linux/net
YOSHIFUJI Hideaki / 吉藤英明 b7dc8c3959 ndisc: Remove unused space at tail of skb for ndisc messages. (TAKE 3)
Currently, the size of skb allocated for NDISC is MAX_HEADER +
LL_RESERVED_SPACE(dev) + packet length + dev->needed_tailroom,
but only LL_RESERVED_SPACE(dev) bytes is "reserved" for headers.
As a result, the skb looks like this (after construction of the
message):

head       data                   tail                       end
+--------------------------------------------------------------+
+           |                      |          |                |
+--------------------------------------------------------------+
|<-hlen---->|<---ipv6 packet------>|<--tlen-->|<--MAX_HEADER-->|
    =LL_                               = dev
     RESERVED_                           ->needed_
     SPACE(dev)                            tailroom

As the name implies, "MAX_HEADER" is used for headers, and should
be "reserved" in prior to packet construction.  Or, if some space
is really required at the tail of ther skb, it should be
explicitly documented.

We have several option after construction of NDISC message:

Option 1:

head       data                   tail       end
+---------------------------------------------+
+           |                      |          |
+---------------------------------------------+
|<-hlen---->|<---ipv6 packet------>|<--tlen-->|
   =LL_                                = dev
    RESERVED_                           ->needed_
    SPACE(dev)                            tailroom

Option 2:

head            data                   tail       end
+--------------------------------------------------+
+                |                      |          |
+--------------------------------------------------+
|<--MAX_HEADER-->|<---ipv6 packet------>|<--tlen-->|
                                            = dev
                                             ->needed_
                                               tailroom

Option 3:

head                        data                   tail       end
+--------------------------------------------------------------+
+                |           |                      |          |
+--------------------------------------------------------------+
|<--MAX_HEADER-->|<-hlen---->|<---ipv6 packet------>|<--tlen-->|
                    =LL_                                = dev
                     RESERVED_                          ->needed_
                     SPACE(dev)                           tailroom

Our tunnel drivers try expanding headroom and the space for tunnel
encapsulation was not a mandatory space -- so we are not seeing
bugs here --, but just for optimization for performance critial
situations.

Since NDISC messages are not performance critical unlike TCP,
and as we know outgoing device, LL_RESERVED_SPACE(dev) should be
just enough for the device in most (if not all) cases:
  LL_RESERVED_SPACE(dev) <= LL_MAX_HEADER <= MAX_HEADER
Note that LL_RESERVED_SPACE(dev) is also enough for NDISC over
SIT (e.g., ISATAP).

So, I think Option 1 is just fine here.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-01-04 15:16:44 -08:00
..
9p virtio: 9p: correctly pass physical address to userspace for high pages 2012-10-22 18:19:36 +10:30
802
8021q vlan: remove usage of dev->master in __vlan_find_dev_deep() 2013-01-04 13:31:50 -08:00
appletalk
atm atm: use scnprintf() instead of sprintf() 2012-12-17 20:50:51 -08:00
ax25
batman-adv net: remove unnecessary NET_ADDR_RANDOM "bitclean" 2013-01-03 22:37:36 -08:00
bluetooth Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid 2012-12-13 12:00:48 -08:00
bridge bridge: remove usage of netdev_set_master() 2013-01-04 13:31:50 -08:00
caif caif_usb: Make the driver name check more efficient 2012-12-09 00:34:02 -05:00
can Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2012-12-13 12:00:02 -08:00
ceph Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client 2012-12-20 14:00:13 -08:00
core net: remove no longer used netdev_set_bond_master() and netdev_set_master() 2013-01-04 13:31:50 -08:00
dcb net: Allow DCBnl to use other namespaces besides init_net 2012-12-10 14:09:01 -05:00
dccp inet: Fix kmemleak in tcp_v4/6_syn_recv_sock and dccp_v4/6_request_recv_sock 2012-12-14 13:14:07 -05:00
decnet net: Push capable(CAP_NET_ADMIN) into the rtnl methods 2012-11-18 20:32:44 -05:00
dns_resolver Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security 2012-12-16 15:40:50 -08:00
dsa dsa: Hide core config options; make drivers select what they need 2012-11-26 17:10:44 -05:00
ethernet net: remove unnecessary NET_ADDR_RANDOM "bitclean" 2013-01-03 22:37:36 -08:00
ieee802154 6lowpan: consider checksum bytes in fragmentation threshold 2012-11-30 12:19:24 -05:00
ipv4 tcp: make proc_tcp_fastopen_key static 2012-12-28 20:32:36 -08:00
ipv6 ndisc: Remove unused space at tail of skb for ndisc messages. (TAKE 3) 2013-01-04 15:16:44 -08:00
ipx
irda TTY/Serial merge for 3.8-rc1 2012-12-11 14:08:47 -08:00
iucv
key net: Allow userns root to control llc, netfilter, netlink, packet, and xfrm 2012-11-18 20:32:45 -05:00
l2tp Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2012-11-10 18:32:51 -05:00
lapb
llc net: Allow userns root to control llc, netfilter, netlink, packet, and xfrm 2012-11-18 20:32:45 -05:00
mac80211 Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2012-12-13 12:00:02 -08:00
mac802154 mac802154: fix destructon ordering for ieee802154 devices 2012-12-14 13:14:07 -05:00
netfilter Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2012-12-13 12:00:02 -08:00
netlabel
netlink netlink: validate addr_len on bind 2012-12-17 20:50:51 -08:00
netrom
nfc nfc: remove noisy message from llcp_sock_sendmsg 2012-12-13 12:58:10 -05:00
openvswitch net: remove unnecessary NET_ADDR_RANDOM "bitclean" 2013-01-03 22:37:36 -08:00
packet net: Allow userns root to control llc, netfilter, netlink, packet, and xfrm 2012-11-18 20:32:45 -05:00
phonet net: Push capable(CAP_NET_ADMIN) into the rtnl methods 2012-11-18 20:32:44 -05:00
rds IB/rds: suppress incompatible protocol when version is known 2012-12-26 15:17:37 -08:00
rfkill Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2012-12-13 12:00:02 -08:00
rose
rxrpc
sched net: sched: integer overflow fix 2012-12-22 00:03:00 -08:00
sctp sctp: make sctp_addr_wq_timeout_handler static 2012-12-28 20:32:36 -08:00
sunrpc Merge branch 'for-3.8' of git://linux-nfs.org/~bfields/linux 2012-12-20 14:04:11 -08:00
tipc tipc: refactor accept() code for improved readability 2012-12-07 17:23:24 -05:00
unix net: Don't export sysctls to unprivileged users 2012-11-18 20:30:55 -05:00
wanrouter
wimax
wireless CONFIG_HOTPLUG removal from networking core 2012-12-22 00:03:00 -08:00
x25
xfrm Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next 2012-11-22 15:25:55 -05:00
compat.c
Kconfig
Makefile ipv6: Preserve ipv6 functionality needed by NET 2012-11-18 02:34:00 -05:00
nonet.c
socket.c cgroup: net_cls: Rework update socket logic 2012-10-26 03:40:51 -04:00
sysctl_net.c user_ns: get rid of duplicate code in net_ctl_permissions 2012-11-18 20:32:45 -05:00