[lttng-dev] [PATCH lttng-ust] Add ctor/dtor priorities for tracepoints/events

[Olivier Dion]

On Mon, 13 Jul 2020, Mathieu Desnoyers <mathieu.desnoyers at efficios.com> wrote:
> ----- On Jul 12, 2020, at 11:49 AM, Olivier Dion olivier.dion at polymtl.ca wrote:
>
>> On Sun, 12 Jul 2020, Mathieu Desnoyers <mathieu.desnoyers at efficios.com> wrote:
>>> ----- On Jul 11, 2020, at 11:29 AM, lttng-dev lttng-dev at lists.lttng.org wrote:
>>>
>>>> Some library might want to generate events in their ctor/dtor.  If
>>>> LTTng initialize/finalize its tracepoints/events at the wrong time,
>>>> events are lost.
>>>> 
>>>> Order of execution of the ctor/dtor is determined by priority.  When
>>>> some priorities are equal, the order of execution seems to be
>>>> determined by:
>>>> 
>>>>	   a) Order of appearance if in the same compilation unit
>>>> 
>>>>	   b) Order of link if in different compilation units
>>>> 
>>>>	   c) Order of load by ld-linux.so or dlopen(3) for
>>>>	      share objects
>>>
>>> I recall different rules about constructor priorities. Can you provide
>>> links to documentation stating the priority order you describe above ?
>> 
>> I haven't found any documentation on that.  This is purely empirical.
>> Although I'm sure that we can dig something if chatting on GCC's IRC.
>
> If it is not documented, then I am reluctant on depending on a behavior
> which may be what happens today, but may not be the same for past/future
> toolchains.

Agree.

>>> Also, we should compare two approaches to fulfill your goal:
>>> one alternative would be to have application/library constructors
>>> explicitly call tracepoint constructors if they wish to use them.
>> 
>> I would prefer this way.  The former solution might not work in some
>> cases (e.g. with LD_PRELOAD and priority =101) and I prefer explicit
>> initialization in that case.
>> 
>> I don't see any cons for the second approach, except making the symbols
>> table a few bytes larger.  I'll post a patch soon so we can compare and
>> try to find more documentation on ctor priority.
>
> And users will have to explicitly call the constructor on which they
> depend, but I don't see it as a huge burden.

The burden is small indeed.  But users should pay close attention to
release the references in a destructor too.

> Beware though that there are a few configurations which can be used for
> probe providers (see lttng-ust(3)).

I'm not following you here.  I don't see any configuration for provider
except TRACEPOINT_LOGLEVEL.  What should I be aware of?

-- 
Olivier Dion
PolyMtl