Re: CPU load on queued_spin_lock_slowpath
Thanks for your advices. I will try to create the erroneous situation by triggering icmp error for existing connection and try non-tcp patch and kernel upgrade respectively. I will report the results at mail list. > On Tue, Feb 6, 2018, 7:10 AM Pablo Neira Ayuso <pa...@netfilter.org> wrote: >> >> On Tue, Feb 06, 2018 at 10:56:20AM +0300, Tugrul Erdogan wrote: >> > Hi All, >> > >> > My server had a locking problem with the logs located below. I can not >> > reproduce this erroneous situation again but I think that there is an >> > active vulnerability at my server because of this error. >> > >> > My server's kernel version is v4.6.4. >> >> Probably this helps you? >> >> commit 49f817d793d1bcc11d721881aac037b996feef5c >> Author: Lin Zhang <xiaolou4...@gmail.com> >> Date: Fri Oct 6 00:44:03 2017 +0800 >> >> netfilter: SYNPROXY: skip non-tcp packet in {ipv4, ipv6}_synproxy_hook >> >> 4.6.4 is rather old, BTW. >> -- >> To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in >> the body of a message to majord...@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
CPU load on queued_spin_lock_slowpath
Hi All, My server had a locking problem with the logs located below. I can not reproduce this erroneous situation again but I think that there is an active vulnerability at my server because of this error. My server's kernel version is v4.6.4. What can be the cause of this error or do you have any opinion about how can I reproduce this logs again? Thanks for your helps. Best regards, Tugrul Feb 5 13:20:42 serv kernel: [] queued_spin_lock_slowpath+0xb/0xf Feb 5 13:20:42 serv kernel: [] _raw_spin_lock_bh+0x2b/0x30 Feb 5 13:20:42 serv kernel: [] connlimit_mt+0x114/0x30 [xt_connlimit] Feb 5 13:20:42 serv kernel: [] ? hashlimit_mt+0x2b7/0x71 [xt_hashlimit] Feb 5 13:20:42 serv kernel: [] ? _raw_spin_unlock_bh+0x1e/0x20 Feb 5 13:20:42 serv kernel: [] ipt_do_table+0x25f/0x710 [ipt_tables] Feb 5 13:20:42 serv kernel: [] ? ipt_do_table+0x332/0x710 [ipt_tables] Feb 5 13:20:42 serv kernel: [] ? tcp_packet+0x39d/0x9a0 Feb 5 13:20:42 serv kernel: [] ? dev_hard_start_xmit+0x22f/0x3e0 Feb 5 13:20:42 serv kernel: [] iptable_mangle_hook+0x37/0x110 [iptable_mangle] Feb 5 13:20:42 serv kernel: [] nf_iterate+0x5d/0x70 Feb 5 13:20:42 serv kernel: [] nf_hook_slow+0x5d/0x70 Feb 5 13:20:42 serv kernel: [] ip_output+0xdb/0xf0 Feb 5 13:20:42 serv kernel: [] ? __ip_local_out+0xa2/0x110 Feb 5 13:20:42 serv kernel: [] ? ip_fragment.constprop.51+0x80/0x80 Feb 5 13:20:42 serv kernel: [] ip_local_out+0x35/0x40 Feb 5 13:20:42 serv kernel: [] synproxy_send_tcp.isra.8+0xca/0xf0 [ipt_SYNPROXY] Feb 5 13:20:42 serv kernel: [] synproxy_recv_client_ack+0x200/0x340 [ipt_SYNPROXY] Feb 5 13:20:42 serv kernel: [] synproxy_tg4+0x11c/0x308 [ipt_SYNPROXY] -- To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
About Kernel SynProxy Performance
Hi All, I have noticed the commit below in the kernel 4.7 change logs. https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=b3d051477cf94e9d71d6acadb8a90de15237b9c1 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=8804b2722dc5d6f9b7ba0a9e812eae9ee5ce95bc Then, I think that this commit might increase SYNPROXY performance of kernel too and updated my kernel from 3.15.9 to 4.10. But I didn't notice any significant change on received pps with netfilter/SYNPROXY module. I wanted to ask you what may be the reason of that. The commit message says that the task test result would have approximately %100 increase. Could you give more details about the reflection of this commit to SYNPROXY module on performance perspective. If this commit not affect SYNPROXY module, I want to take your opinions about is there a way to increase SYNPROXY performace with some changes which are adopted from the commit linked above on SYNPROXY internals. CPU and nic info I used is below: Intel(R) Xeon(R) CPU E5645 @ 2.40GHz 82599ES 10-Gigabit SFI/SFP+ Network Connection and iptables rules -t raw -A PREROUTING -p tcp -m physdev --physdev-in enp7s0f0 -m tcp --tcp-flags FIN,SYN,RST,ACK SYN -j NOTRACK -t filter -A FORWARD -p tcp -m physdev --physdev-in enp7s0f0 -m state --state INVALID,UNTRACKED -j SYNPROXY --sack-perm --timestamp --wscale 7 --mss 1480 Thanks for your kind replies. Best regards, Tugrul -- To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html