Re: rtl8192cu gets confused when scan is aborted by bringing interface down (Re: rtl8192cu goes silent/dead after some time...)

Re: rtl8192cu gets confused when scan is aborted by bringing interface down (Re: rtl8192cu goes silent/dead after some time...)

[Alessandro Lannocca]

Hi, some months ago I reported some bugs about this chipset; now I tried latest stable compat-drivers (3.9-rc2-2-su) with the modified patch by Jussi Kivilinna, and I'm happy to report that it works  with my alfa AWUS036NHR on ubuntu 12.10 with kernel 3.5; The card now sustains multiple connection/disconnection cicles, and doesn't go mute, even after changing mac address.

Some problems still remains: 

-led is always solid, no blinking whatsoever.
-in monitor mode, every client appears as not associated, even when it really is.
-power/signal strength readings (in networkmanager) seem to be fuzzy, almost inverse; nearest APs have lowest signal representation, while farest ones get strongest signal (this could very well be a networkmanager bug/misrepresentation); signal appears normal when the card is connected (only the reading relative to the connected AP)

I know you're probably understaffed and have a lot of work to do, however if you're willing to have a look at this secondary problems (Mr. Finger perhaps), I'm willing to recompile, test and report back to help squash these bugs.

I'm attaching a debug=5 log to show my tests.

As of now, this chipset/card is usable (signal reading is annoying btw)

Thank you all for your work.

Alessandro Lannocca

-----Messaggio originale-----
Da: linux-wireless-owner@vger.kernel.org [mailto:linux-wireless-owner@vger.kernel.org] Per conto di Jussi Kivilinna
Inviato: martedì 12 marzo 2013 20.49
A: Larry Finger
Cc: linux-wireless@vger.kernel.org; 'George0505'
Oggetto: Re: rtl8192cu gets confused when scan is aborted by bringing interface down (Re: rtl8192cu goes silent/dead after some time...)

On 12.03.2013 18:51, Larry Finger wrote:
> On 03/12/2013 04:10 AM, Jussi Kivilinna wrote:
>>
>> That patch alone did not help.. however replacing
>> _rtl92cu_set_check_bssid()
>> with that new rtl92cu_set_check_bssid() does fix the issue (patch 
>> attached).
>>
>> But I think _rtl92cu_set_check_bssid() needs to be rechecked since it 
>> has '(IS_NORMAL_CHIP(rtlhal->version))' conditional etc.
> 
> I doubt that that IS_NORMAL_CHIP business affects us - the test chips 
> are not supposed to be in the wild, but I modified your patch to 
> include that test.
> 

Modified patch works here.

> With that patch I am getting watchdog_wq callbacks that force a 
> reconnection every 6 seconds, but that is due to other changes in my 
> code. The result is a 34% loss of pings, but the connection stays up.
> 

No problems here.. <1% loss and can do large transfers, rtl8192cu being able to reconnect is connection is dropped.

-Jussi
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at  http://vger.kernel.org/majordomo-info.html

R: rtl8192cu gets confused when scan is aborted by bringing interface down (Re: rtl8192cu goes silent/dead after some time...)

[Alessandro Lannocca]

[-- Attachment #1: Type: text/plain, Size: 31 bytes --]

Sorry, attachment was missing.

[-- Attachment #2: wifi.log.gz --]
[-- Type: application/octet-stream, Size: 201267 bytes --]

Re: rtl8192cu gets confused when scan is aborted by bringing interface down (Re: rtl8192cu goes silent/dead after some time...)

[Larry Finger]

On 03/13/2013 10:11 AM, Alessandro Lannocca wrote:
> Hi, some months ago I reported some bugs about this chipset; now I tried latest stable compat-drivers (3.9-rc2-2-su) with the modified patch by Jussi Kivilinna, and I'm happy to report that it works  with my alfa AWUS036NHR on ubuntu 12.10 with kernel 3.5; The card now sustains multiple connection/disconnection cicles, and doesn't go mute, even after changing mac address.
>
> Some problems still remains:
>
> -led is always solid, no blinking whatsoever.
> -in monitor mode, every client appears as not associated, even when it really is.
> -power/signal strength readings (in networkmanager) seem to be fuzzy, almost inverse; nearest APs have lowest signal representation, while farest ones get strongest signal (this could very well be a networkmanager bug/misrepresentation); signal appears normal when the card is connected (only the reading relative to the connected AP)
>
> I know you're probably understaffed and have a lot of work to do, however if you're willing to have a look at this secondary problems (Mr. Finger perhaps), I'm willing to recompile, test and report back to help squash these bugs.
>
> I'm attaching a debug=5 log to show my tests.
>
> As of now, this chipset/card is usable (signal reading is annoying btw)

Thanks for the report. That patch will be pushed upstream in a few minutes and 
will be backported to stable withing a few weeks. I am happy to hear that the 
patch works with your Alfa AWUS036NHR. In my limited testing with that device, 
it failed to connect. I feared that the RTL8192RU differed from the RTL8192CU.

I know that improper operation of the LEDs and screwy signal levels are 
annoying, and I will get to them when I can. At the moment, however, I have more 
important problems to solve.

Thanks,

Larry

R: rtl8192cu gets confused when scan is aborted by bringing interface down (Re: rtl8192cu goes silent/dead after some time...)

[Alessandro Lannocca]

I understand this bugs are lower priority, It's just a huge step forward to have this card in a working state, we'll wait for the rest to get sorted whenever possible.

Thanks

Alessandro Lannocca

-----Messaggio originale-----
Da: Larry Finger [mailto:larry.finger@gmail.com] Per conto di Larry Finger
Inviato: mercoledì 13 marzo 2013 16.26
A: Alessandro Lannocca
Cc: 'Jussi Kivilinna'; linux-wireless@vger.kernel.org; 'George0505'
Oggetto: Re: rtl8192cu gets confused when scan is aborted by bringing interface down (Re: rtl8192cu goes silent/dead after some time...)

On 03/13/2013 10:11 AM, Alessandro Lannocca wrote:
> Hi, some months ago I reported some bugs about this chipset; now I tried latest stable compat-drivers (3.9-rc2-2-su) with the modified patch by Jussi Kivilinna, and I'm happy to report that it works  with my alfa AWUS036NHR on ubuntu 12.10 with kernel 3.5; The card now sustains multiple connection/disconnection cicles, and doesn't go mute, even after changing mac address.
>
> Some problems still remains:
>
> -led is always solid, no blinking whatsoever.
> -in monitor mode, every client appears as not associated, even when it really is.
> -power/signal strength readings (in networkmanager) seem to be fuzzy, 
> almost inverse; nearest APs have lowest signal representation, while 
> farest ones get strongest signal (this could very well be a 
> networkmanager bug/misrepresentation); signal appears normal when the 
> card is connected (only the reading relative to the connected AP)
>
> I know you're probably understaffed and have a lot of work to do, however if you're willing to have a look at this secondary problems (Mr. Finger perhaps), I'm willing to recompile, test and report back to help squash these bugs.
>
> I'm attaching a debug=5 log to show my tests.
>
> As of now, this chipset/card is usable (signal reading is annoying 
> btw)

Thanks for the report. That patch will be pushed upstream in a few minutes and will be backported to stable withing a few weeks. I am happy to hear that the patch works with your Alfa AWUS036NHR. In my limited testing with that device, it failed to connect. I feared that the RTL8192RU differed from the RTL8192CU.

I know that improper operation of the LEDs and screwy signal levels are annoying, and I will get to them when I can. At the moment, however, I have more important problems to solve.

Thanks,

Larry