linux-kernel - Re: Tasks RCU vs Preempt RCU

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: <20180523011911.GD55359@joelaf.mtv.corp.google.com>
Date:   Tue, 22 May 2018 18:19:11 -0700
From:   Joel Fernandes <joel@...lfernandes.org>
To:     "Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>
Cc:     Steven Rostedt <rostedt@...dmis.org>, byungchul.park@....com,
        mathieu.desnoyers@...icios.com,
        Josh Triplett <josh@...htriplett.org>,
        Lai Jiangshan <jiangshanlai@...il.com>,
        linux-kernel@...r.kernel.org, kernel-team@...roid.com
Subject: Re: Tasks RCU vs Preempt RCU

On Tue, May 22, 2018 at 10:47:11AM -0700, Paul E. McKenney wrote:
> On Tue, May 22, 2018 at 01:27:00PM -0400, Steven Rostedt wrote:
> > On Tue, 22 May 2018 09:09:49 -0700
> > "Paul E. McKenney" <paulmck@...ux.vnet.ibm.com> wrote:
> > 
> > > Just for the record, if you guys realy want to take over Tasks RCU,
> > > I have no objections.  For one thing, I don't anticipate any other use
> > > cases for it (famous last words!).  But you break it, you buy it!  ;-)
> > 
> > It really matters how much of a burden is RCU_tasks to RCU itself? If
> > it causes a lot of headache for you, and it prevents you from cleaning
> > up RCU or making it better, then I would be happy to take it out of RCU
> > and maintain it separately myself. But if that's not the case, I'm happy
> > with keeping it within the RCU umbrella. Which brings me to the
> > question of what motivation does Joel have to remove it?
> 
> The burden on me from Tasks RCU has been quite light, so no need for a
> change from my end.
> 
> Over to you, Joel!  ;-)

My motivation was I felt RCU-preempt already did the same thing (which I
still believe it does) so its redundant. Although now I'm convinced from our
earlier discussions that its not feasible to do an rcu_read_lock and
rcu_read_unlock in trampoline code.

Sorry I didn't mean you should really nuke RCU-tasks if it has a purpose, but
I was more trying to understand what its purpose was that RCU-preempt didn't
solve. That's all.

And welcome back from Vacation Steve. I'm about to send v7 of my preempt/irq
tracepoint patches so the timing seems great. I hope you will be able to take
a look at them. thanks!

 - Joel