Worsened ath9k wifi quality recently (since a year ago)

erikstep · 2015-08-30 08:24:45

I'm on an Asus EeePc X101CH.
It's 2 months since I had installed Arch on my netbook. Before I had had lots of distros. My first one'd been Fedora, then Debian Squeeze/Wheezy, eventually Manjaro and (pure) Arch. In the first ones (Fedora and Debian) I used to have good wireless performance, but when I came to rolling-releases wifi got really worse in strenght.

Here you are some informations concerning my network card:

$ lspci -vv -s 02:00.0 (id of my wireless card)

02:00.0 Network controller: Qualcomm Atheros AR9485 Wireless Network Adapter (rev 01)
	Subsystem: Lite-On Communications Inc Device 6627
	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
	Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
	Latency: 0, Cache Line Size: 64 bytes
	Interrupt: pin A routed to IRQ 17
	Region 0: Memory at dfe00000 (64-bit, non-prefetchable) [size=512K]
	Expansion ROM at dfe80000 [disabled] [size=64K]
	Capabilities: <access denied>
	Kernel driver in use: ath9k
	Kernel modules: ath9k

$ iwconfig wlp2s0

wlp2s0    IEEE 802.11bgn  ESSID:"xxxxxxxxxxxx"  (sorry but it's my privacy, and it surely wouldn't help you)
          Mode:Managed  Frequency:2.422 GHz  Access Point: XX:XX:XX:XX:XX:XX   
          Bit Rate=58.5 Mb/s   Tx-Power=15 dBm   
          Retry short limit:7   RTS thr:off   Fragment thr:off
          Power Management:off
          Link Quality=30/70  Signal level=-80 dBm  
          Rx invalid nwid:0  Rx invalid crypt:0  Rx invalid frag:0
          Tx excessive retries:19  Invalid misc:199   Missed beacon:0

$ watch -n 1 cat /proc/net/wireless (average after 60 seconds)

Inter-| sta-|   Quality        |   Discarded packets               | Missed | WE
 face | tus | link level noise |  nwid  crypt   frag  retry   misc | beacon | 22
wlp2s0: 0000   26.  -84.  -256        0      0      0     19    200        0

$ ping -c 20 192.168.1.1

PING 192.168.1.1 (192.168.1.1) 56(84) bytes of data.
64 bytes from 192.168.1.1: icmp_seq=1 ttl=64 time=6.55 ms
64 bytes from 192.168.1.1: icmp_seq=2 ttl=64 time=2.12 ms
64 bytes from 192.168.1.1: icmp_seq=3 ttl=64 time=1.98 ms
64 bytes from 192.168.1.1: icmp_seq=4 ttl=64 time=4.92 ms
64 bytes from 192.168.1.1: icmp_seq=5 ttl=64 time=2.55 ms
[cut...]
--- 192.168.1.1 ping statistics ---
20 packets transmitted, 19 received, 5% packet loss, time 19041ms
rtt min/avg/max/mdev = 1.618/2.701/6.558/1.259 ms

$ ping -c 20 google.it
[cut...]
--- google.it ping statistics ---
20 packets transmitted, 18 received, 10% packet loss, time 23123ms
rtt min/avg/max/mdev = 36.284/38.341/40.726/1.086 ms

What I noticed is that I obtain more or less the same values even when I am 50 centimetres far from my router. link quality is even lower when I'm next to it. I've never reached a value greater than 40/70, and neither does the signal level: -80db is my standard value (and i3status still tells me I've got a signal of 30-40% while my mother's notebook reaches higher values) and gets slightly lower when I'm next to the router (I know that db is a logaritmic value, but don't you think it's got the same value even when I'm 1 centimetre far from my router?).

Once it wasn't like that. I didn't have any problem in connecting, whereas now I require 3-4 attempts to connect when I'm far from my router. In my opinion the fault is of the kernel (the ath9k driver), so doing a downgrade would help, but it's a pity that linux-lts is "too new", so I should draw upon vanilla sources.
What should I do in order to improve wifi quality?

Head_on_a_Stick · 2015-08-30 16:15:33

Have you tried fiddling with the module options for the ath9k driver?

You can try enabling software encryption for a start.
https://wiki.archlinux.org/index.php/Wi … tion#ath9k

erikstep · 2015-08-30 17:54:22

Yes, shortly after posting that I found that page. But it didn't help at all.

Edit:

Just tried linux-lts: as I suspected, it was useless. Maybe changing kernel version is no use?
Maybe, though this makes me really annoyed and puzzled: what is causing this weird behaviour? Why is wifi so unstable?

I'll try a Debian Wheezy live (which uses, if I am right, linux 3.2 <= x <= 3.14) and tell you if something changes

Last edited by erikstep (2015-08-30 19:13:29)

Head_on_a_Stick · 2015-08-30 20:11:18

erikstep wrote:

Why is wifi so unstable?

Is there anything in the logs?

twelveeighty · 2015-08-30 23:10:59

erikstep wrote:

Why is wifi so unstable?

Because some hardware manufacturers do not provide open drivers at all (Broadcom) and/or some do not contribute enough manpower to keep up with the newer kernel versions. Non-rolling distros or ones that do not update the kernel frequently tend to have better wireless stability for some brands because of that flaw.

Check your logs between the older kernels and the current Arch one for that same hardware. There's bound to be something in the logs with failing wireless drivers.

erikstep · 2015-09-01 08:28:13

twelveeighty wrote:

erikstep wrote:
Why is wifi so unstable?
Because some hardware manufacturers do not provide open drivers at all (Broadcom) and/or some do not contribute enough manpower to keep up with the newer kernel versions. Non-rolling distros or ones that do not update the kernel frequently tend to have better wireless stability for some brands because of that flaw.

You're right, that isn't my case though. According to ¹ (look up below) Qualcomm Atheros appears to release FOSS drivers. And, according to ², its hardware seems to be really common in routers and laptops.

By the way: I found out very much time ago that ath9k driver, I'll say it frankly, sucks. Wifi stability is not the only one problem.

Firstly: let's talk about leds. It may be just an esthetic disappoint, but I can't realize why it does not work anymore. In previous kernel versions (I tested it yesterday) it worked. Manually, because you had to do '# echo rfkill0 > /sys/class/led/ath9k-phy0/trigger', but you could have used a script, like ³ in Ubuntu.

Secondly: ad-hoc connection. With Debian I could have an ad-hoc connection with my "home-server", which was suitable when I had to "rsync" lots of thing or wanted a direct WLAN connection. Now I can't. It creates a new ad-hoc connection, but actually it does not.
Sorry, I was wrong because it was I who misunderstood the procedure to create an ad-hoc connection (useless, to be honest, because its radius is less than one inch from the netbook).

Thirdly: monitoring mode doesn't work anymore (but it used to).

Shall I continue? But I can't deal with the fact that there have been so many problems and no developers start working to a big hotfix (they're really critical for me).

Ok... I just need to vent. It isn't your fault, obviously. We'll talk about them in another moment.
Here I am about to tell you what happened yesterday with Debian:
Wifi worked. Nearly well, but not so much as I expected (maybe for nostalgia?).

Anyway: it was more stable but couldn't detect so many SSIDs (actually, only one, while my mother's notebook and my home server about fifteen; consider that my home server uses an usb adapter, a TP-Link TL-WDN3200, which shows, on this netbook, about 18 networks, thank you condominium!).

Here's a little log of dmesg where I tried to connect in a place where I notoriously can't connect:

> [ 1528.939656] wlan0: authenticate with XX:XX:XX:XX:XX:XX (try 1)
> [ 1528.942595] wlan0: authenticated
> [ 1528.960063] wlan0: associate with XX:XX:XX:XX:XX:XX (try 1)
> [ 1528.964304] wlan0: RX ReassocResp from XX:XX:XX:XX:XX:XX (capab=0x411 status=0 aid=1)
> [ 1528.964314] wlan0: associated
> [ 1531.972642] wlan0: deauthenticated from XX:XX:XX:XX:XX:XX (Reason: 2)
> [ 1532.022095] cfg80211: Calling CRDA to update world regulatory domain
> [ 1532.059729] cfg80211: World regulatory domain updated:
> [ 1532.059737] cfg80211:     (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp)
> [ 1532.059745] cfg80211:     (2402000 KHz - 2472000 KHz @ 40000 KHz), (N/A, 2000 mBm)
> [ 1532.059751] cfg80211:     (2457000 KHz - 2482000 KHz @ 40000 KHz), (N/A, 2000 mBm)
> [ 1532.059757] cfg80211:     (2474000 KHz - 2494000 KHz @ 20000 KHz), (N/A, 2000 mBm)
> [ 1532.059764] cfg80211:     (5170000 KHz - 5250000 KHz @ 80000 KHz), (N/A, 2000 mBm)
> [ 1532.059770] cfg80211:     (5735000 KHz - 5835000 KHz @ 80000 KHz), (N/A, 2000 mBm)
> [ 1532.059776] cfg80211:     (57240000 KHz - 63720000 KHz @ 2160000 KHz), (N/A, 0 mBm)

That's the log I obtained today with Arch

> [ 4832.931783] wlp2s0: authenticate with XX:XX:XX:XX:XX:XX
> [ 4832.950435] wlp2s0: direct probe to XX:XX:XX:XX:XX:XX (try 1/3)
> [ 4833.152594] wlp2s0: direct probe to XX:XX:XX:XX:XX:XX (try 2/3)
> [ 4833.355781] wlp2s0: direct probe to XX:XX:XX:XX:XX:XX (try 3/3)
> [ 4833.559036] wlp2s0: authentication with XX:XX:XX:XX:XX:XX timed out

(sometimes I obtain that)
wlp2s0: deauthenticated from XX:XX:XX:XX:XX:XX (Reason: 2=PREV_AUTH_NOT_VALID)

As you can see the logs are radically different (SysV vs. systemd) and say more or less the same thing.

¹ https://wireless.wiki.kernel.org/en/use … #licensing
² https://wireless.wiki.kernel.org/en/users/drivers/ath9k
³ http://ubuntuforums.org/showthread.php?t=966511&page=9

Last edited by erikstep (2015-09-03 10:24:46)

Head_on_a_Stick · 2015-09-01 12:17:10

ArchWiki page I linked above wrote:

In the unlikely event that you have stability issues that trouble you, you could try using the backports-patched package.

erikstep · 2015-09-03 08:12:26

Just tried backports-patched, it didn't work. I mean, wifi works as before installing that patch. Ad-hoc connection still does not work and neither does led. What's more, the i3status bar tells me I'm not connected, but according to NetworkManager I am (I forgot WEXT).

Here's a piece of log with patched modules:

[  587.030743] wlp2s0: authenticate with XX:XX:XX:XX:XX:XX
[  587.046002] wlp2s0: send auth to XX:XX:XX:XX:XX:XX (try 1/3)
[  587.099825] wlp2s0: send auth to XX:XX:XX:XX:XX:XX (try 2/3)
[  587.168365] wlp2s0: send auth to XX:XX:XX:XX:XX:XX (try 3/3)
[  587.200235] wlp2s0: authenticated
[  587.202680] wlp2s0: associate with XX:XX:XX:XX:XX:XX (try 1/3)
[  587.259329] wlp2s0: associate with XX:XX:XX:XX:XX:XX (try 2/3)
[  587.326787] wlp2s0: associate with XX:XX:XX:XX:XX:XX (try 3/3)
[  587.336808] wlp2s0: RX AssocResp from XX:XX:XX:XX:XX:XX (capab=0x411 status=0 aid=1)
[  587.337074] wlp2s0: associated
[  591.386793] wlp2s0: deauthenticated from XX:XX:XX:XX:XX:XX (Reason: 2=PREV_AUTH_NOT_VALID)
[  591.416460] cfg80211: Exceeded CRDA call max attempts. Not calling CRDA
[  615.795741] wlp2s0: authenticate with XX:XX:XX:XX:XX:XX
[  615.819851] wlp2s0: send auth to XX:XX:XX:XX:XX:XX (try 1/3)
[  615.824654] wlp2s0: authenticated
[  615.825500] wlp2s0: associate with XX:XX:XX:XX:XX:XX (try 1/3)
[  615.834048] wlp2s0: RX AssocResp from XX:XX:XX:XX:XX:XX (capab=0x411 status=0 aid=1)
[  615.834269] wlp2s0: associated

Edit: ad-hoc works. No, it does not depend of this patch, but of me, who misunderstood the procedure

Last edited by erikstep (2015-09-03 10:25:56)

erikstep · 2015-09-06 15:55:06

I just tried an external USB wireless adapter. It works without problems: I'm wondering if it's been useful to struggle so much about it without just ignore it and use an external adapter. Yes, it could seem like throwing in the towel but, forgive me, I'm just sick and tired of always roam aimlessly at home only to find a place where signal strength is neither decent.

Anyway, I'd like to thank you for you support, even though there wasn't much to help.

twelveeighty · 2015-09-07 18:58:17

erikstep wrote:

Shall I continue? But I can't deal with the fact that there have been so many problems and no developers start working to a big hotfix (they're really critical for me).

No, please stop. You clearly have no idea who those "developers" are that need to start working for you on that "big critical hotfix". Either start contributing yourself or stop complaining.

Noroimusha · 2015-10-02 20:12:36

I think I have similar issues with ath9k.
After booting, when netctl starts the automatic profile, the strength is around 60% right in front of the router.
However, when I start a manual scan via

# wifi-menu

or

# iw interface_name scan

the signal strength returns to 100%.

Can you try to trigger a manual scan and report if it restores your signal strength?

Last edited by Noroimusha (2015-10-02 20:14:53)

Arch Linux

#1 2015-08-30 08:24:45

Worsened ath9k wifi quality recently (since a year ago)

#2 2015-08-30 16:15:33

Re: Worsened ath9k wifi quality recently (since a year ago)

#3 2015-08-30 17:54:22

Re: Worsened ath9k wifi quality recently (since a year ago)

#4 2015-08-30 20:11:18

Re: Worsened ath9k wifi quality recently (since a year ago)

#5 2015-08-30 23:10:59

Re: Worsened ath9k wifi quality recently (since a year ago)

#6 2015-09-01 08:28:13

Re: Worsened ath9k wifi quality recently (since a year ago)

#7 2015-09-01 12:17:10

Re: Worsened ath9k wifi quality recently (since a year ago)

#8 2015-09-03 08:12:26

Re: Worsened ath9k wifi quality recently (since a year ago)

#9 2015-09-06 15:55:06

Re: Worsened ath9k wifi quality recently (since a year ago)

#10 2015-09-07 18:58:17

Re: Worsened ath9k wifi quality recently (since a year ago)

#11 2015-10-02 20:12:36

Re: Worsened ath9k wifi quality recently (since a year ago)

Board footer