Re: [PATCH 0/7] wifi: ath12k: add support for WoW

[Kalle Valo <kvalo@kernel.org>]

Baochen Qiang <quic_bqiang@quicinc.com> writes:

> On 5/21/2024 5:49 PM, Kalle Valo wrote:
>> Baochen Qiang <quic_bqiang@quicinc.com> writes:
>> 
>>> Add support for WoW (Wakeup on Wireless) funtionalities, this including
>>> magic-pattern, net-detect, disconnect and patterns, HW data filter, ARP
>>> and NS offload, GTK rekey offload.
>>>
>>> Also enable keepalive before suspend to avoid unexpected kick out by AP.
>> 
>> There was a simple conflict in ath12k_core_suspend() but I fixed that
>> (tag ath-pending-202405210818). But I see new warnings:
>> 
>> ./include/linux/fortify-string.h:97:33: warning: '__builtin_memcpy'
>> accessing 18446744073709551605 or more bytes at offsets 0 and 0
>> overlaps 9223372036854775787 bytes at offset -9223372036854775798
>> [-Wrestrict]
>> ./include/linux/fortify-string.h:97:33: warning: '__builtin_memcpy'
>> accessing 18446744073709551605 or more bytes at offsets 0 and 0
>> overlaps 9223372036854775787 bytes at offset -9223372036854775798
>> [-Wrestrict]
>
> I suppose this is run-time warning, right? I don't see them while
> testing, am I missing something? any configuration or special GCC
> version?

Yes, this is a warning from GCC. I use currently GCC 13.2 from
crosstool:

https://wireless.wiki.kernel.org/en/users/drivers/ath10k/codingstyle#tools

The kernel .config I use you should get from our private repository
where I share my NUC configs. But I guess at least this config is
important:

CONFIG_FORTIFY_SOURCE=y

But you might need other configs enabled as well, not sure.

>> drivers/net/wireless/ath/ath12k/core.c:69:5: warning: context
>> imbalance in 'ath12k_core_suspend' - different lock contexts for
>> basic block
>
> This is because the conflict you mentioned above is not fixed in a
> right way: rcu_read_lock() should not be added there in
> ath12k_core_suspend() in commit
> https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git/commit/drivers?h=ath-pending-202405210818&id=2f77ceb7ece056f20d98519f0b1b696711c88908

Ah, my bad. That's what I get when fixing conflicts hastily. Thanks for
finding my error :)

>> And in my suspend tests with WCN7850 the kernel crashes (tried two times
>> and crashed in both):
>> 
>> [  854.087911] wlan0: associated
>> [ 854.139498] wlan0: Limiting TX power to 23 (23 - 0) dBm as
>> advertised by <censored>
>> [ 854.481975] BUG: sleeping function called from invalid context at
>> include/linux/sched/mm.h:315
>> [ 854.483069] in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid:
>> 1694, name: wpa_supplicant
>> [  854.483524] preempt_count: 0, expected: 0
>> [  854.484010] RCU nest depth: 1, expected: 0
>> [  854.484416] 2 locks held by wpa_supplicant/1694:
>> [ 854.484914] #0: ffffffff9c9e2650 (rtnl_mutex){+.+.}-{3:3}, at:
>> rtnetlink_rcv_msg+0x2c8/0xbf0
>> [ 854.485512] #1: ffffffff9c0cd960 (rcu_read_lock){....}-{1:2}, at:
>> atomic_notifier_call_chain+0x28/0x140
>> [  854.486290] CPU: 1 PID: 1694 Comm: wpa_supplicant Not tainted 6.9.0-wt-ath+ #1525
>> [ 854.486707] Hardware name: Intel(R) Client Systems
>> NUC8i7HVK/NUC8i7HVB, BIOS HNKBLi70.86A.0067.2021.0528.1339
>> 05/28/2021
>> [  854.487195] Call Trace:
>> [  854.487467]  <TASK>
>> [  854.487734]  dump_stack_lvl+0xa4/0xe0
>> [  854.488049]  dump_stack+0x10/0x20
>> [  854.488335]  __might_resched+0x363/0x5a0
>> [  854.488628]  ? kmalloc_trace+0x4b/0x2d0
>> [  854.488978]  __might_sleep+0xad/0x160
>> [  854.489282]  kmalloc_trace+0x264/0x2d0
>> [  854.489574]  ? ath12k_mac_arvif_get_arp_ns_offload+0x72/0xd0 [ath12k]
>> [  854.490071]  ath12k_mac_arvif_get_arp_ns_offload+0x72/0xd0 [ath12k]
>> [  854.490543]  ath12k_mac_op_ipv6_changed+0xc7/0x1470 [ath12k]
>> [  854.491117]  ? ieee80211_emulate_change_chanctx+0xe0/0xe0 [mac80211]
>> [  854.491909]  ieee80211_ifa6_changed+0x1c7/0x3a0 [mac80211]
>> [  854.492663]  notifier_call_chain+0x97/0x310
>> [  854.492997]  atomic_notifier_call_chain+0x5d/0x140
>> [  854.493322]  inet6addr_notifier_call_chain+0x16/0x20
>> [  854.493650]  ipv6_add_addr+0x10ec/0x2160
>> [  854.493946]  ? reacquire_held_locks+0x4d0/0x4d0
>> [  854.494294]  ? inet6_dump_ifaddr+0x10/0x10
>> [  854.494599]  ? ip6_route_add+0x75/0x130
>> [  854.494910]  ? __kasan_check_write+0x14/0x20
>> [  854.495214]  ? ip6_route_add+0x8b/0x130
>> [  854.495525]  addrconf_add_linklocal+0xd6/0x300
>> [  854.495965]  ? addrconf_add_linklocal+0xd6/0x300
>> [  854.496297]  ? addrconf_dad_start+0x110/0x110
>> [  854.499635]  ? addrconf_addr_eui48_base+0x80/0xa0
>> [  854.503094]  addrconf_addr_gen.constprop.0+0x155/0x260
>> [  854.506223]  ? mutex_unlock+0xd/0x10
>> [  854.508461]  ? addrconf_add_linklocal+0x300/0x300
>> [  854.510681]  addrconf_init_auto_addrs+0x194/0x4e0
>> [  854.512414]  addrconf_notify+0xabe/0xdd0
>> [  854.514073]  notifier_call_chain+0x97/0x310
>> [  854.515733]  ? __unwind_start+0x1f3/0x3d0
>> [  854.517070]  raw_notifier_call_chain+0x11/0x20
>> [  854.518407]  call_netdevice_notifiers_info+0x8b/0xf0
>> [  854.521761]  netdev_state_change+0xe8/0x120
>> [  854.522882]  ? call_netdevice_notifiers_info+0xf0/0xf0
>> [  854.524019]  netdev_set_operstate+0xdf/0x130
>> [  854.525139]  ? rtnl_put_cacheinfo+0x260/0x260
>> [  854.526216]  set_operstate+0x8a/0x150
>> [  854.527175]  do_setlink+0xf0e/0x2550
>> [  854.528130]  ? rtnl_bridge_getlink+0x870/0x870
>> [  854.529078]  ? check_prev_add+0x2330/0x2330
>> [  854.530011]  ? mark_lock+0xe6/0x1470
>> [  854.530919]  ? check_prev_add+0x2330/0x2330
>> [  854.531710]  ? __lock_acquire+0xd43/0x1dd0
>> [  854.532491]  ? kasan_print_address_stack_frame+0x1d2/0x230
>> [  854.533277]  ? nla_get_range_signed+0x340/0x340
>> [  854.534046]  ? __kasan_check_read+0x11/0x20
>> [  854.534808]  ? mark_lock+0xe6/0x1470
>> [  854.535573]  ? __kasan_check_read+0x11/0x20
>> [  854.536292]  ? __nla_validate_parse+0x1d0/0x400
>> [  854.536980]  rtnl_setlink+0x217/0x330
>> [  854.537650]  ? do_setlink+0x2550/0x2550
>> [  854.538355]  ? rtnetlink_rcv_msg+0x2c8/0xbf0
>> [  854.539010]  ? mutex_lock_io_nested+0x1270/0x1270
>> [  854.539657]  ? reacquire_held_locks+0x4d0/0x4d0
>> [  854.540290]  ? lock_sync+0x1a0/0x1a0
>> [  854.540981]  ? __this_cpu_preempt_check+0x13/0x20
>> [  854.541598]  rtnetlink_rcv_msg+0x315/0xbf0
>> [  854.542278]  ? check_prev_add+0x2330/0x2330
>> [  854.542894]  ? rtnl_link_fill+0x900/0x900
>> [  854.543446]  ? __lock_acquire+0xd43/0x1dd0
>> [  854.544037]  netlink_rcv_skb+0x14c/0x400
>> [  854.544581]  ? rtnl_link_fill+0x900/0x900
>> [  854.545132]  ? netlink_ack+0xd70/0xd70
>> [  854.545851]  ? debug_smp_processor_id+0x17/0x20
>> [  854.546369]  ? __lock_release.isra.0+0x49c/0xae0
>> [  854.546939]  ? netlink_deliver_tap+0x143/0x350
>> [  854.547478]  ? netlink_deliver_tap+0x143/0x350
>> [  854.548062]  ? __this_cpu_preempt_check+0x13/0x20
>> [  854.548620]  ? netlink_deliver_tap+0xc8/0x350
>> [  854.549149]  rtnetlink_rcv+0x10/0x20
>> [  854.549708]  netlink_unicast+0x45c/0x790
>> [  854.550211]  ? netlink_attachskb+0x7f0/0x7f0
>> [  854.550802]  netlink_sendmsg+0x7eb/0xdb0
>> [  854.551302]  ? netlink_unicast+0x790/0x790
>> [  854.551847]  ? unix_ioctl+0x358/0x400
>> [  854.552338]  ? selinux_socket_sendmsg+0x31/0x40
>> [  854.552873]  ? netlink_unicast+0x790/0x790
>> [  854.553360]  __sock_sendmsg+0xc9/0x160
>> [  854.553901]  ? sock_do_ioctl+0xcf/0x1e0
>> [  854.554387]  __sys_sendto+0x1ca/0x290
>> [  854.554925]  ? __ia32_sys_getpeername+0xa0/0xa0
>> [  854.555413]  ? ioctl_has_perm.constprop.0.isra.0+0x276/0x430
>> [  854.556005]  ? sock_ioctl+0x24d/0x540
>> [  854.556497]  ? br_ioctl_call+0x90/0x90
>> [  854.557047]  ? selinux_file_ioctl+0x108/0x250
>> [  854.557538]  ? debug_smp_processor_id+0x17/0x20
>> [  854.558076]  __x64_sys_sendto+0xdb/0x1a0
>> [  854.558588]  ? __this_cpu_preempt_check+0x13/0x20
>> [  854.559170]  ? lockdep_hardirqs_on+0x7d/0x100
>> [  854.559716]  x64_sys_call+0x8a2/0x9e0
>> [  854.560209]  do_syscall_64+0x65/0x130
>> [  854.560807]  entry_SYSCALL_64_after_hwframe+0x4b/0x53
>> [  854.561300] RIP: 0033:0x7f0ef17cb750
>> [ 854.561848] Code: ff eb bc 0f 1f 80 00 00 00 00 f3 0f 1e fa 41 89
>> ca 64 8b 04 25 18 00 00 00 85 c0 75 1d 45 31 c9 45 31 c0 b8 2c 00 00
>> 00 0f 05 <48> 3d 00 f0 ff ff 77 68 c3 0f 1f 80 00 00 00 00 55 48 83
>> ec 20 48
>> [ 854.562946] RSP: 002b:00007ffc4306af98 EFLAGS: 00000246 ORIG_RAX:
>> 000000000000002c
>> [  854.563506] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f0ef17cb750
>> [  854.564122] RDX: 0000000000000028 RSI: 00007ffc4306afb0 RDI: 0000000000000005
>> [  854.564738] RBP: 0000556901de3960 R08: 0000000000000000 R09: 0000000000000000
>> [  854.565298] R10: 0000000000000000 R11: 0000000000000246 R12: 00000000ffffffff
>> [  854.565945] R13: 00007ffc4306afb0 R14: 0000000000000005 R15: 0000556901e2b1f1
>> [  854.566522]  </TASK>
>
> Thanks. there is a race here: ath12k_mac_arvif_get_arp_ns_offload() is
> called by two functions and it allocate memory in the first call. So
> if it is called firstly by ath12k_mac_op_ipv6_changed(), above bug
> shows. But if it is called firstly by ath12k_mac_bss_info_changed(),
> there is no issue. This explains why I did not hit it before.
>
> Anyway, will fix it in next version.

Great, thanks for looking at it.

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches