy approved these changes Nov 8, 2023

View reviewed changes

Shorter code: Use Rfree_slot as monitor.

0c1a934

openjdk bot added the integrated Pull request has been integrated label Nov 9, 2023

openjdk bot closed this Nov 9, 2023

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Nov 9, 2023

TheRealMDoerr deleted the 8316746_lock_stack_V3 branch November 9, 2023 10:14

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

8316746: Top of lock-stack does not match the unlocked object #16406

8316746: Top of lock-stack does not match the unlocked object #16406

TheRealMDoerr commented Oct 27, 2023 •

edited by openjdk bot

Loading

bridgekeeper bot commented Oct 27, 2023

openjdk bot commented Oct 27, 2023

mlbridge bot commented Oct 27, 2023 •

edited

Loading

dean-long commented Oct 27, 2023

TheRealMDoerr commented Oct 29, 2023

TobiHartmann commented Oct 31, 2023

openjdk bot commented Oct 31, 2023

reinrich left a comment

TheRealMDoerr commented Nov 2, 2023

reinrich commented Nov 2, 2023

reinrich left a comment

reinrich Nov 2, 2023

TheRealMDoerr Nov 2, 2023

reinrich Nov 2, 2023

openjdk bot commented Nov 2, 2023 •

edited

Loading

TheRealMDoerr commented Nov 2, 2023

TheRealMDoerr commented Nov 2, 2023

RealLucy left a comment

RealLucy Nov 8, 2023

TheRealMDoerr Nov 8, 2023

RealLucy Nov 8, 2023

RealLucy Nov 8, 2023

TheRealMDoerr commented Nov 8, 2023

TheRealMDoerr commented Nov 9, 2023

openjdk bot commented Nov 9, 2023

openjdk bot commented Nov 9, 2023

8316746: Top of lock-stack does not match the unlocked object #16406

8316746: Top of lock-stack does not match the unlocked object #16406

Conversation

TheRealMDoerr commented Oct 27, 2023 • edited by openjdk bot Loading

Progress

Issue

Reviewers

Reviewing

Webrev

bridgekeeper bot commented Oct 27, 2023

openjdk bot commented Oct 27, 2023

mlbridge bot commented Oct 27, 2023 • edited Loading

Webrevs

dean-long commented Oct 27, 2023

TheRealMDoerr commented Oct 29, 2023

TobiHartmann commented Oct 31, 2023

openjdk bot commented Oct 31, 2023

reinrich left a comment

Choose a reason for hiding this comment

TheRealMDoerr commented Nov 2, 2023

reinrich commented Nov 2, 2023

reinrich left a comment

Choose a reason for hiding this comment

reinrich Nov 2, 2023

Choose a reason for hiding this comment

TheRealMDoerr Nov 2, 2023

Choose a reason for hiding this comment

reinrich Nov 2, 2023

Choose a reason for hiding this comment

openjdk bot commented Nov 2, 2023 • edited Loading

TheRealMDoerr commented Nov 2, 2023

TheRealMDoerr commented Nov 2, 2023

RealLucy left a comment

Choose a reason for hiding this comment

RealLucy Nov 8, 2023

Choose a reason for hiding this comment

TheRealMDoerr Nov 8, 2023

Choose a reason for hiding this comment

RealLucy Nov 8, 2023

Choose a reason for hiding this comment

RealLucy Nov 8, 2023

Choose a reason for hiding this comment

TheRealMDoerr commented Nov 8, 2023

TheRealMDoerr commented Nov 9, 2023

openjdk bot commented Nov 9, 2023

openjdk bot commented Nov 9, 2023

TheRealMDoerr commented Oct 27, 2023 •

edited by openjdk bot

Loading

mlbridge bot commented Oct 27, 2023 •

edited

Loading

openjdk bot commented Nov 2, 2023 •

edited

Loading