Refactor fw reset code to have the unload driver part done on
mlx5_fw_reset_complete_reload(), so if it was called by the PF which
initiated the reload fw activate flow, the unload part will be handled
by the mlx5_devlink_reload_fw_activate() callback itself and not by the
reset event work.
This will be used by the downstream patch to invoke devlink reload
callbacks with devlink lock held.
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
After mlx4 driver is converted to do locked reload,
devlink_region_snapshot_create() may be called from both locked and
unlocked context.
Note that in mlx4 region snapshots could be created on any command
failure. That can happen in any flow that involves commands to FW,
which means most of the driver flows.
So resolve this by removing dependency on devlink->lock for region
snapshots list consistency and introduce new mutex to ensure it.
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
After mlx4 driver is converted to do locked reload, functions to get/put
regions snapshot ID may be called from both locked and unlocked context.
So resolve this by removing dependency on devlink->lock for region
snapshot ID tracking by using internal xa_lock() to maintain
shapshot_ids xa_array consistency.
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Vikas Gupta says:
====================
add framework for selftests in devlink
Add support for selftests in the devlink framework.
Adds a callback .selftests_check and .selftests_run in devlink_ops.
User can add test(s) suite which is subsequently passed to the driver
and driver can opt for running particular tests based on its capabilities.
Patchset adds a flash based test for the bnxt_en driver.
====================
Link: https://lore.kernel.org/r/20220727165721.37959-1-vikas.gupta@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add callbacks
=============
.selftest_check: returns true for flash selftest.
.selftest_run: runs a flash selftest.
Also, refactor NVM APIs so that they can be
used with devlink and ethtool both.
Signed-off-by: Vikas Gupta <vikas.gupta@broadcom.com>
Reviewed-by: Andy Gospodarek <gospo@broadcom.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add a framework for running selftests.
Framework exposes devlink commands and test suite(s) to the user
to execute and query the supported tests by the driver.
Below are new entries in devlink_nl_ops
devlink_nl_cmd_selftests_show_doit/dumpit: To query the supported
selftests by the drivers.
devlink_nl_cmd_selftests_run: To execute selftests. Users can
provide a test mask for executing group tests or standalone tests.
Documentation/networking/devlink/ path is already part of MAINTAINERS &
the new files come under this path. Hence no update needed to the
MAINTAINERS
Signed-off-by: Vikas Gupta <vikas.gupta@broadcom.com>
Reviewed-by: Andy Gospodarek <gospo@broadcom.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Tariq Toukan says:
====================
mlx5e use TLS TX pool to improve connection rate
To offload encryption operations, the mlx5 device maintains state and
keeps track of every kTLS device-offloaded connection. Two HW objects
are used per TX context of a kTLS offloaded connection: a. Transport
interface send (TIS) object, to reach the HW context. b. Data Encryption
Key (DEK) to perform the crypto operations.
These two objects are created and destroyed per TLS TX context, via FW
commands. In total, 4 FW commands are issued per TLS TX context, which
seriously limits the connection rate.
In this series, we aim to save creation and destroy of TIS objects by
recycling them. Upon recycling of a TIS, the HW still needs to be
notified for the re-mapping between a TIS and a context. This is done by
posting WQEs via an SQ, significantly faster API than the FW command
interface.
A pool is used for recycling. The pool dynamically interacts to the load
and connection rate, growing and shrinking accordingly.
Saving the TIS FW commands per context increases connection rate by ~42%,
from 11.6K to 16.5K connections per sec.
Connection rate is still limited by FW bottleneck due to the remaining
per context FW commands (DEK create/destroy). This will soon be addressed
in a followup series. By combining the two series, the FW bottleneck
will be released, and a significantly higher (about 100K connections per
sec) kTLS TX device-offloaded connection rate is reached.
====================
Link: https://lore.kernel.org/r/20220727094346.10540-1-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Let the TLS TX recycle pool be more flexible in size, by continuously
and dynamically allocating and releasing HW resources in response to
changes in the connections rate and load.
Allocate and release pool entries in bulks (16). Use a workqueue to
release/allocate in the background. Allocate a new bulk when the pool
size goes lower than the low threshold (1K). Symmetric operation is done
when the pool size gets greater than the upper threshold (4K).
Every idle pool entry holds: 1 TIS, 1 DEK (HW resources), in addition to
~100 bytes in host memory.
Start with an empty pool to minimize memory and HW resources waste for
non-TLS users that have the device-offload TLS enabled.
Upon a new request, in case the pool is empty, do not wait for a whole bulk
allocation to complete. Instead, trigger an instant allocation of a single
resource to reduce latency.
Performance tests:
Before: 11,684 CPS
After: 16,556 CPS
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The transport interface send (TIS) object is responsible for performing
all transport related operations of the transmit side. The ConnectX HW
uses a TIS object to save and access the TLS crypto information and state
of an offloaded TX kTLS connection.
Before this patch, we used to create a new TIS per connection and destroy
it once it’s closed. Every create and destroy of a TIS is a FW command.
Same applies for the private TLS context, where we used to dynamically
allocate and free it per connection.
Resources recycling reduce the impact of the allocation/free operations
and helps speeding up the connection rate.
In this feature we maintain a pool of TX objects and use it to recycle
the resources instead of re-creating them per connection.
A cached TIS popped from the pool is updated to serve the new connection
via the fast-path HW interface, updating the tls static and progress
params. This is a very fast operation, significantly faster than FW
commands.
On recycling, a WQE fence is required after the context params change.
This guarantees that the data is sent after the context has been
successfully updated in hardware, and that the context modification
doesn't interfere with existing traffic.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Let the caller of mlx5e_ktls_tx_handle_ooo() take care of updating the
stats, according to the returned value. As the switch/case blocks are
already there, this change saves unnecessary branches in the handler.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
TLS TIS objects have a defined role in mapping and reaching the HW TLS
contexts. Some standard TIS attributes (like LAG port affinity) are
not relevant for them.
Use a dedicated TLS TIS create function instead of the generic
mlx5e_create_tis.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Multiple TLS device-offloaded contexts can be added in parallel via
concurrent calls to .tls_dev_add, while calls to .tls_dev_del are
sequential in tls_device_gc_task.
This is not a sustainable behavior. This creates a rate gap between add
and del operations (addition rate outperforms the deletion rate). When
running for enough time, the TLS device resources could get exhausted,
failing to offload new connections.
Replace the single-threaded garbage collector work with a per-context
alternative, so they can be handled on several cores in parallel. Use
a new dedicated destruct workqueue for this.
Tested with mlx5 device:
Before: 22141 add/sec, 103 del/sec
After: 11684 add/sec, 11684 del/sec
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
TLS context destructor can be run in atomic context. Cleanup operations
for device-offloaded contexts could require access and interaction with
the device callbacks, which might sleep. Hence, the cleanup of such
contexts must be deferred and completed inside an async work.
For all others, this is not necessary, as cleanup is atomic. Invoke
cleanup immediately for them, avoiding queueing redundant gc work.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The return from the call to tls_rx_msg_size() is int, it can be
a negative error code, however this is being assigned to an
unsigned long variable 'sz', so making 'sz' an int.
Eliminate the following coccicheck warning:
./net/tls/tls_strp.c:211:6-8: WARNING: Unsigned expression compared with zero: sz < 0
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Link: https://lore.kernel.org/r/20220728031019.32838-1-yang.lee@linux.alibaba.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski says:
====================
tls: rx: follow ups to rx work
A selection of unrelated changes. First some selftest polishing.
Next a change to rcvtimeo handling for locking based on an exchange
with Eric. Follow up to Paolo's comments from yesterday. Last but
not least a fix to a false positive warning, turns out I've been
testing with DEBUG_NET=n this whole time.
====================
Link: https://lore.kernel.org/r/20220727031524.358216-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
I went too far in the accessor conversion, we can't use tls_strp_msg()
after decryption because the message may not be ready. What we care
about on this path is that the output skb is detached, i.e. we didn't
somehow just turn around and used the input skb with its TCP data
still attached. So look at the anchor directly.
Fixes: 84c61fe1a7 ("tls: rx: do not use the standard strparser")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Paolo points out that there seems to be no strong reason strparser
users a single threaded workqueue. Perhaps there were some performance
or pinning considerations? Since we don't know (and it's the slow path)
let's default to the most natural, multi-threaded choice.
Also rename the workqueue to "tls-".
Suggested-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Eric indicates that restarting rcvtimeo on every wait may be fine.
I thought that we should consider it cumulative, and made
tls_rx_reader_lock() return the remaining timeo after acquiring
the reader lock.
tls_rx_rec_wait() gets its timeout passed in by value so it
does not keep track of time previously spent.
Make the lock waiting consistent with tls_rx_rec_wait() - don't
keep track of time spent.
Read the timeo fresh in tls_rx_rec_wait().
It's unclear to me why callers are supposed to cache the value.
Link: https://lore.kernel.org/all/CANn89iKcmSfWgvZjzNGbsrndmCch2HC_EPZ7qmGboDNaWoviNQ@mail.gmail.com/
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add a handful of memory randomizations and precise length checks.
Nothing is really broken here, I did this to increase confidence
when debugging. It does fix a GCC warning, tho. Apparently GCC
recognizes that memory needs to be initialized for send() but
does not recognize that for write().
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Commit 708ac5bea0 ("libbpf: add ksyscall/kretsyscall sections support
for syscall kprobes") added the arch_specific_syscall_pfx() function,
which returns a string representing the architecture in use. As it turns
out this function is currently not aware of Power PC, where NULL is
returned. That's being flagged by the libbpf CI system, which builds for
ppc64le and the compiler sees a NULL pointer being passed in to a %s
format string.
With this change we add representations for two more architectures, for
Power PC and Power PC 64, and also adjust the string format logic to
handle NULL pointers gracefully, in an attempt to prevent similar issues
with other architectures in the future.
Fixes: 708ac5bea0 ("libbpf: add ksyscall/kretsyscall sections support for syscall kprobes")
Signed-off-by: Daniel Müller <deso@posteo.net>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20220728222345.3125975-1-deso@posteo.net
Move the function declaration of mlx5e_init_l2_addr to en/fs.h, rename
to mlx5e_fs_init_l2_addr to align with the fs API functions naming
convention and let it take mlx5e_flow_steering as arguments while keeping
implementation at en_fs.c file. This helps maintain a clean driver code
and avoids unnecessary dependencies.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add inner callee for ndo mlx5e_vlan_rx_add_vid and
mlx5e_vlan_rx_kill_vid, to separate the priv usage from other
flow steering flows.
Move wrapper ndo's into en_main, and split the rest of the functionality
into a separate part inside en_fs.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Separate mlx5e_set_rx_mode into two, and move caller to en_main while
keeping implementation in en_fs in the newly declared function
mlx5e_fs_set_rx_mode. This; to minimize the coupling of flow_steering
to priv.
Add a parallel boolean member vlan_strip_disable to
mlx5e_flow_steering that's updated similarly as its identical in priv,
thus making it possible to adjust the rx_mode work handler to current
changes.
Also, add state_destroy boolean to mlx5e_flow_steering struct which
replaces the old check : !test_bit(MLX5E_STATE_DESTROYING, &priv->state).
This state member is updated accordingly prior to
INIT_WORK(mlx5e_set_rx_mode_work), This is done for similar purposes as
mentioned earlier and to minimize argument passings.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Make flow_steering struct contain mlx5_core_dev such that
it becomes self contained and easier to decouple later on this series.
Let its values be initialized in mlx5e_fs_init().
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Let en_fs report errors via mdev error report API, aka mlx5_core_*
macros, thus replace the netdev API reports.
This to minimize netdev coupling to the flow steering struct.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Make mlx5e_flow_steering member of mlx5e_priv a pointer.
Add dynamic allocation respectively.
Allocate fs for all profiles when initializing profile,
symmetrically deallocate at profile cleanup.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Introduce allocation and de-allocation functions for both flow steering
VLAN and TC as part of fs API.
Add allocations of VLAN and TC as nic profile feature, such that
fs_init() will allocate both VLAN and TC only if they're featured in
the profile. VLAN and TC are relevant for nic_profile only.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Move mlx5e_tc_table struct to en_tc.c thus make it private.
Introduce allocation and deallocation functions as part of the tc API
to allow this switch smoothly.
Convert mlx5e_nic_chain() macro to a function of en_tc.c.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Make fs.tc be a pointer and allocate it dynamically.
Add mlx5e_priv pointer to mlx5e_tc_table, and thus get a work-around to
accessing priv via tc when handling tc events inside mlx5e_tc_netdev_event.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add support for tc action api for police.
Offloading standalone police action without
a tc rule and reporting stats.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
mlx5e_tc_meter_get() to get an existing meter.
mlx5e_tc_meter_update() to update an existing meter without refcount.
mlx5e_tc_meter_replace() to get/create a meter and update if needed.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add red and green counters per meter instance.
TC police action is implemented as a meter instance.
The meter counters represent the police action
notexceed/exceed counters.
TC rules using the same meter instance will use
the same counters.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
To support a TC police action notexceed counter and supporting
actions other than drop/pipe there is a need to create separate ft
and rules per rule and not to use a common one created on eswitch init.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add support for ASO action of type flow metering
on device that supports STEv1.
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Reviewed-by: Hamdan Igbaria <hamdani@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Use skb_inner_tcp_all_headers() instead of skb_tcp_all_headers() when
transmitting an encapsulated packet in mlx5e_tx_get_gso_ihs().
Fixes: 504148fedb ("net: add skb_[inner_]tcp_all_headers helpers")
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
for the release.
Current release - regressions:
- wifi: mac80211: do not abuse fq.lock in ieee80211_do_stop(),
fix taking the lock before its initialized
- Bluetooth: mgmt: fix double free on error path
Current release - new code bugs:
- eth: ice: fix tunnel checksum offload with fragmented traffic
Previous releases - regressions:
- tcp: md5: fix IPv4-mapped support after refactoring, don't take
the pure v6 path
- Revert "tcp: change pingpong threshold to 3", improving detection
of interactive sessions
- mld: fix netdev refcount leak in mld_{query | report}_work() due
to a race
- Bluetooth:
- always set event mask on suspend, avoid early wake ups
- L2CAP: fix use-after-free caused by l2cap_chan_put
- bridge: do not send empty IFLA_AF_SPEC attribute
Previous releases - always broken:
- ping6: fix memleak in ipv6_renew_options()
- sctp: prevent null-deref caused by over-eager error paths
- virtio-net: fix the race between refill work and close,
resulting in NAPI scheduled after close and a BUG()
- macsec:
- fix three netlink parsing bugs
- avoid breaking the device state on invalid change requests
- fix a memleak in another error path
Misc:
- dt-bindings: net: ethernet-controller: rework 'fixed-link' schema
- two more batches of sysctl data race adornment
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
-----BEGIN PGP SIGNATURE-----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=xnan
-----END PGP SIGNATURE-----
Merge tag 'net-5.19-final' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Pull networking fixes from Jakub Kicinski:
"Including fixes from bluetooth and netfilter, no known blockers for
the release.
Current release - regressions:
- wifi: mac80211: do not abuse fq.lock in ieee80211_do_stop(), fix
taking the lock before its initialized
- Bluetooth: mgmt: fix double free on error path
Current release - new code bugs:
- eth: ice: fix tunnel checksum offload with fragmented traffic
Previous releases - regressions:
- tcp: md5: fix IPv4-mapped support after refactoring, don't take the
pure v6 path
- Revert "tcp: change pingpong threshold to 3", improving detection
of interactive sessions
- mld: fix netdev refcount leak in mld_{query | report}_work() due to
a race
- Bluetooth:
- always set event mask on suspend, avoid early wake ups
- L2CAP: fix use-after-free caused by l2cap_chan_put
- bridge: do not send empty IFLA_AF_SPEC attribute
Previous releases - always broken:
- ping6: fix memleak in ipv6_renew_options()
- sctp: prevent null-deref caused by over-eager error paths
- virtio-net: fix the race between refill work and close, resulting
in NAPI scheduled after close and a BUG()
- macsec:
- fix three netlink parsing bugs
- avoid breaking the device state on invalid change requests
- fix a memleak in another error path
Misc:
- dt-bindings: net: ethernet-controller: rework 'fixed-link' schema
- two more batches of sysctl data race adornment"
* tag 'net-5.19-final' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (67 commits)
stmmac: dwmac-mediatek: fix resource leak in probe
ipv6/addrconf: fix a null-ptr-deref bug for ip6_ptr
net: ping6: Fix memleak in ipv6_renew_options().
net/funeth: Fix fun_xdp_tx() and XDP packet reclaim
sctp: leave the err path free in sctp_stream_init to sctp_stream_free
sfc: disable softirqs for ptp TX
ptp: ocp: Select CRC16 in the Kconfig.
tcp: md5: fix IPv4-mapped support
virtio-net: fix the race between refill work and close
mptcp: Do not return EINPROGRESS when subflow creation succeeds
Bluetooth: L2CAP: Fix use-after-free caused by l2cap_chan_put
Bluetooth: Always set event mask on suspend
Bluetooth: mgmt: Fix double free on error path
wifi: mac80211: do not abuse fq.lock in ieee80211_do_stop()
ice: do not setup vlan for loopback VSI
ice: check (DD | EOF) bits on Rx descriptor rather than (EOP | RS)
ice: Fix VSIs unable to share unicast MAC
ice: Fix tunnel checksum offload with fragmented traffic
ice: Fix max VLANs available for VF
netfilter: nft_queue: only allow supported familes and hooks
...
Add support for NETIF_F_LOOPBACK. This feature can be set via:
$ ethtool -K eth0 loopback <on|off>
Feature can be useful for local data path tests.
Acked-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Tested-by: George Kuruvinakunnel <george.kuruvinakunnel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Instead of rather verbose comparison of current netdev->features bits vs
the incoming ones from user, let us compress them by a helper features
set that will be the result of netdev->features XOR features. This way,
current, extensive branches:
if (features & NETIF_F_BIT && !(netdev->features & NETIF_F_BIT))
set_feature(true);
else if (!(features & NETIF_F_BIT) && netdev->features & NETIF_F_BIT)
set_feature(false);
can become:
netdev_features_t changed = netdev->features ^ features;
if (changed & NETIF_F_BIT)
set_feature(!!(features & NETIF_F_BIT));
This is nothing new as currently several other drivers use this
approach, which I find much more convenient.
Acked-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Tested-by: George Kuruvinakunnel <george.kuruvinakunnel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
When trust is turned off for the VF, the expectation is that promiscuous
and allmulticast filters are removed. Currently default VSI filter is not
getting cleared in this flow.
Example:
ip link set enp236s0f0 vf 0 trust on
ip link set enp236s0f0v0 promisc on
ip link set enp236s0f0 vf 0 trust off
/* promiscuous mode is still enabled on VF0 */
Remove switch filters for both cases.
This commit fixes above behavior by removing default VSI filters and
allmulticast filters when vf-true-promisc-support is OFF.
Signed-off-by: Michal Wilczynski <michal.wilczynski@intel.com>
Tested-by: Marek Szlosek <marek.szlosek@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
In current implementation default VSI switch filter is only able to
forward traffic to a single VSI. This limits promiscuous mode with
private flag 'vf-true-promisc-support' to a single VF. Enabling it on
the second VF won't work. Also allmulticast support doesn't seem to be
properly implemented when vf-true-promisc-support is true.
Use standard ice_add_rule_internal() function that already implements
forwarding to multiple VSI's instead of constructing AQ call manually.
Add switch filter for allmulticast mode when vf-true-promisc-support is
enabled. The same filter is added regardless of the flag - it doesn't
matter for this case.
Remove unnecessary fields in switch structure. From now on book keeping
will be done by ice_add_rule_internal().
Refactor unnecessarily passed function arguments.
To test:
1) Create 2 VM's, and two VF's. Attach VF's to VM's.
2) Enable promiscuous mode on both of them and check if
traffic is seen on both of them.
Signed-off-by: Michal Wilczynski <michal.wilczynski@intel.com>
Tested-by: Marek Szlosek <marek.szlosek@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The 82576 PTP implementation still uses .adjfreq instead of using the newer
.adjfine.
This implementation uses a pre-simplified calculation since the base
increment value for the 82576 is just 16 * 2^19. Converting this into
scaled_ppm is tricky, and makes the intent a bit less clear.
Simply convert to the normal flow of multiplying the base increment value
by the scaled_ppm and then dividing by 1000000ULL << 16. This can be
implemented using mul_u64_u64_div_u64 which can avoid the possible overflow
that might occur for large adjustments.
Use of .adjfine can improve the precision of small adjustments and gets us
one driver closer to removing the old implementation from the kernel
entirely.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Convert the ixgbe PTP frequency adjustment implementations from .adjfreq to
.adjfine. This allows using the scaled parts per million adjustment from
the PTP core and results in a more precise adjustment for small
corrections.
To avoid overflow, use mul_u64_u64_div_u64 to perform the calculation. On
X86 platforms, this will use instructions that perform the operations with
128bit intermediate values. For other architectures, the implementation
will limit the loss of precision as much as possible.
This change slightly improves the precision of frequency adjustments for
all ixgbe based devices, and gets us one driver closer to being able to
remove the older .adjfreq implementation from the kernel.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The i40e driver currently implements the .adjfreq handler for frequency
adjustments. This takes the adjustment parameter in parts per billion. The
PTP core supports .adjfine which provides an adjustment in scaled parts per
million. This has a higher resolution and can result in more precise
adjustments for small corrections.
Convert the existing .adjfreq implementation to the newer .adjfine
implementation. This is trivial since it just requires changing the divisor
from 1000000000ULL to (1000000ULL << 16) in the mul_u64_u64_div_u64 call.
This improves the precision of the adjustments and gets us one driver
closer to removing the old .adjfreq support from the kernel.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The i40e device has a different clock rate depending on the current link
speed. This requires using a different increment rate for the PTP clock
registers. For slower link speeds, the base increment value is larger.
Directly multiplying the larger increment value by the parts per billion
adjustment might overflow.
To avoid this, the i40e implementation defaults to using the lower
increment value and then multiplying the adjustment afterwards. This causes
a loss of precision for lower link speeds.
We can fix this by using mul_u64_u64_div_u64 instead of performing the
multiplications using standard C operations. On X86, this will use special
instructions that perform the multiplication and division with 128bit
intermediate values. For other architectures, the fallback implementation
will limit the loss of precision for large values. Small adjustments don't
overflow anyways and won't lose precision at all.
This allows first multiplying the base increment value and then performing
the adjustment calculation, since we no longer fear overflowing. It also
makes it easier to convert to the even more precise .adjfine implementation
in a following change.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The PTP implementation for the e1000e driver uses the older .adjfreq
method. This method takes an adjustment in parts per billion. The newer
.adjfine implementation uses scaled_ppm. The use of scaled_ppm allows for
finer grained adjustments and is preferred over using the older
implementation.
Make use of mul_u64_u64_div_u64 in order to handle possible overflow of the
multiplication used to calculate the desired adjustment to the hardware
increment value.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Naama Meir <naamax.meir@linux.intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The e1000e_phc_adjfreq function validates that the input delta is within
the maximum range. This is already handled by the core PTP code and this is
a duplicate and thus unnecessary check. It also complicates refactoring to
use the newer .adjfine implementation, where the input is no longer
specified in parts per billion. Remove the range validation check.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Naama Meir <naamax.meir@linux.intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The PTP frequency adjustment code needs to determine an appropriate
adjustment given an input scaled_ppm adjustment.
We calculate the adjustment to the register by multiplying the base
(nominal) increment value by the scaled_ppm and then dividing by the
scaled one million value.
For very large adjustments, this might overflow. To avoid this, both the
scaled_ppm and divisor values are downshifted.
We can avoid that on X86 architectures by using mul_u64_u64_div_u64. This
helper function will perform the multiplication and division with 128bit
intermediate values. We know that scaled_ppm is never larger than the
divisor so this operation will never result in an overflow.
This improves the accuracy of the calculations for large adjustment values
on X86. It is likely an improvement on other architectures as well because
the default implementation of mul_u64_u64_div_u64 is smarter than the
original approach taken in the ice code.
Additionally, this implementation is easier to read, using fewer local
variables and lines of code to implement.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
If mediatek_dwmac_clks_config() fails, then call stmmac_remove_config_dt()
before returning. Otherwise it is a resource leak.
Fixes: fa4b3ca60e ("stmmac: dwmac-mediatek: fix clock issue")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/YuJ4aZyMUlG6yGGa@kili
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Moshe Shemesh