Red Hat Bugzilla – Bug 506372
ath5k driver dropping packets
Last modified: 2010-06-28 09:04:40 EDT
Description of problem:
When using ath5k driver on an Atheros AR2413 driver network access is intermittent. Presuming that packets are being dropped/lost.
Version-Release number of selected component (if applicable):
Fedora 11 with kernel 184.108.40.206-167.fc11.i686.PAE
Everytime by just connecting and attempting to access network.
Steps to Reproduce:
1. Enable wireless networking
2. Tried to access websites or ssh to other server - works occasionally
Access works occasionally - if it does, slowly.
On Fedora 10 with madwifi drivers worked fine.
Running lspci | grep Atheros gives:
09:04.0 Ethernet controller: Atheros Communications Inc. AR2413 802.11bg NIC (rev 01)
Will attempt building madwifi for Fedora 11 to see if it makes a difference.
I seem to have the exact problem with my AR928X card.
Can you provide wireless traces using wireshare (or equivalent) on another box?
Not sure what 'wireshare' is - did you mean wireshark?
If so I will get out another box this weekend with wireless and try and gather some data.
For reference, the madwifi drivers work fine on the same box.
wireshark -- sorry, typo!
try checking with iwconfig to see if your wlan card is doing a network scan while you see the connection stall. This is the problem I am seeing. If this is the case, I think there was another bugzilla thread on this (but I can't find it now). I remember that there someone was saying that things should get better in 2.6.31 kernel. I am running F12 beta witth 2.6.31 and I can see that things got better but I still occasionally see that the connection stalls.
03:00.0 Ethernet controller: Atheros Communications Inc. AR242x 802.11abg Wireless PCI Express Adapter (rev 01)
I'm seeing similar problems with 220.127.116.11-170.2.113.fc10.i686 in an Acer notebook that has 'Ethernet controller: Atheros Communications Inc. AR242x 802.11abg Wireless PCI Express Adapter (rev 01)'. When the network connection dies, I see the following in dmesg:
wlan0: no IPv6 routers present
wlan0: authenticate with AP 00:16:b6:da:8a:44
wlan0: associate with AP 00:16:b6:da:8a:44
wlan0: RX ReassocResp from 00:16:b6:da:8a:44 (capab=0x421 status=0 aid=1)
wlan0: No ProbeResp from current AP 00:16:b6:da:8a:44 - assume out of range
The problem still persists on up to date FC12 with 18.104.22.168-174.fc12.i686.PAE.
Really annoying problem. Is there any way to forbid networkmanager for doing a scan? (I am guessing it is the networkmanager doing this)
Can you try one of the later stable releases?
With compat-wireless-2.6.33-rc1 things are much better (didn't try earlier ones):
248 packets transmitted, 246 received, 0% packet loss, time 248215ms
rtt min/avg/max/mdev = 0.844/14.598/225.449/40.783 ms
The worst response times used to be 1000 - 3000 ms. The response occasionally goes to abt 200 ms for some individual packets. Great improvement!
Would be great if we could get a fix for this for current kernels. Recent kernel upgrade took things back to the old problem and had to reinstall wireless-compat again.
F12 will hopefully soon get a 2.6.32-based kernel as an update. You can give kernel-22.214.171.124-30.fc12 a try, get it from Koji:
If it fixes the problem, we just need to wait for the update to be done.
If it does not fix the problem, we'll need to identify which patches between 2.6.32 and 2.6.33-rc1 are needed. I'll make a test kernel build with some likely candidates applied if necessary.
With 126.96.36.199-30.fc12.i686.PAE the behavior is different. The response times do not go occasionally but it drops packets completely:
1714 packets transmitted, 1605 received, 6% packet loss, time 1716024ms
rtt min/avg/max/mdev = 0.782/2.795/1378.196/34.484 ms, pipe 2
I was pinging my basestation. With the compat-wireless that I mentioned above this did not happen. So something is still not quite right.
Jussi, could you download this kernel:
and this tarball with several built ath5k modules with picked patches from 2.6.33-rc1:
Install the kernel and test it. The expected result is the same as in your comment #12.
Then unpack the tarball where you'll find several directories called "ath-$number". In every one of these there are module files ath5k.ko and ath.ko.
Try using each pair of the modules and observe which one fixes the bug:
Please try ath-33 first. It corresponds to the version from 2.6.33-rc1, so it is expected to work. Then do a binary search through the lower numbered ones to find which is the first good one.
Well, ath-33 does not work the same way as the compat stuff did. In fact it hardly works at all. Pinging my base station gives:
115 packets transmitted, 95 received, 17% packet loss, time 114932ms
rtt min/avg/max/mdev = 78.118/89.205/297.613/36.288 ms
With the standard ath5k in kernel 188.8.131.52-39.fc12.i686.PAE I got:
301 packets transmitted, 277 received, 7% packet loss, time 300757ms
rtt min/avg/max/mdev = 1.183/1.495/4.325/0.371 ms
Should I try the latest compat-wireless with this particular kernel again?
Interesting... I unloaded and loaded the drivers again and I see different behavior (not correct either...):
120 packets transmitted, 100 received, 16% packet loss, time 119324ms
rtt min/avg/max/mdev = 1.217/1.560/2.831/0.391 ms
Packet loss is about the same but the response time is smaller.
Also, this is what I see in dmesg (if it helps..):
ath5k 0000:03:00.0: PCI INT A disabled
ath5k 0000:03:00.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17
ath5k 0000:03:00.0: setting latency timer to 64
ath5k 0000:03:00.0: registered as 'phy5'
ath: EEPROM regdomain: 0x65
ath: EEPROM indicates we should expect a direct regpair map
ath: Country alpha2 being used: 00
ath: Regpair used: 0x65
phy5: Selected rate control algorithm 'minstrel'
ath5k phy5: Atheros AR5414 chip found (MAC: 0xa3, PHY: 0x61)
Quick test also showed that none of the smaller number ath's helped in reducing the packet loss either.
This message is a reminder that Fedora 11 is nearing its end of life.
Approximately 30 (thirty) days from now Fedora will stop maintaining
and issuing updates for Fedora 11. It is Fedora's policy to close all
bug reports from releases that are no longer maintained. At that time
this bug will be closed as WONTFIX if it remains open with a Fedora
'version' of '11'.
Package Maintainer: If you wish for this bug to remain open because you
plan to fix it in a currently maintained version, simply change the 'version'
to a later Fedora version prior to Fedora 11's end of life.
Bug Reporter: Thank you for reporting this issue and we are sorry that
we may not be able to fix it before Fedora 11 is end of life. If you
would still like to see this bug fixed and are able to reproduce it
against a later version of Fedora please change the 'version' of this
bug to the applicable version. If you are unable to change the version,
please add a comment here and someone will do it for you.
Although we aim to fix as many bugs as possible during every release's
lifetime, sometimes those efforts are overtaken by events. Often a
more recent Fedora release includes newer upstream software that fixes
bugs or makes them obsolete.
The process we are following is described here:
Fedora 11 changed to end-of-life (EOL) status on 2010-06-25. Fedora 11 is
no longer maintained, which means that it will not receive any further
security or bug fix updates. As a result we are closing this bug.
If you can reproduce this bug against a currently maintained version of
Fedora please feel free to reopen this bug against that version.
Thank you for reporting this bug and we are sorry it could not be fixed.