[PATCH] ath10k: fix wmi mgmt tx queue full due to race condition

[Brian Norris]

Hi,

On Sun, Dec 20, 2020 at 5:53 PM Miaoqing Pan <miaoqing at codeaurora.org> wrote:
>
> Failed to transmit wmi management frames:
>
> [84977.840894] ath10k_snoc a000000.wifi: wmi mgmt tx queue is full
> [84977.840913] ath10k_snoc a000000.wifi: failed to transmit packet, dropping: -28
> [84977.840924] ath10k_snoc a000000.wifi: failed to submit frame: -28
> [84977.840932] ath10k_snoc a000000.wifi: failed to transmit frame: -28
>
> This issue is caused by race condition between skb_dequeue and
> __skb_queue_tail. The queue of ‘wmi_mgmt_tx_queue’ is protected by a
> different lock: ar->data_lock vs list->lock, the result is no protection.

Nice catch!

> --- a/drivers/net/wireless/ath/ath10k/mac.c
> +++ b/drivers/net/wireless/ath/ath10k/mac.c
> @@ -3763,23 +3763,16 @@ bool ath10k_mac_tx_frm_has_freq(struct ath10k *ar)
>  static int ath10k_mac_tx_wmi_mgmt(struct ath10k *ar, struct sk_buff *skb)
>  {
>         struct sk_buff_head *q = &ar->wmi_mgmt_tx_queue;
> -       int ret = 0;
> -
> -       spin_lock_bh(&ar->data_lock);
>
>         if (skb_queue_len(q) == ATH10K_MAX_NUM_MGMT_PENDING) {

I believe you should be switching this to use skb_queue_len_lockless()
too. And this still probably leaves a TOCTOU race; maybe we should use
">=" here, in case we queue a few SKBs simultaneously? It doesn't seem
like we actually have a hard limit here, but it still seems like we
shouldn't leave this potential inconsistency.

Brian

>                 ath10k_warn(ar, "wmi mgmt tx queue is full\n");
> -               ret = -ENOSPC;
> -               goto unlock;
> +               return -ENOSPC;
>         }
>
> -       __skb_queue_tail(q, skb);
> +       skb_queue_tail(q, skb);
>         ieee80211_queue_work(ar->hw, &ar->wmi_mgmt_tx_work);