-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello,
First, sorry for my english. I have to get better it.
I am trying to develop a new protocol for the linux kernel 2.6.26
similar to RSTP by my free time. Because of it I am searching the bridging
code. I began from the function netif_receive_skb (I don't need to know
anything about the hardware or the interruptions) looking for the first
functions connected with the spanning tree protocol implementation.
I found this:
1- The function netif_receive_skb calls to: skb = handle_bridge(skb,
&pt_prev, &ret, orig_dev)
2- In handle_bridge(skb, &pt_prev, &ret, orig_dev) I found:
return
br_handle_frame_hook(port, skb)
3- In br_handle_frame I found: if (NF_HOOK(PF_BRIDGE, NF_BR_LOCAL_IN, skb,
skb->dev, NULL, br_handle_local_finish))
4- In br_handle_local_finish I didn't find anything.
I have been reading the book "Understanding Linux Network Internals
(2005)" but this book (I think) is for the kernel 2.6.16. In this book the
author speaks about the function br_stp_handle_bpdu. With the kernel 2.6.26
in spite of that function you can find br_handle_local_finish.
I continued searching in netif_receive_skb. In this function after the
call to handle_bridge I found this code:
2016 skb = handle_bridge(skb, &pt_prev, &ret, orig_dev);
2107 if (!skb)
2108 goto out;
2109 skb = handle_macvlan(skb, &pt_prev, &ret, orig_dev);
2110 if (!skb)
2111 goto out;
2112
2113 type = skb->protocol;
2114 list_for_each_entry_rcu(ptype,
2115 &ptype_base[ntohs(type) &
PTYPE_HASH_MASK], list) {
2116 if (ptype->type == type &&
2117 (!ptype->dev || ptype->dev == skb->dev)) {
2118 if (pt_prev)
2119 ret = deliver_skb(skb, pt_prev,
orig_dev);
2120 pt_prev = ptype;
2121 }
2122 }
2123
2124 if (pt_prev) {
2125 ret = pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
2126 } else {
2127 kfree_skb(skb);
2128 /* Jamal, now you will not able to escape explaining
2129 * me how you were going to use this. :-)
2130 */
2131 ret = NET_RX_DROP;
2132 }
My question (probably stupid question, sorry but this is my first time
with the linux kernel) is: Where connect the SPANING TREE PROTOCOL
functions with the bridging code?
Thanks in advance.
- --
Gustavo Mart?n Morcuende
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Use GnuPG with Firefox : http://getfiregpg.org (Version: 0.7.8)
iEYEARECAAYFAkqMsJMACgkQ/bktI8Hg05BBtACeI3MSI3HgY2rX5p+zoOi6NOKA
uxEAn1WAk3W+a9r8gsvHWhRGr78hMrih
=YALX
-----END PGP SIGNATURE-----
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.linux-foundation.org/pipermail/bridge/attachments/20090820/6da988a9/attachment.htm
On Thu, 20 Aug 2009 03:56:16 +0200 Gustavo Martin <gu.martinm at gmail.com> wrote:> -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Hello, > > First, sorry for my english. I have to get better it. > I am trying to develop a new protocol for the linux kernel 2.6.26 > similar to RSTP by my free time. Because of it I am searching the bridging > code. I began from the function netif_receive_skb (I don't need to know > anything about the hardware or the interruptions) looking for the first > functions connected with the spanning tree protocol implementation. > I found this: > > 1- The function netif_receive_skb calls to: skb = handle_bridge(skb, > &pt_prev, &ret, orig_dev) > 2- In handle_bridge(skb, &pt_prev, &ret, orig_dev) I found: return > br_handle_frame_hook(port, skb) > 3- In br_handle_frame I found: if (NF_HOOK(PF_BRIDGE, NF_BR_LOCAL_IN, skb, > skb->dev, NULL, br_handle_local_finish)) > 4- In br_handle_local_finish I didn't find anything. > > I have been reading the book "Understanding Linux Network Internals > (2005)" but this book (I think) is for the kernel 2.6.16. In this book the > author speaks about the function br_stp_handle_bpdu. With the kernel 2.6.26 > in spite of that function you can find br_handle_local_finish. > > I continued searching in netif_receive_skb. In this function after the > call to handle_bridge I found this code: > > 2016 skb = handle_bridge(skb, &pt_prev, &ret, orig_dev); > 2107 if (!skb) > 2108 goto out; > 2109 skb = handle_macvlan(skb, &pt_prev, &ret, orig_dev); > 2110 if (!skb) > 2111 goto out; > 2112 > 2113 type = skb->protocol; > 2114 list_for_each_entry_rcu(ptype, > 2115 &ptype_base[ntohs(type) & > PTYPE_HASH_MASK], list) { > 2116 if (ptype->type == type && > 2117 (!ptype->dev || ptype->dev == skb->dev)) { > 2118 if (pt_prev) > 2119 ret = deliver_skb(skb, pt_prev, > orig_dev); > 2120 pt_prev = ptype; > 2121 } > 2122 } > 2123 > 2124 if (pt_prev) { > 2125 ret = pt_prev->func(skb, skb->dev, pt_prev, orig_dev); > 2126 } else { > 2127 kfree_skb(skb); > 2128 /* Jamal, now you will not able to escape explaining > 2129 * me how you were going to use this. :-) > 2130 */ > 2131 ret = NET_RX_DROP; > 2132 } > > > My question (probably stupid question, sorry but this is my first time > with the linux kernel) is: Where connect the SPANING TREE PROTOCOL > functions with the bridging code?net/bridge/br_stp_*.c You might be better off looking at the RSTP code in user space. See: http://git.kernel.org/?p=linux/kernel/git/shemminger/rstp.git;a=summary To get: git clone git://git.kernel.org/pub/scm/linux/kernel/git/shemminger/rstp.git