nf_unregister_hook problem

nf_unregister_hook problem

[gmoliveira]

hi all,

im trying to develop a software to emulate network partitioning to  
evaluate distributed applications. in the code i need to specify some  
temporal triggers like:

init_timer(&foo);
start_trigger.function = some_function;
start_trigger.data = arguments;
start_trigger.expires = jiffies + msecs_to_jiffies(HZ * 1000);
add_timer(&foo);

im using it to register a netfilter hook and it works perfectly.  
however, when i need to set up a trigger do unregister the netfilter  
hook, the system gets itself slow, dmesg returns lots of information  
about BUG SCHEDULING and stuff like that.

im using this:

init_timer(&foo2);
stop_trigger.function = nf_unregister_hook;
stop_trigger.data = &nf_hook;
stop_trigger.expires = jiffies + msecs_to_jiffies(HZ * 5000);
add_timer(&foo2);

it works to every function until now, unless nf_unregister_hook.

best regards,
Gustavo.

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

Re: nf_unregister_hook problem

[Jan Engelhardt]

On Tuesday 2009-06-23 20:47, gmoliveira@inf.ufrgs.br wrote:
>
>im using this:
>
>init_timer(&foo2);
>stop_trigger.function = nf_unregister_hook;
>stop_trigger.data = &nf_hook;
>stop_trigger.expires = jiffies + msecs_to_jiffies(HZ * 5000);
>add_timer(&foo2);
>
>it works to every function until now, unless nf_unregister_hook.

Does the problem go away if you call it indirectly? Like:


stop_trigger.function = myfunc;
...

myfunc(void)
{
	nf_unregister_hook(&nf_hook);
}

Re: nf_unregister_hook problem

[Bruno Fiss]

--- Em ter, 23/6/09, Jan Engelhardt <jengelh@medozas.de> escreveu:


>
> Does the problem go away if you call it indirectly? Like:
>
>
> stop_trigger.function = myfunc;
> ...
>
> myfunc(void)
> {
>     nf_unregister_hook(&nf_hook);
> }
>


Hi,

I've got a very similar question, about scheduling nf_register_hook, and since I'm already calling it indirectly, I think this could prove useful to both of us. I'd appreciate if anybody could help.

This is what I'm trying to do:

void foo () {
        unsigned long parameter;
        ...
        DEFINE_TIMER( timer, secfoo, jiffies + 250, parameter);
        ...
        add_timer(timer);
}

void secfoo( unsigned long param ) {
        ...
        nf_register_hook(...);
        ...
}

If I comment the "add_timer(timer);" and put "secfoo( parameter );" instead, it works out, though it doesn't help me, because I wanted to schedule it. The way it is, I get a kernel panic.

Good day,

Bruno Fiss


      Veja quais são os assuntos do momento no Yahoo! +Buscados
http://br.maisbuscados.yahoo.com
--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: nf_unregister_hook problem

[bcfiss]

Sorry for duplicates.

--- Em ter, 23/6/09, Jan Engelhardt <jengelh@medozas.de> escreveu:

>
> Does the problem go away if you call it indirectly? Like:
>
>
> stop_trigger.function = myfunc;
> ...
>
> myfunc(void)
> {
>     nf_unregister_hook(&nf_hook);
> }
>


Hi,

I've got a very similar question, about scheduling nf_register_hook,  
and since I'm already calling it indirectly, I think this could prove  
useful to both of us. I'd appreciate if anybody could help.

This is what I'm trying to do:

void foo () {
         unsigned long parameter;
         ...
         DEFINE_TIMER( timer, secfoo, jiffies + 250, parameter);
         ...
         add_timer(timer);
}

void secfoo( unsigned long param ) {
         ...
         nf_register_hook(...);
         ...
}

If I comment the "add_timer(timer);" and put "secfoo( parameter );"  
instead, it works out, though it doesn't help me, because I wanted to  
schedule it. The way it is, I get a kernel panic.

Good day,

Bruno Fiss

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

Re: nf_unregister_hook problem

[gmoliveira]

Hi Jan, thanks for your attention..
yes, i've already tried this way you said but the problem remains the same.
declaring nf_unregister_hook(&nf_hook) in __module_exit function worked well,
but requires user intervention. The problem consists in unregister the  
hook by timers...misunderstandable.

here goes and print screen with the "infinite loop error"
when occurs the BUG caused by nf_unregister_hook handled by a timer.

http://inf.ufrgs.br/~gmoliveira/screen.png

if anyone had a similar problem, pleasa contact me.

Citando Jan Engelhardt <jengelh@medozas.de>:

>
> On Tuesday 2009-06-23 20:47, gmoliveira@inf.ufrgs.br wrote:
>>
>> im using this:
>>
>> init_timer(&foo2);
>> stop_trigger.function = nf_unregister_hook;
>> stop_trigger.data = &nf_hook;
>> stop_trigger.expires = jiffies + msecs_to_jiffies(HZ * 5000);
>> add_timer(&foo2);
>>
>> it works to every function until now, unless nf_unregister_hook.
>
> Does the problem go away if you call it indirectly? Like:
>
>
> stop_trigger.function = myfunc;
> ...
>
> myfunc(void)
> {
> 	nf_unregister_hook(&nf_hook);
> }
>
>



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.