qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
* Re: [Qemu-devel] freeze host when injecting NMIs in the guest, at least in 4.4-rc4+
       [not found] ` <20151210162733.GA6587@pd.tnic>
@ 2015-12-10 16:44   ` Borislav Petkov
  2015-12-10 16:49     ` Paolo Bonzini
  0 siblings, 1 reply; 4+ messages in thread
From: Borislav Petkov @ 2015-12-10 16:44 UTC (permalink / raw)
  To: kvm; +Cc: Paolo Bonzini, lkml, qemu-devel

Yap,

this is clearly a qemu/kvm issue. Lemme remove ext4 folks from CC. So
here's what happens:

I boot a kvm guest, connect to its monitor (qemu is started with
"-monitor pty") and on the monitor I issue a couple of times the "nmi"
command. It doesn't explode immediately but it happens pretty often and
when it happens, the *host*(!) freezes with some nasty corruption, see
below.

Thoughts, suggestions, ideas?

...
[   13.564192] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[  111.524485] kvm: zapping shadow pages for mmio generation wraparound
[  166.414836] EXT4-fs (sdb1): mounted filesystem with ordered data mode. Opts: (null)
[  172.408126] kvm: zapping shadow pages for mmio generation wraparound
[  200.912827] BUG: unable to handle kernel NULL pointer dereference at           (null)
[  200.920716] IP: [<          (null)>]           (null)
[  200.925796] PGD 0 
[  200.927833] Oops: 0010 [#1] PREEMPT SMP 
[  200.931258] Modules linked in: binfmt_misc ipv6 vfat fat fuse dm_crypt dm_mod kvm_amd kvm irqq
[  200.931274] CPU: 0 PID: 3936 Comm: qemu-system-x86 Not tainted 4.4.0-rc4+ #1
[  200.931275] Hardware name: To be filled by O.E.M. To be filled by O.E.M./M5A97 EVO R2.0, BIOS3
[  200.931276] task: ffff880037ba0000 ti: ffff8804227d8000 task.ti: ffff8804227d8000
[  200.931277] RIP: 0010:[<0000000000000000>]  [<          (null)>]           (null)
[  200.931278] RSP: 0018:ffff8804227dbcb0  EFLAGS: 00010292
[  200.931278] RAX: ffff880037ba0000 RBX: 0000000000000292 RCX: 0000000000000003
[  200.931279] RDX: ffff880037ba0750 RSI: 0000000000000000 RDI: ffff880037ba0001
[  200.931280] RBP: ffff8804227dbd08 R08: 0000000000000001 R09: 00000011f89b6000
[  200.931280] R10: ffffffff825c4fd0 R11: 0000000000000000 R12: 0000000000000000
[  200.931281] R13: 0000000000000000 R14: 0000000000000001 R15: 0000000000000000
[  200.931282] FS:  00007fec0faba700(0000) GS:ffff88042c600000(0000) knlGS:0000000000000000
[  200.931283] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[  200.931283] CR2: 0000000000000000 CR3: 00000000b6566000 CR4: 00000000000406f0
[  200.931283] Stack:
[  200.931285]  ffffffff8121004e ffff880400000000 ffffffff00000000 0000000000000292
[  200.931287]  00000000a0305d3c ffff880414612280 ffff880414612210 ffff8804227dbe90
[  200.931288]  ffff880414612130 ffff880037ba0000 ffffffff8121004e ffff8804227dbd90
[  200.931288] Call Trace:
[  200.931293]  [<ffffffff8121004e>] ? ext4_file_write_iter+0x8e/0x470
[  200.931295]  [<ffffffff8121004e>] ? ext4_file_write_iter+0x8e/0x470
[  200.931298]  [<ffffffff816de106>] mutex_lock_nested+0x66/0x3d0
[  200.931299]  [<ffffffff8121004e>] ? ext4_file_write_iter+0x8e/0x470
[  200.931301]  [<ffffffff8121004e>] ? ext4_file_write_iter+0x8e/0x470
[  200.931303]  [<ffffffff810a916a>] ? __lock_acquire+0x55a/0x1f80
[  200.931305]  [<ffffffff8121004e>] ext4_file_write_iter+0x8e/0x470
[  200.931307]  [<ffffffff811ac78e>] ? __fget+0x11e/0x210
[  200.931309]  [<ffffffff810a2cb4>] ? percpu_down_read+0x44/0x90
[  200.931311]  [<ffffffff8118d47d>] __vfs_write+0xad/0xe0
[  200.931313]  [<ffffffff8118dcd5>] vfs_write+0xb5/0x1a0
[  200.931315]  [<ffffffff8118e912>] SyS_write+0x52/0xc0
[  200.931317]  [<ffffffff816e2d5b>] entry_SYSCALL_64_fastpath+0x16/0x6f
[  200.931319] Code:  Bad RIP value.
[  200.931320] RIP  [<          (null)>]           (null)
[  200.931321]  RSP <ffff8804227dbcb0>
[  200.931321] CR2: 0000000000000000
[  200.939353] ---[ end trace 70675f54590d7755 ]---
[  200.939370] note: qemu-system-x86[3936] exited with preempt_count 1


-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [Qemu-devel] freeze host when injecting NMIs in the guest, at least in 4.4-rc4+
  2015-12-10 16:44   ` [Qemu-devel] freeze host when injecting NMIs in the guest, at least in 4.4-rc4+ Borislav Petkov
@ 2015-12-10 16:49     ` Paolo Bonzini
  2015-12-10 16:53       ` Borislav Petkov
  0 siblings, 1 reply; 4+ messages in thread
From: Paolo Bonzini @ 2015-12-10 16:49 UTC (permalink / raw)
  To: Borislav Petkov, kvm; +Cc: lkml, qemu-devel



On 10/12/2015 17:44, Borislav Petkov wrote:
> Yap,
> 
> this is clearly a qemu/kvm issue. Lemme remove ext4 folks from CC. So
> here's what happens:
> 
> I boot a kvm guest, connect to its monitor (qemu is started with
> "-monitor pty") and on the monitor I issue a couple of times the "nmi"
> command. It doesn't explode immediately but it happens pretty often and
> when it happens, the *host*(!) freezes with some nasty corruption, see
> below.
> 
> Thoughts, suggestions, ideas?

Can you try it on Intel?

Paolo

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [Qemu-devel] freeze host when injecting NMIs in the guest, at least in 4.4-rc4+
  2015-12-10 16:49     ` Paolo Bonzini
@ 2015-12-10 16:53       ` Borislav Petkov
  2015-12-10 17:34         ` Paolo Bonzini
  0 siblings, 1 reply; 4+ messages in thread
From: Borislav Petkov @ 2015-12-10 16:53 UTC (permalink / raw)
  To: Paolo Bonzini; +Cc: lkml, kvm, qemu-devel

On Thu, Dec 10, 2015 at 05:49:10PM +0100, Paolo Bonzini wrote:
> Can you try it on Intel?

Just did, there it splats even when booting the guest, without even
injecting NMIs:

[  113.233992] ===============================
[  113.238192] [ INFO: suspicious RCU usage. ]
[  113.242393] 4.4.0-rc4+ #1 Not tainted
[  113.246056] -------------------------------
[  113.250257] arch/x86/kvm/trace.h:29 suspicious rcu_dereference_check() usage!
[  113.257400] 
               other info that might help us debug this:

[  113.265432] 
               RCU used illegally from idle CPU!
               rcu_scheduler_active = 1, debug_locks = 0
[  113.276321] RCU used illegally from extended quiescent state!
[  113.282083] 1 lock held by qemu-system-x86/2275:
[  113.286711]  #0:  (&vcpu->mutex){+.+.+.}, at: [<ffffffffa01d472c>] vcpu_load+0x1c/0x80 [kvm]
[  113.295270] 
               stack backtrace:
[  113.299644] CPU: 4 PID: 2275 Comm: qemu-system-x86 Not tainted 4.4.0-rc4+ #1
[  113.306706] Hardware name: Dell Inc. Precision T3600/0PTTT9, BIOS A13 05/11/2014
[  113.314122]  0000000000000001 ffff88042f263d28 ffffffff813c2cfc ffff880432d53000
[  113.321575]  ffff88042f263d58 ffffffff810c5157 ffff88042f268000 0000000000000000
[  113.329027]  0000000000000000 0000000000000001 ffff88042f263e10 ffffffffa01ee7c8
[  113.336483] Call Trace:
[  113.338939]  [<ffffffff813c2cfc>] dump_stack+0x4e/0x82
[  113.344098]  [<ffffffff810c5157>] lockdep_rcu_suspicious+0xe7/0x120
[  113.350385]  [<ffffffffa01ee7c8>] kvm_arch_vcpu_ioctl_run+0x16f8/0x19c0 [kvm]
[  113.357544]  [<ffffffffa01ed1a0>] ? kvm_arch_vcpu_ioctl_run+0xd0/0x19c0 [kvm]
[  113.364695]  [<ffffffffa01d4b32>] kvm_vcpu_ioctl+0x342/0x700 [kvm]
[  113.370896]  [<ffffffff810c4a7d>] ? __lock_is_held+0x4d/0x70
[  113.376583]  [<ffffffff812351ae>] ? __fget+0xfe/0x200
[  113.381662]  [<ffffffff812291f1>] do_vfs_ioctl+0x301/0x550
[  113.387156]  [<ffffffff8123531a>] ? __fget_light+0x2a/0x90
[  113.392654]  [<ffffffff81229481>] SyS_ioctl+0x41/0x70
[  113.397739]  [<ffffffff8185cb36>] entry_SYSCALL_64_fastpath+0x16/0x7a
[  113.755008] kvm: zapping shadow pages for mmio generation wraparound

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [Qemu-devel] freeze host when injecting NMIs in the guest, at least in 4.4-rc4+
  2015-12-10 16:53       ` Borislav Petkov
@ 2015-12-10 17:34         ` Paolo Bonzini
  0 siblings, 0 replies; 4+ messages in thread
From: Paolo Bonzini @ 2015-12-10 17:34 UTC (permalink / raw)
  To: Borislav Petkov; +Cc: lkml, kvm, qemu-devel



On 10/12/2015 17:53, Borislav Petkov wrote:
> Just did, there it splats even when booting the guest, without even
> injecting NMIs:
> 
> [  113.233992] ===============================
> [  113.238192] [ INFO: suspicious RCU usage. ]
> [  113.242393] 4.4.0-rc4+ #1 Not tainted
> [  113.246056] -------------------------------
> [  113.250257] arch/x86/kvm/trace.h:29 suspicious rcu_dereference_check() usage!
> [  113.257400] 
>                other info that might help us debug this:
> 
> [  113.265432] 
>                RCU used illegally from idle CPU!
>                rcu_scheduler_active = 1, debug_locks = 0
> [  113.276321] RCU used illegally from extended quiescent state!
> [  113.282083] 1 lock held by qemu-system-x86/2275:
> [  113.286711]  #0:  (&vcpu->mutex){+.+.+.}, at: [<ffffffffa01d472c>] vcpu_load+0x1c/0x80 [kvm]
> [  113.295270] 
>                stack backtrace:
> [  113.299644] CPU: 4 PID: 2275 Comm: qemu-system-x86 Not tainted 4.4.0-rc4+ #1
> [  113.306706] Hardware name: Dell Inc. Precision T3600/0PTTT9, BIOS A13 05/11/2014
> [  113.314122]  0000000000000001 ffff88042f263d28 ffffffff813c2cfc ffff880432d53000
> [  113.321575]  ffff88042f263d58 ffffffff810c5157 ffff88042f268000 0000000000000000
> [  113.329027]  0000000000000000 0000000000000001 ffff88042f263e10 ffffffffa01ee7c8
> [  113.336483] Call Trace:
> [  113.338939]  [<ffffffff813c2cfc>] dump_stack+0x4e/0x82
> [  113.344098]  [<ffffffff810c5157>] lockdep_rcu_suspicious+0xe7/0x120
> [  113.350385]  [<ffffffffa01ee7c8>] kvm_arch_vcpu_ioctl_run+0x16f8/0x19c0 [kvm]
> [  113.357544]  [<ffffffffa01ed1a0>] ? kvm_arch_vcpu_ioctl_run+0xd0/0x19c0 [kvm]
> [  113.364695]  [<ffffffffa01d4b32>] kvm_vcpu_ioctl+0x342/0x700 [kvm]
> [  113.370896]  [<ffffffff810c4a7d>] ? __lock_is_held+0x4d/0x70
> [  113.376583]  [<ffffffff812351ae>] ? __fget+0xfe/0x200
> [  113.381662]  [<ffffffff812291f1>] do_vfs_ioctl+0x301/0x550
> [  113.387156]  [<ffffffff8123531a>] ? __fget_light+0x2a/0x90
> [  113.392654]  [<ffffffff81229481>] SyS_ioctl+0x41/0x70
> [  113.397739]  [<ffffffff8185cb36>] entry_SYSCALL_64_fastpath+0x16/0x7a
> [  113.755008] kvm: zapping shadow pages for mmio generation wraparound

Wow, this must have been there forever...

Paolo

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2015-12-10 17:34 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <20151210140404.GA3049@nazgul.tnic>
     [not found] ` <20151210162733.GA6587@pd.tnic>
2015-12-10 16:44   ` [Qemu-devel] freeze host when injecting NMIs in the guest, at least in 4.4-rc4+ Borislav Petkov
2015-12-10 16:49     ` Paolo Bonzini
2015-12-10 16:53       ` Borislav Petkov
2015-12-10 17:34         ` Paolo Bonzini

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).