kvm: GPF in irq_bypass_unregister_consumer

[Dmitry Vyukov]

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

[Dmitry Vyukov]

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 ]---

[Wanpeng Li]

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", “0xae01ul” etc here?

Regards,
Wanpeng Li

[Dmitry Vyukov]

strace can shed some light on it:
$ strace -f -e ioctl ./a.out

[Wanpeng Li]

Thanks, I will look into this report.

Regards,
Wanpeng Li