public inbox for pve-devel@lists.proxmox.com
 help / color / mirror / Atom feed
From: David Riley <d.riley@proxmox.com>
To: pve-devel@lists.proxmox.com
Subject: [PATCH access-control/manager/network v4 0/6] fix #7294: pool: add SDN VNets as pool members
Date: Wed, 18 Nov 2026 12:57:51 +0100	[thread overview]
Message-ID: <20261118115757.94919-1-d.riley@proxmox.com> (raw)

Maintainer Note:

This patch requires version bumps:
* pve-access-control
* pve-manager
* pve-network

Please also ensure that the correct version is set in the pve-manager
api patch where 'assert_min_cluster_version' (Pool.pm) is called.

This series does not depend on '[PATCH access-control/network v6 0/8]
fix #7520: sdn: prune orphaned ACLs and handle VNet migrations' [1]
anymore as this has been applied recently.

---

Differences from v3:
- New: Patch 1: rpcenv: add SDN.Allocate to modify permission check
- Fix: Properly enforce `SDN.Allocate` permissions when modifying
  pools (Patch 5/6)
- Fix: Relocated the VLAN-awareness check. Previously, if a VNet had
  its VLAN-awareness disabled while it was a member of a pool, users
  were prevented from removing it from that pool.
  (Patch 5/6)
- Fix: Change default Vlan Tag dropdown text from 'All' to 'untagged'
  to reflect the actual behavior.

Differences from v2:
- Add/Improve: Instead of allowing full vnet access (tagged and
  untagged using the base path 'sdn/zones/<zone>/<vnet>' there is now
  an explicit path for this 'sdn/zones/<zone>/<vnet>/*' this allows
  access to untagged and tagged traffic of this specific vnet. In the
  WebUI this is reflected by a 'propagate' flag.
  Using just 'sdn/zones/<zone>/<vnet>' only allows for untagged use of
  this vnet, so trying to add a NIC with tag 1 will fail.
  Thanks @Daniel K. for the great suggestion!
- Refactor: migrate vnet zone in pool. Passing in a array of path
  strings instead of each single path. This avoids locking the
  user.cfg for each path. Passing in strings because the underlying
  patch series was adapted. [1]
- Refactor: remove vnet from pool. Similar to migrations this now
  also gets passed an array of path strings and only locks the
  configuration once.
- Reorder: patches to improve the review flow
- Remove: #7294 from unrelated commits
- Fix: various nits
- Add: maintainer note about bumping the version to ensure the
  gatekeeper for this feature works properly


Differences from v1:
- Access: Fix permissions propagation.
  Pool ACL paths are setup without propagation, therefore checking
  /sdn/zones/<zone>/<vnet>/<tag> fails even if the user has the
  permission for the base path /sdn/zones/<zone>/<vnet>.
  To allow this, the roles of the base VNet path are looked up if
  the exact tagged path is not found in the pool (see patch 5/9).
- API: Add a unified property string format for VNets 
  (zone=<zone>,vnet=<vnet>,tag=<tag>]), ensuring that zone and vnet
  are strictly required and coupled.
- API CLI: Add typetext to the vnet format for better error messages
  in the CLI
- API: The registered format validation for VNets and Zones now also
  check the length.
- API: Add membership checks for network resources during add/delete
  operations (matching storage/VM behavior).
- Relocated version helpers from PVE::Cluster to PVE::Tools
- UI: Fix light mode. The icons are now light grey to match the
  storage and vm icons.
- Series is now based on v4 instead of v3 of [1]
- Minor refactors

=== Original Cover Letter ===

This series implements support for adding SDN VNets to resource pools,
resolving #7294 [0]. This series depends on the v4 'fix #7520: sdn:
prune orphaned ACLs and handle VNet migrations' [1].

It does not, however, add zones as pool members as requested in #7294.
Zones currently share ACL paths for managing the zone itself and
allocating VNets within it. This makes self-service VNet management
without also granting zone management (and its associated 
side-effects) difficult.

This patch series extends the pool section in the user.cfg and
introduces a new network property to the pool configuration which will
hold VNet entries:
* vnet/<zone>/<vnet>
* vnet/<zone>/<vnet>/<vlan>

The type prefix allows future extension to other network resource 
types.

To prevent potential data loss from overwriting newly added VNets, a
cluster-version check is added which ensures all nodes are running a
version that supports this feature. Note: The hardcoded version guard
should be updated to match the final target release when being
applied.

The existing version check helpers were moved from `qemu-server` to a
new module within `pve-cluster` to make them available for this
implementation, and any future developments that require gatekeeping.
Appropriate attribution has been included for the relocated code.
Please let me know if this organizational move aligns with current
design preferences or additional adjustments are needed.

[0] https://bugzilla.proxmox.com/show_bug.cgi?id=7294
[1] https://lore.proxmox.com/pve-devel/20260716104247.29047-1-d.riley@proxmox.com/


pve-access-control:

David Riley (1):
  rpcenv: add SDN.Allocate to modify permission check

 src/PVE/RPCEnvironment.pm | 2 ++
 1 file changed, 2 insertions(+)


pve-network:

David Riley (2):
  sdn: register api formats for zones and vnets
  fix #7294: sdn: vnet: update pool members on vnet migration and
    deletion

 src/PVE/Network/SDN.pm              |  6 ++++++
 src/PVE/Network/SDN/VnetPlugin.pm   | 26 +++++++++++++++++++++++---
 src/PVE/Network/SDN/Zones/Plugin.pm | 26 +++++++++++++++++++++++---
 3 files changed, 52 insertions(+), 6 deletions(-)


pve-manager:

David Riley (3):
  ui: replace var with let to match style guide for variable declaration
  fix #7294: api: pool: add SDN VNets as pool members
  fix #7294: ui: pool: add SDN VNets as pool members

 PVE/API2/Pool.pm                 | 151 ++++++++++++++++++++++++++-
 www/css/ext6-pve.css             |  15 +++
 www/manager6/Utils.js            |   1 +
 www/manager6/grid/PoolMembers.js | 173 ++++++++++++++++++++++++++++---
 4 files changed, 321 insertions(+), 19 deletions(-)


Summary over all repositories:
  8 files changed, 375 insertions(+), 25 deletions(-)

-- 
Generated by murpp 0.11.0




             reply	other threads:[~2026-11-18 11:58 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-18 11:57 David Riley [this message]
2026-11-18 11:57 ` [PATCH pve-access-control v4 1/6] rpcenv: add SDN.Allocate to modify permission check David Riley
2026-11-18 11:57 ` [PATCH pve-network v4 2/6] sdn: register api formats for zones and vnets David Riley
2026-11-18 11:57 ` [PATCH pve-network v4 3/6] fix #7294: sdn: vnet: update pool members on vnet migration and deletion David Riley
2026-11-18 11:57 ` [PATCH pve-manager v4 4/6] ui: replace var with let to match style guide for variable declaration David Riley
2026-11-18 11:57 ` [PATCH pve-manager v4 5/6] fix #7294: api: pool: add SDN VNets as pool members David Riley
2026-11-18 11:57 ` [PATCH pve-manager v4 6/6] fix #7294: ui: " David Riley
2026-11-19 12:49 ` superseded: [PATCH access-control/manager/network v4 0/6] fix #7294: " David Riley

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=20261118115757.94919-1-d.riley@proxmox.com \
    --to=d.riley@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