[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v4 2/4] util/qemu-sockets.c: Split host:port parsing out of i
|
From: |
Samuel Thibault |
|
Subject: |
Re: [PATCH v4 2/4] util/qemu-sockets.c: Split host:port parsing out of inet_parse |
|
Date: |
Mon, 1 Mar 2021 17:26:23 +0100 |
|
User-agent: |
NeoMutt/20170609 (1.8.3) |
Doug Evans, le lun. 01 mars 2021 08:07:19 -0800, a ecrit:
> Are there any users that this functional change would break?
> [Previously the empty address meant qemu would only listen on ipv4 addr-any.]
One case that could be broken would be a user having already another
service listening on ipv6-only along qemu listening on ipv4-only. But I
find this very little probable.
> What if a user wants only ipv4 addr-any (or only ipv6 addr-any) ?
"0.0.0.0" would get ipv4 addr-any.
Without anything done in particular, "::" would get both ipv6 and
ipv4. We could make libslirp enable the IPV6ONLY flag to avoid that, and
make qemu pass an AF_UNSPEC address for the ipv4+ipv6 case, in which
case libslirp wouldn't set IPV6ONLY.
make that ipv6-only through a flag passed to
> What does hostfwd "::12345-6.7.8.9:10" mean?
> Does the presence of the empty host address mean forward both ipv4 and ipv6 to
> guest ipv4 6.7.8.9?
I'd say so, yes.
Samuel