[Devel] Re: Signals to cinit

Sukadev Bhattiprolu sukadev at linux.vnet.ibm.com
Thu Nov 13 11:10:39 PST 2008


Oleg Nesterov [oleg at redhat.com] wrote:
| (lkml cced because containers list's archive is not useable)
| 
| On 11/10, Oleg Nesterov wrote:
| >
| > On 11/01, sukadev at linux.vnet.ibm.com wrote:
| > >
| > > Other approaches to try ?
| >
| > I think we should try to do something simple, even if not perfect. Because
| > most users do not care about this problem since they do not use containers
| > at all. It would be very sad to add intrusive changes to the code.
| >
| > I think we should fix another problem first. send_signal()->copy_siginfo()
| > path must be changed anyway, when the signal comes from the parent ns we
| > report the "wrong" si_code/si_pid, yes? So, somehow send_signal() must
| > have "bool from_parent_ns" (or whatever) annyway.
| >
| > Now, let's forget forget for a moment that send_signal()->__sigqueue_alloc()
| > can fail.
| >
| > I think we should encode this "from_parent_ns" into "struct siginfo". I do
| > not think it is good idea to extend this structure, I think we can introduce
| > SI_FROM_PARENT_NS or we perhaps can use "SI_FROMUSER(info) && info->si_pid == 0".
| > Or something. yes, sys_rt_sigqueueinfo() is problematic...
| >
| > Now, copy_process(CLONE_NEWPID) sets child->signal |= SIGNAL_UNKILLABLE, this
| > protects cinit from unwanted signals. Then we change get_signal_to_deliver()
| >
| > 	-	if (unlikely(signal->flags & SIGNAL_UNKILLABLE) &&
| > 	+	if (unlikely(signal->flags & SIGNAL_UNKILLABLE) && !siginfo_from_parent_ns(info)
| >
| > and now we can kill cinit from parent ns. This needs more checks if we want
| > to stop/strace it, but perhaps this is enough for the start. Note that we
| > do not need to change complete_signal(), at least for now, the code under
| > "if (sig_fatal(p, sig)" is just optimization.
| >
| >
| > So, afaics, the only real problem is how we can handle the case when
| > __sigqueue_alloc() fails. I think for the start we can just return
| > -ENOMEM in this case (when from_parent_ns == T). Then we can improve
| > this behaviour. We can change complete_signal() to ensure that the
| > fatal signal from the upper ns always kills cinit, and in this case
| > we ignore the the failed __sigqueue_alloc(). This way at least SIGKILL
| > always works.
| >
| > Yes, this is not perfect, and it is very possible I missed something
| > else. But simple.
| 
| But how can send_signal() know that the signal comes from the upper ns?
| This is not trivial, we can't blindly use current to check. The signal
| can be sent from irq/workqueue/etc.
| 
| Perhaps we can start with something like the patch below. Not that I like
| it very much though. We should really place this code under
| CONFIG_I_DO_CARE_ABOUT_NAMESPACES ;)
| 
| Oleg.
| 
| --- K-IS/kernel/signal.c~T	2008-11-10 19:21:17.000000000 +0100
| +++ K-IS/kernel/signal.c	2008-11-10 20:31:24.000000000 +0100
| @@ -798,11 +798,19 @@ static inline int legacy_queue(struct si
|  	return (sig < SIGRTMIN) && sigismember(&signals->signal, sig);
|  }
| 
| +#define SIG_FROM_USER	INT_MIN		/* MSB */
| +

Not necessarily for the problem at hand, but in the long run, is it
worth isolating kernel's siginfo from user's siginfo_t (which is pretty
much carved in stone).

Like the existing 'struct k_sigaction' and 'struct sigaction' have
a 'struct k_siginfo' that is a superset of 'siginfo_t' ?

That might give us more flexibility in passing any additional flags/
values we need in the kernel ?
_______________________________________________
Containers mailing list
Containers at lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers




More information about the Devel mailing list