Re: KMSAN: kernel-infoleak in kvm_vcpu_write_guest_page

[Liran Alon]

> On 7 Nov 2018, at 14:10, Alexander Potapenko <glider@xxxxxxxxxx> wrote:
> 
> On Wed, Nov 7, 2018 at 2:38 AM syzbot
> <syzbot+ded1696f6b50b615b630@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
>> 
>> Hello,
>> 
>> syzbot found the following crash on:
>> 
>> HEAD commit:    88b95ef4c780 kmsan: use MSan assembly instrumentation
>> git tree:       https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_google_kmsan.git_master&d=DwIFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=NpjK8vgD4UmRGNWTNe6YFA-AJTZp9VlD0oMFvKFV25I&s=A5G9_UvV5TpBoJitbGWS2VXmfVMXFUgggq64QM-6nqc&e=
>> console output: https://urldefense.proofpoint.com/v2/url?u=https-3A__syzkaller.appspot.com_x_log.txt-3Fx-3D12505e33400000&d=DwIFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=NpjK8vgD4UmRGNWTNe6YFA-AJTZp9VlD0oMFvKFV25I&s=ZGVw04dRMjYdKZf4amN1yl3IheljZZq6V9h3mO9U-vM&e=
>> kernel config:  https://urldefense.proofpoint.com/v2/url?u=https-3A__syzkaller.appspot.com_x_.config-3Fx-3D8df5fc509a1b351b&d=DwIFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=NpjK8vgD4UmRGNWTNe6YFA-AJTZp9VlD0oMFvKFV25I&s=OUIhmSMzBSMhswtebZqyLnc6SkAagVPBGr0EmCLL2Fg&e=
>> dashboard link: https://urldefense.proofpoint.com/v2/url?u=https-3A__syzkaller.appspot.com_bug-3Fextid-3Dded1696f6b50b615b630&d=DwIFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=NpjK8vgD4UmRGNWTNe6YFA-AJTZp9VlD0oMFvKFV25I&s=jeCou6OWbHHIf190FBGsLr1wGsDvBWlpKnfNMmqGIqI&e=
>> compiler:       clang version 8.0.0 (trunk 343298)
>> syz repro:      https://urldefense.proofpoint.com/v2/url?u=https-3A__syzkaller.appspot.com_x_repro.syz-3Fx-3D15ce62f5400000&d=DwIFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=NpjK8vgD4UmRGNWTNe6YFA-AJTZp9VlD0oMFvKFV25I&s=PVi1m-uNP3XRO4XbDZJicGiqXdVmOPFDxCP20jmXuAs&e=
>> C reproducer:   https://urldefense.proofpoint.com/v2/url?u=https-3A__syzkaller.appspot.com_x_repro.c-3Fx-3D174efca3400000&d=DwIFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=NpjK8vgD4UmRGNWTNe6YFA-AJTZp9VlD0oMFvKFV25I&s=XzvJtd3__2LEBvhAi4F6RTLbxV9mELkY07bMDSDLRMc&e=
>> 
>> IMPORTANT: if you fix the bug, please add the following tag to the commit:
>> Reported-by: syzbot+ded1696f6b50b615b630@xxxxxxxxxxxxxxxxxxxxxxxxx
>> 
>> ==================================================================
>> BUG: KMSAN: kernel-infoleak in __copy_to_user include/linux/uaccess.h:121
>> [inline]
>> BUG: KMSAN: kernel-infoleak in __kvm_write_guest_page
>> arch/x86/kvm/../../../virt/kvm/kvm_main.c:1849 [inline]
>> BUG: KMSAN: kernel-infoleak in kvm_vcpu_write_guest_page+0x39a/0x510
>> arch/x86/kvm/../../../virt/kvm/kvm_main.c:1870
>> CPU: 0 PID: 7918 Comm: syz-executor542 Not tainted 4.19.0+ #77
>> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
>> Google 01/01/2011
>> Call Trace:
>>  __dump_stack lib/dump_stack.c:77 [inline]
>>  dump_stack+0x32d/0x480 lib/dump_stack.c:113
>>  kmsan_report+0x1a2/0x2e0 mm/kmsan/kmsan.c:911
>>  kmsan_internal_check_memory+0x34c/0x430 mm/kmsan/kmsan.c:991
>>  kmsan_copy_to_user+0x85/0xe0 mm/kmsan/kmsan_hooks.c:552
>>  __copy_to_user include/linux/uaccess.h:121 [inline]
>>  __kvm_write_guest_page arch/x86/kvm/../../../virt/kvm/kvm_main.c:1849
>> [inline]
>>  kvm_vcpu_write_guest_page+0x39a/0x510
>> arch/x86/kvm/../../../virt/kvm/kvm_main.c:1870
>>  nested_release_vmcs12 arch/x86/kvm/vmx.c:8441 [inline]
>>  handle_vmptrld+0x2384/0x26b0 arch/x86/kvm/vmx.c:8907
>>  vmx_handle_exit+0x1e81/0xbac0 arch/x86/kvm/vmx.c:10128
>>  vcpu_enter_guest arch/x86/kvm/x86.c:7667 [inline]
>>  vcpu_run arch/x86/kvm/x86.c:7730 [inline]
>>  kvm_arch_vcpu_ioctl_run+0xac32/0x11d80 arch/x86/kvm/x86.c:7930
>>  kvm_vcpu_ioctl+0xfb1/0x1f90 arch/x86/kvm/../../../virt/kvm/kvm_main.c:2590
>>  do_vfs_ioctl+0xf77/0x2d30 fs/ioctl.c:46
>>  ksys_ioctl fs/ioctl.c:702 [inline]
>>  __do_sys_ioctl fs/ioctl.c:709 [inline]
>>  __se_sys_ioctl+0x1da/0x270 fs/ioctl.c:707
>>  __x64_sys_ioctl+0x4a/0x70 fs/ioctl.c:707
>>  do_syscall_64+0xcf/0x110 arch/x86/entry/common.c:291
>>  entry_SYSCALL_64_after_hwframe+0x63/0xe7
>> RIP: 0033:0x44b6e9
>> Code: e8 dc e6 ff ff 48 83 c4 18 c3 0f 1f 80 00 00 00 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 0f 83 2b ff fb ff c3 66 2e 0f 1f 84 00 00 00 00
>> RSP: 002b:00007f096b292ce8 EFLAGS: 00000206 ORIG_RAX: 0000000000000010
>> RAX: ffffffffffffffda RBX: 00000000006e3c48 RCX: 000000000044b6e9
>> RDX: 0000000000000000 RSI: 000000000000ae80 RDI: 0000000000000005
>> RBP: 00000000006e3c40 R08: 0000000000000000 R09: 0000000000000000
>> R10: 0000000000000000 R11: 0000000000000206 R12: 00000000006e3c4c
>> R13: 00007ffd978aeb2f R14: 00007f096b2939c0 R15: 00000000006e3d4c
>> 
>> Uninit was created at:
>>  kmsan_save_stack_with_flags mm/kmsan/kmsan.c:252 [inline]
>>  kmsan_internal_poison_shadow+0xc8/0x1e0 mm/kmsan/kmsan.c:177
>>  kmsan_kmalloc+0x98/0x110 mm/kmsan/kmsan_hooks.c:104
>>  __kmalloc+0x14c/0x4d0 mm/slub.c:3789
>>  kmalloc include/linux/slab.h:518 [inline]
>>  enter_vmx_operation+0x980/0x1a90 arch/x86/kvm/vmx.c:8278
>>  vmx_set_nested_state+0xc3a/0x1530 arch/x86/kvm/vmx.c:14045
>>  kvm_arch_vcpu_ioctl+0x4fc9/0x73a0 arch/x86/kvm/x86.c:4057
>>  kvm_vcpu_ioctl+0xca3/0x1f90 arch/x86/kvm/../../../virt/kvm/kvm_main.c:2742
>>  do_vfs_ioctl+0xf77/0x2d30 fs/ioctl.c:46
>>  ksys_ioctl fs/ioctl.c:702 [inline]
>>  __do_sys_ioctl fs/ioctl.c:709 [inline]
>>  __se_sys_ioctl+0x1da/0x270 fs/ioctl.c:707
>>  __x64_sys_ioctl+0x4a/0x70 fs/ioctl.c:707
>>  do_syscall_64+0xcf/0x110 arch/x86/entry/common.c:291
>>  entry_SYSCALL_64_after_hwframe+0x63/0xe7
>> 
>> Bytes 1000-4095 of 4096 are uninitialized
>> Memory access of size 4096 starts at ffff8801b5157000
>> ==================================================================
> This appears to be a real bug in KVM.
> Please see a simplified reproducer attached.

If I understand the above trace correctly:
1) When guest enters VMX operation, cached_vmcs12 is allocated in enter_vmx_operation() by kmalloc()
2) When guest will execute VMPTRLD, handle_vmptrld() will first release current vmcs12 by nested_release_vmcs12()
    and only then copy newly loaded vmcs12 to cached_vmcs12 by memcpy().
3) The bug theoretically is that nested_release_vmcs12() copies entire cached_vmcs12 to guest memory which is
     not initialised in case this is the very first VMPTRLD executed by guest.
     But this case should not happen because nested_release_vmcs12() starts by checking if current_vmptr is equal to -1
     and vmx_create_vcpu() set current_vmptr to -1ull.
     
However, one case where (3) could happen is by vmx_set_nested_state().
In case kvm_state->vmx.vmcs_pa is valid, set_current_vmptr() is called which sets current_vmptr.
However, it is possible that copy_from_user() which copies into cached_vmcs12 will fail.
Thus, leaving us with current_vmptr != -1 but with uninitialised cached_vmcs12.
Therefore, next VMPTRLD will indeed leak into guest the uninitialised content of cached_vmcs12.

What I described seems to be a bug to me (correct me if Iâm wrong) but it doesnât seem to be the
same issue described here as we can see that attached reproducer donât use KVM_SET_NESTED_STATE ioctl.

-Liran