linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: Kumar Gala <galak@kernel.crashing.org>
To: Scott Wood <scottwood@freescale.com>
Cc: netdev@vger.kernel.org, jgarzik@pobox.com, linuxppc-dev@ozlabs.org
Subject: Re: [PATCH v2] gianfar: Add I/O barriers when touching buffer descriptor ownership.
Date: Wed, 2 May 2007 21:10:17 -0500 (CDT)	[thread overview]
Message-ID: <Pine.LNX.4.64.0705022109350.1690@localhost.localdomain> (raw)
In-Reply-To: <463902EB.8070401@freescale.com>

On Wed, 2 May 2007, Scott Wood wrote:

> Kumar Gala wrote:
> > Why doesn't marking the bdp pointer volatile resolve the issue in
> > gfar_clean_rx_ring() to ensure load ordering?
>
> Because that only addresses compiler reordering (and does so in a rather
> clumsy way -- not all accesses need to be strongly ordered), not hardware
> reordering.
>
> -Scott
>

So what about some thing like this where we do the read only once?

- k

diff --git a/drivers/net/gianfar.c b/drivers/net/gianfar.c
index a06d8d1..9cd7d1e 100644
--- a/drivers/net/gianfar.c
+++ b/drivers/net/gianfar.c
@@ -1438,31 +1438,35 @@ int gfar_clean_rx_ring(struct net_device *dev, int rx_work_limit)
 {
 	struct rxbd8 *bdp;
 	struct sk_buff *skb;
-	u16 pkt_len;
+	u16 pkt_len, status;
+	u32 bd_info;
 	int howmany = 0;
 	struct gfar_private *priv = netdev_priv(dev);

 	/* Get the first full descriptor */
 	bdp = priv->cur_rx;

-	while (!((bdp->status & RXBD_EMPTY) || (--rx_work_limit < 0))) {
+	bd_info = *bdp;
+
+	status = bd_info >> 16;
+	/* Remove the FCS from the packet length */
+	pkt_len = (bd_info & 0xffff) - 4;
+
+	while (!((status & RXBD_EMPTY) || (--rx_work_limit < 0))) {
 		skb = priv->rx_skbuff[priv->skb_currx];

-		if (!(bdp->status &
+		if (!(status &
 		      (RXBD_LARGE | RXBD_SHORT | RXBD_NONOCTET
 		       | RXBD_CRCERR | RXBD_OVERRUN | RXBD_TRUNCATED))) {
 			/* Increment the number of packets */
 			priv->stats.rx_packets++;
 			howmany++;

-			/* Remove the FCS from the packet length */
-			pkt_len = bdp->length - 4;
-
 			gfar_process_frame(dev, skb, pkt_len);

 			priv->stats.rx_bytes += pkt_len;
 		} else {
-			count_errors(bdp->status, priv);
+			count_errors(status, priv);

 			if (skb)
 				dev_kfree_skb_any(skb);
@@ -1480,7 +1484,7 @@ int gfar_clean_rx_ring(struct net_device *dev, int rx_work_limit)
 		priv->rx_skbuff[priv->skb_currx] = skb;

 		/* Update to the next pointer */
-		if (bdp->status & RXBD_WRAP)
+		if (status & RXBD_WRAP)
 			bdp = priv->rx_bd_base;
 		else
 			bdp++;
@@ -1490,6 +1494,11 @@ int gfar_clean_rx_ring(struct net_device *dev, int rx_work_limit)
 		    (priv->skb_currx +
 		     1) & RX_RING_MOD_MASK(priv->rx_ring_size);

+		bd_info = *bdp;
+
+		status = bd_info >> 16;
+		/* Remove the FCS from the packet length */
+		pkt_len = (bd_info & 0xffff) - 4;
 	}

 	/* Update the current rxbd pointer to be the next one */

  reply	other threads:[~2007-05-03  2:14 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-05-02 19:57 [PATCH v2] gianfar: Add I/O barriers when touching buffer descriptor ownership Scott Wood
2007-05-02 20:09 ` Kumar Gala
2007-05-02 20:12   ` Scott Wood
2007-05-02 20:20     ` Kumar Gala
2007-05-02 20:40       ` Scott Wood
2007-05-02 21:23         ` Kumar Gala
2007-05-02 21:30           ` Scott Wood
2007-05-03  2:10             ` Kumar Gala [this message]
2007-05-03 16:00               ` Scott Wood
2007-05-03 16:38                 ` Segher Boessenkool
2007-05-02 23:42         ` Segher Boessenkool
2007-05-04 22:13         ` Linas Vepstas
2007-05-04 23:24           ` Olof Johansson
2007-05-05  0:41           ` Segher Boessenkool

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=Pine.LNX.4.64.0705022109350.1690@localhost.localdomain \
    --to=galak@kernel.crashing.org \
    --cc=jgarzik@pobox.com \
    --cc=linuxppc-dev@ozlabs.org \
    --cc=netdev@vger.kernel.org \
    --cc=scottwood@freescale.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).