Symptoms
- DisplayManager (Owned by client thread) doesn't give up Framebuffer references before the server thread destroys the machine
- Recent RVVM JNI bindings cause the framebuffer to be recycled immediately
- Crash on access to recycled framebuffer VMA from Java side via a still-living ByteBuffer
Solutions
Improve RVVM JNI framebuffer state recycling logic (Instead of freeing on RVVM side, implement cleaner callback in Direct ByteBuffer handed off to Java code?)
@Kamillaova do you perhaps have a good idea how to implement direct ByteBuffer cleanup callback into JNI C side?
Symptoms
Solutions
Improve RVVM JNI framebuffer state recycling logic (Instead of freeing on RVVM side, implement
cleanercallback in Direct ByteBuffer handed off to Java code?)@Kamillaova do you perhaps have a good idea how to implement direct ByteBuffer cleanup callback into JNI C side?