Re: [PATCH 3/4] mwifiex: remove_bss_prio_lock

[Kalle Valo <kvalo@codeaurora.org>]

Avinash Patil <patila@marvell.com> writes:

> From: Zhaoyang Liu <liuzy@marvell.com>
>
> This patch does away with spinlock in
> mwifiex_wmm_get_highest_priolist_ptr in order to improve TP.
>
> Signed-off-by: Zhaoyang Liu <liuzy@marvell.com>
> Signed-off-by: Shengzhen Li <szli@marvell.com>
> Signed-off-by: Cathy Luo <cluo@marvell.com>
> Signed-off-by: Amitkumar Karwar <patila@marvell.com>
> Signed-off-by: Avinash Patil <patila@marvell.com>

[...]

> @@ -727,6 +730,25 @@ static int mwifiex_deinit_priv_params(struct mwifiex_private *priv)
>  	}
>  
>  	mwifiex_deauthenticate(priv, NULL);
> +
> +	spin_lock_irqsave(&adapter->main_proc_lock, flags);
> +	adapter->main_locked = true;
> +	if (adapter->mwifiex_processing) {
> +		spin_unlock_irqrestore(&adapter->main_proc_lock, flags);
> +		flush_workqueue(adapter->workqueue);
> +	} else {
> +		spin_unlock_irqrestore(&adapter->main_proc_lock, flags);
> +	}
> +
> +	spin_lock_irqsave(&adapter->rx_proc_lock, flags);
> +	adapter->rx_locked = true;
> +	if (adapter->rx_processing) {
> +		spin_unlock_irqrestore(&adapter->rx_proc_lock, flags);
> +		flush_workqueue(adapter->rx_workqueue);
> +	} else {
> +		spin_unlock_irqrestore(&adapter->rx_proc_lock, flags);
> +	}

And here the horrible hack just grows to other functions.

-- 
Kalle Valo