Messages in this thread | | | Date | Mon, 24 Aug 2009 11:00:58 +0200 | From | Ingo Molnar <> | Subject | Re: [PATCH -tip] v3 Consolidate sparse and lockdep declarations in include/linux/rcupdate.h |
| |
* Lai Jiangshan <laijs@cn.fujitsu.com> wrote:
> Paul E. McKenney wrote: > > From: Paul E. McKenney <paulmck@linux.vnet.ibm.com> > > > > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> > > --- > > include/linux/rcupdate.h | 46 ++++++++++++++++++++++++++++++++++++++++--- > > include/linux/rcupreempt.h | 4 +- > > include/linux/rcutree.h | 18 ----------------- > > 3 files changed, 44 insertions(+), 24 deletions(-) > > > > diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h > > index e920f0f..9d85ee1 100644 > > --- a/include/linux/rcupdate.h > > +++ b/include/linux/rcupdate.h > > @@ -80,6 +80,16 @@ extern int rcu_scheduler_active; > > (ptr)->next = NULL; (ptr)->func = NULL; \ > > } while (0) > > > > +#ifdef CONFIG_DEBUG_LOCK_ALLOC > > +extern struct lockdep_map rcu_lock_map; > > +# define rcu_read_acquire() \ > > + lock_acquire(&rcu_lock_map, 0, 0, 2, 1, NULL, _THIS_IP_) > > +# define rcu_read_release() lock_release(&rcu_lock_map, 1, _THIS_IP_) > > +#else > > +# define rcu_read_acquire() do { } while (0) > > +# define rcu_read_release() do { } while (0) > > +#endif > > + > > /** > > * rcu_read_lock - mark the beginning of an RCU read-side critical section. > > * > > @@ -109,7 +119,12 @@ extern int rcu_scheduler_active; > > * > > * It is illegal to block while in an RCU read-side critical section. > > */ > > -#define rcu_read_lock() __rcu_read_lock() > > +static inline void rcu_read_lock(void) > > +{ > > + __rcu_read_lock(); > > + __acquire(RCU); > > + rcu_read_acquire(); > > +} > > > > /** > > * rcu_read_unlock - marks the end of an RCU read-side critical section. > > @@ -126,7 +141,12 @@ extern int rcu_scheduler_active; > > * used as well. RCU does not care how the writers keep out of each > > * others' way, as long as they do so. > > */ > > -#define rcu_read_unlock() __rcu_read_unlock() > > +static inline void rcu_read_unlock(void) > > +{ > > + rcu_read_release(); > > + __release(RCU); > > + __rcu_read_unlock(); > > +} > > > > /** > > * rcu_read_lock_bh - mark the beginning of a softirq-only RCU critical section > > @@ -139,14 +159,24 @@ extern int rcu_scheduler_active; > > * can use just rcu_read_lock(). > > * > > */ > > -#define rcu_read_lock_bh() __rcu_read_lock_bh() > > +static inline void rcu_read_lock_bh(void) > > +{ > > + __rcu_read_lock_bh(); > > + __acquire(RCU_BH); > > + rcu_read_acquire(); > > +} > > > > /* > > * rcu_read_unlock_bh - marks the end of a softirq-only RCU critical section > > * > > * See rcu_read_lock_bh() for more information. > > */ > > -#define rcu_read_unlock_bh() __rcu_read_unlock_bh() > > +static inline void rcu_read_unlock_bh(void) > > +{ > > + rcu_read_release(); > > + __release(RCU_BH); > > + __rcu_read_unlock_bh(); > > +} > > > > /** > > * rcu_read_lock_sched - mark the beginning of a RCU-classic critical section > > @@ -160,10 +190,14 @@ extern int rcu_scheduler_active; > > static inline void rcu_read_lock_sched(void) > > { > > preempt_disable(); > > + __acquire(RCU_SCHED); > > + rcu_read_acquire(); > > } > > static inline void rcu_read_lock_sched_notrace(void) > > It need be marked "notrace" too. > > > { > > preempt_disable_notrace(); > > + __acquire(RCU_SCHED); > > + rcu_read_acquire(); > > } > > > > It may cause infinity recursion. > rcu_read_acquire() calls rcu_read_lock_sched_notrace() > before current->lockdep_recursion is set to 1 when tracing in on, > thus infinity recursion occurs.
ah, this would explain the crash i reported in:
http://lkml.org/lkml/2009/8/23/170
Ingo
| |