netdev - Re: [PATCH net] ip6_tunnel: fix GRE6 segmentation

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: <18a17215-e4e4-a1ef-b2d0-a934dca2b21c@gmail.com>
Date:   Thu, 24 Jun 2021 08:36:12 -0600
From:   David Ahern <dsahern@...il.com>
To:     Guillaume Nault <gnault@...hat.com>
Cc:     Jakub Kicinski <kuba@...nel.org>, davem@...emloft.net,
        netdev@...r.kernel.org, yoshfuji@...ux-ipv6.org,
        dsahern@...nel.org, vfedorenko@...ek.ru
Subject: Re: [PATCH net] ip6_tunnel: fix GRE6 segmentation

On 6/24/21 7:39 AM, Guillaume Nault wrote:
> On Wed, Jun 23, 2021 at 12:28:05PM -0600, David Ahern wrote:
>> On 6/23/21 10:14 AM, Jakub Kicinski wrote:
>>> Noob question, why do we need that 2 sec wait with IPv6 sometimes?
>>> I've seen it randomly in my local testing as well I wasn't sure if 
>>> it's a bug or expected.
>>
>> It is to let IPv6 DAD to complete otherwise the address will not be
>> selected as a source address. That typically results in test failures.
>> There are sysctl settings that can prevent the race and hence the need
>> for the sleep.
> 
> But Jakub's script uses "nodad" in the "ip address add ..." commands.
> Isn't that supposed to disable DAD entirely for the new address?
> Why would it need an additional "sleep 2"?
> 

it should yes. I think the selftests have acquired a blend of nodad,
sysctl and sleep. I'm sure it could be cleaned up and made consistent.