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:	Wed, 22 Jan 2014 11:34:40 -0800
From:	"Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>
To:	Oleg Nesterov <oleg@...hat.com>
Cc:	Dave Jones <davej@...hat.com>,
	Peter Zijlstra <peterz@...radead.org>,
	Alan Stern <stern@...land.harvard.edu>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	Ingo Molnar <mingo@...nel.org>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Steven Rostedt <rostedt@...dmis.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	linux-kernel@...r.kernel.org
Subject: Re: uninline rcu_lock_acquire/etc ?

On Wed, Jan 22, 2014 at 07:31:25PM +0100, Oleg Nesterov wrote:
> On 01/21, Paul E. McKenney wrote:
> >
> > On Tue, Jan 21, 2014 at 08:39:09PM +0100, Oleg Nesterov wrote:
> > > On 01/21, Oleg Nesterov wrote:
> > > >
> > > > But I agreed that the code looks simpler with bitfields, so perhaps
> > > > this patch is better.
> > >
> > > Besides, I guess the major offender is rcu...
> > >
> > > Paul, can't we do something like below? Saves 19.5 kilobytes,
> > >
> > > 	-       5255131 2974376 10125312        18354819        1181283 vmlinux
> > > 	+	5235227 2970344 10125312        18330883        117b503 vmlinux
> > >
> > > probably we can also uninline rcu_lockdep_assert()...
> >
> > Looks mostly plausible, some questions inline below.
> 
> Thanks!
> 
> > >  static inline void rcu_read_lock(void)
> > >  {
> > > -	__rcu_read_lock();
> > >  	__acquire(RCU);
> > > -	rcu_lock_acquire(&rcu_lock_map);
> > > -	rcu_lockdep_assert(rcu_is_watching(),
> > > -			   "rcu_read_lock() used illegally while idle");
> > > +	__rcu_read_lock();
> > > +	rcu_lock_acquire();
> >
> > Not sure why __rcu_read_lock() needs to be in any particular order
> > with respect to the sparse __acquire(RCU), but should work either way.
> > Same question about the other reorderings of similar statements.
> 
> I did this unconsciously and for no reason, will revert this accidental
> change.
> 
> > >  static inline void rcu_read_lock_sched(void)
> > >  {
> > > -	preempt_disable();
> > >  	__acquire(RCU_SCHED);
> > > -	rcu_lock_acquire(&rcu_sched_lock_map);
> > > +	preempt_disable();
> > > +	rcu_lock_acquire_sched();
> > >  	rcu_lockdep_assert(rcu_is_watching(),
> > >  			   "rcu_read_lock_sched() used illegally while idle");
> >
> > The above pair of lines (rcu_lockdep_assert()) should also be removed,
> > correct?
> 
> yes, sure, thanks,
> 
> > > @@ -862,8 +867,8 @@ static inline void rcu_read_lock_sched(void)
> > >  /* Used by lockdep and tracing: cannot be traced, cannot call lockdep. */
> > >  static inline notrace void rcu_read_lock_sched_notrace(void)
> > >  {
> > > -	preempt_disable_notrace();
> > >  	__acquire(RCU_SCHED);
> > > +	preempt_disable_notrace();
> >
> > I cannot help repeating myself on this one...  ;-)
> >
> > Why the change in order?
> 
> see above ;)
> 
> > > --- a/kernel/rcu/update.c
> > > +++ b/kernel/rcu/update.c
> > > @@ -333,4 +333,47 @@ static int __init check_cpu_stall_init(void)
> > >  }
> > >  early_initcall(check_cpu_stall_init);
> > >
> > > +#if defined(CONFIG_DEBUG_LOCK_ALLOC) || defined(CONFIG_PROVE_RCU)
> > > +
> > > +static void ck_rcu_is_watching(const char *message)
> > > +{
> > > +	rcu_lockdep_assert(rcu_is_watching(), message);
> > > +}
> > > +
> > > +void rcu_lock_acquire(void)
> > > +{
> > > +	__rcu_lock_acquire(&rcu_lock_map, _RET_IP_);
> > > +	ck_rcu_is_watching("rcu_read_lock() used illegally while idle");
> > > +}
> > > +
> > > +void rcu_lock_release(void)
> > > +{
> > > +	ck_rcu_is_watching("rcu_read_unlock() used illegally while idle");
> > > +	__rcu_lock_release(&rcu_lock_map, _RET_IP_);
> > > +}
> > > ...
> 
> Also, this all should be exported. And I think cleanuped somehow.

I would be happy to take a patch with the above issues fixed.

							Thanx, Paul

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists