netdev - Re: [PATCH v5 8/9] wifi: wilc1000: Register wiphy after reading out chipid

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <7a938ca9-8099-4901-9f05-c3347c38fc53@denx.de>
Date: Tue, 10 Sep 2024 12:53:02 +0200
From: Marek Vasut <marex@...x.de>
To: Alexis Lothoré <alexis.lothore@...tlin.com>,
 linux-wireless@...r.kernel.org
Cc: "David S. Miller" <davem@...emloft.net>,
 Adham Abozaeid <adham.abozaeid@...rochip.com>,
 Ajay Singh <ajay.kathat@...rochip.com>,
 Claudiu Beznea <claudiu.beznea@...on.dev>, Conor Dooley
 <conor+dt@...nel.org>, Eric Dumazet <edumazet@...gle.com>,
 Jakub Kicinski <kuba@...nel.org>, Kalle Valo <kvalo@...nel.org>,
 Krzysztof Kozlowski <krzk+dt@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
 Rob Herring <robh@...nel.org>, devicetree@...r.kernel.org,
 netdev@...r.kernel.org
Subject: Re: [PATCH v5 8/9] wifi: wilc1000: Register wiphy after reading out
 chipid

On 9/10/24 12:08 PM, Alexis Lothoré wrote:
> On 9/9/24 21:29, Marek Vasut wrote:
>> Register wiphy after reading out chipid, so the chipid can be
>> used to determine chip features and not advertise WPA3/SAE
>> support to userspace on WILC3000. Note that wilc_netdev_cleanup()
>> will deregister the wiphy in fail path.
>>
>> Signed-off-by: Marek Vasut <marex@...x.de>
> 
> [...]
> 
>> @@ -1804,14 +1803,8 @@ static struct wilc *wilc_create_wiphy(struct device *dev)
>>   				BIT(NL80211_IFTYPE_P2P_GO) |
>>   				BIT(NL80211_IFTYPE_P2P_CLIENT);
>>   	wiphy->flags |= WIPHY_FLAG_HAS_REMAIN_ON_CHANNEL;
>> -	wiphy->features |= NL80211_FEATURE_SAE;
>>   	set_wiphy_dev(wiphy, dev);
>>   	wl->wiphy = wiphy;
>> -	ret = wiphy_register(wiphy);
>> -	if (ret) {
>> -		wiphy_free(wiphy);
>> -		return NULL;
>> -	}
> 
> If I am reading the patch correctly, there are still some failure paths in
> wilc_cfg80211_init which try to call wiphy_unregister on the (not registered
> anymore in there) wphy.
>>   	return wl;
>>   }
>>   
>> @@ -1861,6 +1854,14 @@ int wilc_cfg80211_init(struct wilc **wilc, struct device *dev, int io_type,
>>   }
>>   EXPORT_SYMBOL_GPL(wilc_cfg80211_init);
>>   
>> +int wilc_cfg80211_register(struct wilc *wilc)
>> +{
>> +	wilc->wiphy->features |= NL80211_FEATURE_SAE;
> 
> Even if I get the general need, it feels weird to have parts of the wphy init
> performed in wilc_create_wiphy, and some parts (the features field) here.
> Wouldn't it work to just move wilc_create_wiphy content here, since wphy will
> not be usable anyway before eventually registering it ?
That's what I thought initially too, but look closely at 
wilc_create_wiphy():

struct wilc *wilc_create_wiphy(struct device *dev)
{
...
struct wiphy *wiphy;
struct wilc *wl;
...
wiphy = wiphy_new(&wilc_cfg80211_ops, sizeof(*wl));
...
wl = wiphy_priv(wiphy); // <----------- HERE , *wl is struct wilc
...
return wl;
}

That 'struct wilc' is allocated as part of wiphy_new() and used all 
around the place before we reach wiphy_register() much later on.