linux-kernel - Re: [PATCH v6 1/4] rcu: Make call

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: <YzMTRrMKCnEMl/ov@google.com>
Date:   Tue, 27 Sep 2022 15:14:14 +0000
From:   Joel Fernandes <joel@...lfernandes.org>
To:     Uladzislau Rezki <urezki@...il.com>
Cc:     rcu@...r.kernel.org, linux-kernel@...r.kernel.org,
        rushikesh.s.kadam@...el.com, neeraj.iitr10@...il.com,
        frederic@...nel.org, paulmck@...nel.org, rostedt@...dmis.org
Subject: Re: [PATCH v6 1/4] rcu: Make call_rcu() lazy to save power

On Tue, Sep 27, 2022 at 04:59:44PM +0200, Uladzislau Rezki wrote:
[...]
> > >    systemd-udevd-642     [050] .....    17.349832: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-665     [061] .....    17.349834: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-675     [012] .....    17.349835: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-675     [012] .....    17.349853: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-642     [050] .....    17.349861: __call_rcu_common: -> 0x0: __dentry_kill+0x140/0x180 <- 0x2
> > >    systemd-udevd-675     [012] .....    17.349873: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-665     [061] .....    17.349879: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-665     [061] .....    17.350007: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-665     [061] .....    17.350011: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-665     [061] .....    17.350080: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-665     [061] .....    17.350175: __call_rcu_common: -> ____fput+0x0/0x10: task_work_run+0x5c/0x90 <- 0x0
> > >    systemd-udevd-665     [061] .....    17.350362: dev_change_name: --> renamed from eth0
> > > <snip>
> > > 
> > > First delay:
> > > 
> > > <snip>
> > > systemd-udevd-645     [053] .....     2.339024: __call_rcu_common: -> 0x0: __dentry_kill+0x140/0x180 <- 0x2
> > >   kworker/0:3-546     [000] d..1.     6.329516: __call_rcu_common: -> 0x0: queue_rcu_work+0x2b/0x40 <- 0xffff8c70effe40a0
> > > <snip>
> > > 
> > > __dentry_kill() function and after 4 seconds there is another one queue_rcu_work().
> > > I have checked the __dentry_kill() if it can do any sync talk with RCU but from the
> > > first glance i do not see anything critical. But more attention is required.
> > 
> > Can you log rcu_barrier() as well? It could be that the print is just a side
> > effect of something else that is not being printed.
> > 
> It has nothing to do with rcu_barrier() in my case. Also i have checked
> the synchronize_rcu() it also works as expected, i.e. it is not a
> blocking reason.
> 
> Have you tried my config?

Yes I am in the process of trying it.

thanks,

 - Joel