Re: [PATCH] ath6kl: TCMD responds through testmode events

[Kalle Valo <kvalo@qca.qualcomm.com>]

On 12/20/2011 01:02 AM, Thomas Pedersen wrote:
> ath6kl now no longer knows what is is sending and / or receiving through
> cfg80211_testmode, and simply passes opaque buffers between userspace
> and the firmware.

Patch title could be a bit more descriptive, like: "send TCMD events
through testmode events"

> @@ -30,7 +31,6 @@ enum ath6kl_tm_attr {
>  
>  enum ath6kl_tm_cmd {
>  	ATH6KL_TM_CMD_TCMD		= 0,
> -	ATH6KL_TM_CMD_RX_REPORT		= 1,
>  };

Please leave the enum but add a comment saying that it's not used anymore.

> +	skb = cfg80211_testmode_alloc_event_skb(ar->wiphy, buf_len, GFP_ATOMIC);

Why atomic?

> +	if (!skb) {
> +		ath6kl_dbg(ATH6KL_DBG_WMI,
> +			   "failed to allocate testmode rx skb!\n");

Please use ath6kl_warn()

> +	NLA_PUT_U32(skb, ATH6KL_TM_ATTR_CMD, ATH6KL_TM_CMD_TCMD);
> +	NLA_PUT(skb, ATH6KL_TM_ATTR_DATA, buf_len, buf);
> +	cfg80211_testmode_event(skb, GFP_ATOMIC);

Why atomic?

>  nla_put_failure:
> -	ret = -ENOBUFS;
> -	goto out;
> +	kfree_skb(skb);
> +	ath6kl_dbg(ATH6KL_DBG_WMI, "nla_put failed on testmode rx skb!\n");

ath6kl_warn()

Kalle