[PATCH net] Revert "openvswitch: switch to per-action label counting in conntrack"

[PATCH net] Revert "openvswitch: switch to per-action label counting in conntrack"

[Xin Long]

Currently, ovs_ct_set_labels() is only called for confirmed conntrack
entries (ct) within ovs_ct_commit(). However, if the conntrack entry
does not have the labels_ext extension, attempting to allocate it in
ovs_ct_get_conn_labels() for a confirmed entry triggers a warning in
nf_ct_ext_add():

  WARN_ON(nf_ct_is_confirmed(ct));

This happens when the conntrack entry is created externally before OVS
increments net->ct.labels_used. The issue has become more likely since
commit fcb1aa5163b1 ("openvswitch: switch to per-action label counting
in conntrack"), which changed to use per-action label counting and
increment net->ct.labels_used when a flow with ct action is added.

Since there’s no straightforward way to fully resolve this issue at the
moment, this reverts the commit to avoid breaking existing use cases.

Fixes: fcb1aa5163b1 ("openvswitch: switch to per-action label counting in conntrack")
Reported-by: Jianbo Liu <jianbol@nvidia.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 net/openvswitch/conntrack.c | 30 ++++++++++++++++++------------
 net/openvswitch/datapath.h  |  3 +++
 2 files changed, 21 insertions(+), 12 deletions(-)

diff --git a/net/openvswitch/conntrack.c b/net/openvswitch/conntrack.c
index 3bb4810234aa..e573e9221302 100644
--- a/net/openvswitch/conntrack.c
+++ b/net/openvswitch/conntrack.c
@@ -1368,8 +1368,11 @@ bool ovs_ct_verify(struct net *net, enum ovs_key_attr attr)
 	    attr == OVS_KEY_ATTR_CT_MARK)
 		return true;
 	if (IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS) &&
-	    attr == OVS_KEY_ATTR_CT_LABELS)
-		return true;
+	    attr == OVS_KEY_ATTR_CT_LABELS) {
+		struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
+
+		return ovs_net->xt_label;
+	}
 
 	return false;
 }
@@ -1378,7 +1381,6 @@ int ovs_ct_copy_action(struct net *net, const struct nlattr *attr,
 		       const struct sw_flow_key *key,
 		       struct sw_flow_actions **sfa,  bool log)
 {
-	unsigned int n_bits = sizeof(struct ovs_key_ct_labels) * BITS_PER_BYTE;
 	struct ovs_conntrack_info ct_info;
 	const char *helper = NULL;
 	u16 family;
@@ -1407,12 +1409,6 @@ int ovs_ct_copy_action(struct net *net, const struct nlattr *attr,
 		return -ENOMEM;
 	}
 
-	if (nf_connlabels_get(net, n_bits - 1)) {
-		nf_ct_tmpl_free(ct_info.ct);
-		OVS_NLERR(log, "Failed to set connlabel length");
-		return -EOPNOTSUPP;
-	}
-
 	if (ct_info.timeout[0]) {
 		if (nf_ct_set_timeout(net, ct_info.ct, family, key->ip.proto,
 				      ct_info.timeout))
@@ -1581,7 +1577,6 @@ static void __ovs_ct_free_action(struct ovs_conntrack_info *ct_info)
 	if (ct_info->ct) {
 		if (ct_info->timeout[0])
 			nf_ct_destroy_timeout(ct_info->ct);
-		nf_connlabels_put(nf_ct_net(ct_info->ct));
 		nf_ct_tmpl_free(ct_info->ct);
 	}
 }
@@ -2006,9 +2001,17 @@ struct genl_family dp_ct_limit_genl_family __ro_after_init = {
 
 int ovs_ct_init(struct net *net)
 {
-#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
+	unsigned int n_bits = sizeof(struct ovs_key_ct_labels) * BITS_PER_BYTE;
 	struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
 
+	if (nf_connlabels_get(net, n_bits - 1)) {
+		ovs_net->xt_label = false;
+		OVS_NLERR(true, "Failed to set connlabel length");
+	} else {
+		ovs_net->xt_label = true;
+	}
+
+#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
 	return ovs_ct_limit_init(net, ovs_net);
 #else
 	return 0;
@@ -2017,9 +2020,12 @@ int ovs_ct_init(struct net *net)
 
 void ovs_ct_exit(struct net *net)
 {
-#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
 	struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
 
+#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
 	ovs_ct_limit_exit(net, ovs_net);
 #endif
+
+	if (ovs_net->xt_label)
+		nf_connlabels_put(net);
 }
diff --git a/net/openvswitch/datapath.h b/net/openvswitch/datapath.h
index 365b9bb7f546..9ca6231ea647 100644
--- a/net/openvswitch/datapath.h
+++ b/net/openvswitch/datapath.h
@@ -160,6 +160,9 @@ struct ovs_net {
 #if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
 	struct ovs_ct_limit_info *ct_limit_info;
 #endif
+
+	/* Module reference for configuring conntrack. */
+	bool xt_label;
 };
 
 /**
-- 
2.47.1

Re: [ovs-dev] [PATCH net] Revert "openvswitch: switch to per-action label counting in conntrack"

[Aaron Conole]

Xin Long <lucien.xin@gmail.com> writes:

> Currently, ovs_ct_set_labels() is only called for confirmed conntrack
> entries (ct) within ovs_ct_commit(). However, if the conntrack entry
> does not have the labels_ext extension, attempting to allocate it in
> ovs_ct_get_conn_labels() for a confirmed entry triggers a warning in
> nf_ct_ext_add():
>
>   WARN_ON(nf_ct_is_confirmed(ct));
>
> This happens when the conntrack entry is created externally before OVS
> increments net->ct.labels_used. The issue has become more likely since
> commit fcb1aa5163b1 ("openvswitch: switch to per-action label counting
> in conntrack"), which changed to use per-action label counting and
> increment net->ct.labels_used when a flow with ct action is added.
>
> Since there’s no straightforward way to fully resolve this issue at the
> moment, this reverts the commit to avoid breaking existing use cases.
>
> Fixes: fcb1aa5163b1 ("openvswitch: switch to per-action label counting in conntrack")
> Reported-by: Jianbo Liu <jianbol@nvidia.com>
> Signed-off-by: Xin Long <lucien.xin@gmail.com>
> ---

I did a quick test using the case provided by Jianbo and I wasn't able
to generate the warning.  If possible, I'd like Jianbo to confirm that
it works as well.

Acked-by: Aaron Conole <aconole@redhat.com>

>  net/openvswitch/conntrack.c | 30 ++++++++++++++++++------------
>  net/openvswitch/datapath.h  |  3 +++
>  2 files changed, 21 insertions(+), 12 deletions(-)
>
> diff --git a/net/openvswitch/conntrack.c b/net/openvswitch/conntrack.c
> index 3bb4810234aa..e573e9221302 100644
> --- a/net/openvswitch/conntrack.c
> +++ b/net/openvswitch/conntrack.c
> @@ -1368,8 +1368,11 @@ bool ovs_ct_verify(struct net *net, enum ovs_key_attr attr)
>  	    attr == OVS_KEY_ATTR_CT_MARK)
>  		return true;
>  	if (IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS) &&
> -	    attr == OVS_KEY_ATTR_CT_LABELS)
> -		return true;
> +	    attr == OVS_KEY_ATTR_CT_LABELS) {
> +		struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
> +
> +		return ovs_net->xt_label;
> +	}
>  
>  	return false;
>  }
> @@ -1378,7 +1381,6 @@ int ovs_ct_copy_action(struct net *net, const struct nlattr *attr,
>  		       const struct sw_flow_key *key,
>  		       struct sw_flow_actions **sfa,  bool log)
>  {
> -	unsigned int n_bits = sizeof(struct ovs_key_ct_labels) * BITS_PER_BYTE;
>  	struct ovs_conntrack_info ct_info;
>  	const char *helper = NULL;
>  	u16 family;
> @@ -1407,12 +1409,6 @@ int ovs_ct_copy_action(struct net *net, const struct nlattr *attr,
>  		return -ENOMEM;
>  	}
>  
> -	if (nf_connlabels_get(net, n_bits - 1)) {
> -		nf_ct_tmpl_free(ct_info.ct);
> -		OVS_NLERR(log, "Failed to set connlabel length");
> -		return -EOPNOTSUPP;
> -	}
> -
>  	if (ct_info.timeout[0]) {
>  		if (nf_ct_set_timeout(net, ct_info.ct, family, key->ip.proto,
>  				      ct_info.timeout))
> @@ -1581,7 +1577,6 @@ static void __ovs_ct_free_action(struct ovs_conntrack_info *ct_info)
>  	if (ct_info->ct) {
>  		if (ct_info->timeout[0])
>  			nf_ct_destroy_timeout(ct_info->ct);
> -		nf_connlabels_put(nf_ct_net(ct_info->ct));
>  		nf_ct_tmpl_free(ct_info->ct);
>  	}
>  }
> @@ -2006,9 +2001,17 @@ struct genl_family dp_ct_limit_genl_family __ro_after_init = {
>  
>  int ovs_ct_init(struct net *net)
>  {
> -#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
> +	unsigned int n_bits = sizeof(struct ovs_key_ct_labels) * BITS_PER_BYTE;
>  	struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
>  
> +	if (nf_connlabels_get(net, n_bits - 1)) {
> +		ovs_net->xt_label = false;
> +		OVS_NLERR(true, "Failed to set connlabel length");
> +	} else {
> +		ovs_net->xt_label = true;
> +	}
> +
> +#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
>  	return ovs_ct_limit_init(net, ovs_net);
>  #else
>  	return 0;
> @@ -2017,9 +2020,12 @@ int ovs_ct_init(struct net *net)
>  
>  void ovs_ct_exit(struct net *net)
>  {
> -#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
>  	struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
>  
> +#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
>  	ovs_ct_limit_exit(net, ovs_net);
>  #endif
> +
> +	if (ovs_net->xt_label)
> +		nf_connlabels_put(net);
>  }
> diff --git a/net/openvswitch/datapath.h b/net/openvswitch/datapath.h
> index 365b9bb7f546..9ca6231ea647 100644
> --- a/net/openvswitch/datapath.h
> +++ b/net/openvswitch/datapath.h
> @@ -160,6 +160,9 @@ struct ovs_net {
>  #if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
>  	struct ovs_ct_limit_info *ct_limit_info;
>  #endif
> +
> +	/* Module reference for configuring conntrack. */
> +	bool xt_label;
>  };
>  
>  /**

Re: [ovs-dev] [PATCH net] Revert "openvswitch: switch to per-action label counting in conntrack"

[Jianbo Liu]

On 3/12/2025 10:25 PM, Aaron Conole wrote:
> Xin Long <lucien.xin@gmail.com> writes:
> 
>> Currently, ovs_ct_set_labels() is only called for confirmed conntrack
>> entries (ct) within ovs_ct_commit(). However, if the conntrack entry
>> does not have the labels_ext extension, attempting to allocate it in
>> ovs_ct_get_conn_labels() for a confirmed entry triggers a warning in
>> nf_ct_ext_add():
>>
>>    WARN_ON(nf_ct_is_confirmed(ct));
>>
>> This happens when the conntrack entry is created externally before OVS
>> increments net->ct.labels_used. The issue has become more likely since
>> commit fcb1aa5163b1 ("openvswitch: switch to per-action label counting
>> in conntrack"), which changed to use per-action label counting and
>> increment net->ct.labels_used when a flow with ct action is added.
>>
>> Since there’s no straightforward way to fully resolve this issue at the
>> moment, this reverts the commit to avoid breaking existing use cases.
>>
>> Fixes: fcb1aa5163b1 ("openvswitch: switch to per-action label counting in conntrack")
>> Reported-by: Jianbo Liu <jianbol@nvidia.com>
>> Signed-off-by: Xin Long <lucien.xin@gmail.com>
>> ---
> 
> I did a quick test using the case provided by Jianbo and I wasn't able
> to generate the warning.  If possible, I'd like Jianbo to confirm that
> it works as well.

Yes, it works. The warning is gone after applying this revert. Thanks!

> 
> Acked-by: Aaron Conole <aconole@redhat.com>
> 
>>   net/openvswitch/conntrack.c | 30 ++++++++++++++++++------------
>>   net/openvswitch/datapath.h  |  3 +++
>>   2 files changed, 21 insertions(+), 12 deletions(-)
>>
>> diff --git a/net/openvswitch/conntrack.c b/net/openvswitch/conntrack.c
>> index 3bb4810234aa..e573e9221302 100644
>> --- a/net/openvswitch/conntrack.c
>> +++ b/net/openvswitch/conntrack.c
>> @@ -1368,8 +1368,11 @@ bool ovs_ct_verify(struct net *net, enum ovs_key_attr attr)
>>   	    attr == OVS_KEY_ATTR_CT_MARK)
>>   		return true;
>>   	if (IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS) &&
>> -	    attr == OVS_KEY_ATTR_CT_LABELS)
>> -		return true;
>> +	    attr == OVS_KEY_ATTR_CT_LABELS) {
>> +		struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
>> +
>> +		return ovs_net->xt_label;
>> +	}
>>   
>>   	return false;
>>   }
>> @@ -1378,7 +1381,6 @@ int ovs_ct_copy_action(struct net *net, const struct nlattr *attr,
>>   		       const struct sw_flow_key *key,
>>   		       struct sw_flow_actions **sfa,  bool log)
>>   {
>> -	unsigned int n_bits = sizeof(struct ovs_key_ct_labels) * BITS_PER_BYTE;
>>   	struct ovs_conntrack_info ct_info;
>>   	const char *helper = NULL;
>>   	u16 family;
>> @@ -1407,12 +1409,6 @@ int ovs_ct_copy_action(struct net *net, const struct nlattr *attr,
>>   		return -ENOMEM;
>>   	}
>>   
>> -	if (nf_connlabels_get(net, n_bits - 1)) {
>> -		nf_ct_tmpl_free(ct_info.ct);
>> -		OVS_NLERR(log, "Failed to set connlabel length");
>> -		return -EOPNOTSUPP;
>> -	}
>> -
>>   	if (ct_info.timeout[0]) {
>>   		if (nf_ct_set_timeout(net, ct_info.ct, family, key->ip.proto,
>>   				      ct_info.timeout))
>> @@ -1581,7 +1577,6 @@ static void __ovs_ct_free_action(struct ovs_conntrack_info *ct_info)
>>   	if (ct_info->ct) {
>>   		if (ct_info->timeout[0])
>>   			nf_ct_destroy_timeout(ct_info->ct);
>> -		nf_connlabels_put(nf_ct_net(ct_info->ct));
>>   		nf_ct_tmpl_free(ct_info->ct);
>>   	}
>>   }
>> @@ -2006,9 +2001,17 @@ struct genl_family dp_ct_limit_genl_family __ro_after_init = {
>>   
>>   int ovs_ct_init(struct net *net)
>>   {
>> -#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
>> +	unsigned int n_bits = sizeof(struct ovs_key_ct_labels) * BITS_PER_BYTE;
>>   	struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
>>   
>> +	if (nf_connlabels_get(net, n_bits - 1)) {
>> +		ovs_net->xt_label = false;
>> +		OVS_NLERR(true, "Failed to set connlabel length");
>> +	} else {
>> +		ovs_net->xt_label = true;
>> +	}
>> +
>> +#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
>>   	return ovs_ct_limit_init(net, ovs_net);
>>   #else
>>   	return 0;
>> @@ -2017,9 +2020,12 @@ int ovs_ct_init(struct net *net)
>>   
>>   void ovs_ct_exit(struct net *net)
>>   {
>> -#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
>>   	struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
>>   
>> +#if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
>>   	ovs_ct_limit_exit(net, ovs_net);
>>   #endif
>> +
>> +	if (ovs_net->xt_label)
>> +		nf_connlabels_put(net);
>>   }
>> diff --git a/net/openvswitch/datapath.h b/net/openvswitch/datapath.h
>> index 365b9bb7f546..9ca6231ea647 100644
>> --- a/net/openvswitch/datapath.h
>> +++ b/net/openvswitch/datapath.h
>> @@ -160,6 +160,9 @@ struct ovs_net {
>>   #if	IS_ENABLED(CONFIG_NETFILTER_CONNCOUNT)
>>   	struct ovs_ct_limit_info *ct_limit_info;
>>   #endif
>> +
>> +	/* Module reference for configuring conntrack. */
>> +	bool xt_label;
>>   };
>>   
>>   /**
> 

Re: [PATCH net] Revert "openvswitch: switch to per-action label counting in conntrack"

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net.git (main)
by Paolo Abeni <pabeni@redhat.com>:

On Sat,  8 Mar 2025 13:05:43 -0500 you wrote:
> Currently, ovs_ct_set_labels() is only called for confirmed conntrack
> entries (ct) within ovs_ct_commit(). However, if the conntrack entry
> does not have the labels_ext extension, attempting to allocate it in
> ovs_ct_get_conn_labels() for a confirmed entry triggers a warning in
> nf_ct_ext_add():
> 
>   WARN_ON(nf_ct_is_confirmed(ct));
> 
> [...]

Here is the summary with links:
  - [net] Revert "openvswitch: switch to per-action label counting in conntrack"
    https://git.kernel.org/netdev/net/c/1063ae07383c

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html