linux-stable.git/net/netfilter, branch linux-5.14.y

netfilter: nfnetlink_queue: fix OOB when mac header was cleared

2021-11-17T10:04:36+00:00

[ Upstream commit 5648b5e1169ff1d6d6a46c35c0b5fbebd2a5cbb2 ]

On 64bit platforms the MAC header is set to 0xffff on allocation and
also when a helper like skb_unset_mac_header() is called.

dev_parse_header may call skb_mac_header() which assumes valid mac offset:

 BUG: KASAN: use-after-free in eth_header_parse+0x75/0x90
 Read of size 6 at addr ffff8881075a5c05 by task nf-queue/1364
 Call Trace:
  memcpy+0x20/0x60
  eth_header_parse+0x75/0x90
  __nfqnl_enqueue_packet+0x1a61/0x3380
  __nf_queue+0x597/0x1300
  nf_queue+0xf/0x40
  nf_hook_slow+0xed/0x190
  nf_hook+0x184/0x440
  ip_output+0x1c0/0x2a0
  nf_reinject+0x26f/0x700
  nfqnl_recv_verdict+0xa16/0x18b0
  nfnetlink_rcv_msg+0x506/0xe70

The existing code only works if the skb has a mac header.

Fixes: 2c38de4c1f8da7 ("netfilter: fix looped (broad|multi)cast's MAC handling")
Signed-off-by: Florian Westphal 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

netfilter: nft_dynset: relax superfluous check on set updates

2021-11-17T10:04:04+00:00

[ Upstream commit 7b1394892de8d95748d05e3ee41e85edb4abbfa1 ]

Relax this condition to make add and update commands idempotent for sets
with no timeout. The eval function already checks if the set element
timeout is available and updates it if the update command is used.

Fixes: 22fe54d5fefc ("netfilter: nf_tables: add support for dynamic set updates")
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

netfilter: conntrack: set on IPS_ASSURED if flows enters internal stream state

2021-11-17T10:03:56+00:00

[ Upstream commit b7b1d02fc43925a4d569ec221715db2dfa1ce4f5 ]

The internal stream state sets the timeout to 120 seconds 2 seconds
after the creation of the flow, attach this internal stream state to the
IPS_ASSURED flag for consistent event reporting.

Before this patch:

      [NEW] udp      17 30 src=10.246.11.13 dst=216.239.35.0 sport=37282 dport=123 [UNREPLIED] src=216.239.35.0 dst=10.246.11.13 sport=123 dport=37282
   [UPDATE] udp      17 30 src=10.246.11.13 dst=216.239.35.0 sport=37282 dport=123 src=216.239.35.0 dst=10.246.11.13 sport=123 dport=37282
   [UPDATE] udp      17 30 src=10.246.11.13 dst=216.239.35.0 sport=37282 dport=123 src=216.239.35.0 dst=10.246.11.13 sport=123 dport=37282 [ASSURED]
  [DESTROY] udp      17 src=10.246.11.13 dst=216.239.35.0 sport=37282 dport=123 src=216.239.35.0 dst=10.246.11.13 sport=123 dport=37282 [ASSURED]

Note IPS_ASSURED for the flow not yet in the internal stream state.

after this update:

      [NEW] udp      17 30 src=10.246.11.13 dst=216.239.35.0 sport=37282 dport=123 [UNREPLIED] src=216.239.35.0 dst=10.246.11.13 sport=123 dport=37282
   [UPDATE] udp      17 30 src=10.246.11.13 dst=216.239.35.0 sport=37282 dport=123 src=216.239.35.0 dst=10.246.11.13 sport=123 dport=37282
   [UPDATE] udp      17 120 src=10.246.11.13 dst=216.239.35.0 sport=37282 dport=123 src=216.239.35.0 dst=10.246.11.13 sport=123 dport=37282 [ASSURED]
  [DESTROY] udp      17 src=10.246.11.13 dst=216.239.35.0 sport=37282 dport=123 src=216.239.35.0 dst=10.246.11.13 sport=123 dport=37282 [ASSURED]

Before this patch, short-lived UDP flows never entered IPS_ASSURED, so
they were already candidate flow to be deleted by early_drop under
stress.

Before this patch, IPS_ASSURED is set on regardless the internal stream
state, attach this internal stream state to IPS_ASSURED.

packet #1 (original direction) enters NEW state
packet #2 (reply direction) enters ESTABLISHED state, sets on IPS_SEEN_REPLY
paclet #3 (any direction) sets on IPS_ASSURED (if 2 seconds since the
          creation has passed by).

Reported-by: Maciej Żenczykowski 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

netfilter: Kconfig: use 'default y' instead of 'm' for bool config option

2021-10-27T07:59:48+00:00

commit 77076934afdcd46516caf18ed88b2f88025c9ddb upstream.

This option, NF_CONNTRACK_SECMARK, is a bool, so it can never be 'm'.

Fixes: 33b8e77605620 ("[NETFILTER]: Add CONFIG_NETFILTER_ADVANCED option")
Signed-off-by: Vegard Nossum 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Greg Kroah-Hartman

netfilter: ipvs: make global sysctl readonly in non-init netns

2021-10-27T07:59:35+00:00

[ Upstream commit 174c376278949c44aad89c514a6b5db6cee8db59 ]

Because the data pointer of net/ipv4/vs/debug_level is not updated per
netns, it must be marked as read-only in non-init netns.

Fixes: c6d2d445d8de ("IPVS: netns, final patch enabling network name space.")
Signed-off-by: Antoine Tenart 
Acked-by: Julian Anastasov 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

netfilter: nf_tables: skip netdev events generated on netns removal

2021-10-27T07:59:34+00:00

[ Upstream commit 68a3765c659f809dcaac20030853a054646eb739 ]

syzbot reported following (harmless) WARN:

 WARNING: CPU: 1 PID: 2648 at net/netfilter/core.c:468
  nft_netdev_unregister_hooks net/netfilter/nf_tables_api.c:230 [inline]
  nf_tables_unregister_hook include/net/netfilter/nf_tables.h:1090 [inline]
  __nft_release_basechain+0x138/0x640 net/netfilter/nf_tables_api.c:9524
  nft_netdev_event net/netfilter/nft_chain_filter.c:351 [inline]
  nf_tables_netdev_event+0x521/0x8a0 net/netfilter/nft_chain_filter.c:382

reproducer:
unshare -n bash -c 'ip link add br0 type bridge; nft add table netdev t ; \
 nft add chain netdev t ingress \{ type filter hook ingress device "br0" \
 priority 0\; policy drop\; \}'

Problem is that when netns device exit hooks create the UNREGISTER
event, the .pre_exit hook for nf_tables core has already removed the
base hook.  Notifier attempts to do this again.

The need to do base hook unregister unconditionally was needed in the past,
because notifier was last stage where reg->dev dereference was safe.

Now that nf_tables does the hook removal in .pre_exit, this isn't
needed anymore.

Reported-and-tested-by: syzbot+154bd5be532a63aa778b@syzkaller.appspotmail.com
Fixes: 767d1216bff825 ("netfilter: nftables: fix possible UAF over chains from packet path in netns")
Signed-off-by: Florian Westphal 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

netfilter: xt_IDLETIMER: fix panic that occurs when timer_type has garbage value

2021-10-27T07:59:34+00:00

[ Upstream commit 902c0b1887522a099aa4e1e6b4b476c2fe5dd13e ]

Currently, when the rule related to IDLETIMER is added, idletimer_tg timer
structure is initialized by kmalloc on executing idletimer_tg_create
function. However, in this process timer->timer_type is not defined to
a specific value. Thus, timer->timer_type has garbage value and it occurs
kernel panic. So, this commit fixes the panic by initializing
timer->timer_type using kzalloc instead of kmalloc.

Test commands:
    # iptables -A OUTPUT -j IDLETIMER --timeout 1 --label test
    $ cat /sys/class/xt_idletimer/timers/test
      Killed

Splat looks like:
    BUG: KASAN: user-memory-access in alarm_expires_remaining+0x49/0x70
    Read of size 8 at addr 0000002e8c7bc4c8 by task cat/917
    CPU: 12 PID: 917 Comm: cat Not tainted 5.14.0+ #3 79940a339f71eb14fc81aee1757a20d5bf13eb0e
    Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1ubuntu1.1 04/01/2014
    Call Trace:
     dump_stack_lvl+0x6e/0x9c
     kasan_report.cold+0x112/0x117
     ? alarm_expires_remaining+0x49/0x70
     __asan_load8+0x86/0xb0
     alarm_expires_remaining+0x49/0x70
     idletimer_tg_show+0xe5/0x19b [xt_IDLETIMER 11219304af9316a21bee5ba9d58f76a6b9bccc6d]
     dev_attr_show+0x3c/0x60
     sysfs_kf_seq_show+0x11d/0x1f0
     ? device_remove_bin_file+0x20/0x20
     kernfs_seq_show+0xa4/0xb0
     seq_read_iter+0x29c/0x750
     kernfs_fop_read_iter+0x25a/0x2c0
     ? __fsnotify_parent+0x3d1/0x570
     ? iov_iter_init+0x70/0x90
     new_sync_read+0x2a7/0x3d0
     ? __x64_sys_llseek+0x230/0x230
     ? rw_verify_area+0x81/0x150
     vfs_read+0x17b/0x240
     ksys_read+0xd9/0x180
     ? vfs_write+0x460/0x460
     ? do_syscall_64+0x16/0xc0
     ? lockdep_hardirqs_on+0x79/0x120
     __x64_sys_read+0x43/0x50
     do_syscall_64+0x3b/0xc0
     entry_SYSCALL_64_after_hwframe+0x44/0xae
    RIP: 0033:0x7f0cdc819142
    Code: c0 e9 c2 fe ff ff 50 48 8d 3d 3a ca 0a 00 e8 f5 19 02 00 0f 1f 44 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 0f 05 <48> 3d 00 f0 ff ff 77 56 c3 0f 1f 44 00 00 48 83 ec 28 48 89 54 24
    RSP: 002b:00007fff28eee5b8 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
    RAX: ffffffffffffffda RBX: 0000000000020000 RCX: 00007f0cdc819142
    RDX: 0000000000020000 RSI: 00007f0cdc032000 RDI: 0000000000000003
    RBP: 00007f0cdc032000 R08: 00007f0cdc031010 R09: 0000000000000000
    R10: 0000000000000022 R11: 0000000000000246 R12: 00005607e9ee31f0
    R13: 0000000000000003 R14: 0000000000020000 R15: 0000000000020000

Fixes: 68983a354a65 ("netfilter: xtables: Add snapshot of hardidletimer target")
Signed-off-by: Juhee Kang 
Reviewed-by: Florian Westphal 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

netfilter: nf_nat_masquerade: defer conntrack walk to work queue

2021-10-17T08:44:49+00:00

[ Upstream commit 7970a19b71044bf4dc2c1becc200275bdf1884d4 ]

The ipv4 and device notifiers are called with RTNL mutex held.
The table walk can take some time, better not block other RTNL users.

'ip a' has been reported to block for up to 20 seconds when conntrack table
has many entries and device down events are frequent (e.g., PPP).

Reported-and-tested-by: Martin Zaharinov 
Signed-off-by: Florian Westphal 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

netfilter: nf_nat_masquerade: make async masq_inet6_event handling generic

2021-10-17T08:44:49+00:00

[ Upstream commit 30db406923b9285a9bac06a6af5e74bd6d0f1d06 ]

masq_inet6_event is called asynchronously from system work queue,
because the inet6 notifier is atomic and nf_iterate_cleanup can sleep.

The ipv4 and device notifiers call nf_iterate_cleanup directly.

This is legal, but these notifiers are called with RTNL mutex held.
A large conntrack table with many devices coming and going will have severe
impact on the system usability, with 'ip a' blocking for several seconds.

This change places the defer code into a helper and makes it more
generic so ipv4 and ifdown notifiers can be converted to defer the
cleanup walk as well in a follow patch.

Signed-off-by: Florian Westphal 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

netfilter: nf_tables: honor NLM_F_CREATE and NLM_F_EXCL in event notification

2021-10-13T07:41:56+00:00

[ Upstream commit 6fb721cf781808ee2ca5e737fb0592cc68de3381 ]

Include the NLM_F_CREATE and NLM_F_EXCL flags in netlink event
notifications, otherwise userspace cannot distiguish between create and
add commands.

Fixes: 96518518cc41 ("netfilter: add nftables")
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin