Notices


Reply
Thread Tools
Posts: 2,802 | Thanked: 4,489 times | Joined on Nov 2007
#511
This isn't DT's fault, but after installing on my new(ish) RX48 I see this weirdness:

Code:
Nokia-N810-WiMAX-43-7:/usr/share/Diablo-Turbo/backup# df .
Filesystem           1k-blocks      Used Available Use% Mounted on
rootfs                  255488    249480      6008  98% /
Nokia-N810-WiMAX-43-7:/usr/share/Diablo-Turbo/backup# ls -l zImage-backup 
-rw-r--r--    1 root     root    926102833 Sep 22 08:52 zImage-backup
Nokia-N810-WiMAX-43-7:/usr/share/Diablo-Turbo/backup# du zImage-backup 
904398	zImage-backup
any ideas that can explain it that do not involve JFFS2 corruption?

Edit: Ah, found it, bug in the perl snippet in the post-install script - the truncate call unpacks the already-unpacked value and gets a bogus result. The resulting file is sparse, and du reports its size correctly but only until the next reboot.

Last edited by lma; 2011-09-22 at 11:22.
 

The Following User Says Thank You to lma For This Useful Post:
Posts: 1,101 | Thanked: 1,183 times | Joined on Aug 2008 @ Spain
#512
Originally Posted by lma View Post
This isn't DT's fault, but after installing on my new(ish) RX48 I see this weirdness:

Code:
Nokia-N810-WiMAX-43-7:/usr/share/Diablo-Turbo/backup# df .
Filesystem           1k-blocks      Used Available Use% Mounted on
rootfs                  255488    249480      6008  98% /
Nokia-N810-WiMAX-43-7:/usr/share/Diablo-Turbo/backup# ls -l zImage-backup 
-rw-r--r--    1 root     root    926102833 Sep 22 08:52 zImage-backup
Nokia-N810-WiMAX-43-7:/usr/share/Diablo-Turbo/backup# du zImage-backup 
904398	zImage-backup
any ideas that can explain it that do not involve JFFS2 corruption?

Edit: Ah, found it, bug in the perl snippet in the post-install script - the truncate call unpacks the already-unpacked value and gets a bogus result. The resulting file is sparse, and du reports its size correctly but only until the next reboot.
Aw! Sh*t!
Now everybody whose first DT kernel is kernel-dt-sd_2.6.21-3 has a corrupt zImage backup.
Well, they don't have any reason to uninstall the DT kernel, do they?
And if they do, just reflashing a good zImage after uninstalling will fix the problem.

Last edited by maacruz; 2011-09-22 at 17:07.
 

The Following User Says Thank You to maacruz For This Useful Post:
Posts: 2,802 | Thanked: 4,489 times | Joined on Nov 2007
#513
Originally Posted by maacruz View Post
Aw! Sh*t!
Now everybody whose first DT kernel is kernel-dt-sd_2.6.21-3 has a corrupt zImage backup.
Well, it's still a beta right? ;-)

At least the ones that are larger than 2MiB should be safe to truncate and write back to /dev/mtd2, and most truncated ones should still contain the version string so the right version could be fetched from somewhere else.
 

The Following User Says Thank You to lma For This Useful Post:
Posts: 1,101 | Thanked: 1,183 times | Joined on Aug 2008 @ Spain
#514
Announce: maemo-pan update (extras-devel)

I have found that maemo-pan stopped working with DT due to a too simple string processing of dbus-send output in the /usr/lib/maemo-pan/pan-control script.

I have uploaded a new release (1.0.2) of maemo-pan to extras-devel to fix this issue. This new release works with both old Diablo and DT.
 

The Following 2 Users Say Thank You to maacruz For This Useful Post:
tso's Avatar
Posts: 4,783 | Thanked: 1,252 times | Joined on Aug 2007 @ norway
#515
either asui or maemo-pan got a bug. When ASUI attempts to disconnect a PAN connection it appears to do so but the connection stays active.
__________________
Be warned, posts are often line of thoughts at highway speeds...
 
Posts: 858 | Thanked: 891 times | Joined on Sep 2010
#516
Originally Posted by tso View Post
either asui or maemo-pan got a bug. When ASUI attempts to disconnect a PAN connection it appears to do so but the connection stays active.
I don't have a cell phone so someone else will need to figure this out. ASUI only listens for a disconnect signal and if PAN isn't sending it...
 
tso's Avatar
Posts: 4,783 | Thanked: 1,252 times | Joined on Aug 2007 @ norway
#517
from what i can tell, maemo-pan listen for a connect and disconnect on a dummy connection and then do the actual connection management behind the scene.

I guess ASUI fails to send a expected signal, and so the actual disconnect is not triggered.

this because the maemo connection manager triggers the correct behavior.
__________________
Be warned, posts are often line of thoughts at highway speeds...
 
Posts: 1,101 | Thanked: 1,183 times | Joined on Aug 2008 @ Spain
#518
Originally Posted by auouymous View Post
I don't have a cell phone so someone else will need to figure this out. ASUI only listens for a disconnect signal and if PAN isn't sending it...
maemo-pan doesn't send any signal when disconnecting.
In fact, pan-daemon listens to system bus for the "dummy" IAP status change signals, and then runs the pan-control script with start/stop as argument. This scripts handles all connecting and disconnecting through dbus method calls to bluez, but the bluetooth connection can also break for being out of range, and in this case, the IAP remains "dummy".
If ASUI needs any signal, we can tweak the script, but I'm not sure how to handle the "out of range" situation.
 
Posts: 858 | Thanked: 891 times | Joined on Sep 2010
#519
Originally Posted by maacruz View Post
If ASUI needs any signal, we can tweak the script, but I'm not sure how to handle the "out of range" situation.
Or tell me what needs to be done in ASUI to support it as is.
 
Posts: 2,802 | Thanked: 4,489 times | Joined on Nov 2007
#520
Originally Posted by maacruz View Post
PLEASE INSTALL THIS KERNEL AND REPORT UPTIME AND CRASHES (IF ANY)
Woke up this morning to find the device had rebooted during the night. Uptime was about 4 days, /proc/bootreason says sw_rst, but syslogd had stopped logging due to 99% full rootfs :-(
 

The Following User Says Thank You to lma For This Useful Post:
Reply

Tags
chinook, diablo, new life, os2008

Thread Tools

 
Forum Jump


All times are GMT. The time now is 17:24.