Re: [2.6.33-rc5] starting emacs makes lockdep warning

[ebiederm@xmission.com (Eric W. Biederman)]

KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> writes:

>> > On Tue, Jan 26, 2010 at 11:20 AM, KOSAKI Motohiro
>> > <kosaki.motohiro@jp.fujitsu.com> wrote:
>> > > Hi
>> > >
>> > > Current linus tree made following lockdep warning when starting emacs command.
>> > > Is this known issue?
>> > >
>> > >
>> > > =========================================================
>> > > [ INFO: possible irq lock inversion dependency detected ]
>> > > 2.6.33-rc5 #77
>> > > ---------------------------------------------------------
>> > > emacs/1609 just changed the state of lock:
>> > >  (&(&tty->ctrl_lock)->rlock){+.....}, at: [<ffffffff8127c648>] tty_fasync+0xe8/0x190
>> > > but this lock took another, HARDIRQ-unsafe lock in the past:
>> > >  (&(&sighand->siglock)->rlock){-.....}
>> > >
>> > > and interrupts could create inverse lock ordering between them.
>> > >
>> > >
>> > 
>> > Hey,
>> > 
>> > does reverting commit 703625118 help?
>> 
>> Seems solved.
>> 
>> Thanks.
>
> I'm sorry.
> I forgot to cc related person at last mail.
>
> Greg, can you please consider revert commit 703625118?

It looks like f_modown needs to do irqsave irqrestore to be safely
called in this context.  My apologies for missing this when I
originally made the suggestion.

As for the other comments I would be very surprised if lock_kernel()
offers any real protection.

I really don't understand what it is talking about siglock being
irq unsafe, that seems wrong on oh so many levels.

Eric

n