[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <52e7f9d3-1c4b-b37c-8a02-0bab3a0c8229@kernel.org>
Date: Sat, 11 Jun 2022 10:55:37 -0600
From: David Ahern <dsahern@...nel.org>
To: Jakub Kicinski <kuba@...nel.org>,
Richard Gobert <richardbgobert@...il.com>
Cc: davem@...emloft.net, yoshfuji@...ux-ipv6.org, pabeni@...hat.com,
netdev@...r.kernel.org
Subject: Re: [PATCH] net: error on unsupported IP setsockopts in IPv6
On 6/10/22 11:16 PM, Jakub Kicinski wrote:
> On Thu, 9 Jun 2022 17:26:10 +0200 Richard Gobert wrote:
>> The IP_TTL and IP_TOS sockopts are unsupported for IPv6 sockets,
>> this bug was reported previously to the kernel bugzilla [1].
>> Make the IP_TTL and IP_TOS sockopts return an error for AF_INET6 sockets.
>>
>> [1] https://bugzilla.kernel.org/show_bug.cgi?id=212585
>
> This is a little risky because applications may set both v4 and v6
> options and expect the correct one to "stick". Obviously it's not
> the way we would have written this code today, but is there any harm?
> Also why just those two options?
agreed, I do not believe this can be changed. The values will not be
used for ipv6 packets, so the exposure should just be a matter of
allowing the setting on a socket.
Powered by blists - more mailing lists