On 17.11.2016 22:44, Cong Wang wrote:
> On Sun, Oct 9, 2016 at 8:14 PM, Al Viro <v...@zeniv.linux.org.uk> wrote:
>> E.g what will happen if some code does a read on AF_UNIX socket with
>> some local mutex held? AFAICS, there are exactly two callers of
>> freezable_schedule_timeout() - this one and one in XFS; the latter is
>> in a kernel thread where we do have good warranties about the locking
>> environment, but here it's in the bleeding ->recvmsg/->splice_read and
>> for those assumption that caller doesn't hold any locks is pretty
>> strong, especially since it's not documented anywhere.
>>
>> What's going on there?
>
> Commit 2b15af6f95 ("af_unix: use freezable blocking calls in read")
> converts schedule_timeout() to its freezable version, it was probably correct
> at that time, but later, commit 2b514574f7e88c8498027ee366
> ("net: af_unix: implement splice for stream af_unix sockets") breaks its
> requirement for a freezable sleep:
>
> commit 0f9548ca10916dec166eaf74c816bded7d8e611d
>
> lockdep: check that no locks held at freeze time
>
> We shouldn't try_to_freeze if locks are held. Holding a lock can cause a
> deadlock if the lock is later acquired in the suspend or hibernate path
> (e.g. by dpm). Holding a lock can also cause a deadlock in the case of
> cgroup_freezer if a lock is held inside a frozen cgroup that is later
> acquired by a process outside that group.
>
> So probably we just need to revert commit 2b15af6f95 now.
>
> I am going to send a revert for at least -net and -stable, since Dmitry
> saw this warning again.
I am not an expert on freezing but this looks around right from the
freezer code. Awesome, thanks a lot for spotting this one!
