linux-kernel - Re: [PATCH v10 2/3] tracepoint: Make rcuidle tracepoint callers use SRCU

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: <20180718001625.GA154046@joelaf.mtv.corp.google.com>
Date:   Tue, 17 Jul 2018 17:16:25 -0700
From:   Joel Fernandes <joel@...lfernandes.org>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     linux-kernel@...r.kernel.org, kernel-team@...roid.com,
        Boqun Feng <boqun.feng@...il.com>,
        Byungchul Park <byungchul.park@....com>,
        Erick Reyes <erickreyes@...gle.com>,
        Ingo Molnar <mingo@...hat.com>,
        Julia Cartwright <julia@...com>,
        Masami Hiramatsu <mhiramat@...nel.org>,
        Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
        Namhyung Kim <namhyung@...nel.org>,
        Paul McKenney <paulmck@...ux.vnet.ibm.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Thomas Glexiner <tglx@...utronix.de>,
        Todd Kjos <tkjos@...gle.com>,
        Tom Zanussi <tom.zanussi@...ux.intel.com>,
        Will Deacon <will.deacon@....com>
Subject: Re: [PATCH v10 2/3] tracepoint: Make rcuidle tracepoint callers use
 SRCU

On Tue, Jul 17, 2018 at 08:25:28PM +0200, Peter Zijlstra wrote:
> On Fri, Jul 13, 2018 at 02:55:46PM -0700, Joel Fernandes wrote:
> > +		/* keep srcu and sched-rcu usage consistent */		\
> > +		preempt_disable_notrace();				\
> > +									\
> > +		/*							\
> > +		 * For rcuidle callers, use srcu since sched-rcu	\
> > +		 * doesn't work from the idle path.			\
> > +		 */							\
> > +		if (rcuidle)						\
> > +			idx = srcu_read_lock_notrace(&tracepoint_srcu);	\
> > +		else							\
> > +			rcu_read_lock_sched_notrace();			\
> 
> That else is completely superfluous. rcu_read_lock_sched_notrace() :=
> prempt_disable_notrace().

I have a patch as below on top of this series. Thanks for the suggestion.

Steve, Peter,

Is it Ok with you to take the below patch ontop of this series? I avoided
resending the 3-patch series since this is the only change, but let me know
what you prefer or makes it easy for you.

------->8--------

From: Joel Fernandes <joelaf@...gle.com>
Subject: [PATCH] tracepoint: Remove extra unneeded rcu_read_lock_sched_notrace

A great idea was provided that rcu_read_lock_sched_notrace being the
equivalent of preempt_disable_notrace is unnecessary to call in
tracepoint code, since we already call preempt_disable_notrace in
advance. So lets remove the extra call.

Suggested-by: Peter Zijlstra <peterz@...radead.org>
Signed-off-by: Joel Fernandes <joelaf@...gle.com>
---
 include/linux/tracepoint.h | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h
index 97e1d365a817..6e7bc6ebfcd8 100644
--- a/include/linux/tracepoint.h
+++ b/include/linux/tracepoint.h
@@ -161,8 +161,6 @@ extern void syscall_unregfunc(void);
 		 */							\
 		if (rcuidle)						\
 			idx = srcu_read_lock_notrace(&tracepoint_srcu);	\
-		else							\
-			rcu_read_lock_sched_notrace();			\
 									\
 		it_func_ptr = rcu_dereference_raw((tp)->funcs);		\
 									\
@@ -176,8 +174,6 @@ extern void syscall_unregfunc(void);
 									\
 		if (rcuidle)						\
 			srcu_read_unlock_notrace(&tracepoint_srcu, idx);\
-		else							\
-			rcu_read_unlock_sched_notrace();		\
 									\
 		preempt_enable_notrace();				\
 	} while (0)
-- 
2.18.0.203.gfac676dfb9-goog