[PATCH net 0/6] xsk: harden TX metadata validation against races

[PATCH net 0/6] xsk: harden TX metadata validation against races

[Stanislav Fomichev]

Cen Zhang reported a KASAN out-of-bounds read when AF_XDP is configured
with a TX metadata area smaller than struct xsk_tx_metadata. The metadata
is also shared with user space, so reading its flags more than once can
produce inconsistent validation and processing decisions.

Require enough space for the flags and one request field, validate the
launch-time field against the configured metadata length, and use one
snapshot of the flags while processing each request. Carry the validated
decision through completion handling so later user-space changes cannot
enable an unrequested completion timestamp.

Reported-by: AutonomousCodeSecurity@microsoft.com
Reported-by: Cen Zhang (Microsoft) <blbllhy@gmail.com>
Link: https://lore.kernel.org/netdev/20260720155210.34229-1-blbllhy@gmail.com/

Stanislav Fomichev (6):
  xsk: require at least 16 bytes of TX metadata
  xsk: pass TX metadata pointer by reference
  xsk: clear metadata pointer when no timestamp is requested
  xsk: validate launch-time metadata size
  xsk: move xsk_tx_metadata_request() to xdp_sock_drv.h
  xsk: validate metadata when processing requests

 drivers/net/ethernet/intel/igc/igc_main.c     |  2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en.h  |  2 +-
 .../net/ethernet/mellanox/mlx5/core/en/xdp.c  | 15 ++--
 .../net/ethernet/mellanox/mlx5/core/en/xdp.h  |  4 +-
 .../ethernet/mellanox/mlx5/core/en/xsk/tx.c   |  2 +-
 .../net/ethernet/stmicro/stmmac/stmmac_main.c |  4 +-
 include/net/libeth/xsk.h                      |  2 +-
 include/net/xdp_sock.h                        | 45 ++---------
 include/net/xdp_sock_drv.h                    | 77 ++++++++++++++++---
 net/xdp/xdp_umem.c                            |  2 +
 net/xdp/xsk.c                                 | 11 ++-
 net/xdp/xsk_buff_pool.c                       |  6 +-
 12 files changed, 101 insertions(+), 71 deletions(-)

-- 
2.53.0-Meta

[PATCH net 1/6] xsk: require at least 16 bytes of TX metadata

[Stanislav Fomichev]

AF_XDP accepts a TX metadata length as small as eight bytes, but every
supported request needs the flags plus at least one eight-byte request
field. Such short metadata also lets the kernel read beyond the registered
area.

Require 16 bytes rather than sizeof(struct xsk_tx_metadata) to preserve
compatibility with applications that do not use launch-time metadata.

Fixes: 341ac980eab9 ("xsk: Support tx_metadata_len")
Reported-by: AutonomousCodeSecurity@microsoft.com
Reported-by: Cen Zhang (Microsoft) <blbllhy@gmail.com>
Link: https://lore.kernel.org/netdev/20260720155210.34229-1-blbllhy@gmail.com/
Signed-off-by: Stanislav Fomichev <sdf@fomichev.me>
---
 net/xdp/xdp_umem.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/xdp/xdp_umem.c b/net/xdp/xdp_umem.c
index 58da2f4f4397..cd8643360eb3 100644
--- a/net/xdp/xdp_umem.c
+++ b/net/xdp/xdp_umem.c
@@ -210,6 +210,8 @@ static int xdp_umem_reg(struct xdp_umem *umem, struct xdp_umem_reg *mr)
 	if (mr->flags & XDP_UMEM_TX_METADATA_LEN) {
 		if (mr->tx_metadata_len >= 256 || mr->tx_metadata_len % 8)
 			return -EINVAL;
+		if (mr->tx_metadata_len < 16)
+			return -EINVAL;
 		umem->tx_metadata_len = mr->tx_metadata_len;
 	}
 
-- 
2.53.0-Meta

[PATCH net 2/6] xsk: pass TX metadata pointer by reference

[Stanislav Fomichev]

Completion handling needs to know whether a timestamp was requested when
the metadata was processed. Let xsk_tx_metadata_request() update the
caller's metadata pointer so that decision can be carried forward without
rereading user-controlled flags.

This only changes the interface; behavior remains unchanged.

Fixes: ca4419f15abd ("xsk: Add launch time hardware offload support to XDP Tx metadata")
Cc: Cen Zhang (Microsoft) <blbllhy@gmail.com>
Signed-off-by: Stanislav Fomichev <sdf@fomichev.me>
---
 drivers/net/ethernet/intel/igc/igc_main.c          |  2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en.h       |  2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c   | 14 +++++++++-----
 drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h   |  4 ++--
 .../net/ethernet/mellanox/mlx5/core/en/xsk/tx.c    |  2 +-
 drivers/net/ethernet/stmicro/stmmac/stmmac_main.c  |  2 +-
 include/net/libeth/xsk.h                           |  2 +-
 include/net/xdp_sock.h                             |  8 +++++---
 8 files changed, 21 insertions(+), 15 deletions(-)

diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c
index 2c9e2dfd8499..5a18d00deb4f 100644
--- a/drivers/net/ethernet/intel/igc/igc_main.c
+++ b/drivers/net/ethernet/intel/igc/igc_main.c
@@ -3082,7 +3082,7 @@ static void igc_xdp_xmit_zc(struct igc_ring *ring)
 		meta_req.tx_buffer = bi;
 		meta_req.meta = meta;
 		meta_req.used_desc = 0;
-		xsk_tx_metadata_request(meta, &igc_xsk_tx_metadata_ops,
+		xsk_tx_metadata_request(&meta, &igc_xsk_tx_metadata_ops,
 					&meta_req);
 
 		/* xsk_tx_metadata_request() may have updated next_to_use */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index d507289096c2..e12cb41c4b97 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -483,7 +483,7 @@ typedef int (*mlx5e_fp_xmit_xdp_frame_check)(struct mlx5e_xdpsq *);
 typedef bool (*mlx5e_fp_xmit_xdp_frame)(struct mlx5e_xdpsq *,
 					struct mlx5e_xmit_data *,
 					int,
-					struct xsk_tx_metadata *);
+					struct xsk_tx_metadata **);
 
 struct mlx5e_xdpsq {
 	/* data path */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c
index d8c7cb8837d7..dac5e9d4c8bd 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c
@@ -452,11 +452,11 @@ INDIRECT_CALLABLE_SCOPE int mlx5e_xmit_xdp_frame_check_mpwqe(struct mlx5e_xdpsq
 
 INDIRECT_CALLABLE_SCOPE bool
 mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xmit_data *xdptxd,
-		     int check_result, struct xsk_tx_metadata *meta);
+		     int check_result, struct xsk_tx_metadata **meta);
 
 INDIRECT_CALLABLE_SCOPE bool
 mlx5e_xmit_xdp_frame_mpwqe(struct mlx5e_xdpsq *sq, struct mlx5e_xmit_data *xdptxd,
-			   int check_result, struct xsk_tx_metadata *meta)
+			   int check_result, struct xsk_tx_metadata **meta)
 {
 	struct mlx5e_tx_mpwqe *session = &sq->mpwqe;
 	struct mlx5e_xdpsq_stats *stats = sq->stats;
@@ -504,7 +504,10 @@ mlx5e_xmit_xdp_frame_mpwqe(struct mlx5e_xdpsq *sq, struct mlx5e_xmit_data *xdptx
 		 * and it's safe to complete it at any time.
 		 */
 		mlx5e_xdp_mpwqe_session_start(sq);
-		xsk_tx_metadata_request(meta, &mlx5e_xsk_tx_metadata_ops, &session->wqe->eth);
+		if (meta)
+			xsk_tx_metadata_request(meta,
+						&mlx5e_xsk_tx_metadata_ops,
+						&session->wqe->eth);
 	}
 
 	mlx5e_xdp_mpwqe_add_dseg(sq, p, stats);
@@ -535,7 +538,7 @@ INDIRECT_CALLABLE_SCOPE int mlx5e_xmit_xdp_frame_check(struct mlx5e_xdpsq *sq)
 
 INDIRECT_CALLABLE_SCOPE bool
 mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xmit_data *xdptxd,
-		     int check_result, struct xsk_tx_metadata *meta)
+		     int check_result, struct xsk_tx_metadata **meta)
 {
 	struct mlx5e_xmit_data_frags *xdptxdf =
 		container_of(xdptxd, struct mlx5e_xmit_data_frags, xd);
@@ -649,7 +652,8 @@ mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xmit_data *xdptxd,
 
 	sq->pc += num_wqebbs;
 
-	xsk_tx_metadata_request(meta, &mlx5e_xsk_tx_metadata_ops, eseg);
+	if (meta)
+		xsk_tx_metadata_request(meta, &mlx5e_xsk_tx_metadata_ops, eseg);
 
 	sq->doorbell_cseg = cseg;
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h
index 3c54f8962664..5b8f4094d553 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h
@@ -114,11 +114,11 @@ extern const struct xsk_tx_metadata_ops mlx5e_xsk_tx_metadata_ops;
 INDIRECT_CALLABLE_DECLARE(bool mlx5e_xmit_xdp_frame_mpwqe(struct mlx5e_xdpsq *sq,
 							  struct mlx5e_xmit_data *xdptxd,
 							  int check_result,
-							  struct xsk_tx_metadata *meta));
+							  struct xsk_tx_metadata **meta));
 INDIRECT_CALLABLE_DECLARE(bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq,
 						    struct mlx5e_xmit_data *xdptxd,
 						    int check_result,
-						    struct xsk_tx_metadata *meta));
+						    struct xsk_tx_metadata **meta));
 INDIRECT_CALLABLE_DECLARE(int mlx5e_xmit_xdp_frame_check_mpwqe(struct mlx5e_xdpsq *sq));
 INDIRECT_CALLABLE_DECLARE(int mlx5e_xmit_xdp_frame_check(struct mlx5e_xdpsq *sq));
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c
index 8aeab4b21035..3d19dad8f868 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c
@@ -105,7 +105,7 @@ bool mlx5e_xsk_tx(struct mlx5e_xdpsq *sq, unsigned int budget)
 
 		ret = INDIRECT_CALL_2(sq->xmit_xdp_frame, mlx5e_xmit_xdp_frame_mpwqe,
 				      mlx5e_xmit_xdp_frame, sq, &xdptxd,
-				      check_result, meta);
+				      check_result, &meta);
 		if (unlikely(!ret)) {
 			if (sq->mpwqe.wqe)
 				mlx5e_xdp_mpwqe_complete(sq);
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index 151c77713025..3f5111969c41 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -2747,7 +2747,7 @@ static bool stmmac_xdp_xmit_zc(struct stmmac_priv *priv, u32 queue, u32 budget)
 		meta_req.set_ic = &set_ic;
 		meta_req.tbs = tx_q->tbs;
 		meta_req.edesc = &tx_q->dma_entx[entry];
-		xsk_tx_metadata_request(meta, &stmmac_xsk_tx_metadata_ops,
+		xsk_tx_metadata_request(&meta, &stmmac_xsk_tx_metadata_ops,
 					&meta_req);
 		if (set_ic) {
 			tx_q->tx_count_frames = 0;
diff --git a/include/net/libeth/xsk.h b/include/net/libeth/xsk.h
index 82b5d21aae87..e2fa6bf6b1b3 100644
--- a/include/net/libeth/xsk.h
+++ b/include/net/libeth/xsk.h
@@ -205,7 +205,7 @@ __libeth_xsk_xmit_fill_buf_md(const struct xdp_desc *xdesc,
 	BUILD_BUG_ON(!__builtin_constant_p(tmo == libeth_xsktmo));
 	tmo = tmo == libeth_xsktmo ? &__libeth_xsktmo : tmo;
 
-	xsk_tx_metadata_request(ctx.meta, tmo, &desc);
+	xsk_tx_metadata_request(&ctx.meta, tmo, &desc);
 
 	return desc;
 }
diff --git a/include/net/xdp_sock.h b/include/net/xdp_sock.h
index 8b51876efbed..06c081feff42 100644
--- a/include/net/xdp_sock.h
+++ b/include/net/xdp_sock.h
@@ -153,17 +153,19 @@ static inline void xsk_tx_metadata_to_compl(struct xsk_tx_metadata *meta,
 /**
  *  xsk_tx_metadata_request - Evaluate AF_XDP TX metadata at submission
  *  and call appropriate xsk_tx_metadata_ops operation.
- *  @meta: pointer to AF_XDP metadata area
+ *  @pmeta: pointer to pointer to AF_XDP metadata area
  *  @ops: pointer to struct xsk_tx_metadata_ops
  *  @priv: pointer to driver-private aread
  *
  *  This function should be called by the networking device when
  *  it prepares AF_XDP egress packet.
  */
-static inline void xsk_tx_metadata_request(const struct xsk_tx_metadata *meta,
+static inline void xsk_tx_metadata_request(struct xsk_tx_metadata **pmeta,
 					   const struct xsk_tx_metadata_ops *ops,
 					   void *priv)
 {
+	const struct xsk_tx_metadata *meta = *pmeta;
+
 	if (!meta)
 		return;
 
@@ -231,7 +233,7 @@ static inline void xsk_tx_metadata_to_compl(struct xsk_tx_metadata *meta,
 {
 }
 
-static inline void xsk_tx_metadata_request(struct xsk_tx_metadata *meta,
+static inline void xsk_tx_metadata_request(struct xsk_tx_metadata **pmeta,
 					   const struct xsk_tx_metadata_ops *ops,
 					   void *priv)
 {
-- 
2.53.0-Meta

[PATCH net 3/6] xsk: clear metadata pointer when no timestamp is requested

[Stanislav Fomichev]

User space can change metadata flags after request processing. Rereading
them during completion can therefore make the kernel write a timestamp
that was not requested when the packet was submitted.

Clear the metadata pointer during request processing unless timestamp
completion is requested. Completion handling can then use the pointer
itself instead of rereading the flags.

On the mlx5 multi-packet WQE path metadata is evaluated per batch:
xsk_tx_metadata_request() runs only for the descriptor that starts a
session, just like the checksum offload that is applied once through the
shared WQE. Only that descriptor's pointer is reset, so completion
handling can record a timestamp for the other descriptors of the session
regardless of their own XDP_TXMD_FLAGS_TIMESTAMP bit. The write stays
inside the metadata area; the single-WQE, other zero-copy, and generic
paths reset the pointer per descriptor and are unaffected.

Fixes: ca4419f15abd ("xsk: Add launch time hardware offload support to XDP Tx metadata")
Cc: Cen Zhang (Microsoft) <blbllhy@gmail.com>
Signed-off-by: Stanislav Fomichev <sdf@fomichev.me>
---
 include/net/xdp_sock.h | 14 ++++++++++----
 net/xdp/xsk.c          |  2 ++
 2 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/include/net/xdp_sock.h b/include/net/xdp_sock.h
index 06c081feff42..2b2eb9b9d580 100644
--- a/include/net/xdp_sock.h
+++ b/include/net/xdp_sock.h
@@ -141,13 +141,16 @@ INDIRECT_CALLABLE_DECLARE(void xsk_destruct_skb(struct sk_buff *));
 static inline void xsk_tx_metadata_to_compl(struct xsk_tx_metadata *meta,
 					    struct xsk_tx_metadata_compl *compl)
 {
+	compl->tx_timestamp = NULL;
+
 	if (!meta)
 		return;
 
-	if (meta->flags & XDP_TXMD_FLAGS_TIMESTAMP)
-		compl->tx_timestamp = &meta->completion.tx_timestamp;
-	else
-		compl->tx_timestamp = NULL;
+	/* we can only arrive here if the completion timestamp has been
+	 * requested via XDP_TXMD_FLAGS_TIMESTAMP, see xsk_tx_metadata_request
+	 */
+
+	compl->tx_timestamp = &meta->completion.tx_timestamp;
 }
 
 /**
@@ -182,6 +185,9 @@ static inline void xsk_tx_metadata_request(struct xsk_tx_metadata **pmeta,
 		if (meta->flags & XDP_TXMD_FLAGS_CHECKSUM)
 			ops->tmo_request_checksum(meta->request.csum_start,
 						  meta->request.csum_offset, priv);
+
+	if (!(meta->flags & XDP_TXMD_FLAGS_TIMESTAMP))
+		*pmeta = NULL;
 }
 
 /**
diff --git a/net/xdp/xsk.c b/net/xdp/xsk.c
index f906d51b6699..fcc6f17f3576 100644
--- a/net/xdp/xsk.c
+++ b/net/xdp/xsk.c
@@ -998,6 +998,8 @@ static int xsk_skb_metadata(struct sk_buff *skb, void *buffer,
 
 	if (meta->flags & XDP_TXMD_FLAGS_LAUNCH_TIME)
 		skb->skb_mstamp_ns = meta->request.launch_time;
+	if (!(meta->flags & XDP_TXMD_FLAGS_TIMESTAMP))
+		meta = NULL;
 	xsk_tx_metadata_to_compl(meta, &skb_shinfo(skb)->xsk_meta);
 
 	return 0;
-- 
2.53.0-Meta

[PATCH net 4/6] xsk: validate launch-time metadata size

[Stanislav Fomichev]

Launch-time metadata extends beyond the first 16 bytes of struct
xsk_tx_metadata. Reject the request when the registered metadata area does
not contain the complete field.

Snapshot the validated flags for the generic transmit path and use that
snapshot for request and completion processing, avoiding inconsistent
decisions if user space changes the flags concurrently.

Note that only xsk_skb_metadata is properly using the flags,
__xsk_buff_get_metadata ignores them. Next commits address that.

Fixes: ca4419f15abd ("xsk: Add launch time hardware offload support to XDP Tx metadata")
Cc: Cen Zhang (Microsoft) <blbllhy@gmail.com>
Signed-off-by: Stanislav Fomichev <sdf@fomichev.me>
---
 include/net/xdp_sock_drv.h | 17 +++++++++++++----
 net/xdp/xsk.c              | 11 ++++++-----
 2 files changed, 19 insertions(+), 9 deletions(-)

diff --git a/include/net/xdp_sock_drv.h b/include/net/xdp_sock_drv.h
index 46797645a0c2..68a787601357 100644
--- a/include/net/xdp_sock_drv.h
+++ b/include/net/xdp_sock_drv.h
@@ -260,21 +260,28 @@ xsk_buff_raw_get_ctx(const struct xsk_buff_pool *pool, u64 addr)
 	0)
 
 static inline bool
-xsk_buff_valid_tx_metadata(const struct xsk_tx_metadata *meta)
+xsk_buff_valid_tx_metadata(const struct xsk_buff_pool *pool,
+			   const struct xsk_tx_metadata *meta, u64 *flags)
 {
-	return !(meta->flags & ~XDP_TXMD_FLAGS_VALID);
+	*flags = READ_ONCE(meta->flags);
+	if (*flags & XDP_TXMD_FLAGS_LAUNCH_TIME)
+		if (pool->tx_metadata_len <
+		    offsetofend(struct xsk_tx_metadata, request.launch_time))
+			return false;
+	return !(*flags & ~XDP_TXMD_FLAGS_VALID);
 }
 
 static inline struct xsk_tx_metadata *
 __xsk_buff_get_metadata(const struct xsk_buff_pool *pool, void *data)
 {
 	struct xsk_tx_metadata *meta;
+	u64 flags;
 
 	if (!pool->tx_metadata_len)
 		return NULL;
 
 	meta = data - pool->tx_metadata_len;
-	if (unlikely(!xsk_buff_valid_tx_metadata(meta)))
+	if (unlikely(!xsk_buff_valid_tx_metadata(pool, meta, &flags)))
 		return NULL; /* no way to signal the error to the user */
 
 	return meta;
@@ -469,7 +476,9 @@ xsk_buff_raw_get_ctx(const struct xsk_buff_pool *pool, u64 addr)
 	return (struct xdp_desc_ctx){ };
 }
 
-static inline bool xsk_buff_valid_tx_metadata(struct xsk_tx_metadata *meta)
+static inline bool
+xsk_buff_valid_tx_metadata(const struct xsk_buff_pool *pool,
+			   const struct xsk_tx_metadata *meta, u64 *flags)
 {
 	return false;
 }
diff --git a/net/xdp/xsk.c b/net/xdp/xsk.c
index fcc6f17f3576..7855ee09c4b6 100644
--- a/net/xdp/xsk.c
+++ b/net/xdp/xsk.c
@@ -967,15 +967,16 @@ static int xsk_skb_metadata(struct sk_buff *skb, void *buffer,
 {
 	struct xsk_tx_metadata *meta = NULL;
 	u16 csum_start, csum_offset;
+	u64 flags;
 
 	if (unlikely(pool->tx_metadata_len == 0))
 		return -EINVAL;
 
 	meta = buffer - pool->tx_metadata_len;
-	if (unlikely(!xsk_buff_valid_tx_metadata(meta)))
+	if (unlikely(!xsk_buff_valid_tx_metadata(pool, meta, &flags)))
 		return -EINVAL;
 
-	if (meta->flags & XDP_TXMD_FLAGS_CHECKSUM) {
+	if (flags & XDP_TXMD_FLAGS_CHECKSUM) {
 		csum_start = READ_ONCE(meta->request.csum_start);
 		csum_offset = READ_ONCE(meta->request.csum_offset);
 
@@ -996,9 +997,9 @@ static int xsk_skb_metadata(struct sk_buff *skb, void *buffer,
 		}
 	}
 
-	if (meta->flags & XDP_TXMD_FLAGS_LAUNCH_TIME)
-		skb->skb_mstamp_ns = meta->request.launch_time;
-	if (!(meta->flags & XDP_TXMD_FLAGS_TIMESTAMP))
+	if (flags & XDP_TXMD_FLAGS_LAUNCH_TIME)
+		skb->skb_mstamp_ns = READ_ONCE(meta->request.launch_time);
+	if (!(flags & XDP_TXMD_FLAGS_TIMESTAMP))
 		meta = NULL;
 	xsk_tx_metadata_to_compl(meta, &skb_shinfo(skb)->xsk_meta);
 
-- 
2.53.0-Meta

[PATCH net 5/6] xsk: move xsk_tx_metadata_request() to xdp_sock_drv.h

[Stanislav Fomichev]

xsk_tx_metadata_request() must validate metadata with
xsk_buff_valid_tx_metadata(), which is defined in xdp_sock_drv.h. Move the
helper there before adding that dependency. All callers already include
the destination header, so this has no functional effect.

Fixes: ca4419f15abd ("xsk: Add launch time hardware offload support to XDP Tx metadata")
Cc: Cen Zhang (Microsoft) <blbllhy@gmail.com>
Signed-off-by: Stanislav Fomichev <sdf@fomichev.me>
---
 include/net/xdp_sock.h     | 43 --------------------------------------
 include/net/xdp_sock_drv.h | 43 ++++++++++++++++++++++++++++++++++++++
 2 files changed, 43 insertions(+), 43 deletions(-)

diff --git a/include/net/xdp_sock.h b/include/net/xdp_sock.h
index 2b2eb9b9d580..6e70b320b399 100644
--- a/include/net/xdp_sock.h
+++ b/include/net/xdp_sock.h
@@ -153,43 +153,6 @@ static inline void xsk_tx_metadata_to_compl(struct xsk_tx_metadata *meta,
 	compl->tx_timestamp = &meta->completion.tx_timestamp;
 }
 
-/**
- *  xsk_tx_metadata_request - Evaluate AF_XDP TX metadata at submission
- *  and call appropriate xsk_tx_metadata_ops operation.
- *  @pmeta: pointer to pointer to AF_XDP metadata area
- *  @ops: pointer to struct xsk_tx_metadata_ops
- *  @priv: pointer to driver-private aread
- *
- *  This function should be called by the networking device when
- *  it prepares AF_XDP egress packet.
- */
-static inline void xsk_tx_metadata_request(struct xsk_tx_metadata **pmeta,
-					   const struct xsk_tx_metadata_ops *ops,
-					   void *priv)
-{
-	const struct xsk_tx_metadata *meta = *pmeta;
-
-	if (!meta)
-		return;
-
-	if (ops->tmo_request_launch_time)
-		if (meta->flags & XDP_TXMD_FLAGS_LAUNCH_TIME)
-			ops->tmo_request_launch_time(meta->request.launch_time,
-						     priv);
-
-	if (ops->tmo_request_timestamp)
-		if (meta->flags & XDP_TXMD_FLAGS_TIMESTAMP)
-			ops->tmo_request_timestamp(priv);
-
-	if (ops->tmo_request_checksum)
-		if (meta->flags & XDP_TXMD_FLAGS_CHECKSUM)
-			ops->tmo_request_checksum(meta->request.csum_start,
-						  meta->request.csum_offset, priv);
-
-	if (!(meta->flags & XDP_TXMD_FLAGS_TIMESTAMP))
-		*pmeta = NULL;
-}
-
 /**
  *  xsk_tx_metadata_complete - Evaluate AF_XDP TX metadata at completion
  *  and call appropriate xsk_tx_metadata_ops operation.
@@ -239,12 +202,6 @@ static inline void xsk_tx_metadata_to_compl(struct xsk_tx_metadata *meta,
 {
 }
 
-static inline void xsk_tx_metadata_request(struct xsk_tx_metadata **pmeta,
-					   const struct xsk_tx_metadata_ops *ops,
-					   void *priv)
-{
-}
-
 static inline void xsk_tx_metadata_complete(struct xsk_tx_metadata_compl *compl,
 					    const struct xsk_tx_metadata_ops *ops,
 					    void *priv)
diff --git a/include/net/xdp_sock_drv.h b/include/net/xdp_sock_drv.h
index 68a787601357..f87c4215673e 100644
--- a/include/net/xdp_sock_drv.h
+++ b/include/net/xdp_sock_drv.h
@@ -271,6 +271,43 @@ xsk_buff_valid_tx_metadata(const struct xsk_buff_pool *pool,
 	return !(*flags & ~XDP_TXMD_FLAGS_VALID);
 }
 
+/**
+ *  xsk_tx_metadata_request - Evaluate AF_XDP TX metadata at submission
+ *  and call appropriate xsk_tx_metadata_ops operation.
+ *  @pmeta: pointer to pointer to AF_XDP metadata area
+ *  @ops: pointer to struct xsk_tx_metadata_ops
+ *  @priv: pointer to driver-private aread
+ *
+ *  This function should be called by the networking device when
+ *  it prepares AF_XDP egress packet.
+ */
+static inline void xsk_tx_metadata_request(struct xsk_tx_metadata **pmeta,
+					   const struct xsk_tx_metadata_ops *ops,
+					   void *priv)
+{
+	const struct xsk_tx_metadata *meta = *pmeta;
+
+	if (!meta)
+		return;
+
+	if (ops->tmo_request_launch_time)
+		if (meta->flags & XDP_TXMD_FLAGS_LAUNCH_TIME)
+			ops->tmo_request_launch_time(meta->request.launch_time,
+						     priv);
+
+	if (ops->tmo_request_timestamp)
+		if (meta->flags & XDP_TXMD_FLAGS_TIMESTAMP)
+			ops->tmo_request_timestamp(priv);
+
+	if (ops->tmo_request_checksum)
+		if (meta->flags & XDP_TXMD_FLAGS_CHECKSUM)
+			ops->tmo_request_checksum(meta->request.csum_start,
+						  meta->request.csum_offset, priv);
+
+	if (!(meta->flags & XDP_TXMD_FLAGS_TIMESTAMP))
+		*pmeta = NULL;
+}
+
 static inline struct xsk_tx_metadata *
 __xsk_buff_get_metadata(const struct xsk_buff_pool *pool, void *data)
 {
@@ -483,6 +520,12 @@ xsk_buff_valid_tx_metadata(const struct xsk_buff_pool *pool,
 	return false;
 }
 
+static inline void xsk_tx_metadata_request(struct xsk_tx_metadata **pmeta,
+					   const struct xsk_tx_metadata_ops *ops,
+					   void *priv)
+{
+}
+
 static inline struct xsk_tx_metadata *
 __xsk_buff_get_metadata(const struct xsk_buff_pool *pool, void *data)
 {
-- 
2.53.0-Meta

[PATCH net 6/6] xsk: validate metadata when processing requests

[Stanislav Fomichev]

The zero-copy path validates TX metadata while obtaining the descriptor
context, then reads it again later when preparing the hardware request.
User space can change the metadata between those operations and bypass the
original validation.

Validate the metadata in xsk_tx_metadata_request() and use the resulting
flags snapshot for every feature check. Read request fields once so all
zero-copy drivers process only values observed after successful
validation.

Fixes: ca4419f15abd ("xsk: Add launch time hardware offload support to XDP Tx metadata")
Cc: Cen Zhang (Microsoft) <blbllhy@gmail.com>
Signed-off-by: Stanislav Fomichev <sdf@fomichev.me>
---
 drivers/net/ethernet/intel/igc/igc_main.c     |  2 +-
 .../net/ethernet/mellanox/mlx5/core/en/xdp.c  |  5 +-
 .../net/ethernet/stmicro/stmmac/stmmac_main.c |  4 +-
 include/net/libeth/xsk.h                      |  2 +-
 include/net/xdp_sock_drv.h                    | 51 ++++++++++---------
 net/xdp/xsk_buff_pool.c                       |  6 +--
 6 files changed, 37 insertions(+), 33 deletions(-)

diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c
index 5a18d00deb4f..7b6cdc037c00 100644
--- a/drivers/net/ethernet/intel/igc/igc_main.c
+++ b/drivers/net/ethernet/intel/igc/igc_main.c
@@ -3082,7 +3082,7 @@ static void igc_xdp_xmit_zc(struct igc_ring *ring)
 		meta_req.tx_buffer = bi;
 		meta_req.meta = meta;
 		meta_req.used_desc = 0;
-		xsk_tx_metadata_request(&meta, &igc_xsk_tx_metadata_ops,
+		xsk_tx_metadata_request(pool, &meta, &igc_xsk_tx_metadata_ops,
 					&meta_req);
 
 		/* xsk_tx_metadata_request() may have updated next_to_use */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c
index dac5e9d4c8bd..77ea51bfbaae 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c
@@ -505,7 +505,7 @@ mlx5e_xmit_xdp_frame_mpwqe(struct mlx5e_xdpsq *sq, struct mlx5e_xmit_data *xdptx
 		 */
 		mlx5e_xdp_mpwqe_session_start(sq);
 		if (meta)
-			xsk_tx_metadata_request(meta,
+			xsk_tx_metadata_request(sq->xsk_pool, meta,
 						&mlx5e_xsk_tx_metadata_ops,
 						&session->wqe->eth);
 	}
@@ -653,7 +653,8 @@ mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xmit_data *xdptxd,
 	sq->pc += num_wqebbs;
 
 	if (meta)
-		xsk_tx_metadata_request(meta, &mlx5e_xsk_tx_metadata_ops, eseg);
+		xsk_tx_metadata_request(sq->xsk_pool, meta,
+					&mlx5e_xsk_tx_metadata_ops, eseg);
 
 	sq->doorbell_cseg = cseg;
 
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index 3f5111969c41..eb666b0bcdd2 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -2747,8 +2747,8 @@ static bool stmmac_xdp_xmit_zc(struct stmmac_priv *priv, u32 queue, u32 budget)
 		meta_req.set_ic = &set_ic;
 		meta_req.tbs = tx_q->tbs;
 		meta_req.edesc = &tx_q->dma_entx[entry];
-		xsk_tx_metadata_request(&meta, &stmmac_xsk_tx_metadata_ops,
-					&meta_req);
+		xsk_tx_metadata_request(pool, &meta,
+					&stmmac_xsk_tx_metadata_ops, &meta_req);
 		if (set_ic) {
 			tx_q->tx_count_frames = 0;
 			stmmac_set_tx_ic(priv, tx_desc);
diff --git a/include/net/libeth/xsk.h b/include/net/libeth/xsk.h
index e2fa6bf6b1b3..5dcc0d7f65b7 100644
--- a/include/net/libeth/xsk.h
+++ b/include/net/libeth/xsk.h
@@ -205,7 +205,7 @@ __libeth_xsk_xmit_fill_buf_md(const struct xdp_desc *xdesc,
 	BUILD_BUG_ON(!__builtin_constant_p(tmo == libeth_xsktmo));
 	tmo = tmo == libeth_xsktmo ? &__libeth_xsktmo : tmo;
 
-	xsk_tx_metadata_request(&ctx.meta, tmo, &desc);
+	xsk_tx_metadata_request(sq->pool, &ctx.meta, tmo, &desc);
 
 	return desc;
 }
diff --git a/include/net/xdp_sock_drv.h b/include/net/xdp_sock_drv.h
index f87c4215673e..b344789f5df8 100644
--- a/include/net/xdp_sock_drv.h
+++ b/include/net/xdp_sock_drv.h
@@ -245,7 +245,7 @@ static inline void *xsk_buff_raw_get_data(struct xsk_buff_pool *pool, u64 addr)
  * details.
  *
  * Return: new &xdp_desc_ctx struct containing desc's DMA address and metadata
- * pointer, if it is present and valid (initialized to %NULL otherwise).
+ * pointer, if it is present (initialized to %NULL otherwise).
  */
 static inline struct xdp_desc_ctx
 xsk_buff_raw_get_ctx(const struct xsk_buff_pool *pool, u64 addr)
@@ -274,54 +274,56 @@ xsk_buff_valid_tx_metadata(const struct xsk_buff_pool *pool,
 /**
  *  xsk_tx_metadata_request - Evaluate AF_XDP TX metadata at submission
  *  and call appropriate xsk_tx_metadata_ops operation.
+ *  @pool: pointer to AF_XDP buffer pool, used to validate the metadata
  *  @pmeta: pointer to pointer to AF_XDP metadata area
  *  @ops: pointer to struct xsk_tx_metadata_ops
- *  @priv: pointer to driver-private aread
+ *  @priv: pointer to driver-private area
  *
  *  This function should be called by the networking device when
  *  it prepares AF_XDP egress packet.
  */
-static inline void xsk_tx_metadata_request(struct xsk_tx_metadata **pmeta,
-					   const struct xsk_tx_metadata_ops *ops,
-					   void *priv)
+static inline void
+xsk_tx_metadata_request(const struct xsk_buff_pool *pool,
+			struct xsk_tx_metadata **pmeta,
+			const struct xsk_tx_metadata_ops *ops, void *priv)
 {
 	const struct xsk_tx_metadata *meta = *pmeta;
+	u64 flags;
 
 	if (!meta)
 		return;
 
+	if (unlikely(!xsk_buff_valid_tx_metadata(pool, meta, &flags))) {
+		*pmeta = NULL;
+		return; /* no way to signal the error to the user */
+	}
+
 	if (ops->tmo_request_launch_time)
-		if (meta->flags & XDP_TXMD_FLAGS_LAUNCH_TIME)
-			ops->tmo_request_launch_time(meta->request.launch_time,
-						     priv);
+		if (flags & XDP_TXMD_FLAGS_LAUNCH_TIME)
+			ops->tmo_request_launch_time(
+				READ_ONCE(meta->request.launch_time), priv);
 
 	if (ops->tmo_request_timestamp)
-		if (meta->flags & XDP_TXMD_FLAGS_TIMESTAMP)
+		if (flags & XDP_TXMD_FLAGS_TIMESTAMP)
 			ops->tmo_request_timestamp(priv);
 
 	if (ops->tmo_request_checksum)
-		if (meta->flags & XDP_TXMD_FLAGS_CHECKSUM)
-			ops->tmo_request_checksum(meta->request.csum_start,
-						  meta->request.csum_offset, priv);
+		if (flags & XDP_TXMD_FLAGS_CHECKSUM)
+			ops->tmo_request_checksum(
+				READ_ONCE(meta->request.csum_start),
+				READ_ONCE(meta->request.csum_offset), priv);
 
-	if (!(meta->flags & XDP_TXMD_FLAGS_TIMESTAMP))
+	if (!(flags & XDP_TXMD_FLAGS_TIMESTAMP))
 		*pmeta = NULL;
 }
 
 static inline struct xsk_tx_metadata *
 __xsk_buff_get_metadata(const struct xsk_buff_pool *pool, void *data)
 {
-	struct xsk_tx_metadata *meta;
-	u64 flags;
-
 	if (!pool->tx_metadata_len)
 		return NULL;
 
-	meta = data - pool->tx_metadata_len;
-	if (unlikely(!xsk_buff_valid_tx_metadata(pool, meta, &flags)))
-		return NULL; /* no way to signal the error to the user */
-
-	return meta;
+	return data - pool->tx_metadata_len;
 }
 
 static inline struct xsk_tx_metadata *
@@ -520,9 +522,10 @@ xsk_buff_valid_tx_metadata(const struct xsk_buff_pool *pool,
 	return false;
 }
 
-static inline void xsk_tx_metadata_request(struct xsk_tx_metadata **pmeta,
-					   const struct xsk_tx_metadata_ops *ops,
-					   void *priv)
+static inline void
+xsk_tx_metadata_request(const struct xsk_buff_pool *pool,
+			struct xsk_tx_metadata **pmeta,
+			const struct xsk_tx_metadata_ops *ops, void *priv)
 {
 }
 
diff --git a/net/xdp/xsk_buff_pool.c b/net/xdp/xsk_buff_pool.c
index a4089480b22b..78c14f106395 100644
--- a/net/xdp/xsk_buff_pool.c
+++ b/net/xdp/xsk_buff_pool.c
@@ -765,11 +765,11 @@ EXPORT_SYMBOL(xp_raw_get_dma);
  * @addr: desc address (from userspace)
  *
  * Helper for getting desc's DMA address and metadata pointer, if present.
- * Saves one call on hotpath, double calculation of the actual address,
- * and inline checks for metadata presence and sanity.
+ * Saves one call on hotpath and double calculation of the actual address.
+ * Metadata is validated later by xsk_tx_metadata_request().
  *
  * Return: new &xdp_desc_ctx struct containing desc's DMA address and metadata
- * pointer, if it is present and valid (initialized to %NULL otherwise).
+ * pointer, if it is present (initialized to %NULL otherwise).
  */
 struct xdp_desc_ctx xp_raw_get_ctx(const struct xsk_buff_pool *pool, u64 addr)
 {
-- 
2.53.0-Meta

Re: [PATCH net 0/6] xsk: harden TX metadata validation against races

[Maciej Fijalkowski]

On Mon, Jul 27, 2026 at 09:19:53AM -0700, Stanislav Fomichev wrote:
> Cen Zhang reported a KASAN out-of-bounds read when AF_XDP is configured
> with a TX metadata area smaller than struct xsk_tx_metadata. The metadata
> is also shared with user space, so reading its flags more than once can
> produce inconsistent validation and processing decisions.
> 
> Require enough space for the flags and one request field, validate the
> launch-time field against the configured metadata length, and use one
> snapshot of the flags while processing each request. Carry the validated
> decision through completion handling so later user-space changes cannot
> enable an unrequested completion timestamp.

Stan,

I'm taking week off and I see one sashiko-reported issue which is not a
pre-existing one. Hope you could clarify it and unless no one picks the
review of this set i'll get back at it next week.

>  static inline struct xsk_tx_metadata *
>  __xsk_buff_get_metadata(const struct xsk_buff_pool *pool, void *data)
>  {
>  	struct xsk_tx_metadata *meta;
> +	u64 flags;
>  
>  	if (!pool->tx_metadata_len)
>  		return NULL;
>  
>  	meta = data - pool->tx_metadata_len;
> -	if (unlikely(!xsk_buff_valid_tx_metadata(meta)))
> +	if (unlikely(!xsk_buff_valid_tx_metadata(pool, meta, &flags)))
>  		return NULL; /* no way to signal the error to the user */
>  
>  	return meta;
The snapshotted flags are validated for size compliance in
xsk_buff_valid_tx_metadata() but then discarded, returning the
un-snapshotted user memory pointer (meta) to the driver.
Later in the zero-copy driver path, xsk_tx_metadata_request() re-reads
meta->flags directly from user memory:
include/net/xdp_sock.h:xsk_tx_metadata_request() {
    ...
    if (meta->flags & XDP_TXMD_FLAGS_LAUNCH_TIME)
        ops->tmo_request_launch_time(meta->request.launch_time, priv);
    ...
}
Does this create a Time-of-Check to Time-of-Use (TOCTOU) race condition in
the zero-copy TX metadata validation where userspace can concurrently enable
launch time after the size validation?

> 
> Reported-by: AutonomousCodeSecurity@microsoft.com
> Reported-by: Cen Zhang (Microsoft) <blbllhy@gmail.com>
> Link: https://lore.kernel.org/netdev/20260720155210.34229-1-blbllhy@gmail.com/
> 
> Stanislav Fomichev (6):
>   xsk: require at least 16 bytes of TX metadata
>   xsk: pass TX metadata pointer by reference
>   xsk: clear metadata pointer when no timestamp is requested
>   xsk: validate launch-time metadata size
>   xsk: move xsk_tx_metadata_request() to xdp_sock_drv.h
>   xsk: validate metadata when processing requests
> 
>  drivers/net/ethernet/intel/igc/igc_main.c     |  2 +-
>  drivers/net/ethernet/mellanox/mlx5/core/en.h  |  2 +-
>  .../net/ethernet/mellanox/mlx5/core/en/xdp.c  | 15 ++--
>  .../net/ethernet/mellanox/mlx5/core/en/xdp.h  |  4 +-
>  .../ethernet/mellanox/mlx5/core/en/xsk/tx.c   |  2 +-
>  .../net/ethernet/stmicro/stmmac/stmmac_main.c |  4 +-
>  include/net/libeth/xsk.h                      |  2 +-
>  include/net/xdp_sock.h                        | 45 ++---------
>  include/net/xdp_sock_drv.h                    | 77 ++++++++++++++++---
>  net/xdp/xdp_umem.c                            |  2 +
>  net/xdp/xsk.c                                 | 11 ++-
>  net/xdp/xsk_buff_pool.c                       |  6 +-
>  12 files changed, 101 insertions(+), 71 deletions(-)
> 
> -- 
> 2.53.0-Meta
>