We periodically ask the neighbouring system to try and resolve
neighbours that are used for nexthops, but aren't currently resolved.
However, 'nud_state' is protected by the neighbour lock, so we shouldn't
access it without taking it. Instead, we can simply check the
'connected' field of the neighbour entry, which we update upon
NEIGH_UPDATE events.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We only add neighbour entries that are also used for nexthops to
'nexthop_neighs_list', so when iterating over this list there's no need
to check that the entry is indeed used for nexthops.
Remove the redundant check.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Up until now we had two interfaces for neighbour related configuration:
ndo_neigh_{construct,destroy} and NEIGH_UPDATE netevents. The ndos were
used to add and remove neighbours from the driver's cache, whereas the
netevent was used to reflect the neighbours into the device's tables.
However, if the NUD state of a neighbour isn't NUD_VALID or if the
neighbour is dead, then there's really no reason for us to keep it
inside our cache. The only exception to this rule are neighbours that
are also used for nexthops, which we periodically refresh to get them
resolved.
We can therefore eliminate the ndo entry point into the driver and
simplify the code, making it similar to the FIB reflection, which is
based solely on events. This also helps us avoid a locking issue, in
which the RIF cache was traversed without proper locking during
insertion into the neigh entry cache.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit 33b1341cd1 ("mlxsw: spectrum_router: Fix handling of
neighbour structure") we no longer use destination IP for neighbour
lookup, so remove it.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We currently associate each neighbour entry with a work item, so it's
not possible to have multiple events queued for the same neighbour
entry. However, this is about to be changed so that the neighbour entry
is only resolved when the work item is scheduled.
The above can result in a mismatch between the kernel's and the device's
neighbour table, unless the associated work items are processed in the
order in which they were submitted.
Do that by migrating the NEIGH_UPDATE work items to be processed in the
ordered workqueue which was recently introduced in mlxsw in commit
a3832b3189 ("mlxsw: core: Create an ordered workqueue for FIB
offload").
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We always use zero delay before queueing a work on the ordered workqueue
('mlxsw_owq'), so use work_struct directly instead of delayable work.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4K Uar series modified the mlx5e driver to use the new bfreg API,
and mistakenly removed the sq->uar_map iomem data path dedicated
pointer, which was meant to be read from xmit path for cache locality
utilization.
Fix that by returning that pointer to the SQ struct.
Fixes: 7309cb4ad71e ("IB/mlx5: Support 4k UAR for libmlx5")
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
ConnectX-5 and later HW generations will report min inline mode ==
MLX5_INLINE_MODE_NONE, which means driver is not required to copy packet
headers to inline fields of TX WQE.
Avoid copy to inline segment in XDP TX routine when HW inline mode doesn't
require it.
This will improve CPU utilization and boost XDP TX performance.
Tested with xdp2 single flow:
CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
HCA: Mellanox Technologies MT28800 Family [ConnectX-5 Ex]
Before: 7.4Mpps
After: 7.8Mpps
Improvement: 5%
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
ConnectX-5 and later HW generations will report min inline mode ==
MLX5_INLINE_MODE_NONE, which means driver is not required to copy packet
headers to inline fields of TX WQE.
When inline is not required, vlan insertion will be handled in the
TX descriptor rather than copy to inline.
For LSO case driver is still required to copy headers, for the HW to
duplicate on wire.
This will improve CPU utilization and boost TX performance.
Tested with pktgen burst single flow:
CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
HCA: Mellanox Technologies MT28800 Family [ConnectX-5 Ex]
Before: 15.1Mpps
After: 17.2Mpps
Improvement: 14%
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Add new TX WQE fields for Connect-X5 vlan insertion support,
type and vlan_tci, when type = MLX5_ETH_WQE_INSERT_VLAN the
HW will insert the vlan and prio fields (vlan_tci) to the packet.
Those bits and the inline header fields are mutually exclusive, and
valid only when:
MLX5_CAP_ETH(mdev, wqe_inline_mode) == MLX5_CAP_INLINE_MODE_NOT_REQUIRED
and MLX5_CAP_ETH(mdev, wqe_vlan_insert),
who will be set in ConnectX-5 and later HW generations.
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
There is a hardware feature that will pad the start or end of a DMA to
be cache line aligned to avoid RMWs on the last cache line. The default
cache line size setting for this feature is 64B. This change configures
the hardware to use 128B alignment on systems with 128B cache lines.
In addition we lower bound MPWRQ stride by HCA cacheline in mlx5e,
MPWRQ stride should be at least the HCA cacheline, the current default
is 64B and in case HCA_CAP.cach_line_128byte capability is set, MPWRQ RX
stride will automatically be aligned to 128B.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
We plan to remove __napi_complete() soon,
this driver is the last user.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete()
We plan to remove __napi_complete() to reduce NAPI complexity.
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete() to :
1) Get support of gro_flush_timeout if opt-in
2) Not rearm interrupts for busy-polling users.
3) use standard NAPI API.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete() to :
1) Get support of gro_flush_timeout if opt-in
2) Not rearm interrupts for busy-polling users.
3) use standard NAPI API.
Note that rx_lock seems to be useless, NAPI logic should
not need this extra care.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete() to :
1) Get support of gro_flush_timeout if opt-in
2) Not rearm interrupts for busy-polling users.
3) use standard NAPI API and get rid of napi_gro_flush()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete() to :
1) Get support of gro_flush_timeout if opt-in
2) Not rearm interrupts for busy-polling users.
3) use standard NAPI API.
4) get rid of baroque code and ease maintenance.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete() to :
1) Get support of gro_flush_timeout if opt-in
2) Not rearm interrupts for busy-polling users.
3) use standard NAPI API.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete() to :
1) Get support of gro_flush_timeout if opt-in
2) Not rearm interrupts for busy-polling users.
3) use standard NAPI API.
4) get rid of baroque code and ease maintenance.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete() to :
1) Get support of gro_flush_timeout if opt-in
2) Not rearm interrupts for busy-polling users.
3) use standard NAPI API.
4) get rid of baroque code and ease maintenance.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete() to :
1) Get support of gro_flush_timeout if opt-in
2) Not rearm interrupts for busy-polling users.
3) use standard NAPI API.
4) Eventually get rid of napi_gro_flush() in the future.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use napi_complete_done() instead of __napi_complete() to :
1) Get support of gro_flush_timeout if opt-in
2) Not rearm interrupts for busy-polling users.
3) use standard NAPI API.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jeff Kirsher says:
====================
40GbE Intel Wired LAN Driver Updates 2017-02-03
This series contains updates to i40e/i40evf only.
Jake fixes up the driver to not call i40e_vsi_kill_vlan() or
i40e_vsi_add_vlan() when the PVID is set or when the VID is less than 1.
Cleaned up a check which really is not needed since there is no real
reason why we cannot just call i40e_del_mac_all_vlan() directly. Renamed
functions to better reflect their actual purpose and how they function
in a more clear manner.
Bimmy cleans up unused/deprecated macros.
Mitch cleans up unused device ids which were intended for use when
running Linux VF drivers under Hyper-V, but found to be not needed.
Then cleaned up a function that is no longer needed since the client
open and close functions were refactored. Adds a sleep without timeout
until the reply from the PF driver has been received since the iWARP
client cannot continue until the operation has been completed.
Tushar Dave fixes an issue seen on SPARC where the use of the 'packed'
directive was causing kernel unaligned errors.
Alex does a refactor to pull some data off of the stack and store it
in the transmit buffer info section of the transmit ring.
Alan fixes a bug which was caused by passing a bad register value to the
firmware, by refactoring the macro INTRL_USEC_TO_REG into a static
inline function. Also added feedback to the user as to the actual
interrupt rate limit being used when it differs from the requested limit.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
In linux-4.5, busy polling was implemented in core
NAPI stack, meaning that all custom implementation can
be removed from drivers.
Not only we remove lot's of code, we also remove one lock
operation in fast path, and allow GRO to do its job.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Acked-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In linux-4.5, busy polling was implemented in core
NAPI stack, meaning that all custom implementation can
be removed from drivers.
Not only we remove lot's of code, we also remove one lock
operation in fast path, and allow GRO to do its job.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Acked-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Extend the existing setup_tc ndo call and allow to offload cls_flower
rules. Only limited set of dissector keys and actions are supported now.
Use previously introduced ACL infrastructure to offload cls_flower rules
to be processed in the HW.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add ACL core infrastructure for Spectrum ASIC. This infra provides an
abstraction layer over specific HW implementations. There are two basic
objects used. One is "rule" and the second is "ruleset" which serves as a
container of multiple rules. In general, within one ruleset the rules are
allowed to have multiple priorities and masks. Each ruleset is bound to
either ingress or egress a of port netdevice.
The initial TCAM implementation is very simple and limited. It utilizes
parman lsort manager to take care of TCAM region layout.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Introduce basic set of Spectrum flexible key blocks. It contains blocks
needed to carry all elements defined so far.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Each entry which is matched during ACL lookup points to an action set.
This action set contains up to three separate actions. If more actions
are needed to be chained, the extended set is created to hold them
in KVD linear area.
This patch implements handling of sets and encoding of actions.
Currectly, only two actions are supported. Drop and forward. Forward
action uses PBS pointer to KVD linear area, so the action code needs to
take care of this as well.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Hardware supports matching on so called "flexible keys". The idea is to
assemble an optimal key to use for matching according to the fields in
packet (elements) requested by user. Certain sets of elements are
combined into pre-defined blocks. There is a picker to find needed blocks.
Keys consist of 1..n blocks.
Alongside with that, an initial portion of elements is introduced in order
to be able to offload basic cls_flower rules.
Picked keys are cached so multiple rules could share them.
There is an encode function provided that takes care of encoding key and
mask values according to given key.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
PEFA register is used for accessing an extended flexible action entry
in the central KVD Linear Database.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The PPBS register retrieves and sets Policy Based Switching Table entries.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The PTCE-V2 register is used for accessing rules within a TCAM region.
It is a new version of PTCE in order to support wider key, mask and
action within a TCAM region.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sometimes it is handy to get a pointer to a char buffer item and use it
direcly to write/read data. So add these helpers.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In linux-4.5, busy polling was implemented in core
NAPI stack, meaning that all custom implementation can
be removed from drivers.
Not only we remove lot of code, we also remove one spin_lock()
from driver fast path.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Ganesh Goudar <ganeshgr@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Compared to custom busy_poll, the generic NAPI one is simpler and
removes a lot of code. It removes one atomic in the fast path (when
busy poll is not in action) since we do not have to use an extra
spinlock.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On large SMP builds, we can run into a build warning:
drivers/net/ethernet/hisilicon/hns/hns_enet.c: In function 'hns_set_irq_affinity.isra.27':
drivers/net/ethernet/hisilicon/hns/hns_enet.c:1242:1: warning: the frame size of 1032 bytes is larger than 1024 bytes [-Wframe-larger-than=]
The solution here is to use cpumask_var_t, which can use dynamic
allocation when CONFIG_CPUMASK_OFFSTACK is enabled.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Without any uld being loaded, uld_txq_info[] will be NULL. uld_send()
is also used for sending control work requests(for eg: setting filter)
that dont require any ulds to be loaded. Hence move uld_txq_info[]
assignment after ctrl_xmit().
Also added a NULL check for uld_txq_info[].
Fixes: 94cdb8bb99 (cxgb4: Add support for dynamic allocation
of resources for ULD).
Signed-off-by: Arjun V <arjun@chelsio.com>
Signed-off-by: Casey Leedom <leedom@chelsio.com>
Signed-off-by: Ganesh Goudar <ganeshgr@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
