Re: kvm: GPF in irq_bypass_unregister_consumer

[Wanpeng Li]

2017-01-03 17:32 GMT+08:00 Dmitry Vyukov <dvyukov@xxxxxxxxxx>:
> On Sat, Nov 12, 2016 at 2:21 AM, Dmitry Vyukov <dvyukov@xxxxxxxxxx> wrote:
>> Hello,
>>
>> The following programs triggers GPF in irq_bypass_unregister_consumer
>> if run in a parallel loop:
>> https://gist.githubusercontent.com/dvyukov/0d18d7b11659187ec3eab22285b4a67e/raw/c1c622ca26ebce0c7b77e3200970ca0f7792cb32/gistfile1.txt
>>
>> On commit 015ed9433be2b476ec7e2e6a9a411a56e3b5b035 (Nov 11).
>>
>> general protection fault: 0000 [#1] SMP KASAN
>> Dumping ftrace buffer:
>>    (ftrace buffer empty)
>> Modules linked in:
>> CPU: 2 PID: 599 Comm: kworker/2:1 Not tainted 4.9.0-rc4+ #41
>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>> Workqueue: kvm-irqfd-cleanup irqfd_shutdown
>> task: ffff88006cb51700 task.stack: ffff88006cbd8000
>> RIP: 0010:[<ffffffff831d3957>]  [<     inline     >] __list_del
>> include/linux/list.h:89
>> RIP: 0010:[<ffffffff831d3957>]  [<     inline     >] list_del
>> include/linux/list.h:107
>> RIP: 0010:[<ffffffff831d3957>]  [<ffffffff831d3957>]
>> irq_bypass_unregister_consumer+0x237/0x360 virt/lib/irqbypass.c:258
>> RSP: 0018:ffff88006cbdfb68  EFLAGS: 00010202
>> RAX: dffffc0000000000 RBX: 0000000000000000 RCX: ffff88006cb51f00
>> RDX: 0000000000000001 RSI: ffff88006cb51f50 RDI: 0000000000000008
>> RBP: ffff88006cbdfb90 R08: 0000000000000000 R09: 0000000000000000
>> R10: 0000000000000000 R11: 0000000000000000 R12: ffff88003caea638
>> R13: 0000000000000000 R14: ffffffff83ece9a0 R15: ffff88006bb0a530
>> FS:  0000000000000000(0000) GS:ffff88006e200000(0000) knlGS:0000000000000000
>> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> CR2: 00000000006e0000 CR3: 0000000003a1d000 CR4: 00000000000026e0
>> Stack:
>>  ffff88003caea4d0 ffff88003caea5f8 1ffff1000d97bf75 ffff88003caea5d0
>>  dffffc0000000000 ffff88006cbdfc30 ffffffff8106d144 0000000000000000
>>  0000000041b58ab3 ffffffff837c38d1 ffffffff8106d030 0000000000000246
>> Call Trace:
>>  [<ffffffff8106d144>] irqfd_shutdown+0x114/0x1a0
>> arch/x86/kvm/../../../virt/kvm/eventfd.c:145
>>  [<ffffffff8129375c>] process_one_work+0x9fc/0x1900 kernel/workqueue.c:2096
>>  [<ffffffff8129474f>] worker_thread+0xef/0x1480 kernel/workqueue.c:2230
>>  [<ffffffff812a7a94>] kthread+0x244/0x2d0 kernel/kthread.c:209
>>  [<ffffffff831ebf6a>] ret_from_fork+0x2a/0x40 arch/x86/entry/entry_64.S:433
>> Code: 48 89 fa 48 c1 ea 03 80 3c 02 00 0f 85 b7 00 00 00 49 8d 7d 08
>> 48 b8 00 00 00 00 00 fc ff df 49 8b 5c 24 08 48 89 fa 48 c1 ea 03 <80>
>> 3c 02 00 0f 85 9d 00 00 00 48 89 da 48 b8 00 00 00 00 00 fc
>> RIP  [<     inline     >] __list_del include/linux/list.h:89
>> RIP  [<     inline     >] list_del include/linux/list.h:107
>> RIP  [<ffffffff831d3957>] irq_bypass_unregister_consumer+0x237/0x360
>> virt/lib/irqbypass.c:258
>>  RSP <ffff88006cbdfb68>
>> ---[ end trace 986ec3b53e4e0338 ]---
>> Kernel panic - not syncing: Fatal exception
>> Dumping ftrace buffer:
>>    (ftrace buffer empty)
>> Kernel Offset: disabled
>> reboot: cpu_has_vmx: ecx=80a02021 1
>
>
> Still happens on c8b4ec8351d21da3299b045b37920e5cf5590793 (Jan 2).
>
> [  480.585892] irq bypass consumer (token ffff88003e205700)
> registration fails: -16
> [  480.686218] BUG: unable to handle kernel NULL pointer dereference
> at 0000000000000008
> [  480.686868] IP: irq_bypass_unregister_consumer+0x8d/0xd0
> [  480.686993] PGD 0
> [  480.686993]
> [  480.686993] Oops: 0002 [#1] SMP
> [  480.686993] Modules linked in:
> [  480.686993] CPU: 3 PID: 585 Comm: kworker/3:1 Not tainted 4.10.0-rc2+ #57
> [  480.686993] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
> BIOS Bochs 01/01/2011
> [  480.686993] Workqueue: kvm-irqfd-cleanup irqfd_shutdown
> [  480.686993] task: ffff88007fa3a8c0 task.stack: ffffc90001524000
> [  480.686993] RIP: 0010:irq_bypass_unregister_consumer+0x8d/0xd0
> [  480.686993] RSP: 0018:ffffc90001527e20 EFLAGS: 00010246
> [  480.686993] RAX: 0000000000000000 RBX: ffff88007dfa5170 RCX: 0000000000000004
> [  480.686993] RDX: 0000000000000000 RSI: ffff88007dfa50e8 RDI: 0000000000000000
> [  480.686993] RBP: ffffc90001527e28 R08: 0000000000000004 R09: ffff88007fd171c0
> [  480.686993] R10: 0000006feb214fc0 R11: 0000000000000000 R12: ffff88007dfa5070
> [  480.686993] R13: ffff88007fd1e300 R14: 00000000000000c0 R15: ffff88007dfa5150
> [  480.686993] FS:  0000000000000000(0000) GS:ffff88007fd00000(0000)
> knlGS:0000000000000000
> [  480.686993] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [  480.686993] CR2: 0000000000000008 CR3: 0000000001e0a000 CR4: 00000000000026e0
> [  480.686993] Call Trace:
> [  480.686993]  irqfd_shutdown+0x52/0x70
> [  480.686993]  process_one_work+0x149/0x3f0
> [  480.686993]  worker_thread+0x126/0x4a0
> [  480.686993]  ? __schedule+0x21c/0x660
> [  480.686993]  kthread+0xfc/0x130
> [  480.686993]  ? rescuer_thread+0x320/0x320
> [  480.686993]  ? kthread_park+0x90/0x90
> [  480.686993]  ret_from_fork+0x25/0x30
> [  480.686993] Code: 30 d7 f0 81 75 0e eb 1a 48 8b 3f 48 81 ff 30 d7
> f0 81 74 0e 48 39 57 10 75 ee 48 89 de e8 cc fb ff ff 48 8b 43 08 48
> 8b 13 31 ff <48> 89 42 08 48 89 10 48 b8 00 01 00 00 00 00 ad de 48 89
> 03 66
> [  480.686993] RIP: irq_bypass_unregister_consumer+0x8d/0xd0 RSP:
> ffffc90001527e20
> [  480.686993] CR2: 0000000000000008
> [  480.686993] ---[ end trace e5a8ddcdf160fccb ]---


  case 1:
    r[2] = syscall(__NR_open, "/dev/kvm", 0x40042ul, 0, 0, 0, 0, 0, 0);
    break;
  case 2:
    r[3] = execute_syscall(__NR_ioctl, r[2], 0xae01ul, 0x0ul, 0, 0, 0,
                           0, 0, 0);
    break;
  case 3:
    r[4] = execute_syscall(__NR_ioctl, r[3], 0xae41ul, 0x3fful, 0, 0, 0,
                           0, 0, 0);
    break;
  case 4:
    r[5] = execute_syscall(__NR_ioctl, r[4], 0xae9aul, 0, 0, 0, 0, 0, 0,
                           0);
    break;
  case 5:
    r[6] = execute_syscall(__NR_eventfd2, 0x8ul, 0x801ul, 0, 0, 0, 0, 0,
                           0, 0);

What's the meaning of parameter "0x40042ul", â0xae01ulâ etc here?

Regards,
Wanpeng Li