Saeed Mahameed says:
====================
pull-request: mlx5-next 2020-08-9
This pulls mlx5-next branch which includes patches already reviewed on
net-next and rdma mailing lists.
1) mlx5 single E-Switch FDB for lag
2) IB/mlx5: Rename is_apu_thread_cq function to is_apu_cq
3) Add DCS caps & fields support
[1] https://patchwork.kernel.org/project/netdevbpf/cover/20210803231959.26513-1-saeed@kernel.org/
[2] 0e3364dab7.1626609184.git.leonro@nvidia.com/
[3] 55e1d69bef.1624258894.git.leonro@nvidia.com/
* 'mlx5-next' of git://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux:
net/mlx5: Lag, Create shared FDB when in switchdev mode
net/mlx5: E-Switch, add logic to enable shared FDB
net/mlx5: Lag, move lag destruction to a workqueue
net/mlx5: Lag, properly lock eswitch if needed
net/mlx5: Add send to vport rules on paired device
net/mlx5: E-Switch, Add event callback for representors
net/mlx5e: Use shared mappings for restoring from metadata
net/mlx5e: Add an option to create a shared mapping
net/mlx5: E-Switch, set flow source for send to uplink rule
RDMA/mlx5: Add shared FDB support
{net, RDMA}/mlx5: Extend send to vport rules
RDMA/mlx5: Fill port info based on the relevant eswitch
net/mlx5: Lag, add initial logic for shared FDB
net/mlx5: Return mdev from eswitch
IB/mlx5: Rename is_apu_thread_cq function to is_apu_cq
net/mlx5: Add DCS caps & fields support
====================
Link: https://lore.kernel.org/r/20210809202522.316930-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Daniel Borkmann says:
====================
bpf-next 2021-08-10
We've added 31 non-merge commits during the last 8 day(s) which contain
a total of 28 files changed, 3644 insertions(+), 519 deletions(-).
1) Native XDP support for bonding driver & related BPF selftests, from Jussi Maki.
2) Large batch of new BPF JIT tests for test_bpf.ko that came out as a result from
32-bit MIPS JIT development, from Johan Almbladh.
3) Rewrite of netcnt BPF selftest and merge into test_progs, from Stanislav Fomichev.
4) Fix XDP bpf_prog_test_run infra after net to net-next merge, from Andrii Nakryiko.
5) Follow-up fix in unix_bpf_update_proto() to enforce socket type, from Cong Wang.
6) Fix bpf-iter-tcp4 selftest to print the correct dest IP, from Jose Blanquicet.
7) Various misc BPF XDP sample improvements, from Niklas Söderlund, Matthew Cover,
and Muhammad Falak R Wani.
* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (31 commits)
bpf, tests: Add tail call test suite
bpf, tests: Add tests for BPF_CMPXCHG
bpf, tests: Add tests for atomic operations
bpf, tests: Add test for 32-bit context pointer argument passing
bpf, tests: Add branch conversion JIT test
bpf, tests: Add word-order tests for load/store of double words
bpf, tests: Add tests for ALU operations implemented with function calls
bpf, tests: Add more ALU64 BPF_MUL tests
bpf, tests: Add more BPF_LSH/RSH/ARSH tests for ALU64
bpf, tests: Add more ALU32 tests for BPF_LSH/RSH/ARSH
bpf, tests: Add more tests of ALU32 and ALU64 bitwise operations
bpf, tests: Fix typos in test case descriptions
bpf, tests: Add BPF_MOV tests for zero and sign extension
bpf, tests: Add BPF_JMP32 test cases
samples, bpf: Add an explict comment to handle nested vlan tagging.
selftests/bpf: Add tests for XDP bonding
selftests/bpf: Fix xdp_tx.c prog section name
net, core: Allow netdev_lower_get_next_private_rcu in bh context
bpf, devmap: Exclude XDP broadcast to master device
net, bonding: Add XDP support to the bonding driver
...
====================
Link: https://lore.kernel.org/r/20210810130038.16927-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch adds skb's frag page recycling support based on
the frag page support in page pool.
The performance improves above 10~20% for single thread iperf
TCP flow with IOMMU disabled when iperf server and irq/NAPI
have a different CPU.
The performance improves about 135%(14Gbit to 33Gbit) for single
thread iperf TCP flow when IOMMU is in strict mode and iperf
server shares the same cpu with irq/NAPI.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently page pool only support page recycling when there
is only one user of the page, and the split page reusing
implemented in the most driver can not use the page pool as
bing-pong way of reusing requires the multi user support in
page pool.
Those reusing or recycling has below limitations:
1. page from page pool can only be used be one user in order
for the page recycling to happen.
2. Bing-pong way of reusing in most driver does not support
multi desc using different part of the same page in order
to save memory.
So add multi-users support and frag page recycling in page
pool to overcome the above limitation.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
For 32 bit systems with 64 bit dma, dma_addr[1] is used to
store the upper 32 bit dma addr, those system should be rare
those days.
For normal system, the dma_addr[1] in 'struct page' is not
used, so we can reuse dma_addr[1] for storing frag count,
which means how many frags this page might be splited to.
In order to simplify the page frag support in the page pool,
the PAGE_POOL_DMA_USE_PP_FRAG_COUNT macro is added to indicate
the 32 bit systems with 64 bit dma, and the page frag support
in page pool is disabled for such system.
The newly added page_pool_set_frag_count() is called to reserve
the maximum frag count before any page frag is passed to the
user. The page_pool_atomic_sub_frag_count_return() is called
when user is done with the page frag.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently, page->pp is cleared and set everytime the page
is recycled, which is unnecessary.
So only set the page->pp when the page is added to the page
pool and only clear it when the page is released from the
page pool.
This is also a preparation to support allocating frag page
in page pool.
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
XDP is implemented in the bonding driver by transparently delegating
the XDP program loading, removal and xmit operations to the bonding
slave devices. The overall goal of this work is that XDP programs
can be attached to a bond device *without* any further changes (or
awareness) necessary to the program itself, meaning the same XDP
program can be attached to a native device but also a bonding device.
Semantics of XDP_TX when attached to a bond are equivalent in such
setting to the case when a tc/BPF program would be attached to the
bond, meaning transmitting the packet out of the bond itself using one
of the bond's configured xmit methods to select a slave device (rather
than XDP_TX on the slave itself). Handling of XDP_TX to transmit
using the configured bonding mechanism is therefore implemented by
rewriting the BPF program return value in bpf_prog_run_xdp. To avoid
performance impact this check is guarded by a static key, which is
incremented when a XDP program is loaded onto a bond device. This
approach was chosen to avoid changes to drivers implementing XDP. If
the slave device does not match the receive device, then XDP_REDIRECT
is transparently used to perform the redirection in order to have
the network driver release the packet from its RX ring. The bonding
driver hashing functions have been refactored to allow reuse with
xdp_buff's to avoid code duplication.
The motivation for this change is to enable use of bonding (and
802.3ad) in hairpinning L4 load-balancers such as [1] implemented with
XDP and also to transparently support bond devices for projects that
use XDP given most modern NICs have dual port adapters. An alternative
to this approach would be to implement 802.3ad in user-space and
implement the bonding load-balancing in the XDP program itself, but
is rather a cumbersome endeavor in terms of slave device management
(e.g. by watching netlink) and requires separate programs for native
vs bond cases for the orchestrator. A native in-kernel implementation
overcomes these issues and provides more flexibility.
Below are benchmark results done on two machines with 100Gbit
Intel E810 (ice) NIC and with 32-core 3970X on sending machine, and
16-core 3950X on receiving machine. 64 byte packets were sent with
pktgen-dpdk at full rate. Two issues [2, 3] were identified with the
ice driver, so the tests were performed with iommu=off and patch [2]
applied. Additionally the bonding round robin algorithm was modified
to use per-cpu tx counters as high CPU load (50% vs 10%) and high rate
of cache misses were caused by the shared rr_tx_counter (see patch
2/3). The statistics were collected using "sar -n dev -u 1 10". On top
of that, for ice, further work is in progress on improving the XDP_TX
numbers [4].
-----------------------| CPU |--| rxpck/s |--| txpck/s |----
without patch (1 dev):
XDP_DROP: 3.15% 48.6Mpps
XDP_TX: 3.12% 18.3Mpps 18.3Mpps
XDP_DROP (RSS): 9.47% 116.5Mpps
XDP_TX (RSS): 9.67% 25.3Mpps 24.2Mpps
-----------------------
with patch, bond (1 dev):
XDP_DROP: 3.14% 46.7Mpps
XDP_TX: 3.15% 13.9Mpps 13.9Mpps
XDP_DROP (RSS): 10.33% 117.2Mpps
XDP_TX (RSS): 10.64% 25.1Mpps 24.0Mpps
-----------------------
with patch, bond (2 devs):
XDP_DROP: 6.27% 92.7Mpps
XDP_TX: 6.26% 17.6Mpps 17.5Mpps
XDP_DROP (RSS): 11.38% 117.2Mpps
XDP_TX (RSS): 14.30% 28.7Mpps 27.4Mpps
--------------------------------------------------------------
RSS: Receive Side Scaling, e.g. the packets were sent to a range of
destination IPs.
[1]: https://cilium.io/blog/2021/05/20/cilium-110#standalonelb
[2]: https://lore.kernel.org/bpf/20210601113236.42651-1-maciej.fijalkowski@intel.com/T/#t
[3]: https://lore.kernel.org/bpf/CAHn8xckNXci+X_Eb2WMv4uVYjO2331UWB2JLtXr_58z0Av8+8A@mail.gmail.com/
[4]: https://lore.kernel.org/bpf/20210805230046.28715-1-maciej.fijalkowski@intel.com/T/#t
Signed-off-by: Jussi Maki <joamaki@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Jay Vosburgh <j.vosburgh@gmail.com>
Cc: Veaceslav Falico <vfalico@gmail.com>
Cc: Andy Gospodarek <andy@greyhouse.net>
Cc: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Cc: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/bpf/20210731055738.16820-4-joamaki@gmail.com
This adds the ndo_xdp_get_xmit_slave hook for transforming XDP_TX
into XDP_REDIRECT after BPF program run when the ingress device
is a bond slave.
The dev_xdp_prog_count is exposed so that slave devices can be checked
for loaded XDP programs in order to avoid the situation where both
bond master and slave have programs loaded according to xdp_state.
Signed-off-by: Jussi Maki <joamaki@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Jay Vosburgh <j.vosburgh@gmail.com>
Cc: Veaceslav Falico <vfalico@gmail.com>
Cc: Andy Gospodarek <andy@greyhouse.net>
Link: https://lore.kernel.org/bpf/20210731055738.16820-3-joamaki@gmail.com
Fix a typo when checking existence of port_type_set function pointer.
Fixes: 82564f6c70 ("devlink: Simplify devlink port API calls")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit a3fe3d01bd ("net/smc: introduce sg-logic for RMBs") introduced
a restriction for RMB allocations as used by SMC-R. However, SMC-D does
not use scatter-gather lists to back its DMBs, yet it was limited by
this restriction, still.
This patch exempts SMC, but limits allocations to the maximum RMB/DMB
size respectively.
Signed-off-by: Stefan Raspl <raspl@linux.ibm.com>
Signed-off-by: Guvenc Gulce <guvenc@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All kernel devlink implementations call to devlink_alloc() during
initialization routine for specific device which is used later as
a parent device for devlink_register().
Such late device assignment causes to the situation which requires us to
call to device_register() before setting other parameters, but that call
opens devlink to the world and makes accessible for the netlink users.
Any attempt to move devlink_register() to be the last call generates the
following error due to access to the devlink->dev pointer.
[ 8.758862] devlink_nl_param_fill+0x2e8/0xe50
[ 8.760305] devlink_param_notify+0x6d/0x180
[ 8.760435] __devlink_params_register+0x2f1/0x670
[ 8.760558] devlink_params_register+0x1e/0x20
The simple change of API to set devlink device in the devlink_alloc()
instead of devlink_register() fixes all this above and ensures that
prior to call to devlink_register() everything already set.
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The driver allocates the spinlock but not initialize it.
Use spin_lock_init() on it to initialize it correctly.
Fixes: aa730a9905 ("net: wwan: Add MHI MBIM network driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Reviewed-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Karsten Graul says:
====================
net/iucv: updates 2021-08-09
Please apply the following iucv patches to netdev's net-next tree.
Remove the usage of register asm statements and replace deprecated
CPU-hotplug functions with the current version.
Use use consume_skb() instead of kfree_skb() to avoid flooding
dropwatch with false-positives, and 2 patches with cleanups.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Using register asm statements has been proven to be very error prone,
especially when using code instrumentation where gcc may add function
calls, which clobbers register contents in an unexpected way.
Therefore get rid of register asm statements in iucv code, even though
there is currently nothing wrong with it. This way we know for sure
that the above mentioned bug class won't be introduced here.
Acked-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use IS_ENABLED(CONFIG_IUCV) to determine whether the iucv_if symbol
is available, and let depmod deal with the module dependency.
This was introduced back with commit 6fcd61f7bf ("af_iucv: use
loadable iucv interface"). And to avoid sprinkling IS_ENABLED() over
all the code, we're keeping the indirection through pr_iucv->...().
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change the good paths to use consume_skb() instead of kfree_skb(). This
avoids flooding dropwatch with false-positives.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean says:
====================
DSA fast ageing fixes/improvements
These are 2 small improvements brought to the DSA fast ageing changes
merged earlier today.
Patch 1 restores the behavior for DSA drivers that don't implement the
.port_bridge_flags function (I don't think there is any breakage due
to the new behavior, but just to be sure). This came as a result of
Andrew's review.
Patch 2 reduces the number of fast ages of a port from 2 to 1 when it
leaves a bridge.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Drivers that support both the toggling of address learning and dynamic
FDB flushing (mv88e6xxx, b53, sja1105) currently need to fast-age a port
twice when it leaves a bridge:
- once, when del_nbp() calls br_stp_disable_port() which puts the port
in the BLOCKING state
- twice, when dsa_port_switchdev_unsync_attrs() calls
dsa_port_clear_brport_flags() which disables address learning
The knee-jerk reaction might be to say "dsa_port_clear_brport_flags does
not need to fast-age the port at all", but the thing is, we still need
both code paths to flush the dynamic FDB entries in different situations.
When a DSA switch port leaves a bonding/team interface that is (still) a
bridge port, no del_nbp() will be called, so we rely on
dsa_port_clear_brport_flags() function to restore proper standalone port
functionality with address learning disabled.
So the solution is just to avoid double the work when both code paths
are called in series. Luckily, DSA already caches the STP port state, so
we can skip flushing the dynamic FDB when we disable address learning
and the STP state is one where no address learning takes place at all.
Under that condition, not flushing the FDB is safe because there is
supposed to not be any dynamic FDB entry at all (they were flushed
during the transition towards that state, and none were learned in the
meanwhile).
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 39f3210154 ("net: dsa: don't fast age standalone ports")
assumed that all standalone ports disable address learning, but if the
switch driver implements .port_fast_age but not .port_bridge_flags (like
ksz9477, ksz8795, lantiq_gswip, lan9303), then that might not actually
be true.
So whereas before, the bridge temporarily walking us through the
BLOCKING STP state meant that the standalone ports had a checkpoint to
flush their baggage and start fresh when they join a bridge, after that
commit they no longer do.
Restore the old behavior for these drivers by checking if the switch can
toggle address learning. If it can't, disregard the "do_fast_age"
argument and unconditionally perform fast ageing on STP state changes.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean says:
====================
Fast ageing support for SJA1105 DSA driver
While adding support for flushing dynamically learned FDB entries in the
sja1105 driver, I noticed a few things that could be improved in DSA.
Most notably, drivers could omit a fast age when address learning is
turned off, which might mean that ports leaving a bridge and becoming
standalone could still have FDB entries pointing towards them. Secondly,
when DSA fast ages a port after the 'learning' flag has been turned off,
the software bridge still has the dynamically learned 'master' FDB
entries installed, and those should be deleted too.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Delete the dynamically learned FDB entries when the STP state changes
and when address learning is disabled.
On sja1105 there is no shorthand SPI command for this, so we need to
walk through the entire FDB to delete.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now that DSA keeps track of the port learning state, it becomes
superfluous to keep an additional variable with this information in the
sja1105 driver. Remove it.
The DSA core's learning state is present in struct dsa_port *dp.
To avoid the antipattern where we iterate through a DSA switch's
ports and then call dsa_to_port to obtain the "dp" reference (which is
bad because dsa_to_port iterates through the DSA switch tree once
again), just iterate through the dst->ports and operate on those
directly.
The sja1105 had an extra use of priv->learn_ena on non-user ports. DSA
does not touch the learning state of those ports - drivers are free to
do what they wish on them. Mark that information with a comment in
struct dsa_port and let sja1105 set dp->learning for cascade ports.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, when DSA performs fast ageing on a port, 'bridge fdb' shows
us that the 'self' entries (corresponding to the hardware bridge, as
printed by dsa_slave_fdb_dump) are deleted, but the 'master' entries
(corresponding to the software bridge) aren't.
Indeed, searching through the bridge driver, neither the
brport_attr_learning handler nor the IFLA_BRPORT_LEARNING handler call
br_fdb_delete_by_port. However, br_stp_disable_port does, which is one
of the paths which DSA uses to trigger a fast ageing process anyway.
There is, however, one other very promising caller of
br_fdb_delete_by_port, and that is the bridge driver's handler of the
SWITCHDEV_FDB_FLUSH_TO_BRIDGE atomic notifier. Currently the s390/qeth
HiperSockets card driver is the only user of this.
I can't say I understand that driver's architecture or interaction with
the bridge, but it appears to not be a switchdev driver in the traditional
sense of the word. Nonetheless, the mechanism it provides is a useful
way for DSA to express the fact that it performs fast ageing too, in a
way that does not change the existing behavior for other drivers.
Cc: Alexandra Winter <wintera@linux.ibm.com>
Cc: Julian Wiedmann <jwi@linux.ibm.com>
Cc: Roopa Prabhu <roopa@nvidia.com>
Cc: Nikolay Aleksandrov <nikolay@nvidia.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On topology changes, stations that were dynamically learned on ports
that are no longer part of the active topology must be flushed - this is
described by clause "17.11 Updating learned station location information"
of IEEE 802.1D-2004.
However, when address learning on the bridge port is turned off in the
first place, there is nothing to flush, so skip a potentially expensive
operation.
We can finally do this now since DSA is aware of the learning state of
its bridged ports.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
