Re: [PATCH] tcm_fc: rcu_deref outside rcu lock/unlock section

["Rustad, Mark D" <mark.d.rustad@intel.com>]

On Aug 18, 2012, at 3:35 PM, Nicholas A. Bellinger wrote:

> On Sat, 2012-08-18 at 16:10 +0400, Denis Efremov wrote:
>> Use rcu_dereference_protected in order to prevent lockdep
>> complaint. Sequel of the patch 863555be
>> 
>> Found by Linux Driver Verification project (linuxtesting.org).
>> 
>> Signed-off-by: Denis Efremov <yefremov.denis@gmail.com>
>> ---
>> drivers/target/tcm_fc/tfc_sess.c |    4 +++-
>> 1 files changed, 3 insertions(+), 1 deletions(-)
>> 
>> diff --git a/drivers/target/tcm_fc/tfc_sess.c b/drivers/target/tcm_fc/tfc_sess.c
>> index 87901fa..3c9e5b5 100644
>> --- a/drivers/target/tcm_fc/tfc_sess.c
>> +++ b/drivers/target/tcm_fc/tfc_sess.c
>> @@ -456,7 +456,9 @@ static void ft_prlo(struct fc_rport_priv *rdata)
>> 	struct ft_tport *tport;
>> 
>> 	mutex_lock(&ft_lport_lock);
>> -	tport = rcu_dereference(rdata->local_port->prov[FC_TYPE_FCP]);
>> +	tport = rcu_dereference_protected(rdata->local_port->prov[FC_TYPE_FCP],
>> +					  lockdep_is_held(&ft_lport_lock));
>> +
>> 	if (!tport) {
>> 		mutex_unlock(&ft_lport_lock);
>> 		return;
> 
> This looks OK to silence lockdep.  CC'ing MDR + Kiran for good measure
> here, and will move from target-pending queue -> master with their ACK.
> 
> Thanks Dennis!
> 
> --nab


Ack. In fact I wonder why I didn't encounter it myself.

-- 
Mark Rustad, LAN Access Division, Intel Corporation