maemo.org - Talk

maemo.org - Talk (https://talk.maemo.org/index.php)
-   Maemo 5 / Fremantle (https://talk.maemo.org/forumdisplay.php?f=40)
-   -   [Announce] kernel-power stable v50 in Extras (https://talk.maemo.org/showthread.php?t=83118)

Estel 2012-07-11 17:16

Re: [Announce] kernel-power stable v50 in Extras
 
QT MobileHotspot works flawlessly with kernel power, no matter of mode selected (WiFi, USB...).

/Estel

ersanpermana 2012-07-11 23:37

Re: [Announce] kernel-power stable v50 in Extras
 
Quote:

Originally Posted by Estel (Post 1236175)
QT MobileHotspot works flawlessly with kernel power, no matter of mode selected (WiFi, USB...).

/Estel

@Estel Ok maybe I can try again. thanks :)

ersanpermana 2012-07-12 01:53

Re: [Announce] kernel-power stable v50 in Extras
 
Yes, it worked for Qt Mobile hotspot, many thanks for Pali and Estel :)

reinob 2012-07-12 07:29

Re: [Announce] kernel-power stable v50 in Extras
 
Quote:

Originally Posted by Estel (Post 1235716)
AFAIK, ShadowJK mentioned once a possibility of overheating chip, when charging with more than 950 mA (he tried faster charging, using charge.sh and temperature change was very troubling - I asked about it, when I considered faster charging due to 2-cell battery and 2A power source).

Soo.. little story here :)

First of all: thanks Estel, and thanks Pali.

Now. I bought an N900 from Estel, with the reinforced USB port. I flashed it (so USB was working OK) and then started removing lots of things from Maemo. At some point something went wrong and the N900 wouldn't charge, either from USB or from dedicated charger.

Plugging the charger usually resulted in receiving a "not sufficient power" (or whatever the exact message was), actually charging for a minute or so, and then just stopping. Other times plugging the charger would actually result in the N900 asking me if I wanted to connect in mass storage mode (!)

With the USB cable it was more or less the same deal. Windows (work) would tell me "device has malfunctioned", and it wouldn't charge.

I thought that maybe the USB port was not properly "reinforced", or that maybe Estel fixed it after it was already broken. But there was still a slight possibility that it was a software issue (I don't keep a log of what I removed, but it's easy enough to diff dpkg list -- the patched stuff is harder to identify, but I'll do it someday).

So, yesterday I installed the KP51 bootimg and modules for use with U-boot (btw. bootmenu entry came with the new OMAPTAG whereas my U-boot still uses REUSEATAGS -- had to patch that, but will update U-boot anyway).

Then I stopped bme, loaded the bq2415x_charger module and plugged the dedicated charger.

The LED went on (solid yellow) and I left it for about 90min. At some point (after ~15min or charging) it felt a little warm but after that nothing I could notice.

When I unplugged it I loaded BME again. The battery had charged from about 300mAh to 600mAh, in about 90min. Meaning ~200mA charging current (didn't check if the mode was "dedicated" or "auto" or whatever -- will do next time).

In short: (1) in-kernel charging definitely works and (2) it may be that my USB-issues were due to software and not to hardware [ but need to investigate further ].

Thanks again!

pali 2012-07-12 08:58

Re: [Announce] kernel-power stable v50 in Extras
 
@reinob: When you load bq2415x_charger module, default mode is auto (it will switch between none, host and dedicated automatically). So it you did not changed mode, it was auto.

reinob 2012-07-12 14:16

Re: [Announce] kernel-power stable v50 in Extras
 
Quote:

Originally Posted by pali (Post 1236465)
@reinob: When you load bq2415x_charger module, default mode is auto (it will switch between none, host and dedicated automatically). So it you did not changed mode, it was auto.

I guessed so, but I didn't check :)

Anyway. I'm going a bit wild now and have purged the following packages:
hald-addon-bme, osso-systemui-actingdead *and* bme-RX-51 (!).

hald-addon-bme was because during charging it was all the time trying to contact BME, which was obviously stopped. So I removed it.

osso-systemui-actingdead requires hald-addon-bme, so it went too.

And then I thought: if the kernel can charge the N900, why do I need BME at all? so I purged it as well.

I had to carefully check /etc/event.d/bme to make sure nothing was fvcked up. I cannot remove it (for now), because other events depend on bme being started (so I fake that it starts).

I then rebooted the N900 (not without fear :) and it rebooted just fine.

Obviously I don't think I can charge the N900 anymore if it's not fully turned on. I will whitelist the bq2415x_charger module so that it can charge while ACTING_DEAD (but I need to check the whole ACTING_DEAD mess). Maybe I'll get rid of that mode and let the N900 boot normally when plugged while being off.

I know it's a bit off-topic. Maybe I'll make a new thread when I have some news.

vi_ 2012-07-12 14:36

Re: [Announce] kernel-power stable v50 in Extras
 
AFAIK qwerty12 wrote a patch that would eliminate the actdead mode.

pali 2012-07-12 14:39

Re: [Announce] kernel-power stable v50 in Extras
 
Quote:

Originally Posted by reinob (Post 1236573)
I guessed so, but I didn't check :)

Anyway. I'm going a bit wild now and have purged the following packages:
hald-addon-bme, osso-systemui-actingdead *and* bme-RX-51 (!).

hald-addon-bme was because during charging it was all the time trying to contact BME, which was obviously stopped. So I removed it.

osso-systemui-actingdead requires hald-addon-bme, so it went too.

And then I thought: if the kernel can charge the N900, why do I need BME at all? so I purged it as well.

I had to carefully check /etc/event.d/bme to make sure nothing was fvcked up. I cannot remove it (for now), because other events depend on bme being started (so I fake that it starts).

I then rebooted the N900 (not without fear :) and it rebooted just fine.

Obviously I don't think I can charge the N900 anymore if it's not fully turned on. I will whitelist the bq2415x_charger module so that it can charge while ACTING_DEAD (but I need to check the whole ACTING_DEAD mess). Maybe I'll get rid of that mode and let the N900 boot normally when plugged while being off.

I know it's a bit off-topic. Maybe I'll make a new thread when I have some news.

freemangordon, started writting hald-addon-bme replacement which will use bq27x00_battery & bq2415x_charger modules.

When all userspace bme apps will be rewritten/replaced I will create one deb package which diable BME (with dpkg-divert) and load bq kernel modules at startup...

reinob 2012-07-13 09:19

Re: [Announce] kernel-power stable v50 in Extras
 
Quote:

Originally Posted by vi_ (Post 1236580)
AFAIK qwerty12 wrote a patch that would eliminate the actdead mode.

Can't seem to find it. But better that way, so I get to play with it a bit (I think I can just get rid of getbootstate and make a fake one).

I'll also consider completely wiping dsme out of the N900 (or making a fake one if dependencies, debianish or real, prevent me from doing that).

joerg_rw 2012-07-13 09:39

Re: [Announce] kernel-power stable v50 in Extras
 
Quote:

Originally Posted by Estel (Post 1235716)
AFAIK, ShadowJK mentioned once a possibility of overheating chip, when charging with more than 950 mA (he tried faster charging, using charge.sh and temperature change was very troubling - I asked about it, when I considered faster charging due to 2-cell battery and 2A power source).

It's possible that I've misunderstood something - so don't get aggressive, like in last answer to my post (despite my tremendous amount of marking, that it's personal view on things, personal opinion, may be not accurate, etc) - but are you sure about that one, i.e complete bulletproofness of charging route design?

/Estel

Charger chip has overtemp protection and will signal that condition as a dedicated error even.
So yes, I'm rather sure there's definitely no way you fry your device by whatever you do to charging hardware via software.

/j


All times are GMT. The time now is 15:04.

vBulletin® Version 3.8.8