----- On Nov 16, 2020, at 12:19 PM, rostedt rostedt@xxxxxxxxxxx wrote:
> On Sat, 14 Nov 2020 21:52:55 -0800
> Matt Mullins <mmullins@xxxxxxx> wrote:
>
>> bpf_link_free is always called in process context, including from a
>> workqueue and from __fput. Neither of these have the ability to
>> propagate an -ENOMEM to the caller.
>>
>
> Hmm, I think the real fix is to not have unregistering a tracepoint probe
> fail because of allocation. We are removing a probe, perhaps we could just
> inject NULL pointer that gets checked via the DO_TRACE loop?
>
> I bet failing an unregister because of an allocation failure causes
> problems elsewhere than just BPF.
>
> Mathieu,
>
> Can't we do something that would still allow to unregister a probe even if
> a new probe array fails to allocate? We could kick off a irq work to try to
> clean up the probe at a later time, but still, the unregister itself should
> not fail due to memory failure.
Currently, the fast path iteration looks like:
struct tracepoint_func *it_func_ptr;
void *it_func;
it_func_ptr = \
rcu_dereference_raw((&__tracepoint_##_name)->funcs); \
do { \
it_func = (it_func_ptr)->func; \
__data = (it_func_ptr)->data; \
((void(*)(void *, proto))(it_func))(__data, args); \
} while ((++it_func_ptr)->func);
So we RCU dereference the array, and iterate on the array until we find a NULL
func. So you could not use NULL to skip items, but you could perhaps reserve
a (void *)0x1UL tombstone for this.
It should ideally be an unlikely branch, and it would be good to benchmark the
change when multiple tracing probes are attached to figure out whether the
overhead is significant when tracing is enabled.
I wonder whether we really mind that much about using slightly more memory
than required after a failed reallocation due to ENOMEM. Perhaps the irq work
is not even needed. Chances are that the irq work would fail again and again if
it's in low memory conditions. So maybe it's better to just keep the tombstone
in place until the next successful callback array reallocation.
Thoughts ?
Thanks,
Mathieu
--
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com
