[PATCH net-next v2] net: sched: cls_api: improve the error message for ID allocation failure

[PATCH net-next v2] net: sched: cls_api: improve the error message for ID allocation failure

[Jakub Kicinski]

We run into an exhaustion problem with the kernel-allocated filter IDs.
Our allocation problem can be fixed on the user space side,
but the error message in this case was quite misleading:

  "Filter with specified priority/protocol not found" (EINVAL)

Specifically when we can't allocate a _new_ ID because filter with
lowest ID already _exists_, saying "filter not found", is confusing.

Kernel allocates IDs in range of 0xc0000 -> 0x8000, giving out ID one
lower than lowest existing in that range. The error message makes sense
when tcf_chain_tp_find() gets called for GET and DEL but for NEW we
need to provide more specific error messages for all three cases:

 - user wants the ID to be auto-allocated but filter with ID 0x8000
   already exists

 - filter already exists and can be replaced, but user asked
   for a protocol change

 - filter doesn't exist

Caller of tcf_chain_tp_insert_unique() doesn't set extack today,
so don't bother plumbing it in.

Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
v1 was sent a while back, I realized this is sitting in my tree
I think I just forgot to send v2.

v2:
 - don't add the extack if we don't find the filter during NEW
v1: https://lore.kernel.org/20240912215306.2060709-1-kuba@kernel.org

CC: jhs@mojatatu.com
CC: xiyou.wangcong@gmail.com
CC: jiri@resnulli.us
---
 net/sched/cls_api.c | 39 +++++++++++++++++++++++++--------------
 1 file changed, 25 insertions(+), 14 deletions(-)

diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 2a7d856cc334..5c8d39b6b107 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -1933,7 +1933,8 @@ static void tcf_chain_tp_remove(struct tcf_chain *chain,
 static struct tcf_proto *tcf_chain_tp_find(struct tcf_chain *chain,
 					   struct tcf_chain_info *chain_info,
 					   u32 protocol, u32 prio,
-					   bool prio_allocate);
+					   bool prio_allocate,
+					   struct netlink_ext_ack *extack);
 
 /* Try to insert new proto.
  * If proto with specified priority already exists, free new proto
@@ -1957,8 +1958,7 @@ static struct tcf_proto *tcf_chain_tp_insert_unique(struct tcf_chain *chain,
 		return ERR_PTR(-EAGAIN);
 	}
 
-	tp = tcf_chain_tp_find(chain, &chain_info,
-			       protocol, prio, false);
+	tp = tcf_chain_tp_find(chain, &chain_info, protocol, prio, false, NULL);
 	if (!tp)
 		err = tcf_chain_tp_insert(chain, &chain_info, tp_new);
 	mutex_unlock(&chain->filter_chain_lock);
@@ -2018,7 +2018,8 @@ static void tcf_chain_tp_delete_empty(struct tcf_chain *chain,
 static struct tcf_proto *tcf_chain_tp_find(struct tcf_chain *chain,
 					   struct tcf_chain_info *chain_info,
 					   u32 protocol, u32 prio,
-					   bool prio_allocate)
+					   bool prio_allocate,
+					   struct netlink_ext_ack *extack)
 {
 	struct tcf_proto **pprev;
 	struct tcf_proto *tp;
@@ -2029,9 +2030,14 @@ static struct tcf_proto *tcf_chain_tp_find(struct tcf_chain *chain,
 	     pprev = &tp->next) {
 		if (tp->prio >= prio) {
 			if (tp->prio == prio) {
-				if (prio_allocate ||
-				    (tp->protocol != protocol && protocol))
+				if (prio_allocate) {
+					NL_SET_ERR_MSG(extack, "Lowest ID from auto-alloc range already in use");
+					return ERR_PTR(-ENOSPC);
+				}
+				if (tp->protocol != protocol && protocol) {
+					NL_SET_ERR_MSG(extack, "Protocol mismatch for filter with specified priority");
 					return ERR_PTR(-EINVAL);
+				}
 			} else {
 				tp = NULL;
 			}
@@ -2312,9 +2318,8 @@ static int tc_new_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
 
 	mutex_lock(&chain->filter_chain_lock);
 	tp = tcf_chain_tp_find(chain, &chain_info, protocol,
-			       prio, prio_allocate);
+			       prio, prio_allocate, extack);
 	if (IS_ERR(tp)) {
-		NL_SET_ERR_MSG(extack, "Filter with specified priority/protocol not found");
 		err = PTR_ERR(tp);
 		goto errout_locked;
 	}
@@ -2539,10 +2544,13 @@ static int tc_del_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
 
 	mutex_lock(&chain->filter_chain_lock);
 	tp = tcf_chain_tp_find(chain, &chain_info, protocol,
-			       prio, false);
-	if (!tp || IS_ERR(tp)) {
+			       prio, false, extack);
+	if (!tp) {
+		err = -ENOENT;
 		NL_SET_ERR_MSG(extack, "Filter with specified priority/protocol not found");
-		err = tp ? PTR_ERR(tp) : -ENOENT;
+		goto errout_locked;
+	} else if (IS_ERR(tp)) {
+		err = PTR_ERR(tp);
 		goto errout_locked;
 	} else if (tca[TCA_KIND] && nla_strcmp(tca[TCA_KIND], tp->ops->kind)) {
 		NL_SET_ERR_MSG(extack, "Specified filter kind does not match existing one");
@@ -2679,11 +2687,14 @@ static int tc_get_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
 
 	mutex_lock(&chain->filter_chain_lock);
 	tp = tcf_chain_tp_find(chain, &chain_info, protocol,
-			       prio, false);
+			       prio, false, extack);
 	mutex_unlock(&chain->filter_chain_lock);
-	if (!tp || IS_ERR(tp)) {
+	if (!tp) {
+		err = -ENOENT;
 		NL_SET_ERR_MSG(extack, "Filter with specified priority/protocol not found");
-		err = tp ? PTR_ERR(tp) : -ENOENT;
+		goto errout;
+	} else if (IS_ERR(tp)) {
+		err = PTR_ERR(tp);
 		goto errout;
 	} else if (tca[TCA_KIND] && nla_strcmp(tca[TCA_KIND], tp->ops->kind)) {
 		NL_SET_ERR_MSG(extack, "Specified filter kind does not match existing one");
-- 
2.47.0

Re: [PATCH net-next v2] net: sched: cls_api: improve the error message for ID allocation failure

[Jamal Hadi Salim]

On Thu, Nov 7, 2024 at 8:02 PM Jakub Kicinski <kuba@kernel.org> wrote:
>
> We run into an exhaustion problem with the kernel-allocated filter IDs.
> Our allocation problem can be fixed on the user space side,
> but the error message in this case was quite misleading:
>
>   "Filter with specified priority/protocol not found" (EINVAL)
>
> Specifically when we can't allocate a _new_ ID because filter with
> lowest ID already _exists_, saying "filter not found", is confusing.
>
> Kernel allocates IDs in range of 0xc0000 -> 0x8000, giving out ID one
> lower than lowest existing in that range. The error message makes sense
> when tcf_chain_tp_find() gets called for GET and DEL but for NEW we
> need to provide more specific error messages for all three cases:
>
>  - user wants the ID to be auto-allocated but filter with ID 0x8000
>    already exists
>
>  - filter already exists and can be replaced, but user asked
>    for a protocol change
>
>  - filter doesn't exist
>
> Caller of tcf_chain_tp_insert_unique() doesn't set extack today,
> so don't bother plumbing it in.
>
> Reviewed-by: Simon Horman <horms@kernel.org>
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>

cheers,
jamal

Re: [PATCH net-next v2] net: sched: cls_api: improve the error message for ID allocation failure

[patchwork-bot+netdevbpf]

Hello:

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

On Thu,  7 Nov 2024 17:02:54 -0800 you wrote:
> We run into an exhaustion problem with the kernel-allocated filter IDs.
> Our allocation problem can be fixed on the user space side,
> but the error message in this case was quite misleading:
> 
>   "Filter with specified priority/protocol not found" (EINVAL)
> 
> Specifically when we can't allocate a _new_ ID because filter with
> lowest ID already _exists_, saying "filter not found", is confusing.
> 
> [...]

Here is the summary with links:
  - [net-next,v2] net: sched: cls_api: improve the error message for ID allocation failure
    https://git.kernel.org/netdev/net-next/c/a58f00ed24b8

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