**** BEGIN LOGGING AT Sun Sep 26 02:59:57 2010 Sep 26 12:35:22 hi Sep 26 12:35:36 ogra: Any smartbook news ? Sep 26 12:38:38 dcordes, http://tosh-ac100.wetpaint.com/thread/4253563/kernel+boot Sep 26 12:39:36 ogra_cmpc: good stuff Sep 26 12:52:13 hi, we also have a channel (another one ;-) -> #ac100 Sep 26 12:54:15 marvin24, awesome Sep 26 12:54:18 * ogra_cmpc joins Sep 26 13:24:38 persia, ^^^^ Sep 26 13:25:41 marvin24_: cool Sep 26 13:25:55 I'm excited, but don't have the attention for another channel just now. Sep 26 13:26:14 persia, i meant the url :) Sep 26 13:26:38 the URL looks very promising indeed. Sep 26 13:26:43 yes Sep 26 13:26:55 i can boot android from a local kernel Sep 26 13:27:03 using this reciepe Sep 26 13:27:07 Can you boot a useful operating system? Sep 26 13:27:30 probably ... the prob is that i cant get any visible console Sep 26 13:27:39 ah Sep 26 13:27:47 common androidism Sep 26 13:27:49 consoles are overrated :) Sep 26 13:28:46 trying the tegra kernel with omap4 rootfs doesnt get me anywhere either Sep 26 13:29:04 * ogra_cmpc just tries with a panda SD here Sep 26 13:30:20 ogra_cmpc: check if you have framebuffer console (fbcon) enabled in your kernel ( /proc/conf.gz ) Sep 26 13:30:38 dcordes, its off Sep 26 13:30:42 thats the prob :) Sep 26 13:31:11 ogra_cmpc: ok. did you also request the sources ? Sep 26 13:31:19 no Sep 26 13:31:52 and i didnt checl the link from the EULA for two weeks either Sep 26 13:32:01 ac100 is a weekend project for me Sep 26 13:32:16 so i'm very slow on it Sep 26 13:32:17 to me all of this is a weekend project ;) Sep 26 16:28:12 persia, so i have a way to run a full ubuntu netboot session ... but no input devices work at all Sep 26 16:28:23 s/netboot/netbook Sep 26 16:51:30 hi all , i have connected usb to serial adapter which is detected as /dev/ttyUSB0 and then i have connected to it a straight null modem cable which i have shorted at 2 and 3 pin number but i am not able print anything in minicom or getting echo of it Sep 26 21:56:43 ogra, Nice progress. My recommendation would be to start an image that allowed ssh so you can get in there, and then play with input-utils to see where things are wrong: sounds like issues with the HID drivers. Sep 26 22:02:10 persia, heh, that would require xinput to work at all Sep 26 22:02:22 the system is still chrooted Sep 26 22:02:29 no udev, no upstart Sep 26 22:03:20 the android system is the most minimal but its still a chrooted android Sep 26 22:03:50 Ah, without udev, I have little advice. But working with ssh/input-utils shouldn't require xinput for a properly constructed image. Sep 26 22:04:22 well, the touchpad works with an xorg.conf Sep 26 22:04:28 And if input-utils can't deal with a device, the chances that xinput knows what to do are strikingly low (unless, for some reason, one is dealing with non-HID devices) Sep 26 22:04:37 the kbd is recognized in the logfile Sep 26 22:04:49 but doesnt hand anything through Sep 26 22:05:09 [ 11283.000] Generic KBD: dropping event due to full queue! Sep 26 22:05:20 thats what i see in the log Sep 26 22:05:30 Touchpad can be fixed with some udev properties then: that's trivial. keyboard probably still requires fiddling with input-utils to make sure there is connectivity, symbol mapping, etc. Sep 26 22:05:44 i dont need to ssh ... androids adb shell is running Sep 26 22:07:45 If you don't need to ssh, install input-utils, and try to probe the keyboard. I've locked up a couple systems doing that with exceedingly exotic input devices, (reboot was always clean), but usually they get enough low-level information that one can make the higher levels work. Might need udev properties. Might need remapping. Might need explicit support in the HID driver, etc. Sep 26 22:10:08 well,, no udev properties possible without udev :) Sep 26 22:11:33 No, but it's possible to determine which properties are required :) Sep 26 22:11:47 * persia really ought file some remapping for the Saitek Commander one of these days. Sep 26 22:12:07 /dev/input/event0 Sep 26 22:12:08 protocol version mismatch (expected 65537, got 65536) Sep 26 22:12:27 thats what i get for :/# input-kbd /dev/input/event0 Sep 26 22:14:43 same thing for input-evente Sep 26 22:14:51 *input-events Sep 26 22:14:56 Bother. I thought I fixed that. Sep 26 22:15:38 well, given the million lines of " Generic KBD: dropping event due to full queue!" in the xorg log ... Sep 26 22:15:59 that might not actually be input-utils fault Sep 26 22:16:46 i wish we still would build the old kbd driver for x Sep 26 22:16:58 so i had actually some alternative to try Sep 26 22:17:17 but there is only evdev nowadays Sep 26 22:18:00 input-utils queries the kernel directly. Nothing to do with X. It makes debugging the new X input scheme *much* easier. This is all good: don't be nostalgic: it used to take forever for those of us who cared to painstakingly map things in the kernel and then *do it all over again* in an X driver. Sep 26 22:18:29 Try `input-kbd 0` rather than /dev/input/event0 Sep 26 22:18:40 Also, try lsinput to make sure 0 is your keyboard :) Sep 26 22:21:40 :/# lsinput Sep 26 22:21:40 /dev/input/event0 Sep 26 22:21:41 protocol version mismatch (expected 65537, got 65536) Sep 26 22:21:55 Right. Kernel bug :) Sep 26 22:22:18 well Sep 26 22:22:33 [12262.640000] init: untracked pid 20189 exited Sep 26 22:22:48 android *is* a kernel bug Sep 26 22:23:01 Yep. Sep 26 22:23:03 * persia decides to sort out why the W axis isn't working on this mouse another day, and goes in search of food. Sep 26 22:23:34 yeah, i'll better finish the day :) **** ENDING LOGGING AT Mon Sep 27 02:59:57 2010