Messages in this thread Patch in this message | | | Date | Tue, 23 Nov 2010 15:40:04 +0200 | From | Sergey Senozhatsky <> | Subject | addrconf: refcnt with IPV6_PRIVACY enabled |
| |
Hello,
I've sent a patch http://lkml.org/lkml/2010/11/21/124 "[PATCH] ipv6: fix inet6_dev refcnt with IPV6_PRIVACY enabled" related to wrong in6_dev refcnt value when IPV6_PRIVACY enabled.
I've took a second look, and question arised - Is it actually necessary to in6_dev_hold/in6_dev_put in ipv6_regen_rndid? In ipv6_regen_rndid we call in6_dev_hold only when mod_timer == 0, and always in6_dev_put.
I've removed check for < 0 and goto from __ipv6_regen_rndid call, since it always return 0.
Please, kindly review V2.
---
net/ipv6/addrconf.c | 10 +++------- 1 files changed, 3 insertions(+), 7 deletions(-)
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 2fc35b3..8187d14 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -425,7 +425,6 @@ static struct inet6_dev * ipv6_add_dev(struct net_device *dev) dev->name); ndev->cnf.use_tempaddr = -1; } else { - in6_dev_hold(ndev); ipv6_regen_rndid((unsigned long) ndev); } #endif @@ -1653,8 +1652,7 @@ static void ipv6_regen_rndid(unsigned long data) if (idev->dead) goto out; - if (__ipv6_regen_rndid(idev) < 0) - goto out; + __ipv6_regen_rndid(idev); expires = jiffies + idev->cnf.temp_prefered_lft * HZ - @@ -1667,13 +1665,11 @@ static void ipv6_regen_rndid(unsigned long data) goto out; } - if (!mod_timer(&idev->regen_timer, expires)) - in6_dev_hold(idev); - + mod_timer(&idev->regen_timer, expires); + out: write_unlock_bh(&idev->lock); rcu_read_unlock_bh(); - in6_dev_put(idev); } static int __ipv6_try_regen_rndid(struct inet6_dev *idev, struct in6_addr *tmpaddr) {
| |