linux-kernel - Re: [BUG v6.0-rc2] lockdep splat on ct_kernel

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:   Mon, 22 Aug 2022 21:40:24 -0400
From:   Steven Rostedt <rostedt@...dmis.org>
To:     LKML <linux-kernel@...r.kernel.org>
Cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        Frederic Weisbecker <fweisbec@...il.com>,
        "Paul E. McKenney" <paulmck@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Jiri Olsa <olsajiri@...il.com>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [BUG v6.0-rc2] lockdep splat on ct_kernel_enter()

On Mon, 22 Aug 2022 20:48:25 -0400
Steven Rostedt <rostedt@...dmis.org> wrote:

> On Mon, 22 Aug 2022 18:38:36 -0400
> Steven Rostedt <rostedt@...dmis.org> wrote:
> 
> > > The warning happens with the commit, and does not for the commit just
> > > before it is applied.
> > >     
> > 
> > Nevermind, I tested the wrong kernel (the grub reboot failed to boot to the
> > new kernel). But it did give me a different lockdep splat when I tested the
> > right kernel :-/
> > 
> > I'll investigate some more.  
> 
> I was close. It was this commit:
> 
> ed4ae5eff4b3 ("rcu: Apply noinstr to rcu_idle_enter() and rcu_idle_exit()")
> 
> that causes the new error message.
> 

And this patch (which is wrong, but will at least let my tests finish
testing my code) makes the warning go away.

-- Steve

diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index 9a5edab5558c..14147d31e651 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -896,13 +896,13 @@ static void noinstr rcu_eqs_exit(bool user)
  * If you add or remove a call to rcu_idle_exit(), be sure to test with
  * CONFIG_RCU_EQS_DEBUG=y.
  */
-void noinstr rcu_idle_exit(void)
+void rcu_idle_exit(void)
 {
 	unsigned long flags;
 
-	raw_local_irq_save(flags);
+	local_irq_save(flags);
 	rcu_eqs_exit(false);
-	raw_local_irq_restore(flags);
+	local_irq_restore(flags);
 }
 EXPORT_SYMBOL_GPL(rcu_idle_exit);