Messages in this thread |  | | Date | Thu, 12 Feb 2015 09:48:52 -0600 | From | Josh Poimboeuf <> | Subject | Re: [RFC PATCH 6/9] livepatch: create per-task consistency model |
| |
On Thu, Feb 12, 2015 at 02:26:42PM +0100, Jiri Slaby wrote: > On 02/12/2015, 04:21 AM, Josh Poimboeuf wrote: > > Ingo, Peter, > > > > Would you have any objections to making task_rq_lock/unlock() non-static > > (or moving them to kernel/sched/sched.h) so they can be called by the > > livepatch code? > > > > To provide some background, I'm looking for a way to temporarily prevent > > a sleeping task from running while its stack is examined, to decide > > whether it can be safely switched to the new patching "universe". For > > more details see klp_transition_task() in the patch below. > > > > Using task_rq_lock() is the most straightforward way I could find to > > achieve that. > > Hi, I cannot speak whether it is the proper way or not. > > But if so, would it make sense to do the opposite: expose an API to walk > through the processes' stack and make the decision? Concretely, move > parts of klp_stacktrace_address_verify_func to sched.c or somewhere in > kernel/sched/ and leave task_rq_lock untouched.
Yeah, it makes sense in theory. But I'm not sure how to do that in a way that prevents races when switching the task's universe. I think we need the rq locked for both the stack walk and the universe switch.
In general, I agree it would be good to find a way to keep the rq locking functions in sched.c.
-- Josh
|  |