Re: [PATCH v5 tip 0/7] tracing: attach eBPF programs to kprobes

[Alexei Starovoitov <ast@plumgrid.com>]

On 3/6/15 5:09 PM, Steven Rostedt wrote:
> On Wed, 4 Mar 2015 15:48:24 -0500
> Steven Rostedt <rostedt@goodmis.org> wrote:
>
>> On Wed, 4 Mar 2015 21:33:16 +0100
>> Ingo Molnar <mingo@kernel.org> wrote:
>>
>>>
>>> * Alexei Starovoitov <ast@plumgrid.com> wrote:
>>>
>>>> On Sun, Mar 1, 2015 at 3:27 PM, Alexei Starovoitov <ast@plumgrid.com> wrote:
>>>>> Peter, Steven,
>>>>> I think this set addresses everything we've discussed.
>>>>> Please review/ack. Thanks!
>>>>
>>>> icmp echo request
>>>
>>> I'd really like to have an Acked-by from Steve (propagated into the
>>> changelogs) before looking at applying these patches.
>>
>> I'll have to look at this tomorrow. I'm a bit swamped with other things
>> at the moment :-/
>>
>
> Just an update. I started looking at it but then was pulled off to do
> other things. I'll make this a priority next week. Sorry for the delay.

There is no rush. Please let me know if I need to clarify anything.
One thing I just caught which I'm planning to address in the follow on
patch is missing 'recursion check'. Since attaching programs to kprobes
means that root may create loops by adding a kprobe somewhere in
the call chain invoked from bpf program. So far I'm thinking to do
simple stack_trace_call()-like check. I don't think it's a blocker
for this set, but if I'm done coding recursion soon, I'll just
roll it in and respin this set :)