ANNOUNCE: Diablo-Turbo first beta available - Page 20 - maemo.org

jmart | # 191 | 2011-03-09, 01:00 | Report

Originally Posted by maacruz View Post

Check that the /etc/rc2.d/S??ramzswap link is there.

The /etc/rc2.d/S31ramzswap file is there.

skatebiker | # 192 | 2011-03-09, 07:52 | Report

Originally Posted by maacruz View Post

Yes

No, this patch doesn't solve issues with faulty hardware or old batteries, I never promised such a thing.

Faulty hardware ?
Old batteries ?

The battery is renewed 6 months ago and how do you think I have 'faulty hardware' ?
It is the same unmodded N810 as I bought it in Sept 2008.

And I do not assume that you promised a proper working it is just a question.

maacruz | # 193 | 2011-03-09, 08:17 | Report

Originally Posted by skatebiker View Post

Faulty hardware ?
Old batteries ?

The two main causes for spontaneus reboots

Originally Posted by

The battery is renewed 6 months ago and how do you think I have 'faulty hardware' ?
It is the same unmodded N810 as I bought it in Sept 2008.

I think it is a hardware fault, either the battery or the device, because you refer spontaneous reboots, independently of the software.
Now, about your new battery.... it is a real Nokia one or another brand, Mugen in particular?

lma | # 194 | 2011-03-09, 10:35 | Report

Originally Posted by maacruz View Post

I do not do perl, too arcane for me to learn, but I'll happyly use your code snippet.
While you're at it, why don't you add the equivalent of "if filesize(Z)<$l: exit(1)" before truncate?

Just for the hell of it, I had a go at doing it in a shell script, using only what's provided by busybox and without "cheating" by using other interpreters like awk, sed etc. It turned out to be surprisingly tricky to get the 32bit int in decimal format without having something like od(1), but here goes anyway:

Code:

#!/bin/sh

dd if=zImage-backup of=l skip=44 count=4 bs=1
b=$(echo $(echo -n '0000' | cmp -l l | cut -c3-5))
read a b c d <<EOF
$b
EOF
rm l

eval $(printf "l=\$(( (((0$d \* 256 + 0$c) \* 256 + 0$b) \* 256 + 0$a) ))")

s=$(echo $(ls -l zImage-backup) | cut -f5 -d" ")
if [ "$s" -gt "$l" ] ; then
        dd if=zImage-backup bs=$l count=1 of=zImage-backup.new && \
        mv -f zImage-backup.new zImage-backup
fi

Improvements welcome ;-)

skatebiker | # 195 | 2011-03-09, 11:10 | Report

Originally Posted by maacruz View Post

The two main causes for spontaneus reboots

I think it is a hardware fault, either the battery or the device, because you refer spontaneous reboots, independently of the software.
Now, about your new battery.... it is a real Nokia one or another brand, Mugen in particular?

The battery is an original Nokia one ordered via eBay.
The only possible hardware fauly might be the MMC1 card which is a 3rd party Micro SD card in a Mini adapter. Can that be qa problem ?
I did not alter anything in the hardware other than replacing the battery and entering an external SD into it. And I never dropped it or spilled water onto it.

In earlier posts I read that the spontaneous reboots were because of a firmware bug which triggers the watchdog inadvertently. And with the better handling of events / watchdog in your improved kernel I hoped this problem would be solved. Now this is not the case then it can be a hardware problem.

Last edited by skatebiker; 2011-03-09 at 11:14.

maacruz | # 196 | 2011-03-09, 12:13 | Report

Originally Posted by skatebiker View Post

The battery is an original Nokia one ordered via eBay.

Ebay, uhmmm, I wouldn't be very sure it is really Nokia. Well, may be the problem is the same as with Mugen ones, loose contacts. Put a piece of thick paper or cardboard at the battery's bottom, so it is hold tight against the device's battery contacts.

Originally Posted by

The only possible hardware fauly might be the MMC1 card which is a 3rd party Micro SD card in a Mini adapter. Can that be qa problem ?

It could be possible, but I think it's unlikely. I'd bet on the battery.

Originally Posted by

In earlier posts I read that the spontaneous reboots were because of a firmware bug which triggers the watchdog inadvertently. And with the better handling of events / watchdog in your improved kernel I hoped this problem would be solved. Now this is not the case then it can be a hardware problem.

Not the watchdog, it is bme what triggers the reboot, and usually happens with old/weak batteries. The watchdog only triggers if the device hangs or becomes so busy that the watchdog isn't refreshed in 20 seconds.
My kernel makes unlikely the later by distributing the cpu more evenly.

maacruz | # 197 | 2011-03-09, 12:20 | Report

Originally Posted by lma View Post

Just for the hell of it, I had a go at doing it in a shell script, using only what's provided by busybox and without "cheating" by using other interpreters like awk, sed etc. It turned out to be surprisingly tricky to get the 32bit int in decimal format without having something like od(1), but here goes anyway:

Code:

#!/bin/sh

dd if=zImage-backup of=l skip=44 count=4 bs=1
b=$(echo $(echo -n '0000' | cmp -l l | cut -c3-5))
read a b c d <<EOF
$b
EOF
rm l

eval $(printf "l=\$(( (((0$d \* 256 + 0$c) \* 256 + 0$b) \* 256 + 0$a) ))")

s=$(echo $(ls -l zImage-backup) | cut -f5 -d" ")
if [ "$s" -gt "$l" ] ; then
        dd if=zImage-backup bs=$l count=1 of=zImage-backup.new && \
        mv -f zImage-backup.new zImage-backup
fi

Improvements welcome ;-)

LOL

Looks like it was inspired by the robodevil, straight from the robohell

I had my own way with hellish shell scripting when I implemented amarok's 1.4 replaygain plugin in bash. Never again

jmart | # 198 | 2011-03-09, 15:07 | Report

Any suggestions on why my ramzswap doesn't seem to be working properly would be appreciated. I think it's an important part of DT and I would like to get it working, but it's a little beyond my scope.

/etc/init.d/ramzswap restart in xterm as root worked, when I typed free I then saw my swap memory. However, after restarting my N810 I am back to 0 swap memory -- until I restart ramzswap in xterm again

Thanks!

maacruz | # 199 | 2011-03-09, 15:31 | Report

Originally Posted by jmart View Post

Any suggestions on why my ramzswap doesn't seem to be working properly would be appreciated. I think it's an important part of DT and I would like to get it working, but it's a little beyond my scope.

/etc/init.d/ramzswap restart in xterm as root worked, when I typed free I then saw my swap memory. However, after restarting my N810 I am back to 0 swap memory -- until I restart ramzswap in xterm again

Thanks!

Please, check if after a fresh restart the /dev/ramzswap0 device is there.

auouymous | # 200 | 2011-03-09, 15:34 | Report

@ skatebiker
Try passing --enable-rd-mode --set-rd-flags=no-lifeguard-reset -R to the flasher and see if your device stops rebooting. This won't reflash, it just turns on R&D mode and disables the the watchdog reset.