| 10 | | That aspect of it doesn't appear to be the case over at least, the kernel stack matches up with the frame entries for me. One point of note btw, at the point where we hit the "final" panic (which, for me at least, is consistently a General Protection Fault at the moment), there are three threads in the team: 1) the main app thread iwhich s perfectly backtraceable and looks sane for a crashed thread (it's waiting in a condition variable spawned from thread_hit_debug_event_internal -> read_port_etc). 2) The pthread function thread which is no longer traceable, but which up until the point where it started getting corrupted was executing the aforementioned vfs operations, and finally a debug task thread is also attached, also waiting on a condition var from debug_nub_thread(). Since only the pthread appears corrupted, and their stacks are, at least here, quite far apart, it seems improbable that one of the operations the other thread triggered would've been responsible for the corruption, and more likely that the thread corrupted its own. |
| | 10 | That aspect of it doesn't appear to be the case over at least, the kernel stack matches up with the frame entries for me. One point of note btw, at the point where we hit the "final" panic (which, for me at least, is consistently a General Protection Fault at the moment), there are three threads in the team: 1) the main app thread which is perfectly backtraceable and looks sane for a crashed thread (it's waiting in a condition variable spawned from thread_hit_debug_event_internal -> read_port_etc). 2) The pthread function thread which is no longer traceable, but which up until the point where it started getting corrupted was executing the aforementioned vfs operations, and finally a debug task thread is also attached, also waiting on a condition var from debug_nub_thread(). Since only the pthread appears corrupted, and their stacks are, at least here, quite far apart, it seems improbable that one of the operations the other thread triggered would've been responsible for the corruption, and more likely that the thread corrupted its own. |
