Project

General

Profile

Actions
Copy link

Support #1787

closed

FreeBSD netmap: [ERRCODE: SC_ERR_NETMAP_READ(264)] - Invalid polling request

Added by Victor Julien almost 8 years ago. Updated almost 8 years ago.

Status:
Closed
Priority:
Normal
Assignee:
Aleksey Katargin
Affected Versions:
Label:

Description

This may be a user issue, as I'm not very familiar with FreeBSD & netmap.

I'm running FreeBSD 10.3 with netmap on a 2 core system. So I would expect 2 worker threads.

netmap:
   # To specify OS endpoint add plus sign at the end (e.g. "eth0+")
 - interface: igb0
   # Number of receive threads. "auto" uses number of RSS queues on interface.
   threads: auto

I start with:

./src/suricata --netmap -v -l /tmp/ -c suricata.yaml

Immediately it starts giving this 'invalid polling request error':

[100147] 20/5/2016 -- 16:04:50 - (util-device.c:218) <Info> (LiveBuildDeviceListCustom) -- Adding interface igb0 from config file
[100147] 20/5/2016 -- 16:04:50 - (runmode-netmap.c:194) <Info> (ParseNetmapConfig) -- Using 1 threads for interface igb0
[100147] 20/5/2016 -- 16:04:50 - (util-runmodes.c:321) <Info> (RunModeSetLiveCaptureWorkersForDevice) -- Going to use 1 thread(s)
[100177] 20/5/2016 -- 16:04:50 - (tmqh-packetpool.c:398) <Info> (PacketPoolInit) -- preallocated 1024 packets. Total memory 2461696
[100147] 20/5/2016 -- 16:04:50 - (runmode-netmap.c:194) <Info> (ParseNetmapConfig) -- Using 1 threads for interface igb0
[100147] 20/5/2016 -- 16:04:50 - (util-runmodes.c:321) <Info> (RunModeSetLiveCaptureWorkersForDevice) -- Going to use 1 thread(s)
[100178] 20/5/2016 -- 16:04:50 - (tmqh-packetpool.c:398) <Info> (PacketPoolInit) -- preallocated 1024 packets. Total memory 2461696
[100147] 20/5/2016 -- 16:04:50 - (runmode-netmap.c:451) <Info> (RunModeIdsNetmapWorkers) -- RunModeIdsNetmapWorkers initialised
[100147] 20/5/2016 -- 16:04:50 - (flow-manager.c:720) <Info> (FlowManagerThreadSpawn) -- using 1 flow manager threads
[100179] 20/5/2016 -- 16:04:50 - (tmqh-packetpool.c:398) <Info> (PacketPoolInit) -- preallocated 1024 packets. Total memory 2461696
[100147] 20/5/2016 -- 16:04:50 - (flow-manager.c:884) <Info> (FlowRecyclerThreadSpawn) -- using 1 flow recycler threads
[100147] 20/5/2016 -- 16:04:50 - (tm-threads.c:2164) <Notice> (TmThreadWaitOnThreadInit) -- all 2 packet processing threads, 4 management threads initialized, engine started.
[100178] 20/5/2016 -- 16:04:50 - (source-netmap.c:918) <Error> (ReceiveNetmapLoop) -- [ERRCODE: SC_ERR_NETMAP_READ(264)] - Invalid polling request
[100178] 20/5/2016 -- 16:04:50 - (source-netmap.c:918) <Error> (ReceiveNetmapLoop) -- [ERRCODE: SC_ERR_NETMAP_READ(264)] - Invalid polling request

AFAICS there should be 2 RSS queues:

# dmesg |grep igb
igb0: <Intel(R) PRO/1000 Network Connection, Version - 2.5.3-k> port 0x2000-0x201f mem 0xe2e00000-0xe2e1ffff,0xe2400000-0xe27fffff,0xe2e40000-0xe2e43fff irq 16 at device 0.0 on pci1
igb0: Using MSIX interrupts with 3 vectors
igb0: Ethernet address: 90:e2:ba:01:7c:38
igb0: Bound queue 0 to cpu 0
igb0: Bound queue 1 to cpu 1
igb0: netmap queues/slots: TX 2/1024, RX 2/1024
igb1: <Intel(R) PRO/1000 Network Connection, Version - 2.5.3-k> port 0x2020-0x203f mem 0xe2e20000-0xe2e3ffff,0xe2800000-0xe2bfffff,0xe2e44000-0xe2e47fff irq 17 at device 0.1 on pci1
igb1: Using MSIX interrupts with 3 vectors
igb1: Ethernet address: 90:e2:ba:01:7c:39
igb1: Bound queue 0 to cpu 0
igb1: Bound queue 1 to cpu 1
igb1: netmap queues/slots: TX 2/1024, RX 2/1024
igb1: link state changed to UP
igb0: link state changed to UP
igb0: permanently promiscuous mode enabled

If I set 'threads: 1' everything works.

Actions
Copy link
 #1

Updated by Victor Julien almost 8 years ago

Actually, looking at Suricata's output above something seems wrong with the runmode setup. It doesn't set up 2 worker mode, but a 2 x 1 worker somehow? Not sure if the effect is different, but this output at least looks suspicious to me.

[100147] 20/5/2016 -- 16:04:50 - (runmode-netmap.c:194) <Info> (ParseNetmapConfig) -- Using 1 threads for interface igb0
[100147] 20/5/2016 -- 16:04:50 - (util-runmodes.c:321) <Info> (RunModeSetLiveCaptureWorkersForDevice) -- Going to use 1 thread(s)
[100177] 20/5/2016 -- 16:04:50 - (tmqh-packetpool.c:398) <Info> (PacketPoolInit) -- preallocated 1024 packets. Total memory 2461696
[100147] 20/5/2016 -- 16:04:50 - (runmode-netmap.c:194) <Info> (ParseNetmapConfig) -- Using 1 threads for interface igb0
[100147] 20/5/2016 -- 16:04:50 - (util-runmodes.c:321) <Info> (RunModeSetLiveCaptureWorkersForDevice) -- Going to use 1 thread(s)
Actions
Copy link
 #2

Updated by Aleksey Katargin almost 8 years ago

Latest master source spawn duplicate worker threads.

netmap:
   # To specify OS endpoint add plus sign at the end (e.g. "eth0+")
 - interface: eth2
   # Number of receive threads. "auto" uses number of RSS queues on interface.
   threads: auto
   # You can use the following variables to activate netmap tap or IPS mode.
   # If copy-mode is set to ips or tap, the traffic coming to the current
   # interface will be copied to the copy-iface interface. If 'tap' is set, the
   # copy is complete. If 'ips' is set, the packet matching a 'drop' action
   # will not be copied.
   # To specify the OS as the copy-iface (so the OS can route packets, or forward
   # to a service running on the same machine) add a plus sign at the end
   # (e.g. "copy-iface: eth0+"). Don't forget to set up a symmetrical eth0+ -> eth0
   # for return packets. Hardware checksumming must be *off* on the interface if
   # using an OS endpoint (e.g. 'ifconfig eth0 -rxcsum -txcsum -rxcsum6 -txcsum6' for FreeBSD
   # or 'ethtool -K eth0 tx off rx off' for Linux).
   copy-mode: tap
   copy-iface: eth3
   # Set to yes to disable promiscuous mode
   # disable-promisc: no
   # Choose checksum verification mode for the interface. At the moment
   # of the capture, some packets may be with an invalid checksum due to
   # offloading to the network card of the checksum computation.
   # Possible values are:
   #  - yes: checksum validation is forced
   #  - no: checksum validation is disabled
   #  - auto: suricata uses a statistical approach to detect when
   #  checksum off-loading is used.
   # Warning: 'checksum-validation' must be set to yes to have any validation
   #checksum-checks: auto
   # BPF filter to apply to this interface. The pcap filter syntax apply here.
   #bpf-filter: port 80 or udp
 - interface: eth3
   threads: auto
   copy-mode: tap
   copy-iface: eth2
   # Put default values here
 - interface: default

Breakpoint 3, RunModeSetLiveCaptureWorkers (ConfigParser=0x5344a3 <ParseNetmapConfig>, ModThreadsCount=0x5354bd <NetmapConfigGeThreadsCount>, recv_mod_name=0x601ea5 "ReceiveNetmap", decode_mod_name=0x601e98 "DecodeNetmap", thread_name=0x602a53 "W", live_dev=0x0)
    at util-runmodes.c:374
374        int nlive = LiveGetDeviceCount();
(gdb) n
378        for (ldev = 0; ldev < nlive; ldev++) {
(gdb) print nlive
$1 = 4
(gdb)
Actions
Copy link
 #3

Updated by Victor Julien almost 8 years ago

  • Status changed from New to Closed

Thanks Aleksey, that was the issue indeed. Fixed in #1768 (https://github.com/inliniac/suricata/pull/2128)

Actions
Copy link

Also available in: Atom PDF