Re: [PATCH] bpf: don't fail kmalloc while releasing raw_tp

[Mathieu Desnoyers <mathieu.desnoyers@efficios.com>]

----- On Nov 16, 2020, at 5:10 PM, rostedt rostedt@goodmis.org wrote:

> On Mon, 16 Nov 2020 16:34:41 -0500 (EST)
> Mathieu Desnoyers <mathieu.desnoyers@efficios.com> wrote:

[...]

>> I think you'll want a WRITE_ONCE(old[i].func, tp_stub_func) here, matched
>> with a READ_ONCE() in __DO_TRACE. This introduces a new situation where the
>> func pointer can be updated and loaded concurrently.
> 
> I thought about this a little, and then only thing we really should worry
> about is synchronizing with those that unregister. Because when we make
> this update, there are now two states. the __DO_TRACE either reads the
> original func or the stub. And either should be OK to call.
> 
> Only the func gets updated and not the data. So what exactly are we worried
> about here?

Indeed with a stub function, I don't see any need for READ_ONCE/WRITE_ONCE.

However, if we want to compare the function pointer to some other value and
conditionally do (or skip) the call, I think you'll need the READ_ONCE/WRITE_ONCE
to make sure the pointer is not re-fetched between comparison and call.

Thanks,

Mathieu

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com