1 |
On Saturday 26 June 2010 14.35.28 Mick wrote: |
2 |
> On Friday 25 June 2010 18:52:18 Enrico Weigelt wrote: |
3 |
> > Hi folks, |
4 |
> > |
5 |
> > |
6 |
> > my Atheros wlan (builtin, internal intenna) is regularily |
7 |
> > loosing link. Reproducible in various different networks. |
8 |
> > At home, my wlan ap is about 2 meter away (within the room), |
9 |
> > so link quality (currently 53) shouldnt be the problem. |
10 |
> > |
11 |
> > Does anyone know what could cause the problem ? |
12 |
> > |
13 |
> > # cat /proc/version |
14 |
> > |
15 |
> > Linux version 2.6.31-gentoo-r10 (root@×××××××××.local) \ |
16 |
> > (gcc version 4.3.4 (Gentoo 4.3.4 p1.0, pie-10.1.5) ) \ |
17 |
> > #1 SMP Wed Jun 2 00:51:13 CEST 2010 |
18 |
> > |
19 |
> > # lspci -v |
20 |
> > ... |
21 |
> > |
22 |
> > 02:00.0 Network controller: Atheros Communications Inc. AR9285 Wireless |
23 |
> > Network Adapter (PCI-Express) (rev 01) Subsystem: Askey Computer Corp. |
24 |
> > Device 7167 |
25 |
> > |
26 |
> > Flags: bus master, fast devsel, latency 0, IRQ 16 |
27 |
> > Memory at f6000000 (64-bit, non-prefetchable) [size=64K] |
28 |
> > Capabilities: [40] Power Management version 3 |
29 |
> > Capabilities: [50] MSI: Enable- Count=1/1 Maskable- 64bit- |
30 |
> > Capabilities: [60] Express Legacy Endpoint, MSI 00 |
31 |
> > Capabilities: [100] Advanced Error Reporting |
32 |
> > Capabilities: [140] Virtual Channel <?> |
33 |
> > Capabilities: [160] Device Serial Number 00-15-17-ff-ff-24-14-12 |
34 |
> > Capabilities: [170] Power Budgeting <?> |
35 |
> > Kernel driver in use: ath9k |
36 |
> > Kernel modules: ath9k |
37 |
> > |
38 |
> > ... |
39 |
> > |
40 |
> > |
41 |
> > Jun 25 19:36:51 excalibur kernel: wlan0: no probe response from AP |
42 |
> > 00:23:08:86:d6:8f - disassociating Jun 25 19:36:51 excalibur |
43 |
> > dhcpcd[10182]: wlan0: carrier lost |
44 |
> |
45 |
> I have seen the same problem with my internal broadcom card and of course a |
46 |
> different driver. If you check google there seem to be pages and pages of |
47 |
> users reporting such a problem on different distros and with different |
48 |
> makes of wireless cards. |
49 |
> |
50 |
> The jury's out as to what's causing this. Is it related to modern cards |
51 |
> with low power management capabilities and how this may interact with the |
52 |
> kernel, or is it something to do with the tolerance built into TTL packets |
53 |
> between the card and the AP? |
54 |
> |
55 |
> With regards to my card I have noticed that at home I stay connected for |
56 |
> hours on end, at work it's a miracle if I stay online for longer than 5 |
57 |
> minutes (both on the same channel). This to me says that the problem is |
58 |
> one of interaction with the router, which points to tolerance on the TTL |
59 |
> packets. |
60 |
> |
61 |
> Of course YMMV ... |
62 |
|
63 |
|
64 |
Last time I had this problem I tracked it to a conflict between wpa_supplicant |
65 |
and the card driver. Basicly the card driver would send a request to assosiate |
66 |
with mac 00:00:....:00 each time a new card was found on the air and |
67 |
wpa_supplicatnt interpreted this as a request from the driver to drop the |
68 |
connection. The driver folks said that this request should be ignored and the |
69 |
wpa_supplicant folks said that this should not be ignored but instead was a |
70 |
disconnect request. Catch 22... |
71 |
|
72 |
Because of this I had to use wep instead of wpa(2) for along time. |
73 |
|
74 |
/Naga |