linux-kernel - Re: [RFC][PATCH v3 6/6] rcu/tree: Use irq_work_queue

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] [day] [month] [year] [list]

Message-ID: <20201029161431.GR2628@hirez.programming.kicks-ass.net>
Date:   Thu, 29 Oct 2020 17:14:31 +0100
From:   Peter Zijlstra <peterz@...radead.org>
To:     "Paul E. McKenney" <paulmck@...nel.org>
Cc:     mingo@...nel.org, linux-kernel@...r.kernel.org, will@...nel.org,
        hch@....de, axboe@...nel.dk, chris@...is-wilson.co.uk,
        davem@...emloft.net, kuba@...nel.org, fweisbec@...il.com,
        oleg@...hat.com, vincent.guittot@...aro.org
Subject: Re: [RFC][PATCH v3 6/6] rcu/tree: Use irq_work_queue_remote()

On Thu, Oct 29, 2020 at 09:04:48AM -0700, Paul E. McKenney wrote:
> On Thu, Oct 29, 2020 at 10:10:53AM +0100, Peter Zijlstra wrote:

> > Dang, clearly TREE01 didn't actually hit any of this code :/ Is there
> > another test I should be running?
> 
> TREE01 is fine, but you have to tell rcutorture to actually generate an
> RCU CPU stall warning.  Like this for a 25-second stall with interrupts
> disabled:
> 
> tools/testing/selftests/rcutorture/bin/kvm.sh --allcpus --duration 3 --configs "10*TREE04" --bootargs "rcutorture.stall_cpu=25 rcutorture.stall_cpu_irqsoff=1" --trust-make

> Of course, to test your change, you also need the grace-period kthread to
> migrate to the stalled CPU just after interrupts are enabled.  For this,
> you need something like an 11-second stall plus something to move the
> grace-period kthread at the right (wrong?) time.  Or just run the above
> commands in a loop on a system with ample storage over night or some such.
> I see about 70MB of storage per run, so disk size shouldn't be too much
> of a problem.

Thanks!, I'll make the above run over night in a loop.