Patchwork [net-next] core: should call pskb_expand_head if skb header is cloned in skb_gso_segment in rx path

login
register
mail settings
Submitter roy.qing.li@gmail.com
Date March 27, 2013, 9:34 a.m.
Message ID <1364376845-5000-1-git-send-email-roy.qing.li@gmail.com>
Download mbox | patch
Permalink /patch/231635/
State Deferred
Delegated to: David Miller
Headers show

Comments

roy.qing.li@gmail.com - March 27, 2013, 9:34 a.m.
From: Li RongQing <roy.qing.li@gmail.com>

12b0004d1d1 (adjust skb_gso_segment() for calling in rx path) tries to kill warnings
by checking if ip_summed is CHECK_NONE or not in rx path, since if skb_gso_segment()
is called on rx path, and ->ip_summed has different meaning.

but this maybe break skb if skb header is cloned, and not expand the header, since when
step into skb_mac_gso_segment(), which will still check ip_summed with CHECKSUM_PARTIAL,
then do gso_send_check(). and after __skb_gso_segment() in queue_gso_packets() of
openvswitch, queue_userspace_packet() still checks ip_summed with CHECKSUM_PARTIAL,
and do checksum.

so I think it is enough to ignore the warning in rx path.

Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
---
 net/core/dev.c |   16 +++-------------
 1 file changed, 3 insertions(+), 13 deletions(-)
David Miller - March 29, 2013, 6:59 p.m.
From: roy.qing.li@gmail.com
Date: Wed, 27 Mar 2013 17:34:05 +0800

> From: Li RongQing <roy.qing.li@gmail.com>
> 
> 12b0004d1d1 (adjust skb_gso_segment() for calling in rx path) tries to kill warnings
> by checking if ip_summed is CHECK_NONE or not in rx path, since if skb_gso_segment()
> is called on rx path, and ->ip_summed has different meaning.
> 
> but this maybe break skb if skb header is cloned, and not expand the header, since when
> step into skb_mac_gso_segment(), which will still check ip_summed with CHECKSUM_PARTIAL,
> then do gso_send_check(). and after __skb_gso_segment() in queue_gso_packets() of
> openvswitch, queue_userspace_packet() still checks ip_summed with CHECKSUM_PARTIAL,
> and do checksum.
> 
> so I think it is enough to ignore the warning in rx path.
> 
> Signed-off-by: Li RongQing <roy.qing.li@gmail.com>

I'm not going back-and-forth on this issue yet another time.

I want to see discussion amongst people involved in this area
before applying a patch like this, and I'm therefore not applying
this patch for now (you'll have to submit it again after the
discussions take place).
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Amerigo Wang - April 1, 2013, 2:18 a.m.
On Mon, 2013-04-01 at 10:07 +0800, RongQing Li wrote:
> Hi Cong:
> 
> Could you give some comments for this patch?

I don't see the original path, sorry.


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/net/core/dev.c b/net/core/dev.c
index de930b7..bf0e586 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -2269,17 +2269,6 @@  struct sk_buff *skb_mac_gso_segment(struct sk_buff *skb,
 }
 EXPORT_SYMBOL(skb_mac_gso_segment);
 
-
-/* openvswitch calls this on rx path, so we need a different check.
- */
-static inline bool skb_needs_check(struct sk_buff *skb, bool tx_path)
-{
-	if (tx_path)
-		return skb->ip_summed != CHECKSUM_PARTIAL;
-	else
-		return skb->ip_summed == CHECKSUM_NONE;
-}
-
 /**
  *	__skb_gso_segment - Perform segmentation on skb.
  *	@skb: buffer to segment
@@ -2294,10 +2283,11 @@  static inline bool skb_needs_check(struct sk_buff *skb, bool tx_path)
 struct sk_buff *__skb_gso_segment(struct sk_buff *skb,
 				  netdev_features_t features, bool tx_path)
 {
-	if (unlikely(skb_needs_check(skb, tx_path))) {
+	if (unlikely(skb->ip_summed != CHECKSUM_PARTIAL)) {
 		int err;
 
-		skb_warn_bad_offload(skb);
+		if (tx_path)
+			skb_warn_bad_offload(skb);
 
 		if (skb_header_cloned(skb) &&
 		    (err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC)))