linux-stable.git/net/xfrm, branch linux-5.7.y Linux kernel stable tree espintcp: handle short messages instead of breaking the encap socket 2020-08-05T07:58:45+00:00 Sabrina Dubroca sd@queasysnail.net 2020-07-29T16:38:42+00:00 22e142108d34c5b802c37c7826799df41a11dafa [ Upstream commit fadd1a63a7b4df295a01fa50b2f4e447542bee59 ] Currently, short messages (less than 4 bytes after the length header) will break the stream of messages. This is unnecessary, since we can still parse messages even if they're too short to contain any usable data. This is also bogus, as keepalive messages (a single 0xff byte), though not needed with TCP encapsulation, should be allowed. This patch changes the stream parser so that short messages are accepted and dropped in the kernel. Messages that contain a valid SPI or non-ESP header are processed as before. Fixes: e27cca96cd68 ("xfrm: add espintcp (RFC 8229)") Reported-by: Andrew Cagney <cagney@libreswan.org> Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit fadd1a63a7b4df295a01fa50b2f4e447542bee59 ]

Currently, short messages (less than 4 bytes after the length header)
will break the stream of messages. This is unnecessary, since we can
still parse messages even if they're too short to contain any usable
data. This is also bogus, as keepalive messages (a single 0xff byte),
though not needed with TCP encapsulation, should be allowed.

This patch changes the stream parser so that short messages are
accepted and dropped in the kernel. Messages that contain a valid SPI
or non-ESP header are processed as before.

Fixes: e27cca96cd68 ("xfrm: add espintcp (RFC 8229)")
Reported-by: Andrew Cagney <cagney@libreswan.org>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
espintcp: recv() should return 0 when the peer socket is closed 2020-08-05T07:58:39+00:00 Sabrina Dubroca sd@queasysnail.net 2020-07-16T08:09:02+00:00 589b5e620160cd5007136f863cedc20b24877f99 [ Upstream commit e229c877cde141a4c46cb603a341ce8c909e9a98 ] man 2 recv says: RETURN VALUE When a stream socket peer has performed an orderly shutdown, the return value will be 0 (the traditional "end-of-file" return). Currently, this works for blocking reads, but non-blocking reads will return -EAGAIN. This patch overwrites that return value when the peer won't send us any more data. Fixes: e27cca96cd68 ("xfrm: add espintcp (RFC 8229)") Reported-by: Andrew Cagney <cagney@libreswan.org> Tested-by: Andrew Cagney <cagney@libreswan.org> Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit e229c877cde141a4c46cb603a341ce8c909e9a98 ]

man 2 recv says:

    RETURN VALUE

    When a stream socket peer has performed an orderly shutdown, the
    return value will be 0 (the traditional "end-of-file" return).

Currently, this works for blocking reads, but non-blocking reads will
return -EAGAIN. This patch overwrites that return value when the peer
won't send us any more data.

Fixes: e27cca96cd68 ("xfrm: add espintcp (RFC 8229)")
Reported-by: Andrew Cagney <cagney@libreswan.org>
Tested-by: Andrew Cagney <cagney@libreswan.org>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
xfrm: policy: match with both mark and mask on user interfaces 2020-08-05T07:58:39+00:00 Xin Long lucien.xin@gmail.com 2020-06-22T08:40:29+00:00 7388c5d12dc5bd7efb2719878be4b0962b15282a [ Upstream commit 4f47e8ab6ab796b5380f74866fa5287aca4dcc58 ] In commit ed17b8d377ea ("xfrm: fix a warning in xfrm_policy_insert_list"), it would take 'priority' to make a policy unique, and allow duplicated policies with different 'priority' to be added, which is not expected by userland, as Tobias reported in strongswan. To fix this duplicated policies issue, and also fix the issue in commit ed17b8d377ea ("xfrm: fix a warning in xfrm_policy_insert_list"), when doing add/del/get/update on user interfaces, this patch is to change to look up a policy with both mark and mask by doing: mark.v == pol->mark.v && mark.m == pol->mark.m and leave the check: (mark & pol->mark.m) == pol->mark.v for tx/rx path only. As the userland expects an exact mark and mask match to manage policies. v1->v2: - make xfrm_policy_mark_match inline and fix the changelog as Tobias suggested. Fixes: 295fae568885 ("xfrm: Allow user space manipulation of SPD mark") Fixes: ed17b8d377ea ("xfrm: fix a warning in xfrm_policy_insert_list") Reported-by: Tobias Brunner <tobias@strongswan.org> Tested-by: Tobias Brunner <tobias@strongswan.org> Signed-off-by: Xin Long <lucien.xin@gmail.com> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 4f47e8ab6ab796b5380f74866fa5287aca4dcc58 ]

In commit ed17b8d377ea ("xfrm: fix a warning in xfrm_policy_insert_list"),
it would take 'priority' to make a policy unique, and allow duplicated
policies with different 'priority' to be added, which is not expected
by userland, as Tobias reported in strongswan.

To fix this duplicated policies issue, and also fix the issue in
commit ed17b8d377ea ("xfrm: fix a warning in xfrm_policy_insert_list"),
when doing add/del/get/update on user interfaces, this patch is to change
to look up a policy with both mark and mask by doing:

  mark.v == pol->mark.v && mark.m == pol->mark.m

and leave the check:

  (mark & pol->mark.m) == pol->mark.v

for tx/rx path only.

As the userland expects an exact mark and mask match to manage policies.

v1->v2:
  - make xfrm_policy_mark_match inline and fix the changelog as
    Tobias suggested.

Fixes: 295fae568885 ("xfrm: Allow user space manipulation of SPD mark")
Fixes: ed17b8d377ea ("xfrm: fix a warning in xfrm_policy_insert_list")
Reported-by: Tobias Brunner <tobias@strongswan.org>
Tested-by: Tobias Brunner <tobias@strongswan.org>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
xfrm: Fix double ESP trailer insertion in IPsec crypto offload. 2020-06-30T19:36:00+00:00 Huy Nguyen huyn@mellanox.com 2020-06-01T21:39:37+00:00 f1f811496c48900a0880c7cd49fedb04b844160d [ Upstream commit 94579ac3f6d0820adc83b5dc5358ead0158101e9 ] During IPsec performance testing, we see bad ICMP checksum. The error packet has duplicated ESP trailer due to double validate_xmit_xfrm calls. The first call is from ip_output, but the packet cannot be sent because netif_xmit_frozen_or_stopped is true and the packet gets dev_requeue_skb. The second call is from NET_TX softirq. However after the first call, the packet already has the ESP trailer. Fix by marking the skb with XFRM_XMIT bit after the packet is handled by validate_xmit_xfrm to avoid duplicate ESP trailer insertion. Fixes: f6e27114a60a ("net: Add a xfrm validate function to validate_xmit_skb") Signed-off-by: Huy Nguyen <huyn@mellanox.com> Reviewed-by: Boris Pismenny <borisp@mellanox.com> Reviewed-by: Raed Salem <raeds@mellanox.com> Reviewed-by: Saeed Mahameed <saeedm@mellanox.com> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 94579ac3f6d0820adc83b5dc5358ead0158101e9 ]

During IPsec performance testing, we see bad ICMP checksum. The error packet
has duplicated ESP trailer due to double validate_xmit_xfrm calls. The first call
is from ip_output, but the packet cannot be sent because
netif_xmit_frozen_or_stopped is true and the packet gets dev_requeue_skb. The second
call is from NET_TX softirq. However after the first call, the packet already
has the ESP trailer.

Fix by marking the skb with XFRM_XMIT bit after the packet is handled by
validate_xmit_xfrm to avoid duplicate ESP trailer insertion.

Fixes: f6e27114a60a ("net: Add a xfrm validate function to validate_xmit_skb")
Signed-off-by: Huy Nguyen <huyn@mellanox.com>
Reviewed-by: Boris Pismenny <borisp@mellanox.com>
Reviewed-by: Raed Salem <raeds@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
xfrm: fix a NULL-ptr deref in xfrm_local_error 2020-05-29T10:10:22+00:00 Xin Long lucien.xin@gmail.com 2020-05-26T09:41:46+00:00 f6a23d85d078c2ffde79c66ca81d0a1dde451649 This patch is to fix a crash: [ ] kasan: GPF could be caused by NULL-ptr deref or user memory access [ ] general protection fault: 0000 [#1] SMP KASAN PTI [ ] RIP: 0010:ipv6_local_error+0xac/0x7a0 [ ] Call Trace: [ ] xfrm6_local_error+0x1eb/0x300 [ ] xfrm_local_error+0x95/0x130 [ ] __xfrm6_output+0x65f/0xb50 [ ] xfrm6_output+0x106/0x46f [ ] udp_tunnel6_xmit_skb+0x618/0xbf0 [ip6_udp_tunnel] [ ] vxlan_xmit_one+0xbc6/0x2c60 [vxlan] [ ] vxlan_xmit+0x6a0/0x4276 [vxlan] [ ] dev_hard_start_xmit+0x165/0x820 [ ] __dev_queue_xmit+0x1ff0/0x2b90 [ ] ip_finish_output2+0xd3e/0x1480 [ ] ip_do_fragment+0x182d/0x2210 [ ] ip_output+0x1d0/0x510 [ ] ip_send_skb+0x37/0xa0 [ ] raw_sendmsg+0x1b4c/0x2b80 [ ] sock_sendmsg+0xc0/0x110 This occurred when sending a v4 skb over vxlan6 over ipsec, in which case skb->protocol == htons(ETH_P_IPV6) while skb->sk->sk_family == AF_INET in xfrm_local_error(). Then it will go to xfrm6_local_error() where it tries to get ipv6 info from a ipv4 sk. This issue was actually fixed by Commit 628e341f319f ("xfrm: make local error reporting more robust"), but brought back by Commit 844d48746e4b ("xfrm: choose protocol family by skb protocol"). So to fix it, we should call xfrm6_local_error() only when skb->protocol is htons(ETH_P_IPV6) and skb->sk->sk_family is AF_INET6. Fixes: 844d48746e4b ("xfrm: choose protocol family by skb protocol") Reported-by: Xiumei Mu <xmu@redhat.com> Signed-off-by: Xin Long <lucien.xin@gmail.com> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> 
This patch is to fix a crash:

  [ ] kasan: GPF could be caused by NULL-ptr deref or user memory access
  [ ] general protection fault: 0000 [#1] SMP KASAN PTI
  [ ] RIP: 0010:ipv6_local_error+0xac/0x7a0
  [ ] Call Trace:
  [ ]  xfrm6_local_error+0x1eb/0x300
  [ ]  xfrm_local_error+0x95/0x130
  [ ]  __xfrm6_output+0x65f/0xb50
  [ ]  xfrm6_output+0x106/0x46f
  [ ]  udp_tunnel6_xmit_skb+0x618/0xbf0 [ip6_udp_tunnel]
  [ ]  vxlan_xmit_one+0xbc6/0x2c60 [vxlan]
  [ ]  vxlan_xmit+0x6a0/0x4276 [vxlan]
  [ ]  dev_hard_start_xmit+0x165/0x820
  [ ]  __dev_queue_xmit+0x1ff0/0x2b90
  [ ]  ip_finish_output2+0xd3e/0x1480
  [ ]  ip_do_fragment+0x182d/0x2210
  [ ]  ip_output+0x1d0/0x510
  [ ]  ip_send_skb+0x37/0xa0
  [ ]  raw_sendmsg+0x1b4c/0x2b80
  [ ]  sock_sendmsg+0xc0/0x110

This occurred when sending a v4 skb over vxlan6 over ipsec, in which case
skb->protocol == htons(ETH_P_IPV6) while skb->sk->sk_family == AF_INET in
xfrm_local_error(). Then it will go to xfrm6_local_error() where it tries
to get ipv6 info from a ipv4 sk.

This issue was actually fixed by Commit 628e341f319f ("xfrm: make local
error reporting more robust"), but brought back by Commit 844d48746e4b
("xfrm: choose protocol family by skb protocol").

So to fix it, we should call xfrm6_local_error() only when skb->protocol
is htons(ETH_P_IPV6) and skb->sk->sk_family is AF_INET6.

Fixes: 844d48746e4b ("xfrm: choose protocol family by skb protocol")
Reported-by: Xiumei Mu <xmu@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
xfrm: fix a warning in xfrm_policy_insert_list 2020-05-25T14:04:12+00:00 Xin Long lucien.xin@gmail.com 2020-05-25T05:53:37+00:00 ed17b8d377eaf6b4a01d46942b4c647378a79bdd This waring can be triggered simply by: # ip xfrm policy update src 192.168.1.1/24 dst 192.168.1.2/24 dir in \ priority 1 mark 0 mask 0x10 #[1] # ip xfrm policy update src 192.168.1.1/24 dst 192.168.1.2/24 dir in \ priority 2 mark 0 mask 0x1 #[2] # ip xfrm policy update src 192.168.1.1/24 dst 192.168.1.2/24 dir in \ priority 2 mark 0 mask 0x10 #[3] Then dmesg shows: [ ] WARNING: CPU: 1 PID: 7265 at net/xfrm/xfrm_policy.c:1548 [ ] RIP: 0010:xfrm_policy_insert_list+0x2f2/0x1030 [ ] Call Trace: [ ] xfrm_policy_inexact_insert+0x85/0xe50 [ ] xfrm_policy_insert+0x4ba/0x680 [ ] xfrm_add_policy+0x246/0x4d0 [ ] xfrm_user_rcv_msg+0x331/0x5c0 [ ] netlink_rcv_skb+0x121/0x350 [ ] xfrm_netlink_rcv+0x66/0x80 [ ] netlink_unicast+0x439/0x630 [ ] netlink_sendmsg+0x714/0xbf0 [ ] sock_sendmsg+0xe2/0x110 The issue was introduced by Commit 7cb8a93968e3 ("xfrm: Allow inserting policies with matching mark and different priorities"). After that, the policies [1] and [2] would be able to be added with different priorities. However, policy [3] will actually match both [1] and [2]. Policy [1] was matched due to the 1st 'return true' in xfrm_policy_mark_match(), and policy [2] was matched due to the 2nd 'return true' in there. It caused WARN_ON() in xfrm_policy_insert_list(). This patch is to fix it by only (the same value and priority) as the same policy in xfrm_policy_mark_match(). Thanks to Yuehaibing, we could make this fix better. v1->v2: - check policy->mark.v == pol->mark.v only without mask. Fixes: 7cb8a93968e3 ("xfrm: Allow inserting policies with matching mark and different priorities") Reported-by: Xiumei Mu <xmu@redhat.com> Signed-off-by: Xin Long <lucien.xin@gmail.com> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> 
This waring can be triggered simply by:

  # ip xfrm policy update src 192.168.1.1/24 dst 192.168.1.2/24 dir in \
    priority 1 mark 0 mask 0x10  #[1]
  # ip xfrm policy update src 192.168.1.1/24 dst 192.168.1.2/24 dir in \
    priority 2 mark 0 mask 0x1   #[2]
  # ip xfrm policy update src 192.168.1.1/24 dst 192.168.1.2/24 dir in \
    priority 2 mark 0 mask 0x10  #[3]

Then dmesg shows:

  [ ] WARNING: CPU: 1 PID: 7265 at net/xfrm/xfrm_policy.c:1548
  [ ] RIP: 0010:xfrm_policy_insert_list+0x2f2/0x1030
  [ ] Call Trace:
  [ ]  xfrm_policy_inexact_insert+0x85/0xe50
  [ ]  xfrm_policy_insert+0x4ba/0x680
  [ ]  xfrm_add_policy+0x246/0x4d0
  [ ]  xfrm_user_rcv_msg+0x331/0x5c0
  [ ]  netlink_rcv_skb+0x121/0x350
  [ ]  xfrm_netlink_rcv+0x66/0x80
  [ ]  netlink_unicast+0x439/0x630
  [ ]  netlink_sendmsg+0x714/0xbf0
  [ ]  sock_sendmsg+0xe2/0x110

The issue was introduced by Commit 7cb8a93968e3 ("xfrm: Allow inserting
policies with matching mark and different priorities"). After that, the
policies [1] and [2] would be able to be added with different priorities.

However, policy [3] will actually match both [1] and [2]. Policy [1]
was matched due to the 1st 'return true' in xfrm_policy_mark_match(),
and policy [2] was matched due to the 2nd 'return true' in there. It
caused WARN_ON() in xfrm_policy_insert_list().

This patch is to fix it by only (the same value and priority) as the
same policy in xfrm_policy_mark_match().

Thanks to Yuehaibing, we could make this fix better.

v1->v2:
  - check policy->mark.v == pol->mark.v only without mask.

Fixes: 7cb8a93968e3 ("xfrm: Allow inserting policies with matching mark and different priorities")
Reported-by: Xiumei Mu <xmu@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
xfrm interface: fix oops when deleting a x-netns interface 2020-04-23T06:09:39+00:00 Nicolas Dichtel nicolas.dichtel@6wind.com 2020-04-22T22:06:45+00:00 c95c5f58b35ef995f66cb55547eee6093ab5fcb8 Here is the steps to reproduce the problem: ip netns add foo ip netns add bar ip -n foo link add xfrmi0 type xfrm dev lo if_id 42 ip -n foo link set xfrmi0 netns bar ip netns del foo ip netns del bar Which results to: [ 186.686395] general protection fault, probably for non-canonical address 0x6b6b6b6b6b6b6bd3: 0000 [#1] SMP PTI [ 186.687665] CPU: 7 PID: 232 Comm: kworker/u16:2 Not tainted 5.6.0+ #1 [ 186.688430] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-1 04/01/2014 [ 186.689420] Workqueue: netns cleanup_net [ 186.689903] RIP: 0010:xfrmi_dev_uninit+0x1b/0x4b [xfrm_interface] [ 186.690657] Code: 44 f6 ff ff 31 c0 5b 5d 41 5c 41 5d 41 5e c3 48 8d 8f c0 08 00 00 8b 05 ce 14 00 00 48 8b 97 d0 08 00 00 48 8b 92 c0 0e 00 00 <48> 8b 14 c2 48 8b 02 48 85 c0 74 19 48 39 c1 75 0c 48 8b 87 c0 08 [ 186.692838] RSP: 0018:ffffc900003b7d68 EFLAGS: 00010286 [ 186.693435] RAX: 000000000000000d RBX: ffff8881b0f31000 RCX: ffff8881b0f318c0 [ 186.694334] RDX: 6b6b6b6b6b6b6b6b RSI: 0000000000000246 RDI: ffff8881b0f31000 [ 186.695190] RBP: ffffc900003b7df0 R08: ffff888236c07740 R09: 0000000000000040 [ 186.696024] R10: ffffffff81fce1b8 R11: 0000000000000002 R12: ffffc900003b7d80 [ 186.696859] R13: ffff8881edcc6a40 R14: ffff8881a1b6e780 R15: ffffffff81ed47c8 [ 186.697738] FS: 0000000000000000(0000) GS:ffff888237dc0000(0000) knlGS:0000000000000000 [ 186.698705] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 186.699408] CR2: 00007f2129e93148 CR3: 0000000001e0a000 CR4: 00000000000006e0 [ 186.700221] Call Trace: [ 186.700508] rollback_registered_many+0x32b/0x3fd [ 186.701058] ? __rtnl_unlock+0x20/0x3d [ 186.701494] ? arch_local_irq_save+0x11/0x17 [ 186.702012] unregister_netdevice_many+0x12/0x55 [ 186.702594] default_device_exit_batch+0x12b/0x150 [ 186.703160] ? prepare_to_wait_exclusive+0x60/0x60 [ 186.703719] cleanup_net+0x17d/0x234 [ 186.704138] process_one_work+0x196/0x2e8 [ 186.704652] worker_thread+0x1a4/0x249 [ 186.705087] ? cancel_delayed_work+0x92/0x92 [ 186.705620] kthread+0x105/0x10f [ 186.706000] ? __kthread_bind_mask+0x57/0x57 [ 186.706501] ret_from_fork+0x35/0x40 [ 186.706978] Modules linked in: xfrm_interface nfsv3 nfs_acl auth_rpcgss nfsv4 nfs lockd grace fscache sunrpc button parport_pc parport serio_raw evdev pcspkr loop ext4 crc16 mbcache jbd2 crc32c_generic 8139too ide_cd_mod cdrom ide_gd_mod ata_generic ata_piix libata scsi_mod piix psmouse i2c_piix4 ide_core 8139cp i2c_core mii floppy [ 186.710423] ---[ end trace 463bba18105537e5 ]--- The problem is that x-netns xfrm interface are not removed when the link netns is removed. This causes later this oops when thoses interfaces are removed. Let's add a handler to remove all interfaces related to a netns when this netns is removed. Fixes: f203b76d7809 ("xfrm: Add virtual xfrm interfaces") Reported-by: Christophe Gouault <christophe.gouault@6wind.com> Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> 
Here is the steps to reproduce the problem:
ip netns add foo
ip netns add bar
ip -n foo link add xfrmi0 type xfrm dev lo if_id 42
ip -n foo link set xfrmi0 netns bar
ip netns del foo
ip netns del bar

Which results to:
[  186.686395] general protection fault, probably for non-canonical address 0x6b6b6b6b6b6b6bd3: 0000 [#1] SMP PTI
[  186.687665] CPU: 7 PID: 232 Comm: kworker/u16:2 Not tainted 5.6.0+ #1
[  186.688430] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-1 04/01/2014
[  186.689420] Workqueue: netns cleanup_net
[  186.689903] RIP: 0010:xfrmi_dev_uninit+0x1b/0x4b [xfrm_interface]
[  186.690657] Code: 44 f6 ff ff 31 c0 5b 5d 41 5c 41 5d 41 5e c3 48 8d 8f c0 08 00 00 8b 05 ce 14 00 00 48 8b 97 d0 08 00 00 48 8b 92 c0 0e 00 00 <48> 8b 14 c2 48 8b 02 48 85 c0 74 19 48 39 c1 75 0c 48 8b 87 c0 08
[  186.692838] RSP: 0018:ffffc900003b7d68 EFLAGS: 00010286
[  186.693435] RAX: 000000000000000d RBX: ffff8881b0f31000 RCX: ffff8881b0f318c0
[  186.694334] RDX: 6b6b6b6b6b6b6b6b RSI: 0000000000000246 RDI: ffff8881b0f31000
[  186.695190] RBP: ffffc900003b7df0 R08: ffff888236c07740 R09: 0000000000000040
[  186.696024] R10: ffffffff81fce1b8 R11: 0000000000000002 R12: ffffc900003b7d80
[  186.696859] R13: ffff8881edcc6a40 R14: ffff8881a1b6e780 R15: ffffffff81ed47c8
[  186.697738] FS:  0000000000000000(0000) GS:ffff888237dc0000(0000) knlGS:0000000000000000
[  186.698705] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  186.699408] CR2: 00007f2129e93148 CR3: 0000000001e0a000 CR4: 00000000000006e0
[  186.700221] Call Trace:
[  186.700508]  rollback_registered_many+0x32b/0x3fd
[  186.701058]  ? __rtnl_unlock+0x20/0x3d
[  186.701494]  ? arch_local_irq_save+0x11/0x17
[  186.702012]  unregister_netdevice_many+0x12/0x55
[  186.702594]  default_device_exit_batch+0x12b/0x150
[  186.703160]  ? prepare_to_wait_exclusive+0x60/0x60
[  186.703719]  cleanup_net+0x17d/0x234
[  186.704138]  process_one_work+0x196/0x2e8
[  186.704652]  worker_thread+0x1a4/0x249
[  186.705087]  ? cancel_delayed_work+0x92/0x92
[  186.705620]  kthread+0x105/0x10f
[  186.706000]  ? __kthread_bind_mask+0x57/0x57
[  186.706501]  ret_from_fork+0x35/0x40
[  186.706978] Modules linked in: xfrm_interface nfsv3 nfs_acl auth_rpcgss nfsv4 nfs lockd grace fscache sunrpc button parport_pc parport serio_raw evdev pcspkr loop ext4 crc16 mbcache jbd2 crc32c_generic 8139too ide_cd_mod cdrom ide_gd_mod ata_generic ata_piix libata scsi_mod piix psmouse i2c_piix4 ide_core 8139cp i2c_core mii floppy
[  186.710423] ---[ end trace 463bba18105537e5 ]---

The problem is that x-netns xfrm interface are not removed when the link
netns is removed. This causes later this oops when thoses interfaces are
removed.

Let's add a handler to remove all interfaces related to a netns when this
netns is removed.

Fixes: f203b76d7809 ("xfrm: Add virtual xfrm interfaces")
Reported-by: Christophe Gouault <christophe.gouault@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
xfrm: call xfrm_output_gso when inner_protocol is set in xfrm_output 2020-04-21T05:30:06+00:00 Xin Long lucien.xin@gmail.com 2020-04-20T13:51:09+00:00 a204aef9fd77dce1efd9066ca4e44eede99cd858 An use-after-free crash can be triggered when sending big packets over vxlan over esp with esp offload enabled: [] BUG: KASAN: use-after-free in ipv6_gso_pull_exthdrs.part.8+0x32c/0x4e0 [] Call Trace: [] dump_stack+0x75/0xa0 [] kasan_report+0x37/0x50 [] ipv6_gso_pull_exthdrs.part.8+0x32c/0x4e0 [] ipv6_gso_segment+0x2c8/0x13c0 [] skb_mac_gso_segment+0x1cb/0x420 [] skb_udp_tunnel_segment+0x6b5/0x1c90 [] inet_gso_segment+0x440/0x1380 [] skb_mac_gso_segment+0x1cb/0x420 [] esp4_gso_segment+0xae8/0x1709 [esp4_offload] [] inet_gso_segment+0x440/0x1380 [] skb_mac_gso_segment+0x1cb/0x420 [] __skb_gso_segment+0x2d7/0x5f0 [] validate_xmit_skb+0x527/0xb10 [] __dev_queue_xmit+0x10f8/0x2320 <--- [] ip_finish_output2+0xa2e/0x1b50 [] ip_output+0x1a8/0x2f0 [] xfrm_output_resume+0x110e/0x15f0 [] __xfrm4_output+0xe1/0x1b0 [] xfrm4_output+0xa0/0x200 [] iptunnel_xmit+0x5a7/0x920 [] vxlan_xmit_one+0x1658/0x37a0 [vxlan] [] vxlan_xmit+0x5e4/0x3ec8 [vxlan] [] dev_hard_start_xmit+0x125/0x540 [] __dev_queue_xmit+0x17bd/0x2320 <--- [] ip6_finish_output2+0xb20/0x1b80 [] ip6_output+0x1b3/0x390 [] ip6_xmit+0xb82/0x17e0 [] inet6_csk_xmit+0x225/0x3d0 [] __tcp_transmit_skb+0x1763/0x3520 [] tcp_write_xmit+0xd64/0x5fe0 [] __tcp_push_pending_frames+0x8c/0x320 [] tcp_sendmsg_locked+0x2245/0x3500 [] tcp_sendmsg+0x27/0x40 As on the tx path of vxlan over esp, skb->inner_network_header would be set on vxlan_xmit() and xfrm4_tunnel_encap_add(), and the later one can overwrite the former one. It causes skb_udp_tunnel_segment() to use a wrong skb->inner_network_header, then the issue occurs. This patch is to fix it by calling xfrm_output_gso() instead when the inner_protocol is set, in which gso_segment of inner_protocol will be done first. While at it, also improve some code around. Fixes: 7862b4058b9f ("esp: Add gso handlers for esp4 and esp6") Reported-by: Xiumei Mu <xmu@redhat.com> Signed-off-by: Xin Long <lucien.xin@gmail.com> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> 
An use-after-free crash can be triggered when sending big packets over
vxlan over esp with esp offload enabled:

  [] BUG: KASAN: use-after-free in ipv6_gso_pull_exthdrs.part.8+0x32c/0x4e0
  [] Call Trace:
  []  dump_stack+0x75/0xa0
  []  kasan_report+0x37/0x50
  []  ipv6_gso_pull_exthdrs.part.8+0x32c/0x4e0
  []  ipv6_gso_segment+0x2c8/0x13c0
  []  skb_mac_gso_segment+0x1cb/0x420
  []  skb_udp_tunnel_segment+0x6b5/0x1c90
  []  inet_gso_segment+0x440/0x1380
  []  skb_mac_gso_segment+0x1cb/0x420
  []  esp4_gso_segment+0xae8/0x1709 [esp4_offload]
  []  inet_gso_segment+0x440/0x1380
  []  skb_mac_gso_segment+0x1cb/0x420
  []  __skb_gso_segment+0x2d7/0x5f0
  []  validate_xmit_skb+0x527/0xb10
  []  __dev_queue_xmit+0x10f8/0x2320 <---
  []  ip_finish_output2+0xa2e/0x1b50
  []  ip_output+0x1a8/0x2f0
  []  xfrm_output_resume+0x110e/0x15f0
  []  __xfrm4_output+0xe1/0x1b0
  []  xfrm4_output+0xa0/0x200
  []  iptunnel_xmit+0x5a7/0x920
  []  vxlan_xmit_one+0x1658/0x37a0 [vxlan]
  []  vxlan_xmit+0x5e4/0x3ec8 [vxlan]
  []  dev_hard_start_xmit+0x125/0x540
  []  __dev_queue_xmit+0x17bd/0x2320  <---
  []  ip6_finish_output2+0xb20/0x1b80
  []  ip6_output+0x1b3/0x390
  []  ip6_xmit+0xb82/0x17e0
  []  inet6_csk_xmit+0x225/0x3d0
  []  __tcp_transmit_skb+0x1763/0x3520
  []  tcp_write_xmit+0xd64/0x5fe0
  []  __tcp_push_pending_frames+0x8c/0x320
  []  tcp_sendmsg_locked+0x2245/0x3500
  []  tcp_sendmsg+0x27/0x40

As on the tx path of vxlan over esp, skb->inner_network_header would be
set on vxlan_xmit() and xfrm4_tunnel_encap_add(), and the later one can
overwrite the former one. It causes skb_udp_tunnel_segment() to use a
wrong skb->inner_network_header, then the issue occurs.

This patch is to fix it by calling xfrm_output_gso() instead when the
inner_protocol is set, in which gso_segment of inner_protocol will be
done first.

While at it, also improve some code around.

Fixes: 7862b4058b9f ("esp: Add gso handlers for esp4 and esp6")
Reported-by: Xiumei Mu <xmu@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
xfrm: espintcp: save and call old ->sk_destruct 2020-04-20T05:34:16+00:00 Sabrina Dubroca sd@queasysnail.net 2020-04-16T15:45:44+00:00 9f0cadc32d738f0f0c8e30be83be7087c7b85ee5 When ESP encapsulation is enabled on a TCP socket, I'm replacing the existing ->sk_destruct callback with espintcp_destruct. We still need to call the old callback to perform the other cleanups when the socket is destroyed. Save the old callback, and call it from espintcp_destruct. Fixes: e27cca96cd68 ("xfrm: add espintcp (RFC 8229)") Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> 
When ESP encapsulation is enabled on a TCP socket, I'm replacing the
existing ->sk_destruct callback with espintcp_destruct. We still need to
call the old callback to perform the other cleanups when the socket is
destroyed. Save the old callback, and call it from espintcp_destruct.

Fixes: e27cca96cd68 ("xfrm: add espintcp (RFC 8229)")
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
xfrm: do pskb_pull properly in __xfrm_transport_prep 2020-04-15T07:50:03+00:00 Xin Long lucien.xin@gmail.com 2020-04-10T09:06:31+00:00 06a0afcfe2f551ff755849ea2549b0d8409fd9a0 For transport mode, when ipv6 nexthdr is set, the packet format might be like: ---------------------------------------------------- | | dest | | | | ESP | ESP | | IP6 hdr| opts.| ESP | TCP | Data | Trailer | ICV | ---------------------------------------------------- and in __xfrm_transport_prep(): pskb_pull(skb, skb->mac_len + sizeof(ip6hdr) + x->props.header_len); it will pull the data pointer to the wrong position, as it missed the nexthdrs/dest opts. This patch is to fix it by using: pskb_pull(skb, skb_transport_offset(skb) + x->props.header_len); as we can be sure transport_header points to ESP header at that moment. It also fixes a panic when packets with ipv6 nexthdr are sent over esp6 transport mode: [ 100.473845] kernel BUG at net/core/skbuff.c:4325! [ 100.478517] RIP: 0010:__skb_to_sgvec+0x252/0x260 [ 100.494355] Call Trace: [ 100.494829] skb_to_sgvec+0x11/0x40 [ 100.495492] esp6_output_tail+0x12e/0x550 [esp6] [ 100.496358] esp6_xmit+0x1d5/0x260 [esp6_offload] [ 100.498029] validate_xmit_xfrm+0x22f/0x2e0 [ 100.499604] __dev_queue_xmit+0x589/0x910 [ 100.502928] ip6_finish_output2+0x2a5/0x5a0 [ 100.503718] ip6_output+0x6c/0x120 [ 100.505198] xfrm_output_resume+0x4bf/0x530 [ 100.508683] xfrm6_output+0x3a/0xc0 [ 100.513446] inet6_csk_xmit+0xa1/0xf0 [ 100.517335] tcp_sendmsg+0x27/0x40 [ 100.517977] sock_sendmsg+0x3e/0x60 [ 100.518648] __sys_sendto+0xee/0x160 Fixes: c35fe4106b92 ("xfrm: Add mode handlers for IPsec on layer 2") Signed-off-by: Xin Long <lucien.xin@gmail.com> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> 
For transport mode, when ipv6 nexthdr is set, the packet format might
be like:

    ----------------------------------------------------
    |        | dest |     |     |      |  ESP    | ESP |
    | IP6 hdr| opts.| ESP | TCP | Data | Trailer | ICV |
    ----------------------------------------------------

and in __xfrm_transport_prep():

  pskb_pull(skb, skb->mac_len + sizeof(ip6hdr) + x->props.header_len);

it will pull the data pointer to the wrong position, as it missed the
nexthdrs/dest opts.

This patch is to fix it by using:

  pskb_pull(skb, skb_transport_offset(skb) + x->props.header_len);

as we can be sure transport_header points to ESP header at that moment.

It also fixes a panic when packets with ipv6 nexthdr are sent over
esp6 transport mode:

  [  100.473845] kernel BUG at net/core/skbuff.c:4325!
  [  100.478517] RIP: 0010:__skb_to_sgvec+0x252/0x260
  [  100.494355] Call Trace:
  [  100.494829]  skb_to_sgvec+0x11/0x40
  [  100.495492]  esp6_output_tail+0x12e/0x550 [esp6]
  [  100.496358]  esp6_xmit+0x1d5/0x260 [esp6_offload]
  [  100.498029]  validate_xmit_xfrm+0x22f/0x2e0
  [  100.499604]  __dev_queue_xmit+0x589/0x910
  [  100.502928]  ip6_finish_output2+0x2a5/0x5a0
  [  100.503718]  ip6_output+0x6c/0x120
  [  100.505198]  xfrm_output_resume+0x4bf/0x530
  [  100.508683]  xfrm6_output+0x3a/0xc0
  [  100.513446]  inet6_csk_xmit+0xa1/0xf0
  [  100.517335]  tcp_sendmsg+0x27/0x40
  [  100.517977]  sock_sendmsg+0x3e/0x60
  [  100.518648]  __sys_sendto+0xee/0x160

Fixes: c35fe4106b92 ("xfrm: Add mode handlers for IPsec on layer 2")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>