In/Out errors :(

neonsigns

The errors are creeping up again. Here is the latest athstat with rxbuf and txbuf at 3000 and 6000 respectively

athstats

9269 tx management frames
275 tx frames discarded prior to association
262 tx failed 'cuz FIFO underrun
2121 tx failed 'cuz bogus xmit rate
1754 tx frames with rts enabled
7519 tx frames with an alternate rate
212 tx frames with 11g protection
483679 rx failed 'cuz of FIFO overrun
2 rx failed 'cuz frame too short
4119 rx management frames
113640 beacon setup failed 'cuz no mbuf
614465541 beacons transmitted
410 periodic calibration failures
75 tx used alternate antenna
Antenna profile:
[2] tx 5749 rx 4174
[3] tx 3530 rx 2081

Any suggestions? Also, the dev.ath.0.rxantenna keeps switching from 1 to 2 and back?

covex

neonsigns, could you post your config file (sysctl)… or at least a section of it with wifi settings you did
i tried g and a, different settings from this thread and still getting tons of in/out errors (btw, "a" mode produces less of'em :-)

neonsigns

sysctl hw.ath

hw.ath.hal.swba_backoff: 0
hw.ath.hal.sw_brt: 10
hw.ath.hal.dma_brt: 2
hw.ath.hal.version: 0.9.17.2
hw.ath.txbuf: 3000
hw.ath.rxbuf: 6000
hw.ath.regdomain: 0
hw.ath.countrycode: 0
hw.ath.xchanmode: 1
hw.ath.outdoor: 1
hw.ath.calibrate: 30
hw.ath.dwell: 200

sysctl dev.ath.0

dev.ath.0.%desc: Atheros 5212
dev.ath.0.%driver: ath
dev.ath.0.%location: slot=8 function=0
dev.ath.0.%pnpinfo: vendor=0x168c device=0x001a subvendor=0x1113 subdevice=0xee24 class=0x020000
dev.ath.0.%parent: pci1
dev.ath.0.smoothing_rate: 95
dev.ath.0.sample_rate: 10
dev.ath.0.countrycode: 0
dev.ath.0.regdomain: 96
dev.ath.0.slottime: 9
dev.ath.0.acktimeout: 48
dev.ath.0.ctstimeout: 48
dev.ath.0.softled: 0
dev.ath.0.ledpin: 0
dev.ath.0.ledon: 0
dev.ath.0.ledidle: 2700
dev.ath.0.txantenna: 0
dev.ath.0.rxantenna: 1
dev.ath.0.diversity: 0
dev.ath.0.txintrperiod: 5
dev.ath.0.diag: 0
dev.ath.0.tpscale: 0
dev.ath.0.tpc: 0
dev.ath.0.tpack: 63
dev.ath.0.tpcts: 63
dev.ath.0.monpass: 24

athstats

10033 tx management frames
522 tx frames discarded prior to association
498 tx failed 'cuz FIFO underrun
3071 tx failed 'cuz bogus xmit rate
2491 tx frames with rts enabled
7547 tx frames with an alternate rate
212 tx frames with 11g protection
1084420 rx failed 'cuz of FIFO overrun
3 rx failed 'cuz frame too short
4277 rx management frames
266364 beacon setup failed 'cuz no mbuf
1822425091 beacons transmitted
932 periodic calibration failures
77 tx used alternate antenna
Antenna profile:
[2] tx 6512 rx 4668
[3] tx 3542 rx 2093

ifconfig -v ath0

ath0: flags=8943 <up,broadcast,running,promisc,simplex,multicast>mtu 1500
inet6 fe80::213:f7ff:fe1f:bffb%ath0 prefixlen 64 scopeid 0x2
ether 00:13:f7:1f:bf:fb
media: IEEE 802.11 Wireless Ethernet autoselect mode 11g <hostap>status: associated
ssid NWIFI channel 11 (2462) bssid xx:xx:xx:xx:xx:xx
authmode WPA1+WPA2/802.11i privacy MIXED deftxkey 3
TKIP 2:128-bit
TKIP 3:128-bit powersavemode OFF powersavesleep 100 txpowmax 33
txpower 63 rtsthreshold 2346 mcastrate 1 fragthreshold 2346 bmiss 7
pureg protmode OFF -wme burst ssid SHOW apbridge dtimperiod 1
bintval 100 -countermeasures</hostap></up,broadcast,running,promisc,simplex,multicast>

eri--

can you please try by forcing the media manually?

GoldServe

I have read somewhere that in freebsd, you can disable te bgscan via ifconfig ath0 -bgscan but it's not available in 6.2. Maybe someone can port over 6.3 and see if this helps.

firestar

I've the same problem about tons of IN/OUT errors. I also have a wifi card based on Atheros 5212.
I've been trying pfSense with the new txbuf and rxbuf and disconnections happen 8-9 hours later the first association instead of every 2 hours as usually occurs without that change in /boot/loader.conf
In my case those parameters in loader.conf make less OUT errors and few IN errors, but unluckily it doesn't solve the problem. Disconnections shouldn't occur due to the short distance between pfSense wifi antenna and client one.
It has been a year wifi connection have disconnection every 2-3 hours :(
As I've said, today wifi connection lasted 9 hours and then disconnection. I bet it will get disconnected with 9 hours from now.
I tought about a problem with hostapd, but now I think that is more driver related.
Moreover, I've two configurations and all of them have wifi cards based on Atheros chipset from different producers: I have the same problem reported in this thread with them. :-\

What can we do? My wifi cards aren't faulty and I think yours, too. How can we report this problem to the maintainer of FreeBSD wireless drivers? Developers, we need your help! ::)

werneman

Hi neonsigns and all,your /boot/loader.conf looks like this ?

autoboot_delay="1"
kern.ipc.nmbclusters="0"
"hw.ath.rxbuf=6000"
"hw.ath.txbuf=3000"

firestar

@werneman:

Hi neonsigns and all,your /boot/loader.conf looks like this ?

Not exactly. Be aware with system files!

autoboot_delay="1"
kern.ipc.nmbclusters="0"
hw.ath.rxbuf="6000"
hw.ath.txbuf="3000"

…or without "

covex

hey guys!
why would i have my in errors skyrocketing and our errors stay at the same level?
see attach…
tnx

1.JPG_thumb

neonsigns

@ermal:

can you please try by forcing the media manually?

ermal, how do I do that?

My loader.conf does not have kern.ipc.nmbclusters="0". What does this setting do?

lsf

[20070116, update 20070212] Systems with very heavy network activity have been observed to have some problems with the kernel memory allocator. Symptoms are processes that get stuck in zonelimit state, or system livelocks. One partial workaround for this problem is to add the following line to /boot/loader.conf and reboot:

kern.ipc.nmbclusters="0"

Setting this to 0 removes the max limit on allocating nmbclusters, this can be a problem if you have very large amounts of connections and not a lot of memory, So use it with caution.

neonsigns

I set the media to OFDM/54Mbps 11g and the kern.ipc.nmbclusters="32768" but I am still getting "In" errors and random disconnections.

firestar

@neonsigns:

but I am still getting "In" errors and random disconnections.

Same problem. Is it ok for a wireless interface?
How does pfSense get those data (in/out errors displayed in Status-Interfaces)? What is the equivalent command from the shell?

Thanks

neonsigns

@firestar:

@neonsigns:

but I am still getting "In" errors and random disconnections.

Same problem. Is it ok for a wireless interface?
How does pfSense get those data (in/out errors displayed in Status-Interfaces)? What is the equivalent command from the shell?

Thanks

Don't know what is causing the errors. Athstats shows that the two biggest errors are:

"rx failed 'cuz of FIFO overrun" and "beacon setup failed 'cuz no mbuf"

Anyone have any insights?

bardelot

Same problem here, very high "in errors" rate.

Compex WLM54SAG 200mW (Chipset should be Atheros AR5414)

Tried increasing the hw.ath.txbuf and hw.ath.rxbuf but it wouldn't help.

$ sysctl -a | grep ath.0
dev.ath.0.%desc: Atheros 5212
dev.ath.0.%driver: ath
dev.ath.0.%location: slot=12 function=0
dev.ath.0.%pnpinfo: vendor=0x168c device=0x001b subvendor=0x168c subdevice=0x2062 class=0x020000
dev.ath.0.%parent: pci0
dev.ath.0.smoothing_rate: 95
dev.ath.0.sample_rate: 10
dev.ath.0.countrycode: 0
dev.ath.0.regdomain: 0
dev.ath.0.slottime: 9
dev.ath.0.acktimeout: 23
dev.ath.0.ctstimeout: 23
dev.ath.0.softled: 0
dev.ath.0.ledpin: 0
dev.ath.0.ledon: 0
dev.ath.0.ledidle: 270
dev.ath.0.txantenna: 1
dev.ath.0.rxantenna: 1
dev.ath.0.diversity: 0
dev.ath.0.txintrperiod: 5
dev.ath.0.diag: 0
dev.ath.0.tpscale: 1
dev.ath.0.tpc: 1
dev.ath.0.tpack: 63
dev.ath.0.tpcts: 63
dev.ath.0.monpass: 24

$ sysctl hw.ath
hw.ath.hal.swba_backoff: 0
hw.ath.hal.sw_brt: 10
hw.ath.hal.dma_brt: 2
hw.ath.hal.version: 0.9.17.2
hw.ath.txbuf: 3000
hw.ath.rxbuf: 6000
hw.ath.regdomain: 0
hw.ath.countrycode: 0
hw.ath.xchanmode: 1
hw.ath.outdoor: 1
hw.ath.calibrate: 30
hw.ath.dwell: 200

jjstecchino

Hi all,
I am experiencing , as many other posting in this thread, a ton of IN and some OUT error on the wireless interface.
Also the performance of the wireless link is sluggish at best. This is unfortunate because pfSense is such an excellent product and would be soooo nice to have an all around firewall/wireless AP working solution.
My setup is an alix 2c3 with a Compex WLM54G 200mW. The chipset is an atheros AR2413. I am running pfSense 1.2 release with the wireless card in AP mode, bridged to the LAN interface.

athstats results are similar to what has been posted by others, a ton of rx failed 'cuz FIFO buffer overrun' and many beacon setup failed 'cuz no mbuf'.

Changing the size of the rx and tx buffer ameliorate but does not solve the problem.

Reading the previous posts it is not clear if the problem we are all experiencing is related to:

a) setup (PC hw vs Embedded platform, mini pci/pci wireless network card).
b) a particular chipset or family of
c) configuration (Access point ?)
d) atheros driver support for that specific chipset/card

Maybe could be useful to ask people having the problem but even more useful people who are NOT experiencing the problem to specify their setup.

The bulk of the errors are receive FIFO buffer overrun which would indicate a problem to timely move out the receive buffer data. It would appear (since we can change the size of the buffer) that the buffer itself is a chunk of main memory allocated by the atheros driver. The inability to timely move this data from the driver allocated memory to the OS would point to a problem in the interaction between FreeBSD kernel and the ath driver however this interaction may be complicated by specific hardware issues.

It would be useful to know from people with working configuration, experiencing good wireless throughput and few or no errors about their setup and configuration.
I am convinced what we are experiencing is a ath driver issue that become evident with specific hw configurations.
By collecting the above info we will know what works and what doesn't. We could submit bug requests to FreeBSD however the issue will not be taken care in useful time.

Being pragmatic it would be easier to buy a card known to work with a specific motherboard and pfSense.

firestar

It would be interesting to setup a FreeBSD 7.0 installation as an access point and try the wifi network for at least a week. This would ensure us it is a problem related to FreeBSD 6.x drivers or anything else in the worst case. Wifi networks are source of trouble. That's why I prefer ethernet cables, though I use pfSense as an AP and I experience this issue. :(
If somebody has enough knowledge to do it, do it a let us know. ;D

jjstecchino

pfSense 1.3 should be out soon and we will be able to try freebsd 7.
In the meantime it would be intresting to know from whoever is not experiencing in/out errors what kind of setup they are using so we can compile a list of working/non working components. In particular it would be intresting to know about what motherboard, wireless card and pfsense version and working/non working status.

KoalaTNR

Hello,

I had the same problems.

Setup 1

Hardware: Wrap, Compex WLM54SAG23, 1 antenna
Software: PfSense 1.2 Release

iMac –> Mac OS X --> very good WLAN performance and few in/out errors
Dell Notebook with Intel Wlan --> Windows --> very poor WLAN performance and many in/out errors

In/out errors 381789/0 in one day

==> with a second antenna I've the same errors.
==> that's not an problem of antenna-settings but rather problem with the wlan-adapters and freebsd.

Setup 2

Hardware: Wrap, Compex WLM54SAG23, 1 antenna
Software: PfSense 1.2 Release (special Build that based on FreeBSD 6.3 instead 6.2)

iMac –> Mac OS X --> very good WLAN performance and few in/out errors
Dell Notebook with Intel Wlan --> Windows --> very good WLAN performance and few in/out errors

In/out errors 216/0 in one day (If I setup the antenna-settings correctly or connect the second antenna than I get no in/out errors)

The 6.3 based PfSense 1.2 can downloaded here:

http://cvs.pfsense.org/~sullrich/testing_images/6/FreeBSD_RELENG_6_3/pfSense_RELENG_1_2/

Greetings
Thomas

covex

If I setup the antenna-settings correctly…

what do you mean by that? could you please post your settings here…