platform/kernel/linux-starfive.git
2 years agowifi: mac80211: mlme: shift some code around
Johannes Berg [Tue, 12 Jul 2022 09:13:56 +0000 (11:13 +0200)]
wifi: mac80211: mlme: shift some code around

We'll need ieee80211_prep_channel() in other code for MLO
later, so move the code up - unchanged for now - to avoid
forward declarations in the future.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: refactor link station setup
Johannes Berg [Tue, 12 Jul 2022 09:06:33 +0000 (11:06 +0200)]
wifi: mac80211: mlme: refactor link station setup

Refactor the code here since we need to have it also for each
link station after association in MLO later.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: move IEEE80211_SDATA_OPERATING_GMODE to link
Johannes Berg [Tue, 12 Jul 2022 08:49:23 +0000 (10:49 +0200)]
wifi: mac80211: move IEEE80211_SDATA_OPERATING_GMODE to link

The flag here is currently per interface, but the way we
set and clear it means it should be per link, so change
it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: make ieee80211_check_rate_mask() link-aware
Johannes Berg [Mon, 11 Jul 2022 13:13:20 +0000 (15:13 +0200)]
wifi: mac80211: make ieee80211_check_rate_mask() link-aware

Change ieee80211_check_rate_mask() to use a link rather than
the sdata and deflink/bss_conf.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: add multi-link element to AUTH frames
Johannes Berg [Tue, 12 Jul 2022 11:38:07 +0000 (13:38 +0200)]
wifi: mac80211: add multi-link element to AUTH frames

When sending an authentication frame from an MLD, include
the multi-link element with the MLD address and use the
link address for transmission.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: clean up supported channels element code
Johannes Berg [Tue, 12 Jul 2022 11:40:02 +0000 (13:40 +0200)]
wifi: mac80211: mlme: clean up supported channels element code

Clean up the code building the supported channels element
a little bit by using a local variable instead of the long
line.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: release channel context on link stop
Johannes Berg [Tue, 12 Jul 2022 11:36:37 +0000 (13:36 +0200)]
wifi: mac80211: release channel context on link stop

When a link is stopped for removal, release the channel
context it may have.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: prohibit DEAUTH_NEED_MGD_TX_PREP in MLO
Johannes Berg [Tue, 12 Jul 2022 11:35:54 +0000 (13:35 +0200)]
wifi: mac80211: prohibit DEAUTH_NEED_MGD_TX_PREP in MLO

For now, prohibit DEAUTH_NEED_MGD_TX_PREP since we can't
really transmit this on a specific link yet as we don't
know which links are active.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: fix some attribute policy entries
Johannes Berg [Tue, 12 Jul 2022 15:33:13 +0000 (17:33 +0200)]
wifi: nl80211: fix some attribute policy entries

The new NL80211_CMD_ADD_LINK_STA and NL80211_CMD_MODIFY_LINK_STA
commands have strict policy validation, so fix the policy so it
can be validated correctly.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: reject fragmented and non-inheritance elements
Johannes Berg [Mon, 11 Jul 2022 12:18:18 +0000 (14:18 +0200)]
wifi: nl80211: reject fragmented and non-inheritance elements

The underlying mac80211 code cannot deal with fragmented
elements for purposes of sorting the elements into the
association frame, so reject those inside the link. We
might want to reject them inside the assoc frame, but
they're used today for FILS, so cannot do that.

The non-inheritance element inside the links similarly
cannot be handled by mac80211, and outside the links it
makes no sense.

Reject both since using them could lead to an incorrect
implementation.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: reject link specific elements on assoc link
Johannes Berg [Mon, 11 Jul 2022 09:53:20 +0000 (11:53 +0200)]
wifi: nl80211: reject link specific elements on assoc link

When we associate, we'll include all the elements for the
link we're sending the association request on in the frame
and the specific ones for other links in the multi-link
element container. Prohibit adding link-specific elements
for the association link.

Fixes: d648c23024bd ("wifi: nl80211: support MLO in auth/assoc")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: set country_elem to NULL
Johannes Berg [Wed, 13 Jul 2022 08:24:23 +0000 (10:24 +0200)]
wifi: cfg80211: set country_elem to NULL

The link loop will always have a valid link so that
it's always set, but static checkers don't always
see that, so set it to NULL explicitly.

Fixes: efbabc116500 ("cfg80211: Indicate MLO connection info in connect and roam callbacks")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: remove link_id parameter from link_info_changed()
Gregory Greenman [Sun, 3 Jul 2022 21:38:22 +0000 (00:38 +0300)]
wifi: mac80211: remove link_id parameter from link_info_changed()

Since struct ieee80211_bss_conf already contains link_id,
passing link_id is not necessary.

Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: replace link_id with link_conf in switch/(un)assign_vif_chanctx()
Gregory Greenman [Sun, 3 Jul 2022 15:04:15 +0000 (18:04 +0300)]
wifi: mac80211: replace link_id with link_conf in switch/(un)assign_vif_chanctx()

Since mac80211 already has a protected pointer to link_conf,
pass it to the driver to avoid additional RCU locking.

Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: advertise MLO support
Johannes Berg [Mon, 11 Jul 2022 08:08:11 +0000 (10:08 +0200)]
wifi: nl80211: advertise MLO support

At least while we don't have any more specific interface
combinations support, add a simple flag for MLO support,
we can keep this later based on something other than the
wiphy flag.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: Support multi link in ieee80211_recalc_min_chandef()
Andrei Otcheretianski [Wed, 22 Jun 2022 13:15:56 +0000 (16:15 +0300)]
wifi: mac80211: Support multi link in ieee80211_recalc_min_chandef()

Recalculate min channel context for the given or all interface
links, depending on the caller. For a station state change, we
need to recalculate all of them since we don't know which link
(or multiple) it might be on.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: don't check carrier in chanctx code
Andrei Otcheretianski [Tue, 14 Jun 2022 07:16:26 +0000 (10:16 +0300)]
wifi: mac80211: don't check carrier in chanctx code

We check here that we don't enable TX (netif_carrier_ok())
before we actually start using some channel context, but to
our knowledge this check has never triggered, and with MLO
it's just wrong since links can be added and removed much
more dynamically than before.

Simply remove the checks, there's no really good way to do
anything that would replace them.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: allow link ID in set_wiphy with frequency
Ilan Peer [Sun, 12 Jun 2022 13:49:45 +0000 (16:49 +0300)]
wifi: nl80211: allow link ID in set_wiphy with frequency

This simplifies hostapd implementation, since it didn't
switch to NL80211_CMD_SET_CHANNEL.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: Allow EAPOL tx from specific link
Andrei Otcheretianski [Thu, 30 Jun 2022 13:43:44 +0000 (16:43 +0300)]
wifi: mac80211: Allow EAPOL tx from specific link

Allow link source address on TX.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: Allow EAPOL frames from link addresses
Andrei Otcheretianski [Thu, 30 Jun 2022 12:37:37 +0000 (15:37 +0300)]
wifi: mac80211: Allow EAPOL frames from link addresses

Allow transmitting EAPOL frames not only from the interface
address (which is the MLD address) but also any link addresses,
in order to support non-MLO stations on AP interfaces.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211/mac80211: Support control port TX from specific link
Andrei Otcheretianski [Thu, 30 Jun 2022 12:27:59 +0000 (15:27 +0300)]
wifi: cfg80211/mac80211: Support control port TX from specific link

In case of authentication with a legacy station, link addressed EAPOL
frames should be sent. Support it.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: Support MLD parameters in nl80211_set_station()
Andrei Otcheretianski [Sun, 26 Jun 2022 07:35:48 +0000 (10:35 +0300)]
wifi: nl80211: Support MLD parameters in nl80211_set_station()

Set the MLD parameters in NL80211_CMD_SET_STATION handling
to be able to change an MLD station.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: check MLO support in authenticate
Johannes Berg [Thu, 7 Jul 2022 16:19:27 +0000 (18:19 +0200)]
wifi: nl80211: check MLO support in authenticate

We should check that MLO connections are supported before
attempting to authenticate with MLO parameters, check that.

Fixes: d648c23024bd ("wifi: nl80211: support MLO in auth/assoc")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: add a helper to fragment an element
Johannes Berg [Thu, 7 Jul 2022 13:28:14 +0000 (15:28 +0200)]
wifi: mac80211: add a helper to fragment an element

The way this works is that you add all the element data,
keeping a pointer to the length field of the element.
Then call this helper function, which will fragment the
element if there was more than 255 bytes in the element,
memmove()ing the data back if needed.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: skip rate statistics for MLD STAs
Johannes Berg [Thu, 7 Jul 2022 13:11:54 +0000 (15:11 +0200)]
wifi: mac80211: skip rate statistics for MLD STAs

For now, skip rate statistics here to avoid warnings in
the called code, we'll need to adjust this to have all
the statistics for link stations.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: set BSS to NULL if IS_ERR()
Johannes Berg [Thu, 7 Jul 2022 13:03:51 +0000 (15:03 +0200)]
wifi: nl80211: set BSS to NULL if IS_ERR()

If the BSS lookup returned an error, set it to NULL so we
don't try to free it.

Fixes: d648c23024bd ("wifi: nl80211: support MLO in auth/assoc")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: add EML/MLD capabilities to per-iftype capabilities
Johannes Berg [Mon, 4 Jul 2022 13:02:33 +0000 (15:02 +0200)]
wifi: nl80211: add EML/MLD capabilities to per-iftype capabilities

We have the per-interface type capabilities, currently for
extended capabilities, add the EML/MLD capabilities there
to have this advertised by the driver.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: better validate link ID for stations
Johannes Berg [Wed, 6 Jul 2022 13:31:55 +0000 (15:31 +0200)]
wifi: nl80211: better validate link ID for stations

If we add a station on an MLD, we need a link ID to see
where it lives (by default). Validate the link ID against
the valid_links.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: fix link manipulation
Johannes Berg [Wed, 6 Jul 2022 12:03:07 +0000 (14:03 +0200)]
wifi: mac80211: fix link manipulation

When we add non-deflink pointers, we need to remove the
link[0] pointer to deflink in case link[0] is not valid
afterwards. Also, we need to add that back when there
are no more valid links. Reorg the code to fix that.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: tighten locking check
Johannes Berg [Wed, 6 Jul 2022 08:09:57 +0000 (10:09 +0200)]
wifi: mac80211: tighten locking check

When we remove a link that doesn't have a channel context,
we don't really need the local->mtx locking. Tighten the
check here.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: clean up links appropriately
Johannes Berg [Wed, 6 Jul 2022 07:57:42 +0000 (09:57 +0200)]
wifi: cfg80211: clean up links appropriately

This was missing earlier, we need to remove links when
interfaces are being destroyed, and we also need to
stop (AP) operations when a link is being destroyed.
Address these issues to remove many warnings that will
otherwise appear in mac80211.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: consider EHT element size in assoc request
Johannes Berg [Fri, 1 Jul 2022 14:22:29 +0000 (16:22 +0200)]
wifi: mac80211: consider EHT element size in assoc request

We need to consider the (maximum) size of the EHT element
we'll add for the association request, otherwise we may run
out of space.

Fixes: 820acc810fb6 ("mac80211: Add EHT capabilities to association/probe request")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: simplify adding ht/vht/he/eht elements
Johannes Berg [Fri, 1 Jul 2022 14:14:33 +0000 (16:14 +0200)]
wifi: mac80211: mlme: simplify adding ht/vht/he/eht elements

The functions currently take a link and check data
from it, but this needs to change for MLO. Simplify
the prototypes by passing only the needed arguments.

Remove the regulatory checks, the warnings shouldn't
trigger, and haven't as far as I know.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: refactor adding custom elements
Johannes Berg [Fri, 1 Jul 2022 13:38:57 +0000 (15:38 +0200)]
wifi: mac80211: refactor adding custom elements

Rework the sorting of custom elements into the association
request by moving the elements before HT/VHT/HE to each
their own function. While at it, fix the placement of the
ones that should be between VHT and HE.

This doesn't fix the placement of elements that should be
between HE and EHT yet, a similar change might be needed
in the future.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: refactor adding rates to assoc request
Johannes Berg [Fri, 1 Jul 2022 12:08:25 +0000 (14:08 +0200)]
wifi: mac80211: refactor adding rates to assoc request

There's some awkward code that really only exists
because we want to optimize the allocation size,
but that's not really all that necessary.

Refactor the code that adds rates to the association
request frame to have a separate function, removing
the goto.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: use only channel width in ieee80211_parse_bitrates()
Johannes Berg [Fri, 1 Jul 2022 12:01:59 +0000 (14:01 +0200)]
wifi: mac80211: use only channel width in ieee80211_parse_bitrates()

For MLO, we may not have a full chandef here later, so change
the API to pass only the width.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: add ieee80211_chanwidth_rate_flags()
Johannes Berg [Fri, 1 Jul 2022 12:01:29 +0000 (14:01 +0200)]
wifi: cfg80211: add ieee80211_chanwidth_rate_flags()

To simplify things when we don't have a full chandef,
add ieee80211_chanwidth_rate_flags().

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: remove redundant condition
Johannes Berg [Fri, 1 Jul 2022 11:40:19 +0000 (13:40 +0200)]
wifi: mac80211: remove redundant condition

Here, ext_capa is checked and can only be non-NULL if
assoc_data->ie_len was set before, so the check here
is redundant.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: don't set link address for station
Johannes Berg [Thu, 30 Jun 2022 12:20:54 +0000 (14:20 +0200)]
wifi: mac80211: don't set link address for station

We need to handle the link addresses for station differently,
they will be determined by the association code, stored, and
then applied when the links are actually created on success,
cfg80211 will fill in the right addresses per the data we're
sending back to it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: fix multi-BSSID element parsing
Johannes Berg [Wed, 29 Jun 2022 11:29:05 +0000 (13:29 +0200)]
wifi: mac80211: fix multi-BSSID element parsing

When parsing a frame containing a multi-BSSID element, we
need to know both the transmitted and non-transmitted BSSID
so we can parse it correctly.

Unfortunately, in quite a number of cases, we got this wrong
and were passing the wrong BSSID or useless information:
 * the mgmt->bssid from a frame is only the transmitted
   BSSID if the frame is a beacon
 * passing just one of the parameters as non-NULL isn't
   useful and ignored

In those case where we need to parse for a specific BSS we
always have a BSS structure pointer, representing the BSS
we need, whether transmitted or not. Thus, pass that pointer
to the parsing function instead of the two BSSIDs.

Also fix two bugs:
 * we need to re-parse all the elements for the other BSS
   when iterating the non-transmitted BSSes in scan
 * we need to parse for the correct BSS when setting up
   the channel data in client code

Fixes: 78ac51f81532 ("mac80211: support multi-bssid")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: move tdls_chan_switch_prohibited to link data
Johannes Berg [Wed, 29 Jun 2022 10:01:41 +0000 (12:01 +0200)]
wifi: mac80211: move tdls_chan_switch_prohibited to link data

This value should be per link, since a TDLS connection is
only established on a given link.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: don't re-parse elems in ieee80211_assoc_success()
Johannes Berg [Wed, 29 Jun 2022 09:52:49 +0000 (11:52 +0200)]
wifi: mac80211: don't re-parse elems in ieee80211_assoc_success()

We're already passing the elems pointer, and have parsed
them from the same frame with exactly the same parameters,
so don't need to do that again.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: replace link_id with link_conf in start/stop_ap()
Gregory Greenman [Wed, 29 Jun 2022 09:22:24 +0000 (12:22 +0300)]
wifi: mac80211: replace link_id with link_conf in start/stop_ap()

When calling start/stop_ap(), mac80211 already has a protected
link_conf pointer. Pass it to the driver, so it shouldn't
handle RCU protection.

Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: refactor elements parsing with parameter struct
Johannes Berg [Tue, 28 Jun 2022 15:49:12 +0000 (17:49 +0200)]
wifi: mac80211: refactor elements parsing with parameter struct

Refactor the element parsing into a version that has
a parameter struct so we can add more parameters more
easily in the future.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: extend cfg80211_rx_assoc_resp() for MLO
Johannes Berg [Tue, 28 Jun 2022 15:17:05 +0000 (17:17 +0200)]
wifi: cfg80211: extend cfg80211_rx_assoc_resp() for MLO

Extend the cfg80211_rx_assoc_resp() to cover multiple
BSSes, the AP MLD address and local link addresses
for MLO.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: put cfg80211_rx_assoc_resp() arguments into a struct
Johannes Berg [Tue, 28 Jun 2022 14:25:37 +0000 (16:25 +0200)]
wifi: cfg80211: put cfg80211_rx_assoc_resp() arguments into a struct

For MLO we'll need a lot more arguments, including all the
BSS pointers and link addresses, so move the data to a struct
to be able to extend it more easily later.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: adjust assoc comeback for MLO
Johannes Berg [Tue, 28 Jun 2022 09:25:38 +0000 (11:25 +0200)]
wifi: cfg80211: adjust assoc comeback for MLO

We only report the BSSID to userspace, so change the
argument from BSS struct pointer to AP address, which
we'll use to carry either the BSSID or AP MLD address.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: unify assoc data event sending
Johannes Berg [Mon, 27 Jun 2022 20:23:42 +0000 (22:23 +0200)]
wifi: mac80211: mlme: unify assoc data event sending

There are a few cases where we send an event to cfg80211
manually, but ieee80211_destroy_assoc_data() also handles
the case of abandoning; some cases don't need an event
and success is handled yet differently.

Unify this by providing a single status argument to the
ieee80211_destroy_assoc_data() function and then handling
all the different cases of events (or no events) there.

This will help simplify the code when MLO support is
added.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: prepare association failure APIs for MLO
Johannes Berg [Mon, 27 Jun 2022 20:09:50 +0000 (22:09 +0200)]
wifi: cfg80211: prepare association failure APIs for MLO

For MLO, we need the ability to report back multiple BSS
structures to release, as well as the AP MLD address (if
attempting to make an MLO connection).

Unify cfg80211_assoc_timeout() and cfg80211_abandon_assoc()
into a new cfg80211_assoc_failure() that gets a structure
parameter with the necessary data.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: remove BSS pointer from cfg80211_disassoc_request
Johannes Berg [Mon, 27 Jun 2022 14:19:18 +0000 (16:19 +0200)]
wifi: cfg80211: remove BSS pointer from cfg80211_disassoc_request

The race described by the comment in mac80211 hasn't existed
since the locking rework to use the same lock and for MLO we
need to pass the AP MLD address, so just pass the BSSID or
AP MLD address instead of the BSS struct pointer, and adjust
all the code accordingly.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: use correct link_sta
Johannes Berg [Mon, 27 Jun 2022 12:17:28 +0000 (14:17 +0200)]
wifi: mac80211: mlme: use correct link_sta

For station capabilities, e.g. TWT, we need to use the correct
link station instead of deflink. Switch the code to do that.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: remove sta argument from ieee80211_config_bw
Johannes Berg [Mon, 27 Jun 2022 12:15:09 +0000 (14:15 +0200)]
wifi: mac80211: mlme: remove sta argument from ieee80211_config_bw

The argument is unused except for NULL checking, but we already
do that anyway, so it's not needed. Remove the argument.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: use ieee80211_get_link_sband()
Johannes Berg [Mon, 27 Jun 2022 12:10:18 +0000 (14:10 +0200)]
wifi: mac80211: mlme: use ieee80211_get_link_sband()

This requires a few more changes.

While at it, also add a warning to ieee80211_get_sband()
to avoid it being used when there are multiple links.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: split IEEE80211_STA_DISABLE_WMM to link data
Johannes Berg [Mon, 27 Jun 2022 11:13:04 +0000 (13:13 +0200)]
wifi: mac80211: split IEEE80211_STA_DISABLE_WMM to link data

If we decide to stop tracking QoS/WMM parameters, then
this should be a per-link decision. Move the flag to
the link instead.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: first adjustments for MLO
Johannes Berg [Fri, 24 Jun 2022 14:18:56 +0000 (16:18 +0200)]
wifi: mac80211: mlme: first adjustments for MLO

Do the first adjustments in the client-side code to pass
the link pointer (instead of sdata) to most places etc.

This is just preparation, so the real MLO patches become
smaller.

Note that this isn't complete, notably there are still
quite a few references to sta->deflink and sta->sta.deflink.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: do IEEE80211_STA_RESET_SIGNAL_AVE per link
Johannes Berg [Fri, 24 Jun 2022 14:28:33 +0000 (16:28 +0200)]
wifi: mac80211: mlme: do IEEE80211_STA_RESET_SIGNAL_AVE per link

Remove the IEEE80211_STA_RESET_SIGNAL_AVE flag and use
a bool instead, but invert the polarity (now calling it
tracking_signal_avg) so we don't have to initialize it,
and put that into the link instead.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: track AP (MLD) address separately
Johannes Berg [Fri, 24 Jun 2022 14:08:39 +0000 (16:08 +0200)]
wifi: mac80211: mlme: track AP (MLD) address separately

To prepare a bit more for MLO in the client code,
track the AP's address (for now only the BSSID, but
will track the AP MLD's address later) separately
from the per-link BSSID.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: remove unused bssid variable
Johannes Berg [Fri, 24 Jun 2022 13:54:38 +0000 (15:54 +0200)]
wifi: mac80211: remove unused bssid variable

This variable is only written to, remove it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: change QoS settings API to take link into account
Johannes Berg [Fri, 24 Jun 2022 13:40:11 +0000 (15:40 +0200)]
wifi: mac80211: change QoS settings API to take link into account

Take the link into account in the QoS settings (EDCA parameters)
APIs.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: expect powersave handling in driver for MLO
Johannes Berg [Fri, 24 Jun 2022 13:14:26 +0000 (15:14 +0200)]
wifi: mac80211: expect powersave handling in driver for MLO

In MLO, expect the driver fully handles powersave handling,
including tracking whether or not a beacon was received,
the DTIM period, etc.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: move ps setting to vif config
Johannes Berg [Fri, 24 Jun 2022 13:02:16 +0000 (15:02 +0200)]
wifi: mac80211: move ps setting to vif config

This really shouldn't be in a per-link config, we don't want
to let anyone control it that way (if anything, link powersave
could be forced through APIs to activate/deactivate a link),
and we don't support powersave in software with devices that
can do MLO.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: provide link ID in link_conf
Johannes Berg [Fri, 24 Jun 2022 12:42:12 +0000 (14:42 +0200)]
wifi: mac80211: provide link ID in link_conf

It might be useful to drivers to be able to pass only the
link_conf pointer, rather than both the pointer and the
link_id; add the link_id to the link_conf to facility that.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: set up/tear down client vif links properly
Johannes Berg [Fri, 24 Jun 2022 08:57:41 +0000 (10:57 +0200)]
wifi: mac80211: set up/tear down client vif links properly

In station/client mode, the link data needs a bit more
initialization and destruction than just zero-init and
kfree() respectively, implement that.

This required some shuffling of the link data handling
in general, as we should set it up in setup and do the
teardown in teardown, otherwise we're asymmetric in
case of interface type changes.

Also stop using kfree_rcu(), we cannot guarantee that
nothing is scheduling things that live within the link
(e.g. the u.mgd.request_smps_work) until we're sure it
cannot be referenced anymore, therefore synchronize
instead. This isn't very efficient, but we can always
optimize it later.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: move ieee80211_request_smps_mgd_work
Johannes Berg [Fri, 24 Jun 2022 09:15:52 +0000 (11:15 +0200)]
wifi: mac80211: move ieee80211_request_smps_mgd_work

This function can be static.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: acquire wdev mutex for dump_survey
Johannes Berg [Mon, 27 Jun 2022 10:49:03 +0000 (12:49 +0200)]
wifi: nl80211: acquire wdev mutex for dump_survey

At least the quantenna driver calls wdev_chandef() here
which now requires the lock, so acquire it.

Fixes: 7b0a0e3c3a88 ("wifi: cfg80211: do some rework towards MLO link APIs")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: fix key lookup
Johannes Berg [Thu, 23 Jun 2022 20:21:28 +0000 (22:21 +0200)]
wifi: mac80211: fix key lookup

With the split into keys[]/deflink.gtk[] arrays, WEP keys are
still installed into the keys[] array, but we didn't look them
up there. This meant they weren't deleted correctly.

Fix this by looking up the key there even if it's not pairwise
so we can be sure we don't have it.

Fixes: bfd8403adddd ("wifi: mac80211: reorg some iface data structs for MLD")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: separate out connection downgrade flags
Johannes Berg [Thu, 23 Jun 2022 10:04:55 +0000 (12:04 +0200)]
wifi: mac80211: separate out connection downgrade flags

Separate out the connection downgrade flags from the ifmgd->flags
and put them into the link information instead. While at it, make
them a separate sparse type so we don't get confused about where
they belong and have static checking on correct handling.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: Align with Draft P802.11be_D2.0
Ilan Peer [Wed, 1 Jun 2022 14:43:34 +0000 (17:43 +0300)]
wifi: mac80211: Align with Draft P802.11be_D2.0

Align the mac80211 implementation with P802.11be_D2.0.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: Align with Draft P802.11be_D1.5
Ilan Peer [Wed, 27 Apr 2022 15:02:10 +0000 (18:02 +0300)]
wifi: mac80211: Align with Draft P802.11be_D1.5

Align the mac80211 implementation with P802.11be_D1.5.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: skip powersave recalc if driver SUPPORTS_DYNAMIC_PS
Johannes Berg [Thu, 23 Jun 2022 07:48:00 +0000 (09:48 +0200)]
wifi: mac80211: skip powersave recalc if driver SUPPORTS_DYNAMIC_PS

There are a few places that check ps_sdata and/or the dynamic
PS timeout, but they're erroneous in case SUPPORTS_DYNAMIC_PS
is set by the driver.

Skip the entire recalculation in this case so we cannot get
into those paths elsewhere, and so we simplify this for the
purpose of implementing MLO.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: debug: omit link if non-MLO connection
Johannes Berg [Wed, 22 Jun 2022 12:30:20 +0000 (14:30 +0200)]
wifi: mac80211: debug: omit link if non-MLO connection

If we don't really have multiple links, omit the link ID from
link debug prints, otherwise we change the format for all of
the existing drivers (most of which might never support MLO),
and also have extra noise in the logs.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211_hwsim: Ack link addressed frames
Andrei Otcheretianski [Wed, 22 Jun 2022 12:02:37 +0000 (15:02 +0300)]
wifi: mac80211_hwsim: Ack link addressed frames

Do address matching with link addresses as well.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: drop BSS elements from assoc trace for now
Johannes Berg [Wed, 22 Jun 2022 09:02:00 +0000 (11:02 +0200)]
wifi: cfg80211: drop BSS elements from assoc trace for now

For multi-link operation, this cannot work as the req->bss pointer
will be NULL, and we'll need to do more work on this to really add
tracing for the MLO case here. Drop the BSS elements for now as
they're not the most useful thing, and it's hard to size things
correctly for the MLO case (without adding a lot of code that's
also executed when tracing isn't enabled.)

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: make cfg80211_auth_request::key_idx signed
Johannes Berg [Wed, 22 Jun 2022 08:17:23 +0000 (10:17 +0200)]
wifi: cfg80211: make cfg80211_auth_request::key_idx signed

We might assign -1 to it in some cases when key is NULL,
which means the key_idx isn't used but can lead to a
warning from static checkers such as smatch.

Make the struct member signed simply to avoid that, we
only need a range of -1..3 anyway.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: enable setting the link address at new station
Shaul Triebitz [Tue, 21 Jun 2022 09:18:31 +0000 (12:18 +0300)]
wifi: nl80211: enable setting the link address at new station

Since for an MLD station the default link is added together
with the add station command, allow also setting the link
MAC address.
Otherwise, it is needed to use the modify link API only
for setting the link MAC address.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: RCU-ify link/link_conf pointers
Johannes Berg [Fri, 17 Jun 2022 20:36:37 +0000 (22:36 +0200)]
wifi: mac80211: RCU-ify link/link_conf pointers

Since links can be added and removed dynamically, we need to
somehow protect the sdata->link[] and vif->link_conf[] array
pointers from disappearing when accessing them without locks.
RCU-ify the pointers to achieve this, which requires quite a
bit of rework.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: hold wdev mutex for station APIs
Johannes Berg [Mon, 20 Jun 2022 13:29:28 +0000 (15:29 +0200)]
wifi: nl80211: hold wdev mutex for station APIs

Since this will need to refer - at least in part - to the link
stations of an MLD, hold the wdev mutex for driver convenience.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: hold wdev mutex for channel switch APIs
Johannes Berg [Mon, 20 Jun 2022 13:28:50 +0000 (15:28 +0200)]
wifi: nl80211: hold wdev mutex for channel switch APIs

Since we deal with links in an MLD here, hold the wdev
mutex now.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: hold wdev mutex in add/mod/del link station
Johannes Berg [Mon, 20 Jun 2022 13:25:56 +0000 (15:25 +0200)]
wifi: nl80211: hold wdev mutex in add/mod/del link station

Since we deal with links, and that requires looking at wdev links,
we should hold the wdev mutex for driver convenience.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: implement callbacks for <add/mod/del>_link_station
Shaul Triebitz [Tue, 14 Jun 2022 14:27:03 +0000 (17:27 +0300)]
wifi: mac80211: implement callbacks for <add/mod/del>_link_station

Implement callbacks for cfg80211 add_link_station, mod_link_station,
and del_link_station API.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211/mac80211: separate link params from station params
Shaul Triebitz [Tue, 14 Jun 2022 10:49:16 +0000 (13:49 +0300)]
wifi: cfg80211/mac80211: separate link params from station params

Put the link_station_parameters structure in the station_parameters
structure (and remove the station_parameters fields already existing
in link_station_parameters).
Now, for an MLD station, the default link is added together with
the station.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: add API to add/modify/remove a link station
Shaul Triebitz [Wed, 8 Jun 2022 09:01:12 +0000 (12:01 +0300)]
wifi: cfg80211: add API to add/modify/remove a link station

Add an API for adding/modifying/removing a link of a station.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: add an ieee80211_get_link_sband
Shaul Triebitz [Sun, 19 Jun 2022 12:08:02 +0000 (15:08 +0300)]
wifi: mac80211: add an ieee80211_get_link_sband

Similar to ieee80211_get_sband but get the sband of the link_conf.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: Remove AP SMPS leftovers
Andrei Otcheretianski [Sun, 19 Jun 2022 11:38:03 +0000 (14:38 +0300)]
wifi: mac80211: Remove AP SMPS leftovers

AP SMPS was removed and not needed anymore. Remove the leftovers.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: Allow MLO TX with link source address
Andrei Otcheretianski [Tue, 14 Jun 2022 14:21:23 +0000 (17:21 +0300)]
wifi: cfg80211: Allow MLO TX with link source address

Management frames are transmitted from link address and not device
address. Allow that.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: Consider MLO links in offchannel logic
Andrei Otcheretianski [Tue, 14 Jun 2022 14:20:04 +0000 (17:20 +0300)]
wifi: mac80211: Consider MLO links in offchannel logic

Check all the MLO links to decide whether offchannel TX is needed.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: rx: accept link-addressed frames
Johannes Berg [Tue, 14 Jun 2022 11:08:42 +0000 (13:08 +0200)]
wifi: mac80211: rx: accept link-addressed frames

When checking whether or not to accept a frame in RX,
take into account the configured link addresses. Also
look up the link station correctly.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: consistently use sdata_dereference()
Johannes Berg [Fri, 17 Jun 2022 13:16:36 +0000 (15:16 +0200)]
wifi: mac80211: consistently use sdata_dereference()

Instead of open-coding it, use sdata_dereference().

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211_hwsim: Support link channel matching on rx
Andrei Otcheretianski [Tue, 14 Jun 2022 10:41:30 +0000 (13:41 +0300)]
wifi: mac80211_hwsim: Support link channel matching on rx

Accept frames from all the links' channels.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: fix mesh airtime link metric estimating
Aditya Kumar Singh [Fri, 1 Jul 2022 13:36:11 +0000 (19:06 +0530)]
wifi: mac80211: fix mesh airtime link metric estimating

ieee80211s_update_metric function uses sta_set_rate_info_tx
function to get struct rate_info data from ieee80211_tx_rate
struct, present in ieee80211_sta->deflink.tx_stats. However,
drivers can skip tx rate calculation by setting rate idx as
-1. Such drivers provides rate_info directly and hence
ieee80211s metric is updated incorrectly since ieee80211_tx_rate
has inconsistent data.

Add fix to use rate_info directly if present instead of
sta_set_rate_info_tx for updating ieee80211s metric.

Signed-off-by: Aditya Kumar Singh <quic_adisi@quicinc.com>
Link: https://lore.kernel.org/r/20220701133611.544-1-quic_adisi@quicinc.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: make 4addr null frames using min_rate for WDS
Lian Chen [Thu, 14 Jul 2022 09:16:36 +0000 (17:16 +0800)]
wifi: mac80211: make 4addr null frames using min_rate for WDS

WDS needs 4addr packets to trigger AP for wlan0.sta creation.
However, the 4addr null frame is sent at a high rate so that
sometimes the AP can't receive it. Switch to using min rate.

Signed-off-by: Lian Chen <lian.chen@mediatek.com>
Link: https://lore.kernel.org/r/20220714091636.59107-1-lian.chen@mediatek.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: cfg80211: use strscpy to replace strlcpy
XueBing Chen [Mon, 11 Jul 2022 14:27:58 +0000 (22:27 +0800)]
wifi: cfg80211: use strscpy to replace strlcpy

The strlcpy should not be used because it doesn't limit the source
length. Preferred is strscpy.

Signed-off-by: XueBing Chen <chenxuebing@jari.cn>
Link: https://lore.kernel.org/r/2d2fcbf7.e33.181eda8e70e.Coremail.chenxuebing@jari.cn
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: exclude multicast packets from AQL pending airtime
Felix Fietkau [Wed, 13 Jul 2022 08:34:44 +0000 (10:34 +0200)]
wifi: mac80211: exclude multicast packets from AQL pending airtime

In AP mode, multicast traffic is handled very differently from normal traffic,
especially if at least one client is in powersave mode.
This means that multicast packets can be buffered a lot longer than normal
unicast packets, and can eat up the AQL budget very quickly because of the low
data rate.
Along with the recent change to maintain a global PHY AQL limit, this can lead
to significant latency spikes for unicast traffic.

Since queueing multicast to hardware is currently not constrained by AQL limits
anyway, let's just exclude it from the AQL pending airtime calculation entirely.

Fixes: 8e4bac067105 ("wifi: mac80211: add a per-PHY AQL limit to improve fairness")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20220713083444.86129-1-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211_hwsim: use 32-bit skb cookie
Johannes Berg [Wed, 13 Jul 2022 19:16:45 +0000 (21:16 +0200)]
wifi: mac80211_hwsim: use 32-bit skb cookie

We won't really have enough skbs to need a 64-bit cookie,
and on 32-bit platforms storing the 64-bit cookie into the
void *rate_driver_data doesn't work anyway. Switch back to
using just a 32-bit cookie and uintptr_t for the type to
avoid compiler warnings about all this.

Fixes: 4ee186fa7e40 ("wifi: mac80211_hwsim: fix race condition in pending packet")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: rsi: remove unused variable
Johannes Berg [Wed, 13 Jul 2022 19:16:37 +0000 (21:16 +0200)]
wifi: rsi: remove unused variable

Remove a variable here that was now set but never used.

Fixes: f276e20b182d ("wifi: mac80211: move interface config to new struct")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agoMerge tag 'mlx5-updates-2022-07-13' of git://git.kernel.org/pub/scm/linux/kernel...
Jakub Kicinski [Fri, 15 Jul 2022 05:11:37 +0000 (22:11 -0700)]
Merge tag 'mlx5-updates-2022-07-13' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2022-07-13

1) Support 802.1ad for bridge offloads

Vlad Buslov Says:
=================

Current mlx5 bridge VLAN offload implementation only supports 802.1Q VLAN
Ethernet protocol. That protocol type is assumed by default and
SWITCHDEV_ATTR_ID_BRIDGE_VLAN_PROTOCOL notification is ignored.

In order to support dynamically setting VLAN protocol handle
SWITCHDEV_ATTR_ID_BRIDGE_VLAN_PROTOCOL notification by flushing FDB and
re-creating VLAN modify header actions with a new protocol. Implement support
for 802.1ad protocol by saving the current VLAN protocol to per-bridge variable
and re-create the necessary flow groups according to its current value (either
use cvlan or svlan flow fields).
==================

2) debugfs to count ongoing FW commands

3) debugfs to query eswitch vport firmware diagnostic counters

4) Add missing meter configuration in flow action

5) Some misc cleanup

* tag 'mlx5-updates-2022-07-13' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5e: Remove the duplicating check for striding RQ when enabling LRO
  net/mlx5e: Move the LRO-XSK check to mlx5e_fix_features
  net/mlx5e: Extend flower police validation
  net/mlx5e: configure meter in flow action
  net/mlx5e: Removed useless code in function
  net/mlx5: Bridge, implement QinQ support
  net/mlx5: Bridge, implement infrastructure for VLAN protocol change
  net/mlx5: Bridge, extract VLAN push/pop actions creation
  net/mlx5: Bridge, rename filter fg to vlan_filter
  net/mlx5: Bridge, refactor groups sizes and indices
  net/mlx5: debugfs, Add num of in-use FW command interface slots
  net/mlx5: Expose vnic diagnostic counters for eswitch managed vports
  net/mlx5: Use software VHCA id when it's supported
  net/mlx5: Introduce ifc bits for using software vhca id
  net/mlx5: Use the bitmap API to allocate bitmaps
====================

Link: https://lore.kernel.org/r/20220713225859.401241-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge branch 'net-devlink-couple-of-trivial-fixes'
Jakub Kicinski [Fri, 15 Jul 2022 04:58:49 +0000 (21:58 -0700)]
Merge branch 'net-devlink-couple-of-trivial-fixes'

Jiri Pirko says:

====================
net: devlink: couple of trivial fixes

Just a couple of trivial fixes I found on the way.
====================

Link: https://lore.kernel.org/r/20220713141853.2992014-1-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: devlink: fix return statement in devlink_port_new_notify()
Jiri Pirko [Wed, 13 Jul 2022 14:18:53 +0000 (16:18 +0200)]
net: devlink: fix return statement in devlink_port_new_notify()

Return directly without intermediate value store at the end of
devlink_port_new_notify() function.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: devlink: fix a typo in function name devlink_port_new_notifiy()
Jiri Pirko [Wed, 13 Jul 2022 14:18:52 +0000 (16:18 +0200)]
net: devlink: fix a typo in function name devlink_port_new_notifiy()

Fix the typo in a name of devlink_port_new_notifiy() function.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: devlink: make devlink_dpipe_headers_register() return void
Jiri Pirko [Wed, 13 Jul 2022 14:18:51 +0000 (16:18 +0200)]
net: devlink: make devlink_dpipe_headers_register() return void

The return value is not used, so change the return value type to void.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>