Update: 6.8.0-78-generic also crashes (different oops)
I reproduced a crash on 6.8.0-78-generic at 2025-09-10 14:39 EDT. This
time the oops is a general protection fault in the slab allocator:
RIP: kmem_cache_alloc+0xce/0x370
Call trace (tail):
alloc_buffer_head → folio_alloc_buffers → create_empty_buffers
→ ext4_block_write_begin → ext4_da_write_begin → vfs_write …
Process at fault time: PID 15336, COMM "JTS-uscryptoLis".
PKRU: 0x55555554. Kdump captured a partial vmcore.
Right before the fault the log shows very frequent virtio_gpu atomic
flips (drm_mode_addfb2 / nonblocking commits), similar to what we see on
6.8.0-79.
Excerpts:
[89133.713100] general protection fault … RIP: kmem_cache_alloc+0xce/0x370
… ext4_block_write_begin+0x4bf/0x570
… (lots of drm_atomic_* activity immediately before)
I’ve attached:
- journalctl.250910.1439.log
- qmp-events.250910.log
- dmesg.202509101440
- serial-console.250910.log
- dump.202509101440 (kdump partial vmcore)
- kdump_lock and kexec_cmd
(tarball: crash-58.tgz)
Notes:
- 6.8.0-79 showed a different signature earlier (“Bad pagetable: 000d” in
do_user_addr_fault from a Java AWT thread). Now 6.8.0-78 hits a GPF in
kmem_cache_alloc during ext4 writes.
- Both crashes occur under a GUI workload in a KVM/QEMU guest with virtio_gpu +
virgl 3D acceleration (needed for 5120x1440). Mesa 23.2.1-* in guest. Xwayland
in use; Xorg stack available too.
Questions / next steps:
- Does this look like memory corruption triggered by virtio_gpu/virgl
(userspace) leading to random kernel splats later, or a kernel-side bug
(DRM/MM/SLUB)?
Environment recap (guest):
- Ubuntu 22.04 (Jammy) HWE kernels 6.8.0-78 and 6.8.0-79 (both crash)
- QEMU/KVM, machine Q35; GPU: virtio GPU [1af4:1050]; virgl on
- Kernel cmdline: workqueue.default_to_unbound=1 console=ttyS0,115200n8
console=tty0 panic=10 loglevel=7 systemd.show_status=1 drm.debug=0x1e
crashkernel=512M-:192M
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2122445
Title:
Jammy HWE 6.8.0-79 in KVM/QEMU guest (virtio_gpu/virgl): kernel oops
“Bad pagetable: 000d” in do_user_addr_fault while running Java
AWT/JavaFX GUI
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux-hwe-6.8/+bug/2122445/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
