Skip to main content
5 events
when toggle format what by license comment
Oct 31, 2024 at 0:35 comment added waltinator Interrupt mode saves the interrupted task's context. When interrupt mode gets interrupted, where would the context (along with the extra "interrupt mode" context) be stored? How would the system return after handling the interrupt's fault? Return to the faulty code? Return to the calling user? Where would it find the user's context?
Oct 26, 2024 at 20:52 comment added Gilles 'SO- stop being evil' That explains why faults in kernel mode should cause a panic. But the question here is, given that most faults in kernel mode actually do not cause a panic in Linux, why do faults in interrupt contexts still cause a panic?
Oct 26, 2024 at 20:01 comment added Elisa K. K. It doesn't answer the question. I know that the system cannot be relied on after a crash. But if that crash happens in non-interrupt context (STILL, IN KERNEL!!!) then my system is not halted: only Oops is triggered. And if the crash happens IN INTERRUPT, the system is ALWAYS HALTED IMMEDIATELY. This design was the topic of the question. NOT the fact that a crashed system can't be relied on.
Oct 26, 2024 at 18:40 history edited ilkkachu CC BY-SA 4.0
I think we can do without jabs like that
Oct 26, 2024 at 18:08 history answered waltinator CC BY-SA 4.0