Hello,
We run CRIU tests on linux-next tree and today we found this issue.
CRIU tests are the set of small programs to check checkpoint/restore
of different primitives (files, sockets, signals, pipes, etc).
https://github.com/xemul/criu/tree/master/test
Each test is executed three times: without namespaces, in a set of all
namespaces except userns, in a set of all namespaces. When a test
passed the preparation tests, it sends a signal to an executer, and
then the executer dumps and restores tests processes, and sends a
signal to the test back to check that everything are restored
correctly.
===================== Run zdtm/transition/unix_sock in ns
======================
Start test
./unix_sock --pidfile=unix_sock.pid --outfile=unix_sock.out
--filename=unix_sock.test
Run criu dump
[ 57.647284] writing to auto_msgmni has no effect
[ 60.730380] criu (2023) used greatest stack depth: 11808 bytes left
Run criu restore
[ 60.993529] BUG: unable to handle kernel NULL pointer dereference at
(null)
[ 60.994221] IP: skb_queue_tail+0x2e/0x50
[ 60.994589] PGD 71070067
[ 60.994590] P4D 71070067
[ 60.994854] PUD 71071067
[ 60.995102] PMD 0
[ 60.995352]
[ 60.995694] Oops: 0002 [#1] SMP
[ 60.996004] CPU: 0 PID: 2053 Comm: unix_sock Not tainted
4.12.0-next-20170713 #6
[ 60.996706] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
BIOS 1.10.2-20170228_101828-anatol 04/01/2014
[ 60.997657] task: ffff880074748c80 task.stack: ffffc90000594000
[ 60.998208] RIP: 0010:skb_queue_tail+0x2e/0x50
[ 60.998614] RSP: 0018:ffffc90000597cf8 EFLAGS: 00010046
[ 60.999132] RAX: 0000000000000246 RBX: ffff88006f3fa0c8 RCX:
0000000000000000
[ 60.999797] RDX: 0000000000000000 RSI: 0000000000000246 RDI:
ffff88006f3fa0dc
[ 61.000455] RBP: ffffc90000597d10 R08: ffffc90000597e50 R09:
0000000000000000
[ 61.001114] R10: ffff880072daea00 R11: ffff88007d002d80 R12:
ffff880072daea00
[ 61.001772] R13: ffff88006f3fa0dc R14: ffff88006f3fa000 R15:
0000000000000001
[ 61.002451] FS: 0000000000000000(0000) GS:ffff88007fc00000(0063)
knlGS:00000000f7f7b380
[ 61.003198] CS: 0010 DS: 002b ES: 002b CR0: 0000000080050033
[ 61.003735] CR2: 0000000000000000 CR3: 000000007106f000 CR4:
00000000000006f0
[ 61.004393] DR0: 0000000000000000 DR1: 0000000000000000 DR2:
0000000000000000
[ 61.005050] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7:
0000000000000400
[ 61.005717] Call Trace:
[ 61.005952] unix_stream_sendmsg+0x1c1/0x380
[ 61.006345] sock_sendmsg+0x33/0x40
[ 61.006667] sock_write_iter+0x7d/0xc0
[ 61.007032] __vfs_write+0xcd/0x120
[ 61.007353] vfs_write+0xac/0x1a0
[ 61.007677] SyS_write+0x41/0xa0
[ 61.007996] do_fast_syscall_32+0x8b/0x15c
[ 61.008371] entry_SYSENTER_compat+0x4c/0x5b
[ 61.008781] RIP: 0023:0xf7f7faf9
[ 61.009082] RSP: 002b:00000000fffd62f8 EFLAGS: 00000246 ORIG_RAX:
0000000000000004
[ 61.009811] RAX: ffffffffffffffda RBX: 0000000000000004 RCX:
00000000fffd6738
[ 61.010453] RDX: 00000000000003e8 RSI: 00000000fffd63b8 RDI:
00000000fffd6749
[ 61.011116] RBP: 00000000fffd6b38 R08: 0000000000000000 R09:
0000000000000000
[ 61.011795] R10: 0000000000000000 R11: 0000000000000000 R12:
0000000000000000
[ 61.012378] R13: 0000000000000000 R14: 0000000000000000 R15:
0000000000000000
[ 61.013027] Code: e5 41 55 4c 8d 6f 14 41 54 53 48 89 fb 4c 89 ef 49
89 f4 e8 85 d3 21 00 48 8b 53 08 49 89 1c 24 4c 89 ef 48 89 c6 49 89 54
24 08 <4c> 89 22 83 43 10 01 4c 89 63 08 e8 22 d4 21 00 5b 41 5c 41 5d
[ 61.014778] RIP: skb_queue_tail+0x2e/0x50 RSP: ffffc90000597cf8
[ 61.015333] CR2: 0000000000000000
[ 61.015639] ---[ end trace efd0a4201d4b29fc ]---
The bug is easily (5/5) reproduced on next-20170713 with the following:
git clone https://github.com/xemul/criu.git
cd criu && git checkout criu-dev
COMPAT_TEST=y make -j5 zdtm
for i in `seq 1 2`; do ./test/zdtm.py run -t zdtm/transition/unix_sock
-f ns ; done
--
Dmitry
