Re: [PATCH v5 3.1.0-rc4-tip 26/26] uprobes: queue signals whilethread is singlestepping.

[Oleg Nesterov]

On 10/12, Srikar Dronamraju wrote:
>
> I think we should be okay if the test exits in UTASK_SSTEP state.

Yes, and afaics we can't avoid this case, at least currently.

But we should move free_uprobe_utask() to mm_release(), or somewhere
else before mm->core_state check in exit_mm().

My main concern is stop/freeze in UTASK_SSTEP state. If nothing else,
debugger can attach to the stopped task and disable the stepping. Or
SIGKILL, it should work in this case.

> > Great. I'll think a bit more and send you the "final" version tomorrow.
> > Assuming we can change sstep_complete() as we discussed, it doesn't need
> > fatal_signal_pending().
>
> Okay.

Sorry. I was busy today. Tomorrow ;)

> > HOWEVER. There is yet another problem. Another thread can, say, unmap()
> > xol_vma. In this case we should ensure that the task can't fault in an
> > endless loop.
>
> Hmm should we add a check in unmap() to see if the vma that we are
> trying to unmap is the xol_vma and if so return?

Oh, I am not sure. You know, I _think_ that perhaps we should do something
diferent in the long term. In particular, this xol page should not have
vma at all. This way we shouldn't worry about unmap/remap/mprotect.
But even if this is possible (I am not really sure), I do not think we
should do this right now.

> Our assumption has been that once an xol_vma has been created, it should
> be around till the process gets killed.

Yes, I see. But afaics this assumption is currently wrong. This means
that we should ensure the evil application can't exploit this fact.

Oleg.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/