[ Upstream commit 65550098c1c4db528400c73acf3e46bfa78d9264 ]
There's a race between rxrpc_sendmsg setting up a call, but then failing to
send anything on it due to an error, and recvmsg() seeing the call
completion occur and trying to return the state to the user.
An assertion fails in rxrpc_recvmsg() because the call has already been
released from the socket and is about to be released again as recvmsg deals
with it. (The recvmsg_q queue on the socket holds a ref, so there's no
problem with use-after-free.)
We also have to be careful not to end up reporting an error twice, in such
a way that both returns indicate to userspace that the user ID supplied
with the call is no longer in use - which could cause the client to
malfunction if it recycles the user ID fast enough.
Fix this by the following means:
(1) When sendmsg() creates a call after the point that the call has been
successfully added to the socket, don't return any errors through
sendmsg(), but rather complete the call and let recvmsg() retrieve
them. Make sendmsg() return 0 at this point. Further calls to
sendmsg() for that call will fail with ESHUTDOWN.
Note that at this point, we haven't send any packets yet, so the
server doesn't yet know about the call.
(2) If sendmsg() returns an error when it was expected to create a new
call, it means that the user ID wasn't used.
(3) Mark the call disconnected before marking it completed to prevent an
oops in rxrpc_release_call().
(4) recvmsg() will then retrieve the error and set MSG_EOR to indicate
that the user ID is no longer known by the kernel.
An oops like the following is produced:
kernel BUG at net/rxrpc/recvmsg.c:605!
...
RIP: 0010:rxrpc_recvmsg+0x256/0x5ae
...
Call Trace:
? __init_waitqueue_head+0x2f/0x2f
____sys_recvmsg+0x8a/0x148
? import_iovec+0x69/0x9c
? copy_msghdr_from_user+0x5c/0x86
___sys_recvmsg+0x72/0xaa
? __fget_files+0x22/0x57
? __fget_light+0x46/0x51
? fdget+0x9/0x1b
do_recvmmsg+0x15e/0x232
? _raw_spin_unlock+0xa/0xb
? vtime_delta+0xf/0x25
__x64_sys_recvmmsg+0x2c/0x2f
do_syscall_64+0x4c/0x78
entry_SYSCALL_64_after_hwframe+0x44/0xa9
Fixes: 357f5ef646 ("rxrpc: Call rxrpc_release_call() on error in rxrpc_new_client_call()")
Reported-by: syzbot+b54969381df354936d96@syzkaller.appspotmail.com
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b39a934ec72fa2b5a74123891f25273a38378b90 ]
The recent patch that substituted a flag on an rxrpc_call for the
connection pointer being NULL as an indication that a call was disconnected
puts the set_bit in the wrong place for service calls. This is only a
problem if a call is implicitly terminated by a new call coming in on the
same connection channel instead of a terminating ACK packet.
In such a case, rxrpc_input_implicit_end_call() calls
__rxrpc_disconnect_call(), which is now (incorrectly) setting the
disconnection bit, meaning that when rxrpc_release_call() is later called,
it doesn't call rxrpc_disconnect_call() and so the call isn't removed from
the peer's error distribution list and the list gets corrupted.
KASAN finds the issue as an access after release on a call, but the
position at which it occurs is confusing as it appears to be related to a
different call (the call site is where the latter call is being removed
from the error distribution list and either the next or pprev pointer
points to a previously released call).
Fix this by moving the setting of the flag from __rxrpc_disconnect_call()
to rxrpc_disconnect_call() in the same place that the connection pointer
was being cleared.
Fixes: 5273a191dca6 ("rxrpc: Fix NULL pointer deref due to call->conn being cleared on disconnect")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5273a191dca65a675dc0bcf3909e59c6933e2831 ]
When a call is disconnected, the connection pointer from the call is
cleared to make sure it isn't used again and to prevent further attempted
transmission for the call. Unfortunately, there might be a daemon trying
to use it at the same time to transmit a packet.
Fix this by keeping call->conn set, but setting a flag on the call to
indicate disconnection instead.
Remove also the bits in the transmission functions where the conn pointer is
checked and a ref taken under spinlock as this is now redundant.
Fixes: 8d94aa381d ("rxrpc: Calls shouldn't hold socket refs")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
rxrpc_put_*conn() calls trace_rxrpc_conn() after they have done the
decrement of the refcount - which looks at the debug_id in the connection
record. But unless the refcount was reduced to zero, we no longer have the
right to look in the record and, indeed, it may be deleted by some other
thread.
Fix this by getting the debug_id out before decrementing the refcount and
then passing that into the tracepoint.
Fixes: 363deeab6d ("rxrpc: Add connection tracepoint and client conn state tracepoint")
Signed-off-by: David Howells <dhowells@redhat.com>
When a local endpoint is ceases to be in use, such as when the kafs module
is unloaded, the kernel will emit an assertion failure if there are any
outstanding client connections:
rxrpc: Assertion failed
------------[ cut here ]------------
kernel BUG at net/rxrpc/local_object.c:433!
and even beyond that, will evince other oopses if there are service
connections still present.
Fix this by:
(1) Removing the triggering of connection reaping when an rxrpc socket is
released. These don't actually clean up the connections anyway - and
further, the local endpoint may still be in use through another
socket.
(2) Mark the local endpoint as dead when we start the process of tearing
it down.
(3) When destroying a local endpoint, strip all of its client connections
from the idle list and discard the ref on each that the list was
holding.
(4) When destroying a local endpoint, call the service connection reaper
directly (rather than through a workqueue) to immediately kill off all
outstanding service connections.
(5) Make the service connection reaper reap connections for which the
local endpoint is marked dead.
Only after destroying the connections can we close the socket lest we get
an oops in a workqueue that's looking at a connection or a peer.
Fixes: 3d18cbb7fd ("rxrpc: Fix conn expiry timers")
Signed-off-by: David Howells <dhowells@redhat.com>
Tested-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license as published by
the free software foundation either version 2 of the license or at
your option any later version
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-or-later
has been chosen to replace the boilerplate/reference in 3029 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190527070032.746973796@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
rxrpc_extract_addr_from_skb() doesn't use the argument that points to the
local endpoint, so remove the argument.
Signed-off-by: David Howells <dhowells@redhat.com>
AF_RXRPC opens an IPv6 socket through which to send and receive network
packets, both IPv6 and IPv4. It currently turns AF_INET addresses into
AF_INET-as-AF_INET6 addresses based on an assumption that this was
necessary; on further inspection of the code, however, it turns out that
the IPv6 code just farms packets aimed at AF_INET addresses out to the IPv4
code.
Fix AF_RXRPC to use AF_INET addresses directly when given them.
Fixes: 7b674e390e ("rxrpc: Fix IPv6 support")
Signed-off-by: David Howells <dhowells@redhat.com>
Fix error distribution by immediately delivering the errors to all the
affected calls rather than deferring them to a worker thread. The problem
with the latter is that retries and things can happen in the meantime when we
want to stop that sooner.
To this end:
(1) Stop the error distributor from removing calls from the error_targets
list so that peer->lock isn't needed to synchronise against other adds
and removals.
(2) Require the peer's error_targets list to be accessed with RCU, thereby
avoiding the need to take peer->lock over distribution.
(3) Don't attempt to affect a call's state if it is already marked complete.
Signed-off-by: David Howells <dhowells@redhat.com>
Make the following changes to improve the robustness of the code that sets
up a new service call:
(1) Cache the rxrpc_sock struct obtained in rxrpc_data_ready() to do a
service ID check and pass that along to rxrpc_new_incoming_call().
This means that I can remove the check from rxrpc_new_incoming_call()
without the need to worry about the socket attached to the local
endpoint getting replaced - which would invalidate the check.
(2) Cache the rxrpc_peer struct, thereby allowing the peer search to be
done once. The peer is passed to rxrpc_new_incoming_call(), thereby
saving the need to repeat the search.
This also reduces the possibility of rxrpc_publish_service_conn()
BUG()'ing due to the detection of a duplicate connection, despite the
initial search done by rxrpc_find_connection_rcu() having turned up
nothing.
This BUG() shouldn't ever get hit since rxrpc_data_ready() *should* be
non-reentrant and the result of the initial search should still hold
true, but it has proven possible to hit.
I *think* this may be due to __rxrpc_lookup_peer_rcu() cutting short
the iteration over the hash table if it finds a matching peer with a
zero usage count, but I don't know for sure since it's only ever been
hit once that I know of.
Another possibility is that a bug in rxrpc_data_ready() that checked
the wrong byte in the header for the RXRPC_CLIENT_INITIATED flag
might've let through a packet that caused a spurious and invalid call
to be set up. That is addressed in another patch.
(3) Fix __rxrpc_lookup_peer_rcu() to skip peer records that have a zero
usage count rather than stopping and returning not found, just in case
there's another peer record behind it in the bucket.
(4) Don't search the peer records in rxrpc_alloc_incoming_call(), but
rather either use the peer cached in (2) or, if one wasn't found,
preemptively install a new one.
Fixes: 8496af50eb ("rxrpc: Use RCU to access a peer's service connection tree")
Signed-off-by: David Howells <dhowells@redhat.com>
There's a check in rxrpc_data_ready() that's checking the CLIENT_INITIATED
flag in the packet type field rather than in the packet flags field.
Fix this by creating a pair of helper functions to check whether the packet
is going to the client or to the server and use them generally.
Fixes: 248f219cb8 ("rxrpc: Rewrite the data and ack handling code")
Signed-off-by: David Howells <dhowells@redhat.com>
rxrpc_find_connection_rcu() initialises variable k twice with the same
information. Remove one of the initialisations.
Signed-off-by: David Howells <dhowells@redhat.com>
While __atomic_add_unless() was originally intended as a building-block
for atomic_add_unless(), it's now used in a number of places around the
kernel. It's the only common atomic operation named __atomic*(), rather
than atomic_*(), and for consistency it would be better named
atomic_fetch_add_unless().
This lack of consistency is slightly confusing, and gets in the way of
scripting atomics. Given that, let's clean things up and promote it to
an official part of the atomics API, in the form of
atomic_fetch_add_unless().
This patch converts definitions and invocations over to the new name,
including the instrumented version, using the following script:
----
git grep -w __atomic_add_unless | while read line; do
sed -i '{s/\<__atomic_add_unless\>/atomic_fetch_add_unless/}' "${line%%:*}";
done
git grep -w __arch_atomic_add_unless | while read line; do
sed -i '{s/\<__arch_atomic_add_unless\>/arch_atomic_fetch_add_unless/}' "${line%%:*}";
done
----
Note that we do not have atomic{64,_long}_fetch_add_unless(), which will
be introduced by later patches.
There should be no functional change as a result of this patch.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Will Deacon <will.deacon@arm.com>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Palmer Dabbelt <palmer@sifive.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/lkml/20180621121321.4761-2-mark.rutland@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull networking updates from David Miller:
1) Support offloading wireless authentication to userspace via
NL80211_CMD_EXTERNAL_AUTH, from Srinivas Dasari.
2) A lot of work on network namespace setup/teardown from Kirill Tkhai.
Setup and cleanup of namespaces now all run asynchronously and thus
performance is significantly increased.
3) Add rx/tx timestamping support to mv88e6xxx driver, from Brandon
Streiff.
4) Support zerocopy on RDS sockets, from Sowmini Varadhan.
5) Use denser instruction encoding in x86 eBPF JIT, from Daniel
Borkmann.
6) Support hw offload of vlan filtering in mvpp2 dreiver, from Maxime
Chevallier.
7) Support grafting of child qdiscs in mlxsw driver, from Nogah
Frankel.
8) Add packet forwarding tests to selftests, from Ido Schimmel.
9) Deal with sub-optimal GSO packets better in BBR congestion control,
from Eric Dumazet.
10) Support 5-tuple hashing in ipv6 multipath routing, from David Ahern.
11) Add path MTU tests to selftests, from Stefano Brivio.
12) Various bits of IPSEC offloading support for mlx5, from Aviad
Yehezkel, Yossi Kuperman, and Saeed Mahameed.
13) Support RSS spreading on ntuple filters in SFC driver, from Edward
Cree.
14) Lots of sockmap work from John Fastabend. Applications can use eBPF
to filter sendmsg and sendpage operations.
15) In-kernel receive TLS support, from Dave Watson.
16) Add XDP support to ixgbevf, this is significant because it should
allow optimized XDP usage in various cloud environments. From Tony
Nguyen.
17) Add new Intel E800 series "ice" ethernet driver, from Anirudh
Venkataramanan et al.
18) IP fragmentation match offload support in nfp driver, from Pieter
Jansen van Vuuren.
19) Support XDP redirect in i40e driver, from Björn Töpel.
20) Add BPF_RAW_TRACEPOINT program type for accessing the arguments of
tracepoints in their raw form, from Alexei Starovoitov.
21) Lots of striding RQ improvements to mlx5 driver with many
performance improvements, from Tariq Toukan.
22) Use rhashtable for inet frag reassembly, from Eric Dumazet.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1678 commits)
net: mvneta: improve suspend/resume
net: mvneta: split rxq/txq init and txq deinit into SW and HW parts
ipv6: frags: fix /proc/sys/net/ipv6/ip6frag_low_thresh
net: bgmac: Fix endian access in bgmac_dma_tx_ring_free()
net: bgmac: Correctly annotate register space
route: check sysctl_fib_multipath_use_neigh earlier than hash
fix typo in command value in drivers/net/phy/mdio-bitbang.
sky2: Increase D3 delay to sky2 stops working after suspend
net/mlx5e: Set EQE based as default TX interrupt moderation mode
ibmvnic: Disable irqs before exiting reset from closed state
net: sched: do not emit messages while holding spinlock
vlan: also check phy_driver ts_info for vlan's real device
Bluetooth: Mark expected switch fall-throughs
Bluetooth: Set HCI_QUIRK_SIMULTANEOUS_DISCOVERY for BTUSB_QCA_ROME
Bluetooth: btrsi: remove unused including <linux/version.h>
Bluetooth: hci_bcm: Remove DMI quirk for the MINIX Z83-4
sh_eth: kill useless check in __sh_eth_get_regs()
sh_eth: add sh_eth_cpu_data::no_xdfar flag
ipv6: factorize sk_wmem_alloc updates done by __ip6_append_data()
ipv4: factorize sk_wmem_alloc updates done by __ip_append_data()
...
rxrpc_local objects cannot be disposed of until all the connections that
point to them have been RCU'd as a connection object holds refcount on the
local endpoint it is communicating through. Currently, this can cause an
assertion failure to occur when a network namespace is destroyed as there's
no check that the RCU destructors for the connections have been run before
we start trying to destroy local endpoints.
The kernel reports:
rxrpc: AF_RXRPC: Leaked local 0000000036a41bc1 {5}
------------[ cut here ]------------
kernel BUG at ../net/rxrpc/local_object.c:439!
Fix this by keeping a count of the live connections and waiting for it to
go to zero at the end of rxrpc_destroy_all_connections().
Fixes: dee46364ce ("rxrpc: Add RCU destruction for connections and calls")
Signed-off-by: David Howells <dhowells@redhat.com>
Fix various issues detected by checker.
Errors:
(*) rxrpc_discard_prealloc() should be using rcu_assign_pointer to set
call->socket.
Warnings:
(*) rxrpc_service_connection_reaper() should be passing NULL rather than 0 to
trace_rxrpc_conn() as the where argument.
(*) rxrpc_disconnect_client_call() should get its net pointer via the
call->conn rather than call->sock to avoid a warning about accessing
an RCU pointer without protection.
(*) Proc seq start/stop functions need annotation as they pass locks
between the functions.
False positives:
(*) Checker doesn't correctly handle of seq-retry lock context balance in
rxrpc_find_service_conn_rcu().
(*) Checker thinks execution may proceed past the BUG() in
rxrpc_publish_service_conn().
(*) Variable length array warnings from SKCIPHER_REQUEST_ON_STACK() in
rxkad.c.
Signed-off-by: David Howells <dhowells@redhat.com>
AF_RXRPC is incorrectly sending back to the server any abort it receives
for a client connection. This is due to the final-ACK offload to the
connection event processor patch. The abort code is copied into the
last-call information on the connection channel and then the event
processor is set.
Instead, the following should be done:
(1) In the case of a final-ACK for a successful call, the ACK should be
scheduled as before.
(2) In the case of a locally generated ABORT, the ABORT details should be
cached for sending in response to further packets related to that
call and no further action scheduled at call disconnect time.
(3) In the case of an ACK received from the peer, the call should be
considered dead, no ABORT should be transmitted at this time. In
response to further non-ABORT packets from the peer relating to this
call, an RX_USER_ABORT ABORT should be transmitted.
(4) In the case of a call killed due to network error, an RX_USER_ABORT
ABORT should be cached for transmission in response to further
packets, but no ABORT should be sent at this time.
Fixes: 3136ef49a1 ("rxrpc: Delay terminal ACK transmission on a client call")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix the rxrpc connection expiry timers so that connections for closed
AF_RXRPC sockets get deleted in a more timely fashion, freeing up the
transport UDP port much more quickly.
(1) Replace the delayed work items with work items plus timers so that
timer_reduce() can be used to shorten them and so that the timer
doesn't requeue the work item if the net namespace is dead.
(2) Don't use queue_delayed_work() as that won't alter the timeout if the
timer is already running.
(3) Don't rearm the timers if the network namespace is dead.
Signed-off-by: David Howells <dhowells@redhat.com>
RxRPC service endpoints expire like they're supposed to by the following
means:
(1) Mark dead rxrpc_net structs (with ->live) rather than twiddling the
global service conn timeout, otherwise the first rxrpc_net struct to
die will cause connections on all others to expire immediately from
then on.
(2) Mark local service endpoints for which the socket has been closed
(->service_closed) so that the expiration timeout can be much
shortened for service and client connections going through that
endpoint.
(3) rxrpc_put_service_conn() needs to schedule the reaper when the usage
count reaches 1, not 0, as idle conns have a 1 count.
(4) The accumulator for the earliest time we might want to schedule for
should be initialised to jiffies + MAX_JIFFY_OFFSET, not ULONG_MAX as
the comparison functions use signed arithmetic.
(5) Simplify the expiration handling, adding the expiration value to the
idle timestamp each time rather than keeping track of the time in the
past before which the idle timestamp must go to be expired. This is
much easier to read.
(6) Ignore the timeouts if the net namespace is dead.
(7) Restart the service reaper work item rather the client reaper.
Signed-off-by: David Howells <dhowells@redhat.com>
Delay terminal ACK transmission on a client call by deferring it to the
connection processor. This allows it to be skipped if we can send the next
call instead, the first DATA packet of which will implicitly ack this call.
Signed-off-by: David Howells <dhowells@redhat.com>
Fix IPv6 support in AF_RXRPC in the following ways:
(1) When extracting the address from a received IPv4 packet, if the local
transport socket is open for IPv6 then fill out the sockaddr_rxrpc
struct for an IPv4-mapped-to-IPv6 AF_INET6 transport address instead
of an AF_INET one.
(2) When sending CHALLENGE or RESPONSE packets, the transport length needs
to be set from the sockaddr_rxrpc::transport_len field rather than
sizeof() on the IPv4 transport address.
(3) When processing an IPv4 ICMP packet received by an IPv6 socket, set up
the address correctly before searching for the affected peer.
Signed-off-by: David Howells <dhowells@redhat.com>
Cache the congestion window setting that was determined during a call's
transmission phase when it finishes so that it can be used by the next call
to the same peer, thereby shortcutting the slow-start algorithm.
The value is stored in the rxrpc_peer struct and is accessed without
locking. Each call takes the value that happens to be there when it starts
and just overwrites the value when it finishes.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Keep the rxrpc_connection struct's idea of the service ID that is exposed
in the protocol separate from the service ID that's used as a lookup key.
This allows the protocol service ID on a client connection to get upgraded
without making the connection unfindable for other client calls that also
would like to use the upgraded connection.
The connection's actual service ID is then returned through recvmsg() by
way of msg_name.
Whilst we're at it, we get rid of the last_service_id field from each
channel. The service ID is per-connection, not per-call and an entire
connection is upgraded in one go.
Signed-off-by: David Howells <dhowells@redhat.com>
Support network namespacing in AF_RXRPC with the following changes:
(1) All the local endpoint, peer and call lists, locks, counters, etc. are
moved into the per-namespace record.
(2) All the connection tracking is moved into the per-namespace record
with the exception of the client connection ID tree, which is kept
global so that connection IDs are kept unique per-machine.
(3) Each namespace gets its own epoch. This allows each network namespace
to pretend to be a separate client machine.
(4) The /proc/net/rxrpc_xxx files are now called /proc/net/rxrpc/xxx and
the contents reflect the namespace.
fs/afs/ should be okay with this patch as it explicitly requires the current
net namespace to be init_net to permit a mount to proceed at the moment. It
will, however, need updating so that cells, IP addresses and DNS records are
per-namespace also.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the following extra tracing information:
(1) Modify the rxrpc_transmit tracepoint to record the Tx window size as
this is varied by the slow-start algorithm.
(2) Modify the rxrpc_rx_ack tracepoint to record more information from
received ACK packets.
(3) Add an rxrpc_rx_data tracepoint to record the information in DATA
packets.
(4) Add an rxrpc_disconnect_call tracepoint to record call disconnection,
including the reason the call was disconnected.
(5) Add an rxrpc_improper_term tracepoint to record implicit termination
of a call by a client either by starting a new call on a particular
connection channel without first transmitting the final ACK for the
previous call.
Signed-off-by: David Howells <dhowells@redhat.com>
Don't store the rxrpc protocol header in sk_buffs on the transmit queue,
but rather generate it on the fly and pass it to kernel_sendmsg() as a
separate iov. This reduces the amount of storage required.
Note that the security header is still stored in the sk_buff as it may get
encrypted along with the data (and doesn't change with each transmission).
Signed-off-by: David Howells <dhowells@redhat.com>
Add a pair of tracepoints, one to track rxrpc_connection struct ref
counting and the other to track the client connection cache state.
Signed-off-by: David Howells <dhowells@redhat.com>
Add CONFIG_AF_RXRPC_IPV6 and make the IPv6 support code conditional on it.
This is then made conditional on CONFIG_IPV6.
Without this, the following can be seen:
net/built-in.o: In function `rxrpc_init_peer':
>> peer_object.c:(.text+0x18c3c8): undefined reference to `ip6_route_output_flags'
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add IPv6 support to AF_RXRPC. With this, AF_RXRPC sockets can be created:
service = socket(AF_RXRPC, SOCK_DGRAM, PF_INET6);
instead of:
service = socket(AF_RXRPC, SOCK_DGRAM, PF_INET);
The AFS filesystem doesn't support IPv6 at the moment, though, since that
requires upgrades to some of the RPC calls.
Note that a good portion of this patch is replacing "%pI4:%u" in print
statements with "%pISpc" which is able to handle both protocols and print
the port.
Signed-off-by: David Howells <dhowells@redhat.com>
Rewrite the data and ack handling code such that:
(1) Parsing of received ACK and ABORT packets and the distribution and the
filing of DATA packets happens entirely within the data_ready context
called from the UDP socket. This allows us to process and discard ACK
and ABORT packets much more quickly (they're no longer stashed on a
queue for a background thread to process).
(2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead
keep track of the offset and length of the content of each packet in
the sk_buff metadata. This means we don't do any allocation in the
receive path.
(3) Jumbo DATA packet parsing is now done in data_ready context. Rather
than cloning the packet once for each subpacket and pulling/trimming
it, we file the packet multiple times with an annotation for each
indicating which subpacket is there. From that we can directly
calculate the offset and length.
(4) A call's receive queue can be accessed without taking locks (memory
barriers do have to be used, though).
(5) Incoming calls are set up from preallocated resources and immediately
made live. They can than have packets queued upon them and ACKs
generated. If insufficient resources exist, DATA packet #1 is given a
BUSY reply and other DATA packets are discarded).
(6) sk_buffs no longer take a ref on their parent call.
To make this work, the following changes are made:
(1) Each call's receive buffer is now a circular buffer of sk_buff
pointers (rxtx_buffer) rather than a number of sk_buff_heads spread
between the call and the socket. This permits each sk_buff to be in
the buffer multiple times. The receive buffer is reused for the
transmit buffer.
(2) A circular buffer of annotations (rxtx_annotations) is kept parallel
to the data buffer. Transmission phase annotations indicate whether a
buffered packet has been ACK'd or not and whether it needs
retransmission.
Receive phase annotations indicate whether a slot holds a whole packet
or a jumbo subpacket and, if the latter, which subpacket. They also
note whether the packet has been decrypted in place.
(3) DATA packet window tracking is much simplified. Each phase has just
two numbers representing the window (rx_hard_ack/rx_top and
tx_hard_ack/tx_top).
The hard_ack number is the sequence number before base of the window,
representing the last packet the other side says it has consumed.
hard_ack starts from 0 and the first packet is sequence number 1.
The top number is the sequence number of the highest-numbered packet
residing in the buffer. Packets between hard_ack+1 and top are
soft-ACK'd to indicate they've been received, but not yet consumed.
Four macros, before(), before_eq(), after() and after_eq() are added
to compare sequence numbers within the window. This allows for the
top of the window to wrap when the hard-ack sequence number gets close
to the limit.
Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also
to indicate when rx_top and tx_top point at the packets with the
LAST_PACKET bit set, indicating the end of the phase.
(4) Calls are queued on the socket 'receive queue' rather than packets.
This means that we don't need have to invent dummy packets to queue to
indicate abnormal/terminal states and we don't have to keep metadata
packets (such as ABORTs) around
(5) The offset and length of a (sub)packet's content are now passed to
the verify_packet security op. This is currently expected to decrypt
the packet in place and validate it.
However, there's now nowhere to store the revised offset and length of
the actual data within the decrypted blob (there may be a header and
padding to skip) because an sk_buff may represent multiple packets, so
a locate_data security op is added to retrieve these details from the
sk_buff content when needed.
(6) recvmsg() now has to handle jumbo subpackets, where each subpacket is
individually secured and needs to be individually decrypted. The code
to do this is broken out into rxrpc_recvmsg_data() and shared with the
kernel API. It now iterates over the call's receive buffer rather
than walking the socket receive queue.
Additional changes:
(1) The timers are condensed to a single timer that is set for the soonest
of three timeouts (delayed ACK generation, DATA retransmission and
call lifespan).
(2) Transmission of ACK and ABORT packets is effected immediately from
process-context socket ops/kernel API calls that cause them instead of
them being punted off to a background work item. The data_ready
handler still has to defer to the background, though.
(3) A shutdown op is added to the AF_RXRPC socket so that the AFS
filesystem can shut down the socket and flush its own work items
before closing the socket to deal with any in-progress service calls.
Future additional changes that will need to be considered:
(1) Make sure that a call doesn't hog the front of the queue by receiving
data from the network as fast as userspace is consuming it to the
exclusion of other calls.
(2) Transmit delayed ACKs from within recvmsg() when we've consumed
sufficiently more packets to avoid the background work item needing to
run.
Signed-off-by: David Howells <dhowells@redhat.com>
Make it possible for the data_ready handler called from the UDP transport
socket to completely instantiate an rxrpc_call structure and make it
immediately live by preallocating all the memory it might need. The idea
is to cut out the background thread usage as much as possible.
[Note that the preallocated structs are not actually used in this patch -
that will be done in a future patch.]
If insufficient resources are available in the preallocation buffers, it
will be possible to discard the DATA packet in the data_ready handler or
schedule a BUSY packet without the need to schedule an attempt at
allocation in a background thread.
To this end:
(1) Preallocate rxrpc_peer, rxrpc_connection and rxrpc_call structs to a
maximum number each of the listen backlog size. The backlog size is
limited to a maxmimum of 32. Only this many of each can be in the
preallocation buffer.
(2) For userspace sockets, the preallocation is charged initially by
listen() and will be recharged by accepting or rejecting pending
new incoming calls.
(3) For kernel services {,re,dis}charging of the preallocation buffers is
handled manually. Two notifier callbacks have to be provided before
kernel_listen() is invoked:
(a) An indication that a new call has been instantiated. This can be
used to trigger background recharging.
(b) An indication that a call is being discarded. This is used when
the socket is being released.
A function, rxrpc_kernel_charge_accept() is called by the kernel
service to preallocate a single call. It should be passed the user ID
to be used for that call and a callback to associate the rxrpc call
with the kernel service's side of the ID.
(4) Discard the preallocation when the socket is closed.
(5) Temporarily bump the refcount on the call allocated in
rxrpc_incoming_call() so that rxrpc_release_call() can ditch the
preallocation ref on service calls unconditionally. This will no
longer be necessary once the preallocation is used.
Note that this does not yet control the number of active service calls on a
client - that will come in a later patch.
A future development would be to provide a setsockopt() call that allows a
userspace server to manually charge the preallocation buffer. This would
allow user call IDs to be provided in advance and the awkward manual accept
stage to be bypassed.
Signed-off-by: David Howells <dhowells@redhat.com>
Condense the terminal states of a call state machine to a single state,
plus a separate completion type value. The value is then set, along with
error and abort code values, only when the call is transitioned to the
completion state.
Helpers are provided to simplify this.
Signed-off-by: David Howells <dhowells@redhat.com>
Improve the management and caching of client rxrpc connection objects.
From this point, client connections will be managed separately from service
connections because AF_RXRPC controls the creation and re-use of client
connections but doesn't have that luxury with service connections.
Further, there will be limits on the numbers of client connections that may
be live on a machine. No direct restriction will be placed on the number
of client calls, excepting that each client connection can support a
maximum of four concurrent calls.
Note that, for a number of reasons, we don't want to simply discard a
client connection as soon as the last call is apparently finished:
(1) Security is negotiated per-connection and the context is then shared
between all calls on that connection. The context can be negotiated
again if the connection lapses, but that involves holding up calls
whilst at least two packets are exchanged and various crypto bits are
performed - so we'd ideally like to cache it for a little while at
least.
(2) If a packet goes astray, we will need to retransmit a final ACK or
ABORT packet. To make this work, we need to keep around the
connection details for a little while.
(3) The locally held structures represent some amount of setup time, to be
weighed against their occupation of memory when idle.
To this end, the client connection cache is managed by a state machine on
each connection. There are five states:
(1) INACTIVE - The connection is not held in any list and may not have
been exposed to the world. If it has been previously exposed, it was
discarded from the idle list after expiring.
(2) WAITING - The connection is waiting for the number of client conns to
drop below the maximum capacity. Calls may be in progress upon it
from when it was active and got culled.
The connection is on the rxrpc_waiting_client_conns list which is kept
in to-be-granted order. Culled conns with waiters go to the back of
the queue just like new conns.
(3) ACTIVE - The connection has at least one call in progress upon it, it
may freely grant available channels to new calls and calls may be
waiting on it for channels to become available.
The connection is on the rxrpc_active_client_conns list which is kept
in activation order for culling purposes.
(4) CULLED - The connection got summarily culled to try and free up
capacity. Calls currently in progress on the connection are allowed
to continue, but new calls will have to wait. There can be no waiters
in this state - the conn would have to go to the WAITING state
instead.
(5) IDLE - The connection has no calls in progress upon it and must have
been exposed to the world (ie. the EXPOSED flag must be set). When it
expires, the EXPOSED flag is cleared and the connection transitions to
the INACTIVE state.
The connection is on the rxrpc_idle_client_conns list which is kept in
order of how soon they'll expire.
A connection in the ACTIVE or CULLED state must have at least one active
call upon it; if in the WAITING state it may have active calls upon it;
other states may not have active calls.
As long as a connection remains active and doesn't get culled, it may
continue to process calls - even if there are connections on the wait
queue. This simplifies things a bit and reduces the amount of checking we
need do.
There are a couple flags of relevance to the cache:
(1) EXPOSED - The connection ID got exposed to the world. If this flag is
set, an extra ref is added to the connection preventing it from being
reaped when it has no calls outstanding. This flag is cleared and the
ref dropped when a conn is discarded from the idle list.
(2) DONT_REUSE - The connection should be discarded as soon as possible and
should not be reused.
This commit also provides a number of new settings:
(*) /proc/net/rxrpc/max_client_conns
The maximum number of live client connections. Above this number, new
connections get added to the wait list and must wait for an active
conn to be culled. Culled connections can be reused, but they will go
to the back of the wait list and have to wait.
(*) /proc/net/rxrpc/reap_client_conns
If the number of desired connections exceeds the maximum above, the
active connection list will be culled until there are only this many
left in it.
(*) /proc/net/rxrpc/idle_conn_expiry
The normal expiry time for a client connection, provided there are
fewer than reap_client_conns of them around.
(*) /proc/net/rxrpc/idle_conn_fast_expiry
The expedited expiry time, used when there are more than
reap_client_conns of them around.
Note that I combined the Tx wait queue with the channel grant wait queue to
save space as only one of these should be in use at once.
Note also that, for the moment, the service connection cache still uses the
old connection management code.
Signed-off-by: David Howells <dhowells@redhat.com>
The main connection list is used for two independent purposes: primarily it
is used to find connections to reap and secondarily it is used to list
connections in procfs.
Split the procfs list out from the reap list. This allows us to stop using
the reap list for client connections when they acquire a separate
management strategy from service collections.
The client connections will not be on a management single list, and sometimes
won't be on a management list at all. This doesn't leave them floating,
however, as they will also be on an rb-tree rooted on the socket so that the
socket can find them to dispatch calls.
Signed-off-by: David Howells <dhowells@redhat.com>
Perform terminal call ACK/ABORT retransmission in the connection processor
rather than in the call processor. With this change, once last_call is
set, no more incoming packets will be routed to the corresponding call or
any earlier calls on that channel (call IDs must only increase on a channel
on a connection).
Further, if a packet's callNumber is before the last_call ID or a packet is
aimed at successfully completed service call then that packet is discarded
and ignored.
Signed-off-by: David Howells <dhowells@redhat.com>
Set the connection expiry time when a connection becomes idle rather than
doing this in rxrpc_put_connection(). This makes the put path more
efficient (it is likely to be called occasionally whilst a connection has
outstanding calls because active workqueue items needs to be given a ref).
The time is also preset in the connection allocator in case the connection
never gets used.
Signed-off-by: David Howells <dhowells@redhat.com>
Drop the channel number (channel) field from the rxrpc_call struct to
reduce the size of the call struct. The field is redundant: if the call is
attached to a connection, the channel can be obtained from there by AND'ing
with RXRPC_CHANNELMASK.
Signed-off-by: David Howells <dhowells@redhat.com>
Move to using RCU access to a peer's service connection tree when routing
an incoming packet. This is done using a seqlock to trigger retrying of
the tree walk if a change happened.
Further, we no longer get a ref on the connection looked up in the
data_ready handler unless we queue the connection's work item - and then
only if the refcount > 0.
Note that I'm avoiding the use of a hash table for service connections
because each service connection is addressed by a 62-bit number
(constructed from epoch and connection ID >> 2) that would allow the client
to engage in bucket stuffing, given knowledge of the hash algorithm.
Peers, however, are hashed as the network address is less controllable by
the client. The total number of peers will also be limited in a future
commit.
Signed-off-by: David Howells <dhowells@redhat.com>
Overhaul the usage count accounting for the rxrpc_connection struct to make
it easier to implement RCU access from the data_ready handler.
The problem is that currently we're using a lock to prevent the garbage
collector from trying to clean up a connection that we're contemplating
unidling. We could just stick incoming packets on the connection we find,
but we've then got a problem that we may race when dispatching a work item
to process it as we need to give that a ref to prevent the rxrpc_connection
struct from disappearing in the meantime.
Further, incoming packets may get discarded if attached to an
rxrpc_connection struct that is going away. Whilst this is not a total
disaster - the client will presumably resend - it would delay processing of
the call. This would affect the AFS client filesystem's service manager
operation.
To this end:
(1) We now maintain an extra count on the connection usage count whilst it
is on the connection list. This mean it is not in use when its
refcount is 1.
(2) When trying to reuse an old connection, we only increment the refcount
if it is greater than 0. If it is 0, we replace it in the tree with a
new candidate connection.
(3) Two connection flags are added to indicate whether or not a connection
is in the local's client connection tree (used by sendmsg) or the
peer's service connection tree (used by data_ready). This makes sure
that we don't try and remove a connection if it got replaced.
The flags are tested under lock with the removal operation to prevent
the reaper from killing the rxrpc_connection struct whilst someone
else is trying to effect a replacement.
This could probably be alleviated by using memory barriers between the
flag set/test and the rb_tree ops. The rb_tree op would still need to
be under the lock, however.
(4) When trying to reap an old connection, we try to flip the usage count
from 1 to 0. If it's not 1 at that point, then it must've come back
to life temporarily and we ignore it.
Signed-off-by: David Howells <dhowells@redhat.com>
Split the service-specific connection code out into into its own file. The
client-specific code has already been split out. This will leave just the
common code in the original file.
Signed-off-by: David Howells <dhowells@redhat.com>
Split the client-specific connection code out into its own file. It will
behave somewhat differently from the service-specific connection code, so
it makes sense to separate them.
Signed-off-by: David Howells <dhowells@redhat.com>
Each channel on a connection has a separate, independent number space from
which to allocate callNumber values. It is entirely possible, for example,
to have a connection with four active calls, each with call number 1.
Note that the callNumber values for any particular channel don't have to
start at 1, but they are supposed to increment monotonically for that
channel from a client's perspective and may not be reused once the call
number is transmitted (until the epoch cycles all the way back round).
Currently, however, call numbers are allocated on a per-connection basis
and, further, are held in an rb-tree. The rb-tree is redundant as the four
channel pointers in the rxrpc_connection struct are entirely capable of
pointing to all the calls currently in progress on a connection.
To this end, make the following changes:
(1) Handle call number allocation independently per channel.
(2) Get rid of the conn->calls rb-tree. This is overkill as a connection
may have a maximum of four calls in progress at any one time. Use the
pointers in the channels[] array instead, indexed by the channel
number from the packet.
(3) For each channel, save the result of the last call that was in
progress on that channel in conn->channels[] so that the final ACK or
ABORT packet can be replayed if necessary. Any call earlier than that
is just ignored. If we've seen the next call number in a packet, the
last one is most definitely defunct.
(4) When generating a RESPONSE packet for a connection, the call number
counter for each channel must be included in it.
(5) When parsing a RESPONSE packet for a connection, the call number
counters contained therein should be used to set the minimum expected
call numbers on each channel.
To do in future commits:
(1) Replay terminal packets based on the last call stored in
conn->channels[].
(2) Connections should be retired before the callNumber space on any
channel runs out.
(3) A server is expected to disregard or reject any new incoming call that
has a call number less than the current call number counter. The call
number counter for that channel must be advanced to the new call
number.
Note that the server cannot just require that the next call that it
sees on a channel be exactly the call number counter + 1 because then
there's a scenario that could cause a problem: The client transmits a
packet to initiate a connection, the network goes out, the server
sends an ACK (which gets lost), the client sends an ABORT (which also
gets lost); the network then reconnects, the client then reuses the
call number for the next call (it doesn't know the server already saw
the call number), but the server thinks it already has the first
packet of this call (it doesn't know that the client doesn't know that
it saw the call number the first time).
Signed-off-by: David Howells <dhowells@redhat.com>
Add RCU destruction for connections and calls as the RCU lookup from the
transport socket data_ready handler is going to come along shortly.
Whilst we're at it, move the cleanup workqueue flushing and RCU barrierage
into the destruction code for the objects that need it (locals and
connections) and add the extra RCU barrier required for connection cleanup.
Signed-off-by: David Howells <dhowells@redhat.com>
When a call is disconnected, clear the call's pointer to the connection and
release the associated ref on that connection. This means that the call no
longer pins the connection and the connection can be discarded even before
the call is.
As the code currently stands, the call struct is effectively pinned by
userspace until userspace has enacted a recvmsg() to retrieve the final
call state as sk_buffs on the receive queue pin the call to which they're
related because:
(1) The rxrpc_call struct contains the userspace ID that recvmsg() has to
include in the control message buffer to indicate which call is being
referred to. This ID must remain valid until the terminal packet is
completely read and must be invalidated immediately at that point as
userspace is entitled to immediately reuse it.
(2) The final ACK to the reply to a client call isn't sent until the last
data packet is entirely read (it's probably worth altering this in
future to be send the ACK as soon as all the data has been received).
This change requires a bit of rearrangement to make sure that the call
isn't going to try and access the connection again after protocol
completion:
(1) Delete the error link earlier when we're releasing the call. Possibly
network errors should be distributed via connections at the cost of
adding in an access to the rxrpc_connection struct.
(2) Remove the call from the connection's call tree before disconnecting
the call. The call tree needs to be removed anyway and incoming
packets delivered by channel pointer instead.
(3) The release call event should be considered last after all other
events have been processed so that we don't need access to the
connection again.
(4) Move the channel_lock taking from rxrpc_release_call() to
rxrpc_disconnect_call() where it will be required in future.
Signed-off-by: David Howells <dhowells@redhat.com>
Turn the connection event and state #define lists into enums and move
outside of the struct definition.
Whilst we're at it, change _SERVER to _SERVICE in those identifiers and add
EV_ into the event name to distinguish them from flags and states.
Also add a symbol indicating the number of states and use that in the state
text array.
Signed-off-by: David Howells <dhowells@redhat.com>
rxkad uses stack memory in SG lists which would not work if stacks were
allocated from vmalloc memory. In fact, in most cases this isn't even
necessary as the stack memory ends up getting copied over to kmalloc
memory.
This patch eliminates all the unnecessary stack memory uses by supplying
the final destination directly to the crypto API. In two instances where a
temporary buffer is actually needed we also switch use a scratch area in
the rxrpc_call struct (only one DATA packet will be being secured or
verified at a time).
Finally there is no need to split a split-page buffer into two SG entries
so code dealing with that has been removed.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Signed-off-by: David Howells <dhowells@redhat.com>
When looking up a client connection to which to route a packet, we need to
check that the packet came from the correct source so that a peer can't try
to muck around with another peer's connection.
Signed-off-by: David Howells <dhowells@redhat.com>
