ar5523 driver

Riccardo Magliocchetti riccardo.magliocchetti at gmail.com
Mon Nov 9 20:04:25 UTC 2009


Hi,

Stefan Lippers-Hollmann ha scritto:
> Hi
> 
> Sorry about the late reply, I've been traveling till this evening.
> 
> On Saturday 07 November 2009, you wrote:
>> Hello,
>>
>> in the hope that this driver will drive my gigaset USB stick 108 
>> (0x129b, 0x160c) i've updated it to compile with latest git. The driver 
>> now compiles fine, loads and unloads without problems. I've added my 
>> device as
> 
> Thank you a lot, I didn't get around to that yet and getting it to compile
> and work with current kernels is the most important aspect to actually do
> real debugging and further development without hitting walls in combination
> with ancient kernels all the time :)
> 
>> +       AR5523_DEVICE_UG(0x129b, 0x160c),       /* Gigaset / USB stick 
>> 108 (CyberTAN Technology) */
>>
>> but the driver fails with: ar5523: probe of 1-1:1.0 failed with error -5
> 
> The firmware is present? Does the device eventually provide windows drivers
> on fake usb-storage and require mode switching (some ar5523 seem to do 
> this)? Could you provide the complete messages upon connecting the wlan 
> card to the already running system?

Yes i have the firmware, with dmesg i have tons of errors i don't have 
in my /var/log/messages. Will try to dig the code to see if it is 
something i touched in the patch.

[ 2053.461071] usb 1-1: new high speed USB device using ehci_hcd and 
address 4
[ 2053.594920] usb 1-1: New USB device found, idVendor=129b, idProduct=160c
[ 2053.594928] usb 1-1: New USB device strings: Mfr=1, Product=2, 
SerialNumber=3
[ 2053.594935] usb 1-1: Product: AR5523
[ 2053.594940] usb 1-1: Manufacturer: Atheros Communications Inc
[ 2053.594945] usb 1-1: SerialNumber: 1.0
[ 2053.595140] usb 1-1: configuration #1 chosen from 1 choice
[ 2058.593095] usb 1-1: failed to reset device - initialization timeout
[ 2058.593105] usb 1-1: could not initialize adapter
[ 2058.593281] usb 1-1: error -1 when resubmitting rx urb
[ 2058.593388] usb 1-1: error -1 when resubmitting rx urb
[ 2058.593525] usb 1-1: error -1 when resubmitting rx urb
[ 2058.593634] usb 1-1: error -1 when resubmitting rx urb
[ 2058.593774] usb 1-1: error -1 when resubmitting rx urb
[ 2058.593910] usb 1-1: error -1 when resubmitting rx urb
[ 2058.594006] usb 1-1: error -1 when resubmitting rx urb
[ 2058.594150] usb 1-1: error -1 when resubmitting rx urb
[ 2058.594276] usb 1-1: error -1 when resubmitting rx urb
[ 2058.594400] usb 1-1: error -1 when resubmitting rx urb
[ 2058.594524] usb 1-1: error -1 when resubmitting rx urb
[ 2058.594650] usb 1-1: error -1 when resubmitting rx urb
[ 2058.594777] usb 1-1: error -1 when resubmitting rx urb
[ 2058.594900] usb 1-1: error -1 when resubmitting rx urb
[ 2058.595024] usb 1-1: error -1 when resubmitting rx urb
[ 2058.595138] usb 1-1: error -1 when resubmitting rx urb
[ 2058.595275] usb 1-1: error -1 when resubmitting rx urb
[ 2058.595400] usb 1-1: error -1 when resubmitting rx urb
[ 2058.595525] usb 1-1: error -1 when resubmitting rx urb
[ 2058.595649] usb 1-1: error -1 when resubmitting rx urb
[ 2058.595775] usb 1-1: error -1 when resubmitting rx urb
[ 2058.595899] usb 1-1: error -1 when resubmitting rx urb
[ 2058.596025] usb 1-1: error -1 when resubmitting rx urb
[ 2058.596150] usb 1-1: error -1 when resubmitting rx urb
[ 2058.596277] usb 1-1: error -1 when resubmitting rx urb
[ 2058.596384] usb 1-1: error -1 when resubmitting rx urb
[ 2058.596510] usb 1-1: error -1 when resubmitting rx urb
[ 2058.596899] usb 1-1: error -1 when resubmitting rx urb
[ 2058.597026] usb 1-1: error -1 when resubmitting rx urb
[ 2058.597138] usb 1-1: error -1 when resubmitting rx urb
[ 2058.597220] ar5523: probe of 1-1:1.0 failed with error -5

> 
> $ md5sum -b /lib/firmware/uath-ar5523.bin
> 3b976a7f9319289ab6c5891bc103a431 */lib/firmware/uath-ar5523.bin

$ md5sum uath-ar5523.bin
3b976a7f9319289ab6c5891bc103a431  uath-ar5523.bin

> I also have a further stability patch in testing, which might eventually 
> also improve support for your device, I'll commit it to svn tonight as 
> well. That said, while this driver has been ported from OpenBSD's uath, 
> it's still in a pretty early and rough state and in need of further 
> debugging and development.
> 
>> What does this mean? Also how do i know if my device is UG or UX?
> 
> UG is 802.11 b/g (super-g),			http://www.atheros.com/pt/AR5005UG.htm
> UX includes 5 GHz operations on top,	http://www.atheros.com/pt/AR5005UX.htm

thanks

>> I'm attaching the diff with kcompat* debian patches applied on top of 
>> svn for someone with more clue than me to review and hopefully test and 
> 
> Thanks, Atheros is trying to keep the driver dual licensed GPL/BSD, in 
> order to share development with *BSD and to allow code sharing with the
> other drivers under drivers/net/wireless/ath/ (ath5k, ath9k, ar9170usb and 
> ath9k_htc), would you agree to committing your patch with the following 
> license/ signed-off header (which follows the way ath5k patches are 
> handled)?
> 
> 
> <patch description, along the lines of: update mac80211 api for 2.6.32-rc6>
> 
> Changes-licensed-under: ISC
> Signed-off-by: Riccardo Magliocchetti <riccardo.magliocchetti at gmail.com>

Sure, will resubmit with problem description and Signed-off-by line once 
you have some feedback for me :)

>> apply. BTW, wouldn't be better to have the driver in staging with at 
>> least the debian patches already applied?
> 
> Not being compatible with current git was precluding it from being merged 
> anywhere so far. This is also the reason why I added minimal support to
> build it out-of-tree and with quilt patch series as part of the preliminary
> packaging, so I could easily test it against (live) ISOs with ancient 
> kernels and run basic "bisecting" in case of newly introduced regressions.
> With that resolved, we can indeed think about getting it in a public tree,
> that may be wireless-testing (as it is based on mac80211) or maybe just 
> Greg's staging tree - I will discuss this with my contact at Atheros as 
> soon as I've had a chance to actually test the patch (hopefully this 
> evening).
> 
> For being merged mainline, the following to-do list is still open:
> - (updating to current kernels <-- thanks a lot)
> - checking for changes/ bugfixes in OpenBSD's uath, since ar5523 has been 
>   forked from it
> - adding wpa (nl80211) support, so far it doesn't support encryption.
> 
>> thanks,
>> riccardo
> 
> Thank you a lot
> 	Stefan Lippers-Hollmann

thanks,
riccardo



More information about the devel mailing list