linux-kernel - Re: [PATCH 0/4] wifi: wilc1000: fix RCU usage

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: <b3c31f4c-e837-47b6-bd0d-e8cf2b9964aa@bootlin.com>
Date: Mon, 19 Feb 2024 17:24:55 +0100
From: Alexis Lothoré <alexis.lothore@...tlin.com>
To: Kalle Valo <kvalo@...nel.org>
Cc: linux-wireless@...r.kernel.org, Ajay Singh <ajay.kathat@...rochip.com>,
 Claudiu Beznea <claudiu.beznea@...on.dev>,
 Thomas Petazzoni <thomas.petazzoni@...tlin.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 0/4] wifi: wilc1000: fix RCU usage

On 2/19/24 17:19, Kalle Valo wrote:
> Alexis Lothoré <alexis.lothore@...tlin.com> writes:
> 
>> This small series aims to fix multiple warnings observed when enabling
>> CONFIG_PROVE_RCU_LIST:
>> - add missing locks to create corresponding critical read sections
>> - fix mix between RCU and SRCU API usage
>>
>> While at it, since SRCU API is already in use in the driver, any fix done
>> on RCU usage was also done with the SRCU variant of RCU API. I do not
>> really get why we are using SRCU in this driver instead of classic RCU, as
>> it seems to be done in any other wireless driver.
> 
> And even more so, no other driver in drivers/net use SRCU.
> 
>> My understanding is that primary SRCU use case is for compatibility
>> with realtime kernel, which needs to be preemptible everywhere. Has
>> the driver been really developped with this constraint in mind ? If
>> you have more details about this, feel free to educate me.
> 
> Alexis, if you have the time I recommend submitting a patchset
> converting wilc1000 to use classic RCU. At least I have a hard time
> understanding why SRCU is needed, especially after seeing the warning
> you found.

If nobody else comes in with a strong argument in favor of keeping SRCU, yes I
can certainly add that to my backlog :)

-- 
Alexis Lothoré, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com