[SRU] CM4 NVMe boot fails due to old start4.elf / fixup4.dat
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
linux-firmware-raspi (Ubuntu) |
Fix Released
|
High
|
Dave Jones | ||
Jammy |
Fix Released
|
High
|
Dave Jones | ||
Kinetic |
Fix Released
|
High
|
Dave Jones | ||
Lunar |
Fix Released
|
High
|
Dave Jones | ||
linux-firmware-raspi2 (Ubuntu) |
Invalid
|
Undecided
|
Unassigned | ||
Jammy |
Invalid
|
Undecided
|
Unassigned | ||
Kinetic |
Invalid
|
Undecided
|
Unassigned | ||
Lunar |
Invalid
|
Undecided
|
Unassigned | ||
linux-raspi (Ubuntu) |
Invalid
|
Undecided
|
Unassigned | ||
Jammy |
Invalid
|
Undecided
|
Unassigned | ||
Kinetic |
Invalid
|
Undecided
|
Unassigned | ||
Lunar |
Invalid
|
Undecided
|
Unassigned |
Bug Description
[ Impact ]
CM4 users wishing to boot directly from NVMe find the initramfs is reported as corrupted. This is due to a bug in the boot firmware reported upstream as:
https:/
https:/
These issues were fixed upstream but Ubuntu now needs an updated linux-firmware-
[ Test Plan ]
* sudo add-apt-repository ppa:waveform/
* sudo apt install linux-firmware-
* sudo reboot
* Verify that the system boots successfully
* Run through the full battery of ISO tests for the relevant image. Specifically, Ubuntu Desktop arm64+raspi, Ubuntu Server arm64+raspi, and Ubuntu Server armhf+raspi under:
- Kinetic dailies: http://
- Jammy dailies: http://
[ Regression Potential ]
The updated package replaces the tertiary boot firmware. An obvious regression potential is therefore that one or more Pi models fail to boot. However, in addition we've encountered other issues which ultimately boiled down to boot firmware in the past (the boot firmware is also responsible for patching the device-tree). Hence the test plan needs to be comprehensive and should be tested on as many models of Pi as is reasonably possible.
Several kind users have already tested the back-ported firmware on the CM4 with NVMe boot (see comments 7 and 8), and I've tested on Pi 3 and 4 models under armhf and arm64 kernels, hence I'm quietly confident but I haven't run the full gamut of ISO tests yet. Still, SRU verification should include all supported models, and ISO tests to guard against regressions due to interactions with differing kernel versions on the SRU'd releases.
[ Original Description ]
Hi there,
there is the issue connected with "Initramfs unpacking failed: Decoding failed" error during Ubuntu boot. See https:/
Thanks in advance!
CVE References
Changed in linux-raspi (Ubuntu): | |
status: | New → Invalid |
Changed in linux-raspi (Ubuntu Jammy): | |
status: | New → Invalid |
Changed in linux-raspi (Ubuntu Kinetic): | |
status: | New → Invalid |
Changed in linux-firmware-raspi2 (Ubuntu Lunar): | |
status: | Confirmed → Invalid |
Changed in linux-firmware-raspi2 (Ubuntu Kinetic): | |
status: | New → Invalid |
Changed in linux-firmware-raspi2 (Ubuntu Jammy): | |
status: | New → Invalid |
Changed in linux-firmware-raspi (Ubuntu Kinetic): | |
status: | New → Confirmed |
Changed in linux-firmware-raspi (Ubuntu Jammy): | |
status: | New → Confirmed |
importance: | Undecided → High |
Changed in linux-firmware-raspi (Ubuntu Kinetic): | |
importance: | Undecided → High |
Changed in linux-firmware-raspi (Ubuntu Lunar): | |
assignee: | nobody → Dave Jones (waveform) |
importance: | Undecided → High |
Changed in linux-firmware-raspi (Ubuntu Jammy): | |
milestone: | none → ubuntu-22.04.2 |
Changed in linux-firmware-raspi (Ubuntu Lunar): | |
milestone: | none → ubuntu-23.04 |
summary: |
- Old start4.elf / fixup4.dat in 22.04 + CM4 NVMe boot fails due to old start4.elf / fixup4.dat |
description: | updated |
summary: |
- CM4 NVMe boot fails due to old start4.elf / fixup4.dat + [SRU] CM4 NVMe boot fails due to old start4.elf / fixup4.dat |
Changed in linux-firmware-raspi (Ubuntu Jammy): | |
assignee: | nobody → Dave Jones (waveform) |
Changed in linux-firmware-raspi (Ubuntu Kinetic): | |
assignee: | nobody → Dave Jones (waveform) |
Status changed to 'Confirmed' because the bug affects multiple users.