netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Tommy Christensen <tommy.christensen@tpack.net>
To: Ben Greear <greearb@candelatech.com>, netdev@oss.sgi.com
Subject: [PATCH] Handle shared SKBs in VLAN receive code
Date: Sat, 11 Oct 2003 01:38:53 +0200	[thread overview]
Message-ID: <3F87430D.2040000@tpack.net> (raw)

[-- Attachment #1: Type: text/plain, Size: 482 bytes --]

As noted by Jonas Munsin, the vlan patch that went into 2.6.0-test7 did a
kfree() rather than a kfree_skb(). Hmm.

Instead of just fixing this up, I suggest we do things the right way:

The VLAN code have long been claiming to handle shared SKBs, without actually
doing so. I have now added the call to skb_share_check().

This enables us to simply do a skb_unshare() when mangling the ethernet header.


Patch is against linux-2.6.0-test7 (applies to 2.4.23-pre7 as well).

-Tommy

[-- Attachment #2: vlan_dev.c.patch --]
[-- Type: text/plain, Size: 1552 bytes --]

diff -ru linux-2.6.0-test7/net/8021q/vlan_dev.c linux-2.6/net/8021q/vlan_dev.c
--- linux-2.6.0-test7/net/8021q/vlan_dev.c	Wed Oct  8 21:24:44 2003
+++ linux-2.6/net/8021q/vlan_dev.c	Sat Oct 11 00:45:28 2003
@@ -74,11 +74,7 @@
 static inline struct sk_buff *vlan_check_reorder_header(struct sk_buff *skb)
 {
 	if (VLAN_DEV_INFO(skb->dev)->flags & 1) {
-		if (skb_shared(skb) || skb_cloned(skb)) {
-			struct sk_buff *nskb = skb_copy(skb, GFP_ATOMIC);
-			kfree(skb);
-			skb = nskb;
-		}
+		skb = skb_unshare(skb, GFP_ATOMIC);
 		if (skb) {
 			/* Lifted from Gleb's VLAN code... */
 			memmove(skb->data - ETH_HLEN,
@@ -121,6 +117,7 @@
 	struct net_device_stats *stats;
 	unsigned short vlan_TCI;
 	unsigned short proto;
+	struct net_device *vlan_dev;
 
 	/* vlan_TCI = ntohs(get_unaligned(&vhdr->h_vlan_TCI)); */
 	vlan_TCI = ntohs(vhdr->h_vlan_TCI);
@@ -144,8 +141,8 @@
 	 */
 
 	spin_lock_bh(&vlan_group_lock);
-	skb->dev = __find_vlan_dev(dev, vid);
-	if (!skb->dev) {
+	vlan_dev = __find_vlan_dev(dev, vid);
+	if (!vlan_dev) {
 		spin_unlock_bh(&vlan_group_lock);
 
 #ifdef VLAN_DEBUG
@@ -156,10 +153,18 @@
 		return -1;
 	}
 
-	skb->dev->last_rx = jiffies;
+	vlan_dev->last_rx = jiffies;
+	stats = vlan_dev_get_stats(vlan_dev);
+
+	skb = skb_share_check(skb, GFP_ATOMIC);
+	if (!skb) {
+		spin_unlock_bh(&vlan_group_lock);
+		stats->rx_dropped++;
+		return NET_RX_DROP;
+	}
+	skb->dev = vlan_dev;
 
 	/* Bump the rx counters for the VLAN device. */
-	stats = vlan_dev_get_stats(skb->dev);
 	stats->rx_packets++;
 	stats->rx_bytes += skb->len;
 


             reply	other threads:[~2003-10-10 23:38 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-10-10 23:38 Tommy Christensen [this message]
2003-10-10 23:52 ` [PATCH] Handle shared SKBs in VLAN receive code Ben Greear
2003-10-11  0:09   ` Tommy Christensen
2003-10-11 19:03     ` Ben Greear
2003-10-11 19:03       ` David S. Miller
2003-10-11 19:55         ` Ben Greear
2003-10-11 19:54           ` David S. Miller
2003-10-12 10:00         ` Tommy Christensen
2003-10-13 17:19           ` David S. Miller

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=3F87430D.2040000@tpack.net \
    --to=tommy.christensen@tpack.net \
    --cc=greearb@candelatech.com \
    --cc=netdev@oss.sgi.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).