linux-kernel - Re: KCSAN: data-race in find_next_bit / rcu_report_exp_cpu

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20191008035834.GB2609633@tardis>
Date:   Tue, 8 Oct 2019 11:58:34 +0800
From:   Boqun Feng <boqun.feng@...il.com>
To:     "Paul E. McKenney" <paulmck@...nel.org>
Cc:     Joel Fernandes <joel@...lfernandes.org>,
        Marco Elver <elver@...gle.com>,
        syzbot <syzbot+134336b86f728d6e55a0@...kaller.appspotmail.com>,
        josh@...htriplett.org, rostedt@...dmis.org,
        mathieu.desnoyers@...icios.com, jiangshanlai@...il.com,
        rcu@...r.kernel.org, a@...table.cc,
        b.a.t.m.a.n@...ts.open-mesh.org, davem@...emloft.net,
        LKML <linux-kernel@...r.kernel.org>, mareklindner@...mailbox.ch,
        netdev@...r.kernel.org, sw@...onwunderlich.de,
        syzkaller-bugs@...glegroups.com
Subject: Re: KCSAN: data-race in find_next_bit / rcu_report_exp_cpu_mult

On Mon, Oct 07, 2019 at 08:33:53PM -0700, Paul E. McKenney wrote:
[...]
> > ---
> >  kernel/rcu/tree_exp.h | 17 +++++++----------
> >  1 file changed, 7 insertions(+), 10 deletions(-)
> > 
> > diff --git a/kernel/rcu/tree_exp.h b/kernel/rcu/tree_exp.h
> > index af7e7b9c86af..fb51752ac9a6 100644
> > --- a/kernel/rcu/tree_exp.h
> > +++ b/kernel/rcu/tree_exp.h
> > @@ -372,12 +372,10 @@ static void sync_rcu_exp_select_node_cpus(struct work_struct *wp)
> >  	raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
> >  
> >  	/* IPI the remaining CPUs for expedited quiescent state. */
> > -	for_each_leaf_node_cpu_mask(rnp, cpu, rnp->expmask) {
> > +	for_each_leaf_node_cpu_mask(rnp, cpu, mask_ofl_ipi) {
> >  		unsigned long mask = leaf_node_cpu_bit(rnp, cpu);
> >  		struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
> >  
> > -		if (!(mask_ofl_ipi & mask))
> > -			continue;
> >  retry_ipi:
> >  		if (rcu_dynticks_in_eqs_since(rdp, rdp->exp_dynticks_snap)) {
> >  			mask_ofl_test |= mask;
> 
> This part I have already on -rcu branch "dev".
> 
> > @@ -389,10 +387,10 @@ static void sync_rcu_exp_select_node_cpus(struct work_struct *wp)
> >  		}
> >  		ret = smp_call_function_single(cpu, rcu_exp_handler, NULL, 0);
> >  		put_cpu();
> > -		if (!ret) {
> > -			mask_ofl_ipi &= ~mask;
> > +		/* the CPU responses the IPI, and it will report QS itself */
> > +		if (!ret)
> >  			continue;
> > -		}
> > +
> >  		/* Failed, raced with CPU hotplug operation. */
> >  		raw_spin_lock_irqsave_rcu_node(rnp, flags);
> >  		if ((rnp->qsmaskinitnext & mask) &&
> > @@ -403,13 +401,12 @@ static void sync_rcu_exp_select_node_cpus(struct work_struct *wp)
> >  			schedule_timeout_uninterruptible(1);
> >  			goto retry_ipi;
> >  		}
> > -		/* CPU really is offline, so we can ignore it. */
> > -		if (!(rnp->expmask & mask))
> > -			mask_ofl_ipi &= ~mask;
> > +		/* CPU really is offline, and we need its QS. */
> > +		if (rnp->expmask & mask)
> > +			mask_ofl_test |= mask;
> >  		raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
> >  	}
> >  	/* Report quiescent states for those that went offline. */
> > -	mask_ofl_test |= mask_ofl_ipi;
> >  	if (mask_ofl_test)
> >  		rcu_report_exp_cpu_mult(rnp, mask_ofl_test, false);
> >  }
> 
> Would you be willing to port this optimization on top of current -rcu
> branch "dev" with an suitably modified commit message?
> 

Sure, will do ;-)

Regards,
Boqun

> 							Thanx, Paul

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)