Comment 5 for bug 1969275

Thanks!

The picture in comment #4 contains the full stack trace.

Some technical analysis for reference. Don't worry about it.
Please see the next comment for suggestions, for now.

...

When returning to userspace from the exit_group() syscall
(as per: orig rax == 0xe7 == 231 -> sys_exit_group; see [1]),
the work to release the last reference to a file closed
(per fput() bits) called into the DRM handling (apparently
the file might be /dev/drm*, I guess), which called into
the _nouveau_driver (the implementation), which owns the
majority of the trace from there.

Finally, the nouveau probably performed an incorrect
memory access (per CR2, showing what seems to be an over
flowed pointer).

This crashed the kernel.

[1]
$ grep 231 git/linux/arch/x86/entry/syscalls/syscall_64.tbl
231 common exit_group sys_exit_group