Re: [BUG net-next] arch/x86/kernel/cpu/bugs.c:2935: "Unpatched return thunk in use. This should not happen!" [STACKTRACE]

From: Michael Roth
Date: Thu Mar 28 2024 - 08:39:03 EST


On Tue, Mar 26, 2024 at 08:15:12PM +0100, Mirsad Todorovac wrote:
> On 3/26/24 11:16, Borislav Petkov wrote:
> > On Wed, Mar 20, 2024 at 02:28:57AM +0100, Mirsad Todorovac wrote:
> > > Please find the kernel .config attached.
> >
> > Thanks, that's one huuuge kernel you're building. :)
> >
> > > I got another one of these "Unpatched thunk" and it seems connected
> > > with selftest/kvm.
> > >
> > > But running selftests/kvm one by one did not trigger the bug.
> >
> > Which commands are you exactly running?
> >
> > I'll try to reproduce here.
>
> I think I have a reproducer here on the latest torvalds vanilla tree (on Ubuntu 22.04 LTS box):
>
> root# tools/testing/selftests/kvm/x86_64/nx_huge_pages_test.sh
> Running test with CAP_SYS_BOOT enabled
> Running as root, skipping nx_huge_pages_test with CAP_SYS_BOOT disabled
> root# git describe
> v6.9-rc1-5-g928a87efa423
> root#

I'm seeing it pretty consistently on kvm/next as well. Not sure if
there's anything special about my config but starting a fairly basic
SVM guest seems to be enough to trigger it for me on the first
invocation of svm_vcpu_run().

It seems to be 2 call-sites, one inside:

amd_clear_divider()

and another inside:

__svm_vcpu_run()

which seems to match up with the decoded stack you posted here. Maybe
the first case would be easiest to focus on? It's a fairly
straight-forward use of ALTERNATIVE():

void noinstr amd_clear_divider(void)
{
asm volatile(ALTERNATIVE("", "div %2\n\t", X86_BUG_DIV0)
:: "a" (0), "d" (0), "r" (1));
}
EXPORT_SYMBOL_GPL(amd_clear_divider);

and it's been that way since before 4461438a84 ("x86/retpoline: Ensure
default return thunk isn't used at runtime") was added. Not sure if
anything else has changed underneath the covers since 4461438a84.

-Mike

>
> > Thx.
>
> Not at all.
>
> The stacktrace for the bug triggered by the above command was:
>
> kernel: [ 101.973612] ------------[ cut here ]------------
> kernel: [ 101.973615] Unpatched return thunk in use. This should not happen!
> kernel: [ 101.973618] WARNING: CPU: 1 PID: 3827 at arch/x86/kernel/cpu/bugs.c:2935 __warn_thunk (./arch/x86/kernel/cpu/bugs.c:2935 (discriminator 3))
> kernel: [ 101.973625] Modules linked in: xfrm_user nf_tables nfnetlink nvme_fabrics binfmt_misc snd_hda_codec_realtek snd_hda_codec_generic snd_hda_scodec_component snd_hda_codec_hdmi snd_hda_intel snd_intel_dspcfg snd_intel_sdw_acpi snd_hda_codec intel_rapl_msr amd_atl snd_hda_core intel_rapl_common nls_iso8859_1 snd_hwdep snd_pcm edac_mce_amd amdgpu crct10dif_pclmul polyval_clmulni snd_seq_midi polyval_generic snd_seq_midi_event ghash_clmulni_intel sha512_ssse3 snd_rawmidi sha256_ssse3 amdxcp sha1_ssse3 drm_exec aesni_intel snd_seq gpu_sched crypto_simd drm_buddy cryptd drm_suballoc_helper drm_ttm_helper snd_seq_device joydev input_leds rapl ttm snd_timer wmi_bmof drm_display_helper cec snd drm_kms_helper k10temp ccp i2c_algo_bit soundcore mac_hid tcp_bbr msr parport_pc ppdev lp parport drm efi_pstore ip_tables x_tables autofs4 btrfs blake2b_generic xor raid6_pq hid_generic nvme r8169 xhci_pci ahci nvme_core crc32_pclmul i2c_piix4 xhci_pci_renesas nvme_auth realtek libahci video wmi gpio_amdpt
> kernel: [ 101.973685] CPU: 1 PID: 3827 Comm: nx_huge_pages_t Not tainted 6.9.0-rc1-torv-00005-g928a87efa423-dirty #36
> kernel: [ 101.973687] Hardware name: ASRock X670E PG Lightning/X670E PG Lightning, BIOS 1.21 04/26/2023
> kernel: [ 101.973688] RIP: 0010:__warn_thunk (./arch/x86/kernel/cpu/bugs.c:2935 (discriminator 3))
> kernel: [ 101.973691] Code: 62 c5 1d 01 83 e3 01 74 0e 48 8b 5d f8 c9 31 f6 31 ff e9 be 98 3b 01 48 c7 c7 98 21 c1 bc c6 05 22 26 8d 02 01 e8 90 aa 07 00 <0f> 0b 48 8b 5d f8 c9 31 f6 31 ff e9 9b 98 3b 01 90 90 90 90 90 90
> All code
> ========
> 0: 62 c5 1d 01 83 (bad)
> 5: e3 01 jrcxz 0x8
> 7: 74 0e je 0x17
> 9: 48 8b 5d f8 mov -0x8(%rbp),%rbx
> d: c9 leave
> e: 31 f6 xor %esi,%esi
> 10: 31 ff xor %edi,%edi
> 12: e9 be 98 3b 01 jmp 0x13b98d5
> 17: 48 c7 c7 98 21 c1 bc mov $0xffffffffbcc12198,%rdi
> 1e: c6 05 22 26 8d 02 01 movb $0x1,0x28d2622(%rip) # 0x28d2647
> 25: e8 90 aa 07 00 call 0x7aaba
> 2a:* 0f 0b ud2 <-- trapping instruction
> 2c: 48 8b 5d f8 mov -0x8(%rbp),%rbx
> 30: c9 leave
> 31: 31 f6 xor %esi,%esi
> 33: 31 ff xor %edi,%edi
> 35: e9 9b 98 3b 01 jmp 0x13b98d5
> 3a: 90 nop
> 3b: 90 nop
> 3c: 90 nop
> 3d: 90 nop
> 3e: 90 nop
> 3f: 90 nop
>
> Code starting with the faulting instruction
> ===========================================
> 0: 0f 0b ud2
> 2: 48 8b 5d f8 mov -0x8(%rbp),%rbx
> 6: c9 leave
> 7: 31 f6 xor %esi,%esi
> 9: 31 ff xor %edi,%edi
> b: e9 9b 98 3b 01 jmp 0x13b98ab
> 10: 90 nop
> 11: 90 nop
> 12: 90 nop
> 13: 90 nop
> 14: 90 nop
> 15: 90 nop
> kernel: [ 101.973692] RSP: 0018:ffffbbd90580fc90 EFLAGS: 00010046
> kernel: [ 101.973694] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
> kernel: [ 101.973695] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
> kernel: [ 101.973696] RBP: ffffbbd90580fc98 R08: 0000000000000000 R09: 0000000000000000
> kernel: [ 101.973697] R10: 0000000000000000 R11: 0000000000000000 R12: ffff9964e4b7d4f0
> kernel: [ 101.973698] R13: 0000000000000000 R14: 0000000000000000 R15: ffff9964e4b7dc70
> kernel: [ 101.973699] FS: 0000720b95372740(0000) GS:ffff9973d7a80000(0000) knlGS:0000000000000000
> kernel: [ 101.973700] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> kernel: [ 101.973701] CR2: 0000000000000000 CR3: 00000001aea6c000 CR4: 0000000000f50ef0
> kernel: [ 101.973703] PKRU: 55555554
> kernel: [ 101.973703] Call Trace:
> kernel: [ 101.973704] <TASK>
> kernel: [ 101.973706] ? show_regs (./arch/x86/kernel/dumpstack.c:479)
> kernel: [ 101.973709] ? __warn_thunk (./arch/x86/kernel/cpu/bugs.c:2935 (discriminator 3))
> kernel: [ 101.973711] ? __warn (./kernel/panic.c:694)
> kernel: [ 101.973713] ? __warn_thunk (./arch/x86/kernel/cpu/bugs.c:2935 (discriminator 3))
> kernel: [ 101.973715] ? report_bug (./lib/bug.c:201 ./lib/bug.c:219)
> kernel: [ 101.973718] ? irq_work_queue (./kernel/irq_work.c:119)
> kernel: [ 101.973722] ? handle_bug (./arch/x86/kernel/traps.c:218)
> kernel: [ 101.973725] ? exc_invalid_op (./arch/x86/kernel/traps.c:260 (discriminator 1))
> kernel: [ 101.973727] ? asm_exc_invalid_op (././arch/x86/include/asm/idtentry.h:621)
> kernel: [ 101.973731] ? __warn_thunk (./arch/x86/kernel/cpu/bugs.c:2935 (discriminator 3))
> kernel: [ 101.973734] warn_thunk_thunk (./arch/x86/entry/entry.S:48)
> kernel: [ 101.973738] svm_vcpu_enter_exit (././include/linux/kvm_host.h:547 ./arch/x86/kvm/svm/svm.c:4115)
> kernel: [ 101.973740] svm_vcpu_run (././arch/x86/include/asm/cpufeature.h:171 ./arch/x86/kvm/svm/svm.c:4186)
> kernel: [ 101.973744] kvm_arch_vcpu_ioctl_run (./arch/x86/kvm/x86.c:11008 ./arch/x86/kvm/x86.c:11211 ./arch/x86/kvm/x86.c:11437)
> kernel: [ 101.973747] ? srso_alias_return_thunk (./arch/x86/lib/retpoline.S:181)
> kernel: [ 101.973750] ? srso_alias_return_thunk (./arch/x86/lib/retpoline.S:181)
> kernel: [ 101.973752] ? kvm_vm_stats_read (./arch/x86/kvm/../../../virt/kvm/kvm_main.c:5066)
> kernel: [ 101.973755] kvm_vcpu_ioctl (./arch/x86/kvm/../../../virt/kvm/kvm_main.c:4464)
> kernel: [ 101.973757] ? srso_alias_return_thunk (./arch/x86/lib/retpoline.S:181)
> kernel: [ 101.973759] ? trace_hardirqs_on_prepare (./kernel/trace/trace_preemptirq.c:47 ./kernel/trace/trace_preemptirq.c:42)
> kernel: [ 101.973761] ? srso_alias_return_thunk (./arch/x86/lib/retpoline.S:181)
> kernel: [ 101.973763] ? syscall_exit_to_user_mode (./kernel/entry/common.c:221)
> kernel: [ 101.973765] ? srso_alias_return_thunk (./arch/x86/lib/retpoline.S:181)
> kernel: [ 101.973767] ? do_syscall_64 (././arch/x86/include/asm/cpufeature.h:171 ./arch/x86/entry/common.c:98)
> kernel: [ 101.973770] __x64_sys_ioctl (./fs/ioctl.c:51 ./fs/ioctl.c:904 ./fs/ioctl.c:890 ./fs/ioctl.c:890)
> kernel: [ 101.973773] do_syscall_64 (./arch/x86/entry/common.c:52 ./arch/x86/entry/common.c:83)
> kernel: [ 101.973775] ? srso_alias_return_thunk (./arch/x86/lib/retpoline.S:181)
> kernel: [ 101.973777] ? irqentry_exit (./kernel/entry/common.c:367)
> kernel: [ 101.973778] ? srso_alias_return_thunk (./arch/x86/lib/retpoline.S:181)
> kernel: [ 101.973780] entry_SYSCALL_64_after_hwframe (./arch/x86/entry/entry_64.S:129)
> kernel: [ 101.973782] RIP: 0033:0x720b9511a94f
> kernel: [ 101.973798] 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 <41> 89 c0 3d 00 f0 ff ff 77 1f 48 8b 44 24 18 64 48 2b 04 25 28 00
> All code
> ========
> 0: 00 48 89 add %cl,-0x77(%rax)
> 3: 44 24 18 rex.R and $0x18,%al
> 6: 31 c0 xor %eax,%eax
> 8: 48 8d 44 24 60 lea 0x60(%rsp),%rax
> d: c7 04 24 10 00 00 00 movl $0x10,(%rsp)
> 14: 48 89 44 24 08 mov %rax,0x8(%rsp)
> 19: 48 8d 44 24 20 lea 0x20(%rsp),%rax
> 1e: 48 89 44 24 10 mov %rax,0x10(%rsp)
> 23: b8 10 00 00 00 mov $0x10,%eax
> 28: 0f 05 syscall
> 2a:* 41 89 c0 mov %eax,%r8d <-- trapping instruction
> 2d: 3d 00 f0 ff ff cmp $0xfffff000,%eax
> 32: 77 1f ja 0x53
> 34: 48 8b 44 24 18 mov 0x18(%rsp),%rax
> 39: 64 fs
> 3a: 48 rex.W
> 3b: 2b .byte 0x2b
> 3c: 04 25 add $0x25,%al
> 3e: 28 00 sub %al,(%rax)
>
> Code starting with the faulting instruction
> ===========================================
> 0: 41 89 c0 mov %eax,%r8d
> 3: 3d 00 f0 ff ff cmp $0xfffff000,%eax
> 8: 77 1f ja 0x29
> a: 48 8b 44 24 18 mov 0x18(%rsp),%rax
> f: 64 fs
> 10: 48 rex.W
> 11: 2b .byte 0x2b
> 12: 04 25 add $0x25,%al
> 14: 28 00 sub %al,(%rax)
> kernel: [ 101.973799] RSP: 002b:00007ffd786b9ca0 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
> kernel: [ 101.973801] RAX: ffffffffffffffda RBX: 0000000000600000 RCX: 0000720b9511a94f
> kernel: [ 101.973802] RDX: 0000000000000000 RSI: 000000000000ae80 RDI: 0000000000000005
> kernel: [ 101.973803] RBP: 0000720b953726c0 R08: 000000000041b228 R09: 0000000000000000
> kernel: [ 101.973804] R10: 0000720b951d8882 R11: 0000000000000246 R12: 000000000c9b18c0
> kernel: [ 101.973805] R13: 000000000c9b18c0 R14: 0000000000000000 R15: 0000000000000064
> kernel: [ 101.973809] </TASK>
> kernel: [ 101.973810] ---[ end trace 0000000000000000 ]---
>
> NOTE: Cc:-ed author of the reproducer for these results.
> NOTE 2: The stacktrace is only displayed once, repeating the reproducer doesn't work until the next reboot.
>
> Sending the latest config as well attached:
>
> Best regards,
> Mirsad Todorovac