Created attachment 306447 [details]
Settings of the VM

Alright, it's not a regression in the kernel but caused by a bios update (I guess).
I get the same on my previous kernel 6.9.0-rc1.
Both my 6.9.0-rc1 6.10.0-rc2 kernels are vanilla builds from kernel.org (unpatched).

After updating the bios/firmware of my mainboard Asus ROG Zenith II Extreme from 1802 to 2102, it always seems to spawn the error:

[ 1150.380137] ------------[ cut here ]------------
[ 1150.380141] Unpatched return thunk in use. This should not happen!
[ 1150.380144] WARNING: CPU: 3 PID: 4849 at arch/x86/kernel/cpu/bugs.c:2935 __warn_thunk+0x40/0x50
[ 1150.380152] Modules linked in: veth rpcsec_gss_krb5 auth_rpcgss nfsv4 nfs lockd grace netfs ebtable_filter ebtables ip_set ip6table_raw iptable_raw ip6table_filter ip6_tables iptable_filter nf_tables scsi_transport_iscsi bonding tls softdog sunrpc nfnetlink_log nfnetlink binfmt_misc amd_atl intel_rapl_msr intel_rapl_common edac_mce_amd kvm_amd kvm crct10dif_pclmul polyval_clmulni polyval_generic ghash_clmulni_intel sha256_ssse3 sha1_ssse3 aesni_intel eeepc_wmi crypto_simd asus_wmi cryptd platform_profile sparse_keymap asus_ec_sensors video pcspkr rapl ccp mxm_wmi wmi_bmof k10temp mac_hid vfio_pci vfio_pci_core vfio_iommu_type1 vfio iommufd vhost_net vhost vhost_iotlb tap nct6775 nct6775_core hwmon_vid lm75 drm efi_pstore dmi_sysfs ip_tables x_tables autofs4 btrfs blake2b_generic xor raid6_pq simplefb dm_thin_pool dm_persistent_data dm_bio_prison dm_bufio libcrc32c igb xhci_pci atlantic nvme ahci crc32_pclmul xhci_pci_renesas i2c_algo_bit libahci dca macsec nvme_core xhci_hcd i2c_piix4 wmi
[ 1150.380266] CPU: 3 PID: 4849 Comm: CPU 0/KVM Not tainted 6.9.0-rc1 #1
[ 1150.380269] Hardware name: ASUS System Product Name/ROG ZENITH II EXTREME, BIOS 2102 02/16/2024
[ 1150.380271] RIP: 0010:__warn_thunk+0x40/0x50
[ 1150.380275] Code: 96 f1 fe 00 83 e3 01 74 0e 48 8b 5d f8 c9 31 f6 31 ff e9 43 1c 08 01 48 c7 c7 b8 f2 f4 9e c6 05 56 61 4c 02 01 e8 00 b1 07 00 <0f> 0b 48 8b 5d f8 c9 31 f6 31 ff e9 20 1c 08 01 90 90 90 90 90 90
[ 1150.380278] RSP: 0018:ffffb478c2ce3ca8 EFLAGS: 00010046
[ 1150.380281] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
[ 1150.380283] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
[ 1150.380285] RBP: ffffb478c2ce3cb0 R08: 0000000000000000 R09: 0000000000000000
[ 1150.380287] R10: 0000000000000000 R11: 0000000000000000 R12: ffff91f80e948000
[ 1150.380289] R13: 0000000000000000 R14: ffffb478c4ab5000 R15: ffff91f80e948038
[ 1150.380291] FS:  00007f74baa006c0(0000) GS:ffff9216bd980000(0000) knlGS:0000000000000000
[ 1150.380293] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1150.380295] CR2: 0000000000000000 CR3: 0000000106750000 CR4: 0000000000350ef0
[ 1150.380298] Call Trace:
[ 1150.380300]  <TASK>
[ 1150.380304]  ? show_regs+0x6c/0x80
[ 1150.380309]  ? __warn+0x88/0x140
[ 1150.380312]  ? __warn_thunk+0x40/0x50
[ 1150.380316]  ? report_bug+0x182/0x1b0
[ 1150.380322]  ? handle_bug+0x46/0x90
[ 1150.380325]  ? exc_invalid_op+0x18/0x80
[ 1150.380329]  ? asm_exc_invalid_op+0x1b/0x20
[ 1150.380336]  ? __warn_thunk+0x40/0x50
[ 1150.380341]  ? __warn_thunk+0x40/0x50
[ 1150.380344]  warn_thunk_thunk+0x16/0x30
[ 1150.380351]  svm_vcpu_enter_exit+0x71/0xc0 [kvm_amd]
[ 1150.380364]  svm_vcpu_run+0x1e7/0x850 [kvm_amd]
[ 1150.380377]  kvm_arch_vcpu_ioctl_run+0xca3/0x16d0 [kvm]
[ 1150.380458]  kvm_vcpu_ioctl+0x295/0x800 [kvm]
[ 1150.380522]  ? srso_return_thunk+0x5/0x5f
[ 1150.380526]  ? __x64_sys_ioctl+0xbb/0xf0
[ 1150.380530]  ? srso_return_thunk+0x5/0x5f
[ 1150.380533]  ? syscall_exit_to_user_mode+0x75/0x1b0
[ 1150.380537]  ? srso_return_thunk+0x5/0x5f
[ 1150.380541]  ? do_syscall_64+0x84/0x140
[ 1150.380544]  ? srso_return_thunk+0x5/0x5f
[ 1150.380547]  ? do_syscall_64+0x84/0x140
[ 1150.380550]  ? switch_fpu_return+0x50/0xe0
[ 1150.380555]  __x64_sys_ioctl+0xa3/0xf0
[ 1150.380559]  do_syscall_64+0x78/0x140
[ 1150.380563]  ? srso_return_thunk+0x5/0x5f
[ 1150.380566]  ? do_syscall_64+0x84/0x140
[ 1150.380569]  entry_SYSCALL_64_after_hwframe+0x6c/0x74
[ 1150.380573] RIP: 0033:0x7f74cbb8cc5b
[ 1150.380592] Code: 00 48 89 44 24 18 31 c0 48 8d 44 24 60 c7 04 24 10 00 00 00 48 89 44 24 08 48 8d 44 24 20 48 89 44 24 10 b8 10 00 00 00 0f 05 <89> c2 3d 00 f0 ff ff 77 1c 48 8b 44 24 18 64 48 2b 04 25 28 00 00
[ 1150.380595] RSP: 002b:00007f74ba9fb060 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[ 1150.380598] RAX: ffffffffffffffda RBX: 000056062f0cf7e0 RCX: 00007f74cbb8cc5b
[ 1150.380600] RDX: 0000000000000000 RSI: 000000000000ae80 RDI: 000000000000001f
[ 1150.380602] RBP: 000000000000ae80 R08: 0000000000000000 R09: 0000000000000000
[ 1150.380604] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
[ 1150.380606] R13: 0000000000000007 R14: 00007ffc10a8f820 R15: 00007f74ba200000
[ 1150.380612]  </TASK>
[ 1150.380613] ---[ end trace 0000000000000000 ]---

This happens when just starting the VM.
Command line: BOOT_IMAGE=/boot/vmlinuz-6.9.0-rc1 root=/dev/mapper/pve-root ro quiet iommu=pt amd_iommu=on kvm_amd.npt=1 video=vesafb:off video=efifb:off video=simplefb:off nomodeset initcall_blacklist=sysfb_init modprobe.blacklist=nouveau modprobe.blacklist=amdgpu modprobe.blacklist=radeon modprobe.blacklist=nvidia amd_pstate=guided

I've attached a screenshot of the settings for the VM.

I believe the new bios updates the AGESA firmware from version:
V9CastlePeakPI-SP3r3-1.0.0.9
To:
CastlePeakPI-SP3r3 1.0.0.A (2023-11-21)

On Mon, Jun 10, 2024, bugzilla-daemon@kernel.org wrote:
> https://bugzilla.kernel.org/show_bug.cgi?id=218949
> 
> --- Comment #2 from Gino Badouri (badouri.g@gmail.com) ---
> Alright, it's not a regression in the kernel but caused by a bios update (I
> guess).
> I get the same on my previous kernel 6.9.0-rc1.

The WARNs are not remotely the same.  The below issue in svm_vcpu_enter_exit()
was resolved in v6.9 final[1].

The lockdep warnings in track_pfn_remap() and remap_pfn_range_notrack() is a
known issue in vfio_pci_mmap_fault(), with an in-progress fix[2] that is destined
for 6.10.

[1] https://lore.kernel.org/all/1d10cd73-2ae7-42d5-a318-2f9facc42bbe@alu.unizg.hr
[2] https://lore.kernel.org/all/20240530045236.1005864-1-alex.williamson@redhat.com

> Both my 6.9.0-rc1 6.10.0-rc2 kernels are vanilla builds from kernel.org
> (unpatched).
> 
> After updating the bios/firmware of my mainboard Asus ROG Zenith II Extreme
> from 1802 to 2102, it always seems to spawn the error:
> 
> [ 1150.380137] ------------[ cut here ]------------
> [ 1150.380141] Unpatched return thunk in use. This should not happen!
> [ 1150.380144] WARNING: CPU: 3 PID: 4849 at arch/x86/kernel/cpu/bugs.c:2935
> __warn_thunk+0x40/0x50

...

> [ 1150.380266] CPU: 3 PID: 4849 Comm: CPU 0/KVM Not tainted 6.9.0-rc1 #1
> [ 1150.380269] Hardware name: ASUS System Product Name/ROG ZENITH II EXTREME,
> BIOS 2102 02/16/2024
> [ 1150.380271] RIP: 0010:__warn_thunk+0x40/0x50

...

> [ 1150.380298] Call Trace:
> [ 1150.380300]  <TASK>
> [ 1150.380344]  warn_thunk_thunk+0x16/0x30
> [ 1150.380351]  svm_vcpu_enter_exit+0x71/0xc0 [kvm_amd]
> [ 1150.380364]  svm_vcpu_run+0x1e7/0x850 [kvm_amd]
> [ 1150.380377]  kvm_arch_vcpu_ioctl_run+0xca3/0x16d0 [kvm]
> [ 1150.380458]  kvm_vcpu_ioctl+0x295/0x800 [kvm]

Hi Sean!

It always amazes me how fast you guys can find the patches/reports for certain bug reports :)

The WARNING happened on 6.9.0-rc1 (so before the final release).

For the pfn warnings I've applied the patchset from https://lore.kernel.org/all/20240530045236.1005864-1-alex.williamson@redhat.com on top of 6.10-rc3 and it's completely fixed, they're gone now.

I've tested Call of Duty MW3 in Windows 11 with the NVIDIA GPU passed through the VM and I didn't notice any performance or stability difference with or without the patch.

Thanks a lot!