Re: kvm: GPF in kvm_irq_map_gsi

[Dmitry Vyukov]

On Tue, May 31, 2016 at 11:56 AM, Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
>
>
> On 15/02/2016 14:30, Dmitry Vyukov wrote:
>>     *(uint32_t*)0x2000a6b9 = (uint32_t)0x3e;
>>     *(uint16_t*)0x2000a6bd = (uint16_t)0x8;
>>     *(uint8_t*)0x2000a6bf = (uint8_t)0x8d4;
>>     *(uint8_t*)0x2000a6c0 = (uint8_t)0xffffffffffff5fe9;
>>     *(uint8_t*)0x2000a6c1 = (uint8_t)0x80000001;
>>     *(uint8_t*)0x2000a6c2 = (uint8_t)0x0;
>>     *(uint8_t*)0x2000a6c3 = (uint8_t)0xbe2;
>>     *(uint8_t*)0x2000a6c4 = (uint8_t)0x9;
>>     *(uint8_t*)0x2000a6c5 = (uint8_t)0x7ff;
>>     *(uint8_t*)0x2000a6c6 = (uint8_t)0x1;
>>     *(uint8_t*)0x2000a6c7 = (uint8_t)0x1f;
>>     *(uint8_t*)0x2000a6c8 = (uint8_t)0x1d8;
>>     *(uint16_t*)0x2000a6c9 = (uint16_t)0x8;
>
> This field (.channels[0].count_load_time) should be uint64_t.  I
> understand that it's all random, but it makes it even harder to follow
> what's going on...

Thanks! Fixed in:
https://github.com/google/syzkaller/commit/1816c21f2bed9b6015f1af1c5423779d2c6e58ad