JDK-8211387 : [Zero] atomic_copy64: Use ldrexd for atomic reads on ARMv7
  • Type: Bug
  • Component: hotspot
  • Sub-Component: runtime
  • Affected Version: 8u172,12
  • Priority: P4
  • Status: Resolved
  • Resolution: Fixed
  • OS: linux
  • Submitted: 2018-10-02
  • Updated: 2020-09-22
  • Resolved: 2018-10-09
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 12 JDK 8
11.0.2Fixed 12 b15Fixed 8u202Fixed
Related Reports
Relates :  
JDK-8253464 - ARM32 Zero: atomic_copy64 is incorrect, breaking volatile stores
Description
The Zero routine for atomic_copy64 should use the same as is done for JDK 7. JDK 7 Zero maintained at IcedTea uses ldrexd for atomic reads on ARMv7:

https://icedtea.classpath.org/hg/icedtea7-forest/hotspot/rev/b0a871360c6b

The Zero port in latest JDK should use the same.
Comments
Fix Request: Requesting downport approval for JDK 11u so as to keep all active Zero versions in sync. The intention is to get this into JDK 8u where there is no ARM 32 JIT in upstream 8u. That's where this fix makes the most sense. If the fix is in 8u, but not in JDK 11u, then the JDK regresses on update. Risk is minimal as this is a Zero-only change which AFAIK isn't being built by Oracle. The JDK 12 changeset imports into JDK 11u as is.
09-10-2018
review-thread: http://mail.openjdk.java.net/pipermail/hotspot-dev/2018-October/034512.html
04-10-2018
webrev: http://cr.openjdk.java.net/~sgehwolf/webrevs/JDK-8211387/webrev.01/
02-10-2018