In preparation for adding Geneve Rx offload support, refactor the
current VXLAN offload flow to be a bit more generic so that it will be
easier to add the new Geneve code. The fm10k hardware supports one VXLAN
and one Geneve tunnel, so we will eventually treat the VXLAN and Geneve
tunnels identically. To this end, factor out the code that handles the
current list so that we can use the generic flow for both tunnels in the
next patch.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

In the event of a surprise remove, we expect the driver to go down,
which includes calling .stop_hw(). However, this function will return an
error because the queues won't appear to cleanly disable. Prevent this
and avoid the unnecessary checks by just returning when
FM10K_REMOVED(hw->hw_addr) is true.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

In the event of an uncorrectable AER error occurring when the driver has
not loaded, the recovery routines are not done. This is done because
future loads of the driver may not be aware of the IO state and may not
be able to recover at all. In this case, when we next load the driver it
fails due to what appears to be a surprise remove event. Instead, add
a check to ensure that the device is in the normal IO state before
continuing to probe. This allows us to give a more descriptive message
of what is wrong.

Without this change, the driver will attempt to probe up to our first
call of .reset_hw() which will be unable to read registers and act as if
a surprise remove event occurred.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

When fm10k_poll fully cleans rings it returns 0. This is incorrect as it
messes up the budget accounting in the core NAPI code. Fix this by
returning actual work done, capped at budget - 1 since the core doesn't
expect a return of the full budget when the driver modifies the NAPI
status.

Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Venkatesh Srinivas <venkateshs@google.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

While technically not needed, as all our uses of ACCESS_ONCE are scalar
types, we already use READ_ONCE in a few places, and for code
readability we can swap all the uses of the older ACCESS_ONCE into
READ_ONCE.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

The function is only used in fm10k_ethtool.c, so make it static.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

A previous patch added support to check for hardware Tx pending in the
fm10k_down routine. This support was intended to ensure that we
accurately check what the hardware state is. However, checking for Tx
hangs in this manor during the hotpath results in a large performance
hit. Avoid this by making the hotpath check use the SW counters instead.

Fixes: a0f53cf49cb0 ("fm10k: use actual hardware registers when checking for pending Tx", 2016-06-08)
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

A previous patch removed the pci_disable_device() call in
.io_error_detected. This call corresponded to a pci_enable_device_mem()
call within .io_slot_reset handler. Change the call here to
a pci_reenable_device() so that it does not increment and leak the
enable_cnt reference count for the device. Without this change, VF
devices may fail during an unbind/bind, and we'll never zero the
reference counter for the pci_dev structure.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

The addition of VLAN support caused a possible use of uninitialized
data if we encounter a zero TCA_FLOWER_KEY_ETH_TYPE key, as pointed
out by "gcc -Wmaybe-uninitialized":

net/sched/cls_flower.c: In function 'fl_change':
net/sched/cls_flower.c:366:22: error: 'ethertype' may be used uninitialized in this function [-Werror=maybe-uninitialized]

This changes the code to only set the ethertype field if it
was nonzero, as before the patch.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 9399ae9a ("net_sched: flower: Add vlan support")
Cc: Hadar Hen Zion <hadarh@mellanox.com>
Cc: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The newly added reset logic uses helper functions for the MMIO that
may fail. However, when the read operation fails, we end up writing
back uninitialized data to the register, as gcc warns:

drivers/net/ethernet/apm/xgene/xgene_enet_xgmac.c: In function 'xgene_enet_link_state':
drivers/net/ethernet/apm/xgene/xgene_enet_xgmac.c:213:2: error: 'data' may be used uninitialized in this function [-Werror=maybe-uninitialized]
drivers/net/ethernet/apm/xgene/xgene_enet_xgmac.c:209:6: note: 'data' was declared here
  u32 data;

We already print a warning to the console log if that happens,
the best alternative that I can see is skip the rest of the reset
sequence if the register value cannot be read: Most likely the
write would fail as well, and if it succeeded, worse things could
happen.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 3eb7cb9d ("drivers: net: xgene: XFI PCS reset when link is down")
Cc: Fushen Chen <fchen@apm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch enhances ethtool link mode bitmap to include
missing interface modes for 1G/10G speeds

Changes:
1000baseX is the mode introduced to cover all 1G Fiber cases.
All modes under 1000BaseX i.e. 1000BASE-SX, 1000BASE-LX, 1000BASE-LX10
and 1000BASE-BX10 are not explicitly defined at this moment.
10G CR,SR,LR and ER link modes are included for 10G speed..

Issue:
ethtool on  1G/10G SFP port reports Base-T
as this port supports 1000baseX,10G CR, SR and LR modes.

root@tor-02$ ethtool swp1
Settings for swp1:
        Supported ports: [ FIBRE ]
        Supported link modes:   1000baseT/Full
                                10000baseT/Full
        Supported pause frame use: Symmetric Receive-only
        Supports auto-negotiation: Yes
        Advertised link modes:  1000baseT/Full
        Advertised pause frame use: No
        Advertised auto-negotiation: No
        Speed: 10000Mb/s
        Duplex: Full
        Port: FIBRE
        PHYAD: 0
        Transceiver: external
        Auto-negotiation: off
        Current message level: 0x00000000 (0)

        Link detected: yes

After fix:
root@tor-02$ ethtool swp1
Settings for swp1:
        Supported ports: [ FIBRE ]
        Supported link modes:   1000baseX/Full
                                10000baseCR/Full
                                10000baseSR/Full
                                10000baseLR/Full
                                10000baseER/Full
        Supported pause frame use: Symmetric Receive-only
        Supports auto-negotiation: Yes
        Advertised link modes:  1000baseT/Full
        Advertised pause frame use: No
        Advertised auto-negotiation: No
        Speed: 10000Mb/s
        Duplex: Full
        Port: FIBRE
        PHYAD: 0
        Transceiver: external
        Auto-negotiation: off
        Current message level: 0x00000000 (0)
        Link detected: yes
Signed-off-by: Vidya Sagar Ravipati <vidya@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

After resume from hibernate on arm64, any amd-xgbe devices that were
running when we hibernated are reported as down, even when it is not.

Re-plugging the cables does not cause the interface to come back, the
link must be marked as down then up via 'ip set link' using the serial
console.

This happens because the device has been power-cycled and possibly
re-initialised by firmware, whereas the driver's memory structures have
been restored from the hibernate image and the two do not agree.

Schedule a restart of the device after powerup in case the world changed
while we were asleep.
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When TCP operates in lossy environments (between 1 and 10 % packet
losses), many SACK blocks can be exchanged, and I noticed we could
drop them on busy senders, if these SACK blocks have to be queued
into the socket backlog.

While the main cause is the poor performance of RACK/SACK processing,
we can try to avoid these drops of valuable information that can lead to
spurious timeouts and retransmits.

Cause of the drops is the skb->truesize overestimation caused by :

- drivers allocating ~2048 (or more) bytes as a fragment to hold an
  Ethernet frame.

- various pskb_may_pull() calls bringing the headers into skb->head
  might have pulled all the frame content, but skb->truesize could
  not be lowered, as the stack has no idea of each fragment truesize.

The backlog drops are also more visible on bidirectional flows, since
their sk_rmem_alloc can be quite big.

Let's add some room for the backlog, as only the socket owner
can selectively take action to lower memory needs, like collapsing
receive queues or partial ofo pruning.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Trivial fix to spelling mistake in dev_warn message.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Trivial fix to spelling mistake in dev_warn message.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Trivial fix to spelling mistake in dev_err message.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Tom Herbert says:

====================
strp: Generalize stream parser to work with other socket types

Add a read_sock protocol operation function that allows something like
tcp_read_sock to be called for other protocol types.

Specific changes in this patch set:
  - Add read_sock function to proto_ops. This has the same signature as
    tcp_read_sock. sk_read_actor_t is also defined in net.h.
  - Set peek_len and read_sock proto_op functions for TCPv4 and TCPv6
    stream ops.
  - Remove references to tcp in strparser.
  - Call peek_len and read_sock operations from strparser instead of
    calling TCP specific functions.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

kcm and strparser need to work with any type of stream socket not just
TCP. Eliminate references to TCP and call generic proto_ops functions of
read_sock and peek_len. Also in strp_init check if the socket support
the proto_ops read_sock and peek_len.
Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In inet_stream_ops we set read_sock to tcp_read_sock and peek_len to
tcp_peek_len (which is just a stub function that calls tcp_inq).
Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add new function in proto_ops structure. This includes moving the
typedef got sk_read_actor into net.h and removing the definition from
tcp.h.
Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix to return a negative error code from the cpsw_fill_rx_channels()
error handling case instead of 0, as done elsewhere in this function.

Fixes: ce52c744 ("net: ethernet: ti: cpsw: add ethtool channels support")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This header file is not used in vxlan.c file.
Signed-off-by: Zhu Yanjun <zyjzyj2000@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fixes the following sparse warning:

drivers/crypto/chelsio/chcr_algo.c:593:5: warning:
 symbol 'cxgb4_is_crypto_q_full' was not declared. Should it be static?
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

MC_CMD_TRIGGER_INTERRUPT does not work on the SFC9140, as used in the
sfn7x42q and sfn7x24f.
Check for this using the MCDI workaround mechanism.
The command is only used during self test.  If it's not supported, skip
the interrupt test.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

nic_data was already initialised to the right thing, no need to assign
it again.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

TX size bins were not supported on the 7000's 40G MAC, but the 8000 series
does support them and the MCPU advertises that via a new capability bit.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Richard Alpe says:

====================
tipc: introduce UDP replicast

This series introduces UDP replicast. A concept where we emulate multicast by
sending multiple unicast messages to configured peers. This allows TIPC to be
used in environments where IP multicast is disabled.

There is a corresponding patch series for the tipc user space tool that
allows a user to add remote addresses to the replicast list.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

When using replicast a UDP bearer can have an arbitrary amount of
remote ip addresses associated with it. This means we cannot simply
add all remote ip addresses to an existing bearer data message as it
might fill the message, leaving us with a truncated message that we
can't safely resume. To handle this we introduce the new netlink
command TIPC_NL_UDP_GET_REMOTEIP. This command is intended to be
called when the bearer data message has the
TIPC_NLA_UDP_MULTI_REMOTEIP flag set, indicating there are more than
one remote ip (replicast).
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add UDP bearer options to netlink bearer get message. This is used by
the tipc user space tool to display UDP options.

The UDP bearer information is passed using either a sockaddr_in or
sockaddr_in6 structs. This means the user space receiver should
intermediately store the retrieved data in a large enough struct
(sockaddr_strage) before casting to the proper IP version type.
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Automatically learn UDP remote IP addresses of communicating peers by
looking at the source IP address of incoming TIPC link configuration
messages (neighbor discovery).

This makes configuration slightly easier and removes the problematic
scenario where a node receives directly addressed neighbor discovery
messages sent using replicast which the node cannot "reply" to using
mutlicast, leaving the link FSM in a limbo state.
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch introduces UDP replicast. A concept where we emulate
multicast by sending multiple unicast messages to configured peers.

The purpose of replicast is mainly to be able to use TIPC in cloud
environments where IP multicast is disabled. Using replicas to unicast
multicast messages is costly as we have to copy each skb and send the
copies individually.
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add a function to check if a tipc UDP media address is a multicast
address or not. This is a purely cosmetic change.
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Split the UDP send function into two. One callback that prepares the
skb and one transmit function that sends the skb. This will come in
handy in later patches, when we introduce UDP replicast.
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Split the UDP netlink parse function so that it only parses one
netlink attribute at the time. This makes the parse function more
generic and allow future UDP API functions to use it for parsing.
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

And while at it, remove the unecessary writing of zeroes to the CPU_MASK_CLEAR
register since it has no functional use.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add tso/tso6 support to the alx driver.
Based on information from the downstream driver at github.com/qca/alx
Signed-off-by: Tobias Regnery <tobias.regnery@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Nelson Chang says:

====================
net: ethernet: mediatek: modify to use the PDMA for Ethernet RX

This series have some modifications and refines to support Ethernet RX by the PDMA.

changes since v4:
- Remove the redundant OR operation in mtk_hw_init()

changes since v3:
- Add GDM hardware settings to send packets to PDMA for RX

changes since v2:
- Fix the bugs of PDMA cpu index and interrupt settings in mtk_poll_rx()

changes since v1:
- Modify to use the PDMA instead of the QDMA for Ethernet RX
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Because we change to use the PDMA as the Ethernet RX DMA engine,
the patch modifies to set GDM to send packets to PDMA for RX.
Acked-by: John Crispin <john@phrozen.org>
Signed-off-by: Nelson Chang <nelson.chang@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>