Re: [PATCH 0/8] btrfs: assemble a multi-device root filesystem without an initrd
From: Christian Brauner
Date: Fri Nov 20 2026 - 07:13:44 EST
On Mon, Nov 16, 2026 at 11:43:16PM +0100, Ismael Luceno wrote:
> This series lets btrfs assemble such a root filesystem in the kernel:
>
> * mount devtmpfs on the initial root before the root filesystem, so
> device nodes exist to open the other devices by name (init);
> * let a filesystem tell whether it is mounting the root device and read
> rootwait, and resolve root=UUID=/root=LABEL= values that name the
> filesystem rather than a partition (init/VFS);
Nope, sorry. The init and VFS parts of this are not acceptable. Scanning
for devices and resolving filesystem UUIDs and labels is exactly what an
initramfs is for. That's userspace policy and it doesn't belong in the
kernel.
The kernel can already find its root without an initrd: root=PARTUUID=
and root=PARTLABEL=. If you need a multi-device root, use an initramfs.
A ->find_by_spec() method in struct file_system_type that only exists to
parse root= isn't going to happen either. And with BTRFS_ROOT_SCAN
defaulting to y and selecting DEVTMPFS, 2/8 changes early boot for
everyone who has btrfs enabled.
So I agree with Qu.