Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Fri, 28 Jan 2022 08:15:47 -0800
From:   "Paul E. McKenney" <paulmck@kernel.org>
To:     Mark Rutland <mark.rutland@arm.com>
Cc:     Sven Schnelle <svens@linux.ibm.com>,
        Steven Rostedt <rostedt@goodmis.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Ingo Molnar <mingo@kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        Yinan Liu <yinan@linux.alibaba.com>,
        Ard Biesheuvel <ardb@kernel.org>,
        Kees Cook <keescook@chromium.org>,
        Sachin Sant <sachinp@linux.ibm.com>,
        linuxppc-dev@lists.ozlabs.org,
        Russell King <linux@armlinux.org.uk>,
        linux-arm-kernel@lists.infradead.org, hca@linux.ibm.com,
        linux-s390@vger.kernel.org
Subject: Re: ftrace hangs waiting for rcu
Message-ID: <20220128161547.GL4285@paulmck-ThinkPad-P17-Gen-1>
Reply-To: paulmck@kernel.org
References: <20220127114249.03b1b52b@gandalf.local.home>
 <YfLjIOlGfFmbh1Zv@FVFF77S0Q05N>
 <yt9dy231yq90.fsf_-_@linux.ibm.com>
 <YfQCohKWJg9H+uID@FVFF77S0Q05N>
 <yt9dee4rn8q7.fsf@linux.ibm.com>
 <YfQVzba5thVs+qap@FVFF77S0Q05N>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YfQVzba5thVs+qap@FVFF77S0Q05N>
Precedence: bulk

On Fri, Jan 28, 2022 at 04:11:57PM +0000, Mark Rutland wrote:
> On Fri, Jan 28, 2022 at 05:08:48PM +0100, Sven Schnelle wrote:
> > Hi Mark,
> > 
> > Mark Rutland <mark.rutland@arm.com> writes:
> > 
> > > On arm64 I bisected this down to:
> > >
> > >   7a30871b6a27de1a ("rcu-tasks: Introduce ->percpu_enqueue_shift for dynamic queue selection")
> > >
> > > Which was going wrong because ilog2() rounds down, and so the shift was wrong
> > > for any nr_cpus that was not a power-of-two. Paul had already fixed that in
> > > rcu-next, and just sent a pull request to Linus:
> > >
> > >   https://lore.kernel.org/lkml/20220128143251.GA2398275@paulmck-ThinkPad-P17-Gen-1/
> > >
> > > With that applied, I no longer see these hangs.
> > >
> > > Does your s390 test machine have a non-power-of-two nr_cpus, and does that fix
> > > the issue for you?
> > 
> > We noticed the PR from Paul and are currently testing the fix. So far
> > it's looking good. The configuration where we have seen the hang is a
> > bit unusual:
> > 
> > - 16 physical CPUs on the kvm host
> > - 248 logical CPUs inside kvm
> 
> Aha! 248 is notably *NOT* a power of two, and in this case the shift would be
> wrong (ilog2() would give 7, when we need a shift of 8).
> 
> So I suspect you're hitting the same issue as I was.

And apparently no one runs -next on systems having a non-power-of-two
number of CPUs.  ;-)

							Thanx, Paul

> Thanks,
> Mark.
> 
> > - debug kernel both on the host and kvm guest
> > 
> > So things are likely a bit slow in the kvm guest. Interesting is that
> > the number of CPUs is even. But maybe RCU sees an odd number of CPUs
> > and gets confused before all cpus are brought up. Have to read code/test
> > to see whether that could be possible.
> > 
> > Thanks for investigating!
> > Sven