Init functions should be placed at the end of
files in most cases to avoid forward declarations
for static functions.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Simplify the code by deduplicating structure
definitions and code.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

The structure is being set up in 2 places.
Deduplicate the code by creating a helper.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

This command is not used anymore and most firmware
revisions do not seem to handle it well. Channel
switching is done via vdev restarting.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Firmware reports the number of RF chains so use
that for initialization of supp_{tx,rx}_chainmask
instead of using a macro for 3x3 chips.

This should make tx/rx chainmask reports correct
for chips other than 3x3.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Some firmware revisions don't wait for beacon tx
completion before sending another SWBA event. This
could lead to hardware using old (freed) beacon
data in some cases, e.g. tx credit starvation
combined with missed TBTT. This is very very rare.

On non-IOMMU-enabled hosts this could be a
possible security issue because hw could beacon
some random data on the air.  On IOMMU-enabled
hosts DMAR faults would occur in most cases and
target device would crash.

Since there are no beacon tx completions (implicit
nor explicit) propagated to host the only
workaround for this is to allocate a DMA-coherent
buffer for a lifetime of a vif and use it for all
beacon tx commands. Worst case for this approach
is some beacons may become corrupted, e.g. garbled
IEs or out-of-date TIM bitmap.

Keep the original beacon-related code as-is in
case future firmware revisions solve this problem
so that the old path can be easily re-enabled with
a fw_feature flag.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

As suggeested by checkpatch:

WARNING: Prefer ether_addr_copy() over memcpy() if the Ethernet addresses are __aligned(2)

In wmi.c I had to change due to sparse warnings copying of struct wmi_mac_addr
from form &cmd->peer_macaddr.addr to cmd->peer_macaddr.addr. In
ath10k_wmi_set_ap_ps_param() I also added the missing ".addr" to the copy
command.
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Makes it more readable and fixes checkpatch warnings:

drivers/net/wireless/ath/ath10k/debug.c:593: WARNING: quoted string split across lines
drivers/net/wireless/ath/ath10k/debug.c:594: WARNING: quoted string split across lines
drivers/net/wireless/ath/ath10k/debug.c:595: WARNING: quoted string split across lines
drivers/net/wireless/ath/ath10k/debug.c:596: WARNING: quoted string split across lines
drivers/net/wireless/ath/ath10k/debug.c:597: WARNING: quoted string split across lines
drivers/net/wireless/ath/ath10k/debug.c:599: WARNING: quoted string split across lines
drivers/net/wireless/ath/ath10k/debug.c:600: WARNING: quoted string split across lines
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Fixes checkpatch warnings:

ath10k/htc.c:49: WARNING: Possible unnecessary 'out of memory' message
ath10k/htc.c:810: WARNING: Possible unnecessary 'out of memory' message
ath10k/htt.h:1034: CHECK: Please use a blank line after function/struct/union/enum declarations
ath10k/htt_rx.c:135: CHECK: Unnecessary parentheses around htt->rx_ring.alloc_idx.vaddr
ath10k/htt_rx.c:173: CHECK: Unnecessary parentheses around htt->rx_ring.alloc_idx.vaddr
ath10k/pci.c:633: WARNING: macros should not use a trailing semicolon
ath10k/wmi.c:3594: WARNING: quoted string split across lines
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Fixes checkpatch warnings:

WARNING: else is not generally useful after a break or return
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Fixes a checkpatch warning:

WARNING: void function return statements are not generally useful
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Fixes checkpatch warnings:

CHECK: No space is necessary after a cast
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Fixes checkpatch warnings:

WARNING: Missing a blank line after declarations

Please note that some of the cases I fixed by moving the variable declarations
to the beginning of the function, which is the preferred style in ath10k.
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Fixes checkpatch warnings:

CHECK: Please don't use multiple blank lines
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

CHECK: Alignment should match open parenthesis
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Fixes checkpatch warning:

ath10k/htt_rx.c:519: CHECK: Alignment should match open parenthesis
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

The function may return an -ENODEV if debugfs is
disabled in kernel. This should originally be
guarded by ath10k's Kconfig but it still makes
sense to check for the non-NULL errno return
value.
Reported-by: NMatteo Croce <matteo@openwrt.org>
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

10.1 firmware does not have an official way to
cause assert on purpose, but it can be done with
carefully crafted WMI command.  This is a different
kind of crash from the 'hard' crash, which is
a bad memory dereference.

Different crashes decode in different manners, so
this will help the crash-report testing as well as
offer better ways to test firmware failure and
recovery.

kvalo: move the wmi command creation to debug.c, modify
the info print
Signed-off-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Add testmode interface for starting and using UTF firmware which is used to run
factory tests. This is implemented by adding new state ATH10K_STATE_UTF and user
space can enable this state with ATH10K_TM_CMD_UTF_START command. To go back to
normal mode user space can send ATH10K_TM_CMD_UTF_STOP.
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

We need this function to send wmi packets from testmode.c.
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

ATH10K_DEBUGFS must depend on DEBUG_FS, otherwise
ath10k will generate an invalid pointer on module load.
Signed-off-by: NMatteo Croce <matteo@openwrt.org>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

On 32bit systems the bitmap was too small and it
was overwritten partially by the stat completion
structure. This was visible with 10.2 firmware
only due to it using a few of the last service
ids.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

The fw_crash_data was allocated too late. Upon
early firmware crash, before registering to
mac80211, it was possible to crash the whole
system:

 ath10k_pci 0000:00:05.0: device has crashed during init
 BUG: unable to handle kernel NULL pointer dereference at           (null)
 IP: [<ffffffffa0058005>] ath10k_debug_get_new_fw_crash_data+0x15/0x30 [ath10k_core]
 PGD 0
 Oops: 0002 [#1] SMP
 Modules linked in: ath10k_pci(O) ath10k_core(O) ath [last unloaded: ath]
 CPU: 3 PID: 29 Comm: kworker/u8:1 Tainted: G           O   3.17.0-rc2-wl-ath+ #447
 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
 Workqueue: ath10k_wq ath10k_core_register_work [ath10k_core]
 task: ffff88001eb01ad0 ti: ffff88001eb60000 task.ti: ffff88001eb60000
 RIP: 0010:[<ffffffffa0058005>]  [<ffffffffa0058005>] ath10k_debug_get_new_fw_crash_data+0x15/0x30 [ath10k_core]
 RSP: 0018:ffff88001eb63ce8  EFLAGS: 00010246
 RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
 RDX: 0000000000000000 RSI: ffffc90001a09030 RDI: 0000000000000001
 RBP: ffff88001eb63cf0 R08: 0000000000000000 R09: ffff8800000bb200
 R10: 00000000000001e2 R11: ffff88001eb638de R12: ffff88001d7459a0
 R13: ffff88001d746ab0 R14: 00000000fffe14d4 R15: ffff88001d747c60
 FS:  0000000000000000(0000) GS:ffff88001fd80000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
 CR2: 0000000000000000 CR3: 000000001df34000 CR4: 00000000000006e0
 Stack:
  ffff88001d7459a0 ffff88001eb63d58 ffffffffa0083bbe ffff880000000010
  ffff88001eb63d68 ffff88001eb63d18 0000000000000002 0000000000059010
  ffffffffa0086fef 00000000deadbeef ffff88001d747a28 ffff88001d7459a0
 Call Trace:
  [<ffffffffa0083bbe>] ath10k_pci_fw_crashed_dump+0x2e/0xd0 [ath10k_pci]
  [<ffffffffa0085410>] __ath10k_pci_hif_power_up+0x5f0/0x700 [ath10k_pci]
  [<ffffffffa0085550>] ath10k_pci_hif_power_up+0x30/0xe0 [ath10k_pci]
  [<ffffffffa005bc7b>] ath10k_core_register_work+0x2b/0x520 [ath10k_core]
  [<ffffffff810689cc>] process_one_work+0x18c/0x3f0
  [<ffffffff81069011>] worker_thread+0x121/0x4a0
  [<ffffffff81068ef0>] ? rescuer_thread+0x2c0/0x2c0
  [<ffffffff8106daf2>] kthread+0xd2/0xf0
  [<ffffffff8106da20>] ? kthread_create_on_node+0x170/0x170
  [<ffffffff81857cfc>] ret_from_fork+0x7c/0xb0
  [<ffffffff8106da20>] ? kthread_create_on_node+0x170/0x170
 Code: 8b 40 38 48 c7 80 00 01 00 00 00 00 00 00 5b 5d c3 0f 1f 44 00 00 0f 1f 44 00 00 55 48 89 e5 53 48 8b 9f 90 1d 00 00 48 8d 7b 01 <c6> 03 01 e8 e3 ec 2b e1 48 8d 7b 18 e8 6a 4f 05 e1 48 89 d8 5b
 RIP  [<ffffffffa0058005>] ath10k_debug_get_new_fw_crash_data+0x15/0x30 [ath10k_core]
  RSP <ffff88001eb63ce8>
 CR2: 0000000000000000
 ---[ end trace 5d0ed15b050bcc1f ]---
 Kernel panic - not syncing: Fatal exception in interrupt
 Kernel Offset: 0x0 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffff9fffffff)
 ---[ end Kernel panic - not syncing: Fatal exception in interrupt

To prevent that split debug functions and allocate
fw_crash_data earlier.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Nothing important was being overwritten so it
didn't yield any bugs yet.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

This makes it easier to log and debug via tracing
with more than 1 ath10k device on a system.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Remove the ugly _access functions. Being explicit
is a good thing.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

This prevents some fw revisions from crashing in
many cases when user is trying to run a
promiscuous station interface (e.g. sniffing,
4addr bridge).
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Fix some cases where monitor start failure left
the driver in a confused state.

This also makes the monitor code simpler.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Commit 5c771e74
introduced a regression. On some systems spurious
interrupts could schedule a tasklet while tearing
down leading to, e.g.:

 BUG: unable to handle kernel paging request at fe589030
 IP: [<c1316fb0>] ioread32+0x30/0x40
 ...
 Call Trace:
  [<fe576c1b>] ath10k_pci_tasklet+0x1b/0x60 [ath10k_pci]
  [<c1053fbe>] tasklet_action+0x9e/0xb0
  [<c10534f1>] __do_softirq+0xf1/0x3f0
  [<c1053400>] ? ftrace_raw_event_irq_handler_entry+0xa0/0xa0
  [<c1004999>] do_softirq_own_stack+0x29/0x40
  <IRQ>
  [<c1053a76>] irq_exit+0x86/0xb0
 ...
  [<c132d522>] do_pci_disable_device+0x52/0x60
  [<c132d57f>] pci_disable_device+0x4f/0xb0
  [<c132a961>] ? __pci_set_master+0x51/0x80
  [<fe5740b3>] ath10k_pci_release+0x33/0x40 [ath10k_pci]
  [<fe575d4b>] ath10k_pci_remove+0x7b/0x90 [ath10k_pci]
Reported-by: NKalle Valo <kvalo@qca.qualcomm.com>
Tested-by: NKalle Valo <kvalo@qca.qualcomm.com>
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

If a station was reassociated, i.e. due to change
of supported rates update via sta_rc_update() the
num_legacy_stations would be (incorrectly) bumped
up leading to unbalanced usage of the var. This in
turn could lock rtscts protection up as enabled.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Recent changes done to start/restart sequences
broke hw recovery in some hw configurations. The
pci transport was stopped twice however due to a
workaround in the pci disabling code the
disable/enable for first msi interrupt was not
balanced. This ended up with irqs not being
properly re-enabled and the following print out
during recovery:

 ath10k: failed to receive control response completion, polling..
 ath10k: Service connect timeout: -110
 ath10k: Could not init core: -110

Legacy interrupt mode was unaffected while msi
ranged mode would be partially crippled (it would
miss fw indication interrupts but otherwise it
worked fine).

This fixes completely broken fw recovery for a
single msi interrupt mode and fixes subsequent fw
crash reports for msi range interrupt mode.
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

According to RCU_INIT_POINTER()'s block comment 3.a, it can be used if
"3.   The referenced data structure has already been exposed to readers either
at compile time or via rcu_assign_pointer() -and-
 a.   You have not made -any- reader-visible changes to this structure since
then".

This case fulfills the conditions above because between the rcu_dereference()
call (cvif = rcu_dereference(ar->beacon_iter);) and the rcu_assign_pointer()
call there is no update of the "cvif" variable.
Therefore, this patch makes the replacement.

The following Coccinelle semantic patch was used:
@@
identifier v;
@@

v = rcu_dereference(...);
... when != rcu_dereference(...);
    when != v = ...;
    when != (<+...v...+>)++;
    when != \(memcpy\|memset\)(...);
(
- rcu_assign_pointer
+ RCU_INIT_POINTER
  (..., v);
|
 if(...) {
... when != v = ...;
- rcu_assign_pointer
+ RCU_INIT_POINTER
  (..., v);
... when any
 }
)

Because there are cases where between a “rcu_dereference()” call and a
“rcu_assign_pointer()” call might be updates of the value that interests us,
the Coccinelle semantic patch ignores them and replaces with
"RCU_INIT_POINTER()" only when the update is not happening.
Signed-off-by: NAndreea-Cristina Bernat <bernat.ada@gmail.com>
Acked-by: NChristian Lamparter <chunkeey@googlemail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Setup the channel context correctly. Not doing this was
causing beacon loss in a P2P-GO/STA concurrent setup.
Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NCanek Peláez Valdés <canek@ciencias.unam.mx>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

If a new context is being added in addition to the current one,
then send the ASSIGN event to abort a running scan since
the addition of a context is usually followed by VIF
assignment and further operations.
Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

There is no reason why managed/p2p interfaces have to
be limited to one. IBSS is the only type that needs
a restriction.
Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This will be useful in handling addition/change of new
channel contexts.
Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>