DPDK patches and discussions
 help / color / mirror / Atom feed
From: Olivier Matz <olivier.matz@6wind.com>
To: dev@dpdk.org
Cc: thomas.monjalon@6wind.com, konstantin.ananyev@intel.com,
	wenzhuo.lu@intel.com, helin.zhang@intel.com
Subject: [dpdk-dev] [RFC 4/9] net/ixgbe: optimize Rx queue descriptor count
Date: Thu, 24 Nov 2016 10:54:16 +0100	[thread overview]
Message-ID: <1479981261-19512-5-git-send-email-olivier.matz@6wind.com> (raw)
In-Reply-To: <1479981261-19512-1-git-send-email-olivier.matz@6wind.com>

Use a binary search algorithm to find the first empty DD bit. The
ring-empty and ring-full cases are managed separately as they are more
likely to happen.

PR=52423
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Ivan Boule <ivan.boule@6wind.com>
---
 drivers/net/ixgbe/ixgbe_rxtx.c | 63 ++++++++++++++++++++++++++++++++----------
 1 file changed, 48 insertions(+), 15 deletions(-)

diff --git a/drivers/net/ixgbe/ixgbe_rxtx.c b/drivers/net/ixgbe/ixgbe_rxtx.c
index 1a8ea5f..07509b4 100644
--- a/drivers/net/ixgbe/ixgbe_rxtx.c
+++ b/drivers/net/ixgbe/ixgbe_rxtx.c
@@ -2852,25 +2852,58 @@ ixgbe_dev_rx_queue_setup(struct rte_eth_dev *dev,
 uint32_t
 ixgbe_dev_rx_queue_count(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 {
-#define IXGBE_RXQ_SCAN_INTERVAL 4
-	volatile union ixgbe_adv_rx_desc *rxdp;
+	volatile uint32_t *status;
 	struct ixgbe_rx_queue *rxq;
-	uint32_t desc = 0;
+	uint32_t offset, interval, nb_hold, resolution;
+	int32_t idx;
 
 	rxq = dev->data->rx_queues[rx_queue_id];
-	rxdp = &(rxq->rx_ring[rxq->rx_tail]);
-
-	while ((desc < rxq->nb_rx_desc) &&
-		(rxdp->wb.upper.status_error &
-			rte_cpu_to_le_32(IXGBE_RXDADV_STAT_DD))) {
-		desc += IXGBE_RXQ_SCAN_INTERVAL;
-		rxdp += IXGBE_RXQ_SCAN_INTERVAL;
-		if (rxq->rx_tail + desc >= rxq->nb_rx_desc)
-			rxdp = &(rxq->rx_ring[rxq->rx_tail +
-				desc - rxq->nb_rx_desc]);
-	}
 
-	return desc;
+	/* check if ring empty */
+	idx = rxq->rx_tail;
+	status = &rxq->rx_ring[idx].wb.upper.status_error;
+	if (!(*status & rte_cpu_to_le_32(IXGBE_RXDADV_STAT_DD)))
+		return 0;
+
+	/* decrease the precision if ring is large */
+	if (rxq->nb_rx_desc <= 256)
+		resolution = 4;
+	else
+		resolution = 16;
+
+	/* check if ring full */
+#ifdef RTE_IXGBE_INC_VECTOR
+	if (rxq->rx_using_sse)
+		nb_hold = rxq->rxrearm_nb;
+	else
+#endif
+		nb_hold = rxq->nb_rx_hold;
+
+	idx = rxq->rx_tail - nb_hold - resolution;
+	if (idx < 0)
+		idx += rxq->nb_rx_desc;
+	status = &rxq->rx_ring[idx].wb.upper.status_error;
+	if (*status & rte_cpu_to_le_32(IXGBE_RXDADV_STAT_DD))
+		return rxq->nb_rx_desc;
+
+	/* use a binary search */
+	interval = (rxq->nb_rx_desc - nb_hold) >> 1;
+	offset = interval;
+
+	do {
+		idx = rxq->rx_tail + offset;
+		if (idx >= rxq->nb_rx_desc)
+			idx -= rxq->nb_rx_desc;
+
+		interval >>= 1;
+		status = &rxq->rx_ring[idx].wb.upper.status_error;
+		if (*status & rte_cpu_to_le_32(IXGBE_RXDADV_STAT_DD))
+			offset += interval;
+		else
+			offset -= interval;
+	} while (interval >= resolution);
+
+	return offset;
 }
 
 int
-- 
2.8.1

  parent reply	other threads:[~2016-11-24  9:59 UTC|newest]

Thread overview: 72+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-11-24  9:54 [dpdk-dev] [RFC 0/9] get Rx and Tx used descriptors Olivier Matz
2016-11-24  9:54 ` [dpdk-dev] [RFC 1/9] ethdev: clarify api comments of rx queue count Olivier Matz
2016-11-24 10:52   ` Ferruh Yigit
2016-11-24 11:13     ` Olivier Matz
2016-11-24  9:54 ` [dpdk-dev] [RFC 2/9] ethdev: move queue id check in generic layer Olivier Matz
2016-11-24 10:59   ` Ferruh Yigit
2016-11-24 13:05     ` Olivier Matz
2016-11-24  9:54 ` [dpdk-dev] [RFC 3/9] ethdev: add handler for Tx queue descriptor count Olivier Matz
2016-11-24  9:54 ` Olivier Matz [this message]
2016-11-24  9:54 ` [dpdk-dev] [RFC 5/9] net/ixgbe: " Olivier Matz
2016-11-24  9:54 ` [dpdk-dev] [RFC 6/9] net/igb: optimize rx " Olivier Matz
2016-11-24  9:54 ` [dpdk-dev] [RFC 7/9] net/igb: add handler for tx " Olivier Matz
2016-11-24  9:54 ` [dpdk-dev] [RFC 8/9] net/e1000: optimize rx " Olivier Matz
2016-11-24  9:54 ` [dpdk-dev] [RFC 9/9] net/e1000: add handler for tx " Olivier Matz
2017-01-13 16:44 ` [dpdk-dev] [RFC 0/9] get Rx and Tx used descriptors Olivier Matz
2017-01-13 17:32   ` Richardson, Bruce
2017-01-17  8:24     ` Olivier Matz
2017-01-17 13:56       ` Bruce Richardson
2017-03-01 17:19 ` [dpdk-dev] [PATCH 0/6] get status of Rx and Tx descriptors Olivier Matz
2017-03-01 17:19   ` [dpdk-dev] [PATCH 1/6] ethdev: add descriptor status API Olivier Matz
2017-03-01 18:22     ` Andrew Rybchenko
2017-03-02 13:57       ` Olivier Matz
2017-03-02 14:19         ` Andrew Rybchenko
2017-03-02 14:54           ` Olivier Matz
2017-03-02 15:05             ` Andrew Rybchenko
2017-03-02 15:14               ` Olivier Matz
2017-03-01 17:19   ` [dpdk-dev] [PATCH 2/6] net/ixgbe: implement " Olivier Matz
2017-03-01 17:19   ` [dpdk-dev] [PATCH 3/6] net/e1000: implement descriptor status API (igb) Olivier Matz
2017-03-02  1:28     ` Lu, Wenzhuo
2017-03-02 13:58       ` Olivier Matz
2017-03-01 17:19   ` [dpdk-dev] [PATCH 4/6] net/e1000: implement descriptor status API (em) Olivier Matz
2017-03-02  1:22     ` Lu, Wenzhuo
2017-03-02 14:46       ` Olivier Matz
2017-03-03  1:15         ` Lu, Wenzhuo
2017-03-01 17:19   ` [dpdk-dev] [PATCH 5/6] net/mlx5: implement descriptor status API Olivier Matz
2017-03-02  7:56     ` Nélio Laranjeiro
2017-03-01 17:19   ` [dpdk-dev] [PATCH 6/6] net/i40e: " Olivier Matz
2017-03-01 18:02   ` [dpdk-dev] [PATCH 0/6] get status of Rx and Tx descriptors Andrew Rybchenko
2017-03-02 13:40     ` Olivier Matz
2017-03-06 10:41       ` Thomas Monjalon
2017-03-01 18:07   ` Stephen Hemminger
2017-03-02 13:43     ` Olivier Matz
2017-03-06 10:41       ` Thomas Monjalon
2017-03-02 15:32   ` Bruce Richardson
2017-03-02 16:14     ` Olivier Matz
2017-03-03 16:18       ` Venkatesan, Venky
2017-03-03 16:45         ` Olivier Matz
2017-03-03 18:46           ` Venkatesan, Venky
2017-03-04 20:45             ` Olivier Matz
2017-03-06 11:02               ` Thomas Monjalon
2017-03-07 15:59   ` [dpdk-dev] [PATCH v2 " Olivier Matz
2017-03-07 15:59     ` [dpdk-dev] [PATCH v2 1/6] ethdev: add descriptor status API Olivier Matz
2017-03-09 11:49       ` Andrew Rybchenko
2017-03-21  8:32       ` Yang, Qiming
2017-03-24 12:49         ` Olivier Matz
2017-03-27  1:28           ` Yang, Qiming
2017-03-07 15:59     ` [dpdk-dev] [PATCH v2 2/6] net/ixgbe: implement " Olivier Matz
2017-03-07 15:59     ` [dpdk-dev] [PATCH v2 3/6] net/e1000: implement descriptor status API (igb) Olivier Matz
2017-03-08  1:17       ` Lu, Wenzhuo
2017-03-07 15:59     ` [dpdk-dev] [PATCH v2 4/6] net/e1000: implement descriptor status API (em) Olivier Matz
2017-03-07 15:59     ` [dpdk-dev] [PATCH v2 5/6] net/mlx5: implement descriptor status API Olivier Matz
2017-03-07 15:59     ` [dpdk-dev] [PATCH v2 6/6] net/i40e: " Olivier Matz
2017-03-08  1:17       ` Lu, Wenzhuo
2017-03-29  8:36     ` [dpdk-dev] [PATCH v3 0/6] get status of Rx and Tx descriptors Olivier Matz
2017-03-29  8:36       ` [dpdk-dev] [PATCH v3 1/6] ethdev: add descriptor status API Olivier Matz
2017-03-29  8:36       ` [dpdk-dev] [PATCH v3 2/6] net/ixgbe: implement " Olivier Matz
2017-03-29  8:36       ` [dpdk-dev] [PATCH v3 3/6] net/e1000: implement descriptor status API (igb) Olivier Matz
2017-03-29  8:36       ` [dpdk-dev] [PATCH v3 4/6] net/e1000: implement descriptor status API (em) Olivier Matz
2017-03-29  8:36       ` [dpdk-dev] [PATCH v3 5/6] net/mlx5: implement descriptor status API Olivier Matz
2017-03-29  8:36       ` [dpdk-dev] [PATCH v3 6/6] net/i40e: " Olivier Matz
2017-03-30 13:30       ` [dpdk-dev] [PATCH v3 0/6] get status of Rx and Tx descriptors Thomas Monjalon
2017-04-19 15:50         ` Ferruh Yigit

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=1479981261-19512-5-git-send-email-olivier.matz@6wind.com \
    --to=olivier.matz@6wind.com \
    --cc=dev@dpdk.org \
    --cc=helin.zhang@intel.com \
    --cc=konstantin.ananyev@intel.com \
    --cc=thomas.monjalon@6wind.com \
    --cc=wenzhuo.lu@intel.com \
    /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).