syzbot found the following issue on:
==================================================================
BUG: KASAN: use-after-free in tipc_named_reinit+0x94f/0x9b0
net/tipc/name_distr.c:413
Read of size 8 at addr ffff88805299a000 by task kworker/1:9/23764
CPU: 1 PID: 23764 Comm: kworker/1:9 Not tainted
5.18.0-rc4-syzkaller-00878-g17d49e6e8012 #0
Hardware name: Google Compute Engine/Google Compute Engine,
BIOS Google 01/01/2011
Workqueue: events tipc_net_finalize_work
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0xcd/0x134 lib/dump_stack.c:106
print_address_description.constprop.0.cold+0xeb/0x495
mm/kasan/report.c:313
print_report mm/kasan/report.c:429 [inline]
kasan_report.cold+0xf4/0x1c6 mm/kasan/report.c:491
tipc_named_reinit+0x94f/0x9b0 net/tipc/name_distr.c:413
tipc_net_finalize+0x234/0x3d0 net/tipc/net.c:138
process_one_work+0x996/0x1610 kernel/workqueue.c:2289
worker_thread+0x665/0x1080 kernel/workqueue.c:2436
kthread+0x2e9/0x3a0 kernel/kthread.c:376
ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:298
</TASK>
[...]
==================================================================
In the commit
d966ddcc38 ("tipc: fix a deadlock when flushing scheduled work"),
the cancel_work_sync() function just to make sure ONLY the work
tipc_net_finalize_work() is executing/pending on any CPU completed before
tipc namespace is destroyed through tipc_exit_net(). But this function
is not guaranteed the work is the last queued. So, the destroyed instance
may be accessed in the work which will try to enqueue later.
In order to completely fix, we re-order the calling of cancel_work_sync()
to make sure the work tipc_net_finalize_work() was last queued and it
must be completed by calling cancel_work_sync().
Reported-by: syzbot+47af19f3307fc9c5c82e@syzkaller.appspotmail.com
Fixes: d966ddcc38 ("tipc: fix a deadlock when flushing scheduled work")
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is to use "struct work_struct" for the finalize work queue
instead of "struct tipc_net_work", as it can get the "net" and "addr"
from tipc_net's other members and there is no need to add extra net
and addr in tipc_net by defining "struct tipc_net_work".
Note that it's safe to get net from tn->bcl as bcl is always released
after the finalize work queue is done.
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On some host, a crash could be triggered simply by repeating these
commands several times:
# modprobe tipc
# tipc bearer enable media udp name UDP1 localip 127.0.0.1
# rmmod tipc
[] BUG: unable to handle kernel paging request at ffffffffc096bb00
[] Workqueue: events 0xffffffffc096bb00
[] Call Trace:
[] ? process_one_work+0x1a7/0x360
[] ? worker_thread+0x30/0x390
[] ? create_worker+0x1a0/0x1a0
[] ? kthread+0x116/0x130
[] ? kthread_flush_work_fn+0x10/0x10
[] ? ret_from_fork+0x35/0x40
When removing the TIPC module, the UDP tunnel sock will be delayed to
release in a work queue as sock_release() can't be done in rtnl_lock().
If the work queue is schedule to run after the TIPC module is removed,
kernel will crash as the work queue function cleanup_beareri() code no
longer exists when trying to invoke it.
To fix it, this patch introduce a member wq_count in tipc_net to track
the numbers of work queues in schedule, and wait and exit until all
work queues are done in tipc_exit_net().
Fixes: d0f91938be ("tipc: add ip/udp media type")
Reported-by: Shuang Li <shuali@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
dist_queue is no longer used since commit 37922ea4a3
("tipc: permit overlapping service ranges in name table")
Acked-by: Jon Maloy <jmaloy@redhat.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Hoang Huu Le <hoang.h.le@dektech.com.au>
Link: https://lore.kernel.org/r/20201028032712.31009-1-hoang.h.le@dektech.com.au
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In the commit fdeba99b1e
("tipc: fix use-after-free in tipc_bcast_get_mode"), we're trying
to make sure the tipc_net_finalize_work work item finished if it
enqueued. But calling flush_scheduled_work() is not just affecting
above work item but either any scheduled work. This has turned out
to be overkill and caused to deadlock as syzbot reported:
======================================================
WARNING: possible circular locking dependency detected
5.9.0-rc2-next-20200828-syzkaller #0 Not tainted
------------------------------------------------------
kworker/u4:6/349 is trying to acquire lock:
ffff8880aa063d38 ((wq_completion)events){+.+.}-{0:0}, at: flush_workqueue+0xe1/0x13e0 kernel/workqueue.c:2777
but task is already holding lock:
ffffffff8a879430 (pernet_ops_rwsem){++++}-{3:3}, at: cleanup_net+0x9b/0xb10 net/core/net_namespace.c:565
[...]
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(pernet_ops_rwsem);
lock(&sb->s_type->i_mutex_key#13);
lock(pernet_ops_rwsem);
lock((wq_completion)events);
*** DEADLOCK ***
[...]
v1:
To fix the original issue, we replace above calling by introducing
a bit flag. When a namespace cleaned-up, bit flag is set to zero and:
- tipc_net_finalize functionial just does return immediately.
- tipc_net_finalize_work does not enqueue into the scheduled work queue.
v2:
Use cancel_work_sync() helper to make sure ONLY the
tipc_net_finalize_work() stopped before releasing bcbase object.
Reported-by: syzbot+d5aa7e0385f6a5d0f4fd@syzkaller.appspotmail.com
Fixes: fdeba99b1e ("tipc: fix use-after-free in tipc_bcast_get_mode")
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: Hoang Huu Le <hoang.h.le@dektech.com.au>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The tipc prefix for log messages generated by tipc was
removed in commit 07f6c4bc04 ("tipc: convert tipc reference
table to use generic rhashtable").
This is still a useful prefix so add it back.
Signed-off-by: Matt Bennett <matt.bennett@alliedtelesis.co.nz>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit offers an option to encrypt and authenticate all messaging,
including the neighbor discovery messages. The currently most advanced
algorithm supported is the AEAD AES-GCM (like IPSec or TLS). All
encryption/decryption is done at the bearer layer, just before leaving
or after entering TIPC.
Supported features:
- Encryption & authentication of all TIPC messages (header + data);
- Two symmetric-key modes: Cluster and Per-node;
- Automatic key switching;
- Key-expired revoking (sequence number wrapped);
- Lock-free encryption/decryption (RCU);
- Asynchronous crypto, Intel AES-NI supported;
- Multiple cipher transforms;
- Logs & statistics;
Two key modes:
- Cluster key mode: One single key is used for both TX & RX in all
nodes in the cluster.
- Per-node key mode: Each nodes in the cluster has one specific TX key.
For RX, a node requires its peers' TX key to be able to decrypt the
messages from those peers.
Key setting from user-space is performed via netlink by a user program
(e.g. the iproute2 'tipc' tool).
Internal key state machine:
Attach Align(RX)
+-+ +-+
| V | V
+---------+ Attach +---------+
| IDLE |---------------->| PENDING |(user = 0)
+---------+ +---------+
A A Switch| A
| | | |
| | Free(switch/revoked) | |
(Free)| +----------------------+ | |Timeout
| (TX) | | |(RX)
| | | |
| | v |
+---------+ Switch +---------+
| PASSIVE |<----------------| ACTIVE |
+---------+ (RX) +---------+
(user = 1) (user >= 1)
The number of TFMs is 10 by default and can be changed via the procfs
'net/tipc/max_tfms'. At this moment, as for simplicity, this file is
also used to print the crypto statistics at runtime:
echo 0xfff1 > /proc/sys/net/tipc/max_tfms
The patch defines a new TIPC version (v7) for the encryption message (-
backward compatibility as well). The message is basically encapsulated
as follows:
+----------------------------------------------------------+
| TIPCv7 encryption | Original TIPCv2 | Authentication |
| header | packet (encrypted) | Tag |
+----------------------------------------------------------+
The throughput is about ~40% for small messages (compared with non-
encryption) and ~9% for large messages. With the support from hardware
crypto i.e. the Intel AES-NI CPU instructions, the throughput increases
upto ~85% for small messages and ~55% for large messages.
By default, the new feature is inactive (i.e. no encryption) until user
sets a key for TIPC. There is however also a new option - "TIPC_CRYPTO"
in the kernel configuration to enable/disable the new code when needed.
MAINTAINERS | add two new files 'crypto.h' & 'crypto.c' in tipc
Acked-by: Ying Xue <ying.xue@windreiver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, TIPC transports intra-node user data messages directly
socket to socket, hence shortcutting all the lower layers of the
communication stack. This gives TIPC very good intra node performance,
both regarding throughput and latency.
We now introduce a similar mechanism for TIPC data traffic across
network namespaces located in the same kernel. On the send path, the
call chain is as always accompanied by the sending node's network name
space pointer. However, once we have reliably established that the
receiving node is represented by a namespace on the same host, we just
replace the namespace pointer with the receiving node/namespace's
ditto, and follow the regular socket receive patch though the receiving
node. This technique gives us a throughput similar to the node internal
throughput, several times larger than if we let the traffic go though
the full network stacks. As a comparison, max throughput for 64k
messages is four times larger than TCP throughput for the same type of
traffic.
To meet any security concerns, the following should be noted.
- All nodes joining a cluster are supposed to have been be certified
and authenticated by mechanisms outside TIPC. This is no different for
nodes/namespaces on the same host; they have to auto discover each
other using the attached interfaces, and establish links which are
supervised via the regular link monitoring mechanism. Hence, a kernel
local node has no other way to join a cluster than any other node, and
have to obey to policies set in the IP or device layers of the stack.
- Only when a sender has established with 100% certainty that the peer
node is located in a kernel local namespace does it choose to let user
data messages, and only those, take the crossover path to the receiving
node/namespace.
- If the receiving node/namespace is removed, its namespace pointer
is invalidated at all peer nodes, and their neighbor link monitoring
will eventually note that this node is gone.
- To ensure the "100% certainty" criteria, and prevent any possible
spoofing, received discovery messages must contain a proof that the
sender knows a common secret. We use the hash mix of the sending
node/namespace for this purpose, since it can be accessed directly by
all other namespaces in the kernel. Upon reception of a discovery
message, the receiver checks this proof against all the local
namespaces'hash_mix:es. If it finds a match, that, along with a
matching node id and cluster id, this is deemed sufficient proof that
the peer node in question is in a local namespace, and a wormhole can
be opened.
- We should also consider that TIPC is intended to be a cluster local
IPC mechanism (just like e.g. UNIX sockets) rather than a network
protocol, and hence we think it can justified to allow it to shortcut the
lower protocol layers.
Regarding traceability, we should notice that since commit 6c9081a391
("tipc: add loopback device tracking") it is possible to follow the node
internal packet flow by just activating tcpdump on the loopback
interface. This will be true even for this mechanism; by activating
tcpdump on the involved nodes' loopback interfaces their inter-name
space messaging can easily be tracked.
v2:
- update 'net' pointer when node left/rejoined
v3:
- grab read/write lock when using node ref obj
v4:
- clone traffics between netns to loopback
Suggested-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since node internal messages are passed directly to the socket, it is not
possible to observe those messages via tcpdump or wireshark.
We now remedy this by making it possible to clone such messages and send
the clones to the loopback interface. The clones are dropped at reception
and have no functional role except making the traffic visible.
The feature is enabled if network taps are active for the loopback device.
pcap filtering restrictions require the messages to be presented to the
receiving side of the loopback device.
v3 - Function dev_nit_active used to check for network taps.
- Procedure netif_rx_ni used to send cloned messages to loopback device.
Signed-off-by: John Rutherford <john.rutherford@dektech.com.au>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is to fix a dst defcnt leak, which can be reproduced by doing:
# ip net a c; ip net a s; modprobe tipc
# ip net e s ip l a n eth1 type veth peer n eth1 netns c
# ip net e c ip l s lo up; ip net e c ip l s eth1 up
# ip net e s ip l s lo up; ip net e s ip l s eth1 up
# ip net e c ip a a 1.1.1.2/8 dev eth1
# ip net e s ip a a 1.1.1.1/8 dev eth1
# ip net e c tipc b e m udp n u1 localip 1.1.1.2
# ip net e s tipc b e m udp n u1 localip 1.1.1.1
# ip net d c; ip net d s; rmmod tipc
and it will get stuck and keep logging the error:
unregister_netdevice: waiting for lo to become free. Usage count = 1
The cause is that a dst is held by the udp sock's sk_rx_dst set on udp rx
path with udp_early_demux == 1, and this dst (eventually holding lo dev)
can't be released as bearer's removal in tipc pernet .exit happens after
lo dev's removal, default_device pernet .exit.
"There are two distinct types of pernet_operations recognized: subsys and
device. At creation all subsys init functions are called before device
init functions, and at destruction all device exit functions are called
before subsys exit function."
So by calling register_pernet_device instead to register tipc_net_ops, the
pernet .exit() will be invoked earlier than loopback dev's removal when a
netns is being destroyed, as fou/gue does.
Note that vxlan and geneve udp tunnels don't have this issue, as the udp
sock is released in their device ndo_stop().
This fix is also necessary for tipc dst_cache, which will hold dsts on tx
path and I will introduce in my next patch.
Reported-by: Li Shuang <shuali@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Error message printed:
modprobe: ERROR: could not insert 'tipc': Address family not
supported by protocol.
when modprobe tipc after the following patch: switch order of
device registration, commit 7e27e8d613
("tipc: switch order of device registration to fix a crash")
Because sock_create_kern(net, AF_TIPC, ...) called by
tipc_topsrv_create_listener() in the initialization process
of tipc_init_net(), so tipc_socket_init() must be execute before that.
Meanwhile, tipc_net_id need to be initialized when sock_create()
called, and tipc_socket_init() is no need to be called for each namespace.
I add a variable tipc_topsrv_net_ops, and split the
register_pernet_subsys() of tipc into two parts, and split
tipc_socket_init() with initialization of pernet params.
By the way, I fixed resources rollback error when tipc_bcast_init()
failed in tipc_init_net().
Fixes: 7e27e8d613 ("tipc: switch order of device registration to fix a crash")
Signed-off-by: Junwei Hu <hujunwei4@huawei.com>
Reported-by: Wang Wang <wangwang2@huawei.com>
Reported-by: syzbot+1e8114b61079bfe9cbc5@syzkaller.appspotmail.com
Reviewed-by: Kang Zhou <zhoukang7@huawei.com>
Reviewed-by: Suanming Mou <mousuanming@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Error message printed:
modprobe: ERROR: could not insert 'tipc': Address family not
supported by protocol.
when modprobe tipc after the following patch: switch order of
device registration, commit 7e27e8d613
("tipc: switch order of device registration to fix a crash")
Because sock_create_kern(net, AF_TIPC, ...) is called by
tipc_topsrv_create_listener() in the initialization process
of tipc_net_ops, tipc_socket_init() must be execute before that.
I move tipc_socket_init() into function tipc_init_net().
Fixes: 7e27e8d613
("tipc: switch order of device registration to fix a crash")
Signed-off-by: Junwei Hu <hujunwei4@huawei.com>
Reported-by: Wang Wang <wangwang2@huawei.com>
Reviewed-by: Kang Zhou <zhoukang7@huawei.com>
Reviewed-by: Suanming Mou <mousuanming@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When tipc is loaded while many processes try to create a TIPC socket,
a crash occurs:
PANIC: Unable to handle kernel paging request at virtual
address "dfff20000000021d"
pc : tipc_sk_create+0x374/0x1180 [tipc]
lr : tipc_sk_create+0x374/0x1180 [tipc]
Exception class = DABT (current EL), IL = 32 bits
Call trace:
tipc_sk_create+0x374/0x1180 [tipc]
__sock_create+0x1cc/0x408
__sys_socket+0xec/0x1f0
__arm64_sys_socket+0x74/0xa8
...
This is due to race between sock_create and unfinished
register_pernet_device. tipc_sk_insert tries to do
"net_generic(net, tipc_net_id)".
but tipc_net_id is not initialized yet.
So switch the order of the two to close the race.
This can be reproduced with multiple processes doing socket(AF_TIPC, ...)
and one process doing module removal.
Fixes: a62fbccecd ("tipc: make subscriber server support net namespace")
Signed-off-by: Junwei Hu <hujunwei4@huawei.com>
Reported-by: Wang Wang <wangwang2@huawei.com>
Reviewed-by: Xiaogang Wang <wangxiaogang3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As a preparation for introducing a smooth switching between replicast
and broadcast method for multicast message, We have to introduce a new
capability flag TIPC_MCAST_RBCTL to handle this new feature.
During a cluster upgrade a node can come back with this new capabilities
which also must be reflected in the cluster capabilities field.
The new feature is only applicable if all node in the cluster supports
this new capability.
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Synchronous pernet_operations are not allowed anymore.
All are asynchronous. So, drop the structure member.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a 32-bit node address is generated from a 128-bit identifier,
there is a risk of collisions which must be discovered and handled.
We do this as follows:
- We don't apply the generated address immediately to the node, but do
instead initiate a 1 sec trial period to allow other cluster members
to discover and handle such collisions.
- During the trial period the node periodically sends out a new type
of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast,
to all the other nodes in the cluster.
- When a node is receiving such a message, it must check that the
presented 32-bit identifier either is unused, or was used by the very
same peer in a previous session. In both cases it accepts the request
by not responding to it.
- If it finds that the same node has been up before using a different
address, it responds with a DSC_TRIAL_FAIL_MSG containing that
address.
- If it finds that the address has already been taken by some other
node, it generates a new, unused address and returns it to the
requester.
- During the trial period the requesting node must always be prepared
to accept a failure message, i.e., a message where a peer suggests a
different (or equal) address to the one tried. In those cases it
must apply the suggested value as trial address and restart the trial
period.
This algorithm ensures that in the vast majority of cases a node will
have the same address before and after a reboot. If a legacy user
configures the address explicitly, there will be no trial period and
messages, so this protocol addition is completely backwards compatible.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We add a 128-bit node identity, as an alternative to the currently used
32-bit node address.
For the sake of compatibility and to minimize message header changes
we retain the existing 32-bit address field. When not set explicitly by
the user, this field will be filled with a hash value generated from the
much longer node identity, and be used as a shorthand value for the
latter.
We permit either the address or the identity to be set by configuration,
but not both, so when the address value is set by a legacy user the
corresponding 128-bit node identity is generated based on the that value.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TIPC looks concentrated in itself, and other pernet_operations
seem not touching its entities.
tipc_net_ops look pernet-divided, and they should be safe to
be executed in parallel for several net the same time.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Make struct pernet_operations::id unsigned.
There are 2 reasons to do so:
1)
This field is really an index into an zero based array and
thus is unsigned entity. Using negative value is out-of-bound
access by definition.
2)
On x86_64 unsigned 32-bit data which are mixed with pointers
via array indexing or offsets added or subtracted to pointers
are preffered to signed 32-bit data.
"int" being used as an array index needs to be sign-extended
to 64-bit before being used.
void f(long *p, int i)
{
g(p[i]);
}
roughly translates to
movsx rsi, esi
mov rdi, [rsi+...]
call g
MOVSX is 3 byte instruction which isn't necessary if the variable is
unsigned because x86_64 is zero extending by default.
Now, there is net_generic() function which, you guessed it right, uses
"int" as an array index:
static inline void *net_generic(const struct net *net, int id)
{
...
ptr = ng->ptr[id - 1];
...
}
And this function is used a lot, so those sign extensions add up.
Patch snipes ~1730 bytes on allyesconfig kernel (without all junk
messing with code generation):
add/remove: 0/0 grow/shrink: 70/598 up/down: 396/-2126 (-1730)
Unfortunately some functions actually grow bigger.
This is a semmingly random artefact of code generation with register
allocator being used differently. gcc decides that some variable
needs to live in new r8+ registers and every access now requires REX
prefix. Or it is shifted into r12, so [r12+0] addressing mode has to be
used which is longer than [r8]
However, overall balance is in negative direction:
add/remove: 0/0 grow/shrink: 70/598 up/down: 396/-2126 (-1730)
function old new delta
nfsd4_lock 3886 3959 +73
tipc_link_build_proto_msg 1096 1140 +44
mac80211_hwsim_new_radio 2776 2808 +32
tipc_mon_rcv 1032 1058 +26
svcauth_gss_legacy_init 1413 1429 +16
tipc_bcbase_select_primary 379 392 +13
nfsd4_exchange_id 1247 1260 +13
nfsd4_setclientid_confirm 782 793 +11
...
put_client_renew_locked 494 480 -14
ip_set_sockfn_get 730 716 -14
geneve_sock_add 829 813 -16
nfsd4_sequence_done 721 703 -18
nlmclnt_lookup_host 708 686 -22
nfsd4_lockt 1085 1063 -22
nfs_get_client 1077 1050 -27
tcf_bpf_init 1106 1076 -30
nfsd4_encode_fattr 5997 5930 -67
Total: Before=154856051, After=154854321, chg -0.00%
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TIPC based clusters are by default set up with full-mesh link
connectivity between all nodes. Those links are expected to provide
a short failure detection time, by default set to 1500 ms. Because
of this, the background load for neighbor monitoring in an N-node
cluster increases with a factor N on each node, while the overall
monitoring traffic through the network infrastructure increases at
a ~(N * (N - 1)) rate. Experience has shown that such clusters don't
scale well beyond ~100 nodes unless we significantly increase failure
discovery tolerance.
This commit introduces a framework and an algorithm that drastically
reduces this background load, while basically maintaining the original
failure detection times across the whole cluster. Using this algorithm,
background load will now grow at a rate of ~(2 * sqrt(N)) per node, and
at ~(2 * N * sqrt(N)) in traffic overhead. As an example, each node will
now have to actively monitor 38 neighbors in a 400-node cluster, instead
of as before 399.
This "Overlapping Ring Supervision Algorithm" is completely distributed
and employs no centralized or coordinated state. It goes as follows:
- Each node makes up a linearly ascending, circular list of all its N
known neighbors, based on their TIPC node identity. This algorithm
must be the same on all nodes.
- The node then selects the next M = sqrt(N) - 1 nodes downstream from
itself in the list, and chooses to actively monitor those. This is
called its "local monitoring domain".
- It creates a domain record describing the monitoring domain, and
piggy-backs this in the data area of all neighbor monitoring messages
(LINK_PROTOCOL/STATE) leaving that node. This means that all nodes in
the cluster eventually (default within 400 ms) will learn about
its monitoring domain.
- Whenever a node discovers a change in its local domain, e.g., a node
has been added or has gone down, it creates and sends out a new
version of its node record to inform all neighbors about the change.
- A node receiving a domain record from anybody outside its local domain
matches this against its own list (which may not look the same), and
chooses to not actively monitor those members of the received domain
record that are also present in its own list. Instead, it relies on
indications from the direct monitoring nodes if an indirectly
monitored node has gone up or down. If a node is indicated lost, the
receiving node temporarily activates its own direct monitoring towards
that node in order to confirm, or not, that it is actually gone.
- Since each node is actively monitoring sqrt(N) downstream neighbors,
each node is also actively monitored by the same number of upstream
neighbors. This means that all non-direct monitoring nodes normally
will receive sqrt(N) indications that a node is gone.
- A major drawback with ring monitoring is how it handles failures that
cause massive network partitionings. If both a lost node and all its
direct monitoring neighbors are inside the lost partition, the nodes in
the remaining partition will never receive indications about the loss.
To overcome this, each node also chooses to actively monitor some
nodes outside its local domain. Those nodes are called remote domain
"heads", and are selected in such a way that no node in the cluster
will be more than two direct monitoring hops away. Because of this,
each node, apart from monitoring the member of its local domain, will
also typically monitor sqrt(N) remote head nodes.
- As an optimization, local list status, domain status and domain
records are marked with a generation number. This saves senders from
unnecessarily conveying unaltered domain records, and receivers from
performing unneeded re-adaptations of their node monitoring list, such
as re-assigning domain heads.
- As a measure of caution we have added the possibility to disable the
new algorithm through configuration. We do this by keeping a threshold
value for the cluster size; a cluster that grows beyond this value
will switch from full-mesh to ring monitoring, and vice versa when
it shrinks below the value. This means that if the threshold is set to
a value larger than any anticipated cluster size (default size is 32)
the new algorithm is effectively disabled. A patch set for altering the
threshold value and for listing the table contents will follow shortly.
- This change is fully backwards compatible.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are two flow control mechanisms in TIPC; one at link level that
handles network congestion, burst control, and retransmission, and one
at connection level which' only remaining task is to prevent overflow
in the receiving socket buffer. In TIPC, the latter task has to be
solved end-to-end because messages can not be thrown away once they
have been accepted and delivered upwards from the link layer, i.e, we
can never permit the receive buffer to overflow.
Currently, this algorithm is message based. A counter in the receiving
socket keeps track of number of consumed messages, and sends a dedicated
acknowledge message back to the sender for each 256 consumed message.
A counter at the sending end keeps track of the sent, not yet
acknowledged messages, and blocks the sender if this number ever reaches
512 unacknowledged messages. When the missing acknowledge arrives, the
socket is then woken up for renewed transmission. This works well for
keeping the message flow running, as it almost never happens that a
sender socket is blocked this way.
A problem with the current mechanism is that it potentially is very
memory consuming. Since we don't distinguish between small and large
messages, we have to dimension the socket receive buffer according
to a worst-case of both. I.e., the window size must be chosen large
enough to sustain a reasonable throughput even for the smallest
messages, while we must still consider a scenario where all messages
are of maximum size. Hence, the current fix window size of 512 messages
and a maximum message size of 66k results in a receive buffer of 66 MB
when truesize(66k) = 131k is taken into account. It is possible to do
much better.
This commit introduces an algorithm where we instead use 1024-byte
blocks as base unit. This unit, always rounded upwards from the
actual message size, is used when we advertise windows as well as when
we count and acknowledge transmitted data. The advertised window is
based on the configured receive buffer size in such a way that even
the worst-case truesize/msgsize ratio always is covered. Since the
smallest possible message size (from a flow control viewpoint) now is
1024 bytes, we can safely assume this ratio to be less than four, which
is the value we are now using.
This way, we have been able to reduce the default receive buffer size
from 66 MB to 2 MB with maintained performance.
In order to keep this solution backwards compatible, we introduce a
new capability bit in the discovery protocol, and use this throughout
the message sending/reception path to always select the right unit.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Nametable updates received from the network that cannot be applied
immediately are placed on a defer queue. This queue is global to the
TIPC module, which might cause problems when using TIPC in containers.
To prevent nametable updates from escaping into the wrong namespace,
we make the queue pernet instead.
Signed-off-by: Erik Hugne <erik.hugne@gmail.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The broadcast transmission link is currently instantiated when the
network subsystem is started, i.e., on order from user space via netlink.
This forces the broadcast transmission code to do unnecessary tests for
the existence of the transmission link, as well in single mode node as
in network mode.
In this commit, we do instead create the link during initialization of
the name space, and remove it when it is stopped. The fact that the
transmission link now has a guaranteed longer life cycle than any of its
potential clients paves the way for further code simplifcations
and optimizations.
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a topology server accepts a connection request from its client,
it allocates a connection instance and a tipc_subscriber structure
object. The former is used to communicate with client, and the latter
is often treated as a subscriber which manages all subscription events
requested from a same client. When a topology server receives a request
of subscribing name services from a client through the connection, it
creates a tipc_subscription structure instance which is seen as a
subscription recording what name services are subscribed. In order to
manage all subscriptions from a same client, topology server links
them into the subscrp_list of the subscriber. So subscriber and
subscription completely represents different meanings respectively,
but function names associated with them make us so confused that we
are unable to easily tell which function is against subscriber and
which is to subscription. So we want to eliminate the confusion by
renaming them.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When remove TIPC module, there is a warning to remind us that a slab
object is leaked like:
root@localhost:~# rmmod tipc
[ 19.056226] =============================================================================
[ 19.057549] BUG TIPC (Not tainted): Objects remaining in TIPC on kmem_cache_close()
[ 19.058736] -----------------------------------------------------------------------------
[ 19.058736]
[ 19.060287] INFO: Slab 0xffffea0000519a00 objects=23 used=1 fp=0xffff880014668b00 flags=0x100000000004080
[ 19.061915] INFO: Object 0xffff880014668000 @offset=0
[ 19.062717] kmem_cache_destroy TIPC: Slab cache still has objects
This is because the listening socket of TIPC topology server is not
closed before TIPC proto handler is unregistered with proto_unregister().
However, as the socket is closed in tipc_exit_net() which is called by
unregister_pernet_subsys() during unregistering TIPC namespace operation,
the warning can be eliminated if calling unregister_pernet_subsys() is
moved before calling proto_unregister().
Fixes: e05b31f4bf ("tipc: make tipc socket support net namespace")
Reviewed-by: Erik Hugne <erik.hugne@ericsson.com>
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add TIPC_CMD_NOOP to compat layer and remove the old framework.
All legacy nl commands are now converted to the compat layer in
netlink_compat.c.
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Erik Hugne <erik.hugne@ericsson.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The new netlink API is no longer "v2" but rather the standard API and
the legacy API is now "nl compat". We split them into separate
start/stop and put them in different files in order to further
distinguish them.
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Erik Hugne <erik.hugne@ericsson.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After namespace is supported, each namespace should own its private
random value. So the global variable representing the random value
must be moved to tipc_net structure.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Tested-by: Tero Aho <Tero.Aho@coriant.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TIPC establishes one subscriber server which allows users to subscribe
their interesting name service status. After tipc supports namespace,
one dedicated tipc stack instance is created for each namespace, and
each instance can be deemed as one independent TIPC node. As a result,
subscriber server must be built for each namespace.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Tested-by: Tero Aho <Tero.Aho@coriant.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If net namespace is supported in tipc, each namespace will be treated
as a separate tipc node. Therefore, every namespace must own its
private tipc node address. This means the "tipc_own_addr" global
variable of node address must be moved to tipc_net structure to
satisfy the requirement. It's turned out that users also can assign
node address for every namespace.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Tested-by: Tero Aho <Tero.Aho@coriant.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TIPC name table is used to store the mapping relationship between
TIPC service name and socket port ID. When tipc supports namespace,
it allows users to publish service names only owned by a certain
namespace. Therefore, every namespace must have its private name
table to prevent service names published to one namespace from being
contaminated by other service names in another namespace. Therefore,
The name table global variable (ie, nametbl) and its lock must be
moved to tipc_net structure, and a parameter of namespace must be
added for necessary functions so that they can obtain name table
variable defined in tipc_net structure.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Tested-by: Tero Aho <Tero.Aho@coriant.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now tipc socket table is statically allocated as a global variable.
Through it, we can look up one socket instance with port ID, insert
a new socket instance to the table, and delete a socket from the
table. But when tipc supports net namespace, each namespace must own
its specific socket table. So the global variable of socket table
must be redefined in tipc_net structure. As a concequence, a new
socket table will be allocated when a new namespace is created, and
a socket table will be deallocated when namespace is destroyed.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Tested-by: Tero Aho <Tero.Aho@coriant.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Global variables associated with node table are below:
- node table list (node_htable)
- node hash table list (tipc_node_list)
- node table lock (node_list_lock)
- node number counter (tipc_num_nodes)
- node link number counter (tipc_num_links)
To make node table support namespace, above global variables must be
moved to tipc_net structure in order to keep secret for different
namespaces. As a consequence, these variables are allocated and
initialized when namespace is created, and deallocated when namespace
is destroyed. After the change, functions associated with these
variables have to utilize a namespace pointer to access them. So
adding namespace pointer as a parameter of these functions is the
major change made in the commit.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Tested-by: Tero Aho <Tero.Aho@coriant.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Involve namespace infrastructure, make the "tipc_net_id" global
variable aware of per namespace, and rename it to "net_id". In
order that the conversion can be successfully done, an instance
of networking namespace must be passed to relevant functions,
allowing them to access the "net_id" variable of per namespace.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Tested-by: Tero Aho <Tero.Aho@coriant.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Only the works of initializing and shutting down tipc module are done
in core.h and core.c files, so all stuffs which are not closely
associated with the two tasks should be moved to appropriate places.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Tested-by: Tero Aho <Tero.Aho@coriant.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove redundant wrapper functions like tipc_core_start() and
tipc_core_stop(), and directly move them to their callers, such
as tipc_init() and tipc_exit(), having us clearly know what are
really done in both initialization and deinitialzation functions.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Tested-by: Tero Aho <Tero.Aho@coriant.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As tipc reference table is statically allocated, its memory size
requested on stack initialization stage is quite big even if the
maximum port number is just restricted to 8191 currently, however,
the number already becomes insufficient in practice. But if the
maximum ports is allowed to its theory value - 2^32, its consumed
memory size will reach a ridiculously unacceptable value. Apart from
this, heavy tipc users spend a considerable amount of time in
tipc_sk_get() due to the read-lock on ref_table_lock.
If tipc reference table is converted with generic rhashtable, above
mentioned both disadvantages would be resolved respectively: making
use of the new resizable hash table can avoid locking on the lookup;
smaller memory size is required at initial stage, for example, 256
hash bucket slots are requested at the beginning phase instead of
allocating the entire 8191 slots in old mode. The hash table will
grow if entries exceeds 75% of table size up to a total table size
of 1M, and it will automatically shrink if usage falls below 30%,
but the minimum table size is allowed down to 256.
Also converts ref_table_lock to a separate mutex to protect hash table
mutations on write side. Lastly defers the release of the socket
reference using call_rcu() to allow using an RCU read-side protected
call to rhashtable_lookup().
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Erik Hugne <erik.hugne@ericsson.com>
Cc: Thomas Graf <tgraf@suug.ch>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
The reference table is now 'socket aware' instead of being generic,
and has in reality become a socket internal table. In order to be
able to minimize the API exposed by the socket layer towards the rest
of the stack, we now move the reference table definitions and functions
into the file socket.c, and rename the functions accordingly.
There are no functional changes in this commit.
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Reviewed-by: Erik Hugne <erik.hugne@ericsson.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We move the inline functions in the file port.h to socket.c, and modify
their names accordingly.
We move struct tipc_port and some macros to socket.h.
Finally, we remove the file port.h.
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Reviewed-by: Erik Hugne <erik.hugne@ericsson.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Memory overhead when allocating big buffers for data transfer may
be quite significant. E.g., truesize of a 64 KB buffer turns out
to be 132 KB, 2 x the requested size.
This invalidates the "worst case" calculation we have been
using to determine the default socket receive buffer limit,
which is based on the assumption that 1024x64KB = 67MB buffers
may be queued up on a socket.
Since TIPC connections cannot survive hitting the buffer limit,
we have to compensate for this overhead.
We do that in this commit by dividing the fix connection flow
control window from 1024 (2*512) messages to 512 (2*256). Since
older version nodes send out acks at 512 message intervals,
compatibility with such nodes is guaranteed, although performance
may be non-optimal in such cases.
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In the previous commits of this series, we removed all asynchronous
actions which were based on the tasklet handler - "tipc_k_signal()".
So the moment has now come when we can completely remove the tasklet
handler infrastructure. That is done with this commit.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Erik Hugne <erik.hugne@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Due to the lacking of any credential, it's allowed to accept commands
requested from remote nodes to query the local node status, which is
prone to involve potential security risks. Instead, if we login to
a remote node with ssh command, this approach is not only more safe
than the remote management feature, but also it can give us more
permissions like changing the remote node configuration. So it's
reasonable for us to obsolete the remote management feature now.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Erik Hugne <erik.hugne@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/wireless/ath/ath9k/recv.c
drivers/net/wireless/mwifiex/pcie.c
net/ipv6/sit.c
The SIT driver conflict consists of a bug fix being done by hand
in 'net' (missing u64_stats_init()) whilst in 'net-next' a helper
was created (netdev_alloc_pcpu_stats()) which takes care of this.
The two wireless conflicts were overlapping changes.
Signed-off-by: David S. Miller <davem@davemloft.net>
Accidentally a side effect is involved by commit 6e967adf7(tipc:
relocate common functions from media to bearer). Now tipc stack
handler of receiving packets from netdevices as well as netdevice
notification handler are registered when bearer is enabled rather
than tipc module initialization stage, but the two handlers are
both unregistered in tipc module exit phase. If tipc module is
inserted and then immediately removed, the following warning
message will appear:
"dev_remove_pack: ffffffffa0380940 not found"
This is because in module insertion stage tipc stack packet handler
is not registered at all, but in module exit phase dev_remove_pack()
needs to remove it. Of course, dev_remove_pack() cannot find tipc
protocol handler from the kernel protocol handler list so that the
warning message is printed out.
But if registering the two handlers is adjusted from enabling bearer
phase into inserting module stage, the warning message will be
eliminated. Due to this change, tipc_core_start_net() and
tipc_core_stop_net() can be deleted as well.
Reported-by: Wang Weidong <wangweidong1@huawei.com>
Cc: Jon Maloy <jon.maloy@ericsson.com>
Cc: Erik Hugne <erik.hugne@ericsson.com>
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When tipc module is inserted, many tipc components are initialized
one by one. During the initialization period, if one of them is
failed, tipc_core_stop() will be called to stop all components
whatever corresponding components are created or not. To avoid to
release uncreated ones, relevant components have to add necessary
enabled flags indicating whether they are created or not.
But in the initialization stage, if one component is unsuccessfully
created, we will just destroy successfully created components before
the failed component instead of all components. All enabled flags
defined in components, in turn, become redundant. Additionally it's
also unnecessary to identify whether table.types is NULL in
tipc_nametbl_stop() because name stable has been definitely created
successfully when tipc_nametbl_stop() is called.
Cc: Jon Maloy <jon.maloy@ericsson.com>
Cc: Erik Hugne <erik.hugne@ericsson.com>
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In our ongoing effort to simplify the TIPC locking structure,
we see a need to remove the linked list for tipc_links
in the bearer. This can be explained as follows.
Currently, we have three different ways to access a link,
via three different lists/tables:
1: Via a node hash table:
Used by the time-critical outgoing/incoming data paths.
(e.g. link_send_sections_fast() and tipc_recv_msg() ):
grab net_lock(read)
find node from node hash table
grab node_lock
select link
grab bearer_lock
send_msg()
release bearer_lock
release node lock
release net_lock
2: Via a global linked list for nodes:
Used by configuration commands (link_cmd_set_value())
grab net_lock(read)
find node and link from global node list (using link name)
grab node_lock
update link
release node lock
release net_lock
(Same locking order as above. No problem.)
3: Via the bearer's linked link list:
Used by notifications from interface (e.g. tipc_disable_bearer() )
grab net_lock(write)
grab bearer_lock
get link ptr from bearer's link list
get node from link
grab node_lock
delete link
release node lock
release bearer_lock
release net_lock
(Different order from above, but works because we grab the
outer net_lock in write mode first, excluding all other access.)
The first major goal in our simplification effort is to get rid
of the "big" net_lock, replacing it with rcu-locks when accessing
the node list and node hash array. This will come in a later patch
series.
But to get there we first need to rewrite access methods ##2 and 3,
since removal of net_lock would introduce three major problems:
a) In access method #2, we access the link before taking the
protecting node_lock. This will not work once net_lock is gone,
so we will have to change the access order. We will deal with
this in a later commit in this series, "tipc: add node lock
protection to link found by link_find_link()".
b) When the outer protection from net_lock is gone, taking
bearer_lock and node_lock in opposite order of method 1) and 2)
will become an obvious deadlock hazard. This is fixed in the
commit ("tipc: remove bearer_lock from tipc_bearer struct")
later in this series.
c) Similar to what is described in problem a), access method #3
starts with using a link pointer that is unprotected by node_lock,
in order to via that pointer find the correct node struct and
lock it. Before we remove net_lock, this access order must be
altered. This is what we do with this commit.
We can avoid introducing problem problem c) by even here using the
global node list to find the node, before accessing its links. When
we loop though the node list we use the own bearer identity as search
criteria, thus easily finding the links that are associated to the
resetting/disabling bearer. It should be noted that although this
method is somewhat slower than the current list traversal, it is in
no way time critical. This is only about resetting or deleting links,
something that must be considered relatively infrequent events.
As a bonus, we can get rid of the mutual pointers between links and
bearers. After this commit, pointer dependency go in one direction
only: from the link to the bearer.
This commit pre-empts introduction of problem c) as described above.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
