| duuude | is it feasible to apply security patches to fremantle kernel? | 10:25 |
|---|---|---|
| sicelo | if they can apply to the kernel sauce ;-) | 10:33 |
| bencoh | you could always try, but fremantle runs a 2.6.28 kernel from 11y ago, so ... | 10:41 |
| bencoh | wait, 12~13yo even | 10:42 |
| KotCzarny | o.O | 10:42 |
| KotCzarny | doesnt fremantle run recent kernels? | 10:42 |
| KotCzarny | erm | 10:42 |
| KotCzarny | nvm, mind barf | 10:42 |
| bencoh | cssu runs a patched kernel, some fixes were merged over time, but even then you'd have the same issues trying to apply recent patches | 10:43 |
| bencoh | maemo-leste runs an almost-mainline kernel ("almost" because of pvr and a few pending patches) | 10:44 |
| bencoh | and support is getting pretty good these days :) | 10:45 |
| duuude | is power management still messy? | 11:00 |
| bencoh | it should be much better than, say, half a year ago | 11:03 |
| bencoh | It might not be on-par with fremantle yet though, but iirc it's pretty close | 11:04 |
| duuude | really? what changed? | 11:04 |
| bencoh | AFAIR: work was done on the OFF mode, and there was something about proactive memory compaction waking up the device more than necessary | 11:05 |
| duuude | does it still empty the entire battery? | 11:06 |
| bencoh | ? | 11:06 |
| duuude | last time it emptied the entire battery when discharging | 11:07 |
| duuude | turning the device on again was painful | 11:07 |
| bencoh | you mean, under the trickle charge threshold? | 11:07 |
| bencoh | (ie "deep" discharge) | 11:07 |
| duuude | yes | 11:08 |
| bencoh | I think this could happen with fremantle as well, except that fremantle's bme would poweroff way before that happens | 11:09 |
| bencoh | but if you replace bme with its opensource counterpart (kernel driver + some userspace script iirc) I think you'd have the same issue as in leste | 11:09 |
| duuude | ~bme | 11:10 |
| bencoh | I don't remember what cssu uses nowadays, I'm not certain opensource battery management was pushed to cssu-stable, it might be only in -devel (and maybe -testing) | 11:11 |
| bencoh | https://wiki.maemo.org/N900_Software_BME | 11:11 |
| bencoh | I think that page lacks info on the final replacement solution but apart from that it looks pretty detailed and accurate | 11:13 |
| Wizzup | yeah so there's still pm work to do in mainline kernel, it does enter off mode with some patches but some drivers keep it awake | 11:16 |
| Wizzup | it's not clear which parts do | 11:16 |
| duuude | can one have this bme thing on leste? | 11:17 |
| bencoh | the closed-source one you mean? I highly doubt it, and a better solution would be to just poweroff device at 10~20% instead of letting it reach close-to-empty state | 11:18 |
| Wizzup | mce has this already | 11:19 |
| Wizzup | (on leste) | 11:19 |
| bencoh | configurable you mean? | 11:20 |
| bencoh | or a hardcoded 10~20%? | 11:20 |
| Wizzup | there is some hardcoded voltage limit I think | 11:25 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!