Arch Linux

Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines

Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!
Tasklist

FS#78035 - [linux] kernel NULL pointer dereference, address: 0000000000000030 on ucsi_acpi_notify

Attached to Project: Arch Linux
Opened by Pascal Bryner (Veldora) - Wednesday, 29 March 2023, 11:29 GMT
Last edited by Toolybird (Toolybird) - Tuesday, 04 April 2023, 22:16 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To No-one
Architecture x86_64
Severity Medium
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:
The Linux kernel ends in a kernel NULL pointer dereference, address: 0000000000000030 on ucsi_acpi_notify when a usb-c device has been unplugged, which causes the complete system to freeze.

The first we've noticed this problem was with kernel version 6.0.12.arch1-1.
According to older forum-posts, the problem already existed in kernel 5.19.

There has been a merge, which fixes the problem in the current next-kernel.
Problem: the next-kernel fixes won't be available in Linux 6.3.
Either the patch will be backported to the current LTS by the kernel-devs or when Linux 6.4 releases.
Until then, a patch solves this problem.

The PKGBUILD of the kernel has been modified by us to add the patches, attached to this bug-report, internally.
It runs stable and the freezes didn't appear anymore.

Additional info:
* package version(s)
linux 6.0.12.arch1-1 until linux 6.2.8.arch1-1
* config and/or log files etc.
* link to upstream bug report, if any
20230308154244.722337-4-hdegoede@redhat.com/T/"> https://lore.kernel.org/stable/20230308154244.722337-4-hdegoede@redhat.com/T/
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=f87fb985452ab2083967103ac00bfd68fb182764

Steps to reproduce:
- Unclear, but many employees could just unplug a usb-c device and the system froze.
   ucsi.patch (3.2 KiB)
   ucsi_acpi.patch (0.4 KiB)
This task depends upon

Closed by  Toolybird (Toolybird)
Tuesday, 04 April 2023, 22:16 GMT
Reason for closing:  Fixed
Additional comments about closing:  linux 6.2.9.arch1-1 (will be fully fixed in 6.2.10)
Comment by Pascal Bryner (Veldora) - Wednesday, 29 March 2023, 11:51 GMT
For those who seem to think that following issue is the same than this one, they're not.
https://bugs.archlinux.org/task/75666

The mentiond bugreport has completely different stacktraces than we do.
I'll attach one of ours and it's clear, that our Kernel Oops is caused by a ucsi_acpi_notify, while his is caused by __check_heap_object or assoc_array_find.
   ucsi_acpi_notify_kernel_oops.... (4.2 KiB)
Comment by loqs (loqs) - Wednesday, 29 March 2023, 13:00 GMT
Two of the fixes are queued for 6.2.9 [1][2] and 6.1.22 [3][4] leaving [5] outstanding which failed to be apply cleanly [6][7].

[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/commit/queue-6.2/usb-ucsi-fix-null-pointer-deref-in-ucsi_connector_change.patch?id=90ef78c458a64a4c56394d1f2862d74a18a320ee
[2] https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/commit/queue-6.2/usb-ucsi_acpi-increase-the-command-completion-timeout.patch?id=90ef78c458a64a4c56394d1f2862d74a18a320ee
[3] https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/commit/queue-6.1/usb-ucsi-fix-null-pointer-deref-in-ucsi_connector_change.patch?id=2507425120390e74caf717f198e6d40f12390416
[4] https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/commit/queue-6.1/usb-ucsi_acpi-increase-the-command-completion-timeout.patch?id=2507425120390e74caf717f198e6d40f12390416
[5] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=0482c34ec6f8557e06cd0f8e2d0e20e8ede6a22c
[6] https://lore.kernel.org/stable/16800048817970%40kroah.com/
[7] https://lore.kernel.org/stable/16800048821504%40kroah.com/
Comment by Pascal Bryner (Veldora) - Wednesday, 29 March 2023, 14:35 GMT
Ha, this is brilliant!
Thank you for noting that.
Any ideas, if there are any plans available, when 6.2.9 will release on the arch repos?
Comment by Toolybird (Toolybird) - Friday, 31 March 2023, 02:26 GMT
6.2.9 is now in the [testing] repo (minus the "failed-to-apply" patch alluded to by @loqs). Does it solve the problem?
Comment by Pascal Bryner (Veldora) - Friday, 31 March 2023, 06:28 GMT
We need to test it first, if the 6.2.9 works as intended without the freezes. If so, I'll request closure of this issue.
Otherwise, I'll write again here, if the freezes reappear again.

In the worst case, we can still switch back to our modified 6.2.8 kernel and keep it updated with the upstream stable kernel with the patches, until the rest of the patches get merged aswell.
Comment by loqs (loqs) - Monday, 03 April 2023, 19:21 GMT
The third patch is now queued for 6.2.10 [1] and 6.1.23 [2] in case it is needed.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/commit/queue-6.2/usb-ucsi-fix-ucsi-connector-race.patch?id=807bc41323eb2f015d22b410472c4d991b79a771
[2] https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/commit/queue-6.1/usb-ucsi-fix-ucsi-connector-race.patch?id=1ee74e16d22ac5d117ffd37be9b8cc5ec04104f4
Comment by Pascal Bryner (Veldora) - Tuesday, 04 April 2023, 09:04 GMT
Wonderful news.
For now it seems that the 6.2.9 is working on one device without any problems.
But i.m.o it's good to know, that this patch will make it into 6.2.10.

Loading...