The descriptor queues for sending (SDQs) and receiving (RDQs) packets
are managed by two counters - producer and consumer - which are both
16-bit in size. A queue is considered full when the difference between
the two equals the queue's maximum number of descriptors.

However, if the producer counter overflows, then it's possible for the
full queue check to fail, as it doesn't take the overflow into account.
In such a case, descriptors already passed to the device - but for which
a completion has yet to be posted - will be overwritten, thereby causing
undefined behavior. The above can be achieved under heavy load (~30
netperf instances).

Fix that by casting the subtraction result to u16, preventing it from
being treated as a signed integer.

Fixes: eda6500a ("mlxsw: Add PCI bus implementation")
Signed-off-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since we only support one VLAN filtering bridge we need to associate a
reference count with it, so that when the last port netdev leaves it, we
would know that a different bridge can be offloaded to hardware.

When a LAG device is memeber in a bridge and port netdevs are leaving
the LAG, we should always decrement the bridge's reference count, as it's
incremented for any port in the LAG.

Fixes: 4dc236c3 ("mlxsw: spectrum: Handle port leaving LAG while bridged")
Signed-off-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

priv->mdio->irq used to be allocated and required freeing, but it
is now a fixed sized array and should no longer be free'd.

Issue detected using static analysis with CoverityScan

Fixes: e7f4dc35 ("mdio: Move allocation of interrupts into core")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Reviewed-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

priv->mdio->irq used to be allocated and required freeing, but it
is now a fixed sized array and should no longer be free'd.

Issue detected using static analysis with CoverityScan

Fixes: e7f4dc35 ("mdio: Move allocation of interrupts into core")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Reviewed-by: NTobias Klauser <tklauser@distanz.ch>
Reviewed-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Used to be allocated and required freeing, but now
priv->mdio->irq is now a fixed sized array and should no longer be
free'd.

Issue detected using static analysis with CoverityScan

Fixes: e7f4dc35 ("mdio: Move allocation of interrupts into core")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Reviewed-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Enable workaround for MPC8548E erratum eTSEC 106,
"Excess delays when transmitting TOE=1 large frames".
(see commit 53fad773 "gianfar: Enable eTSEC-20 erratum w/a
for P2020 Rev1")

This erratum was fixed in Rev 3.1.x.
Signed-off-by: NAtsushi Nemoto <nemoto@toshiba-tops.co.jp>
Acked-by: NClaudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Only two bits (RX0OVR and RX1OVR) are writable in EFLG, write is useless
if these bits aren't set.
Signed-off-by: NEd Spiridonov <edo.rus@gmail.com>
Signed-off-by: NMarc Kleine-Budde <mkl@pengutronix.de>

When the interface is opened (in be_open()) the routine
be_enable_if_filters() must be called to switch on the basic filtering
capabilities of an interface that are not changed at run-time.
These include the flags UNTAGGED, BROADCAST and PASS_L3L4_ERRORS.
Other flags such as MULTICAST and PROMISC must be enabled later by
be_set_rx_mode() based on the state in the netdev/adapter struct.

be_enable_if_filters() routine is wrongly trying to enable MULTICAST flag
without checking the current adapter state. This can cause the RX_FILTER
cmds to the FW to fail.  This patch fixes this problem by only enabling
the basic filtering flags in be_enable_if_filters().

The VF must be able to issue RX_FILTER cmd with any filter flag, as long
as the PF allowed those flags (if_cap_flags) in the iface it provisioned
for the VF. This rule is applicable even when the VF doesn't have the
FILTMGMT privilege. There is a bug in BE3 FW that wrongly fails RX_FILTER
multicast programming cmds on VFs that don't have FILTMGMT privilege.
This patch also helps in insulating the VF driver from be_open failures due
to the FW bug. A fix for the BE3 FW issue will be available in
versions >= 11.0.283.0 and 10.6.334.0
Reported-by: NIvan Vecera <ivecera@redhat.com>
Signed-off-by: NVenkat Duvvuru <venkatkumar.duvvuru@avagotech.com>
Signed-off-by: NSathya Perla <sathya.perla@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We accidentally return IS_ERR(priv->base) which is 1 instead of
PTR_ERR(priv->base) which is the error code.

Fixes: 6c821bd9 ('net: Add MOXA ART SoCs ethernet driver')
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Acked-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When signalling to metadata consumers that the metadata_dst entry
carries additional GBP extension data for vxlan (TUNNEL_VXLAN_OPT),
the dst's vxlan_metadata information is populated, but options_len
is left to zero. F.e. in ovs, ovs_flow_key_extract() checks for
options_len before extracting the data through ip_tunnel_info_opts_get().

Geneve uses ip_tunnel_info_opts_set() helper in receive path, which
sets options_len internally, vxlan however uses ip_tunnel_info_opts(),
so when filling vxlan_metadata, we do need to update options_len.

Fixes: 4c222798 ("ip-tunnel: Use API to access tunnel metadata options.")
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Acked-by: NThomas Graf <tgraf@suug.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This change will also make Coverity happy by avoiding a theoretical NULL
pointer dereference; yet another reason is to use the above helper function
to tighten the code and make it more readable.
Signed-off-by: NWolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: NSimon Horman <horms+renesas@verge.net.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This change will also make Coverity happy by avoiding a theoretical NULL
pointer dereference; yet another reason is to use the above helper function
to tighten the code and make it more readable.
Signed-off-by: NWolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: NSimon Horman <horms+renesas@verge.net.au>
Acked-by: NGeert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The ibmvnic_capability struct was defined incorrectly. The last two
elements of the struct are in the wrong order.  In addition, the number
element should be 64-bit. Byteswapping functions are updated
as well.
Signed-off-by: NThomas Falcon <tlfalcon@linux.vnet.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The MC74xx and EM74xx modules use different IDs by default, according
to the Lenovo EM7455 driver for Windows.
Signed-off-by: NBjørn Mork <bjorn@mork.no>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently, in a case of error, dev_err is using fman->dev
before its initialization and "(NULL device *)" is printed.
This patch fixes this issue.
Signed-off-by: NIgal Liberman <igal.liberman@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch manages the case when you have an Ethernet MAC with
a "fixed link", and not connected to a normal MDIO-managed PHY device.

The test of phy_bus_name was not helpful because it was never affected
and replaced by the mdio test node.
Signed-off-by: NGabriel Fernandez <gabriel.fernandez@linaro.org>
Acked-by: NGiuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This reverts the changed init order from commit 3647bc35
("dwc_eth_qos: Reset hardware before PHY start") and makes another fix
for the race.

It turned out that the reset state machine of the dwceqos hardware
requires PHY clocks to be present in order to complete the reset
cycle.

To plug the race with the phy state machine we defer link speed
setting until the hardware init has finished.
Signed-off-by: NLars Persson <larper@axis.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since debug is hardcoded to 3, the defaults in the DWCEQOS_MSG_DEFAULT
macro are never used, which does not seem to be the intended behaviour
here.  Set debug to -1 like other drivers so that DWCEQOS_MSG_DEFAULT is
actually used by default.
Signed-off-by: NRabin Vincent <rabinv@axis.com>
Signed-off-by: NLars Persson <larper@axis.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since we are in non-atomic context here we can pass GFP_KERNEL to
dma_alloc_coherent(). This enables use of the CMA.
Signed-off-by: NRabin Vincent <rabinv@axis.com>
Signed-off-by: NLars Persson <larper@axis.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

To prepare for using the CMA, we can not be in atomic context when
de-allocating DMA buffers.

The tx lock was needed only to protect the hw reset against the xmit
handler. Now we briefly grab the tx lock while stopping the queue to
make sure no thread is inside or will enter the xmit handler.
Signed-off-by: NLars Persson <larper@axis.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The xmit handler and the tx_reclaim tasklet had a race on the tx_free
variable which could lead to a tx timeout if tx_free was updated after
the tx complete interrupt.
Signed-off-by: NRabin Vincent <rabinv@axis.com>
Signed-off-by: NLars Persson <larper@axis.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The VF administrative mac addresses (stored in the PF driver) are
initialized to zero when the PF driver starts up.

These addresses may be modified in the PF driver through ndo calls
initiated by iproute2 or libvirt.

While we allow the PF/host to change the VF admin mac address from zero
to a valid unicast mac, we do not allow restoring the VF admin mac to
zero. We currently only allow changing this mac to a different unicast mac.

This leads to problems when libvirt scripts are used to deal with
VF mac addresses, and libvirt attempts to revoke the mac so this
host will not use it anymore.

Fix this by allowing resetting a VF administrative MAC back to zero.

Fixes: 8f7ba3ca ('net/mlx4: Add set VF mac address support')
Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il>
Reported-by: NMoshe Levi <moshele@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The limit of 63 is only for virtual functions while the actual enforcement
was for VFs plus physical functions, fix that.

Fixes: e57968a1 ('net/mlx4_core: Support the HA mode for SRIOV VFs too')
Signed-off-by: NMoni Shoua <monis@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In the mac and vlan register/unregister/replace functions, the driver locks
the mac table mutex (or vlan table mutex) on both ports.

We move to use mutex_lock_nested() to prevent warnings, such as the one below.

[ 101.828445] =============================================
[ 101.834820] [ INFO: possible recursive locking detected ]
[ 101.841199] 4.5.0-rc2+  #49 Not tainted
[ 101.850251] ---------------------------------------------
[ 101.856621] modprobe/3054 is trying to acquire lock:
[ 101.862514] (&table->mutex#2){+.+.+.}, at: [<ffffffffa079c10e>] __mlx4_register_mac+0x87e/0xa90 [mlx4_core]
[ 101.874598]
[ 101.874598] but task is already holding lock:
[ 101.881703] (&table->mutex#2){+.+.+.}, at: [<ffffffffa079c0f0>] __mlx4_register_mac+0x860/0xa90 [mlx4_core]
[ 101.893776]
[ 101.893776] other info that might help us debug this:
[ 101.901658] Possible unsafe locking scenario:
[ 101.901658]
[ 101.908859] CPU0
[ 101.911923] ----
[ 101.914985] lock(&table->mutex#2);
[ 101.919595] lock(&table->mutex#2);
[ 101.924199]
[ 101.924199] * DEADLOCK *
[ 101.924199]
[ 101.931643] May be due to missing lock nesting notation

Fixes: 5f61385d ('net/mlx4_core: Keep VLAN/MAC tables mirrored in multifunc HA mode')
Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il>
Suggested-by: NDoron Tsur <doront@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Using the HW VPort counters for traffic (rx/tx packets/bytes)
statistics is wrong. This is because frames dropped due to steering or
out of buffer will be counted as received. To fix that, we move to use
the packet/bytes accounting done by the driver for what the netdev
reports out.

Fixes: f62b8bb8 ('net/mlx5: Extend mlx5_core to support [...]')
Signed-off-by: NGal Pressman <galp@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Sum up rx/tx bytes in software as we do for rx/tx packets, to be reported
in upcoming statistics fix.
Signed-off-by: NGal Pressman <galp@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Upon changing num_channels, reset the RSS indirection table to
match the new value.

Fixes: 2d75b2bc ('net/mlx5e: Add ethtool RSS configuration options')
Signed-off-by: NTariq Toukan <tariqt@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We should modify TIRs explicitly to apply the new RSS configuration.
The light ndo close/open calls do not "refresh" them.

Fixes: 2d75b2bc ('net/mlx5e: Add ethtool RSS configuration options')
Signed-off-by: NTariq Toukan <tariqt@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Readers/Writers lock for SW timecounter was acquired without disabling
interrupts on local CPU.

The problematic scenario:
* HW timestamping is enabled
* Timestamp overflow periodic service task is running on local CPU and
  holding write_lock for SW timecounter
* Completion arrives, triggers interrupt for local CPU.
  Interrupt routine calls napi_schedule(), which triggers rx/tx
  skb process.
  An attempt to read SW timecounter using read_lock is done, which is
  already locked by a writer on the same CPU and cause soft lockup.

Add irqsave/irqrestore for when using the readers/writers lock for
writing.

Fixes: ef9814de ('net/mlx5e: Add HW timestamping (TS) support')
Signed-off-by: NEran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Ethtool LRO enable/disable is broken, as of today we only modify TCP
TIRs in order to apply the requested configuration.

Hardware requires that all TIRs pointing to the same RQ should share the
same LRO configuration. For that all other TIRs' LRO fields must be
modified as well.

Fixes: 5c50368f ('net/mlx5e: Light-weight netdev open/stop')
Signed-off-by: NTariq Toukan <tariqt@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

With the MLX5E_CQ_HAS_CQES optimization flag, the following buggy
flow might occur:
- Suppose RX is always busy, TX has a single packet every second.
- We poll a single TX cqe and clear its flag.
- We never arm it again as RX is always busy.
- TX CQ flag is never changed, and new TX cqes are not polled.

We revert this optimization.

Fixes: e586b3b0 ('net/mlx5: Ethernet Datapath files')
Signed-off-by: NTariq Toukan <tariqt@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For RTL8168G/RTL8168H/RTL8411B/RTL8107E, enable this flag to eliminate
message "AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0002
address=0x0000000000003000 flags=0x0050] in dmesg.
Signed-off-by: NChunhao Lin <hau@realtek.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix a case of misleading code indentation.
Signed-off-by: NMichal Schmidt <mschmidt@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For consistency with other event data structs and to lessen
the chance of a mistake should one of the reserved fields become
used in the future, define the reserved fields as little-endian.
Signed-off-by: NMichal Schmidt <mschmidt@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There were no missing endianness conversions in this case, but the
fields of struct cfc_del_event_data should be defined as little-endian
to get rid of the ugly (__force __le32) casts.
Signed-off-by: NMichal Schmidt <mschmidt@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It's not really a bug, but it was odd that bnx2x_eq_int() read the
message data as if it were a cfc_del_event regardless of the event type.
It's cleaner to access only the appropriate member of union event_data
after checking the event opcode.
Signed-off-by: NMichal Schmidt <mschmidt@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

On ppc64 the PF did not receive messages from VFs correctly.

Fields of struct vf_pf_event_data are little-endian.
Signed-off-by: NMichal Schmidt <mschmidt@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When a VF is sending a message to the PF, it needs to trigger the PF
to tell it the message is ready.
The trigger did not work on ppc64. No interrupt appeared in the PF.

The bug is due to confusion about the layout of struct trigger_vf_zone.
In bnx2x_send_msg2pf() the trigger is written using writeb(), not
writel(), so the attempt to define the struct with a reversed layout on
big-endian is counter-productive.
Signed-off-by: NMichal Schmidt <mschmidt@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

bnx2x crashes during the initialization of the 8021q module on ppc64.
The bug is a missing conversion from le32 in
bnx2x_handle_classification_eqe() when obtaining the cid value from
struct eth_event_data.

The fields in struct eth_event_data should all be declared as
little-endian and conversions added where missing.
Signed-off-by: NMichal Schmidt <mschmidt@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Disable auto negotiation on init to properly detect an already plugged
cable at boot.

At boot, when the phy is started, it is in the PHY_UP state.
However, if a cable is plugged at boot, because auto negociation is already
enabled at the time we get the first interrupt, the phy is already running.
But the state machine then switches from PHY_UP to PHY_AN and calls
phy_start_aneg(). phy_start_aneg() will not do anything because aneg is
already enabled on the phy. It will then wait for a interrupt before going
further. This interrupt will never happen unless the cable is unplugged and
then replugged.

It was working properly before 321beec5 (net: phy: Use interrupts when
available in NOLINK state) because switching to NOLINK meant starting
polling the phy, even if IRQ were enabled.

Fixes: 321beec5 (net: phy: Use interrupts when available in NOLINK state)
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>