1. Added support to filter packets based on IP fragment.
For IPv4 packets check for ip_flag == 0x20 (more fragment bit set).
For IPv6 packets check for next_header == 0x2c (next_header set to
'fragment header for IPv6')
2. Added configuration support from both "ethtool ntuple" and "tc flower".
Signed-off-by: NSuman Ghosh <sumang@marvell.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

1. If a profile does not support DMAC extraction then avoid installing NPC
flow rules for unicast. Similarly, if LXMB(L2 and L3) extraction is not
supported by the profile then avoid installing broadcast and multicast
rules.
2. Allow MCAM entry insertion for promiscuous mode.
3. For the profiles where DMAC is not extracted in MKEX key default
unicast entry installed by AF is not valid. Hence do not use action
from the AF installed default unicast entry for such cases.
4. Adjacent packet header fields in a packet like IP header source
and destination addresses or UDP/TCP header source port and destination
can be extracted together in MKEX profile. Therefore MKEX profile can be
configured to in two ways:
	a. Total of 4 bytes from start of UDP header(src port
	   + destination port)
	or
	b. Two bytes from start and two bytes from offset 2
Signed-off-by: NSuman Ghosh <sumang@marvell.com>
Signed-off-by: NSubbaraya Sundeep <sbhatta@marvell.com>
Link: https://lore.kernel.org/r/20221118053329.2288486-1-sumang@marvell.comSigned-off-by: NPaolo Abeni <pabeni@redhat.com>

octep_get_mac_addr() can fail because send mbox message failed. If this
happens, octep_dev->mac_addr will be zero. It should not continue to
initialize. Add exception handling for octep_get_mac_addr() to fix it.

Fixes: 862cd659 ("octeon_ep: Add driver framework and device initialization")
Signed-off-by: NZiyang Xuan <william.xuanziyang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When occur unsupported_dev and mbox init errors, it did not free oct->conf
and iounmap() oct->mmio[i].hw_addr. That would trigger memory leak problem.
Add kfree() for oct->conf and iounmap() for oct->mmio[i].hw_addr under
unsupported_dev and mbox init errors to fix the problem.

Fixes: 862cd659 ("octeon_ep: Add driver framework and device initialization")
Signed-off-by: NZiyang Xuan <william.xuanziyang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

octep_get_link_status() can fail because send mbox message failed, then
octep_get_link_status() will return ret less than 0. Excute octep_link_up()
as long as ret is not equal to 0 in octep_open() now. That is not correct.

The value type of link.state is enum octep_ctrl_net_state. Positive value
represents up. Excute octep_link_up() when ret is bigger than 0.

Fixes: 862cd659 ("octeon_ep: Add driver framework and device initialization")
Signed-off-by: NZiyang Xuan <william.xuanziyang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

octep_napi_add() and octep_napi_enable() are all after
netif_set_real_num_{tx,rx}_queues() in octep_open(), so it is unnecessary
napi rollback under set_queues_err. Delete them to fix it.

Fixes: 37d79d05 ("octeon_ep: add Tx/Rx processing and interrupt support")
Signed-off-by: NZiyang Xuan <william.xuanziyang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

skb->vlan_present seems redundant.

We can instead derive it from this boolean expression:

vlan_present = skb->vlan_proto != 0 || skb->vlan_tci != 0

Add a new union, to access both fields in a single load/store
when possible.

	union {
		u32	vlan_all;
		struct {
		__be16	vlan_proto;
		__u16	vlan_tci;
		};
	};

This allows following patch to remove a conditional test in GRO stack.

Note:
  We move remcsum_offload to keep TC_AT_INGRESS_MASK
  and SKB_MONO_DELIVERY_TIME_MASK unchanged.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Acked-by: NYonghong Song <yhs@fb.com>
Acked-by: NMartin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

Bump MIN version to reflect support of new platform (AC5X family devices).
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add support for the following AC5x Marvell Prestera PP family devices:
  98DX7312M (12x25G / 8x25G + 1x100G);
  98DX3500  (24x1G + 6x25G);
  98DX3501  (16x1G + 6x10G);
  98DX3510  (48x1G + 6x25G);
  98DX3520  (24x2.5G + 6x25G);

Known issues:
- FW reload doesn't work (rmmod/modprobe sequence).
Co-developed-by: NVadym Kochan <vadym.kochan@plvision.eu>
Signed-off-by: NVadym Kochan <vadym.kochan@plvision.eu>
Signed-off-by: NMaksym Glubokiy <maksym.glubokiy@plvision.eu>
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use defines with proper device names instead of device-id in pci-devices
listing.
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When failed to init rxq or txq in mv643xx_eth_open() for opening device,
napi isn't disabled. When open mv643xx_eth device next time, it will
trigger a BUG_ON() in napi_enable(). Compile tested only.

Fixes: 2257e05c ("mv643xx_eth: get rid of receive-side locking")
Signed-off-by: NZhengchao Shao <shaozhengchao@huawei.com>
Link: https://lore.kernel.org/r/20221109025432.80900-1-shaozhengchao@huawei.comSigned-off-by: NPaolo Abeni <pabeni@redhat.com>

When prestera_sdma_switch_init() failed, the memory pointed to by
sw->rxtx isn't released. Fix it. Only be compiled, not be tested.

Fixes: 501ef306 ("net: marvell: prestera: Add driver for Prestera family ASIC devices")
Signed-off-by: NZhengchao Shao <shaozhengchao@huawei.com>
Reviewed-by: NVadym Kochan <vadym.kochan@plvision.eu>
Link: https://lore.kernel.org/r/20221108025607.338450-1-shaozhengchao@huawei.comSigned-off-by: NJakub Kicinski <kuba@kernel.org>

Current way of checking available SQE count which is based on
HW updated SQB count could result in driver submitting an SQE
even before CQE for the previously transmitted SQE at the same
index is processed in NAPI resulting losing SKB pointers,
hence a leak. Fix this by checking a consumer index which
is updated once CQE is processed.

Fixes: 3ca6c4c8 ("octeontx2-pf: Add packet transmission support")
Signed-off-by: NRatheesh Kannoth <rkannoth@marvell.com>
Reviewed-by: NSunil Kovvuri Goutham <sgoutham@marvell.com>
Link: https://lore.kernel.org/r/20221107033505.2491464-1-rkannoth@marvell.comSigned-off-by: NPaolo Abeni <pabeni@redhat.com>

If CONFIG_MACSEC=m and CONFIG_OCTEONTX2_PF=y, it leads a build error:

  ld: drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.o: in function `otx2_pfaf_mbox_up_handler':
  otx2_pf.c:(.text+0x181c): undefined reference to `cn10k_handle_mcs_event'
  ld: drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.o: in function `otx2_probe':
  otx2_pf.c:(.text+0x437e): undefined reference to `cn10k_mcs_init'
  ld: drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.o: in function `otx2_remove':
  otx2_pf.c:(.text+0x5031): undefined reference to `cn10k_mcs_free'
  ld: drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.o: in function `otx2_mbox_up_handler_mcs_intr_notify':
  otx2_pf.c:(.text+0x5f11): undefined reference to `cn10k_handle_mcs_event'

Make CONFIG_OCTEONTX2_PF depends on CONFIG_MACSEC to fix it. Because
it has empty stub functions of cn10k, CONFIG_OCTEONTX2_PF can be enabled
if CONFIG_MACSEC is disabled

Fixes: c54ffc73 ("octeontx2-pf: mcs: Introduce MACSEC hardware offloading")
Reported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221105063442.2013981-1-yangyingliang@huawei.comSigned-off-by: NJakub Kicinski <kuba@kernel.org>

Virtually all conventional network drivers are now converted to use
phylink_generic_validate() - only DSA drivers and fman_memac remain,
so lets remove the necessity for network drivers to explicitly set
this member, and default to phylink_generic_validate() when unset.
This is possible as .validate must currently be set.

Any remaining instances that have not been addressed by this patch can
be fixed up later.
Signed-off-by: NRussell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/E1or0FZ-001tRa-DI@rmk-PC.armlinux.org.ukSigned-off-by: NJakub Kicinski <kuba@kernel.org>

Support mode switch properly, which is not available before.

If SoC has two Ethernet controllers, by setting both of them into MII
mode, the first controller enters GMII mode, while the second
controller is effectively disabled. This requires configuring (and
maybe enabling) the second controller in the device tree, even though
it cannot be used.
Signed-off-by: NDavid Yang <mmyangfl@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add support for Octeon device CNF95N.
CNF95N is a Octeon Fusion family product with same PCI NIC
characteristics as CN93 which is currently supported by the driver.

update supported device list in Documentation.
Signed-off-by: NVeerasenareddy Burru <vburru@marvell.com>
Link: https://lore.kernel.org/r/20221103060600.1858-1-vburru@marvell.comSigned-off-by: NJakub Kicinski <kuba@kernel.org>

In scenarios where multiple errors have occurred
for a SQ before SW starts handling error interrupt,
SQ_CTX[OP_INT] may get overwritten leading to
NIX_LF_SQ_OP_INT returning incorrect value.
To workaround this read LMT, MNQ and SQ individual
error status registers to determine the cause of error.

Fixes: 4ff7d148 ("octeontx2-pf: Error handling support")
Signed-off-by: NRatheesh Kannoth <rkannoth@marvell.com>
Reviewed-by: NSunil Kovvuri Goutham <sgoutham@marvell.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix build errors when MACSEC=m and OCTEONTX2_PF=y by having
OCTEONTX2_PF depend on MACSEC if it is enabled. By adding
"|| !MACSEC", this means that MACSEC is not required -- it can
be disabled for this driver.

drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.o: in function `otx2_remove':
../drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c:(.text+0x2fd0): undefined reference to `cn10k_mcs_free'
mips64-linux-ld: drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.o: in function `otx2_mbox_up_handler_mcs_intr_notify':
../drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c:(.text+0x4610): undefined reference to `cn10k_handle_mcs_event'
Reported-by: Nkernel test robot <lkp@intel.com>
Fixes: c54ffc73 ("octeontx2-pf: mcs: Introduce MACSEC hardware offloading")
Signed-off-by: NRandy Dunlap <rdunlap@infradead.org>
Cc: Subbaraya Sundeep <sbhatta@marvell.com>
Cc: Sunil Goutham <sgoutham@marvell.com>
Cc: Geetha sowjanya <gakula@marvell.com>
Cc: hariprasad <hkelam@marvell.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Remove ndo_get_devlink_port which is no longer used alongside with the
implementations in drivers.
Signed-off-by: NJiri Pirko <jiri@nvidia.com>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

Benefit from the previously implemented tracking of netdev events in
devlink code and instead of calling  devlink_port_type_eth_set() and
devlink_port_type_clear() to set devlink port type and link to related
netdev, use SET_NETDEV_DEVLINK_PORT() macro to assign devlink_port
pointer to netdevice which is about to be registered.
Signed-off-by: NJiri Pirko <jiri@nvidia.com>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

1. It is possible to have custom mkex profiles which do not extract
DMAC at all into the key. Hence allow mkex profiles which do not
have DMAC to be loaded into MCAM hardware. This patch also adds
debugging prints needed to identify profiles with wrong
configuration.

2. If a mkex profile set "l2l3mb" field for Rx interface,
then Rx multicast and broadcast entry should be configured.
Signed-off-by: NSuman Ghosh <sumang@marvell.com>
Link: https://lore.kernel.org/r/20221031090856.1404303-1-sumang@marvell.comSigned-off-by: NJakub Kicinski <kuba@kernel.org>

Variable i is just being incremented and it's never used anywhere else. The
variable and the increment are redundant so remove it.
Signed-off-by: NColin Ian King <colin.i.king@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Now that the 32bit UP oddity is gone and 32bit uses always a sequence
count, there is no need for the fetch_irq() variants anymore.

Convert to the regular interface.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Acked-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

fix rxsc and txsc not getting freed before going out of scope

Fixes: c54ffc73 ("octeontx2-pf: mcs: Introduce MACSEC hardware offloading")
Signed-off-by: NManank Patel <pmanank200502@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The __prestera_nexthop_group_create() function returns NULL on error
and the prestera_nexthop_group_get() returns error pointers.  Fix these
two checks.

Fixes: 0a23ae23 ("net: marvell: prestera: Add router nexthops ABI")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/Y0bWq+7DoKK465z8@kiliSigned-off-by: NJakub Kicinski <kuba@kernel.org>

Fixes: 13defa27 ("net: marvell: prestera: Add matchall support")
Signed-off-by: NMaksym Glubokiy <maksym.glubokiy@plvision.eu>
Link: https://lore.kernel.org/r/20221006190600.881740-1-maksym.glubokiy@plvision.euSigned-off-by: NJakub Kicinski <kuba@kernel.org>

In error path after calling cn10k_mcs_init(), cn10k_mcs_free() need
be called to avoid memory leak.

Fixes: c54ffc73 ("octeontx2-pf: mcs: Introduce MACSEC hardware offloading")
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If alloc_mem() fails in mcs_register_interrupts(), it should return error
code.

Fixes: 6c635f78 ("octeontx2-af: cn10k: mcs: Handle MCS block interrupts")
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add the missing unlock in some error paths.

Fixes: c54ffc73 ("octeontx2-pf: mcs: Introduce MACSEC hardware offloading")
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If you try to create a 'mirror' ACL rule on a port that already has a
mirror rule, prestera_span_rule_add() will fail with EEXIST error.

This forces rollback procedure which destroys existing mirror rule on
hardware leaving it visible in linux.

Add an explicit check for EEXIST to prevent the deletion of the existing
rule but keep user seeing error message:

  $ tc filter add dev sw1p1 ... skip_sw action mirred egress mirror dev sw1p2
  $ tc filter add dev sw1p1 ... skip_sw action mirred egress mirror dev sw1p3
  RTNETLINK answers: File exists
  We have an error talking to the kernel

Fixes: 13defa27 ("net: marvell: prestera: Add matchall support")
Signed-off-by: NSerhiy Boiko <serhiy.boiko@plvision.eu>
Signed-off-by: NMaksym Glubokiy <maksym.glubokiy@plvision.eu>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Semicolon is not required after curly braces.

Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=2332Reported-by: NAbaci Robot <abaci@linux.alibaba.com>
Signed-off-by: NYang Li <yang.lee@linux.alibaba.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We poll nexthops in HW and call for each active nexthop appropriate
neighbour.

Also we provide implicity neighbour resolving.
For example, user have added nexthop route:
  # ip route add 5.5.5.5 via 1.1.1.2
But neighbour 1.1.1.2 doesn't exist. In this case we will try to call
neigh_event_send, even if there is no traffic.
This is useful, when you have add route, which will be used after some
time but with a lot of traffic (burst). So, we has prepared, offloaded
route in advance.
Co-developed-by: NTaras Chornyi <tchornyi@marvell.com>
Signed-off-by: NTaras Chornyi <tchornyi@marvell.com>
Co-developed-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NYevhen Orlov <yevhen.orlov@plvision.eu>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

Move forward and use new PRESTERA_FIB_TYPE_UC_NH to provide basic
nexthop routes support.
Provide deinitialization sequence for all created router objects.

Limitations:
- Only "local" and "main" tables supported
- Only generic interfaces supported for router (no bridges or vlans)
Co-developed-by: NTaras Chornyi <tchornyi@marvell.com>
Signed-off-by: NTaras Chornyi <tchornyi@marvell.com>
Co-developed-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NYevhen Orlov <yevhen.orlov@plvision.eu>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

Actual handler will be added in next patches
Co-developed-by: NTaras Chornyi <tchornyi@marvell.com>
Signed-off-by: NTaras Chornyi <tchornyi@marvell.com>
Co-developed-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NYevhen Orlov <yevhen.orlov@plvision.eu>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

This will be used to implement nexthops related logic in next patches.
Also try to keep ipv4/6 abstraction to be able to reuse helpers for ipv6
in the future.
Co-developed-by: NTaras Chornyi <tchornyi@marvell.com>
Signed-off-by: NTaras Chornyi <tchornyi@marvell.com>
Co-developed-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NYevhen Orlov <yevhen.orlov@plvision.eu>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

Add macros to determine IP address length (internal driver types).
This will be used in next patches for nexthops logic.
Co-developed-by: NTaras Chornyi <tchornyi@marvell.com>
Signed-off-by: NTaras Chornyi <tchornyi@marvell.com>
Co-developed-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NYevhen Orlov <yevhen.orlov@plvision.eu>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

Flushing workqueues ensures, that no more pending works, related to just
unregistered or deinitialized notifiers. After that we can free memory.

Delayed wq will be used for neighbours in next patches.
Co-developed-by: NTaras Chornyi <tchornyi@marvell.com>
Signed-off-by: NTaras Chornyi <tchornyi@marvell.com>
Co-developed-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NYevhen Orlov <yevhen.orlov@plvision.eu>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

This will, ensure, that there is no more, preciously allocated fib_cache
entries left after deinit.
Will be used to free allocated resources of nexthop routes, that points
to "not our" port (e.g. eth0).
Signed-off-by: NYevhen Orlov <yevhen.orlov@plvision.eu>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>

Do explicity cleanup on router_hw_fini, to ensure, that all allocated
objects cleaned. This will be used in cases,
when upper layer (cache) is not mapped to router_hw layer.
Co-developed-by: NTaras Chornyi <tchornyi@marvell.com>
Signed-off-by: NTaras Chornyi <tchornyi@marvell.com>
Co-developed-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NOleksandr Mazur <oleksandr.mazur@plvision.eu>
Signed-off-by: NYevhen Orlov <yevhen.orlov@plvision.eu>
Signed-off-by: NJakub Kicinski <kuba@kernel.org>