Previously when the sender fails to send (original) data packet or
window probes due to congestion in the local host (e.g. throttling
in qdisc), it'll retry within an RTO or two up to 500ms.

In low-RTT networks such as data-centers, RTO is often far below
the default minimum 200ms. Then local host congestion could trigger
a retry storm pouring gas to the fire. Worse yet, the probe counter
(icsk_probes_out) is not properly updated so the aggressive retry
may exceed the system limit (15 rounds) until the packet finally
slips through.

On such rare events, it's wise to retry more conservatively
(500ms) and update the stats properly to reflect these incidents
and follow the system limit. Note that this is consistent with
the behaviors when a keep-alive probe or RTO retry is dropped
due to local congestion.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Previously when the sender fails to retransmit a data packet on
timeout due to congestion in the local host (e.g. throttling in
qdisc), it'll retry within an RTO up to 500ms.

In low-RTT networks such as data-centers, RTO is often far
below the default minimum 200ms (and the cap 500ms). Then local
host congestion could trigger a retry storm pouring gas to the
fire. Worse yet, the retry counter (icsk_retransmits) is not
properly updated so the aggressive retry may exceed the system
limit (15 rounds) until the packet finally slips through.

On such rare events, it's wise to retry more conservatively (500ms)
and update the stats properly to reflect these incidents and follow
the system limit. Note that this is consistent with the behavior
when a keep-alive probe is dropped due to local congestion.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Previously we use the next unsent skb's timestamp to determine
when to abort a socket stalling on window probes. This no longer
works as skb timestamp reflects the last instead of the first
transmission.

Instead we can estimate how long the socket has been stalling
with the probe count and the exponential backoff behavior.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Create a helper to model TCP exponential backoff for the next patch.
This is pure refactor w no behavior change.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch addresses a corner issue on timeout behavior of a
passive Fast Open socket.  A passive Fast Open server may write
and close the socket when it is re-trying SYN-ACK to complete
the handshake. After the handshake is completely, the server does
not properly stamp the recovery start time (tp->retrans_stamp is
0), and the socket may abort immediately on the very first FIN
timeout, instead of retying until it passes the system or user
specified limit.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Previously TCP socket's retrans_stamp is not set if the
retransmission has failed to send. As a result if a socket is
experiencing local issues to retransmit packets, determining when
to abort a socket is complicated w/o knowning the starting time of
the recovery since retrans_stamp may remain zero.

This complication causes sub-optimal behavior that TCP may use the
latest, instead of the first, retransmission time to compute the
elapsed time of a stalling connection due to local issues. Then TCP
may disrecard TCP retries settings and keep retrying until it finally
succeed: not a good idea when the local host is already strained.

The simple fix is to always timestamp the start of a recovery.
It's worth noting that retrans_stamp is also used to compare echo
timestamp values to detect spurious recovery. This patch does
not break that because retrans_stamp is still later than when the
original packet was sent.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Previously TCP skbs are not always timestamped if the transmission
failed due to memory or other local issues. This makes deciding
when to abort a socket tricky and complicated because the first
unacknowledged skb's timestamp may be 0 on TCP timeout.

The straight-forward fix is to always timestamp skb on every
transmission attempt. Also every skb retransmission needs to be
flagged properly to avoid RTT under-estimation. This can happen
upon receiving an ACK for the original packet and the a previous
(spurious) retransmission has failed.

It's worth noting that this reverts to the old time-stamping
style before commit 8c72c65b ("tcp: update skb->skb_mstamp more
carefully") which addresses a problem in computing the elapsed time
of a stalled window-probing socket. The problem will be addressed
differently in the next patches with a simpler approach.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Previously TCP only warns if its RTO timer fires and the
retransmission queue is empty, but it'll cause null pointer
reference later on. It's better to avoid such catastrophic failure
and simply exit with a warning.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This driver implements open-coded versions of phy_read_mmd() and
phy_write_mmd() for KSZ9031. That's not needed, let's use the
phylib functions directly.

This is compile-tested only because I have no such hardware.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

There is a plan to build the kernel with -Wimplicit-fallthrough and
this place in the code produced a warning (W=1).

This commit removes the following warning:

  include/linux/device.h:1480:5: warning: this statement may fall through [-Wimplicit-fallthrough=]
  drivers/net/ethernet/davicom/dm9000.c:397:3: note: in expansion of macro 'dev_dbg'
  drivers/net/ethernet/davicom/dm9000.c:398:2: note: here
Signed-off-by: Mathieu Malaterre <malat@debian.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

The udp-tunnel setup allows binding sockets to a network device. Prefer
the new SO_BINDTOIFINDEX to avoid temporarily resolving the device-name
just to look it up in the ioctl again.
Reviewed-by: Tom Gundersen <teg@jklm.no>
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The udp-tunnel setup allows binding sockets to a network device. Prefer
the new SO_BINDTOIFINDEX to avoid temporarily resolving the device-name
just to look it up in the ioctl again.
Reviewed-by: Tom Gundersen <teg@jklm.no>
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This introduces a new generic SOL_SOCKET-level socket option called
SO_BINDTOIFINDEX. It behaves similar to SO_BINDTODEVICE, but takes a
network interface index as argument, rather than the network interface
name.

User-space often refers to network-interfaces via their index, but has
to temporarily resolve it to a name for a call into SO_BINDTODEVICE.
This might pose problems when the network-device is renamed
asynchronously by other parts of the system. When this happens, the
SO_BINDTODEVICE might either fail, or worse, it might bind to the wrong
device.

In most cases user-space only ever operates on devices which they
either manage themselves, or otherwise have a guarantee that the device
name will not change (e.g., devices that are UP cannot be renamed).
However, particularly in libraries this guarantee is non-obvious and it
would be nice if that race-condition would simply not exist. It would
make it easier for those libraries to operate even in situations where
the device-name might change under the hood.

A real use-case that we recently hit is trying to start the network
stack early in the initrd but make it survive into the real system.
Existing distributions rename network-interfaces during the transition
from initrd into the real system. This, obviously, cannot affect
devices that are up and running (unless you also consider moving them
between network-namespaces). However, the network manager now has to
make sure its management engine for dormant devices will not run in
parallel to these renames. Particularly, when you offload operations
like DHCP into separate processes, these might setup their sockets
early, and thus have to resolve the device-name possibly running into
this race-condition.

By avoiding a call to resolve the device-name, we no longer depend on
the name and can run network setup of dormant devices in parallel to
the transition off the initrd. The SO_BINDTOIFINDEX ioctl plugs this
race.
Reviewed-by: Tom Gundersen <teg@jklm.no>
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This fixes recvmsg() to be able to peek across multiple tls records.
Without this patch, the tls's selftests test case
'recv_peek_large_buf_mult_recs' fails. Each tls receive context now
maintains a 'rx_list' to retain incoming skb carrying tls records. If a
tls record needs to be retained e.g. for peek case or for the case when
the buffer passed to recvmsg() has a length smaller than decrypted
record length, then it is added to 'rx_list'. Additionally, records are
added in 'rx_list' if the crypto operation runs in async mode. The
records are dequeued from 'rx_list' after the decrypted data is consumed
by copying into the buffer passed to recvmsg(). In case, the MSG_PEEK
flag is used in recvmsg(), then records are not consumed or removed
from the 'rx_list'.
Signed-off-by: Vakul Garg <vakul.garg@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Johan Hovold says:

====================
net: dsa: lantiq_gswip: probe fixes and remove cleanup

This series fix a few issues found through inspection when fixing up new
bad uses of of_find_compatible_node() that have crept in since 4.19.

Note that these have only been compile tested.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The platform-device driver data is set on successful probe and will
never be NULL on remove (or we have much bigger problems).
Signed-off-by: Johan Hovold <johan@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use the new of_get_compatible_child() helper to look up child nodes to
avoid ever matching non-child nodes elsewhere in the tree.

Also fix up the related struct device_node leaks.

Fixes: 14fceff4 ("net: dsa: Add Lantiq / Intel DSA driver for vrx200")
Cc: stable <stable@vger.kernel.org>	# 4.20
Cc: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johan Hovold <johan@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure to disable and deregister the switch on late probe errors to
avoid use-after-free when the device-resource-managed switch is freed.

Fixes: 14fceff4 ("net: dsa: Add Lantiq / Intel DSA driver for vrx200")
Cc: stable <stable@vger.kernel.org>	# 4.20
Cc: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johan Hovold <johan@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

The X2 family of NICs (based on the SFC9250) have additional
MTD partitions for firmware and configuration. This includes
partitions that are read-only.

The NICs also have extended versions of the NVRAM interface,
allowing more detailed status information to be returned.
Signed-off-by: Bert Kenward <bkenward@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

TLS test cases recv_partial & recv_peek_large_buf_mult_recs expect to
receive a certain amount of data and then compare it against known
strings using memcmp. To prevent recvmsg() from returning lesser than
expected number of bytes (compared in memcmp), MSG_WAITALL needs to be
passed in recvmsg().
Signed-off-by: Vakul Garg <vakul.garg@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When requesting the PHY driver module fails we'll bind the genphy
driver later. This isn't obvious to the user and may cause, depending
on the PHY, different types of issues. Therefore check the return code
of request_module(). Note that we only check for failures in loading
the module, not whether a module exists for the respective PHY ID.

v2:
- add comment explaining what is checked and what is not
- return error from phy_device_create() if loading module fails
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fixes the following sparse warning:

 net/tls/tls_sw.c:1023:5: warning:
 symbol 'tls_sw_do_sendpage' was not declared. Should it be static?
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There are no in-tree callers.
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Function sk_msg_clone has been modified to merge the data from source sg
entry to destination sg entry if the cloned data resides in same page
and is contiguous to the end entry of destination sk_msg. This improves
kernel tls throughput to the tune of 10%.

When the user space tls application calls sendmsg() with MSG_MORE, it leads
to calling sk_msg_clone() with new data being cloned placed continuous to
previously cloned data. Without this optimization, a new SG entry in
the destination sk_msg i.e. rec->msg_plaintext in tls_clone_plaintext_msg()
gets used. This leads to exhaustion of sg entries in rec->msg_plaintext
even before a full 16K of allowable record data is accumulated. Hence we
lose oppurtunity to encrypt and send a full 16K record.

With this patch, the kernel tls can accumulate full 16K of record data
irrespective of the size of data passed in sendmsg() with MSG_MORE.
Signed-off-by: Vakul Garg <vakul.garg@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

One of the more common cases of allocation size calculations is finding
the size of a structure that has a zero-sized array at the end, along
with memory for some number of elements for that array. For example:

struct foo {
    int stuff;
    struct boo entry[];
};

instance = devm_kzalloc(dev, sizeof(struct foo) + count * sizeof(struct boo), GFP_KERNEL);

Instead of leaving these open-coded and prone to type mistakes, we can
now use the new struct_size() helper:

instance = devm_kzalloc(dev, struct_size(instance, entry, count), GFP_KERNEL);

This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We are already checking in phy_detach() that the PHY driver is of
generic kind (1G or 10G) and we are going to make use of that in the SFP
layer as well for 1000BaseT SFP modules, so expose helper functions to
return that information.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Florian Fainelli says:

====================
net: dsa: Split platform data to header file

This patch series decouples the DSA platform data structures from
net/dsa.h which was getting used for all sorts of DSA related
structures.

It would probably make sense for this series to go via David's net-next
tree to avoid conflicts on the ARM part, since we cannot obviously
include a header that does not yet exist.

No functional changes intended.
====================
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

b53 and mv88e6xxx support passing platform_data, and now that we have
split the platform_data portion from the main net/dsa.h header file,
include only the relevant parts.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Now that we have split the DSA platform data structures from the main
net/dsa.h header file, include only the relevant header file.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Instead of having net/dsa.h contain both the internal switch tree/driver
structures, split the relevant platform_data parts into
include/linux/platform_data/dsa.h and make that header be included by
net/dsa.h in order not to break any setup. A subsequent set of patches
will update code including net/dsa.h to include only the platform_data
header.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In order to avoid frequent system interrupts when sending and
receiving packets. we replace disable_irq_nosync/enable_irq
with hinic_set_msix_state(), hinic_set_msix_state is used to
access memory mapped hinic devices.
Signed-off-by: Xue Chaojing <xuechaojing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There is not currently way to infer the port number through sysfs that
is being used as the CPU port number. Overlay a ndo_get_phys_port_name()
operation onto the DSA master network device in order to retrieve that
information.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Since 83c0afae ("net: dsa: Add new binding implementation"), DSA is
no longer a platform device exclusively and can support registering DSA
switches from other bus drivers (PCI, USB, I2C, etc.).
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

One of the more common cases of allocation size calculations is finding the
size of a structure that has a zero-sized array at the end, along with memory
for some number of elements for that array. For example:

struct foo {
    int stuff;
    struct boo entry[];
};

instance = kvzalloc(sizeof(struct foo) + count * sizeof(struct boo), GFP_KERNEL);

Instead of leaving these open-coded and prone to type mistakes, we can now
use the new struct_size() helper:

instance = kvzalloc(struct_size(instance, entry, count), GFP_KERNEL);

This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

One of the more common cases of allocation size calculations is finding the
size of a structure that has a zero-sized array at the end, along with memory
for some number of elements for that array. For example:

struct foo {
    int stuff;
    struct boo entry[];
};

instance = kzalloc(sizeof(struct foo) + count * sizeof(struct boo), GFP_KERNEL);

Instead of leaving these open-coded and prone to type mistakes, we can now
use the new struct_size() helper:

instance = kzalloc(struct_size(instance, entry, count), GFP_KERNEL);

This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There's no need to and one shouldn't include asm/irq.h directly.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Some time ago phydev_info() and friends have been added. They allow to
improve and simplify logging.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This workaround attempt helped for some but not all affected users.
With commit 11287b69 ("r8169: load Realtek PHY driver module
before r8169") we have a better workaround now, so we an remove
the first attempt.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Jakub Kicinski says:

====================
nfp: flower: improve flower resilience

This series contains mostly changes which improve nfp flower
offload's resilience, but are too large or risky to push into net.

Fred makes the driver waits for flower FW responses uninterruptible,
and a little longer (~40ms).

Pieter adds support for cards with multiple rule memories.

John reworks the MAC offloads.  He says:
> When potential tunnel end-point MACs are offloaded, they are assigned an
> index. This index may be associated with a port number meaning that if a
> packet matches an offloaded MAC address on the card, then the ingress
> port for that MAC can also be verified. In the case of shared MACs (e.g.
> on a linux bond) there may be situations where this index maps to only
> one of the ports that share the MAC.
>
> The idea of 'global' MAC indexes are supported that bypass the check on
> ingress port on the NFP. The patchset tracks shared MACs and assigns
> global indexes to these. It also ensures that port based indexes are
> re-applied if a single port becomes the only user of an offloaded MAC.
>
> Other patches in the set aim to tidy code without changing functionality.
> There is also a delete offload message introduced to ensure that MACs no
> longer in use in kernel space are removed from the firmware lookup tables.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

A MAC address is not necessarily a unique identifier for a netdev. Drivers
such as Linux bonds, for example, can apply the same MAC address to the
upper layer device and all lower layer devices.

NFP MAC offload for tunnel decap includes port verification for reprs but
also supports the offload of non-repr MAC addresses by assigning 'global'
indexes to these. This means that the FW will not verify the incoming port
of a packet matching this destination MAC.

Modify the MAC offload logic to assign global indexes based on MAC address
instead of net device (as it currently does). Use this to allow multiple
devices to share the same MAC. In other words, if a repr shares its MAC
address with another device then give the offloaded MAC a global index
rather than associate it with an ingress port. Track this so that changes
can be reverted as MACs stop being shared.

Implement this by removing the current list based assignment of global
indexes and replacing it with an rhashtable that maps an offloaded MAC
address to the number of devices sharing it, distributing global indexes
based on this.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>