[dpdk-stable] patch 'net/ixgbe: fix vector Rx' has been queued to LTS release 18.11.11

Kevin Traynor ktraynor at redhat.com
Wed Nov 18 17:35:39 CET 2020
Previous message: [dpdk-stable] patch 'net/i40e: fix QinQ flow pattern to allow non full mask' has been queued to LTS release 18.11.11
Next message: [dpdk-stable] patch 'net/i40e: fix vector Rx' has been queued to LTS release 18.11.11
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi,

FYI, your patch has been queued to LTS release 18.11.11

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 11/24/20. 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/kevintraynor/dpdk-stable-queue

This queued commit can be viewed at:
https://github.com/kevintraynor/dpdk-stable-queue/commit/b7a03919c8e63b053c7453888e6a4fc325bf2221

Thanks.

Kevin.

---
>From b7a03919c8e63b053c7453888e6a4fc325bf2221 Mon Sep 17 00:00:00 2001
From: Jeff Guo <jia.guo at intel.com>
Date: Fri, 16 Oct 2020 17:44:27 +0800
Subject: [PATCH] net/ixgbe: fix vector Rx
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

[ upstream commit 5a3cca342417f7d4be4ac4b6225ca509617777c4 ]

The limitation of burst size in vector rx was removed, since it should
retrieve as much received packets as possible. And also the scattered
receive path should use a wrapper function to achieve the goal of
burst maximizing.

Bugzilla ID: 516
Fixes: b20971b6cca0 ("net/ixgbe: implement vector driver for ARM")
Fixes: 0e51f9dc4860 ("net/ixgbe: rename x86 vector driver file")

Signed-off-by: Jeff Guo <jia.guo at intel.com>
Tested-by: Feifei Wang <feifei.wang2 at arm.com>
Acked-by: Morten Brørup <mb at smartsharesystems.com>
Acked-by: Qi Zhang <qi.z.zhang at intel.com>
---
 drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c | 61 +++++++++++++++----------
 drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c  | 47 +++++++++++++------
 2 files changed, 70 insertions(+), 38 deletions(-)

diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
index d967236c44..da6ee2368d 100644
--- a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
+++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
@@ -131,15 +131,4 @@ desc_to_olflags_v(uint8x16x2_t sterr_tmp1, uint8x16x2_t sterr_tmp2,
 }
 
-/*
- * vPMD raw receive routine, only accept(nb_pkts >= RTE_IXGBE_DESCS_PER_LOOP)
- *
- * Notice:
- * - nb_pkts < RTE_IXGBE_DESCS_PER_LOOP, just return no packet
- * - nb_pkts > RTE_IXGBE_MAX_RX_BURST, only scan RTE_IXGBE_MAX_RX_BURST
- *   numbers of DD bit
- * - floor align nb_pkts to a RTE_IXGBE_DESC_PER_LOOP power-of-two
- * - don't support ol_flags for rss and csum err
- */
-
 #define IXGBE_VPMD_DESC_DD_MASK		0x01010101
 #define IXGBE_VPMD_DESC_EOP_MASK	0x02020202
@@ -207,4 +196,11 @@ desc_to_ptype_v(uint64x2_t descs[4], uint16_t pkt_type_mask,
 }
 
+/**
+ * vPMD raw receive routine, only accept(nb_pkts >= RTE_IXGBE_DESCS_PER_LOOP)
+ *
+ * Notice:
+ * - nb_pkts < RTE_IXGBE_DESCS_PER_LOOP, just return no packet
+ * - floor align nb_pkts to a RTE_IXGBE_DESC_PER_LOOP power-of-two
+ */
 static inline uint16_t
 _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
@@ -227,7 +223,4 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 				 rxq->crc_len, 0, 0, 0};
 
-	/* nb_pkts shall be less equal than RTE_IXGBE_MAX_RX_BURST */
-	nb_pkts = RTE_MIN(nb_pkts, RTE_IXGBE_MAX_RX_BURST);
-
 	/* nb_pkts has to be floor-aligned to RTE_IXGBE_DESCS_PER_LOOP */
 	nb_pkts = RTE_ALIGN_FLOOR(nb_pkts, RTE_IXGBE_DESCS_PER_LOOP);
@@ -383,11 +376,9 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 }
 
-/*
+/**
  * vPMD receive routine, only accept(nb_pkts >= RTE_IXGBE_DESCS_PER_LOOP)
  *
  * Notice:
  * - nb_pkts < RTE_IXGBE_DESCS_PER_LOOP, just return no packet
- * - nb_pkts > RTE_IXGBE_MAX_RX_BURST, only scan RTE_IXGBE_MAX_RX_BURST
- *   numbers of DD bit
  * - floor align nb_pkts to a RTE_IXGBE_DESC_PER_LOOP power-of-two
  * - don't support ol_flags for rss and csum err
@@ -400,5 +391,5 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
 }
 
-/*
+/**
  * vPMD receive routine that reassembles scattered packets
  *
@@ -406,11 +397,9 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
  * - don't support ol_flags for rss and csum err
  * - nb_pkts < RTE_IXGBE_DESCS_PER_LOOP, just return no packet
- * - nb_pkts > RTE_IXGBE_MAX_RX_BURST, only scan RTE_IXGBE_MAX_RX_BURST
- *   numbers of DD bit
  * - floor align nb_pkts to a RTE_IXGBE_DESC_PER_LOOP power-of-two
  */
-uint16_t
-ixgbe_recv_scattered_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
-		uint16_t nb_pkts)
+static uint16_t
+ixgbe_recv_scattered_burst_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
+			       uint16_t nb_pkts)
 {
 	struct ixgbe_rx_queue *rxq = rx_queue;
@@ -444,4 +433,30 @@ ixgbe_recv_scattered_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
 }
 
+/**
+ * vPMD receive routine that reassembles scattered packets.
+ */
+uint16_t
+ixgbe_recv_scattered_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
+			      uint16_t nb_pkts)
+{
+	uint16_t retval = 0;
+
+	while (nb_pkts > RTE_IXGBE_MAX_RX_BURST) {
+		uint16_t burst;
+
+		burst = ixgbe_recv_scattered_burst_vec(rx_queue,
+						       rx_pkts + retval,
+						       RTE_IXGBE_MAX_RX_BURST);
+		retval += burst;
+		nb_pkts -= burst;
+		if (burst < RTE_IXGBE_MAX_RX_BURST)
+			return retval;
+	}
+
+	return retval + ixgbe_recv_scattered_burst_vec(rx_queue,
+						       rx_pkts + retval,
+						       nb_pkts);
+}
+
 static inline void
 vtx1(volatile union ixgbe_adv_tx_desc *txdp,
diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c b/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c
index 599ba30e51..03eacfd86b 100644
--- a/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c
+++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c
@@ -303,11 +303,9 @@ desc_to_ptype_v(__m128i descs[4], uint16_t pkt_type_mask,
 }
 
-/*
+/**
  * vPMD raw receive routine, only accept(nb_pkts >= RTE_IXGBE_DESCS_PER_LOOP)
  *
  * Notice:
  * - nb_pkts < RTE_IXGBE_DESCS_PER_LOOP, just return no packet
- * - nb_pkts > RTE_IXGBE_MAX_RX_BURST, only scan RTE_IXGBE_MAX_RX_BURST
- *   numbers of DD bit
  * - floor align nb_pkts to a RTE_IXGBE_DESC_PER_LOOP power-of-two
  */
@@ -345,7 +343,4 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 	uint8_t vlan_flags;
 
-	/* nb_pkts shall be less equal than RTE_IXGBE_MAX_RX_BURST */
-	nb_pkts = RTE_MIN(nb_pkts, RTE_IXGBE_MAX_RX_BURST);
-
 	/* nb_pkts has to be floor-aligned to RTE_IXGBE_DESCS_PER_LOOP */
 	nb_pkts = RTE_ALIGN_FLOOR(nb_pkts, RTE_IXGBE_DESCS_PER_LOOP);
@@ -557,11 +552,9 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 }
 
-/*
+/**
  * vPMD receive routine, only accept(nb_pkts >= RTE_IXGBE_DESCS_PER_LOOP)
  *
  * Notice:
  * - nb_pkts < RTE_IXGBE_DESCS_PER_LOOP, just return no packet
- * - nb_pkts > RTE_IXGBE_MAX_RX_BURST, only scan RTE_IXGBE_MAX_RX_BURST
- *   numbers of DD bit
  * - floor align nb_pkts to a RTE_IXGBE_DESC_PER_LOOP power-of-two
  */
@@ -573,16 +566,14 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
 }
 
-/*
+/**
  * vPMD receive routine that reassembles scattered packets
  *
  * Notice:
  * - nb_pkts < RTE_IXGBE_DESCS_PER_LOOP, just return no packet
- * - nb_pkts > RTE_IXGBE_MAX_RX_BURST, only scan RTE_IXGBE_MAX_RX_BURST
- *   numbers of DD bit
  * - floor align nb_pkts to a RTE_IXGBE_DESC_PER_LOOP power-of-two
  */
-uint16_t
-ixgbe_recv_scattered_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
-		uint16_t nb_pkts)
+static uint16_t
+ixgbe_recv_scattered_burst_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
+			       uint16_t nb_pkts)
 {
 	struct ixgbe_rx_queue *rxq = rx_queue;
@@ -616,4 +607,30 @@ ixgbe_recv_scattered_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
 }
 
+/**
+ * vPMD receive routine that reassembles scattered packets.
+ */
+uint16_t
+ixgbe_recv_scattered_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
+			      uint16_t nb_pkts)
+{
+	uint16_t retval = 0;
+
+	while (nb_pkts > RTE_IXGBE_MAX_RX_BURST) {
+		uint16_t burst;
+
+		burst = ixgbe_recv_scattered_burst_vec(rx_queue,
+						       rx_pkts + retval,
+						       RTE_IXGBE_MAX_RX_BURST);
+		retval += burst;
+		nb_pkts -= burst;
+		if (burst < RTE_IXGBE_MAX_RX_BURST)
+			return retval;
+	}
+
+	return retval + ixgbe_recv_scattered_burst_vec(rx_queue,
+						       rx_pkts + retval,
+						       nb_pkts);
+}
+
 static inline void
 vtx1(volatile union ixgbe_adv_tx_desc *txdp,
-- 
2.26.2

---
  Diff of the applied patch vs upstream commit (please double-check if non-empty:
---
--- -	2020-11-18 16:33:38.942559122 +0000
+++ 0053-net-ixgbe-fix-vector-Rx.patch	2020-11-18 16:33:37.978215093 +0000
@@ -1 +1 @@
-From 5a3cca342417f7d4be4ac4b6225ca509617777c4 Mon Sep 17 00:00:00 2001
+From b7a03919c8e63b053c7453888e6a4fc325bf2221 Mon Sep 17 00:00:00 2001
@@ -8,0 +9,2 @@
+[ upstream commit 5a3cca342417f7d4be4ac4b6225ca509617777c4 ]
+
@@ -17 +18,0 @@
-Cc: stable at dpdk.org
@@ -29 +30 @@
-index aa27ee1777..4c81ae9dcf 100644
+index d967236c44..da6ee2368d 100644
@@ -45,0 +47 @@
+ #define IXGBE_VPMD_DESC_DD_MASK		0x01010101
@@ -47 +48,0 @@
- #define IXGBE_UINT8_BIT			(CHAR_BIT * sizeof(uint8_t))
@@ -135 +136 @@
-index 586a261808..90c076825a 100644
+index 599ba30e51..03eacfd86b 100644
Previous message: [dpdk-stable] patch 'net/i40e: fix QinQ flow pattern to allow non full mask' has been queued to LTS release 18.11.11
Next message: [dpdk-stable] patch 'net/i40e: fix vector Rx' has been queued to LTS release 18.11.11
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the stable mailing list