[PATCH net v2] net/sched: cls_flower: exact-match ERSPAN key when no mask supplied

[PATCH net v2] net/sched: cls_flower: exact-match ERSPAN key when no mask supplied

[Jamal Hadi Salim]

Commit 292207809486 ("net: sched: fix erspan_opt settings in cls_flower")
removed the unconditional memset of the erspan_metadata mask in
fl_set_erspan_opt(). On the mask-parse pass, when userspace supplies the
encap option key but no mask (TCA_FLOWER_KEY_ENC_OPTS without
TCA_FLOWER_KEY_ENC_OPTS_MASK), the !depth early return leaves the mask
zeroed -- a wildcard. A filter keyed on an ERSPAN index (v1), or dir/hwid
(v2), then matches every ERSPAN packet's values instead of exact-matching.

Default the match fields on the no-mask path the same way the with-mask
branches default them: v1 index to 0xff, v2 dir and hwid to their
exact-match masks. The union cannot just be memset to 0xff -- that would
re-mask the overlapping timestamp/sgt bytes of the v2 header that the
commit deliberately left to 0.

A mask and a key that declare different ERSPAN versions describe different
union layouts, so the mask would silently wildcard the key's matched
fields. Reject that combination. Have fl_set_erspan_opt() report the
parsed key's version back through a u8 pointer so the mask pass uses it
without re-reading the union.

This is a follow-up to commit fee10655709c ("net/sched: cls_flower:
validate mask pointer after nla_next()").

Conditions to recreate the bug: add an ERSPAN encap-opts flower filter
with a key but no mask (raw netlink, TCA_FLOWER_KEY_ENC_OPTS present,
TCA_FLOWER_KEY_ENC_OPTS_MASK absent); dump the filter and observe the
match mask is zero (wildcard) instead of exact for the supplied fields.
For the mismatch case, pair a v1 key with a v2 mask (or vice versa) and
observe the request is accepted with a mask of the other version.

Fixes: 292207809486 ("net: sched: fix erspan_opt settings in cls_flower")
Reported-by: Sashiko (gemini) <sashiko-bot@kernel.org>
Closes: https://netdev-ai.bots.linux.dev/sashiko/#/patchset/20260826025123.62758-1-ljp1205831794@gmail.com
Link: https://lore.kernel.org/netdev/20260826025123.62758-1-ljp1205831794@gmail.com/
Reported-by: Sashiko (nipa) <sashiko-bot@kernel.org>
Link: https://lore.kernel.org/netdev/179003833955.2160803.1077681716391858800@kernel.org/
Suggested-by: Xin Long <lucien.xin@gmail.com>
Tested-by: hybris <hybris@mojatatu.ai>
Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
---
v1 -> v2:
- Scope the key's version default to the no-mask path only. v1 assigned
  the key's version to the mask unconditionally, which changed the
  with-mask parse path for a mask nest that declared no version (nipa M1).
- Reject a mask whose declared version differs from the key's; the two
  versions index different union layouts, so the mismatch silently
  wildcarded the key's matched fields (nipa M2).
- Set the key's parsed version through a u8 pointer argument rather than
  re-reading it from the union at the call site (Xin Long).

 net/sched/cls_flower.c | 39 ++++++++++++++++++++++++++++++++++-----
 1 file changed, 34 insertions(+), 5 deletions(-)

diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 1cefea571efd..0c4beff18d68 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -1358,17 +1358,29 @@ static int fl_set_vxlan_opt(const struct nlattr *nla, struct fl_flow_key *key,
 
 static int fl_set_erspan_opt(const struct nlattr *nla, struct fl_flow_key *key,
 			     int depth, int option_len,
-			     struct netlink_ext_ack *extack)
+			     struct netlink_ext_ack *extack, u8 *key_ver)
 {
 	struct nlattr *tb[TCA_FLOWER_KEY_ENC_OPT_ERSPAN_MAX + 1];
 	struct erspan_metadata *md;
 	int err;
 
 	md = (struct erspan_metadata *)&key->enc_opts.data[key->enc_opts.len];
-	md->version = 1;
 
-	if (!depth)
+	if (!depth) {
+		/* No mask attribute supplied: exact-match every field the
+		 * key set, keyed off the key's own version so the v2 dir/hwid
+		 * bytes are masked without touching the overlapping
+		 * timestamp/sgt bytes of the union (see 292207809486).
+		 */
+		md->version = *key_ver;
+		if (md->version == 1) {
+			memset(&md->u.index, 0xff, sizeof(md->u.index));
+		} else if (md->version == 2) {
+			md->u.md2.dir = 1;
+			set_hwid(&md->u.md2, 0xff);
+		}
 		return sizeof(*md);
+	}
 
 	if (nla_type(nla) != TCA_FLOWER_KEY_ENC_OPTS_ERSPAN) {
 		NL_SET_ERR_MSG(extack, "Non-erspan option type for mask");
@@ -1387,6 +1399,22 @@ static int fl_set_erspan_opt(const struct nlattr *nla, struct fl_flow_key *key,
 
 	if (tb[TCA_FLOWER_KEY_ENC_OPT_ERSPAN_VER])
 		md->version = nla_get_u8(tb[TCA_FLOWER_KEY_ENC_OPT_ERSPAN_VER]);
+	else
+		md->version = 1;
+
+	/* The key and mask must describe the same union layout: the
+	 * per-version default below only exact-matches the fields of its own
+	 * version, so a mask version that differs from the key's would
+	 * silently wildcard the key's matched fields.  A mask without a
+	 * version keeps the historical v1 default.
+	 */
+	if (option_len && md->version != *key_ver) {
+		NL_SET_ERR_MSG(extack, "Tunnel key erspan option ver mismatch");
+		return -EINVAL;
+	}
+
+	if (!option_len)
+		*key_ver = md->version;
 
 	if (md->version == 1) {
 		if (!option_len && !tb[TCA_FLOWER_KEY_ENC_OPT_ERSPAN_INDEX]) {
@@ -1509,6 +1537,7 @@ static int fl_set_enc_opt(struct nlattr **tb, struct fl_flow_key *key,
 {
 	const struct nlattr *nla_enc_key, *nla_opt_key, *nla_opt_msk = NULL;
 	int err, option_len, key_depth, msk_depth = 0;
+	u8 key_ver = 1;
 
 	err = nla_validate_nested_deprecated(tb[TCA_FLOWER_KEY_ENC_OPTS],
 					     TCA_FLOWER_KEY_ENC_OPTS_MAX,
@@ -1607,7 +1636,7 @@ static int fl_set_enc_opt(struct nlattr **tb, struct fl_flow_key *key,
 			key->enc_opts.dst_opt_type = IP_TUNNEL_ERSPAN_OPT_BIT;
 			option_len = fl_set_erspan_opt(nla_opt_key, key,
 						       key_depth, option_len,
-						       extack);
+						       extack, &key_ver);
 			if (option_len < 0)
 				return option_len;
 
@@ -1618,7 +1647,7 @@ static int fl_set_enc_opt(struct nlattr **tb, struct fl_flow_key *key,
 			mask->enc_opts.dst_opt_type = IP_TUNNEL_ERSPAN_OPT_BIT;
 			option_len = fl_set_erspan_opt(nla_opt_msk, mask,
 						       msk_depth, option_len,
-						       extack);
+						       extack, &key_ver);
 			if (option_len < 0)
 				return option_len;
 
-- 
2.43.0

Re: [PATCH net v2] net/sched: cls_flower: exact-match ERSPAN key when no mask supplied

[Simon Horman]

On Thu, Sep 24, 2026 at 08:18:41AM -0400, Jamal Hadi Salim wrote:
> Commit 292207809486 ("net: sched: fix erspan_opt settings in cls_flower")
> removed the unconditional memset of the erspan_metadata mask in
> fl_set_erspan_opt(). On the mask-parse pass, when userspace supplies the
> encap option key but no mask (TCA_FLOWER_KEY_ENC_OPTS without
> TCA_FLOWER_KEY_ENC_OPTS_MASK), the !depth early return leaves the mask
> zeroed -- a wildcard. A filter keyed on an ERSPAN index (v1), or dir/hwid
> (v2), then matches every ERSPAN packet's values instead of exact-matching.
> 
> Default the match fields on the no-mask path the same way the with-mask
> branches default them: v1 index to 0xff, v2 dir and hwid to their
> exact-match masks. The union cannot just be memset to 0xff -- that would
> re-mask the overlapping timestamp/sgt bytes of the v2 header that the
> commit deliberately left to 0.
> 
> A mask and a key that declare different ERSPAN versions describe different
> union layouts, so the mask would silently wildcard the key's matched
> fields. Reject that combination. Have fl_set_erspan_opt() report the
> parsed key's version back through a u8 pointer so the mask pass uses it
> without re-reading the union.
> 
> This is a follow-up to commit fee10655709c ("net/sched: cls_flower:
> validate mask pointer after nla_next()").
> 
> Conditions to recreate the bug: add an ERSPAN encap-opts flower filter
> with a key but no mask (raw netlink, TCA_FLOWER_KEY_ENC_OPTS present,
> TCA_FLOWER_KEY_ENC_OPTS_MASK absent); dump the filter and observe the
> match mask is zero (wildcard) instead of exact for the supplied fields.
> For the mismatch case, pair a v1 key with a v2 mask (or vice versa) and
> observe the request is accepted with a mask of the other version.
> 
> Fixes: 292207809486 ("net: sched: fix erspan_opt settings in cls_flower")
> Reported-by: Sashiko (gemini) <sashiko-bot@kernel.org>
> Closes: https://netdev-ai.bots.linux.dev/sashiko/#/patchset/20260826025123.62758-1-ljp1205831794@gmail.com
> Link: https://lore.kernel.org/netdev/20260826025123.62758-1-ljp1205831794@gmail.com/
> Reported-by: Sashiko (nipa) <sashiko-bot@kernel.org>
> Link: https://lore.kernel.org/netdev/179003833955.2160803.1077681716391858800@kernel.org/
> Suggested-by: Xin Long <lucien.xin@gmail.com>
> Tested-by: hybris <hybris@mojatatu.ai>
> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
> ---
> v1 -> v2:
> - Scope the key's version default to the no-mask path only. v1 assigned
>   the key's version to the mask unconditionally, which changed the
>   with-mask parse path for a mask nest that declared no version (nipa M1).
> - Reject a mask whose declared version differs from the key's; the two
>   versions index different union layouts, so the mismatch silently
>   wildcarded the key's matched fields (nipa M2).
> - Set the key's parsed version through a u8 pointer argument rather than
>   re-reading it from the union at the call site (Xin Long).

Reviewed-by: Simon Horman <horms@kernel.org>

Re: [PATCH net v2] net/sched: cls_flower: exact-match ERSPAN key when no mask supplied

[Xin Long]

On Thu, Sep 24, 2026 at 8:18 AM Jamal Hadi Salim <jhs@mojatatu.com> wrote:
>
> Commit 292207809486 ("net: sched: fix erspan_opt settings in cls_flower")
> removed the unconditional memset of the erspan_metadata mask in
> fl_set_erspan_opt(). On the mask-parse pass, when userspace supplies the
> encap option key but no mask (TCA_FLOWER_KEY_ENC_OPTS without
> TCA_FLOWER_KEY_ENC_OPTS_MASK), the !depth early return leaves the mask
> zeroed -- a wildcard. A filter keyed on an ERSPAN index (v1), or dir/hwid
> (v2), then matches every ERSPAN packet's values instead of exact-matching.
>
> Default the match fields on the no-mask path the same way the with-mask
> branches default them: v1 index to 0xff, v2 dir and hwid to their
> exact-match masks. The union cannot just be memset to 0xff -- that would
> re-mask the overlapping timestamp/sgt bytes of the v2 header that the
> commit deliberately left to 0.
>
> A mask and a key that declare different ERSPAN versions describe different
> union layouts, so the mask would silently wildcard the key's matched
> fields. Reject that combination. Have fl_set_erspan_opt() report the
> parsed key's version back through a u8 pointer so the mask pass uses it
> without re-reading the union.
>
> This is a follow-up to commit fee10655709c ("net/sched: cls_flower:
> validate mask pointer after nla_next()").
>
> Conditions to recreate the bug: add an ERSPAN encap-opts flower filter
> with a key but no mask (raw netlink, TCA_FLOWER_KEY_ENC_OPTS present,
> TCA_FLOWER_KEY_ENC_OPTS_MASK absent); dump the filter and observe the
> match mask is zero (wildcard) instead of exact for the supplied fields.
> For the mismatch case, pair a v1 key with a v2 mask (or vice versa) and
> observe the request is accepted with a mask of the other version.
>
> Fixes: 292207809486 ("net: sched: fix erspan_opt settings in cls_flower")
> Reported-by: Sashiko (gemini) <sashiko-bot@kernel.org>
> Closes: https://netdev-ai.bots.linux.dev/sashiko/#/patchset/20260826025123.62758-1-ljp1205831794@gmail.com
> Link: https://lore.kernel.org/netdev/20260826025123.62758-1-ljp1205831794@gmail.com/
> Reported-by: Sashiko (nipa) <sashiko-bot@kernel.org>
> Link: https://lore.kernel.org/netdev/179003833955.2160803.1077681716391858800@kernel.org/
> Suggested-by: Xin Long <lucien.xin@gmail.com>
> Tested-by: hybris <hybris@mojatatu.ai>
> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>

Reviewed-by: Xin Long <lucien.xin@gmail.com>

Re: [PATCH net v2] net/sched: cls_flower: exact-match ERSPAN key when no mask supplied

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Thu, 24 Sep 2026 08:18:41 -0400 you wrote:
> Commit 292207809486 ("net: sched: fix erspan_opt settings in cls_flower")
> removed the unconditional memset of the erspan_metadata mask in
> fl_set_erspan_opt(). On the mask-parse pass, when userspace supplies the
> encap option key but no mask (TCA_FLOWER_KEY_ENC_OPTS without
> TCA_FLOWER_KEY_ENC_OPTS_MASK), the !depth early return leaves the mask
> zeroed -- a wildcard. A filter keyed on an ERSPAN index (v1), or dir/hwid
> (v2), then matches every ERSPAN packet's values instead of exact-matching.
> 
> [...]

Here is the summary with links:
  - [net,v2] net/sched: cls_flower: exact-match ERSPAN key when no mask supplied
    https://git.kernel.org/netdev/net-next/c/e0edf25f141a

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