Commit 814d488c61 ("tcp: fix the timid additive increase on stretch
ACKs") fixed a bug where tcp_cong_avoid_ai() would either credit a
connection with an increase of snd_cwnd_cnt, or increase snd_cwnd, but
not both, resulting in cwnd increasing by 1 packet on at most every
alternate invocation of tcp_cong_avoid_ai().
Although the commit correctly implemented the CUBIC algorithm, which
can increase cwnd by as much as 1 packet per 1 packet ACKed (2x per
RTT), in practice that could be too aggressive: in tests on network
paths with small buffers, YouTube server retransmission rates nearly
doubled.
This commit restores CUBIC to a maximum cwnd growth rate of 1 packet
per 2 packets ACKed (1.5x per RTT). In YouTube tests this restored
retransmit rates to low levels.
Testing: This patch has been tested in datacenter netperf transfers
and live youtube.com and google.com servers.
Fixes: 9cd981dcf1 ("tcp: fix stretch ACK bugs in CUBIC")
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The recent change to tcp_cong_avoid_ai() to handle stretch ACKs
introduced a bug where snd_cwnd_cnt could accumulate a very large
value while w was large, and then if w was reduced snd_cwnd could be
incremented by a large delta, leading to a large burst and high packet
loss. This was tickled when CUBIC's bictcp_update() sets "ca->cnt =
100 * cwnd".
This bug crept in while preparing the upstream version of
814d488c61.
Testing: This patch has been tested in datacenter netperf transfers
and live youtube.com and google.com servers.
Fixes: 814d488c61 ("tcp: fix the timid additive increase on stretch ACKs")
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When reading from the error queue, msg_name and msg_control are only
populated for some errors. A new exception for empty timestamp skbs
added a false positive on icmp errors without payload.
`traceroute -M udpconn` only displayed gateways that return payload
with the icmp error: the embedded network headers are pulled before
sock_queue_err_skb, leaving an skb with skb->len == 0 otherwise.
Fix this regression by refining when msg_name and msg_control
branches are taken. The solutions for the two fields are independent.
msg_name only makes sense for errors that configure serr->port and
serr->addr_offset. Test the first instead of skb->len. This also fixes
another issue. saddr could hold the wrong data, as serr->addr_offset
is not initialized in some code paths, pointing to the start of the
network header. It is only valid when serr->port is set (non-zero).
msg_control support differs between IPv4 and IPv6. IPv4 only honors
requests for ICMP and timestamps with SOF_TIMESTAMPING_OPT_CMSG. The
skb->len test can simply be removed, because skb->dev is also tested
and never true for empty skbs. IPv6 honors requests for all errors
aside from local errors and timestamps on empty skbs.
In both cases, make the policy more explicit by moving this logic to
a new function that decides whether to process msg_control and that
optionally prepares the necessary fields in skb->cb[]. After this
change, the IPv4 and IPv6 paths are more similar.
The last case is rxrpc. Here, simply refine to only match timestamps.
Fixes: 49ca0d8bfa ("net-timestamp: no-payload option")
Reported-by: Jan Niehusmann <jan@gondor.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
----
Changes
v1->v2
- fix local origin test inversion in ip6_datagram_support_cmsg
- make v4 and v6 code paths more similar by introducing analogous
ipv4_datagram_support_cmsg
- fix compile bug in rxrpc
Signed-off-by: David S. Miller <davem@davemloft.net>
With some mss values, it is possible tcp_xmit_size_goal() puts
one segment more in TSO packet than tcp_tso_autosize().
We send then one TSO packet followed by one single MSS.
It is not a serious bug, but we can do slightly better, especially
for drivers using netif_set_gso_max_size() to lower gso_max_size.
Using same formula avoids these corner cases and makes
tcp_xmit_size_goal() a bit faster.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Fixes: 605ad7f184 ("tcp: refine TSO autosizing")
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ip_check_defrag() may be used by af_packet to defragment outgoing packets.
skb_network_offset() of af_packet's outgoing packets is not zero.
Signed-off-by: Alexander Drozdov <al.drozdov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
1. For an IPv4 ping socket, ping_check_bind_addr does not check
the family of the socket address that's passed in. Instead,
make it behave like inet_bind, which enforces either that the
address family is AF_INET, or that the family is AF_UNSPEC and
the address is 0.0.0.0.
2. For an IPv6 ping socket, ping_check_bind_addr returns EINVAL
if the socket family is not AF_INET6. Return EAFNOSUPPORT
instead, for consistency with inet6_bind.
3. Make ping_v4_sendmsg and ping_v6_sendmsg return EAFNOSUPPORT
instead of EINVAL if an incorrect socket address structure is
passed in.
4. Make IPv6 ping sockets be IPv6-only. The code does not support
IPv4, and it cannot easily be made to support IPv4 because
the protocol numbers for ICMP and ICMPv6 are different. This
makes connect(::ffff:192.0.2.1) fail with EAFNOSUPPORT instead
of making the socket unusable.
Among other things, this fixes an oops that can be triggered by:
int s = socket(AF_INET, SOCK_DGRAM, IPPROTO_ICMP);
struct sockaddr_in6 sin6 = {
.sin6_family = AF_INET6,
.sin6_addr = in6addr_any,
};
bind(s, (struct sockaddr *) &sin6, sizeof(sin6));
Change-Id: If06ca86d9f1e4593c0d6df174caca3487c57a241
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If an over-MTU UDP datagram is sent through a SOCK_RAW socket to a
UFO-capable device, ip_ufo_append_data() sets skb->ip_summed to
CHECKSUM_PARTIAL unconditionally as all GSO code assumes transport layer
checksum is to be computed on segmentation. However, in this case,
skb->csum_start and skb->csum_offset are never set as raw socket
transmit path bypasses udp_send_skb() where they are usually set. As a
result, driver may access invalid memory when trying to calculate the
checksum and store the result (as observed in virtio_net driver).
Moreover, the very idea of modifying the userspace provided UDP header
is IMHO against raw socket semantics (I wasn't able to find a document
clearly stating this or the opposite, though). And while allowing
CHECKSUM_NONE in the UFO case would be more efficient, it would be a bit
too intrusive change just to handle a corner case like this. Therefore
disallowing UFO for packets from SOCK_DGRAM seems to be the best option.
Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp_should_expand_sndbuf() does not expand the send buffer if we have
filled the congestion window.
However, it should use tcp_packets_in_flight() instead of
tp->packets_out to make this check.
Testing has established that the difference matters a lot if there are
many SACKed packets, causing a needless performance shortfall.
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
skb_copy_bits() returns zero on success and negative value on error,
so it is needed to invert the condition in ip_check_defrag().
Fixes: 1bf3751ec9 ("ipv4: ip_check_defrag must not modify skb before unsharing")
Signed-off-by: Alexander Drozdov <al.drozdov@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking updates from David Miller:
1) Missing netlink attribute validation in nft_lookup, from Patrick
McHardy.
2) Restrict ipv6 partial checksum handling to UDP, since that's the
only case it works for. From Vlad Yasevich.
3) Clear out silly device table sentinal macros used by SSB and BCMA
drivers. From Joe Perches.
4) Make sure the remote checksum code never creates a situation where
the remote checksum is applied yet the tunneling metadata describing
the remote checksum transformation is still present. Otherwise an
external entity might see this and apply the checksum again. From
Tom Herbert.
5) Use msecs_to_jiffies() where applicable, from Nicholas Mc Guire.
6) Don't explicitly initialize timer struct fields, use setup_timer()
and mod_timer() instead. From Vaishali Thakkar.
7) Don't invoke tg3_halt() without the tp->lock held, from Jun'ichi
Nomura.
8) Missing __percpu annotation in ipvlan driver, from Eric Dumazet.
9) Don't potentially perform skb_get() on shared skbs, also from Eric
Dumazet.
10) Fix COW'ing of metrics for non-DST_HOST routes in ipv6, from Martin
KaFai Lau.
11) Fix merge resolution error between the iov_iter changes in vhost and
some bug fixes that occurred at the same time. From Jason Wang.
12) If rtnl_configure_link() fails we have to perform a call to
->dellink() before unregistering the device. From WANG Cong.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (39 commits)
net: dsa: Set valid phy interface type
rtnetlink: call ->dellink on failure when ->newlink exists
com20020-pci: add support for eae single card
vhost_net: fix wrong iter offset when setting number of buffers
net: spelling fixes
net/core: Fix warning while make xmldocs caused by dev.c
net: phy: micrel: disable NAND-tree for KSZ8021, KSZ8031, KSZ8051, KSZ8081
ipv6: fix ipv6_cow_metrics for non DST_HOST case
openvswitch: Fix key serialization.
r8152: restore hw settings
hso: fix rx parsing logic when skb allocation fails
tcp: make sure skb is not shared before using skb_get()
bridge: netfilter: Move sysctl-specific error code inside #ifdef
ipv6: fix possible deadlock in ip6_fl_purge / ip6_fl_gc
ipvlan: add a missing __percpu pcpu_stats
tg3: Hold tp->lock before calling tg3_halt() from tg3_init_one()
bgmac: fix device initialization on Northstar SoCs (condition typo)
qlcnic: Delete existing multicast MAC list before adding new
net/mlx5_core: Fix configuration of log_uar_page_sz
sunvnet: don't change gso data on clones
...
Spelling errors caught by codespell.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
IPv6 can keep a copy of SYN message using skb_get() in
tcp_v6_conn_request() so that caller wont free the skb when calling
kfree_skb() later.
Therefore TCP fast open has to clone the skb it is queuing in
child->sk_receive_queue, as all skbs consumed from receive_queue are
freed using __kfree_skb() (ie assuming skb->users == 1)
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Fixes: 5b7ed0892f ("tcp: move fastopen functions to tcp_fastopen.c")
Signed-off-by: David S. Miller <davem@davemloft.net>
Move memcg_socket_limit_enabled decrement to tcp_destroy_cgroup (called
from memcg_destroy_kmem -> mem_cgroup_sockets_destroy) and zap a bunch of
wrapper functions.
Although this patch moves static keys decrement from __mem_cgroup_free to
mem_cgroup_css_free, it does not introduce any functional changes, because
the keys are incremented on setting the limit (tcp or kmem), which can
only happen after successful mem_cgroup_css_online.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Glauber Costa <glommer@parallels.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujtisu.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull security layer updates from James Morris:
"Highlights:
- Smack adds secmark support for Netfilter
- /proc/keys is now mandatory if CONFIG_KEYS=y
- TPM gets its own device class
- Added TPM 2.0 support
- Smack file hook rework (all Smack users should review this!)"
* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security: (64 commits)
cipso: don't use IPCB() to locate the CIPSO IP option
SELinux: fix error code in policydb_init()
selinux: add security in-core xattr support for pstore and debugfs
selinux: quiet the filesystem labeling behavior message
selinux: Remove unused function avc_sidcmp()
ima: /proc/keys is now mandatory
Smack: Repair netfilter dependency
X.509: silence asn1 compiler debug output
X.509: shut up about included cert for silent build
KEYS: Make /proc/keys unconditional if CONFIG_KEYS=y
MAINTAINERS: email update
tpm/tpm_tis: Add missing ifdef CONFIG_ACPI for pnp_acpi_device
smack: fix possible use after frees in task_security() callers
smack: Add missing logging in bidirectional UDS connect check
Smack: secmark support for netfilter
Smack: Rework file hooks
tpm: fix format string error in tpm-chip.c
char/tpm/tpm_crb: fix build error
smack: Fix a bidirectional UDS connect check typo
smack: introduce a special case for tmpfs in smack_d_instantiate()
...
The unified hierarchy interface for memory cgroups will no longer use "-1"
to mean maximum possible resource value. In preparation for this, make
the string an argument and let the caller supply it.
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: Vladimir Davydov <vdavydov@parallels.com>
Cc: Greg Thelen <gthelen@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Change remote checksum handling to set checksum partial as default
behavior. Added an iflink parameter to configure not using
checksum partial (calling csum_partial to update checksum).
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds infrastructure so that remote checksum offload can
set CHECKSUM_PARTIAL instead of calling csum_partial and writing
the modfied checksum field.
Add skb_remcsum_adjust_partial function to set an skb for using
CHECKSUM_PARTIAL with remote checksum offload. Changed
skb_remcsum_process and skb_gro_remcsum_process to take a boolean
argument to indicate if checksum partial can be set or the
checksum needs to be modified using the normal algorithm.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Properly set GSO types and skb->encapsulation in the UDP tunnel GRO
complete so that packets are properly represented for GSO. This sets
SKB_GSO_UDP_TUNNEL or SKB_GSO_UDP_TUNNEL_CSUM depending on whether
non-zero checksums were received, and sets SKB_GSO_TUNNEL_REMCSUM if
the remote checksum option was processed.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remote checksum offload processing is currently the same for both
the GRO and non-GRO path. When the remote checksum offload option
is encountered, the checksum field referred to is modified in
the packet. So in the GRO case, the packet is modified in the
GRO path and then the operation is skipped when the packet goes
through the normal path based on skb->remcsum_offload. There is
a problem in that the packet may be modified in the GRO path, but
then forwarded off host still containing the remote checksum option.
A remote host will again perform RCO but now the checksum verification
will fail since GRO RCO already modified the checksum.
To fix this, we ensure that GRO restores a packet to it's original
state before returning. In this model, when GRO processes a remote
checksum option it still changes the checksum per the algorithm
but on return from lower layer processing the checksum is restored
to its original value.
In this patch we add define gro_remcsum structure which is passed
to skb_gro_remcsum_process to save offset and delta for the checksum
being changed. After lower layer processing, skb_gro_remcsum_cleanup
is called to restore the checksum before returning from GRO.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Using the IPCB() macro to get the IPv4 options is convenient, but
unfortunately NetLabel often needs to examine the CIPSO option outside
of the scope of the IP layer in the stack. While historically IPCB()
worked above the IP layer, due to the inclusion of the inet_skb_param
struct at the head of the {tcp,udp}_skb_cb structs, recent commit
971f10ec ("tcp: better TCP_SKB_CB layout to reduce cache line misses")
reordered the tcp_skb_cb struct and invalidated this IPCB() trick.
This patch fixes the problem by creating a new function,
cipso_v4_optptr(), which locates the CIPSO option inside the IP header
without calling IPCB(). Unfortunately, this isn't as fast as a simple
lookup so some additional tweaks were made to limit the use of this
new function.
Cc: <stable@vger.kernel.org> # 3.18
Reported-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: Paul Moore <pmoore@redhat.com>
Tested-by: Casey Schaufler <casey@schaufler-ca.com>
Packetization Layer Path MTU Discovery works separately beside
Path MTU Discovery at IP level, different net namespace has
various requirements on which one to chose, e.g., a virutalized
container instance would require TCP PMTU to probe an usable
effective mtu for underlying tunnel, while the host would
employ classical ICMP based PMTU to function.
Hence making TCP PMTU mechanism per net namespace to decouple
two functionality. Furthermore the probe base MSS should also
be configured separately for each namespace.
Signed-off-by: Fan Du <fan.du@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If a server has enabled Fast Open and it receives a pure SYN-data
packet (without a Fast Open option), it won't accept the data but it
incorrectly returns a SYN-ACK with a Fast Open cookie and also
increments the SNMP stat LINUX_MIB_TCPFASTOPENPASSIVEFAIL.
This patch makes the server include a Fast Open cookie in SYN-ACK
only if the SYN has some Fast Open option (i.e., when client
requests or presents a cookie).
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Receive Flow Steering is a nice solution but suffers from
hash collisions when a mix of connected and unconnected traffic
is received on the host, when flow hash table is populated.
Also, clearing flow in inet_release() makes RFS not very good
for short lived flows, as many packets can follow close().
(FIN , ACK packets, ...)
This patch extends the information stored into global hash table
to not only include cpu number, but upper part of the hash value.
I use a 32bit value, and dynamically split it in two parts.
For host with less than 64 possible cpus, this gives 6 bits for the
cpu number, and 26 (32-6) bits for the upper part of the hash.
Since hash bucket selection use low order bits of the hash, we have
a full hash match, if /proc/sys/net/core/rps_sock_flow_entries is big
enough.
If the hash found in flow table does not match, we fallback to RPS (if
it is enabled for the rxqueue).
This means that a packet for an non connected flow can avoid the
IPI through a unrelated/victim CPU.
This also means we no longer have to clear the table at socket
close time, and this helps short lived flows performance.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
encap.sport and encap.dport are __be16, use nla_{get,put}_be16 instead
of nla_{get,put}_u16.
Fixes the sparse warnings:
warning: incorrect type in assignment (different base types)
expected restricted __be32 [addressable] [usertype] o_key
got restricted __be16 [addressable] [usertype] i_flags
warning: incorrect type in assignment (different base types)
expected restricted __be16 [usertype] sport
got unsigned short
warning: incorrect type in assignment (different base types)
expected restricted __be16 [usertype] dport
got unsigned short
warning: incorrect type in argument 3 (different base types)
expected unsigned short [unsigned] [usertype] value
got restricted __be16 [usertype] sport
warning: incorrect type in argument 3 (different base types)
expected unsigned short [unsigned] [usertype] value
got restricted __be16 [usertype] dport
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ensure that in state FIN_WAIT2 or TIME_WAIT, where the connection is
represented by a tcp_timewait_sock, we rate limit dupacks in response
to incoming packets (a) with TCP timestamps that fail PAWS checks, or
(b) with sequence numbers that are out of the acceptable window.
We do not send a dupack in response to out-of-window packets if it has
been less than sysctl_tcp_invalid_ratelimit (default 500ms) since we
last sent a dupack in response to an out-of-window packet.
Reported-by: Avery Fay <avery@mixpanel.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ensure that in state ESTABLISHED, where the connection is represented
by a tcp_sock, we rate limit dupacks in response to incoming packets
(a) with TCP timestamps that fail PAWS checks, or (b) with sequence
numbers or ACK numbers that are out of the acceptable window.
We do not send a dupack in response to out-of-window packets if it has
been less than sysctl_tcp_invalid_ratelimit (default 500ms) since we
last sent a dupack in response to an out-of-window packet.
There is already a similar (although global) rate-limiting mechanism
for "challenge ACKs". When deciding whether to send a challence ACK,
we first consult the new per-connection rate limit, and then the
global rate limit.
Reported-by: Avery Fay <avery@mixpanel.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In the SYN_RECV state, where the TCP connection is represented by
tcp_request_sock, we now rate-limit SYNACKs in response to a client's
retransmitted SYNs: we do not send a SYNACK in response to client SYN
if it has been less than sysctl_tcp_invalid_ratelimit (default 500ms)
since we last sent a SYNACK in response to a client's retransmitted
SYN.
This allows the vast majority of legitimate client connections to
proceed unimpeded, even for the most aggressive platforms, iOS and
MacOS, which actually retransmit SYNs 1-second intervals for several
times in a row. They use SYN RTO timeouts following the progression:
1,1,1,1,1,2,4,8,16,32.
Reported-by: Avery Fay <avery@mixpanel.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Helpers for mitigating ACK loops by rate-limiting dupacks sent in
response to incoming out-of-window packets.
This patch includes:
- rate-limiting logic
- sysctl to control how often we allow dupacks to out-of-window packets
- SNMP counter for cases where we rate-limited our dupack sending
The rate-limiting logic in this patch decides to not send dupacks in
response to out-of-window segments if (a) they are SYNs or pure ACKs
and (b) the remote endpoint is sending them faster than the configured
rate limit.
We rate-limit our responses rather than blocking them entirely or
resetting the connection, because legitimate connections can rely on
dupacks in response to some out-of-window segments. For example, zero
window probes are typically sent with a sequence number that is below
the current window, and ZWPs thus expect to thus elicit a dupack in
response.
We allow dupacks in response to TCP segments with data, because these
may be spurious retransmissions for which the remote endpoint wants to
receive DSACKs. This is safe because segments with data can't
realistically be part of ACK loops, which by their nature consist of
each side sending pure/data-less ACKs to each other.
The dupack interval is controlled by a new sysctl knob,
tcp_invalid_ratelimit, given in milliseconds, in case an administrator
needs to dial this upward in the face of a high-rate DoS attack. The
name and units are chosen to be analogous to the existing analogous
knob for ICMP, icmp_ratelimit.
The default value for tcp_invalid_ratelimit is 500ms, which allows at
most one such dupack per 500ms. This is chosen to be 2x faster than
the 1-second minimum RTO interval allowed by RFC 6298 (section 2, rule
2.4). We allow the extra 2x factor because network delay variations
can cause packets sent at 1 second intervals to be compressed and
arrive much closer.
Reported-by: Avery Fay <avery@mixpanel.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/vxlan.c
drivers/vhost/net.c
include/linux/if_vlan.h
net/core/dev.c
The net/core/dev.c conflict was the overlap of one commit marking an
existing function static whilst another was adding a new function.
In the include/linux/if_vlan.h case, the type used for a local
variable was changed in 'net', whereas the function got rewritten
to fix a stacked vlan bug in 'net-next'.
In drivers/vhost/net.c, Al Viro's iov_iter conversions in 'net-next'
overlapped with an endainness fix for VHOST 1.0 in 'net'.
In drivers/net/vxlan.c, vxlan_find_vni() added a 'flags' parameter
in 'net-next' whereas in 'net' there was a bug fix to pass in the
correct network namespace pointer in calls to this function.
Signed-off-by: David S. Miller <davem@davemloft.net>
When we added pacing to TCP, we decided to let sch_fq take care
of actual pacing.
All TCP had to do was to compute sk->pacing_rate using simple formula:
sk->pacing_rate = 2 * cwnd * mss / rtt
It works well for senders (bulk flows), but not very well for receivers
or even RPC :
cwnd on the receiver can be less than 10, rtt can be around 100ms, so we
can end up pacing ACK packets, slowing down the sender.
Really, only the sender should pace, according to its own logic.
Instead of adding a new bit in skb, or call yet another flow
dissection, we tweak skb->truesize to a small value (2), and
we instruct sch_fq to use new helper and not pace pure ack.
Note this also helps TCP small queue, as ack packets present
in qdisc/NIC do not prevent sending a data packet (RPC workload)
This helps to reduce tx completion overhead, ack packets can use regular
sock_wfree() instead of tcp_wfree() which is a bit more expensive.
This has no impact in the case packets are sent to loopback interface,
as we do not coalesce ack packets (were we would detect skb->truesize
lie)
In case netem (with a delay) is used, skb_orphan_partial() also sets
skb->truesize to 1.
This patch is a combination of two patches we used for about one year at
Google.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds skb_remcsum_process and skb_gro_remcsum_process to
perform the appropriate adjustments to the skb when receiving
remote checksum offload.
Updated vxlan and gue to use these functions.
Tested: Ran TCP_RR and TCP_STREAM netperf for VXLAN and GUE, did
not see any change in performance.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
That takes care of the majority of ->sendmsg() instances - most of them
via memcpy_to_msg() or assorted getfrag() callbacks. One place where we
still keep memcpy_fromiovecend() is tipc - there we potentially read the
same data over and over; separate patch, that...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
patch is actually smaller than it seems to be - most of it is unindenting
the inner loop body in tcp_sendmsg() itself...
the bit in tcp_input.c is going to get reverted very soon - that's what
memcpy_from_msg() will become, but not in this commit; let's keep it
reasonably contained...
There's one potentially subtle change here: in case of short copy from
userland, mainline tcp_send_syn_data() discards the skb it has allocated
and falls back to normal path, where we'll send as much as possible after
rereading the same data again. This patch trims SYN+data skb instead -
that way we don't need to copy from the same place twice.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
One deployment requirement of DCTCP is to be able to run
in a DC setting along with TCP traffic. As Glenn Judd's
NSDI'15 paper "Attaining the Promise and Avoiding the Pitfalls
of TCP in the Datacenter" [1] (tba) explains, one way to
solve this on switch side is to split DCTCP and TCP traffic
in two queues per switch port based on the DSCP: one queue
soley intended for DCTCP traffic and one for non-DCTCP traffic.
For the DCTCP queue, there's the marking threshold K as
explained in commit e3118e8359 ("net: tcp: add DCTCP congestion
control algorithm") for RED marking ECT(0) packets with CE.
For the non-DCTCP queue, there's f.e. a classic tail drop queue.
As already explained in e3118e8359, running DCTCP at scale
when not marking SYN/SYN-ACK packets with ECT(0) has severe
consequences as for non-ECT(0) packets, traversing the RED
marking DCTCP queue will result in a severe reduction of
connection probability.
This is due to the DCTCP queue being dominated by ECT(0) traffic
and switches handle non-ECT traffic in the RED marking queue
after passing K as drops, where K is usually a low watermark
in order to leave enough tailroom for bursts. Splitting DCTCP
traffic among several queues (ECN and non-ECN queue) is being
considered a terrible idea in the network community as it
splits single flows across multiple network paths.
Therefore, commit e3118e8359 implements this on Linux as
ECT(0) marked traffic, as we argue that marking all packets
of a DCTCP flow is the only viable solution and also doesn't
speak against the draft.
However, recently, a DCTCP implementation for FreeBSD hit also
their mainline kernel [2]. In order to let them play well
together with Linux' DCTCP, we would need to loosen the
requirement that ECT(0) has to be asserted during the 3WHS as
not implemented in FreeBSD. This simplifies the ECN test and
lets DCTCP work together with FreeBSD.
Joint work with Daniel Borkmann.
[1] https://www.usenix.org/conference/nsdi15/technical-sessions/presentation/judd
[2] 8ad8794452
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Glenn Judd <glenn.judd@morganstanley.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add timestamping option SOF_TIMESTAMPING_OPT_TSONLY. For transmit
timestamps, this loops timestamps on top of empty packets.
Doing so reduces the pressure on SO_RCVBUF. Payload inspection and
cmsg reception (aside from timestamps) are no longer possible. This
works together with a follow on patch that allows administrators to
only allow tx timestamping if it does not loop payload or metadata.
Signed-off-by: Willem de Bruijn <willemb@google.com>
----
Changes (rfc -> v1)
- add documentation
- remove unnecessary skb->len test (thanks to Richard Cochran)
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit be9f4a44e7 ("ipv4: tcp: remove per net tcp_sock")
I tried to address contention on a socket lock, but the solution
I chose was horrible :
commit 3a7c384ffd ("ipv4: tcp: unicast_sock should not land outside
of TCP stack") addressed a selinux regression.
commit 0980e56e50 ("ipv4: tcp: set unicast_sock uc_ttl to -1")
took care of another regression.
commit b5ec8eeac4 ("ipv4: fix ip_send_skb()") fixed another regression.
commit 811230cd85 ("tcp: ipv4: initialize unicast_sock sk_pacing_rate")
was another shot in the dark.
Really, just use a proper socket per cpu, and remove the skb_orphan()
call, to re-enable flow control.
This solves a serious problem with FQ packet scheduler when used in
hostile environments, as we do not want to allocate a flow structure
for every RST packet sent in response to a spoofed packet.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Get rid of nr_cpu_ids and use modern percpu allocation.
Note that the sockets themselves are not yet allocated
using NUMA affinity.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current behavior only passes RTTs from sequentially acked data to CC.
If sender gets a combined ACK for segment 1 and SACK for segment 3, then the
computed RTT for CC is the time between sending segment 1 and receiving SACK
for segment 3.
Pass the minimum computed RTT from any acked data to CC, i.e. time between
sending segment 3 and receiving SACK for segment 3.
Signed-off-by: Kenneth Klette Jonassen <kennetkl@ifi.uio.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
They are all either written once or extremly rarely (e.g. from init
code), so we can move them to the .data..read_mostly section.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
RFC 1191 said, "a host MUST not increase its estimate of the Path
MTU in response to the contents of a Datagram Too Big message."
Signed-off-by: Li Wei <lw@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When I added sk_pacing_rate field, I forgot to initialize its value
in the per cpu unicast_sock used in ip_send_unicast_reply()
This means that for sch_fq users, RST packets, or ACK packets sent
on behalf of TIME_WAIT sockets might be sent to slowly or even dropped
once we reach the per flow limit.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Fixes: 95bd09eb27 ("tcp: TSO packets automatic sizing")
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, it isn't possible to request checksums on the outer UDP
header of tunnels - the TUNNEL_CSUM flag is ignored. This adds
support for requesting that UDP checksums be computed on transmit
and properly reported if they are present on receive.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fixes a bug in CUBIC that causes cwnd to increase slightly
too slowly when multiple ACKs arrive in the same jiffy.
If cwnd is supposed to increase at a rate of more than once per jiffy,
then CUBIC was sometimes too slow. Because the bic_target is
calculated for a future point in time, calculated with time in
jiffies, the cwnd can increase over the course of the jiffy while the
bic_target calculated as the proper CUBIC cwnd at time
t=tcp_time_stamp+rtt does not increase, because tcp_time_stamp only
increases on jiffy tick boundaries.
So since the cnt is set to:
ca->cnt = cwnd / (bic_target - cwnd);
as cwnd increases but bic_target does not increase due to jiffy
granularity, the cnt becomes too large, causing cwnd to increase
too slowly.
For example:
- suppose at the beginning of a jiffy, cwnd=40, bic_target=44
- so CUBIC sets:
ca->cnt = cwnd / (bic_target - cwnd) = 40 / (44 - 40) = 40/4 = 10
- suppose we get 10 acks, each for 1 segment, so tcp_cong_avoid_ai()
increases cwnd to 41
- so CUBIC sets:
ca->cnt = cwnd / (bic_target - cwnd) = 41 / (44 - 41) = 41 / 3 = 13
So now CUBIC will wait for 13 packets to be ACKed before increasing
cwnd to 42, insted of 10 as it should.
The fix is to avoid adjusting the slope (determined by ca->cnt)
multiple times within a jiffy, and instead skip to compute the Reno
cwnd, the "TCP friendliness" code path.
Reported-by: Eyal Perry <eyalpe@mellanox.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change CUBIC to properly handle stretch ACKs in additive increase mode
by passing in the count of ACKed packets to tcp_cong_avoid_ai().
In addition, because we are now precisely accounting for stretch ACKs,
including delayed ACKs, we can now remove the delayed ACK tracking and
estimation code that tracked recent delayed ACK behavior in
ca->delayed_ack.
Reported-by: Eyal Perry <eyalpe@mellanox.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change Reno to properly handle stretch ACKs in additive increase mode
by passing in the count of ACKed packets to tcp_cong_avoid_ai().
In addition, if snd_cwnd crosses snd_ssthresh during slow start
processing, and we then exit slow start mode, we need to carry over
any remaining "credit" for packets ACKed and apply that to additive
increase by passing this remaining "acked" count to
tcp_cong_avoid_ai().
Reported-by: Eyal Perry <eyalpe@mellanox.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp_cong_avoid_ai() was too timid (snd_cwnd increased too slowly) on
"stretch ACKs" -- cases where the receiver ACKed more than 1 packet in
a single ACK. For example, suppose w is 10 and we get a stretch ACK
for 20 packets, so acked is 20. We ought to increase snd_cwnd by 2
(since acked/w = 20/10 = 2), but instead we were only increasing cwnd
by 1. This patch fixes that behavior.
Reported-by: Eyal Perry <eyalpe@mellanox.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Neal Cardwell