Release 3.2 kernel issue tracking thread


I also noticed a bug: i run a rom, i let the screen goes off with the powersaving, then when i bring it on, the graphic are jerky. The sound stays ok.
I too observed and reported these 10-30 seconds of limited system responsiveness. But I do not consider this as a bug, as almost all electronic systems I know of are not immediately ready after wake-up. (For illustration: Just consider yourself at 6:00 AM, it takes a few seconds at least, for others even an hour to get fully operational)
 
Last edited by a moderator:
Not the same as what you reported, it's permanent here and seems to be Picodrive-related, PCSX at least is not affected. Reloading the rom fixes the jerk.
 
I confirm that the composite TV-out displays the colors with the 3.2.14 kernel.
 
Neither of my USB hubs work under 3.2, seem to be fine on the older kernel?


(null)
 
i managed to crash the 3.2 kernel to a unbootable state. while listening to music using audacious, and tab browsing with midori4.2, pandora froze up, and after about 5min, i decided to hard reboot it with pandora button+power switch, and was kindly greeted with a "Failed to load kernel, you may need to reflash the firmware." upon boot. if selecting default boot in menu, it gives me a simple text of "Pandora #" kernel2.6 still boots fine and i see no other problems
 
I've broken the kernel twice now, both from having to hard reset my Pandora when a PND crashed!


(null)
 
Would it be possible to run a simple startup script that replaces the loaded kernel with a backup copy for when this happens? Could cause too many nand writes...


How about loading it to a small ram disk like android? That might complicate with suspend...


I've had to reflash twice because of either corrupted unstartable os or random freezing


Sent from my Nexus S 4G using Tapatalk 2
 
Last edited by a moderator:
system freezes when attaching main power i should add seems to happen if i leave system alone for a bit does not happen straightaway

I noticed the same problem i think it is triggered about half an hour of idle(letting it open with not doing nothing). Tested also with wi-fi enabled and it is the same, You can let any application opened and if you have somewhere to go for a half an hour(10 min more or less) it freezes.


I turned off LCD blinking and tested again : not freezing anymore.I'll keep running some different tests to point this thing down :)
 
Last edited by a moderator:
system freezes when attaching main power i should add seems to happen if i leave system alone for a bit does not happen straightaway
I noticed the same problem i think it is triggered about half an hour of idle(letting it open with not doing nothing). Tested also with wi-fi enabled and it is the same, You can let any application opened and if you have somewhere to go for a half an hour(10 min more or less) it freezes. I turned off LCD blinking and tested again : not freezing anymore.I'll keep running some different tests to point this thing down :)

Same issue as me... I thought that was the case with turning off LCD.. but I couldn`t remember. glad that it`s not just my unit ;)
 
I've been trying to isolate the issues with wifi dropouts previously reported here and here. This is what I've found.


The problem presents as the loss of throughput on what starts as a working wifi connection. Network applications like VNC and Synergy stutter then stop transferring data altogether. The timeframe varies connectivity can last anywhere from a few seconds, to a couple of minutes, before it cuts out when using one of these applications. I have not been able to use either application without such a dropout for more than a couple of minutes under kernel 3.2


Turning off Power management makes no difference.


The issue does occur with security set to WPA2 AES


The issue does not seem to occur with security turned off (though I was reluctant to test it this way for long)


The issue does not occur under the old kernel.


Issue seems to be triggered by heavy traffic within a local wifi network, especially when there is high traffic between the pandora and another local machine.


External traffic doesn't tend to trigger the problem as often.


I have been using ping tests to monitor the status of the connection, I ping my netbook from the pandora in a terminal window and keep an eye on it while testing. What I tend to see there is that I stop receiving ping responses, and then eventually start getting back "Destination Host Unreachable"


The connection always comes good if I trigger a reconnect in Network Manager.


Tonight I setup wireshark on the netbook and observed what happened when the problem kicked in.


It started with a few requests in either direction to retransmit packets. Once the application data ceased, the pings continued. The netbook was receiving every single ping and responding, even though no responses were showing up on the pandora itself. Eventually the pandora seemed to notice it wasn't seeing any incoming traffic and started sending ARP packets requesting the address of the netbook. (the netbook kept responding, to no effect) This was the point where the Destination Host Unreachable errors started.


Basically, it seems to me that the pandora simply stops processing incoming packets, but can still send outgoing packets without a problem.


- Neelix
 
Last edited by a moderator:
Anything on dmesg when this happens?


Could you do



Code:
sudo mkdir /debug

sudo mount -t defbugfs none /debug

and then



Code:
sudo grep . /debug/wl1251/*

sudo grep -v '0$' /debug/wl1251/fw-statistics/*

before and after the problem?


Maybe do it several times when it hits the problem to see if any counters freeze.
 
Last edited by a moderator:
I reproduced this issue several times taking readings before and after each occurrence.


The issue did not result in any dmesg messages, so the only dmesg output occurred when I reconnected.


It was only when I was about to post this Pastebin with my results that I realised I had misunderstood your request, so I did it again and took several readings after the issue started which can be found here. (though I didn't notice any of the counters freezing)


- Neelix
 
Ok it's much over a year since I last looked at that driver, and I can vaguely recall similar issue, i.e. it stopping to receive data, I think it's chip's firmware bug, perhaps a race condition of some sort, when it hits it's unable to decrypt any incoming packets (outgoing ones are still ok).


I've done some hacky code to try to deal with this, but I was not sure it was any good and not ported over to 3.2. Try this driver (heeds SZ beta5a):


http://notaz.gp2x.de/tmp/wl1251_test1.zip



Code:
sudo rmmod wl1251_sdio wl1251

sudo insmod path/to/extracted/wl1251.ko

sudo insmod path/to/extracted/wl1251_sdio.ko
 
My initial testing is quite promising.


The wifi on my primary unit (inigo) inexplicably seems to have become stable on its own, which would normally reduce frustration... except when trying to reproduce the issue to begin with.


So to test this I flashed my second unit (vizzini) to SZb5a. After confirming that the reflashed unit did in fact suffer from the issue I loaded the new driver as described above. The wifi on vizzini now seems stable even under long VNC or Synergy sessions which would have crashed it previously. At this point I recommend seeing if these drivers fix the problem for anyone else.


- Neelix


EDIT: Well so much for fixing itself... while I was playing with VNC on vizzini, the wifi on inigo died. I may have to install the new drivers on inigo after all. :)
 
Last edited by a moderator:
At this point I'd have say that these wifi drivers have definitely fixed the wifi dropouts issue for me. I haven't experienced a single wifi dropout while using them. That includes long sessions with synergy or VNC.


- Neelix
 
Back
Top