aboutsummaryrefslogtreecommitdiffstats
path: root/include/linux/skbuff.h
diff options
context:
space:
mode:
authorHerbert Xu <herbert@gondor.apana.org.au>2007-07-11 01:41:55 -0400
committerDavid S. Miller <davem@davemloft.net>2007-07-11 01:41:55 -0400
commitc6c6e3e05c0b4349824efcdd36650e7be9d5c7c3 (patch)
tree327317560c35fc1c54379c1729c093e3a099de87 /include/linux/skbuff.h
parentcfbba49d80be6cf8d3872b66fc5421f119843b36 (diff)
[NET]: Update comments for skb checksums
Rusty (whose comments we should all study and emulate :) pointed out that our comments for skb checksums are no longer up-to-date. So here is a patch to 1) add the case of partial checksums on input; 2) update partial checksum case to mention csum_start/csum_offset; 3) mention the new IPv6 feature bit. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'include/linux/skbuff.h')
-rw-r--r--include/linux/skbuff.h12
1 files changed, 10 insertions, 2 deletions
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 625d73b07ab7..9391e4a4c344 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -65,13 +65,20 @@
65 * is able to produce some skb->csum, it MUST use COMPLETE, 65 * is able to produce some skb->csum, it MUST use COMPLETE,
66 * not UNNECESSARY. 66 * not UNNECESSARY.
67 * 67 *
68 * PARTIAL: identical to the case for output below. This may occur
69 * on a packet received directly from another Linux OS, e.g.,
70 * a virtualised Linux kernel on the same host. The packet can
71 * be treated in the same way as UNNECESSARY except that on
72 * output (i.e., forwarding) the checksum must be filled in
73 * by the OS or the hardware.
74 *
68 * B. Checksumming on output. 75 * B. Checksumming on output.
69 * 76 *
70 * NONE: skb is checksummed by protocol or csum is not required. 77 * NONE: skb is checksummed by protocol or csum is not required.
71 * 78 *
72 * PARTIAL: device is required to csum packet as seen by hard_start_xmit 79 * PARTIAL: device is required to csum packet as seen by hard_start_xmit
73 * from skb->transport_header to the end and to record the checksum 80 * from skb->csum_start to the end and to record the checksum
74 * at skb->transport_header + skb->csum. 81 * at skb->csum_start + skb->csum_offset.
75 * 82 *
76 * Device must show its capabilities in dev->features, set 83 * Device must show its capabilities in dev->features, set
77 * at device setup time. 84 * at device setup time.
@@ -82,6 +89,7 @@
82 * TCP/UDP over IPv4. Sigh. Vendors like this 89 * TCP/UDP over IPv4. Sigh. Vendors like this
83 * way by an unknown reason. Though, see comment above 90 * way by an unknown reason. Though, see comment above
84 * about CHECKSUM_UNNECESSARY. 8) 91 * about CHECKSUM_UNNECESSARY. 8)
92 * NETIF_F_IPV6_CSUM about as dumb as the last one but does IPv6 instead.
85 * 93 *
86 * Any questions? No questions, good. --ANK 94 * Any questions? No questions, good. --ANK
87 */ 95 */