linux-kernel - Re: Scenario TREE07 with CONFIG_PREEMPT

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]

Date:   Fri, 11 Mar 2022 18:26:20 +0100
From:   Frederic Weisbecker <frederic@...nel.org>
To:     "Paul E. McKenney" <paulmck@...nel.org>
Cc:     linux-kernel@...r.kernel.org
Subject: Re: Scenario TREE07 with CONFIG_PREEMPT_DYNAMIC=n?

On Fri, Mar 11, 2022 at 08:47:58AM -0800, Paul E. McKenney wrote:
> And there is one more issue with this code.  Someone invoking
> get_state_synchronize_rcu_expedited() in one task might naively expect
> that calls to synchronize_rcu_expedited() in some other task would cause
> a later poll_state_synchronize_rcu_expedited() would return true.
> 
> Except that if CONFIG_PREEMPT_NONE=y and there is only one CPU, those
> calls to synchronize_rcu_expedited() won't be helping at all.
> 
> I could imagine poll_state_synchronize_rcu_expedited() setting a
> global flag if there is only one CPU, which could be checked by
> __synchronize_rcu_expedited() and reset.
> 
> Is there a better way?

I would tend to think that in this case, it's the responsibility of the
caller to make sure that the task supposed to start the exp GP has a chance
to run (cond_resched(), etc...).