On 08/16/2016 10:27 AM, christophe leroy wrote: > If I debug a very small app, it gets stuck quickly after the app has > stopped: indeed, the console seems ok but as soon as I try to execute > something simple, like a ps or top, it get stuck. The target still > responds to pings, but nothing else.
This one is a pretty common symptom when the kernel crashes holding a per-process lock of some kind, probably most commonly mmap_sem. Do you have a serial console? Can you do a sysrq-t on it to get stack dumps of the hung processes and ps? It's time to turn on all the debugging options we can find. Can you build a kernel with all this stuff enabled (=y on all of these): CONFIG_SLUB_DEBUG CONFIG_DEBUG_VM CONFIG_DEBUG_PAGEALLOC CONFIG_PAGE_POISONING CONFIG_SLUB_DEBUG_ON CONFIG_KASAN CONFIG_*LOCKUP_DETECTOR CONFIG_DETECT_HUNG_TASK CONFIG_SCHED_DEBUG CONFIG_DEBUG_PREEMPT CONFIG_MAGIC_SYSRQ And all of the spinlock/lockdep stuff enabled too: CONFIG_DEBUG_SPINLOCK=y CONFIG_DEBUG_MUTEXES=y CONFIG_DEBUG_LOCK_ALLOC=y CONFIG_PROVE_LOCKING=y CONFIG_LOCKDEP=y CONFIG_DEBUG_ATOMIC_SLEEP=y CONFIG_STACKTRACE=y CONFIG_DEBUG_LIST=y Let's hope that these turn something up.
