[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAM_iQpVucDgy-=ToNrH1a9h1h7goK33hfY_o30e5Rz1Pqwj4_A@mail.gmail.com>
Date: Tue, 18 Apr 2017 23:23:30 -0700
From: Cong Wang <xiyou.wangcong@...il.com>
To: Jamal Hadi Salim <jhs@...atatu.com>
Cc: Wolfgang Bumiller <w.bumiller@...xmox.com>,
Linux Kernel Network Developers <netdev@...r.kernel.org>,
"David S. Miller" <davem@...emloft.net>
Subject: Re: [PATCH v2 net 2/2] net sched actions: decrement module refcount earlier
On Tue, Apr 18, 2017 at 7:21 PM, Jamal Hadi Salim <jhs@...atatu.com> wrote:
> Indeed. Allocate the cookie before init? That way, we fail early
> and dont need to worry about restoring anything.
No, a->act_cookie needs an action pointer first. ;)
> In the case of a replace, do you really want to call tcf_hash_release?
>
Good point. Probably no, we already call it inside ->init().
Something like this...
diff --git a/net/sched/act_api.c b/net/sched/act_api.c
index 82b1d48d91cc..29ffc348fe2f 100644
--- a/net/sched/act_api.c
+++ b/net/sched/act_api.c
@@ -609,14 +609,12 @@ struct tc_action *tcf_action_init_1(struct net
*net, struct nlattr *nla,
if (cklen > TC_COOKIE_MAX_SIZE) {
err = -EINVAL;
- tcf_hash_release(a, bind);
- goto err_mod;
+ goto err_release;
}
if (nla_memdup_cookie(a, tb) < 0) {
err = -ENOMEM;
- tcf_hash_release(a, bind);
- goto err_mod;
+ goto err_release;
}
}
@@ -629,6 +627,9 @@ struct tc_action *tcf_action_init_1(struct net
*net, struct nlattr *nla,
return a;
+err_release:
+ if (err == ACT_P_CREATED)
+ tcf_hash_release(a, bind);
err_mod:
module_put(a_o->owner);
err_out:
Powered by blists - more mailing lists