I have tested only with linux 4.10.0-24-generic on amd64. The attached repro case:
- compiles a java class
- compiles a C program that initialise the JVM and calls a static method from the java class
- launch the compile C program.
Expected result:
the c program should launch the Java method and print some numbers in the terminal
Actual result:
the c program segfault while initializing the JVM
reverting to linux 4.10.0-22-generic amd64 fixes the problem. Since the only change in the changelog is the Stack Clash fix and since I remember reading that the JVM does tricky stuff with the stack, this is a natural culprit.
It touches any program that initialises a JVM (so for instance Java bindings in Postgresql or Java applets in firefox although the latter have been deprecated and are only supported in firefox ESR now).
I attach the repro-case.
The issue happens with both openjdk-8-jdk from the repository and the standard jdk 8 from oracle.
I have tested only with linux 4.10.0-24-generic on amd64. The attached repro case:
- compiles a java class
- compiles a C program that initialise the JVM and calls a static method from the java class
- launch the compile C program.
Expected result:
the c program should launch the Java method and print some numbers in the terminal
Actual result:
the c program segfault while initializing the JVM
reverting to linux 4.10.0-22-generic amd64 fixes the problem. Since the only change in the changelog is the Stack Clash fix and since I remember reading that the JVM does tricky stuff with the stack, this is a natural culprit.
It touches any program that initialises a JVM (so for instance Java bindings in Postgresql or Java applets in firefox although the latter have been deprecated and are only supported in firefox ESR now).
I attach the repro-case.
The issue happens with both openjdk-8-jdk from the repository and the standard jdk 8 from oracle.