JDK-8248845 : AArch64: stack corruption after spilling vector register
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: 11,15,16,repo-panama
  • Priority: P3
  • Status: Resolved
  • Resolution: Fixed
  • OS: linux
  • CPU: aarch64
  • Submitted: 2020-07-06
  • Updated: 2020-11-19
  • Resolved: 2020-07-08
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 15 JDK 16
11.0.10Fixed 15 b31Fixed 16Fixed
Related Reports
Relates :  
JDK-8076276 - Add support for AVX512
Description
Seen on the vectorIntrinsics branch but should also be possible on jdk/jdk although I haven't found a reliable reproducer. Use the attached TestFloatVector.java and the following command line:

java --add-modules jdk.incubator.vector -XX:-TieredCompilation -Xcomp  -XX:-UseOnStackReplacement TestFloatVector

Note the crash will only happen on vectorIntrinsics *before* the recent commit "Enable call devirtualization during post-parse phase" as this changes the inlining and hides the bug.

#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x0000ffff8d7448cc, pid=225338, tid=225339
#
# JRE version: OpenJDK Runtime Environment (16.0) (fastdebug build 16-internal+0-adhoc.nicgas01.panama-vector)
# Java VM: OpenJDK 64-Bit Server VM (fastdebug 16-internal+0-adhoc.nicgas01.panama-vector, compiled mode, compressed oops, g1 gc, linux-aarch64)
# Problematic frame:
# J 3622 c2 TestFloatVector.fmaTest([F[F[F[F[Z)V (111 bytes) @ 0x0000ffff8d7448cc [0x0000ffff8d744400+0x00000000000004cc]
#
Comments
11u Fix: Patch did not apply cleanly, review approval is here: https://mail.openjdk.java.net/pipermail/jdk-updates-dev/2020-September/003858.html
05-10-2020
From the mailing list discussion, this was caused by JDK-8076276.
23-09-2020
Requesting backport to 11u as the bug is also present there and could cause hard to debug crashes. Patch applies cleanly.
09-07-2020
URL: https://hg.openjdk.java.net/jdk/jdk15/rev/d5be95758352 User: ngasson Date: 2020-07-08 09:22:47 +0000
08-07-2020
Review thread: https://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/2020-July/038886.html
07-07-2020