When register_netdevice() failed in ieee80211_if_add(), ndev->tstats
isn't released. Fix it.

Fixes: 5a490510 ("mac80211: use per-CPU TX/RX statistics")
Signed-off-by: NZhengchao Shao <shaozhengchao@huawei.com>
Link: https://lore.kernel.org/r/20221117064500.319983-1-shaozhengchao@huawei.comSigned-off-by: NJohannes Berg <johannes.berg@intel.com>

All drivers are now using mac80211 internal queues (iTXQs).
Drop mac80211 internal support for the old push path.
Signed-off-by: NAlexander Wetzel <alexander@wetzel-home.de>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Properly handle TX stop for internal queues (iTXQs) within mac80211.

mac80211 must not stop netdev queues when using mac80211 iTXQs.
For these drivers the netdev interface is created with IFF_NO_QUEUE.

While netdev still drops frames for IFF_NO_QUEUE interfaces when we stop
the netdev queues, it also prints a warning when this happens:
Assuming the mac80211 interface is called wlan0 we would get
"Virtual device wlan0 asks to queue packet!" when netdev has to drop a
frame.

This patch is keeping the harmless netdev queue starts for iTXQ drivers.
Signed-off-by: NAlexander Wetzel <alexander@wetzel-home.de>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Unlock before returning -EOPNOTSUPP.

Fixes: 3c06e91b ("wifi: mac80211: Support POWERED_ADDR_CHANGE feature")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Implement an API function and debugfs file to switch
active links.

Also provide an async version of the API so drivers
can call it in arbitrary contexts, e.g. while in the
authorized callback.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

We probably should've done that originally, we already have
about 300 lines of code there, and will add more. Move all
the link code we wrote to a new file.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Adds support in mac80211 for NL80211_EXT_FEATURE_POWERED_ADDR_CHANGE.
The motivation behind this functionality is to fix limitations of
address randomization on frequencies which are disallowed in world
roaming.

The way things work now, if a client wants to randomize their address
per-connection it must power down the device, change the MAC, and
power back up. Here lies a problem since powering down the device
may result in frequencies being disabled (until the regdom is set).
If the desired BSS is on one such frequency the client is unable to
connect once the phy is powered again.

For mac80211 based devices changing the MAC while powered is possible
but currently disallowed (-EBUSY). This patch adds some logic to
allow a MAC change while powered by removing the interface, changing
the MAC, and adding it again. mac80211 will advertise support for
this feature so userspace can determine the best course of action e.g.
disallow address randomization on certain frequencies if not
supported.

There are certain limitations put on this which simplify the logic:
 - No active connection
 - No offchannel work, including scanning.
Signed-off-by: NJames Prestwood <prestwoj@gmail.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Do not allow VLANs to be added to AP interfaces that are
MLDs, this isn't going to work because the link structs
aren't propagated to the VLAN interfaces yet.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Link: https://lore.kernel.org/r/20220902161144.8c88531146e9.If2ef9a3b138d4f16ed2fda91c852da156bdf5e4d@changeidSigned-off-by: NJohannes Berg <johannes.berg@intel.com>

Follow the advice of the below link and prefer 'strscpy' in this
subsystem. Conversion is 1:1 because the return value is not used.
Generated by a coccinelle script.

Link: https://lore.kernel.org/r/CAHk-=wgfRnXz0W3D37d01q3JFkr_i_uTL=V6A6G1oUZcprmknw@mail.gmail.com/Signed-off-by: NWolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Implement key installation and lookup (on TX and RX)
for MLO, so we can use multiple GTKs/IGTKs/BIGTKs.
Co-authored-by: NIlan Peer <ilan.peer@intel.com>
Signed-off-by: NIlan Peer <ilan.peer@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

For an AP interface, set the link BSSID when the link
is initialized.
Signed-off-by: NShaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

lockdep complains use of uninitialized spinlock at ieee80211_do_stop() [1],
for commit f856373e ("wifi: mac80211: do not wake queues on a vif
that is being stopped") guards clear_bit() using fq.lock even before
fq_init() from ieee80211_txq_setup_flows() initializes this spinlock.

According to discussion [2], Toke was not happy with expanding usage of
fq.lock. Since __ieee80211_wake_txqs() is called under RCU read lock, we
can instead use synchronize_rcu() for flushing ieee80211_wake_txqs().

Link: https://syzkaller.appspot.com/bug?extid=eceab52db7c4b961e9d6 [1]
Link: https://lkml.kernel.org/r/874k0zowh2.fsf@toke.dk [2]
Reported-by: Nsyzbot <syzbot+eceab52db7c4b961e9d6@syzkaller.appspotmail.com>
Signed-off-by: NTetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Fixes: f856373e ("wifi: mac80211: do not wake queues on a vif that is being stopped")
Tested-by: Nsyzbot <syzbot+eceab52db7c4b961e9d6@syzkaller.appspotmail.com>
Acked-by: NToke Høiland-Jørgensen <toke@kernel.org>
Signed-off-by: NKalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/9cc9b81d-75a3-3925-b612-9d0ad3cab82b@I-love.SAKURA.ne.jp
[ pick up commit 3598cb6e ("wifi: mac80211: do not abuse fq.lock in ieee80211_do_stop()") from -next]
Link: https://lore.kernel.org/all/87o7xcq6qt.fsf@kernel.org/Signed-off-by: NJakub Kicinski <kuba@kernel.org>

During the code reshuffling, I accidentally set this to
NULL before using it, fix that to fix the link data leak.

Fixes: d3e2439b ("wifi: mac80211: fix link manipulation")
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When adding multiple links, verify that they all have
different addresses.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

lockdep complains use of uninitialized spinlock at ieee80211_do_stop() [1],
for commit f856373e ("wifi: mac80211: do not wake queues on a vif
that is being stopped") guards clear_bit() using fq.lock even before
fq_init() from ieee80211_txq_setup_flows() initializes this spinlock.

According to discussion [2], Toke was not happy with expanding usage of
fq.lock. Since __ieee80211_wake_txqs() is called under RCU read lock, we
can instead use synchronize_rcu() for flushing ieee80211_wake_txqs().

Link: https://syzkaller.appspot.com/bug?extid=eceab52db7c4b961e9d6 [1]
Link: https://lkml.kernel.org/r/874k0zowh2.fsf@toke.dk [2]
Reported-by: Nsyzbot <syzbot+eceab52db7c4b961e9d6@syzkaller.appspotmail.com>
Signed-off-by: NTetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Fixes: f856373e ("wifi: mac80211: do not wake queues on a vif that is being stopped")
Tested-by: Nsyzbot <syzbot+eceab52db7c4b961e9d6@syzkaller.appspotmail.com>
Acked-by: NToke Høiland-Jørgensen <toke@kernel.org>
Signed-off-by: NKalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/9cc9b81d-75a3-3925-b612-9d0ad3cab82b@I-love.SAKURA.ne.jp

When a link is stopped for removal, release the channel
context it may have.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

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: NJohannes Berg <johannes.berg@intel.com>

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: NJohannes Berg <johannes.berg@intel.com>

Take the link into account in the QoS settings (EDCA parameters)
APIs.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

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: NJohannes Berg <johannes.berg@intel.com>

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: NJohannes Berg <johannes.berg@intel.com>

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: NJohannes Berg <johannes.berg@intel.com>

This reverts commits 6a789ba6 and
2433647b.

The virtual time scheduler code has a number of issues:
- queues slowed down by hardware/firmware powersave handling were not properly
  handled.
- on ath10k in push-pull mode, tx queues that the driver tries to pull from
  were starved, causing excessive latency
- delay between tx enqueue and reported airtime use were causing excessively
  bursty tx behavior

The bursty behavior may also be present on the round-robin scheduler, but there
it is much easier to fix without introducing additional regressions
Signed-off-by: NFelix Fietkau <nbd@nbd.name>
Acked-by: NToke Høiland-Jørgensen <toke@toke.dk>
Link: https://lore.kernel.org/r/20220625212411.36675-1-nbd@nbd.nameSigned-off-by: NJohannes Berg <johannes.berg@intel.com>

When a vif is being removed and sdata->bss is cleared, __ieee80211_wake_txqs
can still be called on it, which crashes as soon as sdata->bss is being
dereferenced.
To fix this properly, check for SDATA_STATE_RUNNING before waking queues,
and take the fq lock when setting it (to ensure that __ieee80211_wake_txqs
observes the change when running on a different CPU)
Signed-off-by: NFelix Fietkau <nbd@nbd.name>
Acked-by: NToke Høiland-Jørgensen <toke@kernel.org>
Link: https://lore.kernel.org/r/20220531190824.60019-1-nbd@nbd.nameSigned-off-by: NJohannes Berg <johannes.berg@intel.com>

We need to be able to access these in a race-free way under
traffic while adding/removing them, so RCU-ify the pointers.
This requires passing a link_sta to a lot of functions so
we don't have to do the RCU handling everywhere.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Add the necessary infrastructure, including a new driver
method, to add/remove links to/from an interface.

Also add the missing link address to bss_conf (which we
use as link_conf too), and fill it, in station mode for
now just randomly, in AP mode we get the address from
cfg80211 since the link must be created with an address
first.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Start making some SMPS related code MLD-aware. This isn't
really done yet, but again cuts down our 'deflink' reliance.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Update the code in vht.c and add the link_id parameter where
necessary.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

We don't need to setup lists and work structs every time
we switch the interface type, factor that out into a new
ieee80211_link_init() function and use it.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Make the channel context code MLO aware, along with some
functions that it uses, so that the chan.c file is now
MLD-clean and no longer uses deflink/bss_conf/etc.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Add pointers so we can start using link_id throughout the
code, even if for now only link ID 0 is valid, pointing
to the "built-in" bss_conf, which is used by drivers that
are not aware of MLD.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Split the bss_info_changed method to vif_cfg_changed and
link_info_changed, with the latter getting a link ID.
Also change the 'changed' parameter to u64 already, we
know we need that.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Start reorganizing interface related data structures toward
MLD. The most complex part here is for the keys, since we
have to split the various kinds of GTKs off to the link but
still need to use (for WEP) the other keys as a fallback
even for multicast frames.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

We'll use bss_conf for per-link configuration later, so
move out all the non-link-specific data out into a new
struct ieee80211_vif_cfg used in the vif.

Some adjustments were done with the following spatch:

    @@
    expression sdata;
    struct ieee80211_vif *vifp;
    identifier var = { assoc, ibss_joined, aid, arp_addr_list, arp_addr_cnt, ssid, ssid_len, s1g, ibss_creator };
    @@
    (
    -sdata->vif.bss_conf.var
    +sdata->vif.cfg.var
    |
    -vifp->bss_conf.var
    +vifp->cfg.var
    )

    @bss_conf@
    struct ieee80211_bss_conf *bss_conf;
    identifier var = { assoc, ibss_joined, aid, arp_addr_list, arp_addr_cnt, ssid, ssid_len, s1g, ibss_creator };
    @@
    -bss_conf->var
    +vif_cfg->var

(though more manual fixups were needed, e.g. replacing
"vif_cfg->" by "vif->cfg." in many files.)
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

To add MLD, reuse the bss_conf structure later for per-link
information, so move some things into it that are per link.

Most transformations were done with the following spatch:

    @@
    expression sdata;
    identifier var = { chanctx_conf, mu_mimo_owner, csa_active, color_change_active, color_change_color };
    @@
    -sdata->vif.var
    +sdata->vif.bss_conf.var

    @@
    struct ieee80211_vif *vif;
    identifier var = { chanctx_conf, mu_mimo_owner, csa_active, color_change_active, color_change_color };
    @@
    -vif->var
    +vif->bss_conf.var
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The only driver using this was iwlwifi, where we just removed
the support because it was never really used. Remove the code
from mac80211 as well.

Change-Id: I1667417a5932315ee9d81f5c233c56a354923f09
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

This allows drivers to provide a destination device + info for flow offload
Only supported in combination with 802.3 encap offload
Signed-off-by: NFelix Fietkau <nbd@nbd.name>
Tested-by: NLorenzo Bianconi <lorenzo@kernel.org>
Link: https://lore.kernel.org/r/20211112112223.1209-1-nbd@nbd.nameSigned-off-by: NJohannes Berg <johannes.berg@intel.com>

Since commit a05829a7 ("cfg80211: avoid holding the RTNL when
calling the driver") we've not only been protecting the pointer
to monitor_sdata with the RTNL, but also with the wiphy->mtx. This
is relevant in a number of lockdep assertions, e.g. the one we hit
in ieee80211_set_monitor_channel(). However, we're now protecting
all the assignments/dereferences, even the one in interface iter,
with the wiphy->mtx, so switch over the lockdep assertions to that
lock.

Fixes: a05829a7 ("cfg80211: avoid holding the RTNL when calling the driver")
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Link: https://lore.kernel.org/r/20211112135143.cb8e8ceffef3.Iaa210f16f6904c8a7a24954fb3396da0ef86ec08@changeidSigned-off-by: NJohannes Berg <johannes.berg@intel.com>

Commit 406f42fa ("net-next: When a bond have a massive amount
of VLANs...") introduced a rbtree for faster Ethernet address look
up. To maintain netdev->dev_addr in this tree we need to make all
the writes to it got through appropriate helpers.

Convert mac80211 from memcpy(... ETH_ADDR) to eth_hw_addr_set():

  @@
  expression dev, np;
  @@
  - memcpy(dev->dev_addr, np, ETH_ALEN)
  + eth_hw_addr_set(dev, np)
Signed-off-by: NJakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20211019162816.1384077-1-kuba@kernel.orgSigned-off-by: NJohannes Berg <johannes.berg@intel.com>

Configure multiple BSSID and enhanced multi-BSSID advertisement (EMA)
parameters in mac80211 for AP mode.

For each interface, 'mbssid_tx_vif' points to the transmitting interface of
the MBSSID set. The pointer is set to NULL if MBSSID is disabled.

Function ieee80211_stop() is modified to always bring down all the
non-transmitting interfaces first and the transmitting interface last.
Signed-off-by: NJohn Crispin <john@phrozen.org>
Co-developed-by: NAloka Dixit <alokad@codeaurora.org>
Signed-off-by: NAloka Dixit <alokad@codeaurora.org>
Link: https://lore.kernel.org/r/20210916025437.29138-3-alokad@codeaurora.org
[slightly change logic to be more obvious]
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>