All of lore.kernel.org
 help / color / mirror / Atom feed
* [Patch net-next] net: explain __skb_checksum_complete() with comments
@ 2018-11-26 17:31 Cong Wang
  2018-11-29 21:47 ` David Miller
  0 siblings, 1 reply; 2+ messages in thread
From: Cong Wang @ 2018-11-26 17:31 UTC (permalink / raw)
  To: netdev; +Cc: Cong Wang, Herbert Xu

Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
---
 net/core/dev.c    |  1 +
 net/core/skbuff.c | 18 +++++++++++++++++-
 2 files changed, 18 insertions(+), 1 deletion(-)

diff --git a/net/core/dev.c b/net/core/dev.c
index 368dc3b49dc0..ee0a4ac0bbb6 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -5787,6 +5787,7 @@ __sum16 __skb_gro_checksum_complete(struct sk_buff *skb)
 
 	/* NAPI_GRO_CB(skb)->csum holds pseudo checksum */
 	sum = csum_fold(csum_add(NAPI_GRO_CB(skb)->csum, wsum));
+	/* See comments in __skb_checksum_complete(). */
 	if (likely(!sum)) {
 		if (unlikely(skb->ip_summed == CHECKSUM_COMPLETE) &&
 		    !skb->csum_complete_sw)
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 6f2ea0f0fb75..530097df328f 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -2637,6 +2637,7 @@ __sum16 __skb_checksum_complete_head(struct sk_buff *skb, int len)
 	__sum16 sum;
 
 	sum = csum_fold(skb_checksum(skb, 0, len, skb->csum));
+	/* See comments in __skb_checksum_complete(). */
 	if (likely(!sum)) {
 		if (unlikely(skb->ip_summed == CHECKSUM_COMPLETE) &&
 		    !skb->csum_complete_sw)
@@ -2648,6 +2649,15 @@ __sum16 __skb_checksum_complete_head(struct sk_buff *skb, int len)
 }
 EXPORT_SYMBOL(__skb_checksum_complete_head);
 
+/* This function assumes skb->csum already holds pseudo header's checksum,
+ * which has been changed from the hardware checksum, for example, by
+ * __skb_checksum_validate_complete(). And, the original skb->csum must
+ * have been validated unsuccessfully for CHECKSUM_COMPLETE case.
+ *
+ * It returns non-zero if the recomputed checksum is still invalid, otherwise
+ * zero. The new checksum is stored back into skb->csum unless the skb is
+ * shared.
+ */
 __sum16 __skb_checksum_complete(struct sk_buff *skb)
 {
 	__wsum csum;
@@ -2655,8 +2665,14 @@ __sum16 __skb_checksum_complete(struct sk_buff *skb)
 
 	csum = skb_checksum(skb, 0, skb->len, 0);
 
-	/* skb->csum holds pseudo checksum */
 	sum = csum_fold(csum_add(skb->csum, csum));
+	/* This check is inverted, because we already knew the hardware
+	 * checksum is invalid before calling this function. So, if the
+	 * re-computed checksum is valid instead, then we have a mismatch
+	 * between the original skb->csum and skb_checksum(). This means either
+	 * the original hardware checksum is incorrect or we screw up skb->csum
+	 * when moving skb->data around.
+	 */
 	if (likely(!sum)) {
 		if (unlikely(skb->ip_summed == CHECKSUM_COMPLETE) &&
 		    !skb->csum_complete_sw)
-- 
2.19.1

^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [Patch net-next] net: explain __skb_checksum_complete() with comments
  2018-11-26 17:31 [Patch net-next] net: explain __skb_checksum_complete() with comments Cong Wang
@ 2018-11-29 21:47 ` David Miller
  0 siblings, 0 replies; 2+ messages in thread
From: David Miller @ 2018-11-29 21:47 UTC (permalink / raw)
  To: xiyou.wangcong; +Cc: netdev, herbert

From: Cong Wang <xiyou.wangcong@gmail.com>
Date: Mon, 26 Nov 2018 09:31:26 -0800

> Cc: Herbert Xu <herbert@gondor.apana.org.au>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>

Applied, thanks Cong.

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2018-11-30  9:09 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-11-26 17:31 [Patch net-next] net: explain __skb_checksum_complete() with comments Cong Wang
2018-11-29 21:47 ` David Miller

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.