mail archive of the barebox mailing list
 help / color / mirror / Atom feed
From: Sascha Hauer <s.hauer@pengutronix.de>
To: Barebox List <barebox@lists.infradead.org>
Cc: "Edmund Henniges" <eh@emlix.com>, "Daniel Glöckner" <dg@emlix.com>
Subject: [PATCH 12/16] usbnet: Be more friendly in the receive path
Date: Wed, 11 Mar 2020 15:27:53 +0100	[thread overview]
Message-ID: <20200311142757.25263-13-s.hauer@pengutronix.de> (raw)
In-Reply-To: <20200311142757.25263-1-s.hauer@pengutronix.de>

To recognize if we have a receive packet pending we must set up a USB
bulk transfer. When there's no incoming packet we must wait until the
transfer times out. We do this with every poller call which can
considerably slow down the system. With this patch we do two things
against this:

- lower the timeout for the bulk transfer
- When we haven't received a packet for longer then lower the frequency
  of calling into the USB stack to once every 100ms

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
---
 drivers/net/usb/usbnet.c | 19 ++++++++++++++++++-
 include/usb/usbnet.h     |  3 +++
 2 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c
index f1a49a142c..38aa73047e 100644
--- a/drivers/net/usb/usbnet.c
+++ b/drivers/net/usb/usbnet.c
@@ -123,12 +123,29 @@ static int usbnet_recv(struct eth_device *edev)
 
 	dev_dbg(&edev->dev, "%s\n",__func__);
 
+	/*
+	 * we must let the usb_bulk_msg below timeout before we realize
+	 * that we have no packet received. Since this function runs
+	 * inside a poller we considerably slow down barebox when we
+	 * wait for the timeout too often. To improve this we only poll
+	 * with full speed when we actually have received a packet in the
+	 * last 100ms.
+	 */
+	if (is_timeout(dev->last_pkt_received, 100 * MSECOND) &&
+	    !is_timeout(dev->last_recv_call, 100 * MSECOND)) {
+		return 0;
+	}
+
+	dev->last_recv_call = get_time_ns();
+
 	len = dev->rx_urb_size;
 
-	ret = usb_bulk_msg(dev->udev, dev->in, rx_buf, len, &alen, 100);
+	ret = usb_bulk_msg(dev->udev, dev->in, rx_buf, len, &alen, 2);
 	if (ret)
 		return ret;
 
+	dev->last_pkt_received = get_time_ns();
+
 	if (alen) {
 		if (info->rx_fixup)
 			return info->rx_fixup(dev, rx_buf, alen);
diff --git a/include/usb/usbnet.h b/include/usb/usbnet.h
index 386c2164bd..ee578b81ed 100644
--- a/include/usb/usbnet.h
+++ b/include/usb/usbnet.h
@@ -52,6 +52,9 @@ struct usbnet {
 #		define EVENT_RX_MEMORY	2
 #		define EVENT_STS_SPLIT	3
 #		define EVENT_LINK_RESET	4
+
+	uint64_t last_pkt_received;
+	uint64_t last_recv_call;
 };
 
 #if 0
-- 
2.25.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

  parent reply	other threads:[~2020-03-11 14:28 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-11 14:27 [PATCH v2 00/16] Protect code from pollers Sascha Hauer
2020-03-11 14:27 ` [PATCH 01/16] net: fec_imx: Do not clear MII interrupt during receive Sascha Hauer
2020-03-11 14:27 ` [PATCH 02/16] miitool: Use mdiobus_read() Sascha Hauer
2020-03-11 14:27 ` [PATCH 03/16] net: phy: mdio-mux: Use mdiobus_read/write() Sascha Hauer
2020-03-11 14:27 ` [PATCH 04/16] net: Open ethernet devices explicitly Sascha Hauer
2020-03-11 14:27 ` [PATCH 05/16] Introduce slices Sascha Hauer
2020-03-11 23:51   ` Daniel Glöckner
2020-03-12  8:24     ` Sascha Hauer
2020-03-12  9:18       ` Daniel Glöckner
2020-03-12 10:29         ` Sascha Hauer
2020-03-11 14:27 ` [PATCH 06/16] net: Add a slice to struct eth_device Sascha Hauer
2020-03-11 14:27 ` [PATCH 07/16] net: mdiobus: Add slice Sascha Hauer
2020-03-11 14:27 ` [PATCH 08/16] usb: Add a slice to usb host controllers Sascha Hauer
2020-03-11 14:27 ` [PATCH 09/16] usbnet: Add slice Sascha Hauer
2020-03-11 14:27 ` [PATCH 10/16] net: Call net_poll() in a poller Sascha Hauer
2020-03-11 14:27 ` [PATCH 11/16] net: reply to ping requests Sascha Hauer
2020-03-11 14:27 ` Sascha Hauer [this message]
2020-03-11 14:27 ` [PATCH 13/16] net: phy: Also print link down messages Sascha Hauer
2020-03-11 14:27 ` [PATCH 14/16] net: ifup command: add ethernet device completion Sascha Hauer
2020-03-11 14:27 ` [PATCH 15/16] net: phy: Do not claim the link is up initially Sascha Hauer
2020-03-11 14:27 ` [PATCH 16/16] net: Add ifdown support and command Sascha Hauer

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=20200311142757.25263-13-s.hauer@pengutronix.de \
    --to=s.hauer@pengutronix.de \
    --cc=barebox@lists.infradead.org \
    --cc=dg@emlix.com \
    --cc=eh@emlix.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