Re: [PATCH] mac80211: fix possible deadlock in TX path

[Johannes Berg]

On Sat, 2019-04-27 at 22:41 +0200, Erik Stromdahl wrote:
> This patch fixes a possible deadlock when updating the TX statistics
> (when calling into ieee80211_tx_stats()) from ieee80211_tx_dequeue().
> 
> ieee80211_tx_dequeue() might be called from process context.

I think this really is the problem.

> [<c0cb1864>] (ieee80211_xmit_fast_finish) from [<c0cb4504>] (ieee80211_tx_dequeue+0x30c/0xb9c)
>  r10:d2f1a900 r9:d2d607a4 r8:d2cf20dc r7:d330b29c r6:d2cf2000 r5:d2c342ba
>  r4:d2899d3c
> [<c0cb41f8>] (ieee80211_tx_dequeue) from [<bf057f64>] (ath10k_mac_tx_push_txq+0x78/0x2a4 [ath10k_core])
>  r10:d2d607cc r9:d2fe06a0 r8:00000000 r7:d2fe1e30 r6:d2fe1d38 r5:d2fe1540
>  r4:d2cf20dc
> [<bf057eec>] (ath10k_mac_tx_push_txq [ath10k_core]) from [<bf058364>] (ath10k_mac_tx_push_pending+0x1d4/0x2e0 [ath10k_core])
>  r10:d2cf20dc r9:bf0582b4 r8:bf0b1dba r7:00000002 r6:c1429994 r5:00000000
>  r4:d2fe06a0
> [<bf058190>] (ath10k_mac_tx_push_pending [ath10k_core]) from [<bf0e25a4>] (ath10k_sdio_irq_handler+0x30c/0x4d8 [ath10k_sdio])
>  r10:00005b5a r9:d2fcc040 r8:00180201 r7:d2fe6540 r6:d2fe6a7c r5:00000000
>  r4:d2fe1540

It seems to be entirely ath10k's fault, and quite possibly our
documentation, but we probably should have local_bh_disable() there
rather than try to do u64_stats_update_begin_irqsave() in some path that
really doesn't need it.

This is going to be whack-a-mole otherwise - the TX path in mac80211
really expects to not be interrupted by softirqs.

johannes