Release 3.2 kernel issue tracking thread

[McLovin]

After triggering suspend wifi seems to not come up again. happend several times by now.

Is there no way to tell network manager not to turn on the wifi by default? then it wouldn't matter that the kernel driver was auto-loaded. Or am I missing something?


-Neelix

Should be as easy as right-clicking on the nm-applet and unselecting the "enable wifi"-option


will test that now

 

[notaz]

I did that, both the GUI and /proc/pandora/cpu_opp_max tell me my max OPP is set to 5


Edit: Ok, 880MHz is the highest I can go, as far as I remember that's the highest clock I could go at OPP3. Is it possible my Pandora's ignoring the allowed OPP somehow?

Remove the battery and leave pandora off for 10 minutes then.

 

[McLovin]

confirmed: just right-clicking on nm-applet icon and deselecting "enable wifi" there has dramatical impact. my battery is at ~60% and after deselecting that option and leaving the pandora closed but not suspended for like 10 minutes, expected battery life went from ~60 hours to ~120 hours!!!


@notaz: did you see my problem above? no wifi after awaking from suspend will try again now.

 

[McLovin]

well just "sudo modprobe wl1251_sdio" does the trick and wifi is back. but your "wifi can not be deactivated" problem seems to be no problem at all

 

[peelie]

system freezes when attaching main power i should add seems to happen if i leave system alone for a bit does not happen straightaway

 

[Pickle]

i saw ioquake2 was not working because of controls, it uses the evdev event structure for direct access. Something change with that?

 

[notaz]

system freezes when attaching main power i should add seems to happen if i leave system alone for a bit does not happen straightaway

ugh, all kinds of weird stuff I never get, it charges fine for me from AC..

i saw ioquake2 was not working because of controls, it uses the evdev event structure for direct access. Something change with that?

Should not, unless you hardcoded /dev/input/event* nodes - I wrote in Kernel interface that you shouldn't do that back in 2008. There is also evdev version number change.

 

[fahrstuhl]

The firmware for realtek wifi USB dongles isn't uploaded to the device so my rtl8192cu dongle doesn't work. (Works with 2.6)

This is curious, because my Ralink wireless device seems only to work using the experimental kernel, and doesn't on 2.6.

modinfo rtl8192cu.ko mentions a firmware file rtlwifi/rtl8192cufw.bin which neither I nor the driver are able to locate. (At least it's not in the /lib/firmware/rtlwifi/ directory where I suppose it should be. And find / -name "rtl8192cufw.bin" is unsuccessful.)


It's in the Angstrom repo:


http://www.angstrom-...mware-rtl8192cu


I had to download and install it manually, opkg list doesn't show any linux-firmware packages (and I think no kernel-3.2.1 packages at all?)


Device works more or less, it finds networks now but doesn't connect yet. I'll keep trying.


Works fine now. I'm not sure if that was it but I deactivated wifi channels 12 and 13 on my router.


Also, middleclick clipboard doesn't work for me.

 

[superatarian]

Well my 250 gig USB drive does not work now. The Pandora tells me it cannot recognize the file system, ntfs.


My USB wireless mouse works fine though. Any fix for the USB drive? Worked fine before?

 

[slaeshjag]

did you install the codec pack?

 

[peelie]

system freezes when attaching main power i should add seems to happen if i leave system alone for a bit does not happen straightaway

ugh, all kinds of weird stuff I never get, it charges fine for me from AC..

seems to not freeze now? another thing is that when first connecting power lead i need to connect and reconnect for the red power light to show?


must say even with kinks suspend to ram is great!

 

[NoeleyC]

I've spotted an odd charging issue that seems to be present on both versions of the kernel? Before I plug in a charging cable into my Pandora(I usually use a USB charger with a psp data/charging cable), have never encountered any issues using this method before, but now my Pandora shows the correct %charge and time remaining before I plug it in, then it changes to a charging indication(time remaining is 00:00), but when I plug the cable out again, this doesn't go back to showing me how much time is remaining? I don't recall this happening before?


Update, it does eventually show the correct value, just have to leave it for a minute or two before it shows the right value again?

 

[DynaMight]

I retested all the non-working emulators/games on the stable kernel and you can remove Streets of Rage remake. Once configured correctly it worked fine on the beta kernel. All the others, for me atleast, do not work on the beta kernel but work perfectly fine on the stable kernel.


It does seem slower at loading on the beta kernel? I havent compared time taken to load a certain game/program but the stable kernel does seem much quicker.


Also I'm connected to my wifi, but all webbrowsers fail to resolve any websites. Works fine on the stable kernel.

 

[superatarian]

did you install the codec pack?

ooooo!!!. Did not know I has to re-do that. Thanks. I will give it a try....

 

[thatgui]

Maybe this seems a bit weird, but my Pandora got louder with the new Kernel. I'm one of unfortunate Pandora owners who got a board that chirrs / buzzes when one of the sdcards is accessed (its origin seems to be in the area between the nubs, on the topside of the pcb) .


This was just a small Problem till now, everyday use was possible except when beeing in a quiet environment - like using in Bed was impossible, as it drove me and my wife crazy.


But now the amplitude has risen at least about the factor 3, you can really here it even in public transport.


Also, is there something writing to the sdcards in an regular intervall ? Now every 3-4 minutes a chirr/buzzing occours for about 20seconds, but only when I boot with the experimental kernel.

 

[Neelix]

Does the LED indicator for the SD card come on when it happens?


- Neelix

 

[Wally]

Using the SD Card utility freezes the Pandora when quitting.

 

[thatgui]

Also, is there something writing to the sdcards in an regular intervall ? Now every 3-4 minutes a chirr/buzzing occours for about 20seconds, but only when I boot with the experimental kernel.

Does the LED indicator for the SD card come on when it happens?


- Neelix

As it seems I have to correct myself, the SD Card Lights are both out, when the periodic buzzing occours, so it does not necessarily has something to do with the SD Cards

 

[fahrstuhl]

I did that, both the GUI and /proc/pandora/cpu_opp_max tell me my max OPP is set to 5


Edit: Ok, 880MHz is the highest I can go, as far as I remember that's the highest clock I could go at OPP3. Is it possible my Pandora's ignoring the allowed OPP somehow?

Remove the battery and leave pandora off for 10 minutes then.

It worked but the problem seems to reoccur as soon as I reboot from 3.2 to 2.6. Does anyone else have this problem?


EDIT:

Also, is there something writing to the sdcards in an regular intervall ? Now every 3-4 minutes a chirr/buzzing occours for about 20seconds, but only when I boot with the experimental kernel.

Does the LED indicator for the SD card come on when it happens?


- Neelix

As it seems I have to correct myself, the SD Card Lights are both out, when the periodic buzzing occours, so it does not necessarily has something to do with the SD Cards

I think I have a similar problem from time to time, but it occurs very rarely and I can't even locate its source. I've always suspected something interfering with the speakers but never investigated.


You could try to plug any 3.5mm plug into the headphone jack (like headphones or a line-in cable). This mechanically disconnects the speakers. If the sounds still occur it would rule the speakers out of being the source.

 

[thatgui]

I think I have a similar problem from time to time, but it occurs very rarely and I can't even locate its source. I've always suspected something interfering with the speakers but never investigated.


You could try to plug any 3.5mm plug into the headphone jack (like headphones or a line-in cable). This mechanically disconnects the speakers. If the sounds still occur it would rule the speakers out of being the source.

Did that already, with no effect. The speakers don't generate it, the noise is cleary emitted between the nubs, on the upperside of the pcb - turning the Pandora upside down or closing the lid lowers the amplitude*.


Edit: *not the amplitude of the signal itself, but the amplitude in my perception