Re: [RFC][Patch 1/4] kprobe fast unregistration

[Masami Hiramatsu <masami dot hiramatsu dot pt at hitachi dot com>]

Hi Christoph and Anil,

Thank you for your comments.

Christoph Hellwig wrote:
> Speeding up the unregistration is a very good idea, but this interface
> is rather horrible.  It's almost a receipe for users to get it wrong.
Keshavamurthy, Anil S wrote:
> I agree with Christop that the interface is horrible and error prone.

OK, I agree. I had chosen a confusable name.

> However, I see the use case where people want to disable the probes quickly and
> would like to reenable them again. Looking closely at your patch,
> I think this can be acheived.

Thank you.

> Here is my suggestion.
> 
>> Here is an example code.
>> --
>> struct kprobes *p;
>> for_each_probe(p) {
>> 	unregister_kprobe_fast(p);
>         ^^^^^^^^^^^^^^^^^^^^^^
> Change this to disable_kprobe(p), which is essentially the same as
> what you have implemented. And also provide an opposite function
> to reenable_kprobe(p) which enables the disabled probe again.

I'd like to change that to prepare_to_unregister_kprobe(p) instead of
disable_kprobe(p).

I think Josh and other people want interfaces to disable/reenable all
probes at once when the sysrq is pressed.
So, IMHO, these interfaces should use a global (and per-cpu?) flag which
controls whether kprobes calls user-defined handler or not, instead of
self-modifying.
For example,

if (p && p->pre_handler && kprobe_enable) {
                           ^^^^^^^^^^^^^
   ...
}

But, I think this would be another story.
I'd like to discuss this topic in other mails.

>> }
>> commit_kprobes();
>   ^^^^^^^^^^^^^^
> Change this to unregister_disabled_kprobes(), which essentially 
> unregisters all the disabled probes.

And also, I'd like to change it to unregister_prepared_kprobes().
What would you think about this idea?

Thanks,


> 
> Thanks,
> Anil Keshavamurthy
> 
-- 
Masami HIRAMATSU
Linux Technology Center
Hitachi, Ltd., Systems Development Laboratory
E-mail: masami.hiramatsu.pt@hitachi.com