linux.git/net/ipv4/ping.c, branch v6.6 Linux kernel source tree inet: move inet->recverr to inet->inet_flags 2023-08-16T10:09:17+00:00 Eric Dumazet edumazet@google.com 2023-08-16T08:15:35+00:00 6b5f43ea08150e7ff72f734545101c58489ead5b IP_RECVERR socket option can now be set/get without locking the socket. This patch potentially avoid data-races around inet->recverr. Signed-off-by: Eric Dumazet <edumazet@google.com> Acked-by: Soheil Hassas Yeganeh <soheil@google.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> 
IP_RECVERR socket option can now be set/get without locking the socket.

This patch potentially avoid data-races around inet->recverr.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
inet: introduce inet->inet_flags 2023-08-16T10:09:16+00:00 Eric Dumazet edumazet@google.com 2023-08-16T08:15:33+00:00 c274af2242693f59f00851b3660a21b10bcd76cc Various inet fields are currently racy. do_ip_setsockopt() and do_ip_getsockopt() are mostly holding the socket lock, but some (fast) paths do not. Use a new inet->inet_flags to hold atomic bits in the series. Remove inet->cmsg_flags, and use instead 9 bits from inet_flags. Signed-off-by: Eric Dumazet <edumazet@google.com> Acked-by: Soheil Hassas Yeganeh <soheil@google.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> 
Various inet fields are currently racy.

do_ip_setsockopt() and do_ip_getsockopt() are mostly holding
the socket lock, but some (fast) paths do not.

Use a new inet->inet_flags to hold atomic bits in the series.

Remove inet->cmsg_flags, and use instead 9 bits from inet_flags.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
ping: Stop using RTO_ONLINK. 2023-05-24T07:22:06+00:00 Guillaume Nault gnault@redhat.com 2023-05-22T14:37:57+00:00 726de790f66029a7654b3e748f8d3e7888a30ae5 Define a new helper to figure out the correct route scope to use on TX, depending on socket configuration, ancillary data and send flags. Use this new helper to properly initialise the scope in flowi4_init_output(), instead of overriding tos with the RTO_ONLINK flag. The objective is to eventually remove RTO_ONLINK, which will allow converting .flowi4_tos to dscp_t. Signed-off-by: Guillaume Nault <gnault@redhat.com> Reviewed-by: David Ahern <dsahern@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> 
Define a new helper to figure out the correct route scope to use on TX,
depending on socket configuration, ancillary data and send flags.

Use this new helper to properly initialise the scope in
flowi4_init_output(), instead of overriding tos with the RTO_ONLINK
flag.

The objective is to eventually remove RTO_ONLINK, which will allow
converting .flowi4_tos to dscp_t.

Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
ping: Convert hlist_nulls to plain hlist. 2023-05-13T19:04:31+00:00 Kuniyuki Iwashima kuniyu@amazon.com 2023-05-10T21:54:43+00:00 f1b5dfe63f6a9eb17948cbaee4da4b66f51ac794 Since introduced in commit c319b4d76b9e ("net: ipv4: add IPPROTO_ICMP socket kind"), ping socket does not use SLAB_TYPESAFE_BY_RCU nor check nulls marker in loops. Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: David S. Miller <davem@davemloft.net> 
Since introduced in commit c319b4d76b9e ("net: ipv4: add IPPROTO_ICMP
socket kind"), ping socket does not use SLAB_TYPESAFE_BY_RCU nor check
nulls marker in loops.

Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ping: Fix potentail NULL deref for /proc/net/icmp. 2023-04-05T01:56:58+00:00 Kuniyuki Iwashima kuniyu@amazon.com 2023-04-03T19:49:59+00:00 ab5fb73ffa01072b4d8031cc05801fa1cb653bee After commit dbca1596bbb0 ("ping: convert to RCU lookups, get rid of rwlock"), we use RCU for ping sockets, but we should use spinlock for /proc/net/icmp to avoid a potential NULL deref mentioned in the previous patch. Let's go back to using spinlock there. Note we can convert ping sockets to use hlist instead of hlist_nulls because we do not use SLAB_TYPESAFE_BY_RCU for ping sockets. Fixes: dbca1596bbb0 ("ping: convert to RCU lookups, get rid of rwlock") Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
After commit dbca1596bbb0 ("ping: convert to RCU lookups, get rid
of rwlock"), we use RCU for ping sockets, but we should use spinlock
for /proc/net/icmp to avoid a potential NULL deref mentioned in
the previous patch.

Let's go back to using spinlock there.

Note we can convert ping sockets to use hlist instead of hlist_nulls
because we do not use SLAB_TYPESAFE_BY_RCU for ping sockets.

Fixes: dbca1596bbb0 ("ping: convert to RCU lookups, get rid of rwlock")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-12-09T02:19:59+00:00 Jakub Kicinski kuba@kernel.org 2022-12-09T00:07:53+00:00 837e8ac871499d337212e2161c046f5adf1bad33 No conflicts. Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
No conflicts.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
inet: ping: use hlist_nulls rcu iterator during lookup 2022-12-01T11:42:46+00:00 Florian Westphal fw@strlen.de 2022-11-29T14:06:44+00:00 c25b7a7a565e5eeb2459b37583eea67942057511 ping_lookup() does not acquire the table spinlock, so iteration should use hlist_nulls_for_each_entry_rcu(). Spotted during code review. Fixes: dbca1596bbb0 ("ping: convert to RCU lookups, get rid of rwlock") Cc: Eric Dumazet <edumazet@google.com> Signed-off-by: Florian Westphal <fw@strlen.de> Link: https://lore.kernel.org/r/20221129140644.28525-1-fw@strlen.de Signed-off-by: Paolo Abeni <pabeni@redhat.com> 
ping_lookup() does not acquire the table spinlock, so iteration should
use hlist_nulls_for_each_entry_rcu().

Spotted during code review.

Fixes: dbca1596bbb0 ("ping: convert to RCU lookups, get rid of rwlock")
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://lore.kernel.org/r/20221129140644.28525-1-fw@strlen.de
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
net: Return errno in sk->sk_prot->get_port(). 2022-11-21T13:05:39+00:00 Kuniyuki Iwashima kuniyu@amazon.com 2022-11-18T18:25:06+00:00 7a7160edf1bfde25422262fb26851cef65f695d3 We assume the correct errno is -EADDRINUSE when sk->sk_prot->get_port() fails, so some ->get_port() functions return just 1 on failure and the callers return -EADDRINUSE instead. However, mptcp_get_port() can return -EINVAL. Let's not ignore the error. Note the only exception is inet_autobind(), all of whose callers return -EAGAIN instead. Fixes: cec37a6e41aa ("mptcp: Handle MP_CAPABLE options for outgoing connections") Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Signed-off-by: David S. Miller <davem@davemloft.net> 
We assume the correct errno is -EADDRINUSE when sk->sk_prot->get_port()
fails, so some ->get_port() functions return just 1 on failure and the
callers return -EADDRINUSE instead.

However, mptcp_get_port() can return -EINVAL.  Let's not ignore the error.

Note the only exception is inet_autobind(), all of whose callers return
-EAGAIN instead.

Fixes: cec37a6e41aa ("mptcp: Handle MP_CAPABLE options for outgoing connections")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
inet: ping: fix recent breakage 2022-10-12T08:10:02+00:00 Eric Dumazet edumazet@google.com 2022-10-11T21:27:29+00:00 0d24148bd276ead5708ef56a4725580555bb48a3 Blamed commit broke the assumption used by ping sendmsg() that allocated skb would have MAX_HEADER bytes in skb->head. This patch changes the way ping works, by making sure the skb head contains space for the icmp header, and adjusting ping_getfrag() which was desperate about going past the icmp header :/ This is adopting what UDP does, mostly. syzbot is able to crash a host using both kfence and following repro in a loop. fd = socket(AF_INET6, SOCK_DGRAM, IPPROTO_ICMPV6) connect(fd, {sa_family=AF_INET6, sin6_port=htons(0), sin6_flowinfo=htonl(0), inet_pton(AF_INET6, "::1", &sin6_addr), sin6_scope_id=0}, 28 sendmsg(fd, {msg_name=NULL, msg_namelen=0, msg_iov=[ {iov_base="\200\0\0\0\23\0\0\0\0\0\0\0\0\0"..., iov_len=65496}], msg_iovlen=1, msg_controllen=0, msg_flags=0}, 0 When kfence triggers, skb->head only has 64 bytes, immediately followed by struct skb_shared_info (no extra headroom based on ksize(ptr)) Then icmpv6_push_pending_frames() is overwriting first bytes of skb_shinfo(skb), making nr_frags bigger than MAX_SKB_FRAGS, and/or setting shinfo->gso_size to a non zero value. If nr_frags is mangled, a crash happens in skb_release_data() If gso_size is mangled, we have the following report: lo: caps=(0x00000516401d7c69, 0x00000516401d7c69) WARNING: CPU: 0 PID: 7548 at net/core/dev.c:3239 skb_warn_bad_offload+0x119/0x230 net/core/dev.c:3239 Modules linked in: CPU: 0 PID: 7548 Comm: syz-executor268 Not tainted 6.0.0-syzkaller-02754-g557f050166e5 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/22/2022 RIP: 0010:skb_warn_bad_offload+0x119/0x230 net/core/dev.c:3239 Code: 70 03 00 00 e8 58 c3 24 fa 4c 8d a5 e8 00 00 00 e8 4c c3 24 fa 4c 89 e9 4c 89 e2 4c 89 f6 48 c7 c7 00 53 f5 8a e8 13 ac e7 01 <0f> 0b 5b 5d 41 5c 41 5d 41 5e e9 28 c3 24 fa e8 23 c3 24 fa 48 89 RSP: 0018:ffffc9000366f3e8 EFLAGS: 00010282 RAX: 0000000000000000 RBX: ffff88807a9d9d00 RCX: 0000000000000000 RDX: ffff8880780c0000 RSI: ffffffff8160f6f8 RDI: fffff520006cde6f RBP: ffff888079952000 R08: 0000000000000005 R09: 0000000000000000 R10: 0000000000000400 R11: 0000000000000000 R12: ffff8880799520e8 R13: ffff88807a9da070 R14: ffff888079952000 R15: 0000000000000000 FS: 0000555556be6300(0000) GS:ffff8880b9a00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000020010000 CR3: 000000006eb7b000 CR4: 00000000003506f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: <TASK> gso_features_check net/core/dev.c:3521 [inline] netif_skb_features+0x83e/0xb90 net/core/dev.c:3554 validate_xmit_skb+0x2b/0xf10 net/core/dev.c:3659 __dev_queue_xmit+0x998/0x3ad0 net/core/dev.c:4248 dev_queue_xmit include/linux/netdevice.h:3008 [inline] neigh_hh_output include/net/neighbour.h:530 [inline] neigh_output include/net/neighbour.h:544 [inline] ip6_finish_output2+0xf97/0x1520 net/ipv6/ip6_output.c:134 __ip6_finish_output net/ipv6/ip6_output.c:195 [inline] ip6_finish_output+0x690/0x1160 net/ipv6/ip6_output.c:206 NF_HOOK_COND include/linux/netfilter.h:291 [inline] ip6_output+0x1ed/0x540 net/ipv6/ip6_output.c:227 dst_output include/net/dst.h:445 [inline] ip6_local_out+0xaf/0x1a0 net/ipv6/output_core.c:161 ip6_send_skb+0xb7/0x340 net/ipv6/ip6_output.c:1966 ip6_push_pending_frames+0xdd/0x100 net/ipv6/ip6_output.c:1986 icmpv6_push_pending_frames+0x2af/0x490 net/ipv6/icmp.c:303 ping_v6_sendmsg+0xc44/0x1190 net/ipv6/ping.c:190 inet_sendmsg+0x99/0xe0 net/ipv4/af_inet.c:819 sock_sendmsg_nosec net/socket.c:714 [inline] sock_sendmsg+0xcf/0x120 net/socket.c:734 ____sys_sendmsg+0x712/0x8c0 net/socket.c:2482 ___sys_sendmsg+0x110/0x1b0 net/socket.c:2536 __sys_sendmsg+0xf3/0x1c0 net/socket.c:2565 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7f21aab42b89 Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 41 15 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007fff1729d038 EFLAGS: 00000246 ORIG_RAX: 000000000000002e RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 00007f21aab42b89 RDX: 0000000000000000 RSI: 0000000020000180 RDI: 0000000000000003 RBP: 0000000000000000 R08: 000000000000000d R09: 000000000000000d R10: 000000000000000d R11: 0000000000000246 R12: 00007fff1729d050 R13: 00000000000f4240 R14: 0000000000021dd1 R15: 00007fff1729d044 </TASK> Fixes: 47cf88993c91 ("net: unify alloclen calculation for paged requests") Reported-by: syzbot <syzkaller@googlegroups.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Pavel Begunkov <asml.silence@gmail.com> Cc: Lorenzo Colitti <lorenzo@google.com> Cc: Willem de Bruijn <willemb@google.com> Cc: Maciej Żenczykowski <maze@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> 
Blamed commit broke the assumption used by ping sendmsg() that
allocated skb would have MAX_HEADER bytes in skb->head.

This patch changes the way ping works, by making sure
the skb head contains space for the icmp header,
and adjusting ping_getfrag() which was desperate
about going past the icmp header :/

This is adopting what UDP does, mostly.

syzbot is able to crash a host using both kfence and following repro in a loop.

fd = socket(AF_INET6, SOCK_DGRAM, IPPROTO_ICMPV6)
connect(fd, {sa_family=AF_INET6, sin6_port=htons(0), sin6_flowinfo=htonl(0),
		inet_pton(AF_INET6, "::1", &sin6_addr), sin6_scope_id=0}, 28
sendmsg(fd, {msg_name=NULL, msg_namelen=0, msg_iov=[
		{iov_base="\200\0\0\0\23\0\0\0\0\0\0\0\0\0"..., iov_len=65496}],
		msg_iovlen=1, msg_controllen=0, msg_flags=0}, 0

When kfence triggers, skb->head only has 64 bytes, immediately followed
by struct skb_shared_info (no extra headroom based on ksize(ptr))

Then icmpv6_push_pending_frames() is overwriting first bytes
of skb_shinfo(skb), making nr_frags bigger than MAX_SKB_FRAGS,
and/or setting shinfo->gso_size to a non zero value.

If nr_frags is mangled, a crash happens in skb_release_data()

If gso_size is mangled, we have the following report:

lo: caps=(0x00000516401d7c69, 0x00000516401d7c69)
WARNING: CPU: 0 PID: 7548 at net/core/dev.c:3239 skb_warn_bad_offload+0x119/0x230 net/core/dev.c:3239
Modules linked in:
CPU: 0 PID: 7548 Comm: syz-executor268 Not tainted 6.0.0-syzkaller-02754-g557f050166e5 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/22/2022
RIP: 0010:skb_warn_bad_offload+0x119/0x230 net/core/dev.c:3239
Code: 70 03 00 00 e8 58 c3 24 fa 4c 8d a5 e8 00 00 00 e8 4c c3 24 fa 4c 89 e9 4c 89 e2 4c 89 f6 48 c7 c7 00 53 f5 8a e8 13 ac e7 01 <0f> 0b 5b 5d 41 5c 41 5d 41 5e e9 28 c3 24 fa e8 23 c3 24 fa 48 89
RSP: 0018:ffffc9000366f3e8 EFLAGS: 00010282
RAX: 0000000000000000 RBX: ffff88807a9d9d00 RCX: 0000000000000000
RDX: ffff8880780c0000 RSI: ffffffff8160f6f8 RDI: fffff520006cde6f
RBP: ffff888079952000 R08: 0000000000000005 R09: 0000000000000000
R10: 0000000000000400 R11: 0000000000000000 R12: ffff8880799520e8
R13: ffff88807a9da070 R14: ffff888079952000 R15: 0000000000000000
FS: 0000555556be6300(0000) GS:ffff8880b9a00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000020010000 CR3: 000000006eb7b000 CR4: 00000000003506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>
gso_features_check net/core/dev.c:3521 [inline]
netif_skb_features+0x83e/0xb90 net/core/dev.c:3554
validate_xmit_skb+0x2b/0xf10 net/core/dev.c:3659
__dev_queue_xmit+0x998/0x3ad0 net/core/dev.c:4248
dev_queue_xmit include/linux/netdevice.h:3008 [inline]
neigh_hh_output include/net/neighbour.h:530 [inline]
neigh_output include/net/neighbour.h:544 [inline]
ip6_finish_output2+0xf97/0x1520 net/ipv6/ip6_output.c:134
__ip6_finish_output net/ipv6/ip6_output.c:195 [inline]
ip6_finish_output+0x690/0x1160 net/ipv6/ip6_output.c:206
NF_HOOK_COND include/linux/netfilter.h:291 [inline]
ip6_output+0x1ed/0x540 net/ipv6/ip6_output.c:227
dst_output include/net/dst.h:445 [inline]
ip6_local_out+0xaf/0x1a0 net/ipv6/output_core.c:161
ip6_send_skb+0xb7/0x340 net/ipv6/ip6_output.c:1966
ip6_push_pending_frames+0xdd/0x100 net/ipv6/ip6_output.c:1986
icmpv6_push_pending_frames+0x2af/0x490 net/ipv6/icmp.c:303
ping_v6_sendmsg+0xc44/0x1190 net/ipv6/ping.c:190
inet_sendmsg+0x99/0xe0 net/ipv4/af_inet.c:819
sock_sendmsg_nosec net/socket.c:714 [inline]
sock_sendmsg+0xcf/0x120 net/socket.c:734
____sys_sendmsg+0x712/0x8c0 net/socket.c:2482
___sys_sendmsg+0x110/0x1b0 net/socket.c:2536
__sys_sendmsg+0xf3/0x1c0 net/socket.c:2565
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7f21aab42b89
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 41 15 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007fff1729d038 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 00007f21aab42b89
RDX: 0000000000000000 RSI: 0000000020000180 RDI: 0000000000000003
RBP: 0000000000000000 R08: 000000000000000d R09: 000000000000000d
R10: 000000000000000d R11: 0000000000000246 R12: 00007fff1729d050
R13: 00000000000f4240 R14: 0000000000021dd1 R15: 00007fff1729d044
</TASK>

Fixes: 47cf88993c91 ("net: unify alloclen calculation for paged requests")
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Pavel Begunkov <asml.silence@gmail.com>
Cc: Lorenzo Colitti <lorenzo@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Cc: Maciej Żenczykowski <maze@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ipv6: ping: fix wrong checksum for large frames 2022-10-12T08:10:02+00:00 Eric Dumazet edumazet@google.com 2022-10-11T21:27:28+00:00 87445f369cca2965620e79f87145d3d7fa35befd For a given ping datagram, ping_getfrag() is called once per skb fragment. A large datagram requiring more than one page fragment is currently getting the checksum of the last fragment, instead of the cumulative one. After this patch, "ping -s 35000 ::1" is working correctly. Fixes: 6d0bfe226116 ("net: ipv6: Add IPv6 support to the ping socket.") Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Lorenzo Colitti <lorenzo@google.com> Cc: Maciej Żenczykowski <maze@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> 
For a given ping datagram, ping_getfrag() is called once
per skb fragment.

A large datagram requiring more than one page fragment
is currently getting the checksum of the last fragment,
instead of the cumulative one.

After this patch, "ping -s 35000 ::1" is working correctly.

Fixes: 6d0bfe226116 ("net: ipv6: Add IPv6 support to the ping socket.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Lorenzo Colitti <lorenzo@google.com>
Cc: Maciej Żenczykowski <maze@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>