Add write pointer and cycle count registers to smem monitor header.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Add sizeof(struct iwl_fw_ini_error_dump_range) to the header of the dram
monitor.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

AX210 devices will use u64 for the base address to the DRAM monitor
buffer. To support this, change the structure for all device families
so both address sizes fit.

Also move range_data_size to the top of the struct to ease the parsing
of the memory range.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Back in commit 4d339989 ("iwlwifi: mvm: support ibss in dqa mode")
we changed queue selection for IBSS to be:

    if (ieee80211_is_probe_resp(fc) || ieee80211_is_auth(fc) ||
        ieee80211_is_deauth(fc))
            return IWL_MVM_DQA_AP_PROBE_RESP_QUEUE;
    if (info->hw_queue == info->control.vif->cab_queue)
            return info->hw_queue;
    return IWL_MVM_DQA_AP_PROBE_RESP_QUEUE;

Clearly, the thought at the time must've been that mac80211 will
select the hw_queue as the cab_queue, so that we'll return and use
that, where we store the multicast queue for IBSS. This, however,
isn't true because mac80211 doesn't implement powersave for IBSS
and thus selects the normal IBSS interface AC queue (best effort).

This therefore always used the probe response queue, which maps to
the BE FIFO.

In commit cfbc6c4c ("iwlwifi: mvm: support mac80211 TXQs model")
we rethought this code, and as a consequence now started mapping the
multicast traffic to the multicast hardware queue since we no longer
relied on mac80211 selecting the queue, doing it ourselves instead.
This queue is mapped to the MCAST FIFO. however, this isn't actually
enabled/controlled by the firmware in IBSS mode because we don't
implement powersave, and frames from this queue can never go out in
this case.

Therefore, we got queue hang reports such as
https://bugzilla.kernel.org/show_bug.cgi?id=201707

Fix this by mapping the multicast queue to the BE FIFO in IBSS so
that all the frames can go out.

Fixes: cfbc6c4c ("iwlwifi: mvm: support mac80211 TXQs model")
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Add to the fifo dump the registers addresses.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

When initializing or overriding HE band capabilities, no
need to check the band validity.
Trust the calling function to use a valid band.
Signed-off-by: NShaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Add the region id of the collected memory to the header of the memory
region.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Add version to dump header to allow future changes of the dump struct,
once the ini debug flow becomes operational, without breaking backwards
compatibility.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Add the offset to the base address of a memory region to show the actual
addresses being read.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

The driver is using range->start_addr before assigning it a value.
Set value into range->start_addr and then use it.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Enforce domain checking before sending host commands and collecting
memory regions. Currently the driver supports always on domain only.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Differentiate between SW and HW error interrupts and support ini HW
error trigger.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Since maxattr is common, the policy can't really differ sanely,
so make it common as well.

The only user that did in fact manage to make a non-common policy
is taskstats, which has to be really careful about it (since it's
still using a common maxattr!). This is no longer supported, but
we can fake it using pre_doit.

This reduces the size of e.g. nl80211.o (which has lots of commands):

   text	   data	    bss	    dec	    hex	filename
 398745	  14323	   2240	 415308	  6564c	net/wireless/nl80211.o (before)
 397913	  14331	   2240	 414484	  65314	net/wireless/nl80211.o (after)
--------------------------------
   -832      +8       0    -824

Which is obviously just 8 bytes for each command, and an added 8
bytes for the new policy pointer. I'm not sure why the ops list is
counted as .text though.

Most of the code transformations were done using the following spatch:
    @ops@
    identifier OPS;
    expression POLICY;
    @@
    struct genl_ops OPS[] = {
    ...,
     {
    -	.policy = POLICY,
     },
    ...
    };

    @@
    identifier ops.OPS;
    expression ops.POLICY;
    identifier fam;
    expression M;
    @@
    struct genl_family fam = {
            .ops = OPS,
            .maxattr = M,
    +       .policy = POLICY,
            ...
    };

This also gets rid of devlink_nl_cmd_region_read_dumpit() accessing
the cb->data as ops, which we want to change in a later genl patch.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

align to ini debug struct version 1 and enforce version checking.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

This goto out was unnecessary because the out label was immediately
below it.  Remove it.
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Use struct_size() in kzalloc instead of the 'regd_to_copy'
Signed-off-by: NYueHaibing <yuehaibing@huawei.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Separate between ini and legacy dump flows to allow adding ini triggers
that are not supported in the legacy flow and to increase readabilty.

iwl_fw_dbg_ini_collect function is now called with legacy trigger id and
_iwl_fw_dbg_ini_collect is called with ini trigger id.

Also make the actual dumping function static so that any dump collection
will go through iwl_fw_dbg_collect_sync.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Allows to perform monitor dumping on AX210 device family
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

mac80211 sets max amsdu to min supported ht vs vht but TLC only works
with one mode so we can set to the exact mode used (vht/ht)
and enable larger amsdu sizes for vht.
Signed-off-by: NMordechay Goodstein <mordechay.goodstein@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Add support for radio gf4 (CDB radio).
Signed-off-by: NShaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

The layout of the RBD (receive buffer descriptor) isn't quite right,
the hardware ended up being implemented differently. Switch to the
correct RBD layout. While at it, remove the now useless extra defines.

Also, switch the CD (completion descriptor) to the right format, which
is basically just a code cleanup because the only field we really used
(rbid) is still in the same place. We may need fragmentation later if
we ever want to use it.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

In restart flow, the driver requests HW restart from mac80211
and then mac80211 uses a worker to do the restart flow. In that flow a
sync dump is performed. Instead, schedule the dump worker before
requesting HW restart from mac80211. This approach simplifies the
restart flow.
Also, it is neeeded in order to differentiate between the handling of SW
and HW errors in a future commit.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Only add the size of the tx command to the dump file size if it is set
in the dump_mask.
Signed-off-by: NShahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Newer FWs use a new format of the SCAN_OFFLOAD_PROFILES_QUERY_RSP,
which now supports indicating match on an higher number of channels.

Modify the code to support both the old format and the newer one,
based on a FW TLV.
Signed-off-by: NIlan Peer <ilan.peer@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Support REGULATORY_NVM_GET_INFO_RSP_API_S_VER_4.
This API adds the new 6-7GHz channels.
Signed-off-by: NShaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Add UHB (ultra high band) channels and use 16 bit variables
to fit the new channels.
Signed-off-by: NShaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

If number of channels in the driver is greater than
number of scan channels given by firmware TLV, do not
fail scan config, but adjust to firmware's number of channels.
This is helpful for supporting in driver new channels before
it being supported by firmware scan.
Signed-off-by: NShaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

For some reason we never enabled it, but it appears to work fine.
Enable it now.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Set the capabilities flags and inform firmware
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

There are some buggy APs that keeps changing the count while forcing
us to block TX. This eventually results in queue hang, assert, and
disconnection. Detect such APs and disconnect gracefully in advance.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

In case we receive channel switch announcement with immediate
quiet and unknown switching time, we will switch when FW identifies
AP left channel. However, if AP remains on channel, we will
eventually get TX queue hang. Init a work to disconnect if
switch doesn't occur within 1500 milliseconds. Do it also
for a too long channel switch.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Let FW know if quiet mode is on or not. This is needed
in order to disable it in FW when CSA is complete.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Send to FW modify command for every beacon we receive during channel
switch. FW will track the count, and make sure the event is scheduled
in time even if AP changed count.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Report all NO_DATA events to mac80211 so they get captured
in radiotap for usage in sniffer scenarios; map the info
type to a reasonable radiotap type for this.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

In case we receive abort operation for CSA, clean up
our state.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

Support passing to FW delayed beacon count. This
represents the delay the AP can have when moving to
the new channel.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>

After the previous patch, all the callers of ndo_select_queue()
provide as a 'fallback' argument netdev_pick_tx.
The only exceptions are nested calls to ndo_select_queue(),
which pass down the 'fallback' available in the current scope
- still netdev_pick_tx.

We can drop such argument and replace fallback() invocation with
netdev_pick_tx(). This avoids an indirect call per xmit packet
in some scenarios (TCP syn, UDP unconnected, XDP generic, pktgen)
with device drivers implementing such ndo. It also clean the code
a bit.

Tested with ixgbe and CONFIG_FCOE=m

With pktgen using queue xmit:
threads		vanilla 	patched
		(kpps)		(kpps)
1		2334		2428
2		4166		4278
4		7895		8100

 v1 -> v2:
 - rebased after helper's name change
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

My commit 26a7b547 ("mt76x02: set protection according to ht
operation element") enabled by default RTS/CTS protection for OFDM
and CCK traffic, because MT_TX_RTS_CFG_THRESH is configured to non
0xffff by initvals and .set_rts_threshold callback is not called by
mac80211 on initialization, only on user request or during
ieee80211_reconfig() (suspend/resuem or restart_hw).

Enabling RTS/CTS cause some problems when sending probe request
frames by hcxdumptool penetration tool, but I expect it can cause
other issues on different scenarios.

Restore previous setting of RTS/CTS being disabled by default for
OFDM/CCK by changing MT_TX_RTS_CFG_THRESH initvals to 0xffff.

Fixes: 26a7b547 ("mt76x02: set protection according to ht operation element")
Signed-off-by: NStanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>

Fix following schedule while atomic in mt76x02_reset_state
since synchronize_rcu is run inside a RCU section

[44036.944222] mt76x2e 0000:06:00.0: MCU message 31 (seq 3) timed out
[44036.944281] BUG: sleeping function called from invalid context at kernel/rcu/tree_exp.h:818
[44036.944284] in_atomic(): 1, irqs_disabled(): 0, pid: 28066, name: kworker/u4:1
[44036.944287] INFO: lockdep is turned off.
[44036.944292] CPU: 1 PID: 28066 Comm: kworker/u4:1 Tainted: G        W         5.0.0-rc7-wdn-t1+ #7
[44036.944294] Hardware name: Dell Inc. Studio XPS 1340/0K183D, BIOS A11 09/08/2009
[44036.944305] Workqueue: phy1 mt76x02_wdt_work [mt76x02_lib]
[44036.944308] Call Trace:
[44036.944317]  dump_stack+0x67/0x90
[44036.944322]  ___might_sleep.cold.88+0x9f/0xaf
[44036.944327]  rcu_blocking_is_gp+0x13/0x50
[44036.944330]  synchronize_rcu+0x17/0x80
[44036.944337]  mt76_sta_state+0x138/0x1d0 [mt76]
[44036.944349]  mt76x02_wdt_work+0x1c9/0x610 [mt76x02_lib]
[44036.944355]  process_one_work+0x2a5/0x620
[44036.944361]  worker_thread+0x35/0x3e0
[44036.944368]  kthread+0x11c/0x140
[44036.944376]  ret_from_fork+0x3a/0x50
[44036.944384] BUG: scheduling while atomic: kworker/u4:1/28066/0x00000002
[44036.944387] INFO: lockdep is turned off.
[44036.944389] Modules linked in: cmac ctr ccm af_packet snd_hda_codec_hdmi

Introduce __mt76_sta_remove in order to run sta_remove without holding dev->mutex.
Move __mt76_sta_remove outside of RCU section in mt76x02_reset_state

Fixes: e4ebb8b403d1 ("mt76: mt76x2: implement full device restart on watchdog reset")
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>

__sw_hweight8() is only defined if CONFIG_GENERIC_HWEIGHT is enabled.
The function that works on all architectures is hweight8().
Signed-off-by: NFelix Fietkau <nbd@nbd.name>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>