You are not logged in.
I'm running a custom 2.6.32.2 kernel and have not had any problems with ath9k
Offline
I gave up using the wireless in my eeepc back at 2.6.30 because of the constant problems outlined in this thread, and have been using the wired ethernet only. Just now I decided to try it out again thinking this must be close to sorted by now!
It seems some progress has been made, I connected to my router last night and this morning some 8 hours later it is still connected, so that is good news, it used to drop out after a couple of minutes.
BUT...I found another issue. I just tried to toggle the wifi off again and go back to wired with Fn-F2 and the light went off and I thought all was good, but next time I opened firefox...no internet. After a quick check around I found that I had no ethernet. The ethernet adapter had completely disappeared from lspci. The wireless is still there. Fn-F2 again and the ethernet adapter comes back.
So something seems to be borked, possibly with the rfkill system. Toggling the wifi seems to affect the ethernet instead of the wireless. Reboot didn't change anything. I haven't tried disabling the wireless in the bios yet.
Offline
I have eeepc 1005-HA. I had the exactly same problem. Disassociating, low signal, disconnecting. Nothing helped, but ndiswrapper saved my wireless connection. I have atheros AR928x. So i recommend you to use ndiswrapper and windows drivers. Everything works fine, signal strenght is same as in windows and i can finally use linux only
Offline
When I upgraded the [Zen] kernel to 2.6.32, ath5k would cause a soft lock-up or refuse to see access points every other boot. The remedy to this was to disable rfkill, which has caused me grief in the past, and so far, everything has been in working order.
Last edited by Wintervenom (2009-12-23 19:34:13)
Offline
The ath5k driver, since 2.6.32, would freeze my machine up every other boot. The remedy to this was to disable rfkill, and so far, everything has been in working order.
and how did you disable rfkill?
Mikes on AUR
Offline
Wintervenom wrote:The ath5k driver, since 2.6.32, would freeze my machine up every other boot. The remedy to this was to disable rfkill, and so far, everything has been in working order.
and how did you disable rfkill?
I compiled without it, but I believe you can blacklist it in your modprobe configuration.
Last edited by Wintervenom (2009-12-23 19:36:02)
Offline
what ive tested so far is
rfkill unblock all
with this command wifi gets enabled but i cannot connect to encrtypted networks. I will try to disable rfkill module and post here the results.
Mikes on AUR
Offline
Error with 2.6.32.2:
ath5k 0000:06:02.0: PCI INT A -> GSI 22 (level, low) -> IRQ 22
ath5k 0000:06:02.0: registered as 'phy0'
ath5k phy0: Invalid EEPROM checksum 0xb0e2
ath5k phy0: unable to init EEPROM
ath5k 0000:06:02.0: PCI INT A disabled
ath5k: probe of 0000:06:02.0 failed with error -5
Controller:
06:02.0 Ethernet controller: Atheros Communications Inc. AR2413 802.11bg NIC (rev 01)
How to fix it?
Offline
Hi, I queue up with those ones getting constantly disassociated from their network. This has been happening to me for some time now, probably it's because kernel updates like you already said. I'm using two machines, a new netbook with ath9k driver, and a slightly older notebook that uses ath5k. The problem happens on both machines.
Oddly though, it doesn't happen on Ubuntu. I'm using Ubuntu right now on both computers because I just need a working desktop. It might not be relevant, but I noticed some things, and perhaps this might help a little bit. Here's what I noticed on my notebook:
Wireless driver is ath5k, kernel is 2.6.31-16-generic (shipped with Ubuntu). Wireless works flawlessly, signal strength is the strongest, and so far I've never lost the connection. Although sometimes, if I open a website in Epiphany, it needs a little bit to start loading.
Now my netbook, using ath9k, and the kernel also is 2.6.31-16. From the beginning, I had those problems getting disconnected from my network. My network even disappeared from the WLAN list, and sometimes I got connected again if I manually typed in the essid and the passwort. Other times though, this didn't work and I had to either reboot or turn off Wi-Fi and turn it on again. Oh and the signal strength was awfully low, practically non-existent.
I made some research on the Ubuntu bug tracker, and somehow I ended up installing the package "linux-backports-modules-2.6.31-16-generic-pae", that depends on "linux-image-2.6.31-16-generic-pae". These package are even in the official Ubuntu repository, I didn't use someone else's PPA. This is funny because I first wanted to install some packages of another one's PPA that might have helped, but then I noticed those and I thought, first I'll try these ones, and voilà, all my wireless problems are gone, the signal strength is very high now, and I never got disconnected so far. "uname -r" says my kernel now is "2.6.31-16-generic-pae".
Of course this might be Ubuntu-specific, but I wanted to tell you this because someone more familiar with the linux kernel, wireless modules and so on (I ain't) might understand these changes.
I don't even know what this "-pae" means, blame on me, but for now I'm just happy to have working wireless. Of course I'll be even happier if I can use Arch again.
Offline
Thanks for posting Leye!
The one we need to find are possible patches that ubuntu devs use for their kernel. I hope the archlinux kernel package maintainers will find a solution very soon.
Mikes on AUR
Offline
@ Leye
The backports package you installed provides modules from newer kernel versions, made compatible with the standard Ubuntu Kernel (2.6.31 with 9.10). In this case the equivalent for Arch would be to install compat-wireless (or update the kernel to 2.6.32, from testing).
PAE (Physical Address Extension) is important if you want to use more than 4 GB RAM with x86 architecture.
Offline
Slight update/possible solution:
I've also been having problems with "disassociating from AP" etc. disconnects, although solvable with my FN+F2 turn off/on wlan acpi script, I've switched to using ndiswrapper.
So far it seems to work perfectly.. if not faster. I also haven't seen any problems with suspend.I followed the instructions here: http://wiki.archlinux.org/index.php/Wir … diswrapper
(make sure to blacklist the old kernel module and unload it when using ndiswrapper)
Using the AR9285 driver I found on the web (for XP 32-bit.I hope this helps! It took me a while to find a decent download, so here's a link here for people using the same card as me: http://www.nodevice.com/driver/AR9285/get64221.html
It shouldn't take too long to find one for ath5k, etc.
Good luck
Thank you!! I now have working wireless card(AR9285) with no drops etc.. Only problem is i can't set it to monitor mode. I need mad wifi drivers to do that correct? I tried but they didnt work at all. Maybe if i try setting a country code. But at least i have working wireless for now. Although it worked pretty good before with excpet when doing VOIP like skype etc.. I much prefer it always working Easiest way to get ath9k or mad wifi working is too downgrade the kernel it seems?
Last edited by leetstalker (2009-12-30 22:21:35)
Offline
When I upgraded the [Zen] kernel to 2.6.32, ath5k would cause a soft lock-up or refuse to see access points every other boot. The remedy to this was to disable rfkill, which has caused me grief in the past, and so far, everything has been in working order.
If I disable rfkill does this mean there's no way to turn off the wifi radio, bluetooth, etc., like on a plane?
By the way, I tried upgrading to 2.6.32 and ath5k still didn't work, showing no access points, so I went back to 2.6.30 again. Is this really a bug in the kernel that hasn't been fixed for two versions? Wireless doesn't work? Are other distributions having this problem? At least on Ubuntu, with kernel 2.6.31, ath5k works fine for me on the same laptop.
Last edited by cb474 (2010-01-08 08:38:24)
Offline
I'm having disconnect issues also after upgrading to 2.6.32 in Arch. Tried downgrading to 2.6.31 but problems persisted. The machine sees the card, so that hasn't been a problem, only the inability to stay connected for any length of time.
I'm using an Acer Aspire One with an Atheros5001 chipset.
dmesg output
Linux version 2.6.32-ARCH (tobias@T-POWA-LX) (gcc version 4.4.2 20091208 (prerelease) (GCC) ) #1 SMP PREEMPT Thu Jan 7 22:19:56 UTC 2010
ADDRCONF(NETDEV_UP): wlan0: link is not ready
r8169: eth0: link down
ADDRCONF(NETDEV_UP): eth0: link is not ready
ath5k phy0: noise floor calibration timeout (2472MHz)
ath5k phy0: noise floor calibration timeout (2467MHz)
r8169: eth0: link down
ADDRCONF(NETDEV_UP): eth0: link is not ready
ADDRCONF(NETDEV_UP): wlan0: link is not ready
wlan0: direct probe to AP 00:0c:41:84:1f:94 (try 1)
wlan0: direct probe responded
wlan0: authenticate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: authenticated
wlan0: associate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: deauthenticating from 00:0c:41:84:1f:94 by local choice (reason=3)
wlan0: direct probe to AP 00:0c:41:84:1f:94 (try 1)
wlan0: deauthenticating from 00:0c:41:84:1f:94 by local choice (reason=3)
wlan0: direct probe to AP 00:0c:41:84:1f:94 (try 1)
wlan0: RX AssocResp from 00:0c:41:84:1f:94 (capab=0x5 status=0 aid=16)
wlan0: associated
ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
wlan0: deauthenticated from 00:0c:41:84:1f:94 (Reason: 6)
wlan0: no IPv6 routers present
ath5k phy0: noise floor calibration timeout (2467MHz)
r8169: eth0: link down
ADDRCONF(NETDEV_UP): eth0: link is not ready
ADDRCONF(NETDEV_UP): wlan0: link is not ready
wlan0: direct probe to AP 00:0c:41:84:1f:94 (try 1)
wlan0: direct probe responded
wlan0: authenticate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: authenticated
wlan0: associate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: RX AssocResp from 00:0c:41:84:1f:94 (capab=0x5 status=0 aid=17)
wlan0: associated
ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
wlan0: no IPv6 routers present
No probe response from AP 00:0c:41:84:1f:94 after 500ms, disconnecting.
ath5k phy0: noise floor calibration timeout (2472MHz)
ath5k phy0: noise floor calibration timeout (2472MHz)
ath5k phy0: unsupported jumbo
r8169: eth0: link down
ADDRCONF(NETDEV_UP): eth0: link is not ready
wlan0: direct probe to AP 00:0c:41:84:1f:94 (try 1)
ADDRCONF(NETDEV_UP): wlan0: link is not ready
wlan0: direct probe to AP 00:0c:41:84:1f:94 (try 2)
wlan0: direct probe responded
wlan0: authenticate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: authenticated
wlan0: associate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: RX ReassocResp from 00:0c:41:84:1f:94 (capab=0x5 status=0 aid=18)
wlan0: associated
ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
wlan0: no IPv6 routers present
No probe response from AP 00:0c:41:84:1f:94 after 500ms, disconnecting.
r8169: eth0: link down
ADDRCONF(NETDEV_UP): eth0: link is not ready
ADDRCONF(NETDEV_UP): wlan0: link is not ready
wlan0: direct probe to AP 00:0c:41:84:1f:94 (try 1)
wlan0: direct probe to AP 00:0c:41:84:1f:94 (try 2)
wlan0: direct probe responded
wlan0: authenticate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: authenticated
wlan0: associate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: RX ReassocResp from 00:0c:41:84:1f:94 (capab=0x5 status=0 aid=19)
wlan0: associated
ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
wlan0: no IPv6 routers present
No probe response from AP 00:0c:41:84:1f:94 after 500ms, disconnecting.
r8169: eth0: link down
ADDRCONF(NETDEV_UP): eth0: link is not ready
ADDRCONF(NETDEV_UP): wlan0: link is not ready
wlan0: direct probe to AP 00:0c:41:84:1f:94 (try 1)
wlan0: direct probe responded
wlan0: authenticate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: authenticated
wlan0: associate with AP 00:0c:41:84:1f:94 (try 1)
wlan0: RX ReassocResp from 00:0c:41:84:1f:94 (capab=0x5 status=0 aid=20)
wlan0: associated
ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
wlan0: no IPv6 routers present
No probe response from AP 00:0c:41:84:1f:94 after 500ms, disconnecting.
Interestingly, I am not having any wifi issues with Ubuntu 9.09 using kernel 2.6.31-17 and ath5k. Wireless connects and stays connected.
I haven't decided whether to try something else (madwifi, ndiswrapper, downgrade to 2.6.30) or wait to see if it eventually gets fixed in Arch and just continue to use Ubuntu. It's a bummer either way.
Rob
Offline
I don't know if this is relevant, or if it has already been suggested, but I have had to put the following lines in rc.local to get wireless working:
modprobe -r ath5k
modprobe -r ath_pci
modprobe ath_pci
ifconfig ath0 up
dhcpcd ath0
I specifically have to unload both the ath5k and ath_pci modules, and then reload th ath_pci module. Just blacklisting the ath5k module in rc.conf was not effective.
Jay
Offline
I spoke too soon: it froze again, and has been doing so every other boot. Upgraded to Zen 2.6.33 RC. The problem seems to be fixed here.
Offline
...kernel to 2.6.32, ath5k would cause a soft lock-up or refuse to see access points every other boot.
I've been having the same issue with ath5k for the longest time, along with many others who probably didn't realized that ath5k was the culprit. Now, I'm happy to inform you that an upstream fix has been made to the Stable Kernel 2.6.32.3. Here is the issue and fix:
-----------------------------------------------------------------
commit d5086b90c5986d33308b1ee99546a86ff1571d07
Author: Bob Copeland <me@bobcopeland.com>
Date: Mon Dec 21 22:26:48 2009 -0500
ath5k: fix SWI calibration interrupt storm
commit 242ab7ad689accafd5e87ffd22b85cf1bf7fbbef upstream.
The calibration period is now invoked by triggering a software
interrupt from within the ISR by ath5k_hw_calibration_poll()
instead of via a timer.
However, the calibration interval isn't initialized before
interrupts are enabled, so we can have a situation where an
interrupt occurs before the interval is assigned, so the
interval is actually negative. As a result, the ISR will
arm a software interrupt to schedule the tasklet, and then
rearm it when the SWI is processed, and so on, leading to a
softlockup at modprobe time.
Move the initialization order around so the calibration interval
is set before interrupts are active. Another possible fix
is to schedule the tasklet directly from the poll routine,
but I think there are additional plans for the SWI.
Signed-off-by: Bob Copeland <me@bobcopeland.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
-----------------------------------------------------------------
My system hasn't crashed ever since upgrading to kernel 2.6.32.3. Give it a test, everything should work fine now. Please mark this forum topic as "RESOLVED".
Offline
Upgrading to 2.6.32.3 does not solve the problem for me. If I attempt to use ath5k it fails to see access points every time. Madwifi works, but not ath5k.
Also "rfkill unblock all" does not solve the problem for me. Nor does blacklisting the rfkill module in rc.conf (which gets loaded anyway by something). The only solution that works for me is sticking with 2.6.30.
In addition, the problem that ejmarkow describes as being solved (soft lock-ups and ath5k not seeing access points every other boot) is not the problem that this thread was started about, I don't think. This thread was about ath5k not working at all, never seeing any access points.
So I don't think this thread should be marked resolved. I do remain dumbfounded that ath5k can be broken now through two versions of the kernel.
Offline
Upgrading to 2.6.32.3 does not solve the problem for me. If I attempt to use ath5k it fails to see access points every time. Madwifi works, but not ath5k.
Also "rfkill unblock all" does not solve the problem for me. Nor does blacklisting the rfkill module in rc.conf (which gets loaded anyway by something). The only solution that works for me is sticking with 2.6.30.
In addition, the problem that ejmarkow describes as being solved (soft lock-ups and ath5k not seeing access points every other boot) is not the problem that this thread was started about, I don't think. This thread was about ath5k not working at all, never seeing any access points.
So I don't think this thread should be marked resolved. I do remain dumbfounded that ath5k can be broken now through two versions of the kernel.
It seems this the lack of access points AND soft lock-ups were two issues mentioned in this thread relating to ath5k. Upgrading to kernel 2.6.32.3 [which includes the fix mentioned previously] solved the soft lock-up problem for me. Agreed, if the access point issue is still open, then do not mark this thread as resolved as cb474 stated.
Offline
It seems worth asking, is it only me, or are other people still having problems with ath5k on 2.6.32.3?
Offline
In dmesg:
ath5k 0000:0a:03.0: registered as 'phy1'
ath5k phy1: Invalid EEPROM checksum 0x87b6
ath5k phy1: unable to init EEPROM
ath5k: probe of 0000:0a:03.0 failed with error -5
And there is no any devices like wlan0, wifi0, eth* or ath0, after loading madwifi drivers, there is wifi0 device, but i can't scan, connect, etc.
Any solutions here not worked
// OK, fixed with madwifi & ath_pci module
Last edited by mahho (2010-01-18 13:57:41)
Offline
has anyone tried kernel26 2.6.32.5?? do the wireless issues continue?
Mikes on AUR
Offline
the issue still exists.
Pinky, are you pondering what I'm pondering?
Offline
Offline
can't find a package of kernel 2.6.30 anymore.
Pinky, are you pondering what I'm pondering?
Offline