On Wed, Feb 3, 2021 at 2:44 PM Chris Murphy <lists@xxxxxxxxxxxxxxxxx> wrote: > > On Wed, Feb 3, 2021 at 1:26 PM Chris Murphy <lists@xxxxxxxxxxxxxxxxx> wrote: > > > > qemu-kvm VM starts with kernel 5.10.10 but fails with 5.11.0-rc5. > > > > Libvirt folks think this is a kernel bug, and have attached a > > reproducer to the downstream bug report. > > > > "I've managed to reproduce and found that virBPFLoadProg() logs the > > following message: > > > > in-kernel BTF is malformed\nprocessed 0 insns (limit 1000000) > > max_states_per_insn 0 total_states 0 peak_states 0 mark_read 0\n > > " > > > > https://bugzilla.redhat.com/show_bug.cgi?id=1920857#c4 > > Looks like the bug was introduced in 5.11-rc5, the problem doesn't > happen in rc4. As I mention in the downstream bug, I'm unable to > compile a working kernel for bisect between rc4 and rc5 to find out > the exact commit that introduced the problem, due to many messages > like this: > > Feb 03 15:05:47 kernel: failed to validate module [coretemp] BTF: -22 > Feb 03 15:05:47 kernel: failed to validate module [intel_powerclamp] BTF: -22 > Feb 03 15:05:47 kernel: failed to validate module [irqbypass] BTF: -22 > Feb 03 15:05:47 kernel: failed to validate module [intel_powerclamp] BTF: -22 > Feb 03 15:05:47 kernel: failed to validate module > [x86_pkg_temp_thermal] BTF: -22 > The important and very relevant part from the bugzilla: Feb 03 15:06:26 fmac.local kernel: BPF: sched_reset_on_fork type_id=6 bitfield_size=0 bits_offset=0 Feb 03 15:06:26 fmac.local kernel: BPF: Feb 03 15:06:26 fmac.local kernel: BPF:Invalid member bits_offset Feb 03 15:06:26 fmac.local kernel: BPF: Do you have full dmesg with output from the BPF verifier? Also, what's the kernel config? Which compiler and what version, etc, etc? Please help to reproduce this with as much information as possible. Thanks! If you can share the vmlinux itself, that would help as well. > > -- > Chris Murphy
