Produce an ISO from an assembled rootfs #9

New issue

Open

opened 2026-09-27 06:05:39 -03:00 by marcos-mendez · 0 comments

marcos-mendez commented

2026-09-27 06:05:39 -03:00

(Migrated from github.com)

The fork already carries bt-iso, bt-vm, bt-img, bt-openstack, bt-xen
and bt-ec2 from upstream, and the M0 gate built and measured an ISO of
442,499,072 bytes whose squashfs reproduced bit for bit at a fixed epoch. What
is missing is the plumbing: those tasks expect the traditional fab output, and
our path produces layers assembled into a rootfs.

JedMeister/tkl-gen-iso, cloned in the notes tree, already solves hybrid BIOS
and UEFI booting with GRUB over an existing cdroot. Read it before writing
ours.

Closes when a published layer chain becomes a bootable ISO, published with its
digest and signature alongside the templates, and a test boots it. A format
without a boot test breaks in silence, so the test is part of this issue rather
than a follow up. OVA comes after, as its own issue.

The fork already carries `bt-iso`, `bt-vm`, `bt-img`, `bt-openstack`, `bt-xen` and `bt-ec2` from upstream, and the M0 gate built and measured an ISO of 442,499,072 bytes whose squashfs reproduced bit for bit at a fixed epoch. What is missing is the plumbing: those tasks expect the traditional fab output, and our path produces layers assembled into a rootfs. `JedMeister/tkl-gen-iso`, cloned in the notes tree, already solves hybrid BIOS and UEFI booting with GRUB over an existing cdroot. Read it before writing ours. Closes when a published layer chain becomes a bootable ISO, published with its digest and signature alongside the templates, and a test boots it. A format without a boot test breaks in silence, so the test is part of this issue rather than a follow up. OVA comes after, as its own issue.