Release SuperZaxxon - Beta 4 released


So it's not related to ZRAM, and I still don't know *exactly* how to reproduce it, but it doesn't take much actually. Since once it happened every time by just doing wifi on,off,suspend,wake-up,wifi on and nothing else after a fresh reboot, and it vanished after removing the SD card in slot1 (16GB class 6), I just installed firefox on my 32GB class 10, and after the second time of doing the Firefox etc. sequence it hapened again. So it's not related to the SD-card either, and it may take more than one try to reproduce it that way.
Ok finally managed to reproduce it by repeatedly loading new tabs in firefox, toggling wifi and suspending/resuming. Takes 3-10 minutes of this "work" to happen. I have a preliminary fix, can you test it? If so, get http://notaz.gp2x.de/tmp/oe_rootfs_extra.tar.bz2 , then in terminal



Code:
cd /

sudo tar xvf /media/yoursdcard/yourpath/oe_rootfs_extra.tar.bz2

sync

and then reboot.

any time another device on the network is transferring a fair amount of data via wifi, the wifi on the pandora stops responding, and I have to disconnect and reconnect to the network to get it going again.
Unfortunately I can't really do much for this one. Crashes like in F_Slim's case are relatively easy when backtrace is available, but here it would need deep knowledge of this undocumented chip and it's firmware to figure out what's going on and how to fix it. It's a shame this driver has been abandoned in this rather poor state and is no longer developed.
 
I had the same thing with Wifi. see my posts on previous pages. It was the router encryption (Needed to be set to AES) And interference on the router channel. Once I found a clear channel 3rd choice) the issue was resolved. It may not be your problem but it _could_ be!


I have been fine ever since.
 
echo 9 | bc -l


gives me a segfault. What else should i use to do some bash computation ?
 
echo 9 | bc -l


gives me a segfault. What else should i use to do some bash computation ?

That commands work fine in my desktop linux, but seeing the same segment fault on the Pandora.. however


echo 9 | bc however seems to work.. seems to be a problem with the -l flag


I echo'd simple math 9+9, 2-3 and it seems to work fine..


however


echo 15/2 | bc


will return 7.. it seems not to handle decimals or rounds poorly.


Well did a quick google search.. seems it was a known issue.. maybe fixed in a newer update of bc
 
Last edited by a moderator:
Hi,


I'm still having the same issue I had with B3 for wifi. My home box has a WPA2 access point, and I can't connect. I was able to connect only once after flashing the firmware, and never after rebooting. Everytime I try to connect I go into a password dialog loop. This happens ONLY with the 3.2 kernel, I have no trouble at all if I boot the legacy kernel.


I suspect that something in the driver is wrong for WPA, and it reminds me of a similar issue (same symptoms) I ran into on my eeepc with Ubuntu 10.04 a while ago. Here's the post to the fix that worked: http://ubuntuforums.org/showthread.php?t=1476007, in case it could help the devs.


Which chipset/module is used on B4 for the internal wifi?


Apart from this, B4 seems quite stable, much better than B3 IMO ;-)


Only app I can't run properly so far is PicoDrive, tearing happens when scrolling vertically. Will notaz fix this?


Cheers,
 
Notaz,


Are we still using compat-wireless with this kernel, or is it just using the mainline wireless drivers?


Thanks,


gunrock
 
Last edited by a moderator:
The wl1251 driver source code is not available on TI's website :(


I tought I could compare with the Ubuntu issue I had.
 
Ok finally managed to reproduce it by repeatedly loading new tabs in firefox, toggling wifi and suspending/resuming. Takes 3-10 minutes of this "work" to happen. I have a preliminary fix, can you test it? If so, get http://notaz.gp2x.de...s_extra.tar.bz2 , then in terminal

Code:
 cd /

sudo tar xvf /media/yoursdcard/yourpath/oe_rootfs_extra.tar.bz2

sync
and then reboot.
Hey, finally found some time and tested it. So far I have not been able to reproduce the freeze, so it looks like the fix works great. Thanks! :)


Just to be sure what I actually did up there: I guess the archive is a patch for the kernel, and the "sync" command makes sure files currently in use are really overwritten?


And out of interest: what was the cause and the fix (in case it's possible to describe in a few simple words ^_^ )?
 
Hmm it didn`t work for me... just left my Pandora alone for a bit... come back and it`s frozen. reboot :(


THis can happen two or three times a day... I`m not switching wifi on when it happens either!


i`ve not reflashed, I`m using the beta 4 that it was installed with...


Has there been issues with sd cards causing lock ups? because, unless it was coincedence,, it locked up as I was putting a card in the first slot.


Every other time its been sitting there and froze on me..


Darren
 
Is really beta4 installed on new shipped units? Ithought that was beta2 or 3.
 
And out of interest: what was the cause and the fix (in case it's possible to describe in a few simple words ^_^ )?
Getting an interrupt before driver initialization completes and tripping on NULL pointer. Fix was to keep interrupt masked until init is done.

Hmm it didn`t work for me... just left my Pandora alone for a bit... come back and it`s frozen. reboot :(
Does it happen when the screen blanks?


Does it do it if you close and open the lid several times?


Running any PNDs, perhaps 3d/GL related?


So far there have been reported freezes on time change and screen blank, never seen those here. Need a reliable way to reproduce these, otherwise I have no clue what to look for here.
 
Last edited by a moderator:
hmm let me double check on that then!


Release Super Zaxxon booting kernel 3.2 and there`s a beta4 in the top right corner!


..


I might reflash..
 
Last edited by a moderator:
Just had the desktop crash again. I was trying to load google docs in Firefox and suddenly I'm met with a screen full of text. Logging in to terminal 1, dmesg outputs

*snipping a bunch of irrelevant stuff from over an hour ago


[12480.255249] wlan0: deauthenticating from 00:24:b2:59:fd:8f by local choice (reason=3)


[12480.333465] cfg80211: Calling CRDA to update world regulatory domain


[12484.736175] quiet_error: 26 callbacks suppressed


[12484.736175] Buffer I/O error on device loop1, logical block 0


[12484.737487] Buffer I/O error on device loop1, logical block 1


[12484.738616] Buffer I/O error on device loop1, logical block 2


[12484.739746] Buffer I/O error on device loop1, logical block 3


[12484.740844] Buffer I/O error on device loop1, logical block 4


[12484.741851] Buffer I/O error on device loop1, logical block 5


[12484.742828] Buffer I/O error on device loop1, logical block 6


[12484.743713] Buffer I/O error on device loop1, logical block 7


[12484.744567] Buffer I/O error on device loop1, logical block 8


[12484.745422] Buffer I/O error on device loop1, logical block 9

/dev/loop1 is mounted pndmanager. I closed pndmanager a while ago, and now that I think about it I may have ejected the SD card that it exists on thinking it had successfully closed, but I could be mistaken. gparted is loop2 which I also closed but still seems to be mounted.


Firefox isn't mounted, suggesting it successfully crashed and was unmounted. From the pndrun_hdonk_firefox_001.out

** (firefox:2541): WARNING **: Failed to send buffer


The application 'firefox' lost its connection to the display :0.0;


most likely the X server was shut down or you killed/destroyed


the application.


And for completeness, here's the pndrun_pndmanager.out, though I'm pretty sure this was my fault

=======================================================================================


PND : /media/TTT/pandora/menu/pndmanager.pnd


PND_FSTYPE : Squashfs


APPDATADIR : /media/TTT/pandora/appdata/pndmanager


APPDD_FSTYPE : vfat


PND_CPUSPEED : <unset>


EXENAME : scripts/panorama.sh


ARGUMENTS : <unset>


=======================================================================================


[ START ]--- Mount the PND ----------


Mounting : mount -t squashfs -o ro "/dev/loop1" "/mnt/utmp/pndmanager"


Mounting the Union FS : mount -t aufs -o exec,noplink,dirs="/media/TTT/pandora/appdata/pndmanager=rw+nolwh":"/mnt/pnd/pndmanager=rr" none "/mnt/utmp/pndmanager"


[sUCCESS]--- Mount the PND ----------


[ START ]--- Starting the application (scripts/panorama.sh ) ----------


mkdir: cannot create directory `interfaces/CustomPNDManager': Operation not permitted


cp: target `interfaces/CustomPNDManager' is not a directory


QGtkStyle was unable to detect the current GTK+ theme.


Settings are saved in "/mnt/utmp/pndmanager/settings.cfg"


Loaded UI PNDManager created by B-ZaR.


Description:


PND Manager application


PNDManager::crawl


Crawling "/media/TTT"


Crawling "/media/KINGSTON"


Updating remote repository


Updating local repository


Updating package list


PNDManager::updatePackages


Found 594 packages


file:///mnt/utmp/pndmanager/interfaces/PNDManager/SectionItem.qml:17:3: QML Image: Error downloading http://repo.openpandora.org/files/pnd/dinodefense-dinodefense-15443/icon.png - server replied: Not Found


DownloadWorkerSingleton::DownloadWorkerSingleton


DownloadWorker::start


in function :: _pxml_pnd_parse


Invalid XML!


-----


not well-formed (invalid token)


in function :: _pxml_pnd_parse


Invalid XML!


-----


not well-formed (invalid token)


in function :: _pxml_pnd_parse


Invalid XML!


-----


not well-formed (invalid token)


[ FAILED]--- Starting the application (scripts/panorama.sh ) ----------


[ START ]--- Restoring the frame buffer status ----------


[sUCCESS]--- Restoring the frame buffer status ----------


[ START ]--- uMount the PND ----------


ERROR: selected PND() file does not exist!


Usage:


pnd_run.sh -p file.pnd -e cmd [-a args] [-b pndid] [-s path] [-c speed] [-d [path]] [-x] [-m] [-u]


-p file.pnd : Specify the pnd file to execute


-e cmd : Command to run


-a args : Arguments to the command


-b pndid : name of the directory mount-point (/mnt/utmp/pndid) (Default: name of the pnd file)


-s path : Directory in the union to start the command from


-o speed : Set the CPU speed


-d [path] : Use path as source of the overlay. (Default: pandora/appdata/pndid)


-x : Stop X before starting the apps


-m : Only mount the pnd, dont run it (-e become optional)


-u : Only umount the pnd, dont run it (-e become optional)


[ FAILED]--- uMount the PND ----------


=======================================================================================


Return code is : 5


I will try to recreate the error. Once I get here, the only thing I can do is reboot it seems.
 
And out of interest: what was the cause and the fix (in case it's possible to describe in a few simple words ^_^ )?
Getting an interrupt before driver initialization completes and tripping on NULL pointer. Fix was to keep interrupt masked until init is done.
Hmm it didn`t work for me... just left my Pandora alone for a bit... come back and it`s frozen. reboot :(
Does it happen when the screen blanks? Does it do it if you close and open the lid several times? Running any PNDs, perhaps 3d/GL related? So far there have been reported freezes on time change and screen blank, never seen those here. Need a reliable way to reproduce these, otherwise I have no clue what to look for here.

Well it would happen when it was sitting there idle. I would go back to it and it would be froze! but obviously.. it would be in sleep mode.. and when I notice it froze up the screen is on and locked up good!


BUT: I reflashed.. the codec pack auto installed as well :) and I haven`t had a problem all yesterday afternoon and today so far :)
 
Probably one of those weird X crashes, hopefully will be fixed when Xorg is updated..
I just thought it was weird, getting it twice in one day while doing extremely memory heavy things in Firefox. Ejecting a mounted PND (which I thought had been closed) probably didn't help any.


I haven't since been able to get it again. Have caused Firefox to crash numerous times, but not X.
 
Just noticed my wifi doesn`t disable when I close the lid anymore..


since I reflashed ( and fixed my niggles ) whenever I close the lid the wifi light stays on.. and doesn`t need to reconnect after it awaken.


I can`t find the setting for the life of me :(


Darren
 
Back
Top