[SOLVED] Kernel 3.10.17 BUG with sit module?

Youmu · 07-18-2015, 03:10 PM

Hello everyone!
Recently I have encountered a rather strange behavior of my linux machine. Syslog says it is maybe a kernel bug. See it yourself:
*****
1)I was playing with ip tunnels.

Code:

ip tunnel add foo mode sit remote 192.168.1.42

Created an ip tunnel for me. What it additionally creates is a sit0 interface so in ifconfig it looks like:

Code:

foo: flags=144<POINTOPOINT,NOARP>  mtu 1480
        unspec 00-00-00-00-33-35-00-64-00-00-00-00-00-00-00-00  txqueuelen 0  (UNSPEC)
        RX packets 0  bytes 0 (0.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 0  bytes 0 (0.0 B)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 0  (Local Loopback)
        RX packets 0  bytes 0 (0.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 0  bytes 0 (0.0 B)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

sit0: flags=128<NOARP>  mtu 1480
        unspec 00-00-00-00-33-35-00-64-00-00-00-00-00-00-00-00  txqueuelen 0  (UNSPEC)
        RX packets 0  bytes 0 (0.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 0  bytes 0 (0.0 B)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

2)I wanted to delete these interfaces completely, so I typed

Code:

ip link set foo down
ip tunnel del foo

And foo tunnel was gone
3)But sit0 tunnel was there. I could not delete is because it was a kernel's module. So I typed:

Code:

sudo rmmod sit

To get rid of sit module and sit interface.
And guess what was next? - ifconfig command hangs, all networking hung, ip command hung...
4)I have stopped X server and rebooted
5)THAT was in my syslog:

Code:

Jul 18 23:53:30 darkstar kernel: [  274.043055] ------------[ cut here ]------------
Jul 18 23:53:30 darkstar kernel: [  274.043120] kernel BUG at net/core/dev.c:4914!
Jul 18 23:53:30 darkstar kernel: [  274.043165] invalid opcode: 0000 [#1] SMP 
Jul 18 23:53:30 darkstar kernel: [  274.043213] Modules linked in:  sit(-) tunnel4 ip_tunnel bbswitch(O) ipv6 cpufreq_ondemand lp ppdev parport_pc parport btusb bluetooth fuse snd_hda_codec_hdmi snd_hda_codec_realtek joydev uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_core cdc_ether usbnet videodev mii i2c_dev gpio_ich rtsx_pci_sdmmc rtsx_pci_ms mmc_core memstick intel_powerclamp coretemp kvm_intel kvm rtl8192ce crc32_pclmul crc32c_intel rtlwifi snd_hda_intel rtl8192c_common snd_hda_codec ghash_clmulni_intel snd_hwdep microcode snd_pcm mac80211 i915 snd_page_alloc psmouse intel_agp cfg80211 snd_timer intel_gtt drm_kms_helper alx evdev rfkill serio_raw snd i2c_i801 drm mei_me ehci_pci mdio soundcore rtsx_pci lpc_ich mei agpgart xhci_hcd mperf ehci_hcd i2c_algo_bit i2c_core thermal video wmi processor thermal_sys freq_table hwmon battery button ac loop [last unloaded: nvidia]
Jul 18 23:53:30 darkstar kernel: [  274.044175] CPU: 0 PID: 1676 Comm: rmmod Tainted: P           O 3.10.17 #2
Jul 18 23:53:30 darkstar kernel: [  274.044237] Hardware name: Quanta JW6H/JW6H, BIOS DR151 12/10/2012
Jul 18 23:53:30 darkstar kernel: [  274.044297] task: ffff8801504bdbd0 ti: ffff88014e722000 task.ti: ffff88014e722000
Jul 18 23:53:30 darkstar kernel: [  274.044363] RIP: 0010:[<ffffffff81a867fe>]  [<ffffffff81a867fe>] rollback_registered_many+0x1de/0x220
Jul 18 23:53:30 darkstar kernel: [  274.044458] RSP: 0018:ffff88014e723df0  EFLAGS: 00010206
Jul 18 23:53:30 darkstar kernel: [  274.044506] RAX: 0000000000000004 RBX: ffff88014e723dc0 RCX: ffff88014e723e30
Jul 18 23:53:30 darkstar kernel: [  274.044569] RDX: ffff880159d9d070 RSI: ffff88014e723e30 RDI: ffff88014e723e30
Jul 18 23:53:30 darkstar kernel: [  274.044631] RBP: ffff88014e723e08 R08: ffffffff820d3030 R09: 0000000000000001
Jul 18 23:53:30 darkstar kernel: [  274.044694] R10: 0000000000000000 R11: 0000000000000001 R12: ffff880159d9d000
Jul 18 23:53:30 darkstar kernel: [  274.044758] R13: ffff88014e723e30 R14: 0000000000000018 R15: ffff880158e96400
Jul 18 23:53:30 darkstar kernel: [  274.044822] FS:  00007f5baa73b740(0000) GS:ffff88015f200000(0000) knlGS:0000000000000000
Jul 18 23:53:30 darkstar kernel: [  274.044893] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Jul 18 23:53:30 darkstar kernel: [  274.044944] CR2: 00000000006fe030 CR3: 000000015067d000 CR4: 00000000001407f0
Jul 18 23:53:30 darkstar kernel: [  274.045007] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
Jul 18 23:53:30 darkstar kernel: [  274.045069] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Jul 18 23:53:30 darkstar kernel: [  274.045131] Stack:
Jul 18 23:53:30 darkstar kernel: [  274.045153]  ffff88014e723e30 ffff88014e723e30 0000000000000080 ffff88014e723e20
Jul 18 23:53:30 darkstar kernel: [  274.045231]  ffffffff81a86897 0000000000000000 ffff88014e723e68 ffffffffa04fa48a
Jul 18 23:53:30 darkstar kernel: [  274.045308]  ffff880159d9d070 ffff880159d9d070 ffffffff820d3000 ffff88014e723ea8
Jul 18 23:53:30 darkstar kernel: [  274.045386] Call Trace:
Jul 18 23:53:30 darkstar kernel: [  274.045417]  [<ffffffff81a86897>] unregister_netdevice_many+0x17/0x70
Jul 18 23:53:30 darkstar kernel: [  274.045479]  [<ffffffffa04fa48a>] sit_exit_net+0xba/0xe0 [sit]
Jul 18 23:53:30 darkstar kernel: [  274.045536]  [<ffffffff81a81b89>] ops_exit_list.isra.3+0x39/0x60
Jul 18 23:53:30 darkstar kernel: [  274.045594]  [<ffffffff81a81cda>] unregister_pernet_operations+0x9a/0xe0
Jul 18 23:53:30 darkstar kernel: [  274.045657]  [<ffffffff81a81d7a>] ? unregister_pernet_device+0x2a/0x50
Jul 18 23:53:30 darkstar kernel: [  274.045719]  [<ffffffffa04fc803>] ? sit_cleanup+0x2d/0x34 [sit]
Jul 18 23:53:30 darkstar kernel: [  274.045778]  [<ffffffff810b56ff>] ? SyS_delete_module+0x14f/0x2b0
Jul 18 23:53:30 darkstar kernel: [  274.045840]  [<ffffffff810304ae>] ? do_page_fault+0xe/0x10
Jul 18 23:53:30 darkstar kernel: [  274.045894]  [<ffffffff81b6f786>] ? system_call_fastpath+0x1a/0x1f
Jul 18 23:53:30 darkstar kernel: [  274.045951] Code: 49 39 d5 48 8d 42 90 74 1b 0f 1f 40 00 48 8b 90 98 03 00 00 65 ff 0a 48 8b 50 70 49 39 d5 48 8d 42 90 75 e9 5b 41 5c 41 5d 5d c3 <0f> 0b be 59 13 00 00 48 c7 c7 83 7b ed 81 e8 ff 07 5d ff eb 89 
Jul 18 23:53:30 darkstar kernel: [  274.046363] RIP  [<ffffffff81a867fe>] rollback_registered_many+0x1de/0x220
Jul 18 23:53:30 darkstar kernel: [  274.046431]  RSP <ffff88014e723df0>
Jul 18 23:53:30 darkstar kernel: [  274.058018] ---[ end trace 0473e8a3307b3f0b ]---

*****

Code:

Linux darkstar 3.10.17 #2 SMP Wed Oct 23 16:34:38 CDT 2013 x86_64 Intel(R) Core(TM) i7-3517U CPU @ 1.90GHz GenuineIntel GNU/Linux

I will be very grateful if someone just explained to me what was that?
Thanks in advance!
-Youmu

Didier Spaier · 07-18-2015, 03:37 PM

If there's a bug in the kernel it's not Slackware related. But before reporting it to the kernel developers, upgrade to the last 3.10.nn version, at time of writing 3.10.84 according to https://www.kernel.org/

On the other hand why not just do this?

Code:

ifconfig sit0 down

Youmu · 07-18-2015, 03:42 PM

Quote:

Originally Posted by Didier Spaier

If there's a bug in the kernel its not Slackware related. But before reporting it to the kernel developers, upgrade to the last 3.10.nn version, at time of writing 3.10.84 according to https://www.kernel.org/

On the other hand why not just do this?

Code:

ifconfig sit0 down

The interface remains.
On the other hand, sudo ip del sit0 shows that

Code:

delete tunnel "sit0" failed: Operation not permitted

Which is very strange itself.
I don't want to compile and test new kernel - don't have time for that now. Maybe just wait when the right person comes and ask him to test?

Didier Spaier · 07-18-2015, 03:53 PM

Quote:

Originally Posted by Youmu

Maybe just wait when the right person comes and ask him to test?

Why not?

Meanwhile, the wrong person will just ask why keeping the interface hurts, as long as it is down. But you have no obligation to answer

Oh, and just in case: use "su" to get root's privileges instead of sudo. This is Slackware

Youmu · 07-18-2015, 04:17 PM

Quote:

Meanwhile, the wrong person will just ask why keeping the interface hurts, as long as it is down. But you have no obligation to answer

No man, you just won't understand =^_^=

55020 · 07-18-2015, 05:02 PM

With that 'Tainted: P' flag, no 'right person' is going to pay any attention. You'll have to do without nvidia or vmware long enough to grab another panic from the log.

If you don't want to build your own 3.10.84, you can grab the 3.18.11 kernel packages from -current and use them with 14.1 (it works for me).

Youmu · 07-18-2015, 05:16 PM

Quote:

Originally Posted by 55020

With that 'Tainted: P' flag, no 'right person' is going to pay any attention. You'll have to do without nvidia or vmware long enough to grab another panic from the log.

If you don't want to build your own 3.10.84, you can grab the 3.18.11 kernel packages from -current and use them with 14.1 (it works for me).

Quote:

When the kernel is tainted, it means that it is in a state that is unsupported by the community. Most kernel developers will ignore bug reports involving tainted kernels, and...

Okey-dokey
But at last, have the problem repeated for you?

mancha · 07-18-2015, 05:35 PM

Quote:

Originally Posted by Youmu

I will be very grateful if someone just explained to me what was that? Thanks in advance! -Youmu

Hi. I just tried it on 3.10.84 (32-bit) and can't repro: after removing the sit module, ifconfig works and nothing hangs.

Rather than invest time into debugging your oops, I prefer if you test the .84 release and report back.

Note: I also removed the tunnel4 and ip_tunnel modules but only after my above test was done.

--mancha

mancha · 07-18-2015, 05:51 PM

Hi Youmu.

I looked into this a bit more and confirm the bug in 3.10.17: on sit unload it free's the fallback device twice.

This has been fixed in 3.10.29 so any kernel from that point onwards should work fine for you. This serves as another good reminder
for Slackware 14.1 users to keep their kernels relatively up-to-date in the 3.10 tree (3.10.17 is ancient).

--mancha

Youmu · 07-18-2015, 06:19 PM

Quote:

Originally Posted by mancha

Hi Youmu.

I looked into this a bit more and confirm the bug in 3.10.17: on sit unload it free's the fallback device twice.

This has been fixed in 3.10.29 so any kernel from that point onwards should work fine for you. This serves as another good reminder
for Slackware 14.1 users to keep their kernels relatively up-to-date in the 3.10 tree (3.10.17 is ancient).

--mancha

Thanks!
I am compiling my new kernel now.
-Youmu

Youmu · 07-19-2015, 06:19 AM

Youmu here.
Kernel 3.10.84 works perfectly, no sit bug here.

mancha · 07-19-2015, 09:40 AM

Quote:

Originally Posted by Youmu

Youmu here.
Kernel 3.10.84 works perfectly, no sit bug here.

Excellent. Thanks for the update.

--mancha