[PATCH 1/3] Fix overread in PREQ frame processing

[PATCH 1/3] Fix overread in PREQ frame processing

[Masashi Honma]

When the AF flag is enabled, hwmp_preq_frame_process() overreads target_addr
by 2 bytes. Since this occurs within the socket buffer, it does not read across
memory boundaries and therefore poses no security risk; however, we will fix it
as a precaution.

In this fix, a new function mesh_path_parse_request_frame() is established to
separate the implementation of frame format validation and the check for
unsupported features. This is intended to facilitate future work when
implementing the currently unsupported parts.

Assisted-by: Claude:Sonnet 4.6
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
---
 net/mac80211/mesh.h      |  1 +
 net/mac80211/mesh_hwmp.c | 35 +++++++++++++++++++++++++++++++++--
 net/mac80211/parse.c     |  7 +++++--
 3 files changed, 39 insertions(+), 4 deletions(-)

diff --git a/net/mac80211/mesh.h b/net/mac80211/mesh.h
index 3f9664e4e00c..331d2d774196 100644
--- a/net/mac80211/mesh.h
+++ b/net/mac80211/mesh.h
@@ -317,6 +317,7 @@ struct mesh_path *
 mpp_path_lookup_by_idx(struct ieee80211_sub_if_data *sdata, int idx);
 void mesh_path_fix_nexthop(struct mesh_path *mpath, struct sta_info *next_hop);
 void mesh_path_expire(struct ieee80211_sub_if_data *sdata);
+u8 mesh_path_parse_request_frame(const u8 *pos, u8 elen);
 void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 			    struct ieee80211_mgmt *mgmt, size_t len);
 struct mesh_path *
diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c
index 9d89ebcce1c1..4239412d12c7 100644
--- a/net/mac80211/mesh_hwmp.c
+++ b/net/mac80211/mesh_hwmp.c
@@ -45,6 +45,7 @@ static inline u16 u16_field_get(const u8 *preq_elem, int offset, bool ae)
 #define PREQ_IE_ORIG_SN(x)	u32_field_get(x, 13, 0)
 #define PREQ_IE_LIFETIME(x)	u32_field_get(x, 17, AE_F_SET(x))
 #define PREQ_IE_METRIC(x) 	u32_field_get(x, 21, AE_F_SET(x))
+#define PREQ_IE_TARGET_COUNT(x)	(*(AE_F_SET(x) ? x + 31 : x + 25))
 #define PREQ_IE_TARGET_F(x)	(*(AE_F_SET(x) ? x + 32 : x + 26))
 #define PREQ_IE_TARGET_ADDR(x) 	(AE_F_SET(x) ? x + 33 : x + 27)
 #define PREQ_IE_TARGET_SN(x) 	u32_field_get(x, 33, AE_F_SET(x))
@@ -924,6 +925,29 @@ static void hwmp_rann_frame_process(struct ieee80211_sub_if_data *sdata,
 	rcu_read_unlock();
 }
 
+/* IEEE Std 802.11-2016 9.4.2.113 PREQ element */
+u8 mesh_path_parse_request_frame(const u8 *pos, u8 elen)
+{
+	u8 target_count;
+	u8 expected_len;
+
+	target_count = PREQ_IE_TARGET_COUNT(pos);
+	if (unlikely(target_count < 1 || target_count > 20))
+		return IEEE80211_PARSE_ERR_UNEXPECTED_ELEM;
+
+	expected_len = 1 /* Flags */ + 1 /* Hop Count */ + 1 /* Element TTL */ +
+		       4 /* Path Discovery ID */ +
+		       6 /* Originator Mesh STA Address */ +
+		       4 /* Originator HWMP Sequence Number */ +
+		       (AE_F_SET(pos) ? 6 : 0) /* Originator External Address */ +
+		       4 /* Lifetime */ + 4 /* Metric */ + 1 + /* Target Count */ +
+		       target_count * (1 /* Per Target Flags */ +
+		       6 /* Target Address */ + 4 /* Target HWMP Sequence Number */);
+	if (unlikely(elen != expected_len))
+		return IEEE80211_PARSE_ERR_BAD_ELEM_SIZE;
+
+	return 0;
+}
 
 void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 			    struct ieee80211_mgmt *mgmt, size_t len)
@@ -932,6 +956,7 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 	size_t baselen;
 	u32 path_metric;
 	struct sta_info *sta;
+	u8 target_count;
 
 	/* need action_code */
 	if (len < IEEE80211_MIN_ACTION_SIZE(mesh_action))
@@ -955,9 +980,15 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 		return;
 
 	if (elems->preq) {
-		if (elems->preq_len != 37)
-			/* Right now we support just 1 destination and no AE */
+		/* Right now we do not support AE (Address Extension) */
+		if (AE_F_SET(elems->preq))
 			goto free;
+
+		/* Right now we only supports 1 target */
+		target_count = PREQ_IE_TARGET_COUNT(elems->preq);
+		if (target_count != 1)
+			goto free;
+
 		path_metric = hwmp_route_info_get(sdata, mgmt, elems->preq,
 						  MPATH_PREQ);
 		if (path_metric)
diff --git a/net/mac80211/parse.c b/net/mac80211/parse.c
index 5e61457be0f3..7a2abe676361 100644
--- a/net/mac80211/parse.c
+++ b/net/mac80211/parse.c
@@ -547,8 +547,11 @@ _ieee802_11_parse_elems_full(struct ieee80211_elems_parse_params *params,
 				elems->awake_window = (void *)pos;
 			break;
 		case WLAN_EID_PREQ:
-			elems->preq = pos;
-			elems->preq_len = elen;
+			elem_parse_failed = mesh_path_parse_request_frame(pos, elen);
+			if (likely(!elem_parse_failed)) {
+				elems->preq = pos;
+				elems->preq_len = elen;
+			}
 			break;
 		case WLAN_EID_PREP:
 			elems->prep = pos;
-- 
2.43.0

[PATCH 2/3] Fix overread in PREP frame processing

[Masashi Honma]

When the AF flag is enabled, hwmp_prep_frame_process() overreads orig_addr
by 2 bytes. Since this occurs within the socket buffer, it does not read across
memory boundaries and therefore poses no security risk; however, we will fix it
as a precaution.

In this fix, a new function mesh_path_parse_reply_frame() is established to
separate the implementation of frame format validation and the check for
unsupported features. This is intended to facilitate future work when
implementing the currently unsupported parts.

Assisted-by: Claude:Sonnet 4.6
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
---
 net/mac80211/mesh.h      |  1 +
 net/mac80211/mesh_hwmp.c | 24 ++++++++++++++++++++++--
 net/mac80211/parse.c     |  7 +++++--
 3 files changed, 28 insertions(+), 4 deletions(-)

diff --git a/net/mac80211/mesh.h b/net/mac80211/mesh.h
index 331d2d774196..aba7ef59195a 100644
--- a/net/mac80211/mesh.h
+++ b/net/mac80211/mesh.h
@@ -318,6 +318,7 @@ mpp_path_lookup_by_idx(struct ieee80211_sub_if_data *sdata, int idx);
 void mesh_path_fix_nexthop(struct mesh_path *mpath, struct sta_info *next_hop);
 void mesh_path_expire(struct ieee80211_sub_if_data *sdata);
 u8 mesh_path_parse_request_frame(const u8 *pos, u8 elen);
+u8 mesh_path_parse_reply_frame(const u8 *pos, u8 elen);
 void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 			    struct ieee80211_mgmt *mgmt, size_t len);
 struct mesh_path *
diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c
index 4239412d12c7..9f903a2408f7 100644
--- a/net/mac80211/mesh_hwmp.c
+++ b/net/mac80211/mesh_hwmp.c
@@ -949,6 +949,25 @@ u8 mesh_path_parse_request_frame(const u8 *pos, u8 elen)
 	return 0;
 }
 
+/* IEEE Std 802.11-2016 9.4.2.114 PREP element */
+u8 mesh_path_parse_reply_frame(const u8 *pos, u8 elen)
+{
+	u8 expected_len;
+
+	expected_len = 1 /* Flags */ + 1 /* Hop Count */ + 1 /* Element TTL */ +
+		       6 /* Target Mesh STA Address */ +
+		       4 /* Target HWMP Sequence Number */ +
+		       (AE_F_SET(pos) ? 6 : 0) /* Target External Address */ +
+		       4 /* Lifetime */ + 4 /* Metric */ +
+		       6 /* Originator Mesh STA Address */ +
+		       4 /* Originator HWMP Sequence Number */;
+
+	if (unlikely(elen != expected_len))
+		return IEEE80211_PARSE_ERR_BAD_ELEM_SIZE;
+
+	return 0;
+}
+
 void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 			    struct ieee80211_mgmt *mgmt, size_t len)
 {
@@ -996,9 +1015,10 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 						path_metric);
 	}
 	if (elems->prep) {
-		if (elems->prep_len != 31)
-			/* Right now we support no AE */
+		/* Right now we do not support AE (Address Extension) */
+		if (AE_F_SET(elems->prep))
 			goto free;
+
 		path_metric = hwmp_route_info_get(sdata, mgmt, elems->prep,
 						  MPATH_PREP);
 		if (path_metric)
diff --git a/net/mac80211/parse.c b/net/mac80211/parse.c
index 7a2abe676361..cdc549055242 100644
--- a/net/mac80211/parse.c
+++ b/net/mac80211/parse.c
@@ -554,8 +554,11 @@ _ieee802_11_parse_elems_full(struct ieee80211_elems_parse_params *params,
 			}
 			break;
 		case WLAN_EID_PREP:
-			elems->prep = pos;
-			elems->prep_len = elen;
+			elem_parse_failed = mesh_path_parse_reply_frame(pos, elen);
+			if (likely(!elem_parse_failed)) {
+				elems->prep = pos;
+				elems->prep_len = elen;
+			}
 			break;
 		case WLAN_EID_PERR:
 			elems->perr = pos;
-- 
2.43.0

[PATCH 3/3] Fix PERR frame processing

[Masashi Honma]

There are no issues with the PERR processing itself; however, to maintain
consistency with the previous PREQ/PREP code modifications, I will create a new
mesh_path_parse_error_frame() function to separately implement the frame format
validation and the "not supported" check.

Assisted-by: Claude:Sonnet 4.6
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
---
 net/mac80211/mesh.h      |  1 +
 net/mac80211/mesh_hwmp.c | 48 ++++++++++++++++++++++++++++++++++++++--
 net/mac80211/parse.c     |  7 ++++--
 3 files changed, 52 insertions(+), 4 deletions(-)

diff --git a/net/mac80211/mesh.h b/net/mac80211/mesh.h
index aba7ef59195a..610c34e35e64 100644
--- a/net/mac80211/mesh.h
+++ b/net/mac80211/mesh.h
@@ -319,6 +319,7 @@ void mesh_path_fix_nexthop(struct mesh_path *mpath, struct sta_info *next_hop);
 void mesh_path_expire(struct ieee80211_sub_if_data *sdata);
 u8 mesh_path_parse_request_frame(const u8 *pos, u8 elen);
 u8 mesh_path_parse_reply_frame(const u8 *pos, u8 elen);
+u8 mesh_path_parse_error_frame(const u8 *pos, u8 elen);
 void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 			    struct ieee80211_mgmt *mgmt, size_t len);
 struct mesh_path *
diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c
index 9f903a2408f7..2a293b771f79 100644
--- a/net/mac80211/mesh_hwmp.c
+++ b/net/mac80211/mesh_hwmp.c
@@ -62,6 +62,7 @@ static inline u16 u16_field_get(const u8 *preq_elem, int offset, bool ae)
 #define PREP_IE_TARGET_SN(x)	u32_field_get(x, 9, 0)
 
 #define PERR_IE_TTL(x)		(*(x))
+#define PERR_IE_NUMBER_OF_DST(x)	(*(x + 1))
 #define PERR_IE_TARGET_FLAGS(x)	(*(x + 2))
 #define PERR_IE_TARGET_ADDR(x)	(x + 3)
 #define PERR_IE_TARGET_SN(x)	u32_field_get(x, 9, 0)
@@ -968,6 +969,42 @@ u8 mesh_path_parse_reply_frame(const u8 *pos, u8 elen)
 	return 0;
 }
 
+/* IEEE Std 802.11-2016 9.4.2.115 PERR element */
+u8 mesh_path_parse_error_frame(const u8 *pos, u8 elen)
+{
+	u8 number_of_dst;
+	u8 expected_len;
+	const u8 *start;
+	int i;
+
+	number_of_dst = PERR_IE_NUMBER_OF_DST(pos);
+	if (unlikely(number_of_dst < 1 || number_of_dst > 19))
+		return IEEE80211_PARSE_ERR_UNEXPECTED_ELEM;
+
+	start = pos;
+	expected_len = 1 /* Element TTL */ + 1 /* Number of Destinations */;
+	pos += 2;
+
+	for (i = 0; i < number_of_dst; i++) {
+		u8 dst_len;
+
+		if (unlikely(pos - start >= elen))
+			return IEEE80211_PARSE_ERR_BAD_ELEM_SIZE;
+
+		dst_len = 1 /* Flags */ + 6 /* Destination Address */ +
+			  4 /* HWMP Sequence Number */ +
+			  (AE_F_SET(pos) ? 6 : 0) /* Destination External Address */ +
+			  2 /* Reason Code */;
+		expected_len += dst_len;
+		pos += dst_len;
+	}
+
+	if (unlikely(elen != expected_len))
+		return IEEE80211_PARSE_ERR_BAD_ELEM_SIZE;
+
+	return 0;
+}
+
 void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 			    struct ieee80211_mgmt *mgmt, size_t len)
 {
@@ -976,6 +1013,7 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 	u32 path_metric;
 	struct sta_info *sta;
 	u8 target_count;
+	u8 number_of_dst;
 
 	/* need action_code */
 	if (len < IEEE80211_MIN_ACTION_SIZE(mesh_action))
@@ -1026,9 +1064,15 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 						path_metric);
 	}
 	if (elems->perr) {
-		if (elems->perr_len != 15)
-			/* Right now we support only one destination per PERR */
+		/* Right now we support only one destination per PERR */
+		number_of_dst = PERR_IE_NUMBER_OF_DST(elems->perr);
+		if (number_of_dst != 1)
 			goto free;
+
+		/* Right now we do not support AE (Address Extension) */
+		if (PERR_IE_TARGET_FLAGS(elems->perr) & AE_F)
+			goto free;
+
 		hwmp_perr_frame_process(sdata, mgmt, elems->perr);
 	}
 	if (elems->rann)
diff --git a/net/mac80211/parse.c b/net/mac80211/parse.c
index cdc549055242..dfc326db0471 100644
--- a/net/mac80211/parse.c
+++ b/net/mac80211/parse.c
@@ -561,8 +561,11 @@ _ieee802_11_parse_elems_full(struct ieee80211_elems_parse_params *params,
 			}
 			break;
 		case WLAN_EID_PERR:
-			elems->perr = pos;
-			elems->perr_len = elen;
+			elem_parse_failed = mesh_path_parse_error_frame(pos, elen);
+			if (likely(!elem_parse_failed)) {
+				elems->perr = pos;
+				elems->perr_len = elen;
+			}
 			break;
 		case WLAN_EID_RANN:
 			if (elen >= sizeof(struct ieee80211_rann_ie))
-- 
2.43.0

Re: [PATCH 1/3] Fix overread in PREQ frame processing

[Johannes Berg]

On Sat, 2026-05-09 at 07:59 +0900, Masashi Honma wrote:
>  
> +/* IEEE Std 802.11-2016 9.4.2.113 PREQ element */
> +u8 mesh_path_parse_request_frame(const u8 *pos, u8 elen)
> +{
> +	u8 target_count;
> +	u8 expected_len;
> +
> +	target_count = PREQ_IE_TARGET_COUNT(pos);
> +	if (unlikely(target_count < 1 || target_count > 20))
> +		return IEEE80211_PARSE_ERR_UNEXPECTED_ELEM;

not sure the likely/unlikely really is worth anything there - we don't
process *that* many frames.

> +
> +	expected_len = 1 /* Flags */ + 1 /* Hop Count */ + 1 /* Element TTL */ +
> +		       4 /* Path Discovery ID */ +
> +		       6 /* Originator Mesh STA Address */ +
> +		       4 /* Originator HWMP Sequence Number */ +
> +		       (AE_F_SET(pos) ? 6 : 0) /* Originator External Address */ +
> +		       4 /* Lifetime */ + 4 /* Metric */ + 1 + /* Target Count */ +
> +		       target_count * (1 /* Per Target Flags */ +
> +		       6 /* Target Address */ + 4 /* Target HWMP Sequence Number */);
> +	if (unlikely(elen != expected_len))
> +		return IEEE80211_PARSE_ERR_BAD_ELEM_SIZE;

I think there's a case to be made for doing this like many others, in an
inline in ieee80211-mesh.h, and calling it ..._size_ok() with a bool
return.

> +++ b/net/mac80211/parse.c
> @@ -547,8 +547,11 @@ _ieee802_11_parse_elems_full(struct ieee80211_elems_parse_params *params,
>  				elems->awake_window = (void *)pos;
>  			break;
>  		case WLAN_EID_PREQ:
> -			elems->preq = pos;
> -			elems->preq_len = elen;
> +			elem_parse_failed = mesh_path_parse_request_frame(pos, elen);
> +			if (likely(!elem_parse_failed)) {

that would also fix the build issue here if mesh isn't compiled in.

(and also here, not sure about likely/unlikely, doesn't really seem
worth it, and under attack maybe the failure becomes likely?)

johannes

[PATCH 1/3] Fix overread in PREQ frame processing

[Masashi Honma]

When the AF flag is enabled, hwmp_preq_frame_process() overreads target_addr
by 2 bytes. Since this occurs within the socket buffer, it does not read across
memory boundaries and therefore poses no security risk; however, we will fix it
as a precaution.

In this fix, a new function mesh_path_parse_request_frame() is established to
separate the implementation of frame format validation and the check for
unsupported features. This is intended to facilitate future work when
implementing the currently unsupported parts.

Assisted-by: Claude:Sonnet 4.6
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
---
 include/linux/ieee80211-mesh.h | 56 ++++++++++++++++++++++++++++++++++
 net/mac80211/mesh_hwmp.c       | 42 ++++++-------------------
 net/mac80211/parse.c           |  9 ++++--
 3 files changed, 72 insertions(+), 35 deletions(-)

diff --git a/include/linux/ieee80211-mesh.h b/include/linux/ieee80211-mesh.h
index 4b829bcb38b6..8ff68cf539ff 100644
--- a/include/linux/ieee80211-mesh.h
+++ b/include/linux/ieee80211-mesh.h
@@ -34,6 +34,38 @@ struct ieee80211s_hdr {
 #define MESH_FLAGS_AE		0x3
 #define MESH_FLAGS_PS_DEEP	0x4
 
+/* HWMP IE processing macros */
+#define AE_F			(1<<6)
+#define AE_F_SET(x)		(*x & AE_F)
+#define PREQ_IE_FLAGS(x)	(*(x))
+#define PREQ_IE_HOPCOUNT(x)	(*(x + 1))
+#define PREQ_IE_TTL(x)		(*(x + 2))
+#define PREQ_IE_PREQ_ID(x)	u32_field_get(x, 3, 0)
+#define PREQ_IE_ORIG_ADDR(x)	(x + 7)
+#define PREQ_IE_ORIG_SN(x)	u32_field_get(x, 13, 0)
+#define PREQ_IE_LIFETIME(x)	u32_field_get(x, 17, AE_F_SET(x))
+#define PREQ_IE_METRIC(x) 	u32_field_get(x, 21, AE_F_SET(x))
+#define PREQ_IE_TARGET_COUNT(x)	(*(AE_F_SET(x) ? x + 31 : x + 25))
+#define PREQ_IE_TARGET_F(x)	(*(AE_F_SET(x) ? x + 32 : x + 26))
+#define PREQ_IE_TARGET_ADDR(x) 	(AE_F_SET(x) ? x + 33 : x + 27)
+#define PREQ_IE_TARGET_SN(x) 	u32_field_get(x, 33, AE_F_SET(x))
+
+#define PREP_IE_FLAGS(x)	PREQ_IE_FLAGS(x)
+#define PREP_IE_HOPCOUNT(x)	PREQ_IE_HOPCOUNT(x)
+#define PREP_IE_TTL(x)		PREQ_IE_TTL(x)
+#define PREP_IE_ORIG_ADDR(x)	(AE_F_SET(x) ? x + 27 : x + 21)
+#define PREP_IE_ORIG_SN(x)	u32_field_get(x, 27, AE_F_SET(x))
+#define PREP_IE_LIFETIME(x)	u32_field_get(x, 13, AE_F_SET(x))
+#define PREP_IE_METRIC(x)	u32_field_get(x, 17, AE_F_SET(x))
+#define PREP_IE_TARGET_ADDR(x)	(x + 3)
+#define PREP_IE_TARGET_SN(x)	u32_field_get(x, 9, 0)
+
+#define PERR_IE_TTL(x)			(*(x))
+#define PERR_IE_TARGET_FLAGS(x)		(*(x + 2))
+#define PERR_IE_TARGET_ADDR(x)		(x + 3)
+#define PERR_IE_TARGET_SN(x)		u32_field_get(x, 9, 0)
+#define PERR_IE_TARGET_RCODE(x)		u16_field_get(x, 13, 0)
+
 /**
  * enum ieee80211_preq_flags - mesh PREQ element flags
  *
@@ -227,4 +259,28 @@ enum ieee80211_root_mode_identifier {
 	IEEE80211_PROACTIVE_RANN = 4,
 };
 
+/* IEEE Std 802.11-2016 9.4.2.113 PREQ element */
+static inline bool ieee80211_mesh_preq_size_ok(const u8 *pos, u8 elen)
+{
+	u8 target_count;
+	u8 needed;
+
+	target_count = PREQ_IE_TARGET_COUNT(pos);
+	if (target_count < 1 || target_count > 20)
+		return false;
+
+	needed = 1 /* Flags */ + 1 /* Hop Count */ + 1 /* Element TTL */ +
+		 4 /* Path Discovery ID */ +
+		 6 /* Originator Mesh STA Address */ +
+		 4 /* Originator HWMP Sequence Number */ +
+		 (AE_F_SET(pos) ? 6 : 0) /* Originator External Address */ +
+		 4 /* Lifetime */ + 4 /* Metric */ + 1 + /* Target Count */ +
+		 target_count * (1 /* Per Target Flags */ +
+		 6 /* Target Address */ + 4 /* Target HWMP Sequence Number */);
+	if (elen != needed)
+		return false;
+
+	return true;
+}
+
 #endif /* LINUX_IEEE80211_MESH_H */
diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c
index 9d89ebcce1c1..a4984b015995 100644
--- a/net/mac80211/mesh_hwmp.c
+++ b/net/mac80211/mesh_hwmp.c
@@ -35,37 +35,6 @@ static inline u16 u16_field_get(const u8 *preq_elem, int offset, bool ae)
 }
 
 /* HWMP IE processing macros */
-#define AE_F			(1<<6)
-#define AE_F_SET(x)		(*x & AE_F)
-#define PREQ_IE_FLAGS(x)	(*(x))
-#define PREQ_IE_HOPCOUNT(x)	(*(x + 1))
-#define PREQ_IE_TTL(x)		(*(x + 2))
-#define PREQ_IE_PREQ_ID(x)	u32_field_get(x, 3, 0)
-#define PREQ_IE_ORIG_ADDR(x)	(x + 7)
-#define PREQ_IE_ORIG_SN(x)	u32_field_get(x, 13, 0)
-#define PREQ_IE_LIFETIME(x)	u32_field_get(x, 17, AE_F_SET(x))
-#define PREQ_IE_METRIC(x) 	u32_field_get(x, 21, AE_F_SET(x))
-#define PREQ_IE_TARGET_F(x)	(*(AE_F_SET(x) ? x + 32 : x + 26))
-#define PREQ_IE_TARGET_ADDR(x) 	(AE_F_SET(x) ? x + 33 : x + 27)
-#define PREQ_IE_TARGET_SN(x) 	u32_field_get(x, 33, AE_F_SET(x))
-
-
-#define PREP_IE_FLAGS(x)	PREQ_IE_FLAGS(x)
-#define PREP_IE_HOPCOUNT(x)	PREQ_IE_HOPCOUNT(x)
-#define PREP_IE_TTL(x)		PREQ_IE_TTL(x)
-#define PREP_IE_ORIG_ADDR(x)	(AE_F_SET(x) ? x + 27 : x + 21)
-#define PREP_IE_ORIG_SN(x)	u32_field_get(x, 27, AE_F_SET(x))
-#define PREP_IE_LIFETIME(x)	u32_field_get(x, 13, AE_F_SET(x))
-#define PREP_IE_METRIC(x)	u32_field_get(x, 17, AE_F_SET(x))
-#define PREP_IE_TARGET_ADDR(x)	(x + 3)
-#define PREP_IE_TARGET_SN(x)	u32_field_get(x, 9, 0)
-
-#define PERR_IE_TTL(x)		(*(x))
-#define PERR_IE_TARGET_FLAGS(x)	(*(x + 2))
-#define PERR_IE_TARGET_ADDR(x)	(x + 3)
-#define PERR_IE_TARGET_SN(x)	u32_field_get(x, 9, 0)
-#define PERR_IE_TARGET_RCODE(x)	u16_field_get(x, 13, 0)
-
 #define MSEC_TO_TU(x) (x*1000/1024)
 #define SN_GT(x, y) ((s32)(y - x) < 0)
 #define SN_LT(x, y) ((s32)(x - y) < 0)
@@ -932,6 +901,7 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 	size_t baselen;
 	u32 path_metric;
 	struct sta_info *sta;
+	u8 target_count;
 
 	/* need action_code */
 	if (len < IEEE80211_MIN_ACTION_SIZE(mesh_action))
@@ -955,9 +925,15 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 		return;
 
 	if (elems->preq) {
-		if (elems->preq_len != 37)
-			/* Right now we support just 1 destination and no AE */
+		/* Right now we do not support AE (Address Extension) */
+		if (AE_F_SET(elems->preq))
+			goto free;
+
+		/* Right now we only supports 1 target */
+		target_count = PREQ_IE_TARGET_COUNT(elems->preq);
+		if (target_count != 1)
 			goto free;
+
 		path_metric = hwmp_route_info_get(sdata, mgmt, elems->preq,
 						  MPATH_PREQ);
 		if (path_metric)
diff --git a/net/mac80211/parse.c b/net/mac80211/parse.c
index 5e61457be0f3..9e52cc48fc18 100644
--- a/net/mac80211/parse.c
+++ b/net/mac80211/parse.c
@@ -547,8 +547,13 @@ _ieee802_11_parse_elems_full(struct ieee80211_elems_parse_params *params,
 				elems->awake_window = (void *)pos;
 			break;
 		case WLAN_EID_PREQ:
-			elems->preq = pos;
-			elems->preq_len = elen;
+			if (ieee80211_mesh_preq_size_ok(pos, elen)) {
+				elems->preq = pos;
+				elems->preq_len = elen;
+			} else {
+				elem_parse_failed =
+					IEEE80211_PARSE_ERR_BAD_ELEM_SIZE;
+			}
 			break;
 		case WLAN_EID_PREP:
 			elems->prep = pos;
-- 
2.43.0

[PATCH 2/3] Fix overread in PREP frame processing

[Masashi Honma]

When the AF flag is enabled, hwmp_prep_frame_process() overreads orig_addr
by 2 bytes. Since this occurs within the socket buffer, it does not read across
memory boundaries and therefore poses no security risk; however, we will fix it
as a precaution.

In this fix, a new function mesh_path_parse_reply_frame() is established to
separate the implementation of frame format validation and the check for
unsupported features. This is intended to facilitate future work when
implementing the currently unsupported parts.

Assisted-by: Claude:Sonnet 4.6
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
---
 include/linux/ieee80211-mesh.h | 18 ++++++++++++++++++
 net/mac80211/mesh_hwmp.c       |  5 +++--
 net/mac80211/parse.c           |  9 +++++++--
 3 files changed, 28 insertions(+), 4 deletions(-)

diff --git a/include/linux/ieee80211-mesh.h b/include/linux/ieee80211-mesh.h
index 8ff68cf539ff..ddbf475c1cc8 100644
--- a/include/linux/ieee80211-mesh.h
+++ b/include/linux/ieee80211-mesh.h
@@ -283,4 +283,22 @@ static inline bool ieee80211_mesh_preq_size_ok(const u8 *pos, u8 elen)
 	return true;
 }
 
+/* IEEE Std 802.11-2016 9.4.2.114 PREP element */
+static inline bool ieee80211_mesh_prep_size_ok(const u8 *pos, u8 elen)
+{
+	u8 needed;
+
+	needed = 1 /* Flags */ + 1 /* Hop Count */ + 1 /* Element TTL */ +
+		 6 /* Target Mesh STA Address */ +
+		 4 /* Target HWMP Sequence Number */ +
+		 (AE_F_SET(pos) ? 6 : 0) /* Target External Address */ +
+		 4 /* Lifetime */ + 4 /* Metric */ +
+		 6 /* Originator Mesh STA Address */ +
+		 4 /* Originator HWMP Sequence Number */;
+	if (elen != needed)
+		return false;
+
+	return true;
+}
+
 #endif /* LINUX_IEEE80211_MESH_H */
diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c
index a4984b015995..c70cfc2d6299 100644
--- a/net/mac80211/mesh_hwmp.c
+++ b/net/mac80211/mesh_hwmp.c
@@ -941,9 +941,10 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 						path_metric);
 	}
 	if (elems->prep) {
-		if (elems->prep_len != 31)
-			/* Right now we support no AE */
+		/* Right now we do not support AE (Address Extension) */
+		if (AE_F_SET(elems->prep))
 			goto free;
+
 		path_metric = hwmp_route_info_get(sdata, mgmt, elems->prep,
 						  MPATH_PREP);
 		if (path_metric)
diff --git a/net/mac80211/parse.c b/net/mac80211/parse.c
index 9e52cc48fc18..bbd1e1bc77b4 100644
--- a/net/mac80211/parse.c
+++ b/net/mac80211/parse.c
@@ -556,8 +556,13 @@ _ieee802_11_parse_elems_full(struct ieee80211_elems_parse_params *params,
 			}
 			break;
 		case WLAN_EID_PREP:
-			elems->prep = pos;
-			elems->prep_len = elen;
+			if (ieee80211_mesh_prep_size_ok(pos, elen)) {
+				elems->prep = pos;
+				elems->prep_len = elen;
+			} else {
+				elem_parse_failed =
+					IEEE80211_PARSE_ERR_BAD_ELEM_SIZE;
+			}
 			break;
 		case WLAN_EID_PERR:
 			elems->perr = pos;
-- 
2.43.0

[PATCH 3/3] Fix PERR frame processing

[Masashi Honma]

There are no issues with the PERR processing itself; however, to maintain
consistency with the previous PREQ/PREP code modifications, I will create a new
mesh_path_parse_error_frame() function to separately implement the frame format
validation and the "not supported" check.

Assisted-by: Claude:Sonnet 4.6
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
---
 include/linux/ieee80211-mesh.h | 38 ++++++++++++++++++++++++++++++++++
 net/mac80211/mesh_hwmp.c       | 11 ++++++++--
 net/mac80211/parse.c           |  9 ++++++--
 3 files changed, 54 insertions(+), 4 deletions(-)

diff --git a/include/linux/ieee80211-mesh.h b/include/linux/ieee80211-mesh.h
index ddbf475c1cc8..1bca65e49309 100644
--- a/include/linux/ieee80211-mesh.h
+++ b/include/linux/ieee80211-mesh.h
@@ -61,6 +61,7 @@ struct ieee80211s_hdr {
 #define PREP_IE_TARGET_SN(x)	u32_field_get(x, 9, 0)
 
 #define PERR_IE_TTL(x)			(*(x))
+#define PERR_IE_NUMBER_OF_DST(x)	(*(x + 1))
 #define PERR_IE_TARGET_FLAGS(x)		(*(x + 2))
 #define PERR_IE_TARGET_ADDR(x)		(x + 3)
 #define PERR_IE_TARGET_SN(x)		u32_field_get(x, 9, 0)
@@ -301,4 +302,41 @@ static inline bool ieee80211_mesh_prep_size_ok(const u8 *pos, u8 elen)
 	return true;
 }
 
+/* IEEE Std 802.11-2016 9.4.2.115 PERR element */
+static inline bool ieee80211_mesh_perr_size_ok(const u8 *pos, u8 elen)
+{
+	u8 number_of_dst;
+	u8 needed;
+	const u8 *start;
+	int i;
+
+	number_of_dst = PERR_IE_NUMBER_OF_DST(pos);
+	if (number_of_dst < 1 || number_of_dst > 19)
+		return false;
+
+	start = pos;
+	needed = 1 /* Element TTL */ + 1 /* Number of Destinations */;
+	pos += 2;
+
+	for (i = 0; i < number_of_dst; i++) {
+		u8 dst_len;
+
+		if (pos - start >= elen)
+			return false;
+
+		dst_len = 1 /* Flags */ + 6 /* Destination Address */ +
+			  4 /* HWMP Sequence Number */ +
+			  (AE_F_SET(pos) ? 6 : 0)
+			  /* Destination External Address */ +
+			  2 /* Reason Code */;
+		needed += dst_len;
+		pos += dst_len;
+	}
+
+	if (elen != needed)
+		return false;
+
+	return true;
+}
+
 #endif /* LINUX_IEEE80211_MESH_H */
diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c
index c70cfc2d6299..d2295aa54bb4 100644
--- a/net/mac80211/mesh_hwmp.c
+++ b/net/mac80211/mesh_hwmp.c
@@ -902,6 +902,7 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 	u32 path_metric;
 	struct sta_info *sta;
 	u8 target_count;
+	u8 number_of_dst;
 
 	/* need action_code */
 	if (len < IEEE80211_MIN_ACTION_SIZE(mesh_action))
@@ -952,9 +953,15 @@ void mesh_rx_path_sel_frame(struct ieee80211_sub_if_data *sdata,
 						path_metric);
 	}
 	if (elems->perr) {
-		if (elems->perr_len != 15)
-			/* Right now we support only one destination per PERR */
+		/* Right now we support only one destination per PERR */
+		number_of_dst = PERR_IE_NUMBER_OF_DST(elems->perr);
+		if (number_of_dst != 1)
 			goto free;
+
+		/* Right now we do not support AE (Address Extension) */
+		if (PERR_IE_TARGET_FLAGS(elems->perr) & AE_F)
+			goto free;
+
 		hwmp_perr_frame_process(sdata, mgmt, elems->perr);
 	}
 	if (elems->rann)
diff --git a/net/mac80211/parse.c b/net/mac80211/parse.c
index bbd1e1bc77b4..d84e5e12ad24 100644
--- a/net/mac80211/parse.c
+++ b/net/mac80211/parse.c
@@ -565,8 +565,13 @@ _ieee802_11_parse_elems_full(struct ieee80211_elems_parse_params *params,
 			}
 			break;
 		case WLAN_EID_PERR:
-			elems->perr = pos;
-			elems->perr_len = elen;
+			if (ieee80211_mesh_perr_size_ok(pos, elen)) {
+				elems->perr = pos;
+				elems->perr_len = elen;
+			} else {
+				elem_parse_failed =
+					IEEE80211_PARSE_ERR_BAD_ELEM_SIZE;
+			}
 			break;
 		case WLAN_EID_RANN:
 			if (elen >= sizeof(struct ieee80211_rann_ie))
-- 
2.43.0

Re: [PATCH 1/3] Fix overread in PREQ frame processing

[Johannes Berg]

Hi,

Please send new versions with a version tag ('v2' in the subject etc.),
and not in the old thread, that just messes up the threads. Ideally tag
which tree these should go to, though I think given the many other
things I'll put this into wireless-next anyway, unless you strongly feel
it needs to be in wireless?

Also, please use appropriate "wifi: mac80211: " prefix in the subject.

> +#define PREQ_IE_PREQ_ID(x)	u32_field_get(x, 3, 0)

This isn't really right since u32_field_get() exists only within
mesh_hwmp.c ... it's probably better to modernise all this while at it:

#define IEEE80211_MESH_HWMP_FLAGS_AE	0x40

struct ieee80211_mesh_hwmp_target {
	u8 flags;
	u8 addr[ETH_ALEN];
	__le32 sn;
} __packed;

struct ieee80211_mesh_hwmp_element {
	u8 flags;
	u8 hopcount;
	u8 ttl;
	__le32 preq_id;
	u8 orig_addr[ETH_ALEN];
	__le32 orig_sn;

	/* optional AE, lifetime, metric, target */
	u8 variable[];
} __packed;

and then some inlines like

static inline const struct ieee80211_mesh_hwmp_target *
ieee80211_mesh_hwmp_get_target(const struct ieee80211_mesh_hwmp_element *elem)
{
	int offs = 8; // seems wrong

	if (elem->flags & IEEE80211_MESH_HWMP_FLAGS_AE)
		offs += ETH_ALEN;
	
	return (const void *)&elem->variable[offs];
}

static inline u32
ieee80211_mesh_hwmp_get_lifetime(const struct ieee80211_mesh_hwmp_element *elem)
{
	int offs = FIXME; // not sure

	if (elem->flags & IEEE80211_MESH_HWMP_FLAGS_AE)
		offs += ETH_ALEN;
	
	return (const void *)&elem->variable[offs];
}

and restructure the code accordingly?

Anyway, I dunno. Maybe we should just go with your original patch for
now. Maybe I'm also asking more of you than others because you have an
LLM to help ;-)

johannes

Re: [PATCH 1/3] Fix overread in PREQ frame processing

[Masashi Honma]

Thank you for your reviews!

> Please send new versions with a version tag ('v2' in the subject etc.),

Yes. I used --subject-prefix incorrectly, so I will correct it from next time.

> and not in the old thread, that just messes up the threads.

Yes.

> Ideally tag which tree these should go to, though I think given the many other
> things I'll put this into wireless-next anyway, unless you strongly feel
> it needs to be in wireless?

I'm happy to put this into wireless-next.

> Also, please use appropriate "wifi: mac80211: " prefix in the subject.

Yes.

> This isn't really right since u32_field_get() exists only within
> mesh_hwmp.c ... it's probably better to modernise all this while at it:

Ah, yes. Both the build and tests passed, so I overlooked it.

> and restructure the code accordingly?
> Anyway, I dunno. Maybe we should just go with your original patch for
> now. Maybe I'm also asking more of you than others because you have an
> LLM to help ;-)

If it is not urgent, I would like to proceed with the requested restructuring.
Actually, I only used the LLM to find potential vulnerabilities, and I wrote
the code myself :)

2026年5月11日(月) 16:47 Johannes Berg <johannes@sipsolutions.net>:
>
> Hi,
>
> Please send new versions with a version tag ('v2' in the subject etc.),
> and not in the old thread, that just messes up the threads. Ideally tag
> which tree these should go to, though I think given the many other
> things I'll put this into wireless-next anyway, unless you strongly feel
> it needs to be in wireless?
>
> Also, please use appropriate "wifi: mac80211: " prefix in the subject.
>
> > +#define PREQ_IE_PREQ_ID(x)   u32_field_get(x, 3, 0)
>
> This isn't really right since u32_field_get() exists only within
> mesh_hwmp.c ... it's probably better to modernise all this while at it:
>
> #define IEEE80211_MESH_HWMP_FLAGS_AE    0x40
>
> struct ieee80211_mesh_hwmp_target {
>         u8 flags;
>         u8 addr[ETH_ALEN];
>         __le32 sn;
> } __packed;
>
> struct ieee80211_mesh_hwmp_element {
>         u8 flags;
>         u8 hopcount;
>         u8 ttl;
>         __le32 preq_id;
>         u8 orig_addr[ETH_ALEN];
>         __le32 orig_sn;
>
>         /* optional AE, lifetime, metric, target */
>         u8 variable[];
> } __packed;
>
> and then some inlines like
>
> static inline const struct ieee80211_mesh_hwmp_target *
> ieee80211_mesh_hwmp_get_target(const struct ieee80211_mesh_hwmp_element *elem)
> {
>         int offs = 8; // seems wrong
>
>         if (elem->flags & IEEE80211_MESH_HWMP_FLAGS_AE)
>                 offs += ETH_ALEN;
>
>         return (const void *)&elem->variable[offs];
> }
>
> static inline u32
> ieee80211_mesh_hwmp_get_lifetime(const struct ieee80211_mesh_hwmp_element *elem)
> {
>         int offs = FIXME; // not sure
>
>         if (elem->flags & IEEE80211_MESH_HWMP_FLAGS_AE)
>                 offs += ETH_ALEN;
>
>         return (const void *)&elem->variable[offs];
> }
>
> and restructure the code accordingly?
>
> Anyway, I dunno. Maybe we should just go with your original patch for
> now. Maybe I'm also asking more of you than others because you have an
> LLM to help ;-)
>
> johannes

Re: [PATCH 1/3] Fix overread in PREQ frame processing

[Johannes Berg]

On Mon, 2026-05-11 at 17:58 +0900, Masashi Honma wrote:
> > This isn't really right since u32_field_get() exists only within
> > mesh_hwmp.c ... it's probably better to modernise all this while at it:
> 
> Ah, yes. Both the build and tests passed, so I overlooked it.

Yes, it would, but it's basically not self-contained. More of a code
hygiene issue I guess than a real problem right now.

> > and restructure the code accordingly?
> > Anyway, I dunno. Maybe we should just go with your original patch for
> > now. Maybe I'm also asking more of you than others because you have an
> > LLM to help ;-)
> 
> If it is not urgent, I would like to proceed with the requested restructuring.

Oh sure, as far as I'm concerned there's no urgency, I just didn't want
to keep asking you to make changes too much.

> Actually, I only used the LLM to find potential vulnerabilities, and I wrote
> the code myself :)

Oh OK :) Maybe we need that as a kind of Reported-by? Hmm.

Or you could send a separate bug report email, say there Claude found
it, and then do a Closes: link :-p

johannes

Re: [PATCH 1/3] Fix overread in PREQ frame processing

[Masashi Honma]

> Oh sure, as far as I'm concerned there's no urgency, I just didn't want
> to keep asking you to make changes too much.

No problem. The code is getting better and better, so your reviews are very
welcome.

> Oh OK :) Maybe we need that as a kind of Reported-by? Hmm.
> Or you could send a separate bug report email, say there Claude found
> it, and then do a Closes: link :-p

Ha ha, no problem. I am happy to tag Claude because I wouldn't have even been
able to find this issue if I hadn't used an LLM.

2026年5月11日(月) 18:01 Johannes Berg <johannes@sipsolutions.net>:
>
> On Mon, 2026-05-11 at 17:58 +0900, Masashi Honma wrote:
> > > This isn't really right since u32_field_get() exists only within
> > > mesh_hwmp.c ... it's probably better to modernise all this while at it:
> >
> > Ah, yes. Both the build and tests passed, so I overlooked it.
>
> Yes, it would, but it's basically not self-contained. More of a code
> hygiene issue I guess than a real problem right now.
>
> > > and restructure the code accordingly?
> > > Anyway, I dunno. Maybe we should just go with your original patch for
> > > now. Maybe I'm also asking more of you than others because you have an
> > > LLM to help ;-)
> >
> > If it is not urgent, I would like to proceed with the requested restructuring.
>
> Oh sure, as far as I'm concerned there's no urgency, I just didn't want
> to keep asking you to make changes too much.
>
> > Actually, I only used the LLM to find potential vulnerabilities, and I wrote
> > the code myself :)
>
> Oh OK :) Maybe we need that as a kind of Reported-by? Hmm.
>
> Or you could send a separate bug report email, say there Claude found
> it, and then do a Closes: link :-p
>
> johannes