Live data from Hacker News

Enabling IPv6 support for IPv4-only apps on Linux

blog.apnic.net

21–30 of 199 posts

Re: Enabling IPv6 support for IPv4-only apps on Linux

#21

Earlier quoted context omitted.

Also Android doesn't support DHCPv6

What's the benefit of DHCPv6 over SLAAC (genuine question, never looked into DHCPv6 at all)

all of the DHCP options, for example.

SLAAC just tells the endpoint device to self-assign an address and to roll with it. For example, there’s no way to pass in DNS servers with SLAAC.

Re: Enabling IPv6 support for IPv4-only apps on Linux

#22
The main advantage of Linux over other systems is that its available from source and thus can be patched and recompiled easily. Here we see someone using LD_PRELOAD to patch the behavior of a binary, because they rely on distributions pre-compiled packages without means to change the source code themselves. This sort of negates that advantage, right?

Re: Enabling IPv6 support for IPv4-only apps on Linux

#23
post #16
post #12

ssh 64:ff9b::1.3.3.7 What address syntax is that? I mean I can easily guess what it means, but I wasn't aware such mixture exists. Do connect(2) and friends support that?

It's an RFC6052 address.

Thanks for the pointer. Actually RFC6052 just refers to RFC4291 from 2006 in this case. And that simply calls it an alternative representation.

Edit: RFC1884 from 1995 already had it, nothing has changed in that aspect.

Re: Enabling IPv6 support for IPv4-only apps on Linux

#24

Earlier quoted context omitted.

What's the benefit of DHCPv6 over SLAAC (genuine question, never looked into DHCPv6 at all)

all of the DHCP options, for example. SLAAC just tells the endpoint device to self-assign an address and to roll with it. For example, there’s no way to pass in DNS servers with SLAAC.

Can't you provide DNS servers as part of router advertisement packets ? RFC8106

Re: Enabling IPv6 support for IPv4-only apps on Linux

#25
post #24

Earlier quoted context omitted.

all of the DHCP options, for example. SLAAC just tells the endpoint device to self-assign an address and to roll with it. For example, there’s no way to pass in DNS servers with SLAAC.

Can't you provide DNS servers as part of router advertisement packets ? RFC8106

I stand corrected, but I haven't seen this in practice yet.

Re: Enabling IPv6 support for IPv4-only apps on Linux

#26
post #24

Earlier quoted context omitted.

all of the DHCP options, for example. SLAAC just tells the endpoint device to self-assign an address and to roll with it. For example, there’s no way to pass in DNS servers with SLAAC.

Can't you provide DNS servers as part of router advertisement packets ? RFC8106

This requires client support though, which isn't a given.

Re: Enabling IPv6 support for IPv4-only apps on Linux

#27

Earlier quoted context omitted.

What's the benefit of DHCPv6 over SLAAC (genuine question, never looked into DHCPv6 at all)

all of the DHCP options, for example. SLAAC just tells the endpoint device to self-assign an address and to roll with it. For example, there’s no way to pass in DNS servers with SLAAC.

The DNS bit it's not true: there's RFC8106 [1] for this and it's widely supported. You can set a list of DNS servers to use (RDNSS) and a domain suffix for resolving unqualified hostnames (DNSSL), exactly as for DHCP.

[1]: https://datatracker.ietf.org/doc/html/rfc8106

Re: Enabling IPv6 support for IPv4-only apps on Linux

#28

The main advantage of Linux over other systems is that its available from source and thus can be patched and recompiled easily. Here we see someone using LD_PRELOAD to patch the behavior of a binary, because they rely on distributions pre-compiled packages without means to change the source code themselves. This sort of negates that advantage, right?

Having an additional way to effect changes in behavior without recompiling an application hardly negates the benefit you described. They don't say the _cant_ or _couldn't_ change the source themselves either. And what you described is hardly the main advantage of Linux, FreeBSD enables that _much_ easier.

Re: Enabling IPv6 support for IPv4-only apps on Linux

#29

The main advantage of Linux over other systems is that its available from source and thus can be patched and recompiled easily. Here we see someone using LD_PRELOAD to patch the behavior of a binary, because they rely on distributions pre-compiled packages without means to change the source code themselves. This sort of negates that advantage, right?

I don't think openssh upstream would accept a patch that makes ssh interpret ipv4 addresses as anything other than ipv4 addresses, and who wants to carry such a patch downstream forever?
Post reply on HN