public inbox for linux-scsi@vger.kernel.org
 help / color / mirror / Atom feed
From: Robert Love <robert.w.love@intel.com>
To: linux-scsi@vger.kernel.org
Cc: Neil Horman <nhorman@tuxdriver.com>, Vasu Dev <vasu.dev@intel.com>
Subject: [PATCH 09/10] fcoe: reduce contention for fcoe_rx_list lock [v2]
Date: Fri, 09 Mar 2012 14:50:24 -0800	[thread overview]
Message-ID: <20120309225024.6515.5554.stgit@localhost6.localdomain6> (raw)
In-Reply-To: <20120309224937.6515.83801.stgit@localhost6.localdomain6>

From: Neil Horman <nhorman@tuxdriver.com>

There is potentially lots of contention for the rx_list_lock.  On a cpu that is
receiving lots of fcoe traffic, the softirq context has to add and release the
lock for every frame it receives, as does the receiving per-cpu thread.  We can
reduce this contention somewhat by altering the per-cpu threads loop such that
when traffic is detected on the fcoe_rx_list, we splice it to a temporary list.
In this way, we can process multiple skbs while only having to acquire and
release the fcoe_rx_list lock once.

[ Braces around single statement while loop removed by Robert Love
  to satisfy checkpath.pl. ]

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Vasu Dev <vasu.dev@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
---
 drivers/scsi/fcoe/fcoe.c |   22 ++++++++++++++--------
 1 files changed, 14 insertions(+), 8 deletions(-)

diff --git a/drivers/scsi/fcoe/fcoe.c b/drivers/scsi/fcoe/fcoe.c
index d86ca37..58c88b0 100644
--- a/drivers/scsi/fcoe/fcoe.c
+++ b/drivers/scsi/fcoe/fcoe.c
@@ -1471,7 +1471,7 @@ static int fcoe_rcv(struct sk_buff *skb, struct net_device *netdev,
 	 * in softirq context.
 	 */
 	__skb_queue_tail(&fps->fcoe_rx_list, skb);
-	if (fps->fcoe_rx_list.qlen == 1)
+	if (fps->thread->state == TASK_INTERRUPTIBLE)
 		wake_up_process(fps->thread);
 	spin_unlock(&fps->fcoe_rx_list.lock);
 
@@ -1790,23 +1790,29 @@ static int fcoe_percpu_receive_thread(void *arg)
 {
 	struct fcoe_percpu_s *p = arg;
 	struct sk_buff *skb;
+	struct sk_buff_head tmp;
+
+	skb_queue_head_init(&tmp);
 
 	set_user_nice(current, -20);
 
 	while (!kthread_should_stop()) {
 
 		spin_lock_bh(&p->fcoe_rx_list.lock);
-		while ((skb = __skb_dequeue(&p->fcoe_rx_list)) == NULL) {
+		skb_queue_splice_init(&p->fcoe_rx_list, &tmp);
+		spin_unlock_bh(&p->fcoe_rx_list.lock);
+
+		while ((skb = __skb_dequeue(&tmp)) != NULL)
+			fcoe_recv_frame(skb);
+
+		spin_lock_bh(&p->fcoe_rx_list.lock);
+		if (!skb_queue_len(&p->fcoe_rx_list)) {
 			set_current_state(TASK_INTERRUPTIBLE);
 			spin_unlock_bh(&p->fcoe_rx_list.lock);
 			schedule();
 			set_current_state(TASK_RUNNING);
-			if (kthread_should_stop())
-				return 0;
-			spin_lock_bh(&p->fcoe_rx_list.lock);
-		}
-		spin_unlock_bh(&p->fcoe_rx_list.lock);
-		fcoe_recv_frame(skb);
+		} else
+			spin_unlock_bh(&p->fcoe_rx_list.lock);
 	}
 	return 0;
 }


  parent reply	other threads:[~2012-03-09 22:50 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-03-09 22:49 [PATCH 00/10] libfc, libfcoe and fcoe updates for scsi-misc Robert Love
2012-03-09 22:49 ` [PATCH 01/10] fcoe: avoid getting into dev_queue_xmit with bottom halves disabled Robert Love
2012-03-09 23:50   ` Bhanu Prakash Gollapudi
2012-03-10  0:12     ` Love, Robert W
     [not found]       ` <4F5A9C53.5040200-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2012-03-12 22:51         ` Love, Robert W
     [not found]     ` <4F5A9754.6080009-dY08KVG/lbpWk0Htik3J/w@public.gmane.org>
2012-03-11 17:43       ` Neil Horman
2012-03-12  3:25         ` Bhanu Prakash Gollapudi
2012-03-12 10:42           ` Neil Horman
2012-03-09 22:49 ` [PATCH 02/10] fcoe: Ensure fcoe_recv_frame is always called in process context Robert Love
2012-03-09 22:49 ` [PATCH 03/10] libfcoe: Do not sends FDISCs before FLOGI during CVL Robert Love
2012-03-09 22:49 ` [PATCH 04/10] libfc: update fc_host mfs along with updating lport->mfs Robert Love
2012-03-09 22:50 ` [PATCH 05/10] libfcoe: Support extra MAC descriptor to be used as FCoE MAC Robert Love
2012-03-09 22:50 ` [PATCH 06/10] foce: remove bh disable from fcoe sw transport rcv function Robert Love
2012-03-09 22:50 ` [PATCH 07/10] bnx2fc: Remove bh disable in softirq context Robert Love
2012-03-09 22:50 ` [PATCH 08/10] fcoe: remove frame dropping code from fcoe_percpu_clean Robert Love
2012-03-09 22:50 ` Robert Love [this message]
2012-03-09 22:50 ` [PATCH 10/10] libfc: fcoe_transport_create fails in single-CPU environment Robert Love

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=20120309225024.6515.5554.stgit@localhost6.localdomain6 \
    --to=robert.w.love@intel.com \
    --cc=linux-scsi@vger.kernel.org \
    --cc=nhorman@tuxdriver.com \
    --cc=vasu.dev@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