JDK-8292260 : [BACKOUT] JDK-8279219: [REDO] C2 crash when allocating array of size too large
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: 11.0.16,17,18,19,20
  • Priority: P1
  • Status: Resolved
  • Resolution: Fixed
  • OS: generic
  • CPU: generic
  • Submitted: 2022-08-11
  • Updated: 2022-08-22
  • Resolved: 2022-08-12
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
JDK 11 JDK 17 JDK 18 JDK 19 JDK 20 JDK 8
11.0.16.1-oracleFixed 17.0.4.1-oracleFixed 18.0.2.1Fixed 19 b36Fixed 20Fixed 8u345-perfFixed
Related Reports
Duplicate :  
JDK-8291665 - C2: assert compiling SSLEngineInputRecord::decodeInputRecord
Duplicate :  
JDK-8288184 - C2 compilation asserts with "Bad graph detected in compute_lca_of_uses"
Relates :  
JDK-8291919 - [11u] High compiler native memory consumption after JDK-8279219
Relates :  
JDK-8284369 - TestFailedAllocationBadGraph fails with -XX:TieredStopAtLevel < 4
Relates :  
JDK-8279219 - [REDO] C2 crash when allocating array of size too large
Relates :  
JDK-8292301 - [REDO v2] C2 crash when allocating array of size too large
Sub Tasks
JDK-8292396 :  
Release Note: C2 Compilation Errors Unpredictably Crashes JVM - Closed
Description
JDK-8279219 caused regressions in the OpenJDK 11.0.16 and OpenJDK 17.0.4 releases and we are backing it out. See JDK-8291665.

Mailing list discussion:
https://mail.openjdk.org/pipermail/jdk-updates-dev/2022-August/016469.html
Comments
Changeset: 967a28c3 Author: Vladimir Kozlov <kvn@openjdk.org> Date: 2022-08-12 20:24:53 +0000 URL: https://git.openjdk.org/jdk19/commit/967a28c3d85fdde6d5eb48aa0edd8f7597772469
12-08-2022
JDK 19 fix request approved.
12-08-2022
This backout also backs out fix JDK-8284369 in the removed test compiler/allocation/TestFailedAllocationBadGraph.java
12-08-2022
I've created JDK-8292301 for the redo v2 work.
12-08-2022
Fix request for JDK 19. Regressions have been observed with the 11u, 17u, 18, 19, 20. See JDK-8291665 and JDK-8288184. There might be more code-shapes out in the wild which trigger similar issues. Therefore we'd like to back out the patch of JDK-8279219. Backout is not clean because of additional changes in graphKit.cpp and test TestFailedAllocationBadGraph.java. I verified changes with test case from JDK-8288184. Backout passed tier1-5 mach5 testing.
12-08-2022
A pull request was submitted for review. URL: https://git.openjdk.org/jdk19/pull/163 Date: 2022-08-12 13:25:11 +0000
12-08-2022
For the record, this backout will restore the issue listed in the original bug report JDK-8278413 and followed JDK-8279219. We will need new fix (REDO) later.
12-08-2022
Thank you, Severin
12-08-2022
Fix Request (critical for OpenJDK 11.0.16.1 and OpenJDK 17.0.4.1): Regressions have been observed with the 11.0.16 release. See JDK-8291665. While there is a work-around to disable c2 compilation for specific methods, there might be more code-shapes out in the wild which trigger JDK-8291665. Therefore we'd like to back out the patch of JDK-8279219 of the 11.0.16 and 17.0.4 releases and do a patch release of the result. Risk is obviously that the fix of JDK-8279219 would no longer be there and potential other issues caused by the absence of JDK-8279219 (that we don't know yet). Yet it gets us back to where 11.0.15 and 17.0.3 were. It seems the safest option forward. Testing still in progress.
12-08-2022
A pull request was submitted for review. URL: https://git.openjdk.org/jdk17u/pull/350 Date: 2022-08-12 09:20:53 +0000
12-08-2022
A pull request was submitted for review. URL: https://git.openjdk.org/jdk11u/pull/49 Date: 2022-08-12 08:54:03 +0000
12-08-2022
I'm creating the OpenJDK 11.0.16 and OpenJDK 17.0.4 specific revert PRs meanwhile and change to the proper backport sha once it's pushed. This should speed up the process for people to test.
12-08-2022
OK.
12-08-2022
I am taking over this backout to speedup process for JDK 19.
12-08-2022
JDK-8288184 is also caused by JDK-8291919. It reproduced with all releases.
12-08-2022
Are JDK 19 and 20 affected too?
12-08-2022
All affected releases should be listed in "Affected versions". The back out should be done in latest sources (JDK 20) and then backported to all releases where JDK-8279219 was pushed. Sources in all releases should be consistent.
12-08-2022
ILW = same as JDK-8291665 = P3
11-08-2022
Patch will be the anti-delta of JDK-8279219.
11-08-2022