Comment 15 for bug 1379340

I think this patch fixes the issue:

https://lkml.org/lkml/2014/9/22/211

Looking at the stacktrace:

[ 4.690909] divide error: 0000 [#1] SMP
[ 4.690909] Modules linked in: dm_crypt kvm_amd kvm serio_raw isofs crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 lrw gf128mul glue_helper ablk_helper cryptd psmouse floppy
[ 4.690909] CPU: 0 PID: 663 Comm: cloud-init Not tainted 3.13.0-40-generic #69-Ubuntu
[ 4.690909] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
[ 4.690909] task: ffff88001f373000 ti: ffff88001460a000 task.ti: ffff88001460a000
[ 4.690909] RIP: 0010:[<ffffffff8104ed58>] [<ffffffff8104ed58>] kvm_unlock_kick+0xa8/0x100
[ 4.690909] RSP: 0000:ffff88001fc03df0 EFLAGS: 00010046
[ 4.690909] RAX: 0000000000000005 RBX: 0000000000000000 RCX: 0000000000000001
[ 4.690909] RDX: ffffffff81eb1448 RSI: 0000000000000000 RDI: 0000000000000000
[ 4.690909] RBP: ffff88001fc03e10 R08: ffffffff81eb1440 R09: ffff880016000000
[ 4.690909] R10: 0000000000000006 R11: 561488f3089a6867 R12: ffffffff81fc66c0
[ 4.690909] R13: 0000000000000802 R14: 0000000000000001 R15: 00000000000000c2
[ 4.690909] FS: 00007fc269f46740(0000) GS:ffff88001fc00000(0000) knlGS:0000000000000000
[ 4.690909] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 4.690909] CR2: 00007fc2665de050 CR3: 000000001f50f000 CR4: 00000000000406f0
[ 4.690909] Stack:
[ 4.690909] 0000000000000046 0000000000000060 0000000000000046 0000000000000020
[ 4.690909] ffff88001fc03e20 ffffffff81718b53 ffff88001fc03e38 ffffffff817270da
[ 4.690909] ffffffff81fc66c0 ffff88001fc03e70 ffffffff8146de04 ffffffff81fc66c0
[ 4.690909] Call Trace:
[ 4.690909] <IRQ>
[ 4.690909] [<ffffffff81718b53>] __ticket_unlock_slowpath+0x24/0x34
[ 4.690909] [<ffffffff817270da>] _raw_spin_unlock_irqrestore+0x3a/0x40
[ 4.690909] [<ffffffff8146de04>] serial8250_handle_irq.part.14+0x84/0xb0
[ 4.690909] [<ffffffff8146de77>] serial8250_default_handle_irq+0x27/0x30
[ 4.690909] [<ffffffff8146ce73>] serial8250_interrupt+0x63/0xe0
[ 4.690909] [<ffffffff810bf97e>] handle_irq_event_percpu+0x3e/0x1d0
[ 4.690909] [<ffffffff810bfb4d>] handle_irq_event+0x3d/0x60
[ 4.690909] [<ffffffff810c25d7>] handle_edge_irq+0x77/0x130
[ 4.690909] [<ffffffff81015dbe>] handle_irq+0x1e/0x30
[ 4.690909] [<ffffffff8173205d>] do_IRQ+0x4d/0xc0
[ 4.690909] [<ffffffff8172772d>] common_interrupt+0x6d/0x6d
[ 4.690909] <EOI>
[ 4.690909] Code: 66 44 39 e8 75 bd 0f b6 35 36 27 e6 00 40 84 f6 75 2a 83 05 46 27 e6 00 01 48 c7 c0 8a b0 00 00 31 db 0f b7 0c 01 b8 05 00 00 00 <0f> 01 c1 0f 1f 44 00 00 5b 41 5c 41 5d 41 5e 5d c3 89 f0 31 c9
[ 4.690909] RIP [<ffffffff8104ed58>] kvm_unlock_kick+0xa8/0x100
[ 4.690909] RSP <ffff88001fc03df0>

Looking at the objdump we see we get a Divide Error on a vmcall instruction.
In addition we build our kernels with CONFIG_DEBUG_RODATA and PV locking.

static void kvm_kick_cpu(int cpu)
{
        int apicid;
        unsigned long flags = 0;

        apicid = per_cpu(x86_cpu_to_apicid, cpu);
ffffffff8104ed46: 48 c7 c0 8a b0 00 00 mov $0xb08a,%rax

static inline long kvm_hypercall2(unsigned int nr, unsigned long p1,
                                  unsigned long p2)
{
        long ret;
        asm volatile(KVM_HYPERCALL
ffffffff8104ed4d: 31 db xor %ebx,%ebx
        kvm_hypercall2(KVM_HC_KICK_CPU, flags, apicid);
ffffffff8104ed4f: 0f b7 0c 01 movzwl (%rcx,%rax,1),%ecx
ffffffff8104ed53: b8 05 00 00 00 mov $0x5,%eax
ffffffff8104ed58: 0f 01 c1 vmcall
ffffffff8104ed5b: 0f 1f 44 00 00 nopl 0x0(%rax,%rax,1)
                        add_stats(RELEASED_SLOW_KICKED, 1);
                        kvm_kick_cpu(cpu);
                        break;
                }
        }
}