Hi,

On 21.09.2021 11:36, Michael Tremer wrote:
> Hello,
> 
> Could we report this to the right place with the Linux kernel community?

Where would be this "right place"?

https://bugzilla.kernel.org/ or https://lkml.org/?

It happened again today with Core 161 (running 'suricata 5.0.8').

=> see Attachment

Best,
Matthias

> This seems to be a long-standing problem which wouldn’t be very difficult to solve.
> 
> -Michael
> 
>> On 20 Sep 2021, at 16:16, Matthias Fischer <matthias.fischer(a)ipfire.org> wrote:
>> 
>> Hi,
>> 
>> today it happened again - at 1:18am:
>> 
>> ***SNIP***
>> ipfire kernel: refcount_t: underflow; use-after-free.
>> ipfire kernel: WARNING: CPU: 1 PID: 30228 at lib/refcount.c:28
>> refcount_warn_saturate+0xa6/0xf0
>> ipfire kernel: Modules linked in: xt_REDIRECT nfnetlink_queue xt_NFQUEUE
>> xt_MASQUERADE ccm cpufreq_conservative cpufreq_ondemand xt_geoip(O)
>> xt_ipp2p(O) compat_xtables(O) xt_hashlimit xt_mac xt_multiport xt_mark
>> xt_policy xt_TCPMSS xt_conntrack xt_comment ipt_REJECT nf_reject_ipv4
>> xt_LOG xt_limit nf_log_ipv4 nf_log_common iptable_raw iptable_mangle
>> iptable_filter vfat fat rt2800usb rt2x00usb rt2800lib rt2x00lib
>> sch_fq_codel x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel
>> mac80211 at24 kvm regmap_i2c iTCO_wdt iTCO_vendor_support irqbypass
>> crct10dif_pclmul crc32_pclmul cfg80211 ghash_clmulni_intel i2c_i801
>> pcspkr i2c_smbus lpc_ich rfkill mfd_core r8169 realtek libarc4
>> ir_rc6_decoder i2c_algo_bit fb_sys_fops snd_hda_codec_realtek
>> syscopyarea rc_rc6_mce sysfillrect snd_hda_codec_generic sysimgblt
>> nuvoton_cir i2c_core ledtrig_audio rc_core snd_hda_intel
>> snd_intel_dspcfg snd_hda_codec snd_hda_core snd_hwdep acpi_pad snd_pcm
>> snd_timer snd soundcore lp parport_pc parport video
>> ipfire kernel: CPU: 1 PID: 30228 Comm: W-NFQ#0 Tainted: G           O
>>   5.10.55-ipfire #1
>> ipfire kernel: Hardware name: To be filled by O.E.M. To be filled by
>> O.E.M./CRESCENTBAY, BIOS 5.011 04/13/2016
>> ipfire kernel: RIP: 0010:refcount_warn_saturate+0xa6/0xf0
>> ipfire kernel: Code: 05 ba 60 1f 01 01 e8 af a3 52 00 0f 0b c3 80 3d a8
>> 60 1f 01 00 75 95 48 c7 c7 b8 fc 11 ab c6 05 98 60 1f 01 01 e8 90 a3 52
>> 00 <0f> 0b c3 80 3d 87 60 1f 01 00 0f 85 72 ff ff ff 48 c7 c7 10 fd 11
>> ipfire kernel: RSP: 0018:ffffbaba82823950 EFLAGS: 00010282
>> ipfire kernel: RAX: 0000000000000000 RBX: ffff998e77a25500 RCX:
>> 0000000000000027
>> ipfire kernel: RDX: ffff998f47318968 RSI: 0000000000000001 RDI:
>> ffff998f47318960
>> ipfire kernel: RBP: ffffbaba82823a50 R08: 0000000000000000 R09:
>> ffffbaba82823788
>> ipfire kernel: R10: ffffbaba82823780 R11: ffffffffab533dc8 R12:
>> ffff998e77a25500
>> ipfire kernel: R13: ffff998e04193a00 R14: 0000000000000005 R15:
>> ffff998e03a6ea00
>> ipfire kernel: FS:  0000758d85125640(0000) GS:ffff998f47300000(0000)
>> knlGS:0000000000000000
>> ipfire kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> ipfire kernel: CR2: 00007722aefed0a0 CR3: 000000010351e003 CR4:
>> 00000000000706e0
>> ipfire kernel: Call Trace:
>> ipfire kernel:  nf_queue_entry_release_refs+0x82/0xa0
>> ipfire kernel:  nf_reinject+0x7a/0x1e0
>> ipfire kernel:  nfqnl_recv_verdict+0x302/0x4f0 [nfnetlink_queue]
>> ipfire kernel:  nfnetlink_rcv_msg+0x16d/0x2c0
>> ipfire kernel:  ? nfnetlink_net_exit_batch+0x60/0x60
>> ipfire kernel:  netlink_rcv_skb+0x5b/0x100
>> ipfire kernel:  netlink_unicast+0x209/0x2d0
>> ipfire kernel:  netlink_sendmsg+0x23a/0x470
>> ipfire kernel:  sock_sendmsg+0x5e/0x60
>> ipfire kernel:  ____sys_sendmsg+0x258/0x2a0
>> ipfire kernel:  ___sys_sendmsg+0xa3/0xf0
>> ipfire kernel:  __sys_sendmsg+0x81/0xd0
>> ipfire kernel:  do_syscall_64+0x33/0x40
>> ipfire kernel:  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>> ipfire kernel: RIP: 0033:0x758d871bb62d
>> ipfire kernel: Code: 28 89 54 24 1c 48 89 74 24 10 89 7c 24 08 e8 fa ee
>> ff ff 8b 54 24 1c 48 8b 74 24 10 41 89 c0 8b 7c 24 08 b8 2e 00 00 00 0f
>> 05 <48> 3d 00 f0 ff ff 77 33 44 89 c7 48 89 44 24 08 e8 4e ef ff ff 48
>> ipfire kernel: RSP: 002b:0000758d85122f40 EFLAGS: 00000293 ORIG_RAX:
>> 000000000000002e
>> ipfire kernel: RAX: ffffffffffffffda RBX: 0000758d80268dd0 RCX:
>> 0000758d871bb62d
>> ipfire kernel: RDX: 0000000000000000 RSI: 0000758d85122f80 RDI:
>> 0000000000000005
>> ipfire kernel: RBP: 0000758d85122fe0 R08: 0000000000000000 R09:
>> 0000758d8689fde0
>> ipfire kernel: R10: 0000000000000000 R11: 0000000000000293 R12:
>> 0000000000000000
>> ipfire kernel: R13: 0000000000000000 R14: 0000000000000001 R15:
>> 0000000000000070
>> ipfire kernel: ---[ end trace 4c8c047c62e118e2 ]---
>> ***SNAP***
>> 
>> Machine is running without (seen) problems - I didn'T reboot yet.
>> 
>> Best,
>> Matthias
>> 
>> On 09.09.2021 22:36, Peter Müller wrote:
>>> Hello *,
>>> 
>>> just for the records: I noticed this behaviour while testing Core Update 140/141 (in
>>> February 2020) the first time. Not kept track on this, but it does not seem to cause
>>> any harm - at least none I am aware of. But it certainly is not a good thing to see,
>>> either...
>>> 
>>> Please refer to https://lists.ipfire.org/pipermail/development/2020-February/007046.html
>>> for further details.
>>> 
>>> Thanks, and best regards,
>>> Peter Müller
>>> 
>>>> 
>>>> 
>>>>> On 31 Aug 2021, at 16:25, Matthias Fischer <matthias.fischer(a)ipfire.org> wrote:
>>>>> 
>>>>> Hi,
>>>>> 
>>>>> On 31.08.2021 11:56, Michael Tremer wrote:
>>>>>> Hey,
>>>>>> 
>>>>>> This is an oops in the code that injects packets back into the kernel after they have been processed by suricata.
>>>>> 
>>>>> Wow. How did you find this out!?
>>>> 
>>>> There are two key functions in the trace:
>>>> 
>>>>>>> 1 Time(s):  nf_queue_entry_release_refs+0x82/0xa0
>>>>>>> 1 Time(s):  nf_reinject+0x7a/0x1e0
>>>> 
>>>> They are from Netfilter and the NFQUEUE module. We only use that for the IPS.
>>>> 
>>>>>> Was this a one-off or does this happen on a regular basis?
>>>>> 
>>>>> Until now, it only happened once.
>>>>> 
>>>>> It rebooted the machine - just to be sure - and its running "without
>>>>> seen problems" since then. Absolutely normal.
>>>> 
>>>> Would be interesting to see how it behaves if it doesn’t get a reboot.
>>>> 
>>>> This is definitely a bug and needs to be fixed in the Linux kernel.
>>>> 
>>>> -Michael
>>>> 
>>>>> 
>>>>> Best,
>>>>> Matthias
>>>>> 
>>>>>> -Michael
>>>>>> 
>>>>>>> On 27 Aug 2021, at 17:16, Matthias Fischer <matthias.fischer(a)ipfire.org> wrote:
>>>>>>> 
>>>>>>> Hi,
>>>>>>> 
>>>>>>> today I took the usual look in LOG SUMMARY and was surprised finding this:
>>>>>>> 
>>>>>>> ***SNIP***
>>>>>>> Kernel
>>>>>>> 
>>>>>>> WARNING:  Kernel Errors Present
>>>>>>>  WARNING: CPU: 0 PID: 2984 at lib/refcount.c:28 r ...:  1 Time(s)
>>>>>>> 
>>>>>>> 1 Time(s):  ? nfnetlink_net_exit_batch+0x60/0x60
>>>>>>> 1 Time(s):  [last unloaded: hwmon_vid]
>>>>>>> 1 Time(s):  ____sys_sendmsg+0x258/0x2a0
>>>>>>> 1 Time(s):  ___sys_sendmsg+0xa3/0xf0
>>>>>>> 1 Time(s):  __sys_sendmsg+0x81/0xd0
>>>>>>> 1 Time(s):  do_syscall_64+0x33/0x40
>>>>>>> 1 Time(s):  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>>>>>> 1 Time(s):  netlink_rcv_skb+0x5b/0x100
>>>>>>> 1 Time(s):  netlink_sendmsg+0x23a/0x470
>>>>>>> 1 Time(s):  netlink_unicast+0x209/0x2d0
>>>>>>> 1 Time(s):  nf_queue_entry_release_refs+0x82/0xa0
>>>>>>> 1 Time(s):  nf_reinject+0x7a/0x1e0
>>>>>>> 1 Time(s):  nfnetlink_rcv_msg+0x16d/0x2c0
>>>>>>> 1 Time(s):  nfqnl_recv_verdict+0x302/0x4f0 [nfnetlink_queue]
>>>>>>> 1 Time(s):  sock_sendmsg+0x5e/0x60
>>>>>>> 1 Time(s): ------------[ cut here ]------------
>>>>>>> 1 Time(s): ---[ end trace 49e1e291edb98731 ]---
>>>>>>> 1 Time(s): CPU: 0 PID: 2984 Comm: W-NFQ#1 Tainted: G           O
>>>>>>> 5.10.55-ipfire #1
>>>>>>> 1 Time(s): CR2: 000079fa1356a6f8 CR3: 0000000101532004 CR4:
>>>>>>> 00000000000706f0
>>>>>>> 1 Time(s): CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>>>>> 1 Time(s): Call Trace:
>>>>>>> 1 Time(s): Code: 05 ba 60 1f 01 01 e8 af a3 52 00 0f 0b c3 80 3d a8 60
>>>>>>> 1f 01 00 75 95 48 c7 c7 b8 fc 11 b4 c6 05 98 60 1f 01 01 e8 90 a3 52 00
>>>>>>> <0f> 0b c3 80 3d 87 60 1f 01 00 0f 85 72 ff ff ff 48 c7 c7 10 fd 11
>>>>>>> 1 Time(s): Code: 28 89 54 24 1c 48 89 74 24 10 89 7c 24 08 e8 fa ee ff
>>>>>>> ff 8b 54 24 1c 48 8b 74 24 10 41 89 c0 8b 7c 24 08 b8 2e 00 00 00 0f 05
>>>>>>> <48> 3d 00 f0 ff ff 77 33 44 89 c7 48 89 44 24 08 e8 4e ef ff ff 48
>>>>>>> 1 Time(s): FS:  00007144d27d9640(0000) GS:ffff9db887200000(0000)
>>>>>>> knlGS:0000000000000000
>>>>>>> 1 Time(s): Hardware name: To be filled by O.E.M. To be filled by
>>>>>>> O.E.M./CRESCENTBAY, BIOS 5.011 04/13/2016
>>>>>>> 1 Time(s): Modules linked in: xt_REDIRECT nfnetlink_queue xt_NFQUEUE
>>>>>>> xt_MASQUERADE ccm cpufreq_conservative cpufreq_ondemand xt_geoip(O)
>>>>>>> xt_ipp2p(O) compat_xtables(O) xt_hashlimit xt_mac xt_multiport xt_mark
>>>>>>> xt_policy xt_TCPMSS xt_conntrack xt_comment ipt_REJECT nf_reject_ipv4
>>>>>>> xt_LOG xt_limit nf_log_ipv4 nf_log_common iptable_raw iptable_mangle
>>>>>>> iptable_filter vfat fat sch_fq_codel rt2800usb rt2x00usb
>>>>>>> x86_pkg_temp_thermal rt2800lib intel_powerclamp coretemp rt2x00lib
>>>>>>> kvm_intel mac80211 kvm at24 regmap_i2c iTCO_wdt iTCO_vendor_support
>>>>>>> irqbypass crct10dif_pclmul cfg80211 crc32_pclmul ghash_clmulni_intel
>>>>>>> pcspkr i2c_i801 rfkill r8169 lpc_ich i2c_smbus mfd_core realtek libarc4
>>>>>>> ir_rc6_decoder rc_rc6_mce i2c_algo_bit snd_hda_codec_realtek nuvoton_cir
>>>>>>> fb_sys_fops syscopyarea sysfillrect sysimgblt snd_hda_codec_generic
>>>>>>> ledtrig_audio i2c_core rc_core snd_hda_intel snd_intel_dspcfg
>>>>>>> snd_hda_codec snd_hda_core snd_hwdep acpi_pad snd_pcm snd_timer snd
>>>>>>> soundcore lp parport_pc parport video
>>>>>>> 1 Time(s): R10: 0000000000000000 R11: 0000000000000293 R12:
>>>>>>> 0000000000000000
>>>>>>> 1 Time(s): R10: ffffb330821af780 R11: ffffffffb4533dc8 R12:
>>>>>>> ffff9db791371980
>>>>>>> 1 Time(s): R13: 0000000000000000 R14: 0000000000000001 R15:
>>>>>>> 0000000000000070
>>>>>>> 1 Time(s): R13: ffff9db7444d9200 R14: 0000000000000005 R15:
>>>>>>> ffff9db7f15c6b00
>>>>>>> 1 Time(s): RAX: 0000000000000000 RBX: ffff9db791371980 RCX:
>>>>>>> 0000000000000027
>>>>>>> 1 Time(s): RAX: ffffffffffffffda RBX: 00007144c4268dd0 RCX:
>>>>>>> 00007144d507062d
>>>>>>> 1 Time(s): RBP: 00007144d27d6fe0 R08: 0000000000000000 R09:
>>>>>>> 00007144d4754de0
>>>>>>> 1 Time(s): RBP: ffffb330821afa50 R08: 0000000000000000 R09:
>>>>>>> ffffb330821af788
>>>>>>> 1 Time(s): RDX: 0000000000000000 RSI: 00007144d27d6f80 RDI:
>>>>>>> 0000000000000006
>>>>>>> 1 Time(s): RDX: ffff9db887218968 RSI: 0000000000000001 RDI:
>>>>>>> ffff9db887218960
>>>>>>> 1 Time(s): RIP: 0010:refcount_warn_saturate+0xa6/0xf0
>>>>>>> 1 Time(s): RIP: 0033:0x7144d507062d
>>>>>>> 1 Time(s): RSP: 0018:ffffb330821af950 EFLAGS: 00010282
>>>>>>> 1 Time(s): RSP: 002b:00007144d27d6f40 EFLAGS: 00000293 ORIG_RAX:
>>>>>>> 000000000000002e
>>>>>>> 1 Time(s): refcount_t: underflow; use-after-free.
>>>>>>> ***SNAP***
>>>>>>> 
>>>>>>> I don't know exactly how do deal with this - does anyone has an idea
>>>>>>> what this means?
>>>>>>> 
>>>>>>> Besides the machine is up and running with Core 159 / 64bit since 6 days
>>>>>>> now and if I hadn't looked at the logs I would not have noticed it.
>>>>>>> 
>>>>>>> I didn't reboot or changed anything yet- should I do?
>>>>>>> 
>>>>>>> Best,
>>>>>>> Matthias
>>>>>> 
>>>>> 
>>>> 
>>> 
>> 
>