The stack currently detects reordering and avoid spurious
retransmission very well. However the throughput is sub-optimal under
high reordering because cwnd is increased only if the data is deliverd
in order. I.e., FLAG_DATA_ACKED check in tcp_ack().  The more packet
are reordered the worse the throughput is.

Therefore when reordering is proven high, cwnd should advance whenever
the data is delivered regardless of its ordering. If reordering is low,
conservatively advance cwnd only on ordered deliveries in Open state,
and retain cwnd in Disordered state (RFC5681).

Using netperf on a qdisc setup of 20Mbps BW and random RTT from 45ms
to 55ms (for reordering effect). This change increases TCP throughput
by 20 - 25% to near bottleneck BW.

A special case is the stretched ACK with new SACK and/or ECE mark.
For example, a receiver may receive an out of order or ECN packet with
unacked data buffered because of LRO or delayed ACK. The principle on
such an ACK is to advance cwnd on the cummulative acked part first,
then reduce cwnd in tcp_fastretrans_alert().
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>

Commit bd61224b (SolutionEngine7724: fix Ether
support) has a typo in the 'phy_interface' field name of the platform data which
causes build error -- fix it.
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 06a64f91 (SH7619: fix Ether support)  has
a typo in the 'phy_interface' field name of the platform data which causes build
error -- fix it.
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

For some reason, my PCIe RTL8111E onboard NIC on a GA-Z68X-UD3H-B3
motherboard reads as FFs when reading from MMIO with a block size
larger than 7. Therefore change to reading blocks of four bytes.

Ben Hutchings noted that the buffer is large enough to hold all
registers, so now all registers are read.
Signed-off-by: Peter Wu <lekensteyn@gmail.com>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates

This series contains updates to igb and e1000e.

Akeem provides 3 igb patches, the first resets the link when EEE is enabled
or disabled if the link is up.  His second patch changes a register read
which normally stores of the read value to "just-read" so that hardware
can accurately latch the register read.  Lastly, he adds rcu_lock to avoid
a possible race condition with igb_update_stats function.

Mitch provides a fix for SR-IOV, where MSI-X interrupts are required, so
make sure that MSI-X is enabled before allowing the user to turn on SR-IOV.

Alex's igb patch make it so that we limit the lower bound for max_frame_size
to the size of a standard Ethernet frame.  This allows for feature parity
with other Intel based drivers such as ixgbe.

Carolyn adds a SKU for a flashless i210 device and a fix for get_fw_version()
so that it works for all parts for igb.  In addition, she has 2 igb patches
to refactor NVM code to accommodate devices with no flash.  Lastly, she
adds code to check for the failure of pci_disable_link_state() to attempt
to work around a problem found with some systems.

Laura provides the remaining 2 igb patches.  One removing the hard-coded
value for the size of the RETA indirection table, and creates a macro instead
for the RETA indirection table.  The second adds the ethtool callbacks
necessary to change the RETA indirection table from userspace.

Bruce fixes a whitespace issue in a recent commit and resolves a jiffies
comparison warning by using time_after().

Li provides a fix for e1000e to avoid a kernel crash on shutdown by adding
one more check in e1000e_shutdown().  This is due to e1000e_shutdown()
trying to clear correctable errors on the upstream P2P bridge, when under
some cases we do not have the upstream P2P bridge.

v2:
 - fixed patch 11 conditional statement from < to <= based on feedback
   from Ben Hutchings
 - fixed patch 12 patch description (adding the commit summary) based
   on feedback from Sergei Shtylyov
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

WARNING:JIFFIES_COMPARISON: Comparing jiffies is almost always wrong;
prefer time_after, time_before and friends
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

While doing shutdown on the PCI device, the corresponding callback
function e1000e_shutdown() is trying to clear those correctable
errors on the upstream P2P bridge. Unfortunately, we don't have
the upstream P2P bridge under some cases (e.g. PCI-passthrou for
KVM on Power). That leads to kernel crash eventually.

The patch adds one more check on that to avoid kernel crash.
Signed-off-by: Li Zhang <zhlcindy@gmail.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch attempts to work around a problem found with some systems where
the call to pci_diable_link_state_locked() fails.  As a result, ASPM is not,
in fact, disabled.  Changing disable ASPM code to check if state actually
is disabled after the call and, if not, try another way to disable it.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Acked-by: Bruce W. Allan <bruce.w.allan@intel.com>
Tested-by: Pavel Machek <pavel@ucw.cz>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Commit (c96ddb0b e1000e: Use marco instead of digit for defining
e1000_rx_desc_packet_split) moved a define from one file to another but
missed using proper indentation/whitespace.

CC: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch adds the ethtool callbacks necessary to change the RETA
indirection table from userspace.

In order to achieve this, we add the indirection table field (rss_indir_tbl)
in the board specific data structure (struct igb_adapter) to preserve the
values across hardware resets.

The indirection table must be initialized with default values in the
following cases:
	* at module init time
	* when the number of RX queues changes.
For this reason we add a new field (rss_indir_tbl_init) in igb_adapter
that keeps track of the number of RX queues. Whenever the number of RX
queues changes, the rss_indir_tbl is modified and initialized with default
values. The rss_indir_tbl_init is updated accordingly.

CC: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Laura Mihaela Vasilescu <laura.vasilescu@rosedu.org>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

RETA indirection table is used to assign the received data to a CPU
in order to maintain an efficient distribution of network receive
processing across multiple CPUs.

This patch removes the hard-coded value for the size of the indirection
table and defines a new macro.
Signed-off-by: Laura Mihaela Vasilescu <laura.vasilescu@rosedu.org>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch fixes issues found with older parts and older NVM tools in the
display of the version in ethtool.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch adds the specific device id support for versions of i210 that do
not have flash installed.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch refactors NVM read functions in order to accommodate i210 devices
that do not have a flash.  Previously, this was not supported on i210
devices.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch refactors the init_nvm_params functions for 82575 and adds a new
function for the i210/i211 devices in order to configure separately the NVM
functionality for the i210/i211 family.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This change makes it so that we limit the lower bound for max_frame_size to
the size of a standard Ethernet frame.  This allows for feature parity with
other Intel based drivers such as ixgbe.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

MSI-X interrupts are required for SR-IOV operation. Check to make sure
they're enabled before allowing the user to turn on VFs.
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch adds rcu_lock to avoid possible race condition with igb_update_stats
function accessing the rings in free_ q_vector.

CC: Eric Dumazet <edumazet@google.com>
Signed-off-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch changes register read to "just-read" without returning a value
for hardware to accurately latch the register value.
Signed-off-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch resets the link, if link is up - whenever users enable or disable EEE
Signed-off-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Use the wrapper functions for getting and setting the driver data using
platform_device instead of using dev_set_drvdata() with &pdev->dev,
so we can directly pass a struct platform_device.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use the wrapper functions for getting and setting the driver data using
platform_device instead of using deva_set_drvdata() with &pdev->dev,
so we can directly pass a struct platform_device.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Marc Kleine-Budde says:

====================
this pull-request for net-next consists of a series by Alexander
Shiyan, he cleans up the mcp251x driver. As the first patch touches
arch/arm/mach-pxa, it's acked by Haojian Zhuang.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch removes the no_csum_insertion private parameter that is not used anymore
and, also,  the "likely" annotation from the condition that is not in a critical path.
Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Cc: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Unnecessary dev_set_drvdata() is removed, because the driver core
clears the driver data to NULL after device_release or on probe failure.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Unnecessary dev_set_drvdata() is removed, because the driver core
clears the driver data to NULL after device_release or on probe failure.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Unnecessary dev_set_drvdata() is removed, because the driver core
clears the driver data to NULL after device_release or on probe failure.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Acked-by: Michal Simek <monstr@monstr.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>

Unnecessary dev_set_drvdata() is removed, because the driver core
clears the driver data to NULL after device_release or on probe failure.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Unnecessary dev_set_drvdata() is removed, because the driver core
clears the driver data to NULL after device_release or on probe failure.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Unnecessary dev_set_drvdata() is removed, because the driver core
clears the driver data to NULL after device_release or on probe failure.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Unnecessary dev_set_drvdata() is removed, because the driver core
clears the driver data to NULL after device_release or on probe failure.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Unnecessary dev_set_drvdata() is removed, because the driver core
clears the driver data to NULL after device_release or on probe failure.
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pavel Emelyanov says:

====================
tun: Some bits required for tun's checkpoint-restore (v2)

After taking a closer look on tun checkpoint-restore I've found several
issues with the tun's API that make it impossible to dump and restore
the state of tun device and attached tun-files.

The proposed API changes are all about extending the existing ioctl-based
stuff. Patches fit today's net-next.

This v2 has David's comments about patch #1 fixed. All the rest is the same.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The only thing we may have from tun device is the fprog, whic contains
the number of filter elements and a pointer to (user-space) memory
where the elements are. The program itself may not be available if the
device is persistent and detached.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There's a small problem with sk-filters on tun devices. Consider
an application doing this sequence of steps:

fd = open("/dev/net/tun");
ioctl(fd, TUNSETIFF, { .ifr_name = "tun0" });
ioctl(fd, TUNATTACHFILTER, &my_filter);
ioctl(fd, TUNSETPERSIST, 1);
close(fd);

At that point the tun0 will remain in the system and will keep in
mind that there should be a socket filter at address '&my_filter'.

If after that we do

fd = open("/dev/net/tun");
ioctl(fd, TUNSETIFF, { .ifr_name = "tun0" });

we most likely receive the -EFAULT error, since tun_attach() would
try to connect the filter back. But (!) if we provide a filter at
address &my_filter, then tun0 will be created and the "new" filter
would be attached, but application may not know about that.

This may create certain problems to anyone using tun-s, but it's
critical problem for c/r -- if we meet a persistent tun device
with a filter in mind, we will not be able to attach to it to dump
its state (flags, owner, address, vnethdr size, etc.).

The proposal is to allow to attach to tun device (with TUNSETIFF)
w/o attaching the filter to the tun-file's socket. After this
attach app may e.g clean the device by dropping the filter, it
doesn't want to have one, or (in case of c/r) get information
about the device with tun ioctls.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Multiqueue tun devices allow to attach and detach from its queues
while keeping the interface itself set on file.

Knowing this is critical for the checkpoint part of criu project.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Tun devices cannot be created with ifidex user wants, but it's
required by checkpoint-restore project.

Long time ago such ability was implemented for rtnl_ops-based
interface for creating links (9c7dafbf net: Allow to create links
with given ifindex), but the only API for creating and managing
tuntap devices is ioctl-based and is evolving with adding new ones
(cde8b15f tuntap: add ioctl to attach or detach a file form tuntap
device).

Following that trend, here's how a new ioctl that sets the ifindex
for device, that _will_ be created by TUNSETIFF ioctl looks like.
So those who want a tuntap device with the ifindex N, should open
the tun device, call ioctl(fd, TUNSETIFINDEX, &N), then call TUNSETIFF.
If the index N is busy, then the register_netdev will find this out
and the ioctl would be failed with -EBUSY.

If setifindex is not called, then it will be generated as before.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Some variables could have their scope reduced.
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Result of skb_frag_dma_map() and dma_map_single() wasn't checked.
Added a check and proper handling in case of failure.
Moved the mapping to the beginning of mlx4_en_xmit(), before updating
the ring data structure to make error handling easier.
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When hardware gets into error state, must notify user about it.
When QP in error state no traffic will be tx'ed from the attached
tx_ring.

Driver should know how to recover from this unexpected state. I will send later
on the recovery flow, but having the print shouldn't be delayed.
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>