FS#72777 - [linux] after upgrading kernel to linux 5.15.2 boot fails at 'Mount: /new_root: can't find UUID'

Attached to Project: Arch Linux
Opened by nobody (0nobody0) - Wednesday, 17 November 2021, 20:43 GMT
Last edited by Antonio Rojas (arojas) - Monday, 27 December 2021, 14:45 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To Jan Alexander Steffens (heftig)
Architecture x86_64
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

Description:

arch install fails to boot using kernel linux 5.15.2. linux-lts 5.10.79 and linux 5.14.14 boots fine.

laptop gets super hot immediately on boot and hangs for a long time at `loading initial ramdisk` (30-60 seconds) before eventually returning:

:: mounting 'UUID=0a4d77a3-2b16-4248-9610-50a7f56d2c7a' on real root
mount: /new_root: can't find UUID=0a4d77a3-2b16-4248-9610-50a7f56d2c7a`
You are now being dropped into an emergency shell.
sh: can't access tty; job control turned off

setting rd.udev.log.priority and loglevel to 7 reveals the last messages at the time loglevel 3 would be hanging at 'loading initial ramdisk' are:

ata1 SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata2 SATA link up 1.5 Gbps (SStatus 113 SControl 300)

also noticed earlier in the logs:

ata1.00: qc timeout (cmd 0xec)
ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)
ata2.00: qc timeout (cmd 0xa1)
ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)

unfortunately without the root partition mounting logs aren't saved and they scroll too fast to read otherwise. happy to provide further logs if possible.

the system is https://everymac.com/systems/apple/macbook_pro/specs/macbook-pro-core-2-duo-2.53-aluminum-13-mid-2009-sd-firewire-800-unibody-specs.html with a corsair force GT SSD

Additional info:
linux-5.15.2.arch1-1
further description of the issue and system config/logs can be found at https://forum.garudalinux.org/t/mount-new-root-cant-find-uuid-following-linux-zen-kernel-update-on-clean-install/14448
This task depends upon

Closed by  Antonio Rojas (arojas)
Monday, 27 December 2021, 14:45 GMT
Reason for closing:  Fixed
Comment by loqs (loqs) - Wednesday, 17 November 2021, 21:05 GMT
Nvidia ION system? If so please try the patches from [1]

[1] https://lore.kernel.org/linux-pci/20211104180130.3825416-1-maz%40kernel.org/
Comment by nobody (0nobody0) - Thursday, 18 November 2021, 05:51 GMT
thank you. just compiled the kernel with the drivers/pci/msi.c , drivers/pci/quirks.c , and include/linux/pci.h patches from the above link and unfortunately the same behavior continues to occur.

this does appear to be ion:

00:0b.0 SATA controller: NVIDIA Corporation MCP79 AHCI Controller (rev b1) (prog-if 01 [AHCI 1.0])
Subsystem: NVIDIA Corporation Apple iMac 9,1
Flags: bus master, 66MHz, fast devsel, latency 0, IRQ 26
I/O ports at 21d8 [size=8]
I/O ports at 21ec [size=4]
I/O ports at 21d0 [size=8]
I/O ports at 21e8 [size=4]
I/O ports at 21c0 [size=16]
Memory at d3484000 (32-bit, non-prefetchable) [size=8K]
Capabilities: [44] Power Management version 2
Capabilities: [8c] SATA HBA v1.0
Capabilities: [b0] MSI: Enable+ Count=1/8 Maskable- 64bit+
Kernel driver in use: ahci

should i also apply the patch from [2]

[2]87bl41qkrh.wl-maz@kernel.org/#t"> https://lore.kernel.org/all/87bl41qkrh.wl-maz@kernel.org/#t
Comment by nobody (0nobody0) - Thursday, 18 November 2021, 05:54 GMT
apologies for the formatting:

[2] [url]https://lore.kernel.org/all/87bl41qkrh.wl-maz@kernel.org/#t[/url]
Comment by loqs (loqs) - Thursday, 18 November 2021, 06:07 GMT
Please try linux-mainline 5-16-rc1 which already has those fixes applied. Available prebuilt from [2].

If that does not work you may have to contact upstream to let them know the fix does not work for you.

Flyspray interprets any address with @ in it as an email address, you can work around it by replacing @ with %40

[1] https://aur.archlinux.org/packages/linux-mainline/
[2] https://wiki.archlinux.org/title/Unofficial_user_repositories#miffe
Comment by nobody (0nobody0) - Thursday, 18 November 2021, 07:19 GMT
thank you on all counts. adding the patch from [1] to the patches referenced in the first reply above got the patched 5.15.2 kernel to boot.

tested linux-mainline 5-16-rc1 as is from the repo and it boots fine so looks like all's well from here on out.

[1] https://lore.kernel.org/all/87bl41qkrh.wl-maz%40kernel.org
Comment by Jan Alexander Steffens (heftig) - Thursday, 18 November 2021, 21:54 GMT
Is the ID of your controller 10de:cb79 ?
Comment by nobody (0nobody0) - Thursday, 18 November 2021, 23:39 GMT
it shows up as MCP79 AHCI Controller [10de:0ab9] (rev b1)
Comment by nobody (0nobody0) - Thursday, 18 November 2021, 23:42 GMT
thanks for asking - just realized the reason the first set of patches might not have worked is the need to change 0x0ab8 to 0x0ab9 in +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_NVIDIA, 0x0ab8, nvidia_ion_ahci_fixup); ?
Comment by Jan Alexander Steffens (heftig) - Thursday, 18 November 2021, 23:56 GMT
5.15.3.arch1-1 might fix this, I added another line referencing 0ab9
Comment by nobody (0nobody0) - Thursday, 18 November 2021, 23:57 GMT
thank you
Comment by nobody (0nobody0) - Friday, 19 November 2021, 23:47 GMT
to confirm, just installed linux 5.15.3.arch1-1 and it boots fine.
Comment by Magnus (DeArchDev) - Sunday, 26 December 2021, 13:46 GMT
This might be a kernel issue instead of archlinux's. If you have found a fix to this issue or your issue has been fixed, you might want to request a closure to the issue and submit the issue to the linux kernel's team instead.

Loading...