Add vlan push_eth and pop_eth action to the hardware intermediate
representation model which would subsequently allow it to be used
by drivers for offload.
Signed-off-by: Maor Dickman <maord@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Now that devlink ports are protected by the instance lock
it seems natural to pass devlink_port as an argument to
the port_split / port_unsplit callbacks.
This should save the drivers from doing a lookup.
In theory drivers may have supported unsplitting ports
which were not registered prior to this change.
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
It should be familiar and beneficial to expose devlink instance
lock to the drivers. This way drivers can block devlink from
calling them during critical sections without breakneck locking.
Add port helpers, port splitting callbacks will be the first
target.
Use 'devl_' prefix for "explicitly locked" API. Initial RFC used
'__devlink' but that's too much typing.
devl_lock_is_held() is not defined without lockdep, which is
the same behavior as lockdep_is_held() itself.
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The fundamental premise of VRF and l3mdev core code is binding a socket
to a device (l3mdev or netdev with an L3 domain) to indicate L3 scope.
Legacy code resets flowi_oif to the l3mdev losing any original port
device binding. Ben (among others) has demonstrated use cases where the
original port device binding is important and needs to be retained.
This patch handles that by adding a new entry to the common flow struct
that can indicate the l3mdev index for later rule and table matching
avoiding the need to reset flowi_oif.
In addition to allowing more use cases that require port device binds,
this patch brings a few datapath simplications:
1. l3mdev_fib_rule_match is only called when walking fib rules and
always after l3mdev_update_flow. That allows an optimization to bail
early for non-VRF type uses cases when flowi_l3mdev is not set. Also,
only that index needs to be checked for the FIB table id.
2. l3mdev_update_flow can be called with flowi_oif set to a l3mdev
(e.g., VRF) device. By resetting flowi_oif only for this case the
FLOWI_FLAG_SKIP_NH_OIF flag is not longer needed and can be removed,
removing several checks in the datapath. The flowi_iif path can be
simplified to only be called if the it is not loopback (loopback can
not be assigned to an L3 domain) and the l3mdev index is not already
set.
3. Avoid another device lookup in the output path when the fib lookup
returns a reject failure.
Note: 2 functional tests for local traffic with reject fib rules are
updated to reflect the new direct failure at FIB lookup time for ping
rather than the failure on packet path. The current code fails like this:
HINT: Fails since address on vrf device is out of device scope
COMMAND: ip netns exec ns-A ping -c1 -w1 -I eth1 172.16.3.1
ping: Warning: source address might be selected on device other than: eth1
PING 172.16.3.1 (172.16.3.1) from 172.16.3.1 eth1: 56(84) bytes of data.
--- 172.16.3.1 ping statistics ---
1 packets transmitted, 0 received, 100% packet loss, time 0ms
where the test now directly fails:
HINT: Fails since address on vrf device is out of device scope
COMMAND: ip netns exec ns-A ping -c1 -w1 -I eth1 172.16.3.1
ping: connect: No route to host
Signed-off-by: David Ahern <dsahern@kernel.org>
Tested-by: Ben Greear <greearb@candelatech.com>
Link: https://lore.kernel.org/r/20220314204551.16369-1-dsahern@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Follow the established programming model for this driver and provide
shims in the felix DSA driver which call the implementations from the
ocelot switch lib. The ocelot switchdev driver wasn't integrated with
dcbnl due to lack of hardware availability.
The switch doesn't have any fancy QoS classification enabled by default.
The provided getters will create a default-prio app table entry of 0,
and no dscp entry. However, the getters have been made to actually
retrieve the hardware configuration rather than static values, to be
future proof in case DSA will need this information from more call paths.
For default-prio, there is a single field per port, in ANA_PORT_QOS_CFG,
called QOS_DEFAULT_VAL.
DSCP classification is enabled per-port, again via ANA_PORT_QOS_CFG
(field QOS_DSCP_ENA), and individual DSCP values are configured as
trusted or not through register ANA_DSCP_CFG (replicated 64 times).
An untrusted DSCP value falls back to other QoS classification methods.
If trusted, the selected ANA_DSCP_CFG register also holds the QoS class
in the QOS_DSCP_VAL field.
The hardware also supports DSCP remapping (DSCP value X is translated to
DSCP value Y before the QoS class is determined based on the app table
entry for Y) and DSCP packet rewriting. The dcbnl framework, for being
so flexible in other useless areas, doesn't appear to support this.
So this functionality has been left out.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Similar to the port-based default priority, IEEE 802.1Q-2018 allows the
Application Priority Table to define QoS classes (0 to 7) per IP DSCP
value (0 to 63).
In the absence of an app table entry for a packet with DSCP value X,
QoS classification for that packet falls back to other methods (VLAN PCP
or port-based default). The presence of an app table for DSCP value X
with priority Y makes the hardware classify the packet to QoS class Y.
As opposed to the default-prio where DSA exposes only a "set" in
dsa_switch_ops (because the port-based default is the fallback, it
always exists, either implicitly or explicitly), for DSCP priorities we
expose an "add" and a "del". The addition of a DSCP entry means trusting
that DSCP priority, the deletion means ignoring it.
Drivers that already trust (at least some) DSCP values can describe
their configuration in dsa_switch_ops :: port_get_dscp_prio(), which is
called for each DSCP value from 0 to 63.
Again, there can be more than one dcbnl app table entry for the same
DSCP value, DSA chooses the one with the largest configured priority.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The port-based default QoS class is assigned to packets that lack a
VLAN PCP (or the port is configured to not trust the VLAN PCP),
an IP DSCP (or the port is configured to not trust IP DSCP), and packets
on which no tc-skbedit action has matched.
Similar to other drivers, this can be exposed to user space using the
DCB Application Priority Table. IEEE 802.1Q-2018 specifies in Table
D-8 - Sel field values that when the Selector is 1, the Protocol ID
value of 0 denotes the "Default application priority. For use when
application priority is not otherwise specified."
The way in which the dcbnl integration in DSA has been designed has to
do with its requirements. Andrew Lunn explains that SOHO switches are
expected to come with some sort of pre-configured QoS profile, and that
it is desirable for this to come pre-loaded into the DSA slave interfaces'
DCB application priority table.
In the dcbnl design, this is possible because calls to dcb_ieee_setapp()
can be initiated by anyone including being self-initiated by this device
driver.
However, what makes this challenging to implement in DSA is that the DSA
core manages the net_devices (effectively hiding them from drivers),
while drivers manage the hardware. The DSA core has no knowledge of what
individual drivers' QoS policies are. DSA could export to drivers a
wrapper over dcb_ieee_setapp() and these could call that function to
pre-populate the app priority table, however drivers don't have a good
moment in time to do this. The dsa_switch_ops :: setup() method gets
called before the net_devices are created (dsa_slave_create), and so is
dsa_switch_ops :: port_setup(). What remains is dsa_switch_ops ::
port_enable(), but this gets called upon each ndo_open. If we add app
table entries on every open, we'd need to remove them on close, to avoid
duplicate entry errors. But if we delete app priority entries on close,
what we delete may not be the initial, driver pre-populated entries, but
rather user-added entries.
So it is clear that letting drivers choose the timing of the
dcb_ieee_setapp() call is inappropriate. The alternative which was
chosen is to introduce hardware-specific ops in dsa_switch_ops, and
effectively hide dcbnl details from drivers as well. For pre-populating
the application table, dsa_slave_dcbnl_init() will call
ds->ops->port_get_default_prio() which is supposed to read from
hardware. If the operation succeeds, DSA creates a default-prio app
table entry. The method is called as soon as the slave_dev is
registered, but before we release the rtnl_mutex. This is done such that
user space sees the app table entries as soon as it sees the interface
being registered.
The fact that we populate slave_dev->dcbnl_ops with a non-NULL pointer
changes behavior in dcb_doit() from net/dcb/dcbnl.c, which used to
return -EOPNOTSUPP for any dcbnl operation where netdev->dcbnl_ops is
NULL. Because there are still dcbnl-unaware DSA drivers even if they
have dcbnl_ops populated, the way to restore the behavior is to make all
dcbnl_ops return -EOPNOTSUPP on absence of the hardware-specific
dsa_switch_ops method.
The dcbnl framework absurdly allows there to be more than one app table
entry for the same selector and protocol (in other words, more than one
port-based default priority). In the iproute2 dcb program, there is a
"replace" syntactical sugar command which performs an "add" and a "del"
to hide this away. But we choose the largest configured priority when we
call ds->ops->port_set_default_prio(), using __fls(). When there is no
default-prio app table entry left, the port-default priority is restored
to 0.
Link: https://patchwork.kernel.org/project/netdevbpf/patch/20210113154139.1803705-2-olteanv@gmail.com/
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
____napi_schedule() needs to be invoked with disabled interrupts due to
__raise_softirq_irqoff (in order not to corrupt the per-CPU list).
____napi_schedule() needs also to be invoked from an interrupt context
so that the raised-softirq is processed while the interrupt context is
left.
Add lockdep asserts for both conditions.
While this is the second time the irq/softirq check is needed, provide a
generic lockdep_assert_softirq_will_run() which is used by both caller.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Tony Nguyen says:
====================
ice: GTP support in switchdev
Marcin Szycik says:
Add support for adding GTP-C and GTP-U filters in switchdev mode.
To create a filter for GTP, create a GTP-type netdev with ip tool, enable
hardware offload, add qdisc and add a filter in tc:
ip link add $GTP0 type gtp role <sgsn/ggsn> hsize <hsize>
ethtool -K $PF0 hw-tc-offload on
tc qdisc add dev $GTP0 ingress
tc filter add dev $GTP0 ingress prio 1 flower enc_key_id 1337 \
action mirred egress redirect dev $VF1_PR
By default, a filter for GTP-U will be added. To add a filter for GTP-C,
specify enc_dst_port = 2123, e.g.:
tc filter add dev $GTP0 ingress prio 1 flower enc_key_id 1337 \
enc_dst_port 2123 action mirred egress redirect dev $VF1_PR
Note: outer IPv6 offload is not supported yet.
Note: GTP-U with no payload offload is not supported yet.
ICE COMMS package is required to create a filter as it contains GTP
profiles.
Changes in iproute2 [1] are required to be able to add GTP netdev and use
GTP-specific options (QFI and PDU type).
[1] https://lore.kernel.org/netdev/20220211182902.11542-1-wojciech.drewek@intel.com/T
---
v2: Add more CC
v3: Fix mail thread, sorry for spam
v4: Add GTP echo response in gtp module
v5: Change patch order
v6: Add GTP echo request in gtp module
v7: Fix kernel-docs in ice
v8: Remove handling of GTP Echo Response
v9: Add sending of multicast message on GTP Echo Response, fix GTP-C dummy
packet selection
v10: Rebase, fixed most 80 char line limits
v11: Rebase, collect Harald's Reviewed-by on patch 3
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Before adding yet another possibly contended atomic_long_t,
it is time to add per-cpu storage for existing ones:
dev->tx_dropped, dev->rx_dropped, and dev->rx_nohandler
Because many devices do not have to increment such counters,
allocate the per-cpu storage on demand, so that dev_get_stats()
does not have to spend considerable time folding zero counters.
Note that some drivers have abused these counters which
were supposed to be only used by core networking stack.
v4: should use per_cpu_ptr() in dev_get_stats() (Jakub)
v3: added a READ_ONCE() in netdev_core_stats_alloc() (Paolo)
v2: add a missing include (reported by kernel test robot <lkp@intel.com>)
Change in netdev_core_stats_alloc() (Jakub)
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: jeffreyji <jeffreyji@google.com>
Reviewed-by: Brian Vazquez <brianvv@google.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Link: https://lore.kernel.org/r/20220311051420.2608812-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Enable binding the nfp driver to NFP3800 and NFP3803 devices.
The PCIE_SRAM offset is different for the NFP3800 device, which also
only supports a single explicit group.
Changes to Dirk's work:
* 48-bit dma addressing is not ready yet. Keep 40-bit dma addressing
for NFP3800.
Signed-off-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Fei Qin <fei.qin@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Johannes Berg says:
====================
brcmfmac
* add BCM43454/6 support
rtw89
* add support for 160 MHz channels and 6 GHz band
* hardware scan support
iwlwifi
* support UHB TAS enablement via BIOS
* remove a bunch of W=1 warnings
* add support for channel switch offload
* support 32 Rx AMPDU sessions in newer devices
* add support for a couple of new devices
* add support for band disablement via BIOS
mt76
* mt7915 thermal management improvements
* SAR support for more mt76 drivers
* mt7986 wmac support on mt7915
ath11k
* debugfs interface to configure firmware debug log level
* debugfs interface to test Target Wake Time (TWT)
* provide 802.11ax High Efficiency (HE) data via radiotap
ath9k
* use hw_random API instead of directly dumping into random.c
wcn36xx
* fix wcn3660 to work on 5 GHz band
ath6kl
* add device ID for WLU5150-D81
cfg80211/mac80211
* initial EHT (from 802.11be) support
(EHT rates, 320 MHz, larger block-ack)
* support disconnect on HW restart
* tag 'wireless-next-2022-03-11' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next: (247 commits)
mac80211: Add support to trigger sta disconnect on hardware restart
mac80211: fix potential double free on mesh join
mac80211: correct legacy rates check in ieee80211_calc_rx_airtime
nl80211: fix typo of NL80211_IF_TYPE_OCB in documentation
mac80211: Use GFP_KERNEL instead of GFP_ATOMIC when possible
mac80211: replace DEFINE_SIMPLE_ATTRIBUTE with DEFINE_DEBUGFS_ATTRIBUTE
rtw89: 8852c: process logic efuse map
rtw89: 8852c: process efuse of phycap
rtw89: support DAV efuse reading operation
rtw89: 8852c: add chip::dle_mem
rtw89: add page_regs to handle v1 chips
rtw89: add chip_info::{h2c,c2h}_reg to support more chips
rtw89: add hci_func_en_addr to support variant generation
rtw89: add power_{on/off}_func
rtw89: read chip version depends on chip ID
rtw89: pci: use a struct to describe all registers address related to DMA channel
rtw89: pci: add V1 of PCI channel address
rtw89: pci: add struct rtw89_pci_info
rtw89: 8852c: add 8852c empty files
MAINTAINERS: add devicetree bindings entry for mt76
...
====================
Link: https://lore.kernel.org/r/20220311124029.213470-1-johannes@sipsolutions.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Options are as follows: PDU_TYPE:QFI and they refernce to
the fields from the PDU Session Protocol. PDU Session data
is conveyed in GTP-U Extension Header.
GTP-U Extension Header is described in 3GPP TS 29.281.
PDU Session Protocol is described in 3GPP TS 38.415.
PDU_TYPE - indicates the type of the PDU Session Information (4 bits)
QFI - QoS Flow Identifier (6 bits)
# ip link add gtp_dev type gtp role sgsn
# tc qdisc add dev gtp_dev ingress
# tc filter add dev gtp_dev protocol ip parent ffff: \
flower \
enc_key_id 11 \
gtp_opts 1:8/ff:ff \
action mirred egress redirect dev eth0
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Adding GTP device through ip link creates the situation where
GTP instance is not able to send GTP echo requests.
Echo requests are used to check if GTP peer is still alive.
With this patch, gtp_genl_ops are extended by new cmd (GTP_CMD_ECHOREQ)
which allows to send echo request in the given version of GTP
protocol (v0 or v1), from the given ms address to he given
peer. TID is not inclued because in all path management
messages it should be equal to 0.
When GTP echo response is detected, multicast message is
send to everyone in the gtp_genl_family. Message contains
GTP version, ms address and peer address.
Suggested-by: Harald Welte <laforge@gnumonks.org>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
Reviewed-by: Harald Welte <laforge@gnumonks.org>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Adding GTP device through ip link creates the situation where
there is no userspace daemon which would handle GTP messages
(Echo Request for example). GTP-U instance which would not respond
to echo requests would violate GTP specification.
When GTP packet arrives with GTP_ECHO_REQ message type,
GTP_ECHO_RSP is send to the sender. GTP_ECHO_RSP message
should contain information element with GTPIE_RECOVERY tag and
restart counter value. For GTPv1 restart counter is not used
and should be equal to 0, for GTPv0 restart counter contains
information provided from userspace(IFLA_GTP_RESTART_COUNT).
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
Suggested-by: Harald Welte <laforge@gnumonks.org>
Reviewed-by: Harald Welte <laforge@gnumonks.org>
Tested-by: Harald Welte <laforge@gnumonks.org>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Currently, when the user wants to create GTP device, he has to
provide file handles to the sockets created in userspace (IFLA_GTP_FD0,
IFLA_GTP_FD1). This behaviour is not ideal, considering the option of
adding support for GTP device creation through ip link. Ip link
application is not a good place to create such sockets.
This patch allows to create GTP device without providing
IFLA_GTP_FD0 and IFLA_GTP_FD1 arguments. If the user sets
IFLA_GTP_CREATE_SOCKETS attribute, then GTP module takes care
of creating UDP sockets by itself. Sockets are created with the
commonly known UDP ports used for GTP protocol (GTP0_PORT and
GTP1U_PORT). In this case we don't have to provide encap_destroy
because no extra deinitialization is needed, everything is covered
by udp_tunnel_sock_release.
Note: GTP instance created with only this change applied, does
not handle GTP Echo Requests. This is implemented in the following
patch.
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Currently in case of target hardware restart, we just reconfig and
re-enable the security keys and enable the network queues to start
data traffic back from where it was interrupted.
Many ath10k wifi chipsets have sequence numbers for the data
packets assigned by firmware and the mac sequence number will
restart from zero after target hardware restart leading to mismatch
in the sequence number expected by the remote peer vs the sequence
number of the frame sent by the target firmware.
This mismatch in sequence number will cause out-of-order packets
on the remote peer and all the frames sent by the device are dropped
until we reach the sequence number which was sent before we restarted
the target hardware
In order to fix this, we trigger a sta disconnect, in case of target
hw restart. After this there will be a fresh connection and thereby
avoiding the dropping of frames by remote peer.
The right fix would be to pull the entire data path into the host
which is not feasible or would need lots of complex changes and
will still be inefficient.
Tested on ath10k using WCN3990, QCA6174
Signed-off-by: Youghandhar Chintala <youghand@codeaurora.org>
Link: https://lore.kernel.org/r/20220308115325.5246-2-youghand@codeaurora.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Today's implementation of csum_shift() leads to branching based on
parity of 'offset'
000002f8 <csum_block_add>:
2f8: 70 a5 00 01 andi. r5,r5,1
2fc: 41 a2 00 08 beq 304 <csum_block_add+0xc>
300: 54 84 c0 3e rotlwi r4,r4,24
304: 7c 63 20 14 addc r3,r3,r4
308: 7c 63 01 94 addze r3,r3
30c: 4e 80 00 20 blr
Use first bit of 'offset' directly as input of the rotation instead of
branching.
000002f8 <csum_block_add>:
2f8: 54 a5 1f 38 rlwinm r5,r5,3,28,28
2fc: 20 a5 00 20 subfic r5,r5,32
300: 5c 84 28 3e rotlw r4,r4,r5
304: 7c 63 20 14 addc r3,r3,r4
308: 7c 63 01 94 addze r3,r3
30c: 4e 80 00 20 blr
And change to left shift instead of right shift to skip one more
instruction. This has no impact on the final sum.
000002f8 <csum_block_add>:
2f8: 54 a5 1f 38 rlwinm r5,r5,3,28,28
2fc: 5c 84 28 3e rotlw r4,r4,r5
300: 7c 63 20 14 addc r3,r3,r4
304: 7c 63 01 94 addze r3,r3
308: 4e 80 00 20 blr
Seems like only powerpc benefits from a branchless implementation.
Other main architectures like ARM or X86 get better code with
the generic implementation and its branch.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>
The assumption that the first byte in the module mapping dword is the
module number shouldn't be hard-coded in the driver, but come from
mlx5_ifc structs.
While at it, fix the incorrect width for the 'rx_lane' and 'tx_lane'
fields.
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
The MCIA register supports either 12 or 32 dwords, use the correct value
by querying the capability from the MCAM register.
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Few years ago OVS user space made a strange choice in the commit [1]
to define types only valid for the user space inside the copy of a
kernel uAPI header. '#ifndef __KERNEL__' and another attribute was
added later.
This leads to the inevitable clash between user space and kernel types
when the kernel uAPI is extended. The issue was unveiled with the
addition of a new type for IPv6 extension header in kernel uAPI.
When kernel provides the OVS_KEY_ATTR_IPV6_EXTHDRS attribute to the
older user space application, application tries to parse it as
OVS_KEY_ATTR_PACKET_TYPE and discards the whole netlink message as
malformed. Since OVS_KEY_ATTR_IPV6_EXTHDRS is supplied along with
every IPv6 packet that goes to the user space, IPv6 support is fully
broken.
Fixing that by bringing these user space attributes to the kernel
uAPI to avoid the clash. Strictly speaking this is not the problem
of the kernel uAPI, but changing it is the only way to avoid breakage
of the older user space applications at this point.
These 2 types are explicitly rejected now since they should not be
passed to the kernel. Additionally, OVS_KEY_ATTR_TUNNEL_INFO moved
out from the '#ifdef __KERNEL__' as there is no good reason to hide
it from the userspace. And it's also explicitly rejected now, because
it's for in-kernel use only.
Comments with warnings were added to avoid the problem coming back.
(1 << type) converted to (1ULL << type) to avoid integer overflow on
OVS_KEY_ATTR_IPV6_EXTHDRS, since it equals 32 now.
[1] beb75a40fdc2 ("userspace: Switching of L3 packets in L2 pipeline")
Fixes: 28a3f06017 ("net: openvswitch: IPv6: Add IPv6 extension header support")
Link: https://lore.kernel.org/netdev/3adf00c7-fe65-3ef4-b6d7-6d8a0cad8a5f@nvidia.com
Link: beb75a40fd
Reported-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Aaron Conole <aconole@redhat.com>
Link: https://lore.kernel.org/r/20220309222033.3018976-1-i.maximets@ovn.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Marc Kleine-Budde says:
====================
pull-request: can-next 2022-03-10
The first 3 patches are by Oliver Hartkopp, target the CAN ISOTP
protocol and update the CAN frame sending behavior, and increases the
max PDU size to 64 kByte.
The next 2 patches are also by Oliver Hartkopp and update the virtual
VXCAN driver so that CAN frames send into the peer name space show up
as RX'ed CAN frames.
Vincent Mailhol contributes a patch for the etas_es58x driver to fix a
false positive dereference uninitialized variable warning.
2 patches by Ulrich Hecht add r8a779a0 SoC support to the rcar_canfd
driver.
The remaining 21 patches target the gs_usb driver and are by Peter
Fink, Ben Evans, Eric Evenchick and me. This series cleans up the
gs-usb driver, documents some bits of the USB ABI used by the widely
used open source firmware candleLight, adds support for up to 3 CAN
interfaces per USB device, adds CAN-FD support, adds quirks for some
hardware and software workarounds and finally adds support for 2 new
devices.
* tag 'linux-can-next-for-5.18-20220310' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next: (29 commits)
can: gs_usb: add VID/PID for ABE CAN Debugger devices
can: gs_usb: add VID/PID for CES CANext FD devices
can: gs_usb: add extended bt_const feature
can: gs_usb: activate quirks for CANtact Pro unconditionally
can: gs_usb: add quirk for CANtact Pro overlapping GS_USB_BREQ value
can: gs_usb: add usb quirk for NXP LPC546xx controllers
can: gs_usb: add CAN-FD support
can: gs_usb: use union and FLEX_ARRAY for data in struct gs_host_frame
can: gs_usb: support up to 3 channels per device
can: gs_usb: gs_usb_probe(): introduce udev and make use of it
can: gs_usb: document the PAD_PKTS_TO_MAX_PKT_SIZE feature
can: gs_usb: document the USER_ID feature
can: gs_usb: update GS_CAN_FEATURE_IDENTIFY documentation
can: gs_usb: add HW timestamp mode bit
can: gs_usb: gs_make_candev(): call SET_NETDEV_DEV() after handling all bt_const->feature
can: gs_usb: rewrap usb_control_msg() and usb_fill_bulk_urb()
can: gs_usb: rewrap error messages
can: gs_usb: GS_CAN_FLAG_OVERFLOW: make use of BIT()
can: gs_usb: sort include files alphabetically
can: gs_usb: fix checkpatch warning
...
====================
Link: https://lore.kernel.org/r/20220310142903.341658-1-mkl@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pull networking fixes from Jakub Kicinski:
"Including fixes from bluetooth, and ipsec.
Current release - regressions:
- Bluetooth: fix unbalanced unlock in set_device_flags()
- Bluetooth: fix not processing all entries on cmd_sync_work, make
connect with qualcomm and intel adapters reliable
- Revert "xfrm: state and policy should fail if XFRMA_IF_ID 0"
- xdp: xdp_mem_allocator can be NULL in trace_mem_connect()
- eth: ice: fix race condition and deadlock during interface enslave
Current release - new code bugs:
- tipc: fix incorrect order of state message data sanity check
Previous releases - regressions:
- esp: fix possible buffer overflow in ESP transformation
- dsa: unlock the rtnl_mutex when dsa_master_setup() fails
- phy: meson-gxl: fix interrupt handling in forced mode
- smsc95xx: ignore -ENODEV errors when device is unplugged
Previous releases - always broken:
- xfrm: fix tunnel mode fragmentation behavior
- esp: fix inter address family tunneling on GSO
- tipc: fix null-deref due to race when enabling bearer
- sctp: fix kernel-infoleak for SCTP sockets
- eth: macb: fix lost RX packet wakeup race in NAPI receive
- eth: intel stop disabling VFs due to PF error responses
- eth: bcmgenet: don't claim WOL when its not available"
* tag 'net-5.17-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (50 commits)
xdp: xdp_mem_allocator can be NULL in trace_mem_connect().
ice: Fix race condition during interface enslave
net: phy: meson-gxl: improve link-up behavior
net: bcmgenet: Don't claim WOL when its not available
net: arc_emac: Fix use after free in arc_mdio_probe()
sctp: fix kernel-infoleak for SCTP sockets
net: phy: correct spelling error of media in documentation
net: phy: DP83822: clear MISR2 register to disable interrupts
gianfar: ethtool: Fix refcount leak in gfar_get_ts_info
selftests: pmtu.sh: Kill nettest processes launched in subshell.
selftests: pmtu.sh: Kill tcpdump processes launched by subshell.
NFC: port100: fix use-after-free in port100_send_complete
net/mlx5e: SHAMPO, reduce TIR indication
net/mlx5e: Lag, Only handle events from highest priority multipath entry
net/mlx5: Fix offloading with ESWITCH_IPV4_TTL_MODIFY_ENABLE
net/mlx5: Fix a race on command flush flow
net/mlx5: Fix size field in bufferx_reg struct
ax25: Fix NULL pointer dereference in ax25_kill_by_device
net: marvell: prestera: Add missing of_node_put() in prestera_switch_set_base_mac_addr
net: ethernet: lpc_eth: Handle error for clk_enable
...
Saeed Mahameed says:
====================
mlx5-updates-2022-03-09
1) Remove kernel log prints on FW events regarding FW pages management
and replace that with debugfs entries to track FW pages management commands
failures and general stats, we do that for all FW commands in general since
it's the same effort to do so under the already existing debugfs entry for
FW commands.
2) Add support for ConnectX-7 Software managed steering, in other words STEv2
which shares a lot in common with STE V1, the difference is in specific
offsets in the devices, the logic is almost the same, thus we implement
STEv1 and STEv2 in the same file.
* tag 'mlx5-updates-2022-03-09' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
net/mlx5: DR, Add support for ConnectX-7 steering
net/mlx5: DR, Refactor ste_ctx handling for STE v0/1
net/mlx5: DR, Rename action modify fields to reflect naming in HW spec
net/mlx5: DR, Fix handling of different actions on the same STE in STEv1
net/mlx5: DR, Remove unneeded comments
net/mlx5: DR, Add support for matching on Internet Header Length (IHL)
net/mlx5: DR, Align mlx5dv_dr API vport action with FW behavior
net/mlx5: Add debugfs counters for page commands failures
net/mlx5: Add pages debugfs
net/mlx5: Move debugfs entries to separate struct
net/mlx5: Change release_all_pages cap bit location
net/mlx5: Remove redundant error on reclaim pages
net/mlx5: Remove redundant error on give pages
net/mlx5: Remove redundant notify fail on give pages
net/mlx5: Add command failures data to debugfs
net/mlx5e: TC, Fix use after free in mlx5e_clone_flow_attr_for_post_act()
====================
Link: https://lore.kernel.org/r/20220309213755.610202-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The N_As value describes the time a CAN frame needs on the wire when
transmitted by the CAN controller. Even very short CAN FD frames need
arround 100 usecs (bitrate 1Mbit/s, data bitrate 8Mbit/s).
Having N_As to be zero (the former default) leads to 'no CAN frame
separation' when STmin is set to zero by the receiving node. This 'burst
mode' should not be enabled by default as it could potentially dump a high
number of CAN frames into the netdev queue from the soft hrtimer context.
This does not affect the system stability but is just not nice and
cooperative.
With this N_As/frame_txtime value the 'burst mode' is disabled by default.
As user space applications usually do not set the frame_txtime element
of struct can_isotp_options the new in-kernel default is very likely
overwritten with zero when the sockopt() CAN_ISOTP_OPTS is invoked.
To make sure that a N_As value of zero is only set intentional the
value '0' is now interpreted as 'do not change the current value'.
When a frame_txtime of zero is required for testing purposes this
CAN_ISOTP_FRAME_TXTIME_ZERO u32 value has to be set in frame_txtime.
Link: https://lore.kernel.org/all/20220309120416.83514-2-socketcan@hartkopp.net
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Pull xen fixes from Juergen Gross:
"Several Linux PV device frontends are using the grant table interfaces
for removing access rights of the backends in ways being subject to
race conditions, resulting in potential data leaks, data corruption by
malicious backends, and denial of service triggered by malicious
backends:
- blkfront, netfront, scsifront and the gntalloc driver are testing
whether a grant reference is still in use. If this is not the case,
they assume that a following removal of the granted access will
always succeed, which is not true in case the backend has mapped
the granted page between those two operations.
As a result the backend can keep access to the memory page of the
guest no matter how the page will be used after the frontend I/O
has finished. The xenbus driver has a similar problem, as it
doesn't check the success of removing the granted access of a
shared ring buffer.
- blkfront, netfront, scsifront, usbfront, dmabuf, xenbus, 9p,
kbdfront, and pvcalls are using a functionality to delay freeing a
grant reference until it is no longer in use, but the freeing of
the related data page is not synchronized with dropping the granted
access.
As a result the backend can keep access to the memory page even
after it has been freed and then re-used for a different purpose.
- netfront will fail a BUG_ON() assertion if it fails to revoke
access in the rx path.
This will result in a Denial of Service (DoS) situation of the
guest which can be triggered by the backend"
* tag 'xsa396-5.17-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen/netfront: react properly to failing gnttab_end_foreign_access_ref()
xen/gnttab: fix gnttab_end_foreign_access() without page specified
xen/pvcalls: use alloc/free_pages_exact()
xen/9p: use alloc/free_pages_exact()
xen/usb: don't use gnttab_end_foreign_access() in xenhcd_gnttab_done()
xen: remove gnttab_query_foreign_access()
xen/gntalloc: don't use gnttab_query_foreign_access()
xen/scsifront: don't use gnttab_query_foreign_access() for mapped status
xen/netfront: don't use gnttab_query_foreign_access() for mapped status
xen/blkfront: don't use gnttab_query_foreign_access() for mapped status
xen/grant-table: add gnttab_try_end_foreign_access()
xen/xenbus: don't let xenbus_grant_ring() remove grants in error case
Back when tcp_tso_autosize() and TCP pacing were introduced,
our focus was really to reduce burst sizes for long distance
flows.
The simple heuristic of using sk_pacing_rate/1024 has worked
well, but can lead to too small packets for hosts in the same
rack/cluster, when thousands of flows compete for the bottleneck.
Neal Cardwell had the idea of making the TSO burst size
a function of both sk_pacing_rate and tcp_min_rtt()
Indeed, for local flows, sending bigger bursts is better
to reduce cpu costs, as occasional losses can be repaired
quite fast.
This patch is based on Neal Cardwell implementation
done more than two years ago.
bbr is adjusting max_pacing_rate based on measured bandwidth,
while cubic would over estimate max_pacing_rate.
/proc/sys/net/ipv4/tcp_tso_rtt_log can be used to tune or disable
this new feature, in logarithmic steps.
Tested:
100Gbit NIC, two hosts in the same rack, 4K MTU.
600 flows rate-limited to 20000000 bytes per second.
Before patch: (TSO sizes would be limited to 20000000/1024/4096 -> 4 segments per TSO)
~# echo 0 >/proc/sys/net/ipv4/tcp_tso_rtt_log
~# nstat -n;perf stat ./super_netperf 600 -H otrv6 -l 20 -- -K dctcp -q 20000000;nstat|egrep "TcpInSegs|TcpOutSegs|TcpRetransSegs|Delivered"
96005
Performance counter stats for './super_netperf 600 -H otrv6 -l 20 -- -K dctcp -q 20000000':
65,945.29 msec task-clock # 2.845 CPUs utilized
1,314,632 context-switches # 19935.279 M/sec
5,292 cpu-migrations # 80.249 M/sec
940,641 page-faults # 14264.023 M/sec
201,117,030,926 cycles # 3049769.216 GHz (83.45%)
17,699,435,405 stalled-cycles-frontend # 8.80% frontend cycles idle (83.48%)
136,584,015,071 stalled-cycles-backend # 67.91% backend cycles idle (83.44%)
53,809,530,436 instructions # 0.27 insn per cycle
# 2.54 stalled cycles per insn (83.36%)
9,062,315,523 branches # 137422329.563 M/sec (83.22%)
153,008,621 branch-misses # 1.69% of all branches (83.32%)
23.182970846 seconds time elapsed
TcpInSegs 15648792 0.0
TcpOutSegs 58659110 0.0 # Average of 3.7 4K segments per TSO packet
TcpExtTCPDelivered 58654791 0.0
TcpExtTCPDeliveredCE 19 0.0
After patch:
~# echo 9 >/proc/sys/net/ipv4/tcp_tso_rtt_log
~# nstat -n;perf stat ./super_netperf 600 -H otrv6 -l 20 -- -K dctcp -q 20000000;nstat|egrep "TcpInSegs|TcpOutSegs|TcpRetransSegs|Delivered"
96046
Performance counter stats for './super_netperf 600 -H otrv6 -l 20 -- -K dctcp -q 20000000':
48,982.58 msec task-clock # 2.104 CPUs utilized
186,014 context-switches # 3797.599 M/sec
3,109 cpu-migrations # 63.472 M/sec
941,180 page-faults # 19214.814 M/sec
153,459,763,868 cycles # 3132982.807 GHz (83.56%)
12,069,861,356 stalled-cycles-frontend # 7.87% frontend cycles idle (83.32%)
120,485,917,953 stalled-cycles-backend # 78.51% backend cycles idle (83.24%)
36,803,672,106 instructions # 0.24 insn per cycle
# 3.27 stalled cycles per insn (83.18%)
5,947,266,275 branches # 121417383.427 M/sec (83.64%)
87,984,616 branch-misses # 1.48% of all branches (83.43%)
23.281200256 seconds time elapsed
TcpInSegs 1434706 0.0
TcpOutSegs 58883378 0.0 # Average of 41 4K segments per TSO packet
TcpExtTCPDelivered 58878971 0.0
TcpExtTCPDeliveredCE 9664 0.0
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Link: https://lore.kernel.org/r/20220309015757.2532973-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Having the definitions of {__,}tls_driver_ctx() under an #if
guard means code referencing them also needs to rely on the
preprocessor. The protection doesn't appear needed so make the
definitions unconditional.
Fixes: db37bc177d ("net/funeth: add the data path")
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: kernel test robot <lkp@intel.com>
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Dimitris Michailidis <dmichail@fungible.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add support for a new SW format version that is implemented by
ConnectX-7.
Except for several differences, the STEv2 is identical to STEv1, so for
most callbacks the STEv2 context struct will call STEv1 functions.
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Reviewed-by: Alex Vesker <valex@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add the following new debugfs counters for debug and verbosity:
fw_pages_alloc_failed - number of pages FW requested but driver failed
to allocate.
give_pages_dropped - number of pages given to FW, but command give pages
failed by FW.
reclaim_pages_discard - number of pages which were about to reclaim back
and FW failed the command.
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add pages debugfs to expose the following counters for debuggability:
fw_pages_total - How many pages were given to FW and not returned yet.
vfs_pages - For SRIOV, how many pages were given to FW for virtual
functions usage.
host_pf_pages - For ECPF, how many pages were given to FW for external
hosts physical functions usage.
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Move the debugfs entry pointers under priv to their own struct.
Add get function for device debugfs root.
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
mlx5 FW has changed release_all_pages cap bit by one bit offset to
reflect a fix in the FW flow for release_all_pages. The driver should
use the new bit to ensure it calls release_all_pages only if the FW fix
is there.
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add new counters to command interface debugfs to count command failures.
The following counters added:
total_failed - number of times command failed (any kind of failure).
failed_mbox_status - number of times command failed on bad status
returned by FW.
In addition, add data about last command failure to command interface
debugfs:
last_failed_errno - last command failed returned errno.
last_failed_mbox_status - last bad status returned by FW.
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
SHAMPO is an RQ / WQ feature, an indication was added to the TIR in the
first place to enforce suitability between connected TIR and RQ, this
enforcement does not exist in current the Firmware implementation and was
redundant in the first place.
Fixes: 83439f3c37 ("net/mlx5e: Add HW-GRO offload")
Signed-off-by: Ben Ben-Ishay <benishay@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
According to HW spec the field "size" should be 16 bits
in bufferx register.
Fixes: e281682bf2 ("net/mlx5_core: HW data structs/types definitions cleanup")
Signed-off-by: Mohammad Kabat <mohammadkab@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Steffen Klassert says:
====================
pull request (net): ipsec 2022-03-09
1) Fix IPv6 PMTU discovery for xfrm interfaces.
From Lina Wang.
2) Revert failing for policies and states that are
configured with XFRMA_IF_ID 0. It broke a
user configuration. From Kai Lueke.
3) Fix a possible buffer overflow in the ESP output path.
4) Fix ESP GSO for tunnel and BEET mode on inter address
family tunnels.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
We have a number of cases where function returns drop/no drop
decision as a boolean. Now that we want to report the reason
code as well we have to pass extra output arguments.
We can make the reason code evaluate correctly as bool.
I believe we're good to reorder the reasons as they are
reported to user space as strings.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Felix driver declares FDB isolation but puts all standalone ports in
VID 0. This is mostly problem-free as discussed with Alvin here:
https://patchwork.kernel.org/project/netdevbpf/cover/20220302191417.1288145-1-vladimir.oltean@nxp.com/#24763870
however there is one catch. DSA still thinks that FDB entries are
installed on the CPU port as many times as there are user ports, and
this is problematic when multiple user ports share the same MAC address.
Consider the default case where all user ports inherit their MAC address
from the DSA master, and then the user runs:
ip link set swp0 address 00:01:02:03:04:05
The above will make dsa_slave_set_mac_address() call
dsa_port_standalone_host_fdb_add() for 00:01:02:03:04:05 in port 0's
standalone database, and dsa_port_standalone_host_fdb_del() for the old
address of swp0, again in swp0's standalone database.
Both the ->port_fdb_add() and ->port_fdb_del() will be propagated down
to the felix driver, which will end up deleting the old MAC address from
the CPU port. But this is still in use by other user ports, so we end up
breaking unicast termination for them.
There isn't a problem in the fact that DSA keeps track of host
standalone addresses in the individual database of each user port: some
drivers like sja1105 need this. There also isn't a problem in the fact
that some drivers choose the same VID/FID for all standalone ports.
It is just that the deletion of these host addresses must be delayed
until they are known to not be in use any longer, and only the driver
has this knowledge. Since DSA keeps these addresses in &cpu_dp->fdbs and
&cpu_db->mdbs, it is just a matter of walking over those lists and see
whether the same MAC address is present on the CPU port in the port db
of another user port.
I have considered reusing the generic dsa_port_walk_fdbs() and
dsa_port_walk_mdbs() schemes for this, but locking makes it difficult.
In the ->port_fdb_add() method and co, &dp->addr_lists_lock is held, but
dsa_port_walk_fdbs() also acquires that lock. Also, even assuming that
we introduce an unlocked variant of the address iterator, we'd still
need some relatively complex data structures, and a void *ctx in the
dsa_fdb_walk_cb_t which we don't currently pass, such that drivers are
able to figure out, after iterating, whether the same MAC address is or
isn't present in the port db of another port.
All the above, plus the fact that I expect other drivers to follow the
same model as felix where all standalone ports use the same FID, made me
conclude that a generic method provided by DSA is necessary:
dsa_fdb_present_in_other_db() and the mdb equivalent. Felix calls this
from the ->port_fdb_del() handler for the CPU port, when the database
was classified to either a port db, or a LAG db.
For symmetry, we also call this from ->port_fdb_add(), because if the
address was installed once, then installing it a second time serves no
purpose: it's already in hardware in VID 0 and it affects all standalone
ports.
This change moves dsa_db_equal() from switch.c to dsa.c, since it now
has one more caller.
Fixes: 54c3198460 ("net: mscc: ocelot: enforce FDB isolation when VLAN-unaware")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In some edge scenarios, an MPTCP subflows can use a local address
mapped by a "implicit" endpoint created by the in-kernel path manager.
Such endpoints presence can be confusing, as it's creation is hard
to track and will prevent the later endpoint creation from the user-space
using the same address.
Define a new endpoint flag to mark implicit endpoints and allow the
user-space to replace implicit them with user-provided data at endpoint
creation time.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The tracepoint in get_mapping_status() only dumped the incoming mpext
fields. This patch added a new tracepoint in mptcp_sendmsg_frag() to dump
the outgoing mpext too.
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
