netdev - Re: [RFC net-next 04/10] devlink: always check if the devlink instance is registered

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: <234f11c3-b83b-3c7b-2ee6-8e90a761cae1@intel.com>
Date:   Mon, 19 Dec 2022 14:08:11 -0800
From:   Jacob Keller <jacob.e.keller@...el.com>
To:     Jakub Kicinski <kuba@...nel.org>
CC:     <jiri@...nulli.us>, <leon@...nel.org>, <netdev@...r.kernel.org>
Subject: Re: [RFC net-next 04/10] devlink: always check if the devlink
 instance is registered



On 12/19/2022 1:55 PM, Jakub Kicinski wrote:
> On Mon, 19 Dec 2022 09:48:54 -0800 Jacob Keller wrote:
>> On 12/16/2022 5:19 PM, Jakub Kicinski wrote:
>>> Always check under the instance lock whether the devlink instance
>>> is still / already registered.
>>
>> Ok. So now the reference ensures less about whats valid. It guarantees a
>> lock but doesn't ensure that the devlink remains registered unless you
>> acquire the lock and check that the devlink is alive under lock now?
> 
> Correct.
> 
>>> This is a no-op for the most part, as the unregistration path currently
>>> waits for all references. On the init path, however, we may temporarily
>>> open up a race with netdev code, if netdevs are registered before the
>>> devlink instance. This is temporary, the next change fixes it, and this
>>> commit has been split out for the ease of review.
>>>   
>>
>> This means you're adding the problem here, but its fixed in next commit..?
> 
> Yes, I can squash when posting for applying, but TBH I think the clarity
> of the changes outweighs the tiny and transient race.

I would agree. The only reason I could think it to be a problem is if a
bisect lands precisely on this commit and you happen to hit this... what
are the side effects of the race here? If the side effects don't include
significant issues I think its fine.