public inbox for pve-devel@lists.proxmox.com
 help / color / mirror / Atom feed
From: "Fabian Grünbichler" <f.gruenbichler@proxmox.com>
To: Proxmox VE development discussion <pve-devel@lists.proxmox.com>
Subject: Re: [pve-devel] [PATCH container] oci create: fix creating privileged containers
Date: Wed, 26 Nov 2025 09:31:44 +0100	[thread overview]
Message-ID: <1764145853.pp0ipj03q9.astroid@yuna.none> (raw)
In-Reply-To: <20251125141922.165771-1-f.schauer@proxmox.com>

On November 25, 2025 3:19 pm, Filip Schauer wrote:
> Previously, creating privileged containers from OCI images failed with:
> `unable to create CT 123 - Invalid argument`
> 
> This was caused by an empty $id_map being passed to run_in_userns.
> 
> This commit fixes this by making the call to run_in_userns conditional,
> based on whether $id_map is empty or not.
> 
> Reported in the Proxmox forum:
> https://forum.proxmox.com/threads/proxmox-virtual-environment-9-1-available.176255/post-818600
> 
> Signed-off-by: Filip Schauer <f.schauer@proxmox.com>

or we could forbid creating them, since we want to get rid of privileged
containers mid-to-longterm anyway?

> ---
>  src/PVE/LXC/Create.pm | 47 +++++++++++++++++++++++++------------------
>  1 file changed, 27 insertions(+), 20 deletions(-)
> 
> diff --git a/src/PVE/LXC/Create.pm b/src/PVE/LXC/Create.pm
> index dc97327..9956cf9 100644
> --- a/src/PVE/LXC/Create.pm
> +++ b/src/PVE/LXC/Create.pm
> @@ -674,12 +674,17 @@ sub restore_oci_archive {
>  
>      my ($id_map, undef, undef) = PVE::LXC::parse_id_maps($conf);
>      # NOTE: values of $unsafe_oci_config are untrusted! do NOT use them as is, only via the helpers!
> -    my $unsafe_oci_config = PVE::LXC::Namespaces::run_in_userns(
> -        sub {
> -            PVE::RS::OCI::parse_and_extract_image($archive_file, $rootdir);
> -        },
> -        $id_map,
> -    );
> +    my $unsafe_oci_config;
> +    if (@$id_map) {
> +        $unsafe_oci_config = PVE::LXC::Namespaces::run_in_userns(
> +            sub {
> +                PVE::RS::OCI::parse_and_extract_image($archive_file, $rootdir);
> +            },
> +            $id_map,
> +        );
> +    } else {
> +        $unsafe_oci_config = PVE::RS::OCI::parse_and_extract_image($archive_file, $rootdir);
> +    }
>  
>      # should we rather validate this on the rust side already?
>      my $has_ctrl_char = sub { return $_[0] =~ /[\x00-\x08\x10-\x1F\x7F]/; };
> @@ -715,20 +720,22 @@ sub restore_oci_archive {
>          # This will also keep the cases working where a user does know about them and
>          # added MPs at this locations, at they will simply get mounted there correctly then.
>          # TODO: should the folders always be owned by the CT root user though?
> -        PVE::LXC::Namespaces::run_in_userns(
> -            sub {
> -                # we're now in the correct user namespace, but not in the mount namespace, so chroot
> -                # into the rootdir to ensure that make_path is safe from ../ and symlinks!
> -                chroot($rootdir) or die "failed to change root to: $rootdir: $!\n";
> -                chdir('/') or die "failed to change to root directory\n";
> -
> -                for my $path (@data_volume_paths) {
> -                    print "creating base directory for volume at $path\n";
> -                    make_path("/$path"); # chrooted to /$rootdir above already
> -                }
> -            },
> -            $id_map,
> -        );
> +        my $create_volume_paths = sub {
> +            # we're not in the correct mount namespace, so chroot into the rootdir
> +            # to ensure that make_path is safe from ../ and symlinks!
> +            chroot($rootdir) or die "failed to change root to: $rootdir: $!\n";
> +            chdir('/') or die "failed to change to root directory\n";
> +
> +            for my $path (@data_volume_paths) {
> +                print "creating base directory for volume at $path\n";
> +                make_path("/$path"); # chrooted to /$rootdir above already
> +            }
> +        };
> +        if (@$id_map) {
> +            PVE::LXC::Namespaces::run_in_userns($create_volume_paths, $id_map);
> +        } else {
> +            PVE::Tools::run_fork($create_volume_paths);
> +        }
>      }
>  
>      my $init_cmd = [];
> -- 
> 2.47.3
> 
> 
> 
> _______________________________________________
> pve-devel mailing list
> pve-devel@lists.proxmox.com
> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
> 
> 
> 


_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel


  reply	other threads:[~2025-11-26  8:32 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-25 14:19 Filip Schauer
2025-11-26  8:31 ` Fabian Grünbichler [this message]
2025-11-26  8:55   ` Thomas Lamprecht
2025-11-26  8:59     ` Thomas Lamprecht

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=1764145853.pp0ipj03q9.astroid@yuna.none \
    --to=f.gruenbichler@proxmox.com \
    --cc=pve-devel@lists.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Service provided by Proxmox Server Solutions GmbH | Privacy | Legal