Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.kernel > #80831 > unrolled thread

Bug#1032104: Status update

Back to article view | Back to linux.debian.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Bug#1032104: Status update Timothy Pearson <tpearson@raptorengineering.com> - 2023-11-07 18:00 +0100
    Bug#1032104: Status update Timothy Pearson <tpearson@raptorengineering.com> - 2023-11-14 06:30 +0100

#80831 — Bug#1032104: Status update

I have traced this to a regression in the Linux kernel.  The issue appears to be a type of data race that is more likely to occur on ppc64el than on other architectures, but is also likely to affect other architectures.  The issue remains in the latest GIT version of the Linux kernel, and I am working with both upstream and our internal resources to try to isolate the root cause and generate a fix.

In the interim, disabling the io_uring subsystem will allow mariadb to function normally.  Given the nature of the kernel bug, I would recommend disabling io_uring entirely in the kernel configuration for affected systems, as other applications may also be impacted by the data corruption.

[toc] | [next] | [standalone]

#80873

I have traced this bug to a missing memory barrier in the powerpc IPI handling code.  io_uring uses task_work_add() to schedule I/O worker creation, which in turn issues an IPI, and when precise timing conditions are met the inconsistent state between the two CPU cores can lead to corruption of userspace data in RAM.

I have sent a patch upstream, and created a merge request for Debian here:

https://salsa.debian.org/kernel-team/linux/-/merge_requests/907

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.kernel

csiph-web