steven m. parrish wrote:
OS13 is now available for testing.
excellent stuff!
i'm sorry i didn't get a chance to try os12, but i just installed os13.
suspend/resume, and powerd, seem to be working mostly fine, as far as i've tested. (which wasn't all that far.)
there are a few things i noticed that are a bit of a concern, and arise directly from me having been concentrating entirely on XO-1.5, and not considering the effects of some changes on XO-1.
- the sleep/dim/blank timers are tuned for XO-1.5. in os13 sleep happens in 15 seconds, dim happens 45s later, and blank happens 3 minutes after that. this behavior is quite different than anything we've shipped on XO-1 before -- previously we dimmed, then slept. (and never blanked, though we would have liked to.) the only effects of this change i've seen so far is that the wakeup-on-keystroke delay is more noticeable than XO-1.5 (obviously), and on the network view, the access points seem to all disappear more often than they used to. i haven't looked at why that last thing happens. it's certainly annoying. since i think the wlan card stays powered, i don't really understand it.
- the "config_MESH_DURING_SUSPEND" setting is set to "no", which means the wireless won't stay on when the laptop is sleeping. i think ohmd runs some code to decide whether it should leave mesh active or not -- there are "todo" comments to that effect in powerd, but i've never looked at it further. i don't know if the "relay mesh for others while sleeping" feature is used in the field -- i suspect not.
- powerd now integrates most of olpc-pwr-log, which will let us gather battery and charging data from the field much more readily. i don't know that we necessarily want to run this on XO-1 all the time -- it's definitely not free in terms of cpu (even if it only runs every minute or so), or disk. the logs it keeps (which start out in /var/log, and are moved to /home/olpc/olpc-pwr-log) may take more space than XO-1 users would like to give up. (currently capped at 10M)
- in the next releases of powerd, there are even more changes surrounding wake-on-wlan that may have an effect on XO-1. i'll try and be more proactive about keeping XO-1 working, but it will have to be lower priority than XO-1.5, at least for a while.
in any case, all of this points at having separate default configurations for powerd on XO-1 and XO-1.5. i guess this should be handled by the rpm install -- but i'm not sure how. (help appreciated.)
thanks again for all your great work!
paul =--------------------- paul fox, pgf@laptop.org