From: luca.boccassi@gmail.com
To: Alexander Kozyrev <akozyrev@nvidia.com>
Cc: Matan Azrad <matan@nvidia.com>, dpdk stable <stable@dpdk.org>
Subject: patch 'net/mlx5: ignore non-critical syndromes for Rx queue' has been queued to stable release 20.11.8
Date: Thu, 23 Feb 2023 09:36:57 +0000 [thread overview]
Message-ID: <20230223093715.3926893-53-luca.boccassi@gmail.com> (raw)
In-Reply-To: <20230223093715.3926893-1-luca.boccassi@gmail.com>
Hi,
FYI, your patch has been queued to stable release 20.11.8
Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 02/25/23. So please
shout if anyone has objections.
Also note that after the patch there's a diff of the upstream commit vs the
patch applied to the branch. This will indicate if there was any rebasing
needed to apply to the stable branch. If there were code changes for rebasing
(ie: not only metadata diffs), please double check that the rebase was
correctly done.
Queued patches are on a temporary branch at:
https://github.com/bluca/dpdk-stable
This queued commit can be viewed at:
https://github.com/bluca/dpdk-stable/commit/aad5672767479652695f199cdb26dcdae7e5f5e6
Thanks.
Luca Boccassi
---
From aad5672767479652695f199cdb26dcdae7e5f5e6 Mon Sep 17 00:00:00 2001
From: Alexander Kozyrev <akozyrev@nvidia.com>
Date: Fri, 27 Jan 2023 05:22:43 +0200
Subject: [PATCH] net/mlx5: ignore non-critical syndromes for Rx queue
[ upstream commit aa67ed3084588e6ca12e9709a6cab021f0ffeba7 ]
For non-fatal syndromes like LOCAL_LENGTH_ERR, the Rx queue reset
shouldn't be triggered. Rx queue could continue with the next packets
without any recovery. Only three syndromes warrant Rx queue reset:
LOCAL_QP_OP_ERR, LOCAL_PROT_ERR and WR_FLUSH_ERR.
Do not initiate a Rx queue reset in any other cases.
Skip all non-critical error CQEs and continue with packet processing.
Fixes: 88c0733535 ("net/mlx5: extend Rx completion with error handling")
Signed-off-by: Alexander Kozyrev <akozyrev@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
---
drivers/net/mlx5/mlx5_rxtx.c | 123 ++++++++++++++++++++++++-------
drivers/net/mlx5/mlx5_rxtx.h | 5 +-
drivers/net/mlx5/mlx5_rxtx_vec.c | 3 +-
3 files changed, 102 insertions(+), 29 deletions(-)
diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 1422961d62..f7c8b8c076 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -86,7 +86,8 @@ rxq_cq_to_pkt_type(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
static __rte_always_inline int
mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
- uint16_t cqe_cnt, volatile struct mlx5_mini_cqe8 **mcqe);
+ uint16_t cqe_cnt, volatile struct mlx5_mini_cqe8 **mcqe,
+ uint16_t *skip_cnt, bool mprq);
static __rte_always_inline uint32_t
rxq_cq_to_ol_flags(volatile struct mlx5_cqe *cqe);
@@ -983,10 +984,14 @@ mlx5_queue_state_modify(struct rte_eth_dev *dev,
return ret;
}
+#define MLX5_ERROR_CQE_MASK 0x40000000
/* Must be negative. */
-#define MLX5_ERROR_CQE_RET (-1)
+#define MLX5_REGULAR_ERROR_CQE_RET (-5)
+#define MLX5_CRITICAL_ERROR_CQE_RET (-4)
/* Must not be negative. */
#define MLX5_RECOVERY_ERROR_RET 0
+#define MLX5_RECOVERY_IGNORE_RET 1
+#define MLX5_RECOVERY_COMPLETED_RET 2
/**
* Handle a Rx error.
@@ -1004,10 +1009,14 @@ mlx5_queue_state_modify(struct rte_eth_dev *dev,
* Number of CQEs to check for an error.
*
* @return
- * MLX5_RECOVERY_ERROR_RET in case of recovery error, otherwise the CQE status.
+ * MLX5_RECOVERY_ERROR_RET in case of recovery error,
+ * MLX5_RECOVERY_IGNORE_RET in case of non-critical error syndrome,
+ * MLX5_RECOVERY_COMPLETED_RET in case of recovery is completed,
+ * otherwise the CQE status after ignored error syndrome or queue reset.
*/
int
-mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
+mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec,
+ uint16_t err_n, uint16_t *skip_cnt)
{
const uint16_t cqe_n = 1 << rxq->cqe_n;
const uint16_t cqe_mask = cqe_n - 1;
@@ -1022,14 +1031,35 @@ mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
.cqe = &(*rxq->cqes)[(rxq->cq_ci - vec) & cqe_mask],
};
struct mlx5_mp_arg_queue_state_modify sm;
+ bool critical_syndrome = false;
int ret, i;
switch (rxq->err_state) {
+ case MLX5_RXQ_ERR_STATE_IGNORE:
+ ret = check_cqe(u.cqe, cqe_n, rxq->cq_ci - vec);
+ if (ret != MLX5_CQE_STATUS_ERR) {
+ rxq->err_state = MLX5_RXQ_ERR_STATE_NO_ERROR;
+ return ret;
+ }
+ /* Fall-through */
case MLX5_RXQ_ERR_STATE_NO_ERROR:
for (i = 0; i < (int)err_n; i++) {
u.cqe = &(*rxq->cqes)[(rxq->cq_ci - vec - i) & cqe_mask];
- if (MLX5_CQE_OPCODE(u.cqe->op_own) == MLX5_CQE_RESP_ERR)
+ if (MLX5_CQE_OPCODE(u.cqe->op_own) == MLX5_CQE_RESP_ERR) {
+ if (u.err_cqe->syndrome == MLX5_CQE_SYNDROME_LOCAL_QP_OP_ERR ||
+ u.err_cqe->syndrome == MLX5_CQE_SYNDROME_LOCAL_PROT_ERR ||
+ u.err_cqe->syndrome == MLX5_CQE_SYNDROME_WR_FLUSH_ERR)
+ critical_syndrome = true;
break;
+ }
+ }
+ if (!critical_syndrome) {
+ if (rxq->err_state == MLX5_RXQ_ERR_STATE_NO_ERROR) {
+ *skip_cnt = 0;
+ if (i == err_n)
+ rxq->err_state = MLX5_RXQ_ERR_STATE_IGNORE;
+ }
+ return MLX5_RECOVERY_IGNORE_RET;
}
rxq->err_state = MLX5_RXQ_ERR_STATE_NEED_RESET;
/* Fall-through */
@@ -1122,6 +1152,7 @@ mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
}
mlx5_rxq_initialize(rxq);
rxq->err_state = MLX5_RXQ_ERR_STATE_NO_ERROR;
+ return MLX5_RECOVERY_COMPLETED_RET;
}
return ret;
default:
@@ -1141,19 +1172,24 @@ mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
* @param[out] mcqe
* Store pointer to mini-CQE if compressed. Otherwise, the pointer is not
* written.
- *
+ * @param[out] skip_cnt
+ * Number of packets skipped due to recoverable errors.
+ * @param mprq
+ * Indication if it is called from MPRQ.
* @return
- * 0 in case of empty CQE, MLX5_ERROR_CQE_RET in case of error CQE,
- * otherwise the packet size in regular RxQ, and striding byte
- * count format in mprq case.
+ * 0 in case of empty CQE, MLX5_REGULAR_ERROR_CQE_RET in case of error CQE,
+ * MLX5_CRITICAL_ERROR_CQE_RET in case of error CQE lead to Rx queue reset,
+ * otherwise the packet size in regular RxQ,
+ * and striding byte count format in mprq case.
*/
static inline int
mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
- uint16_t cqe_cnt, volatile struct mlx5_mini_cqe8 **mcqe)
+ uint16_t cqe_cnt, volatile struct mlx5_mini_cqe8 **mcqe,
+ uint16_t *skip_cnt, bool mprq)
{
struct rxq_zip *zip = &rxq->zip;
uint16_t cqe_n = cqe_cnt + 1;
- int len;
+ int len = 0, ret = 0;
uint16_t idx, end;
do {
@@ -1202,7 +1238,6 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
* compressed.
*/
} else {
- int ret;
int8_t op_own;
uint32_t cq_ci;
@@ -1210,10 +1245,12 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
if (unlikely(ret != MLX5_CQE_STATUS_SW_OWN)) {
if (unlikely(ret == MLX5_CQE_STATUS_ERR ||
rxq->err_state)) {
- ret = mlx5_rx_err_handle(rxq, 0, 1);
- if (ret == MLX5_CQE_STATUS_HW_OWN ||
- ret == MLX5_RECOVERY_ERROR_RET)
- return MLX5_ERROR_CQE_RET;
+ ret = mlx5_rx_err_handle(rxq, 0, 1, skip_cnt);
+ if (ret == MLX5_CQE_STATUS_HW_OWN)
+ return MLX5_ERROR_CQE_MASK;
+ if (ret == MLX5_RECOVERY_ERROR_RET ||
+ ret == MLX5_RECOVERY_COMPLETED_RET)
+ return MLX5_CRITICAL_ERROR_CQE_RET;
} else {
return 0;
}
@@ -1266,8 +1303,15 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
}
}
if (unlikely(rxq->err_state)) {
+ if (rxq->err_state == MLX5_RXQ_ERR_STATE_IGNORE &&
+ ret == MLX5_CQE_STATUS_SW_OWN) {
+ rxq->err_state = MLX5_RXQ_ERR_STATE_NO_ERROR;
+ return len & MLX5_ERROR_CQE_MASK;
+ }
cqe = &(*rxq->cqes)[rxq->cq_ci & cqe_cnt];
++rxq->stats.idropped;
+ (*skip_cnt) += mprq ? (len & MLX5_MPRQ_STRIDE_NUM_MASK) >>
+ MLX5_MPRQ_STRIDE_NUM_SHIFT : 1;
} else {
return len;
}
@@ -1418,6 +1462,7 @@ mlx5_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
int len = 0; /* keep its value across iterations. */
while (pkts_n) {
+ uint16_t skip_cnt;
unsigned int idx = rq_ci & wqe_cnt;
volatile struct mlx5_wqe_data_seg *wqe =
&((volatile struct mlx5_wqe_data_seg *)rxq->wqes)[idx];
@@ -1456,11 +1501,24 @@ mlx5_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
}
if (!pkt) {
cqe = &(*rxq->cqes)[rxq->cq_ci & cqe_cnt];
- len = mlx5_rx_poll_len(rxq, cqe, cqe_cnt, &mcqe);
- if (len <= 0) {
- rte_mbuf_raw_free(rep);
- if (unlikely(len == MLX5_ERROR_CQE_RET))
+ len = mlx5_rx_poll_len(rxq, cqe, cqe_cnt, &mcqe, &skip_cnt, false);
+ if (unlikely(len & MLX5_ERROR_CQE_MASK)) {
+ if (len == MLX5_CRITICAL_ERROR_CQE_RET) {
+ rte_mbuf_raw_free(rep);
rq_ci = rxq->rq_ci << sges_n;
+ break;
+ }
+ rq_ci >>= sges_n;
+ rq_ci += skip_cnt;
+ rq_ci <<= sges_n;
+ idx = rq_ci & wqe_cnt;
+ wqe = &((volatile struct mlx5_wqe_data_seg *)rxq->wqes)[idx];
+ seg = (*rxq->elts)[idx];
+ cqe = &(*rxq->cqes)[rxq->cq_ci & cqe_cnt];
+ len = len & ~MLX5_ERROR_CQE_MASK;
+ }
+ if (len == 0) {
+ rte_mbuf_raw_free(rep);
break;
}
pkt = seg;
@@ -1684,6 +1742,7 @@ mlx5_rx_burst_mprq(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
uint16_t strd_cnt;
uint16_t strd_idx;
uint32_t byte_cnt;
+ uint16_t skip_cnt;
volatile struct mlx5_mini_cqe8 *mcqe = NULL;
enum mlx5_rqx_code rxq_code;
@@ -1696,14 +1755,26 @@ mlx5_rx_burst_mprq(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
buf = (*rxq->mprq_bufs)[rq_ci & wq_mask];
}
cqe = &(*rxq->cqes)[rxq->cq_ci & cq_mask];
- ret = mlx5_rx_poll_len(rxq, cqe, cq_mask, &mcqe);
+ ret = mlx5_rx_poll_len(rxq, cqe, cq_mask, &mcqe, &skip_cnt, true);
+ if (unlikely(ret & MLX5_ERROR_CQE_MASK)) {
+ if (ret == MLX5_CRITICAL_ERROR_CQE_RET) {
+ rq_ci = rxq->rq_ci;
+ consumed_strd = rxq->consumed_strd;
+ break;
+ }
+ consumed_strd += skip_cnt;
+ while (consumed_strd >= strd_n) {
+ /* Replace WQE if the buffer is still in use. */
+ mprq_buf_replace(rxq, rq_ci & wq_mask);
+ /* Advance to the next WQE. */
+ consumed_strd -= strd_n;
+ ++rq_ci;
+ buf = (*rxq->mprq_bufs)[rq_ci & wq_mask];
+ }
+ cqe = &(*rxq->cqes)[rxq->cq_ci & cq_mask];
+ }
if (ret == 0)
break;
- if (unlikely(ret == MLX5_ERROR_CQE_RET)) {
- rq_ci = rxq->rq_ci;
- consumed_strd = rxq->consumed_strd;
- break;
- }
byte_cnt = ret;
len = (byte_cnt & MLX5_MPRQ_LEN_MASK) >> MLX5_MPRQ_LEN_SHIFT;
MLX5_ASSERT((int)len >= (rxq->crc_present << 2));
diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index 7fa471a651..b5fe330a71 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -91,6 +91,7 @@ enum mlx5_rxq_err_state {
MLX5_RXQ_ERR_STATE_NO_ERROR = 0,
MLX5_RXQ_ERR_STATE_NEED_RESET,
MLX5_RXQ_ERR_STATE_NEED_READY,
+ MLX5_RXQ_ERR_STATE_IGNORE,
};
enum mlx5_rqx_code {
@@ -427,8 +428,8 @@ void mlx5_set_cksum_table(void);
void mlx5_set_swp_types_table(void);
uint16_t mlx5_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n);
void mlx5_rxq_initialize(struct mlx5_rxq_data *rxq);
-__rte_noinline int mlx5_rx_err_handle(struct mlx5_rxq_data *rxq,
- uint8_t vec, uint16_t err_n);
+__rte_noinline int mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec,
+ uint16_t err_n, uint16_t *skip_cnt);
void mlx5_mprq_buf_free_cb(void *addr, void *opaque);
void mlx5_mprq_buf_free(struct mlx5_mprq_buf *buf);
uint16_t mlx5_rx_burst_mprq(void *dpdk_rxq, struct rte_mbuf **pkts,
diff --git a/drivers/net/mlx5/mlx5_rxtx_vec.c b/drivers/net/mlx5/mlx5_rxtx_vec.c
index ca0f585863..4a1dc3c3e1 100644
--- a/drivers/net/mlx5/mlx5_rxtx_vec.c
+++ b/drivers/net/mlx5/mlx5_rxtx_vec.c
@@ -50,6 +50,7 @@ rxq_handle_pending_error(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts,
uint16_t pkts_n)
{
uint16_t n = 0;
+ uint16_t skip_cnt;
unsigned int i;
#ifdef MLX5_PMD_SOFT_COUNTERS
uint32_t err_bytes = 0;
@@ -73,7 +74,7 @@ rxq_handle_pending_error(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts,
rxq->stats.ipackets -= (pkts_n - n);
rxq->stats.ibytes -= err_bytes;
#endif
- mlx5_rx_err_handle(rxq, 1, pkts_n);
+ mlx5_rx_err_handle(rxq, 1, pkts_n, &skip_cnt);
return n;
}
--
2.39.1
---
Diff of the applied patch vs upstream commit (please double-check if non-empty:
---
--- - 2023-02-23 09:36:30.351277669 +0000
+++ 0053-net-mlx5-ignore-non-critical-syndromes-for-Rx-queue.patch 2023-02-23 09:36:28.318171634 +0000
@@ -1 +1 @@
-From aa67ed3084588e6ca12e9709a6cab021f0ffeba7 Mon Sep 17 00:00:00 2001
+From aad5672767479652695f199cdb26dcdae7e5f5e6 Mon Sep 17 00:00:00 2001
@@ -5,0 +6,2 @@
+[ upstream commit aa67ed3084588e6ca12e9709a6cab021f0ffeba7 ]
+
@@ -14 +15,0 @@
-Cc: stable@dpdk.org
@@ -19,2 +20,2 @@
- drivers/net/mlx5/mlx5_rx.c | 123 ++++++++++++++++++++++++-------
- drivers/net/mlx5/mlx5_rx.h | 5 +-
+ drivers/net/mlx5/mlx5_rxtx.c | 123 ++++++++++++++++++++++++-------
+ drivers/net/mlx5/mlx5_rxtx.h | 5 +-
@@ -24,5 +25,5 @@
-diff --git a/drivers/net/mlx5/mlx5_rx.c b/drivers/net/mlx5/mlx5_rx.c
-index 7612d15f01..99a08ef5f1 100644
---- a/drivers/net/mlx5/mlx5_rx.c
-+++ b/drivers/net/mlx5/mlx5_rx.c
-@@ -39,7 +39,8 @@ rxq_cq_to_pkt_type(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
+diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
+index 1422961d62..f7c8b8c076 100644
+--- a/drivers/net/mlx5/mlx5_rxtx.c
++++ b/drivers/net/mlx5/mlx5_rxtx.c
+@@ -86,7 +86,8 @@ rxq_cq_to_pkt_type(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
@@ -38,2 +39,2 @@
-@@ -408,10 +409,14 @@ mlx5_rxq_initialize(struct mlx5_rxq_data *rxq)
- *rxq->rq_db = rte_cpu_to_be_32(rxq->rq_ci);
+@@ -983,10 +984,14 @@ mlx5_queue_state_modify(struct rte_eth_dev *dev,
+ return ret;
@@ -54 +55 @@
-@@ -429,10 +434,14 @@ mlx5_rxq_initialize(struct mlx5_rxq_data *rxq)
+@@ -1004,10 +1009,14 @@ mlx5_queue_state_modify(struct rte_eth_dev *dev,
@@ -71 +72 @@
-@@ -447,14 +456,35 @@ mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
+@@ -1022,14 +1031,35 @@ mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
@@ -108 +109 @@
-@@ -546,6 +576,7 @@ mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
+@@ -1122,6 +1152,7 @@ mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
@@ -116 +117 @@
-@@ -565,19 +596,24 @@ mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
+@@ -1141,19 +1172,24 @@ mlx5_rx_err_handle(struct mlx5_rxq_data *rxq, uint8_t vec, uint16_t err_n)
@@ -147 +148 @@
-@@ -626,7 +662,6 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
+@@ -1202,7 +1238,6 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
@@ -155 +156 @@
-@@ -634,10 +669,12 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
+@@ -1210,10 +1245,12 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
@@ -172 +173 @@
-@@ -690,8 +727,15 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
+@@ -1266,8 +1303,15 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
@@ -188 +189 @@
-@@ -843,6 +887,7 @@ mlx5_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
+@@ -1418,6 +1462,7 @@ mlx5_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
@@ -196 +197 @@
-@@ -881,11 +926,24 @@ mlx5_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
+@@ -1456,11 +1501,24 @@ mlx5_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
@@ -225 +226 @@
-@@ -1095,6 +1153,7 @@ mlx5_rx_burst_mprq(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
+@@ -1684,6 +1742,7 @@ mlx5_rx_burst_mprq(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
@@ -233 +234 @@
-@@ -1107,14 +1166,26 @@ mlx5_rx_burst_mprq(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
+@@ -1696,14 +1755,26 @@ mlx5_rx_burst_mprq(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
@@ -266,5 +267,5 @@
-diff --git a/drivers/net/mlx5/mlx5_rx.h b/drivers/net/mlx5/mlx5_rx.h
-index 4ba53ebc48..6b42e27c89 100644
---- a/drivers/net/mlx5/mlx5_rx.h
-+++ b/drivers/net/mlx5/mlx5_rx.h
-@@ -62,6 +62,7 @@ enum mlx5_rxq_err_state {
+diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
+index 7fa471a651..b5fe330a71 100644
+--- a/drivers/net/mlx5/mlx5_rxtx.h
++++ b/drivers/net/mlx5/mlx5_rxtx.h
+@@ -91,6 +91,7 @@ enum mlx5_rxq_err_state {
@@ -278,2 +279,2 @@
-@@ -286,8 +287,8 @@ int mlx5_hrxq_modify(struct rte_eth_dev *dev, uint32_t hxrq_idx,
-
+@@ -427,8 +428,8 @@ void mlx5_set_cksum_table(void);
+ void mlx5_set_swp_types_table(void);
@@ -285,0 +287 @@
+ void mlx5_mprq_buf_free_cb(void *addr, void *opaque);
@@ -288 +289,0 @@
- uint16_t pkts_n);
@@ -290 +291 @@
-index c6be2be763..667475a93e 100644
+index ca0f585863..4a1dc3c3e1 100644
@@ -293 +294 @@
-@@ -51,6 +51,7 @@ rxq_handle_pending_error(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts,
+@@ -50,6 +50,7 @@ rxq_handle_pending_error(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts,
@@ -301 +302 @@
-@@ -74,7 +75,7 @@ rxq_handle_pending_error(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts,
+@@ -73,7 +74,7 @@ rxq_handle_pending_error(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts,
next prev parent reply other threads:[~2023-02-23 9:40 UTC|newest]
Thread overview: 144+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-23 9:36 patch 'eal/windows: mark memory config as complete' " luca.boccassi
2023-02-23 9:36 ` patch 'kni: fix build on RHEL 9.1' " luca.boccassi
2023-02-23 9:36 ` patch 'doc: fix dependency setup in l2fwd-cat example guide' " luca.boccassi
2023-02-23 9:36 ` patch 'devtools: fix escaped space in grep pattern' " luca.boccassi
2023-02-23 9:36 ` patch 'app/crypto-perf: fix number of segments' " luca.boccassi
2023-02-23 9:36 ` patch 'eventdev/eth_tx: fix devices loop' " luca.boccassi
2023-02-23 9:36 ` patch 'crypto/qat: fix stream cipher direction' " luca.boccassi
2023-02-23 9:36 ` patch 'fbarray: fix metadata dump' " luca.boccassi
2023-02-23 9:36 ` patch 'graph: fix node shrink' " luca.boccassi
2023-02-23 9:36 ` patch 'net/nfp: fix firmware name derived from PCI name' " luca.boccassi
2023-02-23 9:36 ` patch 'app/testpmd: fix interactive mode with no ports' " luca.boccassi
2023-02-23 9:36 ` patch 'examples/qos_sched: fix debug mode' " luca.boccassi
2023-02-23 9:36 ` patch 'build: fix dependencies lookup' " luca.boccassi
2023-02-23 9:36 ` patch 'vdpa/ifc: fix argument compatibility check' " luca.boccassi
2023-02-23 9:36 ` patch 'vdpa/ifc: fix reconnection in SW-assisted live migration' " luca.boccassi
2023-02-23 9:36 ` patch 'vhost: fix net header settings in datapath' " luca.boccassi
2023-02-23 9:36 ` patch 'app/bbdev: add allocation checks' " luca.boccassi
2023-02-23 9:36 ` patch 'baseband/acc: fix memory leak on acc100 close' " luca.boccassi
2023-02-23 9:36 ` patch 'baseband/acc: fix acc100 iteration counter in TB' " luca.boccassi
2023-02-23 9:36 ` patch 'crypto/ccp: remove some printf' " luca.boccassi
2023-02-23 9:36 ` patch 'test/crypto: add missing MAC-I to PDCP vectors' " luca.boccassi
2023-02-23 9:36 ` patch 'compressdev: fix end of driver list' " luca.boccassi
2023-02-23 9:36 ` patch 'net/bnxt: fix Tx queue stats after queue stop and start' " luca.boccassi
2023-02-23 9:36 ` patch 'net/bnxt: fix Rx " luca.boccassi
2023-02-23 9:36 ` patch 'net/bnxt: fix RSS hash in mbuf' " luca.boccassi
2023-02-23 9:36 ` patch 'mem: fix hugepage info mapping' " luca.boccassi
2023-02-23 9:36 ` patch 'raw/ifpga/base: fix init with multi-process' " luca.boccassi
2023-02-23 9:36 ` patch 'telemetry: fix repeat display when callback don't init dict' " luca.boccassi
2023-02-23 9:36 ` patch 'test/mbuf: fix mbuf reset test' " luca.boccassi
2023-02-23 9:36 ` patch 'eventdev/timer: fix overflow' " luca.boccassi
2023-02-23 9:36 ` patch 'vhost: decrease log level for unimplemented requests' " luca.boccassi
2023-02-23 9:36 ` patch 'vhost: fix possible FD leaks' " luca.boccassi
2023-02-23 9:36 ` patch 'vhost: fix possible FD leaks on truncation' " luca.boccassi
2023-02-23 9:36 ` patch 'net/virtio-user: fix device starting failure handling' " luca.boccassi
2023-02-23 9:36 ` patch 'app/testpmd: fix forwarding stats for Tx dropped' " luca.boccassi
2023-02-23 9:36 ` patch 'net/txgbe: fix default signal quality value for KX/KX4' " luca.boccassi
2023-02-23 9:36 ` patch 'net/txgbe: fix packet type to parse from offload flags' " luca.boccassi
2023-02-23 9:36 ` patch 'net/txgbe: fix interrupt loss' " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: fix log about indirection table size' " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: refactor set RSS hash algorithm and key interface' " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: fix RSS key size compatibility' " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: fix clearing RSS configuration' " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: use RSS filter list to check duplicated rule' " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: remove useless code when destroy valid RSS " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: fix warning on flush or destroy " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: fix config struct used for conversion' " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: fix duplicate RSS rule check' " luca.boccassi
2023-02-23 9:36 ` patch 'net/hns3: extract common functions to set Rx/Tx' " luca.boccassi
2023-02-23 9:36 ` patch 'net/sfc: enforce fate action in transfer flow rules' " luca.boccassi
2023-02-23 9:36 ` patch 'net/txgbe: fix Rx buffer size in config register' " luca.boccassi
2023-02-23 9:36 ` patch 'net/mlx5: fix flow sample with ConnectX-5' " luca.boccassi
2023-02-23 9:36 ` patch 'net/mlx5: fix error CQE dumping for vectorized Rx' " luca.boccassi
2023-02-23 9:36 ` luca.boccassi [this message]
2023-02-23 9:36 ` patch 'net/i40e: reduce interrupt interval in multi-driver mode' " luca.boccassi
2023-02-23 9:36 ` patch 'net/ixgbe: fix firmware version consistency' " luca.boccassi
2023-02-23 9:37 ` patch 'net/iavf: add lock for VF commands' " luca.boccassi
2023-02-23 9:37 ` patch 'net/i40e: fix validation of flow transfer attribute' " luca.boccassi
2023-02-23 9:37 ` patch 'net/ice: " luca.boccassi
2023-02-23 9:37 ` patch 'net/iavf: protect insertion in flow list' " luca.boccassi
2023-02-23 9:37 ` patch 'net/ixgbe: enable IPv6 mask in flow rules' " luca.boccassi
2023-02-23 9:37 ` patch 'app/compress-perf: fix some typos' " luca.boccassi
2023-02-23 9:37 ` patch 'app/compress-perf: fix testing single operation' " luca.boccassi
2023-02-23 9:37 ` patch 'net/bnxt: fix link state change interrupt config' " luca.boccassi
2023-02-23 9:37 ` patch 'app/testpmd: fix crash on cleanup' " luca.boccassi
2023-02-23 9:37 ` patch 'eal/freebsd: fix lock in alarm callback' " luca.boccassi
2023-02-23 9:37 ` patch 'reorder: invalidate buffer from ready queue in drain' " luca.boccassi
2023-02-23 9:37 ` patch 'test/reorder: fix double free of drained buffers' " luca.boccassi
2023-02-23 9:37 ` patch 'build: fix toolchain definition' " luca.boccassi
2023-02-23 9:37 ` patch 'eal: use same atomic intrinsics for GCC and clang' " luca.boccassi
2023-02-23 9:37 ` patch 'examples/cmdline: fix build with GCC 12' " luca.boccassi
2023-02-23 9:37 ` patch 'examples/qos_sched: fix Tx port config when link down' " luca.boccassi
2023-03-15 22:45 ` patch 'eal/windows: fix pedantic build' " luca.boccassi
2023-03-15 22:45 ` patch 'doc: fix reference to event timer header' " luca.boccassi
2023-03-15 22:45 ` patch 'test/bbdev: fix crash for non supported HARQ length' " luca.boccassi
2023-03-15 22:45 ` patch 'test/bbdev: extend HARQ tolerance' " luca.boccassi
2023-03-15 22:45 ` patch 'test/bbdev: remove check for invalid opaque data' " luca.boccassi
2023-03-15 22:45 ` patch 'vhost: fix OOB access for invalid vhost ID' " luca.boccassi
2023-03-16 9:27 ` Luca Boccassi
2023-03-16 9:56 ` David Marchand
2023-03-16 10:30 ` Kevin Traynor
2023-03-16 10:36 ` Luca Boccassi
2023-03-16 10:50 ` David Marchand
2023-03-15 22:45 ` patch 'Revert "vhost: fix OOB access for invalid vhost ID"' " luca.boccassi
2023-03-15 22:45 ` patch 'net/virtio: deduce IP length for TSO checksum' " luca.boccassi
2023-03-15 22:46 ` patch 'app/testpmd: fix Tx preparation in checksum engine' " luca.boccassi
2023-03-15 22:46 ` patch 'app/testpmd: fix packet count in IEEE 1588 " luca.boccassi
2023-03-15 22:46 ` patch 'app/testpmd: fix packet transmission in noisy VNF " luca.boccassi
2023-03-15 22:46 ` patch 'net/nfp: fix getting RSS configuration' " luca.boccassi
2023-03-15 22:46 ` patch 'net/ixgbe: fix IPv6 mask in flow director' " luca.boccassi
2023-03-15 22:46 ` patch 'net/i40e: revert link status check on device start' " luca.boccassi
2023-03-15 22:46 ` patch 'net/nfp: fix MTU configuration order' " luca.boccassi
2023-03-15 22:46 ` patch 'kvargs: add API documentation for process callback' " luca.boccassi
2023-03-15 22:46 ` patch 'compressdev: fix empty devargs parsing' " luca.boccassi
2023-03-15 22:46 ` patch 'cryptodev: " luca.boccassi
2023-03-15 22:46 ` patch 'net/virtio: " luca.boccassi
2023-03-15 22:46 ` patch 'raw/skeleton: " luca.boccassi
2023-03-15 22:46 ` patch 'kni: fix possible starvation when mbufs are exhausted' " luca.boccassi
2023-03-15 22:46 ` patch 'cmdline: handle EOF as quit' " luca.boccassi
2023-03-15 22:46 ` patch 'app/testpmd: cleanup cleanly from signal' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: fix possible truncation of hash key when config' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: fix possible truncation of redirection table' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: use hardware config to report hash key' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: use hardware config to report hash types' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: use hardware config to report redirection table' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: separate setting hash algorithm' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: separate setting hash key' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: separate setting redirection table' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: separate setting RSS types' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: separate setting and clearing RSS rule' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: use new RSS rule to configure hardware' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: save hash algo to RSS filter list node' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: allow adding queue buffer size hash rule' " luca.boccassi
2023-03-15 22:46 ` patch 'net/hns3: separate flow RSS config from RSS conf' " luca.boccassi
2023-03-15 22:46 ` patch 'app/crypto-perf: fix test file memory leak' " luca.boccassi
2023-03-15 22:46 ` patch 'app/flow-perf: fix division or module by zero' " luca.boccassi
2023-03-22 0:41 ` patch 'raw/skeleton: fix selftest' " luca.boccassi
2023-03-22 0:41 ` patch 'ring: silence GCC 12 warnings' " luca.boccassi
2023-03-22 0:41 ` patch 'reorder: fix sequence number mbuf field register' " luca.boccassi
2023-03-22 0:41 ` patch 'test: fix segment length in packet generator' " luca.boccassi
2023-03-22 0:41 ` patch 'test/mbuf: fix test with mbuf debug enabled' " luca.boccassi
2023-03-22 0:41 ` patch 'app/testpmd: cleanup cleanly from signal' " luca.boccassi
2023-03-22 0:41 ` patch 'app/testpmd: fix interactive mode on Windows' " luca.boccassi
2023-03-22 0:41 ` patch 'app/compress-perf: fix remaining data for ops' " luca.boccassi
2023-03-22 0:41 ` patch 'app/bbdev: check statistics failure' " luca.boccassi
2023-03-22 0:41 ` patch 'net/vhost: add missing newline in logs' " luca.boccassi
2023-03-22 0:41 ` patch 'net/virtio: remove address width limit for modern devices' " luca.boccassi
2023-03-22 0:41 ` patch 'net/e1000: fix saving of stripped VLAN TCI' " luca.boccassi
2023-03-22 0:41 ` patch 'net/i40e: fix MAC loopback on X722' " luca.boccassi
2023-03-22 0:41 ` patch 'net/iavf: fix device stop during reset' " luca.boccassi
2023-03-22 0:41 ` patch 'net/mlx5: fix hairpin Tx queue reference count' " luca.boccassi
2023-03-22 0:41 ` patch 'common/mlx5: use just sufficient barrier for Arm' " luca.boccassi
2023-03-22 0:41 ` patch 'bus/ifpga: fix devargs handling' " luca.boccassi
2023-03-22 0:41 ` patch 'net/ipn3ke: fix thread exit' " luca.boccassi
2023-03-22 0:41 ` patch 'net/ipn3ke: fix representor name' " luca.boccassi
2023-03-29 1:04 ` patch 'examples/qos_sched: fix config entries in wrong sections' " luca.boccassi
2023-03-29 1:04 ` patch 'net/mlx5: fix build with GCC 12 and ASan' " luca.boccassi
2023-03-29 1:04 ` patch 'net/mlx5: fix CQE dump for Tx' " luca.boccassi
2023-03-29 1:04 ` patch 'net/mlx5: fix sysfs port name translation' " luca.boccassi
2023-03-29 1:04 ` patch 'doc: fix code blocks in cryptodev guide' " luca.boccassi
2023-03-29 1:04 ` patch 'test/crypto: fix statistics error messages' " luca.boccassi
2023-03-29 1:04 ` patch 'pdump: fix build with GCC 12' " luca.boccassi
2023-03-29 1:04 ` patch 'acl: fix crash on PPC64 with GCC 11' " luca.boccassi
2023-03-29 1:04 ` patch 'doc: fix pipeline example path in user guide' " luca.boccassi
2023-03-29 1:04 ` patch 'doc: add Linux capability to access physical addresses' " luca.boccassi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230223093715.3926893-53-luca.boccassi@gmail.com \
--to=luca.boccassi@gmail.com \
--cc=akozyrev@nvidia.com \
--cc=matan@nvidia.com \
--cc=stable@dpdk.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).