After Fedora 33 update; kworker/u32:2+events _unbound killing my cpu

@tom: same for me … kworker issue persist in new kernel 5.10.9 …

… seems our bug found the way to bugzilla@redhat …

https://bugzilla.redhat.com/show_bug.cgi?id=1900099

… and to fedora too …

https://retrace.fedoraproject.org/faf/problems/bthash/?bth=e225dbd17fe5261615a642ea78445ca76ec67763&bth=b56cdce5172e80e5da645568ecd02806eaaed9f4&bth=8b337604620217bfd93768766e42fde6a3413170&bth=efd2e9db1ecfef1af5ced5dc38848af64e320b7a&bth=359258bfd88ca650e72656b233551eea3715c01f&bth=660a89d9e2050a53e74d096f96648a85b130c93e&bth=df8868ab1a6256de4c37bf91a9102e917a7d7e84&bth=02d968d6c629f113b03e619bf8aac7f5fc1f0752&bth=c767e9c3ae963bd565c383dab320bb37dfc9eb1d&bth=dd880b21cd08937bc109ab3654d30d8de9443cc1

… how to patch nouveau driver with DRM timeout causes machine to freeze …

“The upstream patch appears to be some way off making it into mainline before it can be backported to 5.10, so given the scope and effects of the bug, I’m submitting it here to be considered to patch the bullseye kernel in the interim.”

https://www.mail-archive.com/debian-kernel@lists.debian.org/msg120920.html

… and another link which might reflect to our problem …

https://bugzilla.kernel.org/show_bug.cgi?id=210333

… should be fixed in Linux 5.11-rc7 (“Bastian Beranek (1): drm/nouveau/dispnv50: Restore pushing of all data.”) …

https://lkml.org/lkml/2021/1/28/1351

Same issue.
Upgraded recently from Fedora 30 to Fedora 32.
Different kworkers at uncertain conditions (supposedly, after starting VSCode and running it for a while) start to slow down the cores, eventually either unfreezing or hanging down to unusable.
Kernel 5.10.7-100.fc32.x86_64.

“For those wondering: It seems the patch has now been integrated mainline (it is part of 5.11-rc6) and should land in Linux 5.11. Thanks @karolherbst and @skeggsb! Hope that it qualifies for the stable kernel branches as well. ( Bastian Beranek )”

“Quick note: @bastianbeischer’s patch is backported to stable kernel >= v5.10.13: (https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/gpu/drm/nouveau?h=v5.10.13&id=e4d2a196fdc5f7eeab21d3d6a27566f3dc1f4d60)”

… solved …

This issue has nothing to do with the distro as is, more with the kernel modules, specifically the “nouveau” module, as mentioned many times, above.

This had another side-effect for me: whenever I used a multi-threaded compressor such as xz (with the -T8 option) or pbzip2 (with -p8), the compressor wouldn’t use every CPU I threw at it because many were used with the kworker process.

The issue was fixed for me when I switched to the Nvidia-provided driver. It goes against my instincts to use a proprietary driver, but it’s either that or seeing my machine crawl to uselesness.

–JFG