all lists on lists.proxmox.com
 help / color / mirror / Atom feed
From: Wolfgang Bumiller <w.bumiller@proxmox.com>
To: Stoiko Ivanov <s.ivanov@proxmox.com>
Cc: pve-devel@lists.proxmox.com
Subject: Re: [pve-devel] [PATCH common 2/2] daemon: explicitly bind to wildcard address.
Date: Tue, 4 May 2021 13:28:34 +0200	[thread overview]
Message-ID: <20210504112834.by4ugiqr2ts5kc2z@olga.proxmox.com> (raw)
In-Reply-To: <20210504101222.21276-3-s.ivanov@proxmox.com>

On Tue, May 04, 2021 at 12:12:16PM +0200, Stoiko Ivanov wrote:
> with the recent change in pve-manager pveproxy (and spiceproxy)
> try binding to '::' per default. This fails for hosts having disabled
> ipv6 via kernel commandline.
> 
> Our desired behavior of binding on '::' and only falling back to
> '0.0.0.0' in case this is not supported is not directly possible with
> IO::Socket::IP->new (or rather by Socket::GetAddrInfo, which at least
> on my system always returns the v4 wildcard-address first).
> 
> the code now binds to:
> * the provided $host if not undef
> * '::' if $host is not set
> * '0.0.0.0' if $host is not set and binding on '::' yields an error
> 
> Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
> ---
>  src/PVE/Daemon.pm | 18 ++++++++++++++----
>  1 file changed, 14 insertions(+), 4 deletions(-)
> 
> diff --git a/src/PVE/Daemon.pm b/src/PVE/Daemon.pm
> index 79b90ad..f4786b8 100644
> --- a/src/PVE/Daemon.pm
> +++ b/src/PVE/Daemon.pm
> @@ -819,14 +819,24 @@ sub create_reusable_socket {
>  	$socket->fcntl(Fcntl::F_SETFD(), Fcntl::FD_CLOEXEC);
>      } else {
>  
> -	$socket = IO::Socket::IP->new(
> -	    LocalHost => $host,
> +	my %sockargs = (
>  	    LocalPort => $port,
>  	    Listen => SOMAXCONN,
>  	    Proto  => 'tcp',
>  	    GetAddrInfoFlags => 0,
> -	    ReuseAddr => 1) ||
> -	    die "unable to create socket - $@\n";
> +	    ReuseAddr => 1,
> +	);
> +	if (defined($host)) {
> +	    $socket = IO::Socket::IP->new( LocalHost => $host, %sockargs) ||
> +		die "unable to create socket - $@\n";
> +	} else {
> +	    # disabling AF_INET6 (by adding ipv6.disable=1 to the kernel cmdline)
> +	    # causes bind on :: to fail, try 0.0.0.0 in that case
> +	    $socket = IO::Socket::IP->new( LocalHost => '::', %sockargs);
> +
> +	    $socket = IO::Socket::IP->new( LocalHost => '0.0.0.0', %sockargs) if $@;

So apparently this doesn't *die* but just returns `undef` on error. But
we also should not rely on internal implementation details, so we cannot
assume that $@ will be cleared on success.
So we should probably bind the two calls together via `//` (or clear $@
before the first call?)

    $socket = IO::Socket::IP->new( LocalHost => '::', %sockargs)
        // IO::Socket::IP->new( LocalHost => '0.0.0.0', %sockargs);

> +	    die "unable to create socket - $@\n" if !$socket;
> +	}
>  
>  	# we often observe delays when using Nagle algorithm,
>  	# so we disable that to maximize performance
> -- 
> 2.20.1




  reply	other threads:[~2021-05-04 11:29 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-04 10:12 [pve-devel] [PATCH common/manager/http-server/docs] improve binding, docs and access-control for pveproxy/spiceproxy Stoiko Ivanov
2021-05-04 10:12 ` [pve-devel] [PATCH common 1/2] daemon: drop Domain parameter from create_reusable_socket Stoiko Ivanov
2021-05-04 10:12 ` [pve-devel] [PATCH common 2/2] daemon: explicitly bind to wildcard address Stoiko Ivanov
2021-05-04 11:28   ` Wolfgang Bumiller [this message]
2021-05-04 10:12 ` [pve-devel] [PATCH manager 1/1] proxy: fix wildcard address use Stoiko Ivanov
2021-05-04 10:12 ` [pve-devel] [PATCH http-server 1/2] access control: correctly match v4-mapped-v6 addresses Stoiko Ivanov
2021-05-04 10:12 ` [pve-devel] [PATCH http-server 2/2] access control: also include ipv6 in 'all' Stoiko Ivanov
2021-05-04 10:12 ` [pve-devel] [PATCH docs 1/3] pveproxy: add note about bindv6only sysctl Stoiko Ivanov
2021-05-04 10:12 ` [pve-devel] [PATCH docs 2/3] pveproxy: update documentation on 'all' alias Stoiko Ivanov
2021-05-04 10:12 ` [pve-devel] [PATCH docs 3/3] network: shortly document disabling ipv6 support Stoiko Ivanov
2021-05-04 11:25 ` [pve-devel] [PATCH common/manager/http-server/docs] improve binding, docs and access-control for pveproxy/spiceproxy Oguz Bektas
2021-05-05  5:36   ` Thomas Lamprecht
2021-05-05  9:25     ` Oguz Bektas
2021-05-04 12:20 ` Wolfgang Bumiller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210504112834.by4ugiqr2ts5kc2z@olga.proxmox.com \
    --to=w.bumiller@proxmox.com \
    --cc=pve-devel@lists.proxmox.com \
    --cc=s.ivanov@proxmox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.
Service provided by Proxmox Server Solutions GmbH | Privacy | Legal