Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII;
 format=flowed
Content-Transfer-Encoding: 7bit
Date:   Wed, 29 Apr 2020 17:02:24 -0700
From:   Rajkumar Manoharan <rmanohar@codeaurora.org>
To:     Johannes Berg <johannes@sipsolutions.net>
Cc:     linux-wireless@vger.kernel.org
Subject: Re: [PATCH 02/10] cfg80211: validate 6 GHz chandef
In-Reply-To: <73a4e63e16bffb69cd9b62fd904b926dd5278fbf.camel@sipsolutions.net>
References: <1587768108-25248-1-git-send-email-rmanohar@codeaurora.org>
 <1587768108-25248-3-git-send-email-rmanohar@codeaurora.org>
 <73a4e63e16bffb69cd9b62fd904b926dd5278fbf.camel@sipsolutions.net>
Message-ID: <f7fa5618844630a3315acad53765e5e5@codeaurora.org>
User-Agent: Roundcube Webmail/1.3.9
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

On 2020-04-29 07:26, Johannes Berg wrote:
> On Fri, 2020-04-24 at 15:41 -0700, Rajkumar Manoharan wrote:
>> 
>> +static inline bool
>> +cfg80211_chandef_is_6ghz(const struct cfg80211_chan_def *chandef)
>> +{
>> +	return (chandef->center_freq1 > 5940 && chandef->center_freq1 < 
>> 7105);
>> +}
> 
> Seems like this
> 
>> +++ b/net/wireless/chan.c
>> @@ -19,6 +19,29 @@ static bool cfg80211_valid_60g_freq(u32 freq)
>>  	return freq >= 58320 && freq <= 70200;
>>  }
>> 
>> +static bool cfg80211_is_6ghz_freq(u32 freq)
>> +{
>> +	return (freq > 5940 && freq < 7105);
>> +}
> 
> should use this, by also exposing it, or something.
> 
Sure. Export this and remove the above one.

>> +static enum nl80211_chan_width cfg80211_chan_to_bw_6ghz(u8 idx)
>> +{
>> +	/* channels: 1, 5, 9, 13... */
>> +	if ((idx & 0x3) == 0x1)
>> +		return NL80211_CHAN_WIDTH_20;
>> +	/* channels 3, 11, 19... */
>> +	if ((idx & 0x7) == 0x3)
>> +		return NL80211_CHAN_WIDTH_40;
>> +	/* channels 7, 23, 39.. */
>> +	if ((idx & 0xf) == 0x7)
>> +		return NL80211_CHAN_WIDTH_80;
>> +	/* channels 15, 47, 79...*/
>> +	if ((idx & 0x1f) == 0xf)
>> +		return NL80211_CHAN_WIDTH_160;
>> +
>> +	return NL80211_CHAN_WIDTH_20;
>> +}
> 
> We haven't really done that for anything else - is that really
> necessary?
> 
Hmm.. to check whether give center_freq1 chan_idx is allowed to operate 
in given bandwidth.
Similar to center_idx_to_bw_6ghz of hostapd, this API is used to chandef 
bw.

[...]
>> @@ -213,6 +255,10 @@ bool cfg80211_chandef_valid(const struct 
>> cfg80211_chan_def *chandef)
>>  	    !cfg80211_edmg_chandef_valid(chandef))
>>  		return false;
>> 
>> +	if (cfg80211_chandef_is_6ghz(chandef) &&
>> +	    !cfg80211_6ghz_chandef_valid(chandef))
>> +		return false;
> 
> You only get there if it was in range ...
> 
> Not sure about this whole patch, it seems a bit pointless?
> 

Don't we have to check chandef bw? If not, I will drop the change.

-Rajkumar