| --- Day changed Sun May 26 2013 | ||
| @juiceme | Jonni, ping | 00:41 |
|---|---|---|
| @juiceme | rikanee, ping | 00:41 |
| @juiceme | I've been looking through the logs, and it still looks like it is the twl4030 watchdog that either causes the problem or it's an indication of the problem. | 00:44 |
| @juiceme | as far as I have been able to understand, in twl4030_wdt.c it's not allowed to stop the watchdog when WATCHDOG_NOWAYOUT is set, ever. | 00:48 |
| @juiceme | still somebody does that, closes the watchdog file or otherwice releases it. | 00:49 |
| @juiceme | I'd imagine it's the dsme, as there's the loggings: "dsme wdd: Exited main loop, quitting", "dsme wdd: Couldn't remove lockfile" | 00:52 |
| @juiceme | Now what could make dsme-server to throw the spoon to corner? | 00:53 |
| @juiceme | ahh, dsme & libdsme ar on the PR1.3 cdrom... have to take a peek | 01:15 |
| -!- Jonni [~rainisto@2001:41d0:1:b16c::4] has quit [Remote host closed the connection] | 01:46 | |
| -!- Jonni [~rainisto@2001:41d0:1:b16c::4] has joined #ubiboot | 01:51 | |
| -!- Jonni [~rainisto@2001:41d0:1:b16c::4] has quit [Remote host closed the connection] | 01:56 | |
| -!- Jonni [~rainisto@2001:41d0:1:b16c::4] has joined #ubiboot | 01:56 | |
| rikanee | juiceme: I'm guessing some components may need to be patched. | 02:08 |
| -!- rikanee [~rika@unaffiliated/rikanee] has quit [Ping timeout: 252 seconds] | 08:43 | |
| -!- rikanee [~rika@unaffiliated/rikanee] has joined #ubiboot | 08:52 | |
| @juiceme | rikanee, well, I hope it does not need patching of dsme. It would be easiest if the correction could be done to the l2-fix kernel or to ubiboot kernel/utilities. | 12:03 |
| @juiceme | Now what I guess this could be like; when the device first boots with ubiboot, watchdog is initialized and taken into use. when the second boot to Harmattan occurs, the watchdog should be reinitialized to an empty state, whatever that is, like the device was just booted from cold state. | 12:05 |
| @juiceme | now there's propably something going on there in the doggy which causes the problems. | 12:06 |
| @juiceme | Can you share any light on how the watchdog is implemented on the HW level, I know that on kernel layer it's presented as a device that you just need to write something at periodic intervals, but how is the actual HW in the OMAP implemented, and at what state should I set it before I kexec() the 2nd stage kernel? | 12:08 |
| -!- Jonni [~rainisto@2001:41d0:1:b16c::4] has quit [Remote host closed the connection] | 12:53 | |
| -!- Jonni [~rainisto@2001:41d0:1:b16c::4] has joined #ubiboot | 12:54 | |
| valdur55 | juiceme: I found interesting command: act-dead-charging-indicator | 13:19 |
| valdur55 | It's just indicator, nothing more :) If you connect changer, then it will indicate charging. | 13:20 |
| rikanee | juiceme: watchdog is implemented as a function of TWL4030 | 16:58 |
| rikanee | no idea how Nokia configured it though - 'no resets after boot' may have been a security measure | 16:59 |
| rikanee | it needs to get pinged every 60 seconds on the N900 (apparently, less on the N9/50), else it forces a hardware reboot (which it can, because it also manages platform power) | 16:59 |
| @juiceme | Jonni, ping | 23:03 |
| Jonni | pong | 23:04 |
| @juiceme | there's 2 watchdogs in N9, omap_wdt and twl4030_wdt, what's the difference and are both used? | 23:05 |
| @juiceme | am looking into the shutdown, propably I need to disable the watchdogs properly when launching 2nd stage kernel | 23:06 |
| Jonni | do idea :) | 23:06 |
| Jonni | no idea :) | 23:06 |
| Jonni | even | 23:06 |
| @juiceme | :) | 23:06 |
Generated by irclog2html.py 2.12.1 by Marius Gedminas - find it at mg.pov.lt!