Skip to contentSkip to conversation

net/octeontx2: fix corruption in segments list

PATCH Applied [email protected] 3 participants 3 messages Dec 21, 2020 – Jan 12, 2021 Network drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net/octeontx2: fix corruption in segments list

From: Sunil Kumar Kori <[email protected]>

On Tx, lastseg->next is not being reset to null for multi segmented packet and same mbuf can be used on Rx which has a stale mbuf entry into mbuf->next.

On Rx, application receives mbuf with mbuf->next uninitialized though mbuf->nb_segs is correct. Application iterates over all segments using mbuf->next ignoring mbuf->nb_segs which leads to undefined behavior.

So earlier assumption of just having right value in mbuf->nb_segs is enough, is incorrect. Mbuf must contain valid and synced value in nb_segs and next pointer.

Fixes 364eb0e46683 ("net/octeontx2: avoid per packet barrier with multi segment")
---
 drivers/net/octeontx2/otx2_rx.c | 6 ++++++
 drivers/net/octeontx2/otx2_rx.h | 7 +++++--
 2 files changed, 11 insertions(+), 2 deletions(-)
drivers/net/octeontx2/otx2_rx.c +6 -0
index 2da8efe77..ffeade595 100644
@@ -279,6 +279,12 @@ nix_recv_pkts_vector(void *rx_queue, struct rte_mbuf **rx_pkts,
 		vst1q_u64((uint64_t *)mbuf2->rearm_data, rearm2);
 		vst1q_u64((uint64_t *)mbuf3->rearm_data, rearm3);
 
+		/* Update that no more segments */
+		mbuf0->next = NULL;
+		mbuf1->next = NULL;
+		mbuf2->next = NULL;
+		mbuf3->next = NULL;
+
 		/* Store the mbufs to rx_pkts */
 		vst1q_u64((uint64_t *)&rx_pkts[packets], mbuf01);
 		vst1q_u64((uint64_t *)&rx_pkts[packets + 2], mbuf23);
drivers/net/octeontx2/otx2_rx.h +5 -2
index 926f614a4..0ba3d3d96 100644
@@ -215,6 +215,7 @@ nix_cqe_xtract_mseg(const struct nix_rx_parse_s *rx,
 			iova_list = (const rte_iova_t *)(iova_list + 1);
 		}
 	}
+	mbuf->next = NULL;
 }
 
 static __rte_always_inline uint16_t
@@ -330,10 +331,12 @@ otx2_nix_cqe_to_mbuf(const struct nix_cqe_hdr_s *cq, const uint32_t tag,
 	*(uint64_t *)(&mbuf->rearm_data) = val;
 	mbuf->pkt_len = len;
 
-	if (flag & NIX_RX_MULTI_SEG_F)
+	if (flag & NIX_RX_MULTI_SEG_F) {
 		nix_cqe_xtract_mseg(rx, mbuf, val);
-	else
+	} else {
 		mbuf->data_len = len;
+		mbuf->next = NULL;
+	}
 }
 
 #define CKSUM_F NIX_RX_OFFLOAD_CHECKSUM_F
Signature · 1 line
2.25.1
Acked-by on v1 1/1
#1 lore raw
Quoted 70 lines · On Mon, Dec 21, 2020 at 07:33:08PM +0530, [email protected] wrote:

From: Sunil Kumar Kori <[email protected]>

On Tx, lastseg->next is not being reset to null for multi segmented packet and same mbuf can be used on Rx which has a stale mbuf entry into mbuf->next.

On Rx, application receives mbuf with mbuf->next uninitialized though mbuf->nb_segs is correct. Application iterates over all segments using mbuf->next ignoring mbuf->nb_segs which leads to undefined behavior.

So earlier assumption of just having right value in mbuf->nb_segs is enough, is incorrect. Mbuf must contain valid and synced value in nb_segs and next pointer.

Fixes 364eb0e46683 ("net/octeontx2: avoid per packet barrier with multi segment")
---
 drivers/net/octeontx2/otx2_rx.c | 6 ++++++
 drivers/net/octeontx2/otx2_rx.h | 7 +++++--
 2 files changed, 11 insertions(+), 2 deletions(-)
drivers/net/octeontx2/otx2_rx.c +6 -0
index 2da8efe77..ffeade595 100644
@@ -279,6 +279,12 @@ nix_recv_pkts_vector(void *rx_queue, struct rte_mbuf **rx_pkts,
 		vst1q_u64((uint64_t *)mbuf2->rearm_data, rearm2);
 		vst1q_u64((uint64_t *)mbuf3->rearm_data, rearm3);
 
+		/* Update that no more segments */
+		mbuf0->next = NULL;
+		mbuf1->next = NULL;
+		mbuf2->next = NULL;
+		mbuf3->next = NULL;
+
 		/* Store the mbufs to rx_pkts */
 		vst1q_u64((uint64_t *)&rx_pkts[packets], mbuf01);
 		vst1q_u64((uint64_t *)&rx_pkts[packets + 2], mbuf23);
drivers/net/octeontx2/otx2_rx.h +5 -2
index 926f614a4..0ba3d3d96 100644
@@ -215,6 +215,7 @@ nix_cqe_xtract_mseg(const struct nix_rx_parse_s *rx,
 			iova_list = (const rte_iova_t *)(iova_list + 1);
 		}
 	}
+	mbuf->next = NULL;
 }
 
 static __rte_always_inline uint16_t
@@ -330,10 +331,12 @@ otx2_nix_cqe_to_mbuf(const struct nix_cqe_hdr_s *cq, const uint32_t tag,
 	*(uint64_t *)(&mbuf->rearm_data) = val;
 	mbuf->pkt_len = len;
 
-	if (flag & NIX_RX_MULTI_SEG_F)
+	if (flag & NIX_RX_MULTI_SEG_F) {
 		nix_cqe_xtract_mseg(rx, mbuf, val);
-	else
+	} else {
 		mbuf->data_len = len;
+		mbuf->next = NULL;
+	}
 }
 
 #define CKSUM_F NIX_RX_OFFLOAD_CHECKSUM_F
Signature · 1 line
2.25.1
Applied
#2 lore raw
Quoted 16 lines · Nithin Dabilpuram
Quoted 14 lines · On Mon, Dec 21, 2020 at 07:33:08PM +0530, [email protected] wrote:

From: Sunil Kumar Kori <[email protected]>

On Tx, lastseg->next is not being reset to null for multi segmented packet and same mbuf can be used on Rx which has a stale mbuf entry into mbuf->next.

On Rx, application receives mbuf with mbuf->next uninitialized though mbuf->nb_segs is correct. Application iterates over all segments using mbuf->next ignoring mbuf->nb_segs which leads to undefined behavior.

So earlier assumption of just having right value in mbuf->nb_segs is enough, is incorrect. Mbuf must contain valid and synced value in nb_segs and next pointer.

Quoted 8 lines
Quoted 6 lines
Fixes 364eb0e46683 ("net/octeontx2: avoid per packet barrier with multi segment")

Applied to dpdk-next-net-mrvl/for-main. Thanks