Re: [syzbot] [kvm-x86?] WARNING in kvm_arch_can_dequeue_async_page_present

[Sean Christopherson <seanjc@google.com>]

On Fri, Nov 07, 2025, syzbot wrote:
> Hello,
> 
> syzbot found the following issue on:
> 
> HEAD commit:    9c0826a5d9aa Add linux-next specific files for 20251107
> git tree:       linux-next
> console output: https://syzkaller.appspot.com/x/log.txt?x=13a67012580000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=4f8fcc6438a785e7
> dashboard link: https://syzkaller.appspot.com/bug?extid=6bea72f0c8acbde47c55
> compiler:       Debian clang version 20.1.8 (++20250708063551+0c9f909b7976-1~exp1~20250708183702.136), Debian LLD 20.1.8
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=14e110b4580000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=14ab1114580000
> 
> Downloadable assets:
> disk image: https://storage.googleapis.com/syzbot-assets/6b76dc0ec17f/disk-9c0826a5.raw.xz
> vmlinux: https://storage.googleapis.com/syzbot-assets/522b6d2a1d1d/vmlinux-9c0826a5.xz
> kernel image: https://storage.googleapis.com/syzbot-assets/4a58225d70f3/bzImage-9c0826a5.xz
> 
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+6bea72f0c8acbde47c55@syzkaller.appspotmail.com
> 
> kvm_intel: L1TF CPU bug present and SMT on, data leak possible. See CVE-2018-3646 and https://www.kernel.org/doc/html/latest/admin-guide/hw-vuln/l1tf.html for details.
> ------------[ cut here ]------------
> WARNING: arch/x86/kvm/x86.c:13965 at kvm_arch_can_dequeue_async_page_present+0x1a9/0x2f0 arch/x86/kvm/x86.c:13965, CPU#0: syz.0.17/5998
> Modules linked in:
> CPU: 0 UID: 0 PID: 5998 Comm: syz.0.17 Not tainted syzkaller #0 PREEMPT(full) 
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/02/2025
> RIP: 0010:kvm_arch_can_dequeue_async_page_present+0x1a9/0x2f0 arch/x86/kvm/x86.c:13965
> Code: 00 65 48 8b 0d 58 81 72 11 48 3b 4c 24 40 75 21 48 8d 65 d8 5b 41 5c 41 5d 41 5e 41 5f 5d e9 3e af 20 0a cc e8 48 e1 79 00 90 <0f> 0b 90 b0 01 eb c0 e8 4b c1 1d 0a f3 0f 1e fa 4c 8d b3 f8 02 00
> RSP: 0018:ffffc90003167460 EFLAGS: 00010293
> RAX: ffffffff8147eee8 RBX: ffff888030280000 RCX: ffff88807fda1e80
> RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000009
> RBP: ffffc900031674e8 R08: ffff88803028003f R09: 1ffff11006050007
> R10: dffffc0000000000 R11: ffffed1006050008 R12: 1ffff9200062ce8c
> R13: dffffc0000000000 R14: 0000000000000000 R15: dffffc0000000000
> FS:  000055556a8c3500(0000) GS:ffff888125a79000(0000) knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 0000000000000000 CR3: 0000000072cc2000 CR4: 00000000003526f0
> Call Trace:
>  <TASK>
>  kvm_check_async_pf_completion+0x102/0x3c0 virt/kvm/async_pf.c:158
>  vcpu_enter_guest arch/x86/kvm/x86.c:11209 [inline]
>  vcpu_run+0x26be/0x7760 arch/x86/kvm/x86.c:11650
>  kvm_arch_vcpu_ioctl_run+0x116c/0x1cb0 arch/x86/kvm/x86.c:11995
>  kvm_vcpu_ioctl+0x99a/0xed0 virt/kvm/kvm_main.c:4477
>  vfs_ioctl fs/ioctl.c:51 [inline]
>  __do_sys_ioctl fs/ioctl.c:597 [inline]
>  __se_sys_ioctl+0xfc/0x170 fs/ioctl.c:583
>  do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
>  do_syscall_64+0xfa/0xfa0 arch/x86/entry/syscall_64.c:94
>  entry_SYSCALL_64_after_hwframe+0x77/0x7f
> RIP: 0033:0x7f1588b8f6c9
> Code: ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 a8 ff ff ff f7 d8 64 89 01 48
> RSP: 002b:00007ffdfcd816a8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
> RAX: ffffffffffffffda RBX: 00007f1588de5fa0 RCX: 00007f1588b8f6c9
> RDX: 0000000000000000 RSI: 000000000000ae80 RDI: 0000000000000006
> RBP: 00007f1588c11f91 R08: 0000000000000000 R09: 0000000000000000
> R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
> R13: 00007f1588de5fa0 R14: 00007f1588de5fa0 R15: 0000000000000003
>  </TASK>

/facepalm

This is due to a new WARN[*] that was added to yell if KVM tries to process an
async #PF after the guest disables *PV* async #PFs.  But I completely forgot that
KVM uses the same paths for non-PV async #PF (KVM doesn't inject anything into
the guest and instead puts the vCPU into a "synthetic" HLT state so that if an
IRQ/NMI comes along for the vCPU, KVM can immediately delivery the interrupt).

I'll simply drop the patch.

[*] https://lkml.kernel.org/r/20251015033258.50974-2-mlevitsk%40redhat.com

> ---
> This report is generated by a bot. It may contain errors.
> See https://goo.gl/tpsmEJ for more information about syzbot.
> syzbot engineers can be reached at syzkaller@googlegroups.com.
> 
> syzbot will keep track of this issue. See:
> https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
> 
> If the report is already addressed, let syzbot know by replying with:
> #syz fix: exact-commit-title
> 
> If you want syzbot to run the reproducer, reply with:
> #syz test: git://repo/address.git branch-or-commit-hash
> If you attach or paste a git patch, syzbot will apply it before testing.
> 
> If you want to overwrite report's subsystems, reply with:
> #syz set subsystems: new-subsystem
> (See the list of subsystem names on the web dashboard)
> 
> If the report is a duplicate of another one, reply with:
> #syz dup: exact-subject-of-another-report
> 
> If you want to undo deduplication, reply with:
> #syz undup