From: Demi Marie Obenour <demiobenour@gmail.com>
To: Yureka <yuka@yuka.dev>, Alyssa Ross <hi@alyssa.is>
Cc: devel@spectrum-os.org
Subject: Re: [PATCH v1 RFC 1/4] vm/sys/net: remove connman & dbus
Date: Tue, 25 Nov 2025 20:25:14 -0500 [thread overview]
Message-ID: <132e6cc0-25ce-4d81-8087-895237a7e299@gmail.com> (raw)
In-Reply-To: <2cf02c2f-04f9-4983-9017-80a700423bf3@yuka.dev>
[-- Attachment #1.1.1: Type: text/plain, Size: 5718 bytes --]
On 11/25/25 06:37, Yureka wrote:
> On 11/25/25 11:15, Alyssa Ross wrote:
>> Yureka Lilian <yureka@cyberchaos.dev> writes:
>>
>>> In preparation to integrating xdp-forwarder, making the net-vm a net-driver VM.
>>>
>>> Signed-off-by: Yureka Lilian <yureka@cyberchaos.dev>
>>> ---
>>> vm/sys/net/Makefile | 2 +-
>>> vm/sys/net/default.nix | 8 +++-----
>>> vm/sys/net/file-list.mk | 13 +------------
>>> vm/sys/net/image/etc/dbus-1/system.conf | 8 --------
>>> .../etc/s6-rc/connman/dependencies.d/dbus | 0
>>> vm/sys/net/image/etc/s6-rc/connman/run | 19 -------------------
>>> vm/sys/net/image/etc/s6-rc/connman/type | 1 -
>>> .../net/image/etc/s6-rc/connman/type.license | 2 --
>>> .../net/image/etc/s6-rc/dbus/notification-fd | 1 -
>>> .../etc/s6-rc/dbus/notification-fd.license | 2 --
>>> vm/sys/net/image/etc/s6-rc/dbus/run | 10 ----------
>>> vm/sys/net/image/etc/s6-rc/dbus/type | 1 -
>>> vm/sys/net/image/etc/s6-rc/dbus/type.license | 2 --
>>> .../image/etc/s6-rc/ok-all/contents.d/sysctl | 0
>>> vm/sys/net/image/etc/s6-rc/sysctl/type | 1 -
>>> .../net/image/etc/s6-rc/sysctl/type.license | 2 --
>>> vm/sys/net/image/etc/s6-rc/sysctl/up | 4 ----
>>> vm/sys/net/image/etc/sysctl.conf | 4 ----
>>> 18 files changed, 5 insertions(+), 75 deletions(-)
>>> delete mode 100644 vm/sys/net/image/etc/dbus-1/system.conf
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/connman/dependencies.d/dbus
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/connman/run
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/connman/type
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/connman/type.license
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/dbus/notification-fd
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/dbus/notification-fd.license
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/dbus/run
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/dbus/type
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/dbus/type.license
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/ok-all/contents.d/sysctl
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/sysctl/type
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/sysctl/type.license
>>> delete mode 100644 vm/sys/net/image/etc/s6-rc/sysctl/up
>>> delete mode 100644 vm/sys/net/image/etc/sysctl.conf
>> Won't we still need connman or NetworkManager or something to configure
>> Wi-Fi?
>
> Thank you for this excellent question.
Me too!
I have some comments below, but I will be the first to state that
none of them should block merging this patch. The current situation
is definitely unsatisfactory.
> connman and NetworkManager are all-in-one tools for networking. They
> mostly provide a unified interface for the underlying stacks
> (wpa_supplicant/iwd, kernel networking stack, firewall, VPNs). I don't
> expect we can re-use such an all-in-one tool in the driver VM, as the
> driver VM is only responsible for one device and shuffling data from and
> to it. In the Wi-Fi case there is an exception because the device needs
> special configuration, with user input. In the future, I see
> wpa_supplicant as a candidate which can maintain Wi-Fi connections in a
> stateful configuration file, providing both a cli and a GUI which we
> could forward to the user.
NetworkManager has the advantage that good GUI and CLI tools for it
already exist. That said, if we are going to use a daemon directly,
I strongly recommend going with iwd over wpa_supplicant. It has much
better code quality and can handle network configuration itself.
> For more complex networking configuration, I would expect them to be
> done via the VM "graph", where an advanced user could build a chain of
> provider VMs which do 1:n multiplexing (the router I'm currently
> writing), n:1 multiplexing (something like Android's automatic switching
> between Wi-Fi, cellular uplink and cabled connections depending on
> network conditions), and VPN/Tor as a 1:1 provider performing some
> encapsulation.
I don't think that n:1 multiplexing is an advanced configuration.
My home laptop and probably many others has Wi-Fi and Ethernet, and
others have Wi-Fi and USB/Thunderbolt docking stations. Yet other
devices will have cellular data too. It's a mess, and it should work
out of the box.
> In this current patch series the spectrum router takes over the 1:n
> multiplexing that connman was previously used for. For the n:1
> multiplexing needed for network devices with multiple interfaces, I
> expect to just use the next best one which receives router
> advertisements (which should be similar to the previous user experience
> with connman), but in this current patch series the n:1 multiplexing is
> not implemented.
NetworkManager might have more complex policies. If so, it might be
best to reuse it. I don't think this is an area where it is worth
diverging too much from most Linux distros.
> From what I can tell Wi-Fi was not a previously working feature, and
> supporting it in the future via wpa_supplicant would not be much more
> difficult, as we have tested the router in principle supports
> multiplexing multiple apps to one Wi-Fi connection.
Getting Wi-Fi working should be quite simple, but I expect that
choosing between different networks won't be, at least for all of the
types of networks in the wild. I would love to be prove wrong on this.
--
Sincerely,
Demi Marie Obenour (she/her/hers)
[-- Attachment #1.1.2: OpenPGP public key --]
[-- Type: application/pgp-keys, Size: 7253 bytes --]
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
next prev parent reply other threads:[~2025-11-26 1:25 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-24 16:35 [PATCH v1 RFC 0/4] spectrum-router Yureka Lilian
2025-11-24 16:35 ` [PATCH v1 RFC 1/4] vm/sys/net: remove connman & dbus Yureka Lilian
2025-11-25 10:15 ` Alyssa Ross
2025-11-25 11:37 ` Yureka
2025-11-26 1:25 ` Demi Marie Obenour [this message]
2025-11-24 16:35 ` [PATCH v1 RFC 2/4] vm/sys/net: integrate xdp-forwarder Yureka Lilian
2025-11-26 10:53 ` Alyssa Ross
2025-11-24 16:35 ` [PATCH v1 RFC 3/4] tools: add spectrum-router Yureka Lilian
2025-11-24 19:05 ` Demi Marie Obenour
2025-11-24 16:35 ` [PATCH v1 RFC 4/4] host: integrate router Yureka Lilian
2025-11-24 19:10 ` Demi Marie Obenour
2025-11-24 20:06 ` Yureka
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=132e6cc0-25ce-4d81-8087-895237a7e299@gmail.com \
--to=demiobenour@gmail.com \
--cc=devel@spectrum-os.org \
--cc=hi@alyssa.is \
--cc=yuka@yuka.dev \
/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.
Code repositories for project(s) associated with this public inbox
https://spectrum-os.org/git/crosvm
https://spectrum-os.org/git/doc
https://spectrum-os.org/git/mktuntap
https://spectrum-os.org/git/nixpkgs
https://spectrum-os.org/git/spectrum
https://spectrum-os.org/git/ucspi-vsock
https://spectrum-os.org/git/www
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).