netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH net] net/sched: cls_flower: Return invalid for unknown ct_state flags rules
@ 2021-02-04  4:17 wenxu
  2021-02-04 13:38 ` Marcelo Ricardo Leitner
  0 siblings, 1 reply; 4+ messages in thread
From: wenxu @ 2021-02-04  4:17 UTC (permalink / raw)
  To: mleitner, i.maximets; +Cc: netdev

From: wenxu <wenxu@ucloud.cn>

Reject the unknown ct_state flags of cls flower rules. This also make
the userspace like ovs to probe the ct_state flags support in the
kernel.

Signed-off-by: wenxu <wenxu@ucloud.cn>
---
 include/uapi/linux/pkt_cls.h | 9 +++++++++
 net/sched/cls_flower.c       | 4 ++++
 2 files changed, 13 insertions(+)

diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
index ee95f42..1ac8ff8 100644
--- a/include/uapi/linux/pkt_cls.h
+++ b/include/uapi/linux/pkt_cls.h
@@ -591,8 +591,17 @@ enum {
 	TCA_FLOWER_KEY_CT_FLAGS_ESTABLISHED = 1 << 1, /* Part of an existing connection. */
 	TCA_FLOWER_KEY_CT_FLAGS_RELATED = 1 << 2, /* Related to an established connection. */
 	TCA_FLOWER_KEY_CT_FLAGS_TRACKED = 1 << 3, /* Conntrack has occurred. */
+
+	__TCA_FLOWER_KEY_CT_FLAGS_MAX,
 };
 
+#define TCA_FLOWER_KEY_CT_FLAGS_MAX \
+		((__TCA_FLOWER_KEY_CT_FLAGS_MAX - 1) << 1)
+#define TCA_FLOWER_KEY_CT_FLAGS_MASK \
+		(TCA_FLOWER_KEY_CT_FLAGS_MAX - 1)
+#define TCA_FLOWER_KEY_CT_FLAGS_UNKNOWN(v) \
+		((v) & (~TCA_FLOWER_KEY_CT_FLAGS_MASK))
+
 enum {
 	TCA_FLOWER_KEY_ENC_OPTS_UNSPEC,
 	TCA_FLOWER_KEY_ENC_OPTS_GENEVE, /* Nested
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 84f9325..ebee70f 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -1403,6 +1403,10 @@ static int fl_set_key_ct(struct nlattr **tb,
 		fl_set_key_val(tb, &key->ct_state, TCA_FLOWER_KEY_CT_STATE,
 			       &mask->ct_state, TCA_FLOWER_KEY_CT_STATE_MASK,
 			       sizeof(key->ct_state));
+		if (TCA_FLOWER_KEY_CT_FLAGS_UNKNOWN(mask->ct_state)) {
+			NL_SET_ERR_MSG(extack, "invalid ct_state flags");
+			return -EINVAL;
+		}
 	}
 	if (tb[TCA_FLOWER_KEY_CT_ZONE]) {
 		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES)) {
-- 
1.8.3.1


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH net] net/sched: cls_flower: Return invalid for unknown ct_state flags rules
  2021-02-04  4:17 [PATCH net] net/sched: cls_flower: Return invalid for unknown ct_state flags rules wenxu
@ 2021-02-04 13:38 ` Marcelo Ricardo Leitner
  2021-02-04 15:50   ` wenxu
  0 siblings, 1 reply; 4+ messages in thread
From: Marcelo Ricardo Leitner @ 2021-02-04 13:38 UTC (permalink / raw)
  To: wenxu; +Cc: i.maximets, netdev

Hi,

On Thu, Feb 04, 2021 at 12:17:24PM +0800, wenxu@ucloud.cn wrote:
> From: wenxu <wenxu@ucloud.cn>
> 
> Reject the unknown ct_state flags of cls flower rules. This also make
> the userspace like ovs to probe the ct_state flags support in the
> kernel.

That's a good start but it could also do some combination sanity
checks, like ovs does in validate_ct_state(). For example, it does:

      if (state && !(state & CS_TRACKED)) {
          ds_put_format(ds, "%s: invalid connection state: "
                        "If \"trk\" is unset, no other flags are set\n",

...
> --- a/net/sched/cls_flower.c
> +++ b/net/sched/cls_flower.c
> @@ -1403,6 +1403,10 @@ static int fl_set_key_ct(struct nlattr **tb,
>  		fl_set_key_val(tb, &key->ct_state, TCA_FLOWER_KEY_CT_STATE,
>  			       &mask->ct_state, TCA_FLOWER_KEY_CT_STATE_MASK,
>  			       sizeof(key->ct_state));
> +		if (TCA_FLOWER_KEY_CT_FLAGS_UNKNOWN(mask->ct_state)) {
> +			NL_SET_ERR_MSG(extack, "invalid ct_state flags");

cls_flower is inconsistent on this but please use NL_SET_ERR_MSG_MOD instead.

> +			return -EINVAL;
> +		}
>  	}
>  	if (tb[TCA_FLOWER_KEY_CT_ZONE]) {
>  		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES)) {
> -- 
> 1.8.3.1
> 


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH net] net/sched: cls_flower: Return invalid for unknown ct_state flags rules
  2021-02-04 13:38 ` Marcelo Ricardo Leitner
@ 2021-02-04 15:50   ` wenxu
  2021-02-04 18:23     ` Marcelo Ricardo Leitner
  0 siblings, 1 reply; 4+ messages in thread
From: wenxu @ 2021-02-04 15:50 UTC (permalink / raw)
  To: Marcelo Ricardo Leitner; +Cc: i.maximets, netdev


在 2021/2/4 21:38, Marcelo Ricardo Leitner 写道:
> Hi,
>
> On Thu, Feb 04, 2021 at 12:17:24PM +0800, wenxu@ucloud.cn wrote:
>> From: wenxu <wenxu@ucloud.cn>
>>
>> Reject the unknown ct_state flags of cls flower rules. This also make
>> the userspace like ovs to probe the ct_state flags support in the
>> kernel.
> That's a good start but it could also do some combination sanity
> checks, like ovs does in validate_ct_state(). For example, it does:
>
>       if (state && !(state & CS_TRACKED)) {
>           ds_put_format(ds, "%s: invalid connection state: "
>                         "If \"trk\" is unset, no other flags are set\n",
>
So this sanity checks maybe also need to be added in the ovs kernel modules?

The kernel datapath can work without ovs-vswitchd.


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH net] net/sched: cls_flower: Return invalid for unknown ct_state flags rules
  2021-02-04 15:50   ` wenxu
@ 2021-02-04 18:23     ` Marcelo Ricardo Leitner
  0 siblings, 0 replies; 4+ messages in thread
From: Marcelo Ricardo Leitner @ 2021-02-04 18:23 UTC (permalink / raw)
  To: wenxu; +Cc: i.maximets, netdev

On Thu, Feb 04, 2021 at 11:50:53PM +0800, wenxu wrote:
> 
> 在 2021/2/4 21:38, Marcelo Ricardo Leitner 写道:
> > Hi,
> >
> > On Thu, Feb 04, 2021 at 12:17:24PM +0800, wenxu@ucloud.cn wrote:
> >> From: wenxu <wenxu@ucloud.cn>
> >>
> >> Reject the unknown ct_state flags of cls flower rules. This also make
> >> the userspace like ovs to probe the ct_state flags support in the
> >> kernel.
> > That's a good start but it could also do some combination sanity
> > checks, like ovs does in validate_ct_state(). For example, it does:
> >
> >       if (state && !(state & CS_TRACKED)) {
> >           ds_put_format(ds, "%s: invalid connection state: "
> >                         "If \"trk\" is unset, no other flags are set\n",
> >
> So this sanity checks maybe also need to be added in the ovs kernel modules?
> 
> The kernel datapath can work without ovs-vswitchd.

IMHO that would be nice, yes.

  Marcelo


^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2021-02-04 18:25 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-02-04  4:17 [PATCH net] net/sched: cls_flower: Return invalid for unknown ct_state flags rules wenxu
2021-02-04 13:38 ` Marcelo Ricardo Leitner
2021-02-04 15:50   ` wenxu
2021-02-04 18:23     ` Marcelo Ricardo Leitner

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).