The mlx4 driver by default allocates order-3 pages for the ring to
consume in multiple fragments. When the device has an xdp program, this
behavior will prevent tx actions since the page must be re-mapped in
TODEVICE mode, which cannot be done if the page is still shared.
Start by making the allocator configurable based on whether xdp is
running, such that order-0 pages are always used and never shared.
Since this will stress the page allocator, add a simple page cache to
each rx ring. Pages in the cache are left dma-mapped, and in drop-only
stress tests the page allocator is eliminated from the perf report.
Note that setting an xdp program will now require the rings to be
reconfigured.
Before:
26.91% ksoftirqd/0 [mlx4_en] [k] mlx4_en_process_rx_cq
17.88% ksoftirqd/0 [mlx4_en] [k] mlx4_en_alloc_frags
6.00% ksoftirqd/0 [mlx4_en] [k] mlx4_en_free_frag
4.49% ksoftirqd/0 [kernel.vmlinux] [k] get_page_from_freelist
3.21% swapper [kernel.vmlinux] [k] intel_idle
2.73% ksoftirqd/0 [kernel.vmlinux] [k] bpf_map_lookup_elem
2.57% swapper [mlx4_en] [k] mlx4_en_process_rx_cq
After:
31.72% swapper [kernel.vmlinux] [k] intel_idle
8.79% swapper [mlx4_en] [k] mlx4_en_process_rx_cq
7.54% swapper [kernel.vmlinux] [k] poll_idle
6.36% swapper [mlx4_core] [k] mlx4_eq_int
4.21% swapper [kernel.vmlinux] [k] tasklet_action
4.03% swapper [kernel.vmlinux] [k] cpuidle_enter_state
3.43% swapper [mlx4_en] [k] mlx4_en_prepare_rx_desc
2.18% swapper [kernel.vmlinux] [k] native_irq_return_iret
1.37% swapper [kernel.vmlinux] [k] menu_select
1.09% swapper [kernel.vmlinux] [k] bpf_map_lookup_elem
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for the BPF_PROG_TYPE_XDP hook in mlx4 driver.
In tc/socket bpf programs, helpers linearize skb fragments as needed
when the program touches the packet data. However, in the pursuit of
speed, XDP programs will not be allowed to use these slower functions,
especially if it involves allocating an skb.
Therefore, disallow MTU settings that would produce a multi-fragment
packet that XDP programs would fail to access. Future enhancements could
be done to increase the allowable MTU.
The xdp program is present as a per-ring data structure, but as of yet
it is not possible to set at that granularity through any ndo.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sets the bpf program represented by fd as an early filter in the rx path
of the netdev. The fd must have been created as BPF_PROG_TYPE_XDP.
Providing a negative value as fd clears the program. Getting the fd back
via rtnl is not possible, therefore reading of this value merely
provides a bool whether the program is valid on the link or not.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add one new netdev op for drivers implementing the BPF_PROG_TYPE_XDP
filter. The single op is used for both setup/query of the xdp program,
modelled after ndo_setup_tc.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a new bpf prog type that is intended to run in early stages of the
packet rx path. Only minimal packet metadata will be available, hence a
new context type, struct xdp_md, is exposed to userspace. So far only
expose the packet start and end pointers, and only in read mode.
An XDP program must return one of the well known enum values, all other
return codes are reserved for future use. Unfortunately, this
restriction is hard to enforce at verification time, so take the
approach of warning at runtime when such programs are encountered. Out
of bounds return codes should alias to XDP_ABORTED.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
A subsystem may need to store many copies of a bpf program, each
deserving its own reference. Rather than requiring the caller to loop
one by one (with possible mid-loop failure), add a bulk bpf_prog_add
api.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gavin Shan says:
====================
NCSI Support
This series rebases on David's linux-net git repo ("master" branch). It's
to support NCSI stack on drivers/net/ethernet/faraday/ftgmac100.c. The
implementation is based on NCSI spec (version: 1.1.0):
https://www.dmtf.org/sites/default/files/standards/documents/DSP0222_1.1.0.pdf
As the following figure shows and defined in NCSI spec:
* The NC-SI (aka NCSI) is defined as the interface between a (Base)
Management Controller (BMC) and one or multiple Network Interface
Controlers (NIC) on host side. The interface is responsible for providing
external network connectivity for BMC.
* Each BMC can connect to multiple packages, up to 8. Each package can have
multiple channels, up to 32. Every package and channel are identified by
3-bits and 5-bits in NCSI packet.
* NCSI packet, encapsulated in ethernet frame, has 0x88F8 in the protocol
field. The destination MAC address should be 0xFF's while the source MAC
address can be arbitrary one.
* NCSI packets are classified to command, response, AEN (Asynchronous Event Notification).
Commands are sent from BMC to host (NIC) for configuration and
information retrival. Responses, corresponding to commands, are sent from
host to BMC for confirmation and requested information. One command should
have one and only one response. AEN is sent from host to BMC for notification
(e.g. link down on active channel) so that BMC can take appropriate action.
+------------------+ +----------------------------------------------+
| | | Host |
| BMC | | |
| | | +-------------------+ +-------------------+ |
| +---------+ | | | Package-A | | Package-B | |
| | | | | +---------+---------+ +-------------------+ |
| |ftgmac100| | | | Channel | Channel | | Channel | Channel | |
+----+----+----+---+ +-+---------+---------+--+---------+---------+-+
| | |
| | |
+-----------------------------+----------------------+
The series of patches is highlighted as:
The design for the patchset is highlighted as below:
* The network driver uses 3 interfaces exported from NCSI stack:
ncsi_register_dev() - Register (create) a associated NCSI device.
ncsi_start_dev() - Bring up the NCSI device.
ncsi_unregister_dev() - Destroy the registered NCSI device.
* There are several data structures introduced for different objects:
struct ncsi_dev - NCSI device seen by network device driver.
struct ncsi_dev_priv - NCSI device seen by NCSI stack.
struct ncsi_package - NCSI package which can have multiple channels.
struct ncsi_channel - NCSI channel.
* The NCSI stack is driven by workqueue and state machine internally.
* The all available NCSI packages and channels are enumerated (probed) on
the first call to ncsi_start_dev(). The NCSI topology won't change until
the NCSI device is destroyed.
* All available channels will be brought up When the hardware arbitration
is enabled. Otherwise, only one channel is selected as active one. The
NCSI internal is driven by state machine with help of a workqueue. In
the meanwhile, there are 3 states for each channel which can be put into
a queue requesting for configuration or suspending. Channels in the queue
with inactive state set will be configured (bringup) while channels in
the queue with active state will be suspended (teardown). The request
configuration or suspending is being applied on the channel if it's in
invisible state.
* Failover, another inactive channel is selected as active, can happen when
the hardware arbitration is disabled. The failover can be caused by timeout
on link monitor and AEN.
* NCSI stack should be configurable through netlink or another mechanism, it's
not implemented in this patchset. It's something TBD.
* The first NIC driver that is aware of NCSI: drivers/net/ethernet/faraday/ftgmac100.c
Changelog
=========
v2 -> v3:
* Include (one line) change in include/uapi/linux/if_ether.h to fix build
error.
v1 -> v2:
* Support NCSI spec v1.1.0 (3 more commands and 4 hardware arbitration
modes added).
* Enable AEN packets according to the supported list.
* Introduce NCSI channel states and processing queue in order to support
the hardware arbitration.
* The hardware arbitration is supported (tested with emulated environment).
* Introduce link monitor with GLS (Get Link Status) command/response as part
of the error handling defined in NCSI spec.
* Support IPv6 address discovery when CONFIG_IPV6 is enabled.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Bogus PHY interrupts are observed. This masks the PHY interrupt
when the interface works in NCSI mode as there is no attached
PHY under the circumstance.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This matches the driver with devices compatible with "faraday,ftgmac100"
declared in the device tree. Originally, device's name from device
tree for it.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This makes ftgmac100 driver support NCSI mode. The NCSI is enabled
on the interface if property "use-nc-si" or "use-ncsi" is found from
the device node in device tree.
* No PHY device is used when NCSI mode is enabled.
* The NCSI device (struct ncsi_dev) is created when probing the
device while it's enabled/started when the interface is brought
up.
* Hardware IP checksum dosn't work when NCSI mode is enabled. It
is disabled on enabled NCSI.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The device is assigned with random MAC address. It isn't reasonable.
An valid MAC address might have been provided by (uboot) firmware by
device-tree or in chip. It's reasonable to use it to maintain consistency.
This uses the MAC address from device-tree or that in the chip if it's
valid. Otherwise, a random MAC address is given as before.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This introduces two helper functions to create or destroy MDIO
interface. No logical changes introduced except the proper MDIO
names are given when having more than one MDIO bus.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This introduces NCSI AEN packet handlers that result in (A) the
currently active channel is reconfigured; (B) Currently active
channel is deconfigured and disabled, another channel is chosen
as active one and configured. Case (B) won't happen if hardware
arbitration has been enabled, the channel that was in active
state is suspended simply.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This manages NCSI packages and channels:
* The available packages and channels are enumerated in the first
time of calling ncsi_start_dev(). The channels' capabilities are
probed in the meanwhile. The NCSI network topology won't change
until the NCSI device is destroyed.
* There in a queue in every NCSI device. The element in the queue,
channel, is waiting for configuration (bringup) or suspending
(teardown). The channel's state (inactive/active) indicates the
futher action (configuration or suspending) will be applied on the
channel. Another channel's state (invisible) means the requested
action is being applied.
* The hardware arbitration will be enabled if all available packages
and channels support it. All available channels try to provide
service when hardware arbitration is enabled. Otherwise, one channel
is selected as the active one at once.
* When channel is in active state, meaning it's providing service, a
timer started to retrieve the channe's link status. If the channel's
link status fails to be updated in the determined period, the channel
is going to be reconfigured. It's the error handling implementation
as defined in NCSI spec.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The NCSI response packets are sent to MC (Management Controller)
from the remote end. They are responses of NCSI command packets
for multiple purposes: completion status of NCSI command packets,
return NCSI channel's capability or configuration etc.
This defines struct to represent NCSI response packets and introduces
function ncsi_rcv_rsp() which will be used to receive NCSI response
packets and parse them.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The NCSI command packets are sent from MC (Management Controller)
to remote end. They are used for multiple purposes: probe existing
NCSI package/channel, retrieve NCSI channel's capability, configure
NCSI channel etc.
This defines struct to represent NCSI command packets and introduces
function ncsi_xmit_cmd(), which will be used to transmit NCSI command
packet according to the request. The request is represented by struct
ncsi_cmd_arg.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
NCSI spec (DSP0222) defines several objects: package, channel, mode,
filter, version and statistics etc. This introduces the data structs
to represent those objects and implement functions to manage them.
Also, this introduces CONFIG_NET_NCSI for the newly implemented NCSI
stack.
* The user (e.g. netdev driver) dereference NCSI device by
"struct ncsi_dev", which is embedded to "struct ncsi_dev_priv".
The later one is used by NCSI stack internally.
* Every NCSI device can have multiple packages simultaneously, up
to 8 packages. It's represented by "struct ncsi_package" and
identified by 3-bits ID.
* Every NCSI package can have multiple channels, up to 32. It's
represented by "struct ncsi_channel" and identified by 5-bits ID.
* Every NCSI channel has version, statistics, various modes and
filters. They are represented by "struct ncsi_channel_version",
"struct ncsi_channel_stats", "struct ncsi_channel_mode" and
"struct ncsi_channel_filter" separately.
* Apart from AEN (Asynchronous Event Notification), the NCSI stack
works in terms of command and response. This introduces "struct
ncsi_req" to represent a complete NCSI transaction made of NCSI
request and response.
link: https://www.dmtf.org/sites/default/files/standards/documents/DSP0222_1.1.0.pdf
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vivien Didelot says:
====================
net: dsa: mv88e6xxx: Global2 cleanup and STP
The Marvell switches registers are organized in distinct internal SMI
devices, such as PHY, Port, Global 1 or Global 2 registers sets.
Since not all chips support every registers sets or have slightly
differences in them (such as old 88E6060 or new 88E6390 likely to be
supported soon), make the setup code clearer now by removing a few
family checks and adding flags to describe the Global 2 registers map.
This patchset enables basic STP support and bridging on most chips when
getting rid of a few inconsistencies in chip descriptions (patch 1) and
add bridge Ageing Time support to DSA and the mv88e6xxx driver.
Changes v2 -> v3:
- rename mv88e6xxx_update_write to mv88e6xxx_update
- set fastest ageing time in use in the chip for multiple bridges,
tested with a few printk
Changes v1 -> v2:
- add a write helper for pointer-data Update registers
- add ageing time support
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Implement the DSA driver function to configure the bridge ageing time.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
All Marvell switch chips from (88E6060 to 88E6390) have a ATU Control
register containing bits 11:4 to configure an ATU Age Time quotient.
However the coefficient used to calculate the ATU Age Time vary with the
models. E.g. 88E6060, 88E6352 and 88E6390 use respectively 16, 15 and
3.75 seconds.
Add a age_time_coeff to the info structure to handle this and a Global 1
helper to set the default age time of 5 minutes in the setup code.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a new function for DSA drivers to handle the switchdev
SWITCHDEV_ATTR_ID_BRIDGE_AGEING_TIME attribute.
The ageing time is passed as milliseconds.
Also because we can have multiple logical bridges on top of a physical
switch and ageing time are switch-wide, call the driver function with
the fastest ageing time in use on the chip instead of the requested one.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add capability flags to describe the presence of Ingress Rate Limit unit
registers and an helper function to clear it.
In the meantime, fix a few harmless issues:
- 6185 and 6095 don't have such registers (reserved)
- the previous code didn't wait for the IRL operation to complete
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add flags and helpers to describe the presence of Priority Override
Table (POT) related registers and simplify the setup of Global 2.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add flags to describe the presence of Cross-chip Port VLAN Table (PVT)
related registers and simplify the setup of Global 2.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Switches such as 88E6185 as 3 Switch MAC registers in Global 1. Newer
chips such as 88E6352 have freed these registers in favor of an indirect
access in a Switch MAC/WoL/WoF register in Global 2.
Explicit this difference with G1 and G2 helpers and flags.
Also, note that this indirect access is a single-register which doesn't
require to wait for the operation to complete (like Switch MAC, Trunk
Mapping, etc.), in contrary to multi-registers indirect accesses with
several operations and a busy bit.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some switches provide a Rsvd2CPU mechanism used to choose which of the
16 reserved multicast destination addresses matching 01:80:c2:00:00:0x
should be considered as MGMT and thus forwarded to the CPU port.
Other switches extend this mechanism to also configure as MGMT the
additional 16 reserved multicast addresses matching 01:80:c2:00:00:2x.
This mechanism is exposed via two registers in Global 2, and an Rsvd2CPU
enable bit in the management register.
Newer chip (such as 88E6390) has replaced these registers with a new
indirect MGMT mechanism in Global 1.
The patch adds two MV88E6XXX_FLAG_G2_MGMT_EN_{0,2}X flags to describe
the presence of these Global 2 registers. If 88E6390 support is added, a
MV88E6XXX_FLAG_G1_MGMT_CTRL flag will be needed to setup Rsvd2CPU.
Note: all switches still support in parallel the ATU Load operation with
an MGMT Entry State to forward such frames in a less convenient way.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Trunk Mask and Trunk Mapping registers are two Global 2 indirect
accesses to trunking configuration.
Add helpers for these tables and simplify the Global 2 setup.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Device Mapping register is an indirect table access.
Provide helpers to access this table and explicit the checking of the
new DSA_RTABLE_NONE routing table value.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Separate the setup of Global 1 and Global 2 internal SMI devices and add
a flag to describe the presence of this second registers set.
Also rearrange the G1 setup in the registers order.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
All 88E6xxx Marvell switches (even the old not supported yet 88E6060)
have at least an ATU, per-port STP states and VLAN map, to run basic
switch functions such as Spanning Tree and port based VLANs.
Get rid of the related MV88E6XXX_FLAG_{ATU,PORTSTATE,VLANTABLE} flags,
as they are defaults to every chip.
This enables STP on 6185 and removes many inconsistencies on others.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
kernel/trace/bpf_trace.c: In function 'bpf_event_output':
kernel/trace/bpf_trace.c:312: error: unknown field 'next' specified in initializer
kernel/trace/bpf_trace.c:312: warning: missing braces around initializer
kernel/trace/bpf_trace.c:312: warning: (near initialization for 'raw.frag.<anonymous>')
Fixes: 555c8a8623 ("bpf: avoid stack copy and use skb ctx for event output")
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
VLAN and MQ control was doing DMA from the stack. Fix it.
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: "netdev@vger.kernel.org" <netdev@vger.kernel.org>
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
txr->dev_state was not consistently manipulated with the acquisition of
the per-queue lock, after further inspection the lock does not seem
necessary, either the value is read as BNXT_DEV_STATE_CLOSING or 0.
Reported-by: coverity (CID 1339583)
Fixes: c0c050c58d ("bnxt_en: New Broadcom ethernet driver.")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Shmulik Ladkani says:
====================
net: Consider fragmentation of udp tunneled skbs in 'ip_finish_output_gso'
Currently IP fragmentation of GSO segments that exceed dst mtu is
considered only in the ipv4 forwarding case.
There are cases where GSO skbs that are bridged and then udp-tunneled
may have gso_size exceeding the egress device mtu.
It makes sense to fragment them, as in the non GSOed code path.
The exact cases where this behavior is needed is described and addressed
in the 2nd patch.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Given:
- tap0 and vxlan0 are bridged
- vxlan0 stacked on eth0, eth0 having small mtu (e.g. 1400)
Assume GSO skbs arriving from tap0 having a gso_size as determined by
user-provided virtio_net_hdr (e.g. 1460 corresponding to VM mtu of 1500).
After encapsulation these skbs have skb_gso_network_seglen that exceed
eth0's ip_skb_dst_mtu.
These skbs are accidentally passed to ip_finish_output2 AS IS.
Alas, each final segment (segmented either by validate_xmit_skb or by
hardware UFO) would be larger than eth0 mtu.
As a result, those above-mtu segments get dropped on certain networks.
This behavior is not aligned with the NON-GSO case:
Assume a non-gso 1500-sized IP packet arrives from tap0. After
encapsulation, the vxlan datagram is fragmented normally at the
ip_finish_output-->ip_fragment code path.
The expected behavior for the GSO case would be segmenting the
"gso-oversized" skb first, then fragmenting each segment according to
dst mtu, and finally passing the resulting fragments to ip_finish_output2.
'ip_finish_output_gso' already supports this "Slowpath" behavior,
according to the IPSKB_FRAG_SEGS flag, which is only set during ipv4
forwarding (not set in the bridged case).
In order to support the bridged case, we'll mark skbs arriving from an
ingress interface that get udp-encaspulated as "allowed to be fragmented",
causing their network_seglen to be validated by 'ip_finish_output_gso'
(and fragment if needed).
Note the TUNNEL_DONT_FRAGMENT tun_flag is still honoured (both in the
gso and non-gso cases), which serves users wishing to forbid
fragmentation at the udp tunnel endpoint.
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Cc: Florian Westphal <fw@strlen.de>
Signed-off-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This flag indicates whether fragmentation of segments is allowed.
Formerly this policy was hardcoded according to IPSKB_FORWARDED (set by
either ip_forward or ipmr_forward).
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Cc: Florian Westphal <fw@strlen.de>
Signed-off-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan says:
====================
bnxt_en: Add support for NS2 Nitro.
This series adds support for the embedded version of the
ethernet controller (Nitro) in the North Star 2 SoC. There are a number
of features not supported and a software workaround for a hardware rx
bug is required for Nitro A0. Please review.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
A bridge device in NS2 has the same device ID as the ethernet controller.
Add check to avoid probing the bridge device.
Signed-off-by: Prashant Sreedharan <prashant.sreedharan@broadcom.com>
Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allocate special vnic for dropping packets not matching the RX filters.
First vnic is for normal RX packets and the driver will drop all
packets on the 2nd vnic.
Signed-off-by: Prashant Sreedharan <prashant.sreedharan@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allocate napi for special vnic, packets arriving on this
napi will simply be dropped and the buffers will be replenished back
to the HW.
Signed-off-by: Prashant Sreedharan <prashant.sreedharan@broadcom.com>
Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The hardware is unable to drop rx packets not matching the RX filters. To
workaround it, we create a special VNIC and configure the hardware to
direct all packets not matching the filters to it. We then setup the
driver to drop packets received on this VNIC.
This patch creates the infrastructure for this VNIC, reserves a
completion ring, and rx rings. Only shared completion ring mode is
supported. The next 2 patches add a NAPI to handle packets from this
VNIC and the setup of the VNIC.
Signed-off-by: Prashant Sreedharan <prashant.sreedharan@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Nitro A0 has a hardware bug in the rx path. The workaround is to create
a special COS context as a path for non-RSS (non-IP) packets. Without this
workaround, the chip may stall when receiving RSS and non-RSS packets.
Add infrastructure to allow 2 contexts (RSS and CoS) per VNIC. Allocate
and configure the CoS context for Nitro A0.
Signed-off-by: Prashant Sreedharan <prashant.sreedharan@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Nitro is the embedded version of the ethernet controller in the North
Star 2 SoC. Add basic code to recognize the chip ID and disable
the features (ntuple, TPA, ring and port statistics) not supported on
Nitro A0.
Signed-off-by: Prashant Sreedharan <prashant.sreedharan@broadcom.com>
Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Charles-Antoine Couret says:
====================
Marvell phy: fiber interface configuration
Another patchset to manage correctly the fiber link for some concerned Marvell's
phy like 88E1512.
This patchset fixed the commit log for the third and last commits and a comment
in the first commit.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
These functions used standards registers in a different page
for both interfaces: copper and fiber.
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Charles-Antoine Couret <charles-antoine.couret@nexvision.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
To be correctly initilized, the fiber interface needs
to be configured via autonegociation registers which use
some customs options or registers.
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Charles-Antoine Couret <charles-antoine.couret@nexvision.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for the fiber receiver error counter in the
statistics. Rename the current counter which is for copper errors to
phy_receive_errors_copper, so it is easy to distinguish copper from
fiber.
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Charles-Antoine Couret <charles-antoine.couret@nexvision.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
For concerned phy, the fiber link is checked before the copper link.
According to datasheet, the link which is up is enabled.
If both links are down, copper link would be used.
To detect fiber link status, we used the real time status
because of troubles with the copper method.
Tested with Marvell 88E1512.
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Charles-Antoine Couret <charles-antoine.couret@nexvision.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sergei Shtylyov says:
====================
Fix DMA channel misreporting for the Renesas Ethernet drivers
Here's a set of 2 patches against DaveM's 'net.git' repo fixing up the DMA
channel reporting by 'ifconfig'...
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Brenden Blanco