[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAM_iQpUrbikYYfttzEn8AihpSmXtwrdnV-949BzVH_0DdwW37g@mail.gmail.com>
Date: Thu, 8 Sep 2016 10:12:20 -0700
From: Cong Wang <xiyou.wangcong@...il.com>
To: Jamal Hadi Salim <jhs@...atatu.com>
Cc: David Miller <davem@...emloft.net>,
Linux Kernel Network Developers <netdev@...r.kernel.org>
Subject: Re: [PATCH net 1/1] net sched actions: fix GETing actions
On Thu, Sep 8, 2016 at 5:12 AM, Jamal Hadi Salim <jhs@...atatu.com> wrote:
> From: Jamal Hadi Salim <jhs@...atatu.com>
>
> Example of what broke:
> ...add a gact action to drop
> sudo $TC actions add action drop index 10
> ...now retrieve it, looks good
> sudo $TC actions get action gact index 10
> ...retrieve it again and find it is gone!
> sudo $TC actions get action gact index 10
This changelog sucks, it only tells the phenomenon of the bug,
never explains what causes the bug, which is crucial for review.
>
> Fixes: f07fed82ad79 ("net_sched: remove the leftover cleanup_a())
>
> Signed-off-by: Jamal Hadi Salim <jhs@...atatu.com>
> ---
> net/sched/act_api.c | 20 +++++++++++++++++---
> 1 file changed, 17 insertions(+), 3 deletions(-)
>
> diff --git a/net/sched/act_api.c b/net/sched/act_api.c
> index d09d068..d4b4704 100644
> --- a/net/sched/act_api.c
> +++ b/net/sched/act_api.c
> @@ -612,6 +612,8 @@ int tcf_action_init(struct net *net, struct nlattr *nla,
> goto err;
> }
> act->order = i;
> + if (ovr)
> + act->tcfa_refcnt+=1;
This looks like not relevant to commit f07fed82ad79 at all?
> list_add_tail(&act->list, actions);
> }
> return 0;
> @@ -856,6 +858,15 @@ tcf_del_notify(struct net *net, struct nlmsghdr *n, struct list_head *actions,
> return ret;
> }
>
> +static void cleanup_a(struct list_head *actions)
> +{
> + struct tc_action *a, *tmp;
> +
> + list_for_each_entry_safe(a, tmp, actions, list) {
> + list_del(&a->list);
> + }
After my patches, all the action lists are now temporary,
you didn't explain why we should remove an action from
a temporary list, they should just come and go, since
we call ist_add() every time.
Did see any list corruption or kernel crash (your changelog
didn't say so)?
> +}
> +
> static int
> tca_action_gd(struct net *net, struct nlattr *nla, struct nlmsghdr *n,
> u32 portid, int event)
> @@ -883,6 +894,7 @@ tca_action_gd(struct net *net, struct nlattr *nla, struct nlmsghdr *n,
> goto err;
> }
> act->order = i;
> + act->tcfa_refcnt+=1;
Maybe we only need to fixup the refcnt without touching list,
from my quick observation to your bug report (again, you didn't explain).
> list_add_tail(&act->list, &actions);
> }
>
> @@ -892,10 +904,12 @@ tca_action_gd(struct net *net, struct nlattr *nla, struct nlmsghdr *n,
> ret = tcf_del_notify(net, n, &actions, portid);
> if (ret)
> goto err;
> + cleanup_a(&actions);
> return ret;
> }
> err:
> tcf_action_destroy(&actions, 0);
> + cleanup_a(&actions);
> return ret;
> }
>
> @@ -931,10 +945,10 @@ tcf_action_add(struct net *net, struct nlattr *nla, struct nlmsghdr *n,
> LIST_HEAD(actions);
>
> ret = tcf_action_init(net, nla, NULL, NULL, ovr, 0, &actions);
> - if (ret)
> - return ret;
>
> - return tcf_add_notify(net, n, &actions, portid);
> + ret = tcf_add_notify(net, n, &actions, portid);
> + cleanup_a(&actions);
> + return ret;
> }
>
> static int tc_ctl_action(struct sk_buff *skb, struct nlmsghdr *n)
> --
> 1.9.1
>
Powered by blists - more mailing lists