You are not logged in.

#1 2010-08-03 17:14:08

Kane
Member
Registered: 2006-10-08
Posts: 220

Bug in wireless?

EDIT: See last post


Hi guys,
Im using a wireless usb dongle which uses the rt2870 driver.
Every 20 minutes (exactly) i get:

cat /var/log/everything.log | grep lost:

Aug  3 13:04:27 kane dhcpcd[16250]: wlan0: carrier lost
Aug  3 13:24:27 kane dhcpcd[16950]: wlan0: carrier lost
Aug  3 13:44:27 kane dhcpcd[17745]: wlan0: carrier lost
Aug  3 14:04:27 kane dhcpcd[18541]: wlan0: carrier lost
Aug  3 14:24:27 kane dhcpcd[19344]: wlan0: carrier lost
Aug  3 14:44:27 kane dhcpcd[20140]: wlan0: carrier lost
Aug  3 15:04:27 kane dhcpcd[20938]: wlan0: carrier lost
Aug  3 15:24:27 kane dhcpcd[21729]: wlan0: carrier lost
Aug  3 15:44:28 kane dhcpcd[22530]: wlan0: carrier lost
Aug  3 16:04:28 kane dhcpcd[23327]: wlan0: carrier lost
Aug  3 16:24:28 kane dhcpcd[24130]: wlan0: carrier lost
Aug  3 16:44:28 kane dhcpcd[28100]: wlan0: carrier lost
Aug  3 17:04:28 kane dhcpcd[28899]: wlan0: carrier lost
Aug  3 17:24:28 kane dhcpcd[29692]: wlan0: carrier lost

I was originally using networkmanager, once this disconnected, i could not reconnect until i ether unplugged and inserted the dongle again, or until i restarted networkmanager.
However now im using wicd, and it seems to be able to reconnect (usually fast enough that i dont notice its disconnected).

The driver is not the issue as i also have the same problem with another usb dongle (however this one requires ndiswrapper).

Is  this a problem with the access point?
I thought it might be powersaving but i disabled it (ifconfig wlan0 power off)
I also have never noticed a problem when testing it in windows.

Thanks
Kane

Last edited by Kane (2010-08-06 16:31:46)

Offline

#2 2010-08-03 18:51:04

ewaller
Administrator
From: Pasadena, CA
Registered: 2009-07-13
Posts: 19,808

Re: Bug in wireless?

Kane wrote:

The driver is not the issue as i also have the same problem with another usb dongle (however this one requires ndiswrapper).

Is  this a problem with the access point?

Seems likely.  Is there anything else interesting in dmesg around the 'lost' lines.

Tell us some more about the AP.  Do you have admin rights on it?


Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

Offline

#3 2010-08-03 21:58:39

ScottWilson1990
Member
Registered: 2010-01-20
Posts: 33

Re: Bug in wireless?

AP is a Netgear DG834G, full admin rights.

We live in the same house and mine is having a similar problem, but mine is using the Zoom Wireless-G USB 4410B card with ndiswrapper. I've tried using Network-Manager, WICD and Netcfg and all give the same problem. In my everything.log I get a

dhcpcd[6309]: wlan0: carrier lost

error. But Kanes disconnects a hell of a lot more and possibly gives a different error, but i'll leave that for him to discuss.


Scott.

Last edited by ScottWilson1990 (2010-08-03 21:59:53)


"The major difference between a thing that might go wrong and a thing that cannot possibly go wrong is that when a thing that cannot possibly go wrong goes wrong it usually turns out to be impossible to get at or repair..." - Douglas Adams

Offline

#4 2010-08-03 22:11:17

Kane
Member
Registered: 2006-10-08
Posts: 220

Re: Bug in wireless?

Edit: As above

Aug  3 22:48:35 kane kernel: No probe response from AP 00:0f:b5:7c:94:9a after 500ms, disconnecting.
Aug  3 22:48:35 kane dhcpcd[6342]: wlan0: carrier lost
Aug  3 22:48:35 kane kernel: cfg80211: Calling CRDA to update world regulatory domain
Aug  3 22:48:37 kane kernel: wlan0: authenticate with 00:0f:b5:7c:94:9a (try 1)
Aug  3 22:48:37 kane kernel: wlan0: authenticate with 00:0f:b5:7c:94:9a (try 2)
Aug  3 22:48:37 kane kernel: wlan0: authenticate with 00:0f:b5:7c:94:9a (try 3)
Aug  3 22:48:37 kane kernel: wlan0: authentication with 00:0f:b5:7c:94:9a timed out
Aug  3 22:48:38 kane dhcpcd[6342]: received SIGHUP, releasing
Aug  3 22:48:38 kane dhcpcd[6342]: wlan0: removing interface
Aug  3 22:48:38 kane dhcpcd[7395]: sending signal 1 to pid 6342
Aug  3 22:48:38 kane dhcpcd[7395]: waiting for pid 6342 to exit
Aug  3 22:48:41 kane dhcpcd[7408]: dhcpcd not running
Aug  3 22:48:41 kane kernel: r8169 0000:02:00.0: eth0: link down

Then reconnects:

Aug  3 22:48:49 kane dhcpcd[7423]: dhcpcd not running
Aug  3 22:48:52 kane dhcpcd[7429]: dhcpcd not running
Aug  3 22:48:52 kane kernel: r8169 0000:02:00.0: eth0: link down
Aug  3 22:48:52 kane dhcpcd[7436]: dhcpcd not running
Aug  3 22:48:57 kane kernel: wlan0: authenticate with 00:0f:b5:7c:94:9a (try 1)
Aug  3 22:48:57 kane kernel: wlan0: authenticated
Aug  3 22:48:57 kane kernel: wlan0: associate with 00:0f:b5:7c:94:9a (try 1)
Aug  3 22:48:57 kane kernel: wlan0: RX AssocResp from 00:0f:b5:7c:94:9a (capab=0x471 status=0 aid=2)
Aug  3 22:48:57 kane kernel: wlan0: associated
Aug  3 22:48:58 kane dhcpcd[7452]: version 5.2.5 starting
Aug  3 22:48:58 kane dhcpcd[7452]: wlan0: rebinding lease of 192.168.0.6
Aug  3 22:48:58 kane dhcpcd[7452]: wlan0: acknowledged 192.168.0.6 from 192.168.0.1
Aug  3 22:48:58 kane dhcpcd[7452]: wlan0: checking for 192.168.0.6
Aug  3 22:49:03 kane dhcpcd[7452]: wlan0: leased 192.168.0.6 for 86400 seconds
Aug  3 22:49:03 kane dhcpcd[7452]: forked to background, child pid 7473

However i noticed when using wpa_supplicant alone, it behaves similarly to networkmanager (i.e. disconnects and wont reconnect until wpa_supplicant is killed and restarted). Unlike wicd which seems to be able to reconnect without issue.
This also occurs when all encryption is disabled on the router (ruling out issues with wpa).

Thanks

Edit: Wicd and Networkmanager also report 100% signal strength.

Last edited by Kane (2010-08-03 22:16:37)

Offline

#5 2010-08-03 22:21:23

ewaller
Administrator
From: Pasadena, CA
Registered: 2009-07-13
Posts: 19,808

Re: Bug in wireless?

Is your router configured to keep a log? If so, is there any correlation to the logs from your Arch systems?

Do your systems' logs indicate that your drops occur simultaneously on both systems, or are they asynchronous?


Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

Offline

#6 2010-08-03 22:28:55

Kane
Member
Registered: 2006-10-08
Posts: 220

Re: Bug in wireless?

it doesn't allow the ability to log connections.

and the connection drops are asynchronous, the only similarity between the two boxes is they are both about the same distance away from the router.
Therefore looks most likely to be an AP issue.
I guess the next stage is to relocate the AP and do some testing

Offline

#7 2010-08-06 16:43:56

Kane
Member
Registered: 2006-10-08
Posts: 220

Re: Bug in wireless?

Well after some more testing:

Using 2 different computers (one 64bit the other 32bit) with 2 different wireless usb dongles (one uses ndiswrapper the other uses rt28xx usb driver) with 2  different access points (both tried with and without wpa/or any encryption whatsoever).
With the access point within a metre of the dongle
With no issues in windows

It seems most likely there is a bug in the wireless code somewhere? Either that or all 8 combinations just happen to have the same issue tongue (or both network cards have the same issue)

I can rule out wpa_supplicant as it still happens without any encryption, and have tried netcfg, wicd, networkmanager (all still have the same issue) therefore not likely to be a problem with those (unless they all use the same libraries/routines/underlying software?).

some people may not realise their wireless is disconnecting as most of the time it seems to reconnect fairly quickly but checking /var/log/everything.log will show many disconnections if its happening.

Anyone any ideas how to test further?

Cheers
Kane

Last edited by Kane (2010-08-06 16:44:49)

Offline

Board footer

Powered by FluxBB