| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
In order to unload a device driver and reload it, xprtrdma will need
to close a transport's interface adapter, and then call
rpcrdma_ia_open again, possibly finding a different interface
adapter.
Make rpcrdma_ia_open safe to call on the same transport multiple
times.
This is a refactoring change only.
Signed-off-by: Chuck Lever <[email protected]>
Signed-off-by: Anna Schumaker <[email protected]>
|
|
Current NFS clients rely on connection loss to determine when to
retransmit. In particular, for protocols like NFSv4, clients no
longer rely on RPC timeouts to drive retransmission: NFSv4 servers
are required to terminate a connection when they need a client to
retransmit pending RPCs.
When a server is no longer reachable, either because it has crashed
or because the network path has broken, the server cannot actively
terminate a connection. Thus NFS clients depend on transport-level
keepalive to determine when a connection must be replaced and
pending RPCs retransmitted.
However, RDMA RC connections do not have a native keepalive
mechanism. If an NFS/RDMA server crashes after a client has sent
RPCs successfully (an RC ACK has been received for all OTW RDMA
requests), there is no way for the client to know the connection is
moribund.
In addition, new RDMA requests are subject to the RPC-over-RDMA
credit limit. If the client has consumed all granted credits with
NFS traffic, it is not allowed to send another RDMA request until
the server replies. Thus it has no way to send a true keepalive when
the workload has already consumed all credits with pending RPCs.
To address this, forcibly disconnect a transport when an RPC times
out. This prevents moribund connections from stopping the
detection of failover or other configuration changes on the server.
Note that even if the connection is still good, retransmitting
any RPC will trigger a disconnect thanks to this logic in
xprt_rdma_send_request:
/* Must suppress retransmit to maintain credits */
if (req->rl_connect_cookie == xprt->connect_cookie)
goto drop_connection;
req->rl_connect_cookie = xprt->connect_cookie;
Signed-off-by: Chuck Lever <[email protected]>
Signed-off-by: Anna Schumaker <[email protected]>
|
|
xprt_force_disconnect() is already invoked from the socket
transport. I want to invoke xprt_force_disconnect() from the
RPC-over-RDMA transport, which is a separate module from sunrpc.ko.
Signed-off-by: Chuck Lever <[email protected]>
Signed-off-by: Anna Schumaker <[email protected]>
|
|
Trying to create MRs while the transport is being torn down can
cause a crash.
Fixes: e2ac236c0b65 ("xprtrdma: Allocate MRs on demand")
Signed-off-by: Chuck Lever <[email protected]>
Signed-off-by: Anna Schumaker <[email protected]>
|
|
The monitor mode delivery logic makes it hard to add any
kind of filtering in an efficient way, because the monitor
SKB is created first and then passed to all interfaces.
Rewrite the logic to create the monitor SKB the first time
it's actually needed, and then keep delivering it.
Signed-off-by: Johannes Berg <[email protected]>
|
|
When part of a bigger bandwidth (160 MHz) channel falls in DFS
channel range it is possible that the center frequency may not
necessarily be a radar channel. Remove the sanity check on channel
flag for IEEE80211_CHAN_RADAR in regulatory_propagate_dfs_state(),
this should fix the dfs state propagation for non-DFS center freq
which has DFS channels in it's bandwidth, should also fix unnecessary
WARN_ON() spam in regulatory_propagate_dfs_state().
Fixes: 8976672736d6 ("cfg80211: Share Channel DFS state across wiphys of same DFS domain")
Reported-by: Johannes Berg <[email protected]>
Signed-off-by: Vasanthakumar Thiagarajan <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
|
|
In the case getsockopt() is called with PACKET_HDRLEN and optlen < 4
|val| remains uninitialized and the syscall may behave differently
depending on its value, and even copy garbage to userspace on certain
architectures. To fix this we now return -EINVAL if optlen is too small.
This bug has been detected with KMSAN.
Signed-off-by: Alexander Potapenko <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Taking down the loopback device wreaks havoc on IPv6 routing. By
extension, taking down a VRF device wreaks havoc on its table.
Dmitry and Andrey both reported heap out-of-bounds reports in the IPv6
FIB code while running syzkaller fuzzer. The root cause is a dead dst
that is on the garbage list gets reinserted into the IPv6 FIB. While on
the gc (or perhaps when it gets added to the gc list) the dst->next is
set to an IPv4 dst. A subsequent walk of the ipv6 tables causes the
out-of-bounds access.
Andrey's reproducer was the key to getting to the bottom of this.
With IPv6, host routes for an address have the dst->dev set to the
loopback device. When the 'lo' device is taken down, rt6_ifdown initiates
a walk of the fib evicting routes with the 'lo' device which means all
host routes are removed. That process moves the dst which is attached to
an inet6_ifaddr to the gc list and marks it as dead.
The recent change to keep global IPv6 addresses added a new function,
fixup_permanent_addr, that is called on admin up. That function restarts
dad for an inet6_ifaddr and when it completes the host route attached
to it is inserted into the fib. Since the route was marked dead and
moved to the gc list, re-inserting the route causes the reported
out-of-bounds accesses. If the device with the address is taken down
or the address is removed, the WARN_ON in fib6_del is triggered.
All of those faults are fixed by regenerating the host route if the
existing one has been moved to the gc list, something that can be
determined by checking if the rt6i_ref counter is 0.
Fixes: f1705ec197e7 ("net: ipv6: Make address flushing on ifdown optional")
Reported-by: Dmitry Vyukov <[email protected]>
Reported-by: Andrey Konovalov <[email protected]>
Signed-off-by: David Ahern <[email protected]>
Acked-by: Martin KaFai Lau <[email protected]>
Acked-by: Eric Dumazet <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
During removing a bridge device, if the bridge is still up, a new mdb entry
still can be added in br_multicast_add_group() after all mdb entries are
removed in br_multicast_dev_del(). Like the path:
mld_ifc_timer_expire ->
mld_sendpack -> ...
br_multicast_rcv ->
br_multicast_add_group
The new mp's timer will be set up. If the timer expires after the bridge
is freed, it may cause use-after-free panic in br_multicast_group_expired.
BUG: unable to handle kernel NULL pointer dereference at 0000000000000048
IP: [<ffffffffa07ed2c8>] br_multicast_group_expired+0x28/0xb0 [bridge]
Call Trace:
<IRQ>
[<ffffffff81094536>] call_timer_fn+0x36/0x110
[<ffffffffa07ed2a0>] ? br_mdb_free+0x30/0x30 [bridge]
[<ffffffff81096967>] run_timer_softirq+0x237/0x340
[<ffffffff8108dcbf>] __do_softirq+0xef/0x280
[<ffffffff8169889c>] call_softirq+0x1c/0x30
[<ffffffff8102c275>] do_softirq+0x65/0xa0
[<ffffffff8108e055>] irq_exit+0x115/0x120
[<ffffffff81699515>] smp_apic_timer_interrupt+0x45/0x60
[<ffffffff81697a5d>] apic_timer_interrupt+0x6d/0x80
Nikolay also found it would cause a memory leak - the mdb hash is
reallocated and not freed due to the mdb rehash.
unreferenced object 0xffff8800540ba800 (size 2048):
backtrace:
[<ffffffff816e2287>] kmemleak_alloc+0x67/0xc0
[<ffffffff81260bea>] __kmalloc+0x1ba/0x3e0
[<ffffffffa05c60ee>] br_mdb_rehash+0x5e/0x340 [bridge]
[<ffffffffa05c74af>] br_multicast_new_group+0x43f/0x6e0 [bridge]
[<ffffffffa05c7aa3>] br_multicast_add_group+0x203/0x260 [bridge]
[<ffffffffa05ca4b5>] br_multicast_rcv+0x945/0x11d0 [bridge]
[<ffffffffa05b6b10>] br_dev_xmit+0x180/0x470 [bridge]
[<ffffffff815c781b>] dev_hard_start_xmit+0xbb/0x3d0
[<ffffffff815c8743>] __dev_queue_xmit+0xb13/0xc10
[<ffffffff815c8850>] dev_queue_xmit+0x10/0x20
[<ffffffffa02f8d7a>] ip6_finish_output2+0x5ca/0xac0 [ipv6]
[<ffffffffa02fbfc6>] ip6_finish_output+0x126/0x2c0 [ipv6]
[<ffffffffa02fc245>] ip6_output+0xe5/0x390 [ipv6]
[<ffffffffa032b92c>] NF_HOOK.constprop.44+0x6c/0x240 [ipv6]
[<ffffffffa032bd16>] mld_sendpack+0x216/0x3e0 [ipv6]
[<ffffffffa032d5eb>] mld_ifc_timer_expire+0x18b/0x2b0 [ipv6]
This could happen when ip link remove a bridge or destroy a netns with a
bridge device inside.
With Nikolay's suggestion, this patch is to clean up bridge multicast in
ndo_uninit after bridge dev is shutdown, instead of br_dev_delete, so
that netif_running check in br_multicast_add_group can avoid this issue.
v1->v2:
- fix this issue by moving br_multicast_dev_del to ndo_uninit, instead
of calling dev_close in br_dev_delete.
(NOTE: Depends upon b6fe0440c637 ("bridge: implement missing ndo_uninit()"))
Fixes: e10177abf842 ("bridge: multicast: fix handling of temp and perm entries")
Reported-by: Jianwen Ji <[email protected]>
Signed-off-by: Xin Long <[email protected]>
Reviewed-by: Stephen Hemminger <[email protected]>
Signed-off-by: Nikolay Aleksandrov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Commit a149e7c7ce81 ("ipv6: sr: add support for SRH injection through
setsockopt") introduced handling of IPV6_SRCRT_TYPE_4, but at the same
time restricted it to only IPV6_SRCRT_TYPE_0 and
IPV6_SRCRT_TYPE_4. Previously, ipv6_push_exthdr() and fl6_update_dst()
would also handle other values (ie STRICT and TYPE_2).
Restore previous source routing behavior, by handling IPV6_SRCRT_STRICT
and IPV6_SRCRT_TYPE_2 the same way as IPV6_SRCRT_TYPE_0 in
ipv6_push_exthdr() and fl6_update_dst().
Fixes: a149e7c7ce81 ("ipv6: sr: add support for SRH injection through setsockopt")
Signed-off-by: Sabrina Dubroca <[email protected]>
Reviewed-by: Hannes Frederic Sowa <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
This provides a generic SKB based non-optimized XDP path which is used
if either the driver lacks a specific XDP implementation, or the user
requests it via a new IFLA_XDP_FLAGS value named XDP_FLAGS_SKB_MODE.
It is arguable that perhaps I should have required something like
this as part of the initial XDP feature merge.
I believe this is critical for two reasons:
1) Accessibility. More people can play with XDP with less
dependencies. Yes I know we have XDP support in virtio_net, but
that just creates another depedency for learning how to use this
facility.
I wrote this to make life easier for the XDP newbies.
2) As a model for what the expected semantics are. If there is a pure
generic core implementation, it serves as a semantic example for
driver folks adding XDP support.
One thing I have not tried to address here is the issue of
XDP_PACKET_HEADROOM, thanks to Daniel for spotting that. It seems
incredibly expensive to do a skb_cow(skb, XDP_PACKET_HEADROOM) or
whatever even if the XDP program doesn't try to push headers at all.
I think we really need the verifier to somehow propagate whether
certain XDP helpers are used or not.
v5:
- Handle both negative and positive offset after running prog
- Fix mac length in XDP_TX case (Alexei)
- Use rcu_dereference_protected() in free_netdev (kbuild test robot)
v4:
- Fix MAC header adjustmnet before calling prog (David Ahern)
- Disable LRO when generic XDP is installed (Michael Chan)
- Bypass qdisc et al. on XDP_TX and record the event (Alexei)
- Do not perform generic XDP on reinjected packets (DaveM)
v3:
- Make sure XDP program sees packet at MAC header, push back MAC
header if we do XDP_TX. (Alexei)
- Elide GRO when generic XDP is in use. (Alexei)
- Add XDP_FLAG_SKB_MODE flag which the user can use to request generic
XDP even if the driver has an XDP implementation. (Alexei)
- Report whether SKB mode is in use in rtnl_xdp_fill() via XDP_FLAGS
attribute. (Daniel)
v2:
- Add some "fall through" comments in switch statements based
upon feedback from Andrew Lunn
- Use RCU for generic xdp_prog, thanks to Johannes Berg.
Tested-by: Andy Gospodarek <[email protected]>
Tested-by: Jesper Dangaard Brouer <[email protected]>
Tested-by: David Ahern <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Until now in tipc_recv_stream(), we update the received
unacknowledged bytes based on a stack variable and not based on the
actual message size.
If the user buffer passed at tipc_recv_stream() is smaller than the
received skb, the size variable in stack differs from the actual
message size in the skb. This leads to a flow control accounting
error causing permanent congestion.
In this commit, we fix this accounting error by always using the
size of the incoming message.
Fixes: 10724cc7bb78 ("tipc: redesign connection-level flow control")
Signed-off-by: Parthasarathy Bhuvaragan <[email protected]>
Reviewed-by: Jon Maloy <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Until now in tipc_send_stream(), we return -1 when the socket
encounters link congestion even if the socket had successfully
sent partial data. This is incorrect as the application resends
the same the partial data leading to data corruption at
receiver's end.
In this commit, we return the partially sent bytes as the return
value at link congestion.
Fixes: 10724cc7bb78 ("tipc: redesign connection-level flow control")
Signed-off-by: Parthasarathy Bhuvaragan <[email protected]>
Reviewed-by: Jon Maloy <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The ipv6 stub pointer is currently initialized before the ipv6
routing subsystem: a 3rd party can access and use such stub
before the routing data is ready.
Moreover, such pointer is not cleared in case of initialization
error, possibly leading to dangling pointers usage.
This change addresses the above moving the stub initialization
at the end of ipv6 init code.
Fixes: 5f81bd2e5d80 ("ipv6: export a stub for IPv6 symbols used by vxlan")
Signed-off-by: Paolo Abeni <[email protected]>
Acked-by: Cong Wang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Export type of l2tpeth interfaces to userspace
(/sys/class/net/<iface>/uevent).
Signed-off-by: Guillaume Nault <[email protected]>
Acked-by: James Chapman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Export naming scheme used when creating l2tpeth interfaces
(/sys/class/net/<iface>/name_assign_type). This let userspace know if
the device's name has been generated automatically or defined manually.
Signed-off-by: Guillaume Nault <[email protected]>
Acked-by: James Chapman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
per discussion at netconf/netdev:
When we have an action that is capable of branching (example a policer),
we can achieve a continuation of the action graph by programming a
"continue" where we find an exact replica of the same filter rule with a lower
priority and the remainder of the action graph. When you have 100s of thousands
of filters which require such a feature it gets very inefficient to do two
lookups.
This patch completes a leftover feature of action codes. Its time has come.
Example below where a user labels packets with a different skbmark on ingress
of a port depending on whether they have/not exceeded the configured rate.
This mark is then used to make further decisions on some egress port.
#rate control, very low so we can easily see the effect
sudo $TC actions add action police rate 1kbit burst 90k \
conform-exceed pipe/jump 2 index 10
# skbedit index 11 will be used if the user conforms
sudo $TC actions add action skbedit mark 11 ok index 11
# skbedit index 12 will be used if the user does not conform
sudo $TC actions add action skbedit mark 12 ok index 12
#lets bind the user ..
sudo $TC filter add dev $ETH parent ffff: protocol ip prio 8 u32 \
match ip dst 127.0.0.8/32 flowid 1:10 \
action police index 10 \
action skbedit index 11 \
action skbedit index 12
#run a ping -f and see what happens..
#
jhs@foobar:~$ sudo $TC -s filter ls dev $ETH parent ffff: protocol ip
filter pref 8 u32
filter pref 8 u32 fh 800: ht divisor 1
filter pref 8 u32 fh 800::800 order 2048 key ht 800 bkt 0 flowid 1:10 (rule hit 2800 success 1005)
match 7f000008/ffffffff at 16 (success 1005 )
action order 1: police 0xa rate 1Kbit burst 23440b mtu 2Kb action pipe/jump 2 overhead 0b
ref 2 bind 1 installed 207 sec used 122 sec
Action statistics:
Sent 84420 bytes 1005 pkt (dropped 0, overlimits 721 requeues 0)
backlog 0b 0p requeues 0
action order 2: skbedit mark 11 pass
index 11 ref 2 bind 1 installed 204 sec used 122 sec
Action statistics:
Sent 60564 bytes 721 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
action order 3: skbedit mark 12 pass
index 12 ref 2 bind 1 installed 201 sec used 122 sec
Action statistics:
Sent 23856 bytes 284 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
Not bad, about 28% non-conforming packets..
Signed-off-by: Jamal Hadi Salim <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
When recalculating the outer ICMPv6 checksum for a reverse path NATv6
such as ICMPV6_TIME_EXCEED nf_nat_icmpv6_reply_translation() was
accessing data beyond the headlen of the skb for non-linear skb. This
resulted in incorrect ICMPv6 checksum as garbage data was used.
Patch replaces csum_partial() with skb_checksum() which supports
non-linear skbs similar to nf_nat_icmp_reply_translation() from ipv4.
Signed-off-by: Dave Johnson <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
Currently, after adding the following nft rules:
# nft add set x target1 { type ipv4_addr \; flags timeout \;}
# nft add rule x y set add ip daddr timeout 1d @target1 counter
the counters will always be zero despite of the elements are added
to the dynamic set "target1" or not, as we will break the nft expr
traversal unconditionally:
# nft list ruleset
...
set target1 {
...
elements = { 8.8.8.8 expires 23h59m53s}
}
chain output {
...
set add ip daddr timeout 1d @target1 counter packets 0 bytes 0
^ ^
...
}
Since we add the elements to the set successfully, we should continue
to the next expression.
Additionally, if elements are added to "flow table" successfully, we
will _always_ continue to the next expr, even if the operation is
_OP_ADD. So it's better to keep them to be consistent.
Fixes: 22fe54d5fefc ("netfilter: nf_tables: add support for dynamic set updates")
Reported-by: Robert White <[email protected]>
Signed-off-by: Liping Zhang <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
When trying to redirect bridged frames to the bridge device itself or
a bridge port (brouting) via the dnat target then this currently fails:
The ethernet destination of the frame is dnat'ed to the MAC address of
the bridge device or port just fine. However, the IP code drops it in
the beginning of ip_input.c/ip_rcv() as the dnat target left
the skb->pkt_type as PACKET_OTHERHOST.
Fixing this by resetting skb->pkt_type to an appropriate type after
dnat'ing.
Signed-off-by: Linus Lüssing <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
The CAN gateway was not implemented as per-net in the initial network
namespace support by Mario Kicherer (8e8cda6d737d).
This patch enables the CAN gateway to be used in different namespaces.
Signed-off-by: Oliver Hartkopp <[email protected]>
Signed-off-by: Marc Kleine-Budde <[email protected]>
|
|
The CAN_BCM protocol and its procfs entries were not implemented as per-net
in the initial network namespace support by Mario Kicherer (8e8cda6d737d).
This patch adds the missing per-net functionality for the CAN BCM.
Signed-off-by: Oliver Hartkopp <[email protected]>
Signed-off-by: Marc Kleine-Budde <[email protected]>
|
|
The statistics and its proc output was not implemented as per-net in the
initial network namespace support by Mario Kicherer (8e8cda6d737d).
This patch adds the missing per-net statistics for the CAN subsystem.
Signed-off-by: Oliver Hartkopp <[email protected]>
Signed-off-by: Marc Kleine-Budde <[email protected]>
|
|
can_rx_alldev_list is a per-net data structure now. Remove it's definition
here and can_rx_dev_list too.
Signed-off-by: Oliver Hartkopp <[email protected]>
Signed-off-by: Marc Kleine-Budde <[email protected]>
|
|
The namespace support for the CAN subsystem does not need any additional
memory. So when ".size = 0" there's no extra memory allocated by the system.
And therefore ".id" is obsolete too.
Signed-off-by: Oliver Hartkopp <[email protected]>
Signed-off-by: Marc Kleine-Budde <[email protected]>
|
|
The can_rx_alldev_list is a per-net data structure now and allocated in
can_pernet_init(). Make sure the memory is free'd in can_pernet_exit() too.
Signed-off-by: Oliver Hartkopp <[email protected]>
Signed-off-by: Marc Kleine-Budde <[email protected]>
|
|
* Convert both smp and selftest to crypto kpp API
* Remove module ecc as no more required
* Add ecdh_helper functions for wrapping kpp async calls
This patch has been tested *only* with selftest, which is called on
module loading.
Signed-off-by: Salvatore Benedetto <[email protected]>
Signed-off-by: Marcel Holtmann <[email protected]>
|
|
Function nlmsg_new() will return a NULL pointer if there is no enough
memory, and its return value should be checked before it is used.
However, in function tipc_nl_node_get_monitor(), the validation of the
return value of function nlmsg_new() is missed. This patch fixes the
bug.
Signed-off-by: Pan Bian <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Function nla_nest_start() may return a NULL pointer on error. However,
in function lwtunnel_fill_encap(), the return value of nla_nest_start()
is not validated before it is used. This patch checks the return value
of nla_nest_start() against NULL.
Signed-off-by: Pan Bian <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Add support to the tc flower classifier to match based on fields in MPLS
labels (TTL, Bottom of Stack, TC field, Label).
Signed-off-by: Benjamin LaHaise <[email protected]>
Signed-off-by: Benjamin LaHaise <[email protected]>
Reviewed-by: Jakub Kicinski <[email protected]>
Cc: "David S. Miller" <[email protected]>
Cc: Simon Horman <[email protected]>
Cc: Jamal Hadi Salim <[email protected]>
Cc: Cong Wang <[email protected]>
Cc: Jiri Pirko <[email protected]>
Cc: Eric Dumazet <[email protected]>
Cc: Hadar Hen Zion <[email protected]>
Cc: Gao Feng <[email protected]>
Acked-by: Jiri Pirko <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Add support for parsing MPLS flows to the flow dissector in preparation for
adding MPLS match support to cls_flower.
Signed-off-by: Benjamin LaHaise <[email protected]>
Signed-off-by: Benjamin LaHaise <[email protected]>
Reviewed-by: Jakub Kicinski <[email protected]>
Cc: "David S. Miller" <[email protected]>
Cc: Simon Horman <[email protected]>
Cc: Jamal Hadi Salim <[email protected]>
Cc: Cong Wang <[email protected]>
Cc: Jiri Pirko <[email protected]>
Cc: Eric Dumazet <[email protected]>
Cc: Hadar Hen Zion <[email protected]>
Cc: Gao Feng <[email protected]>
Acked-by: Jiri Pirko <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Christoph Paasch from Apple found another firewall issue for TFO:
After successful 3WHS using TFO, server and client starts to exchange
data. Afterwards, a 10s idle time occurs on this connection. After that,
firewall starts to drop every packet on this connection.
The fix for this issue is to extend existing firewall blackhole detection
logic in tcp_write_timeout() by removing the mss check.
Signed-off-by: Wei Wang <[email protected]>
Acked-by: Yuchung Cheng <[email protected]>
Acked-by: Neal Cardwell <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
This counter records the number of times the firewall blackhole issue is
detected and active TFO is disabled.
Signed-off-by: Wei Wang <[email protected]>
Acked-by: Yuchung Cheng <[email protected]>
Acked-by: Neal Cardwell <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Middlebox firewall issues can potentially cause server's data being
blackholed after a successful 3WHS using TFO. Following are the related
reports from Apple:
https://www.nanog.org/sites/default/files/Paasch_Network_Support.pdf
Slide 31 identifies an issue where the client ACK to the server's data
sent during a TFO'd handshake is dropped.
C ---> syn-data ---> S
C <--- syn/ack ----- S
C (accept & write)
C <---- data ------- S
C ----- ACK -> X S
[retry and timeout]
https://www.ietf.org/proceedings/94/slides/slides-94-tcpm-13.pdf
Slide 5 shows a similar situation that the server's data gets dropped
after 3WHS.
C ---- syn-data ---> S
C <--- syn/ack ----- S
C ---- ack --------> S
S (accept & write)
C? X <- data ------ S
[retry and timeout]
This is the worst failure b/c the client can not detect such behavior to
mitigate the situation (such as disabling TFO). Failing to proceed, the
application (e.g., SSL library) may simply timeout and retry with TFO
again, and the process repeats indefinitely.
The proposed solution is to disable active TFO globally under the
following circumstances:
1. client side TFO socket detects out of order FIN
2. client side TFO socket receives out of order RST
We disable active side TFO globally for 1hr at first. Then if it
happens again, we disable it for 2h, then 4h, 8h, ...
And we reset the timeout to 1hr if a client side TFO sockets not opened
on loopback has successfully received data segs from server.
And we examine this condition during close().
The rational behind it is that when such firewall issue happens,
application running on the client should eventually close the socket as
it is not able to get the data it is expecting. Or application running
on the server should close the socket as it is not able to receive any
response from client.
In both cases, out of order FIN or RST will get received on the client
given that the firewall will not block them as no data are in those
frames.
And we want to disable active TFO globally as it helps if the middle box
is very close to the client and most of the connections are likely to
fail.
Also, add a debug sysctl:
tcp_fastopen_blackhole_detect_timeout_sec:
the initial timeout to use when firewall blackhole issue happens.
This can be set and read.
When setting it to 0, it means to disable the active disable logic.
Signed-off-by: Wei Wang <[email protected]>
Acked-by: Yuchung Cheng <[email protected]>
Acked-by: Neal Cardwell <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5-updates-2017-04-22
Sparse and compiler warnings fixes from Stephen Hemminger.
From Roi Dayan and Or Gerlitz, Add devlink and mlx5 support for controlling
E-Switch encapsulation mode, this knob will enable HW support for applying
encapsulation/decapsulation to VF traffic as part of SRIOV e-switch offloading.
====================
Signed-off-by: David S. Miller <[email protected]>
|
|
systemd-sysctl is triggering a suspicious RCU usage message when
net.ipv4.tcp_early_demux or net.ipv4.udp_early_demux is changed via
a sysctl config file:
[ 33.896184] ===============================
[ 33.899558] [ ERR: suspicious RCU usage. ]
[ 33.900624] 4.11.0-rc7+ #104 Not tainted
[ 33.901698] -------------------------------
[ 33.903059] /home/dsa/kernel-2.git/net/ipv4/sysctl_net_ipv4.c:305 suspicious rcu_dereference_check() usage!
[ 33.905724]
other info that might help us debug this:
[ 33.907656]
rcu_scheduler_active = 2, debug_locks = 0
[ 33.909288] 1 lock held by systemd-sysctl/143:
[ 33.910373] #0: (sb_writers#5){.+.+.+}, at: [<ffffffff8123a370>] file_start_write+0x45/0x48
[ 33.912407]
stack backtrace:
[ 33.914018] CPU: 0 PID: 143 Comm: systemd-sysctl Not tainted 4.11.0-rc7+ #104
[ 33.915631] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.7.5-20140531_083030-gandalf 04/01/2014
[ 33.917870] Call Trace:
[ 33.918431] dump_stack+0x81/0xb6
[ 33.919241] lockdep_rcu_suspicious+0x10f/0x118
[ 33.920263] proc_configure_early_demux+0x65/0x10a
[ 33.921391] proc_udp_early_demux+0x3a/0x41
add rcu locking to proc_configure_early_demux.
Fixes: dddb64bcb3461 ("net: Add sysctl to toggle early demux for tcp and udp")
Signed-off-by: David Ahern <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
When arp_notify is set to 1 for either a specific interface or for 'all'
interfaces, gratuitous arp requests are sent. Since ndisc_notify is the
ipv6 equivalent to arp_notify, it should follow the same semantics.
Commit 4a6e3c5def13 ("net: ipv6: send unsolicited NA on admin up") sends
the NA on admin up. The final piece is checking devconf_all->ndisc_notify
in addition to the per device setting. Add it.
Fixes: 5cb04436eef6 ("ipv6: add knob to send unsolicited ND on link-layer address change")
Signed-off-by: David Ahern <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Commit 834184b1f3a4 ("netfilter: defrag: only register defrag
functionality if needed") used the outdated XT_SOCKET_HAVE_IPV6 macro
which was removed earlier in commit 8db4c5be88f6 ("netfilter: move
socket lookup infrastructure to nf_socket_ipv{4,6}.c"). With that macro
never being defined, the xt_socket match emits an "Unknown family 10"
warning when used with IPv6:
WARNING: CPU: 0 PID: 1377 at net/netfilter/xt_socket.c:160 socket_mt_enable_defrag+0x47/0x50 [xt_socket]
Unknown family 10
Modules linked in: xt_socket nf_socket_ipv4 nf_socket_ipv6 nf_defrag_ipv4 [...]
CPU: 0 PID: 1377 Comm: ip6tables-resto Not tainted 4.10.10 #1
Hardware name: [...]
Call Trace:
? __warn+0xe7/0x100
? socket_mt_enable_defrag+0x47/0x50 [xt_socket]
? socket_mt_enable_defrag+0x47/0x50 [xt_socket]
? warn_slowpath_fmt+0x39/0x40
? socket_mt_enable_defrag+0x47/0x50 [xt_socket]
? socket_mt_v2_check+0x12/0x40 [xt_socket]
? xt_check_match+0x6b/0x1a0 [x_tables]
? xt_find_match+0x93/0xd0 [x_tables]
? xt_request_find_match+0x20/0x80 [x_tables]
? translate_table+0x48e/0x870 [ip6_tables]
? translate_table+0x577/0x870 [ip6_tables]
? walk_component+0x3a/0x200
? kmalloc_order+0x1d/0x50
? do_ip6t_set_ctl+0x181/0x490 [ip6_tables]
? filename_lookup+0xa5/0x120
? nf_setsockopt+0x3a/0x60
? ipv6_setsockopt+0xb0/0xc0
? sock_common_setsockopt+0x23/0x30
? SyS_socketcall+0x41d/0x630
? vfs_read+0xfa/0x120
? do_fast_syscall_32+0x7a/0x110
? entry_SYSENTER_32+0x47/0x71
This patch brings the conditional back in line with how the rest of the
file handles IPv6.
Fixes: 834184b1f3a4 ("netfilter: defrag: only register defrag functionality if needed")
Signed-off-by: Peter Tirsek <[email protected]>
Acked-by: Florian Westphal <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
We should acquire the ct->lock before accessing or modifying the
nf_ct_seqadj, as another CPU may modify the nf_ct_seqadj at the same
time during its packet proccessing.
Signed-off-by: Liping Zhang <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
After converting to use rcu for conntrack hash, one CPU may update
the ct->status via ctnetlink, while another CPU may process the
packets and update the ct->status.
So the non-atomic operation "ct->status |= status;" via ctnetlink
becomes unsafe, and this may clear the IPS_DYING_BIT bit set by
another CPU unexpectedly. For example:
CPU0 CPU1
ctnetlink_change_status __nf_conntrack_find_get
old = ct->status nf_ct_gc_expired
- nf_ct_kill
- test_and_set_bit(IPS_DYING_BIT
new = old | status; -
ct->status = new; <-- oops, _DYING_ is cleared!
Now using a series of atomic bit operation to solve the above issue.
Also note, user shouldn't set IPS_TEMPLATE, IPS_SEQ_ADJUST directly,
so make these two bits be unchangable too.
If we set the IPS_TEMPLATE_BIT, ct will be freed by nf_ct_tmpl_free,
but actually it is alloced by nf_conntrack_alloc.
If we set the IPS_SEQ_ADJUST_BIT, this may cause the NULL pointer
deference, as the nfct_seqadj(ct) maybe NULL.
Last, add some comments to describe the logic change due to the
commit a963d710f367 ("netfilter: ctnetlink: Fix regression in CTA_STATUS
processing"), which makes me feel a little confusing.
Fixes: 76507f69c44e ("[NETFILTER]: nf_conntrack: use RCU for conntrack hash")
Signed-off-by: Liping Zhang <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
Currently, ctnetlink_change_conntrack is always protected by _expect_lock,
but this will cause a deadlock when deleting the helper from a conntrack,
as the _expect_lock will be acquired again by nf_ct_remove_expectations:
CPU0
----
lock(nf_conntrack_expect_lock);
lock(nf_conntrack_expect_lock);
*** DEADLOCK ***
May be due to missing lock nesting notation
2 locks held by lt-conntrack_gr/12853:
#0: (&table[i].mutex){+.+.+.}, at: [<ffffffffa05e2009>]
nfnetlink_rcv_msg+0x399/0x6a9 [nfnetlink]
#1: (nf_conntrack_expect_lock){+.....}, at: [<ffffffffa05f2c1f>]
ctnetlink_new_conntrack+0x17f/0x408 [nf_conntrack_netlink]
Call Trace:
dump_stack+0x85/0xc2
__lock_acquire+0x1608/0x1680
? ctnetlink_parse_tuple_proto+0x10f/0x1c0 [nf_conntrack_netlink]
lock_acquire+0x100/0x1f0
? nf_ct_remove_expectations+0x32/0x90 [nf_conntrack]
_raw_spin_lock_bh+0x3f/0x50
? nf_ct_remove_expectations+0x32/0x90 [nf_conntrack]
nf_ct_remove_expectations+0x32/0x90 [nf_conntrack]
ctnetlink_change_helper+0xc6/0x190 [nf_conntrack_netlink]
ctnetlink_new_conntrack+0x1b2/0x408 [nf_conntrack_netlink]
nfnetlink_rcv_msg+0x60a/0x6a9 [nfnetlink]
? nfnetlink_rcv_msg+0x1b9/0x6a9 [nfnetlink]
? nfnetlink_bind+0x1a0/0x1a0 [nfnetlink]
netlink_rcv_skb+0xa4/0xc0
nfnetlink_rcv+0x87/0x770 [nfnetlink]
Since the operations are unrelated to nf_ct_expect, so we can drop the
_expect_lock. Also note, after removing the _expect_lock protection,
another CPU may invoke nf_conntrack_helper_unregister, so we should
use rcu_read_lock to protect __nf_conntrack_helper_find invoked by
ctnetlink_change_helper.
Fixes: ca7433df3a67 ("netfilter: conntrack: seperate expect locking from nf_conntrack_lock")
Signed-off-by: Liping Zhang <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
First, when creating a new ct, we will invoke request_module to try to
load the related inkernel cthelper. So there's no need to call the
request_module again when updating the ct helpinfo.
Second, ctnetlink_change_helper may be called with rcu_read_lock held,
i.e. rcu_read_lock -> nfqnl_recv_verdict -> nfqnl_ct_parse ->
ctnetlink_glue_parse -> ctnetlink_glue_parse_ct ->
ctnetlink_change_helper. But the request_module invocation may sleep,
so we can't call it with the rcu_read_lock held.
Remove it now.
Signed-off-by: Liping Zhang <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
We forget to free dummy elements when deleting the set. So when I was
running nft-test.py, I saw many kmemleak warnings:
kmemleak: 1344 new suspected memory leaks ...
# cat /sys/kernel/debug/kmemleak
unreferenced object 0xffff8800631345c8 (size 32):
comm "nft", pid 9075, jiffies 4295743309 (age 1354.815s)
hex dump (first 32 bytes):
f8 63 13 63 00 88 ff ff 88 79 13 63 00 88 ff ff .c.c.....y.c....
04 0c 00 00 00 00 00 00 00 00 00 00 08 03 00 00 ................
backtrace:
[<ffffffff819059da>] kmemleak_alloc+0x4a/0xa0
[<ffffffff81288174>] __kmalloc+0x164/0x310
[<ffffffffa061269d>] nft_set_elem_init+0x3d/0x1b0 [nf_tables]
[<ffffffffa06130da>] nft_add_set_elem+0x45a/0x8c0 [nf_tables]
[<ffffffffa0613645>] nf_tables_newsetelem+0x105/0x1d0 [nf_tables]
[<ffffffffa05fe6d4>] nfnetlink_rcv+0x414/0x770 [nfnetlink]
[<ffffffff817f0ca6>] netlink_unicast+0x1f6/0x310
[<ffffffff817f10c6>] netlink_sendmsg+0x306/0x3b0
...
Fixes: e920dde516088 ("netfilter: nft_set_bitmap: keep a list of dummy elements")
Signed-off-by: Liping Zhang <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
cthelpers added via nfnetlink may have the same tuple, i.e. except for
the l3proto and l4proto, other fields are all zero. So even with the
different names, we will also fail to add them:
# nfct helper add ssdp inet udp
# nfct helper add tftp inet udp
nfct v1.4.3: netlink error: File exists
So in order to avoid unpredictable behaviour, we should:
1. cthelpers can be selected by nft ct helper obj or xt_CT target, so
report error if duplicated { name, l3proto, l4proto } tuple exist.
2. cthelpers can be selected by nf_ct_tuple_src_mask_cmp when
nf_ct_auto_assign_helper is enabled, so also report error if duplicated
{ l3proto, l4proto, src-port } tuple exist.
Also note, if the cthelper is added from userspace, then the src-port will
always be zero, it's invalid for nf_ct_auto_assign_helper, so there's no
need to check the second point listed above.
Fixes: 893e093c786c ("netfilter: nf_ct_helper: bail out on duplicated helpers")
Signed-off-by: Liping Zhang <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
Conntrack helpers do not check for a potentially clashing conntrack
entry when creating a new expectation. Also, nf_conntrack_in() will
check expectations (via init_conntrack()) only if a conntrack entry
can not be found. The expectation for a packet which also matches an
existing conntrack entry will not be removed by conntrack, and is
currently handled inconsistently by OVS, as OVS expects the
expectation to be removed when the connection tracking entry matching
that expectation is confirmed.
It should be noted that normally an IP stack would not allow reuse of
a 5-tuple of an old (possibly lingering) connection for a new data
connection, so this is somewhat unlikely corner case. However, it is
possible that a misbehaving source could cause conntrack entries be
created that could then interfere with new related connections.
Fix this in the OVS module by deleting the clashing conntrack entry
after an expectation has been matched. This causes the following
nf_conntrack_in() call also find the expectation and remove it when
creating the new conntrack entry, as well as the forthcoming reply
direction packets to match the new related connection instead of the
old clashing conntrack entry.
Fixes: 7f8a436eaa2c ("openvswitch: Add conntrack action")
Reported-by: Yang Song <[email protected]>
Signed-off-by: Jarno Rajahalme <[email protected]>
Acked-by: Joe Stringer <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
There are two cases which causes refcnt leak.
1. When nf_ct_timeout_ext_add failed in xt_ct_set_timeout, it should
free the timeout refcnt.
Now goto the err_put_timeout error handler instead of going ahead.
2. When the time policy is not found, we should call module_put.
Otherwise, the related cthelper module cannot be removed anymore.
It is easy to reproduce by typing the following command:
# iptables -t raw -A OUTPUT -p tcp -j CT --helper ftp --timeout xxx
Signed-off-by: Gao Feng <[email protected]>
Signed-off-by: Liping Zhang <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>
|
|
Add a new optional conntrack action attribute OVS_CT_ATTR_EVENTMASK,
which can be used in conjunction with the commit flag
(OVS_CT_ATTR_COMMIT) to set the mask of bits specifying which
conntrack events (IPCT_*) should be delivered via the Netfilter
netlink multicast groups. Default behavior depends on the system
configuration, but typically a lot of events are delivered. This can be
very chatty for the NFNLGRP_CONNTRACK_UPDATE group, even if only some
types of events are of interest.
Netfilter core init_conntrack() adds the event cache extension, so we
only need to set the ctmask value. However, if the system is
configured without support for events, the setting will be skipped due
to extension not being found.
Signed-off-by: Jarno Rajahalme <[email protected]>
Reviewed-by: Greg Rose <[email protected]>
Acked-by: Joe Stringer <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Fix typo in a comment.
Signed-off-by: Jarno Rajahalme <[email protected]>
Acked-by: Greg Rose <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Otherwise, UDP checksum offloads could corrupt ESP packets by attempting
to calculate UDP checksum when this inner UDP packet is already protected
by IPsec.
One way to reproduce this bug is to have a VM with virtio_net driver (UFO
set to ON in the guest VM); and then encapsulate all guest's Ethernet
frames in Geneve; and then further encrypt Geneve with IPsec. In this
case following symptoms are observed:
1. If using ixgbe NIC, then it will complain with following error message:
ixgbe 0000:01:00.1: partial checksum but l4 proto=32!
2. Receiving IPsec stack will drop all the corrupted ESP packets and
increase XfrmInStateProtoError counter in /proc/net/xfrm_stat.
3. iperf UDP test from the VM with packet sizes above MTU will not work at
all.
4. iperf TCP test from the VM will get ridiculously low performance because.
Signed-off-by: Ansis Atteka <[email protected]>
Co-authored-by: Steffen Klassert <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
David reported that doing the following:
ip li add red type vrf table 10
ip link set dev eth1 vrf red
ip addr add 127.0.0.1/8 dev red
ip link set dev eth1 up
ip li set red up
ping -c1 -w1 -I red 127.0.0.1
ip li del red
when either policy routing IP rules are present or the local table
lookup ip rule is before the l3mdev lookup results in a hang with
these messages:
unregister_netdevice: waiting for red to become free. Usage count = 1
The problem is caused by caching the dst used for sending the packet
out of the specified interface on a local route with a different
nexthop interface. Thus the dst could stay around until the route in
the table the lookup was done is deleted which may be never.
Address the problem by not forcing output device to be the l3mdev in
the flow's output interface if the lookup didn't use the l3mdev. This
then results in the dst using the right device according to the route.
Changes in v2:
- make the dev_out passed in by __ip_route_output_key_hash correct
instead of checking the nh dev if FLOWI_FLAG_SKIP_NH_OIF is set as
suggested by David.
Fixes: 5f02ce24c2696 ("net: l3mdev: Allow the l3mdev to be a loopback")
Reported-by: David Ahern <[email protected]>
Suggested-by: David Ahern <[email protected]>
Signed-off-by: Robert Shearman <[email protected]>
Acked-by: David Ahern <[email protected]>
Tested-by: David Ahern <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|