Bug 217688 - Guest call trace during boot
Summary: Guest call trace during boot
Status: RESOLVED CODE_FIX
Alias: None
Product: Virtualization
Classification: Unclassified
Component: kvm (show other bugs)
Hardware: Intel Linux
: P3 normal
Assignee: virtualization_kvm
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2023-07-20 03:09 UTC by Chen, Fan
Modified: 2023-07-28 02:30 UTC (History)
2 users (show)

See Also:
Kernel Version:
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description Chen, Fan 2023-07-20 03:09:48 UTC
Environment:

Host/guest kernel: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master fdf0eaf11452d72945af3 6.5.0-rc2
Qemu: https://gitlab.com/qemu-project/qemu.git master 361d5397
Host/Guest OS: CentOS Stream 9
Platform: SPR/CLX

Bug detail description: 

Create VM:
qemu-system-x86_64 -accel kvm -smp 8 -m 8192 -cpu host -machine q35 -drive file=centos9.qcow2,if=none,id=virtio-disk0 -device virtio-blk-pci,drive=virtio-disk0,bootindex=0 -daemonize -vnc :3 -device virtio-net-pci,netdev=nic0,mac=00:b9:f9:b2:90:72 -netdev tap,id=nic0,br=virbr0,helper=/usr/local/libexec/qemu-bridge-helper,vhost=on

When VM boot, we can see below Call trace:

[    0.387684] Key type asymmetric registered
[    0.388161] Asymmetric key parser 'x509' registered
[    0.388717] Block layer SCSI generic (bsg) driver version 0.4 loaded (major 246)
[    0.389580] io scheduler mq-deadline registered
[    0.390097] io scheduler kyber registered
[    0.391265] atomic64_test: passed for x86-64 platform with CX8 and with SSE
[    0.392143] unchecked MSR access error: RDMSR from 0xe2 at rIP: 0xffffffffab090378 (native_read_msr+0x8/0x40)
[    0.393213] Call Trace:
[    0.393539]  <TASK>
[    0.393824]  ? ex_handler_msr+0x128/0x140
[    0.394327]  ? fixup_exception+0x89/0x340
[    0.394798]  ? exc_general_protection+0xdc/0x3d0
[    0.395325]  ? asm_exc_general_protection+0x26/0x30
[    0.395873]  ? __pfx_intel_idle_init+0x10/0x10
[    0.396384]  ? native_read_msr+0x8/0x40
[    0.396834]  intel_idle_init_cstates_icpu.constprop.0+0x5e/0x560
[    0.397492]  ? __pfx_intel_idle_init+0x10/0x10
[    0.397997]  intel_idle_vminit.isra.0+0xee/0x1d0
[    0.398517]  do_one_initcall+0x45/0x220
[    0.398971]  do_initcalls+0xac/0x130
[    0.399400]  kernel_init_freeable+0x128/0x1e0
[    0.399896]  ? __pfx_kernel_init+0x10/0x10
[    0.400374]  kernel_init+0x1a/0x1c0
[    0.400790]  ret_from_fork+0x31/0x50
[    0.401219]  ? __pfx_kernel_init+0x10/0x10
[    0.401692]  ret_from_fork_asm+0x1b/0x30
[    0.402151] RIP: 0000:0x0
[    0.402492] Code: Unable to access opcode bytes at 0xffffffffffffffd6.
[    0.403190] RSP: 0000:0000000000000000 EFLAGS: 00000000 ORIG_RAX: 0000000000000000
[    0.404035] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
[    0.404891] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
[    0.405680] RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
[    0.406440] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
[    0.407193] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
[    0.407950]  </TASK>
[    0.408755] input: Power Button as /devices/LNXSYSTM:00/LNXPWRBN:00/input/input0
[    0.409622] ACPI: button: Power Button [PWRF]
[    0.410369] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
[    0.411138] 00:03: ttyS0 at I/O 0x3f8 (irq = 4, base_baud = 115200) is a 16550A
[    0.412241] Non-volatile memory driver v1.3
[    0.412727] Linux agpgart interface v0.103
......

Can be reproduced on latest commit bfa3037d828050 of this mainline linux.git.
And no such issue on host.
Comment 1 Chao Gao 2023-07-20 03:31:54 UTC
this is a known issue. Xiaoyao reported this issue earlier, see https://lore.kernel.org/all/CAJZ5v0gaZHpAri7LRcfpS2TyK_Bsjuxkw9cZUm_uGZAgiub4Jw@mail.gmail.com/
Comment 2 Sean Christopherson 2023-07-20 14:48:43 UTC
Ya, and ttps://lkml.kernel.org/r/CA%2BQYu4qSBdhEgFURu%2BouAf2d_JNPbZgCSUaxCLoGzMqDQOLWsQ%40mail.gmail.com
Comment 3 Chen, Fan 2023-07-28 02:30:06 UTC
Thanks all, cannot be reproduced on 6.5-rc3.

Note You need to log in before you can comment on or make changes to this bug.