[jira] Closed: (HARMONY-5342) [drlvm][jvmti] Test org.apache.harmony.test.stress.jpda.jdwp.scenario.EVENT014.EventTest014 crashes on x86_64

["Gregory Shimansky (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HARMONY-5342?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Gregory Shimansky closed HARMONY-5342.
--------------------------------------

    Resolution: Fixed

Patch to fix this bug is committed at 606790.

The change makes parsing of REX prefixes more flexible. All possible REX combinations should now be allowed including those not mentioned in instruction opcodes of encoder tables.

Also patch fixes a bug when a return is made from "noreturn" function VMBreakPoints::process_native_breakpoint. It cannot return because its return address is undefined in the stack. Instead it should transfer control back to original register context.


> [drlvm][jvmti] Test org.apache.harmony.test.stress.jpda.jdwp.scenario.EVENT014.EventTest014 crashes on x86_64
> -------------------------------------------------------------------------------------------------------------
>
>                 Key: HARMONY-5342
>                 URL: https://issues.apache.org/jira/browse/HARMONY-5342
>             Project: Harmony
>          Issue Type: Bug
>          Components: DRLVM
>         Environment: All x86_64
>            Reporter: Gregory Shimansky
>            Assignee: Gregory Shimansky
>
> This test crashes both on Linux and Windows on x86_64. Command line to run it is:
> <path to Java>/java -Djpda.settings.debuggeeLaunchKind=auto -Dorg.apache.harmony.test.share.stress.ReliabilityRunner.timeToWork=30 -Dorg.apache.harmony.test.share.stress.ReliabilityRunner.timeToAbort=30 -Djpda.settings.timeout=600000 -cp bin/classes:bin/lib:bin/tools/th.jar:../classlib/trunk/depends/jars/junit_3.8.2/junit.jar org.apache.harmony.test.stress.jpda.jdwp.scenario.EVENT014.EventTest014
> in tests/stress/qa from buildtest root.
> I think that the bug is in wrong decoding of some x86_64 instructions by the disassembler. It incorrectly calculates their length and this leads to a crash.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.