linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: sunil.kovvuri@gmail.com (sunil.kovvuri at gmail.com)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH 04/15] net: thunderx: Set queue count based on number of CPUs
Date: Mon, 11 Jul 2016 15:31:46 +0530	[thread overview]
Message-ID: <1468231317-24023-5-git-send-email-sunil.kovvuri@gmail.com> (raw)
In-Reply-To: <1468231317-24023-1-git-send-email-sunil.kovvuri@gmail.com>

From: Sunil Goutham <sgoutham@cavium.com>

81xx has only 4 CPUs, so it doesn't make sense to initialize
entire Qset i.e 8 queues by default. Made changes to queue
initialization to init queues equal to number of CPUs or
8 queues whichever is lesser. Also this will be applicable to
VMs with VNIC VF attached and having less VCPUs

Signed-off-by: Sunil Goutham <sgoutham@cavium.com>
---
 drivers/net/ethernet/cavium/thunder/nic_main.c     | 6 ++++++
 drivers/net/ethernet/cavium/thunder/nicvf_main.c   | 7 +++----
 drivers/net/ethernet/cavium/thunder/nicvf_queues.c | 8 ++++----
 drivers/net/ethernet/cavium/thunder/nicvf_queues.h | 5 +----
 4 files changed, 14 insertions(+), 12 deletions(-)

diff --git a/drivers/net/ethernet/cavium/thunder/nic_main.c b/drivers/net/ethernet/cavium/thunder/nic_main.c
index 4974923..0d81117 100644
--- a/drivers/net/ethernet/cavium/thunder/nic_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nic_main.c
@@ -1009,6 +1009,12 @@ static int nic_num_sqs_en(struct nicpf *nic, int vf_en)
 	int pos, sqs_per_vf = MAX_SQS_PER_VF_SINGLE_NODE;
 	u16 total_vf;
 
+	/* Secondary Qsets are needed only if CPU count is
+	 * morethan MAX_QUEUES_PER_QSET.
+	 */
+	if (num_online_cpus() <= MAX_QUEUES_PER_QSET)
+		return 0;
+
 	/* Check if its a multi-node environment */
 	if (nr_node_ids > 1)
 		sqs_per_vf = MAX_SQS_PER_VF;
diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_main.c b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
index 0c10635..af04d9f 100644
--- a/drivers/net/ethernet/cavium/thunder/nicvf_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
@@ -1537,14 +1537,13 @@ static int nicvf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		goto err_release_regions;
 	}
 
-	qcount = MAX_CMP_QUEUES_PER_QS;
+	qcount = min_t(int, MAX_CMP_QUEUES_PER_QS, num_online_cpus());
 
 	/* Restrict multiqset support only for host bound VFs */
 	if (pdev->is_virtfn) {
 		/* Set max number of queues per VF */
-		qcount = roundup(num_online_cpus(), MAX_CMP_QUEUES_PER_QS);
-		qcount = min(qcount,
-			     (MAX_SQS_PER_VF + 1) * MAX_CMP_QUEUES_PER_QS);
+		qcount = min_t(int, num_online_cpus(),
+			       (MAX_SQS_PER_VF + 1) * MAX_CMP_QUEUES_PER_QS);
 	}
 
 	netdev = alloc_etherdev_mqs(sizeof(struct nicvf), qcount, qcount);
diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_queues.c b/drivers/net/ethernet/cavium/thunder/nicvf_queues.c
index 0ff8e60..e521a94 100644
--- a/drivers/net/ethernet/cavium/thunder/nicvf_queues.c
+++ b/drivers/net/ethernet/cavium/thunder/nicvf_queues.c
@@ -762,10 +762,10 @@ int nicvf_set_qset_resources(struct nicvf *nic)
 	nic->qs = qs;
 
 	/* Set count of each queue */
-	qs->rbdr_cnt = RBDR_CNT;
-	qs->rq_cnt = RCV_QUEUE_CNT;
-	qs->sq_cnt = SND_QUEUE_CNT;
-	qs->cq_cnt = CMP_QUEUE_CNT;
+	qs->rbdr_cnt = DEFAULT_RBDR_CNT;
+	qs->rq_cnt = min_t(u8, MAX_RCV_QUEUES_PER_QS, num_online_cpus());
+	qs->sq_cnt = min_t(u8, MAX_SND_QUEUES_PER_QS, num_online_cpus());
+	qs->cq_cnt = max_t(u8, qs->rq_cnt, qs->sq_cnt);
 
 	/* Set queue lengths */
 	qs->rbdr_len = RCV_BUF_COUNT;
diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_queues.h b/drivers/net/ethernet/cavium/thunder/nicvf_queues.h
index 6673e11..869f338 100644
--- a/drivers/net/ethernet/cavium/thunder/nicvf_queues.h
+++ b/drivers/net/ethernet/cavium/thunder/nicvf_queues.h
@@ -57,10 +57,7 @@
 #define CMP_QUEUE_SIZE6		6ULL /* 64K entries */
 
 /* Default queue count per QS, its lengths and threshold values */
-#define RBDR_CNT		1
-#define RCV_QUEUE_CNT		8
-#define SND_QUEUE_CNT		8
-#define CMP_QUEUE_CNT		8 /* Max of RCV and SND qcount */
+#define DEFAULT_RBDR_CNT	1
 
 #define SND_QSIZE		SND_QUEUE_SIZE2
 #define SND_QUEUE_LEN		(1ULL << (SND_QSIZE + 10))
-- 
2.7.4

  parent reply	other threads:[~2016-07-11 10:01 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-07-11 10:01 [PATCH 00/15] net: thunderx: Add support for 81xx and 83xx sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 01/15] net: thunderx: Moved HW capability info from macros to structure sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 02/15] net: thunderx: Add VNIC's PCI devid on future chips sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 03/15] net: thunderx: Add support for 81xx and 83xx chips sunil.kovvuri at gmail.com
2016-07-11 10:01 ` sunil.kovvuri at gmail.com [this message]
2016-07-11 10:01 ` [PATCH 05/15] net: thunderx: Enable CQE_RX desc's extension fields sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 06/15] net: thunderx: Enable mailbox interrupts on 81xx/83xx sunil.kovvuri at gmail.com
2016-07-11 11:41   ` kbuild test robot
2016-07-11 10:01 ` [PATCH 07/15] net: thunderx: Support for different LMAC types within BGX sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 08/15] net: thunderx: Add 81xx support to BGX driver sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 09/15] net: thunderx: Add QSGMII interface type support sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 10/15] net: thunderx: Add RGMII " sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 11/15] net: thunderx: Use netdev's name for naming VF's interrupts sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 12/15] net: thunderx: Use skb_add_rx_frag() for split buffer Rx pkts sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 13/15] net: thunderx: Improvement for MBX interface debug messages sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 14/15] net: thunderx: Reset RXQ HW stats when interface is brought down sunil.kovvuri at gmail.com
2016-07-11 10:01 ` [PATCH 15/15] net: thunderx: Don't set mac address for secondary Qset VFs sunil.kovvuri at gmail.com

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=1468231317-24023-5-git-send-email-sunil.kovvuri@gmail.com \
    --to=sunil.kovvuri@gmail.com \
    --cc=linux-arm-kernel@lists.infradead.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).