**** BEGIN LOGGING AT Fri Sep 10 02:59:57 2010 Sep 10 05:32:32 lindi-, I applied mem=127M to uboot bootargs, recompiled the kernel with your patch, copied the modules to the device (overwriting old ones), flashed the kernel, ran depmod. Then I caused the kernel panic and had to remove/reinsert battery and turned on the device again. I then ran a.out (from ramdisk-dump.c). Error-message: invalid magic 0x646f6d5f Sep 10 06:00:02 trev2: removing the battery corrupts mem somewhat Sep 10 06:00:09 trev2: isn't there an option to ignore magic? Sep 10 06:00:50 what's valid le elf? Sep 10 06:06:23 PaulF, I read a bit more and it seems like using watchdog to do a reset on panic is the way to go Sep 10 06:06:42 Is watchdog enabled in the kernel config before compile or can I activate it somehow? Sep 10 06:07:36 do gta0x have wdt? Sep 10 06:08:00 don't know what that is Sep 10 06:08:11 whihc? Sep 10 06:08:19 wdt Sep 10 06:08:24 hi,did someone use 0xFFFF Sep 10 06:08:24 oh, watchdog timer Sep 10 06:08:53 well I have no idea Sep 10 06:09:00 I need something to do a reset on kernel panic Sep 10 06:09:06 are there other options? Sep 10 06:09:20 talk to DocScrutinizer Sep 10 06:10:11 DocScrutinizer, have you got any idea on how to do an automatic reset on kernel panic so that memory doesn't get corrupted and I can catch the kernel dump after reboot? Sep 10 06:10:57 I have a kernel patch from lindi handling everything. I just have to make sure memory doesn't get corrupted Sep 10 06:11:34 oh, I misunderstood, I thought you wanted hardware watchdog Sep 10 06:11:44 either linux or the android tree has reboot on panic with panic timer Sep 10 06:11:47 hi DocScrutinizer Sep 10 06:12:00 andorid also can right NAND with dmesg without needing trip through uboot Sep 10 06:13:39 I am trying to fetch the kernel panic error message or at least the dmesg output before the kernel for about 12 hours now Sep 10 06:13:48 I have no debug board Sep 10 06:14:00 I get tired *yawn* Sep 10 06:14:33 thx for your help tmzt_ :) Sep 10 06:16:58 in the patch readme it says: Sep 10 06:17:00 "2) panic=20 and/or watchdog is needed so that user can recover the message after reboot" Sep 10 06:17:01 usb serial on new kernels? Sep 10 06:17:10 what is meant by "panic=20" ? Sep 10 06:17:15 is this a kernel boot argument? Sep 10 06:17:17 add that to the command line Sep 10 06:17:19 yep Sep 10 06:17:24 panic=0 stops the reboot Sep 10 06:17:30 very useful on hacked up android kernels Sep 10 06:17:54 so panic=20 restarts after 20 secs of panicing or what does it do? Sep 10 06:18:46 it restarts after 20 seconds Sep 10 06:18:51 shows the message for 20 seconds Sep 10 06:19:01 panic=0 or missing (unless default) means forever Sep 10 06:19:19 sounds good Sep 10 06:19:22 will try Sep 10 06:28:07 mkay it's panicing now Sep 10 06:28:19 yeah baby! REBOOT! :D Sep 10 06:30:06 now it hangs on boot Sep 10 06:30:08 great Sep 10 06:38:23 btw. does anyone know how to mount the NAND kernel partition? Sep 10 06:39:18 trev2: the watchdog is enabled in kernel by default, and you need user-space app installed. Sep 10 06:39:39 trev2: there's no way to mount the kernel NAND partition because it's not a filesystem, just the kernel image. Sep 10 06:39:56 Where to supply kernel boot arguments to qi then if I boot from NAND? Sep 10 06:40:28 I already tried placing a seperate kernel image and append-xxx file in /boot but it didn't take my commandline arguments Sep 10 06:40:43 trev2: nowhere :( hardcode in the kernel or recompile Qi and hardcode in there. Sep 10 06:40:53 Now that sucks Sep 10 06:41:14 trev2: Qi was developed with the assumption that booting from NAND sucks :) Sep 10 06:41:22 trev2: it's SD-centric. Sep 10 06:41:49 The changes I made to the NOR u-boot promt with setenv bootargs_base xxxxxxxxxx are not being saved permanently, are they? Sep 10 06:43:17 I guess the fastest way would be to replace qi with u-boot and make changes to u-boot NAND bootargs? Sep 10 06:48:07 trev2: yes Sep 10 06:48:14 trev2: yes Sep 10 06:48:31 PaulF, I really appreciate your help. Thank you Sep 10 06:48:37 trev2: the idea of nor u-boot is that it always starts with default env. Sep 10 06:48:39 I feel like I am getting somewhere Sep 10 06:48:40 :) Sep 10 06:50:18 I don't know how your ethic opinion about this is but you might have heared of the playstation 3 hack which enables homebrew software. It has been ported to many portable devices. I'm giving it a go on the Freerunner. I only had to implement 3 functions to port it but now I got a kernel panic and am stuck for many hours because I can't recover the message Sep 10 06:50:23 but now it see the light Sep 10 06:50:51 you have it working on s3c? Sep 10 06:51:22 well it compiles, loads, initializes, starts... but when I initiate the injection, I get a kernel panic Sep 10 06:51:30 so far I haven't been able to recover the message Sep 10 06:51:34 but I think I am close Sep 10 06:53:14 hi DocScrutinizer Sep 10 06:54:39 tmzt_, http://pastebin.com/ZC8FmqJd Sep 10 07:02:56 trev2: #psgroove Sep 10 07:03:24 oops Sep 10 07:03:29 #psfreedom Sep 10 07:06:06 yeah I am in there but they do not care about s3c Sep 10 07:09:22 trev2: i'm all of porting it to FR Sep 10 07:09:45 trev2: i even tried to help KaKaRoTo with n810 port and with musb oddities etc. Sep 10 07:10:01 :) Sep 10 07:10:55 * trev2 puts PaulF on his list Sep 10 07:11:03 not the black one! Sep 10 07:11:16 trev2: the main trick there is to change the device address when needed. Sep 10 07:11:49 I implemented psfreedom_get_address and psfreedom_set_address and the get_endpoint thingy Sep 10 07:12:24 udc_write((tmp | S3C2410_UDC_FUNCADDR_UPDATE), Sep 10 07:12:24 S3C2410_UDC_FUNC_ADDR_REG); Sep 10 07:12:24 s3c2410_udc_set_ep0_de_out(base_addr); Sep 10 07:12:34 yeah Sep 10 07:12:35 I guess this's all that's needed, and it should work :) Sep 10 07:12:38 that's what I used in my code Sep 10 07:12:50 wait, I'll put my code online with pastebin Sep 10 07:12:55 it's real short Sep 10 07:15:21 http://pastebin.com/pvgYLmzj Sep 10 07:15:31 it is my first c attempt so there must be some mistakes Sep 10 07:16:44 trev2: i'm sure you're doing good, that's not an obvious task for a newcomer. Sep 10 07:17:19 I have a lot of Java and a little c++ experience but no c so far Sep 10 07:17:33 I believe if I can grep the kernel panic message, I could make more progress Sep 10 07:18:16 trev2: the kernel panic contains the backtrace, it should help. Sep 10 07:18:51 PaulF, yup I sure hope so! :) Sep 10 07:19:57 Getting back to topic, you said I just have to install watchdog userspace app to make it work because it is enabled by default in the kernel? Sep 10 07:20:23 or more specific the daemon? Sep 10 07:20:37 trev2: and amend the command line to make it auto reboot iirc . Sep 10 07:20:49 trev2: yes, the watchdog daemon. Sep 10 07:21:27 well panic=20 works fine so far Sep 10 07:21:33 is there any disadvantage? Sep 10 07:21:58 trev2: no, as you have kernel panic :) Sep 10 07:22:14 trev2: so, where's the panic message? Sep 10 07:27:01 I didn't accept the commandline arguments. I went into NOR u-boot bootprompt and changed the bootargs_base Sep 10 07:27:03 then I saved Sep 10 07:27:18 it said "writing changes to NAND" Sep 10 07:27:22 didn't work Sep 10 07:27:53 so I was not able to supply the panic=20 arguent and again it did not reboot Sep 10 07:28:05 s/arguent/argument/ Sep 10 07:28:17 trev2: yes, nor u-boot is like that. Sep 10 07:28:55 entering the bootprompt with NAND u-boot didn't work and I thought it would do what it said: "writing changes to NAND" Sep 10 07:29:04 trev2 meant: so I was not able to supply the panic=20 argument and again it did not reboot Sep 10 07:31:45 trev2: with NAND u-boot saving changes should work. Sep 10 07:31:59 trev2: there also are several userspace apps for editing the NAND env. Sep 10 07:32:11 idea, put the psfreedom in uboot :) Sep 10 07:33:05 heh :) Sep 10 07:38:33 booting u-boot in from nand doesn't give me /dev/ttyAM0 Sep 10 07:39:29 trev2: it should while you stay in the menu Sep 10 07:39:51 I am in the menu, there is no /dev/ttyAM0 Sep 10 07:39:57 booting from NOR immediately brings it up Sep 10 07:40:22 morning Sep 10 07:40:27 trev2: ah, probably you need to activate "enable usb console" there Sep 10 07:40:27 morning Sep 10 07:40:32 daniele_athome: morning. Sep 10 07:40:50 PaulF, I thought so but that didn't help Sep 10 07:41:10 trev2: then just edit it from userspace Sep 10 07:41:18 yeah I am on it Sep 10 07:41:27 trev2: i'm sorry but i do not remember how the tool in question is called. Sep 10 07:48:40 Well last time I used it, it totally bugged my config and made u-boot unusable Sep 10 07:48:53 I think I will just use a sdcard based distribution with qi Sep 10 07:49:13 I am sick of fiddling around with unnecessary stuff Sep 10 07:49:42 qi seems more solid Sep 10 07:51:51 it looks so seductive: put rootfs and /boot together with a file full of commandline arguments on a sd card, put it in and smile Sep 10 07:54:56 trev2: yeah, that's the point :) Sep 10 09:09:03 radekp: so reading pcf50633 registers means talking i2c which rules out debug board? Sep 10 09:16:05 lindi-: hi, sorry i have no idea - i am not kernel hacker & hardware guy ;) Sep 10 09:17:32 by measuring the consumption for bug #2349 you definitely qualify :) Sep 10 09:24:26 lindi-: i think i will have to redo the measurement once more Sep 10 09:29:29 lindi-: hm. interesting. i thought yesterday you were not impressed by idea of measurement with video? Sep 10 09:31:18 gena2x: hmm? Sep 10 09:31:47 probably misunderstood me. I just found it too difficult for me to perform :) Sep 10 09:32:05 lindi-: ah. ok :) Sep 10 09:32:32 (and also difficult to do automatically) Sep 10 09:33:01 radekp: I guess we need a function for dumping pcf50633 state Sep 10 09:33:36 lindi-: pcf state is already dumped Sep 10 09:37:31 gena2x: ah, where? Sep 10 09:38:34 "show_dump_regs()"? Sep 10 09:40:00 mrmoku|a`: ping Sep 10 09:40:32 gena2x: /sys/devices/platform/s3c2440-i2c/i2c-adapter/i2c-0/0-0073/dump_regs? Sep 10 09:43:06 radekp: gena2x: is it too late to call show_dump_regs in pcf50633_suspend()? Sep 10 09:43:52 lindi-: if you want to see just pcf registers from userspace http://svn.openmoko.org/developers/werner/pmu/ Sep 10 09:45:05 gena2x: I think we'd like to see pcf registers as late as possible in suspend Sep 10 09:45:33 lindi-: so why did you mentioned /sys path? Sep 10 09:45:56 gena2x: I wasn't sure I was reading the right file :) Sep 10 09:46:28 lindi-: reading from kernel space? Sep 10 09:46:55 gena2x: yep Sep 10 09:47:03 gena2x: did you see bug #2349? Sep 10 09:47:47 lindi-: ah, lars comment Sep 10 09:49:35 lindi-: interesting. i thought pcf state is dumped to kernel log moment before suspending Sep 10 09:50:26 lindi-: but seem not backlight anyway. Sep 10 09:50:38 i can be just some minor problem with pcf backlight - i did echo 1 > /sys/.../blank before started measuring, maybe that can be the problem Sep 10 09:51:05 radekp: you may turn it off via pmu tool i guess Sep 10 09:51:18 but i guess this is not reason Sep 10 09:51:30 agree Sep 10 09:51:38 as backlight is eating much much more than 9mA Sep 10 09:51:54 i have tried to set brightness limit to 0 and still i have 18mA in suspend Sep 10 09:53:27 i have tried to disable all platform devices... Sep 10 09:55:57 radekp: while you are recompiling kernels anyway, could you try something like http://paste.debian.net/89071/ so that we could answer to lars? Sep 10 09:56:23 (that was for 2.6.29) Sep 10 09:58:53 no more /usr/lib/enlightenment in my FR :) at last! :) Sep 10 09:59:04 i'm happy :D Sep 10 09:59:30 daniele_athome: using your own UI now? 8) Sep 10 09:59:45 yeah, as author told - enlightment is not for FR Sep 10 10:00:05 Q-Master: sure! :) Sep 10 10:00:41 daniele_athome: congrats :) Sep 10 10:00:44 * Q-Master is thinking of moving to it, but the keyboard problem is a problem. 8( no time to edit Sep 10 10:00:54 Q-Master, what do you mean? Sep 10 10:01:05 you're the one that wants the foreign keyboard, aren't you? Sep 10 10:01:07 daniele_athome: I need a cyrillic keyboard layout Sep 10 10:01:11 i see Sep 10 10:01:33 the problem is the keys are inside the EDC file... so I should make something easily editable like illume keyboard for example Sep 10 10:01:38 daniele_athome: which UI you found best? Sep 10 10:01:47 gena2x, among which ones? Sep 10 10:01:55 daniele_athome: of course it should be selectable. some sort of programs need both langs Sep 10 10:02:11 Q-Master, but also the number of keys might be different Sep 10 10:02:29 so it's not just a change of characters in the keyboard Sep 10 10:02:30 daniele_athome: yep. russians have much more letters that english ones. Sep 10 10:02:42 daniele_athome: what is 'your own UI'? Sep 10 10:03:05 gena2x, it's a mix of own window manager, own desktop & launcher app, dock panel and a phone application Sep 10 10:03:12 (mokosuite2) Sep 10 10:03:21 all based on EFL Sep 10 10:03:26 but not enlightenment :) Sep 10 10:03:27 jikes Sep 10 10:03:31 ah... Sep 10 10:04:06 lindi-: oki i'll try Sep 10 10:04:07 Q-Master, anyway i should rethink the keyboard at this point... I mean it can only be latin right now... Sep 10 10:05:00 Q-Master, in the meantime you can do a skitch or rapid picture of how (by your point of view) this keyboard should look like Sep 10 10:05:14 s/skitch/mockup/ Sep 10 10:05:16 daniele_athome meant: Q-Master, in the meantime you can do a mockup or rapid picture of how (by your point of view) this keyboard should look like Sep 10 10:05:35 i might implement that as a different EDC on its own Sep 10 10:07:00 daniele_athome: I've edited some time ago this keyboard. (it's vkbd not?) but seems that i've lost my work Sep 10 10:07:22 Q-Master, yes but i fixed some sizes to fit on FR screen, both landscape and portrait Sep 10 10:07:23 daniele_athome: do you have any HTC android device nearby? Sep 10 10:07:32 Q-Master, no Sep 10 10:09:18 daniele_athome: me too. I found some experiments on this keyboard. can you tell me your mail to send you a sketch? Sep 10 10:11:29 (sketch, that's the word :D ) Sep 10 10:12:28 anybody an idea when mrmoku and jama will be back again? Sep 10 10:13:47 yeah... many things need to be bumped :( Sep 10 10:27:13 Hey kernel guys, someone knows whether shr uses udev or not? Sep 10 10:28:03 papa-piet: unstable - not. Sep 10 10:42:04 DocScrutinizer, hi I've some questions: Sep 10 10:42:11 *does 0xFFF work? Sep 10 10:42:25 *how do I install bootmenu-n900 Sep 10 10:42:35 all my attempt to install that app failed Sep 10 10:42:48 more specifically Sep 10 10:42:52 I fail to add the repo Sep 10 10:43:30 [2010-09-09 23:47:26] anyway neither me nor for sure PaulF did ever try 0xFFFF Sep 10 10:44:07 ok Sep 10 10:44:11 don't you think #maemo is a chan more adequate for those questions? Sep 10 10:44:15 0xFFF didn't seem to work Sep 10 10:44:15 ok Sep 10 10:57:37 lindi- gena2x: i have backported backlight driver from 2.6.29 - it was quite easy, but current is unchanged Sep 10 10:58:19 radekp: ok Sep 10 10:58:25 so the problem is elsewhere Sep 10 10:58:39 radekp: however, lars was interested in the pcf registers, can you add that info? Sep 10 11:02:03 radekp: as expected. what's with gadget thing? Sep 10 11:04:23 gena2x: with gadget? Sep 10 11:07:01 [ 1074.795000] s3c2440-usbgadget s3c2440-usbgadget: suspend -> 80.9 Sep 10 11:07:01 [ 1074.795000] gta02_udc_command S3C2410_UDC_P_DISABLE Sep 10 11:07:01 -6mA Sep 10 11:07:36 look very similar to voltage we searching for Sep 10 11:07:48 s/voltage/current/ Sep 10 11:07:48 gena2x meant: look very similar to current we searching for Sep 10 11:33:09 Q-Master, bluetooth daemon doesn't get activated automatically if udev is not present Sep 10 11:33:16 one is supposed to start on its own? Sep 10 11:34:22 daniele_athome: SHR has a problem with bluetooth because of no udev Sep 10 11:35:14 Q-Master, i'm talking about right that: said that udev is unstable, they could at least provide some way to start bluetoothd Sep 10 11:35:22 on boot or something else Sep 10 11:37:08 daniele_athome: heh. dunno what to answer. 8) Sep 10 11:37:12 :) Sep 10 11:37:19 Q-Master, don't worry just an exclamation Sep 10 11:37:20 :) Sep 10 11:41:28 8) Sep 10 11:49:24 daniele_athome: i can tell you that OpenWRT uses a minimal udev replacement that you might want to use on SHR as well. Sep 10 11:49:39 mmm Sep 10 11:49:48 daniele_athome: they use it to activate the user-space firmware helper of course. Sep 10 11:49:59 PaulF, it is just a recompiled stripped version or it's written from scratch? Sep 10 11:50:19 stripped I mean only the needed functions and objs etc. Sep 10 11:52:14 daniele_athome: it's some written from scratch minimal helper, not sure if that's needed for shr as in fact FSO already does have a fw loader plugin. Sep 10 11:52:27 mmm Sep 10 11:52:52 PaulF, what about writing a bluetooth plugin for fsodeviced? Sep 10 11:53:05 that in fact replaces udev and loads bluetoothd Sep 10 11:53:10 daniele_athome: since bluetooth is FSO-activated resourse i'd expect FSO to somehow start it. Probably it should "ping" bluetooth after enabling the resource to get bluetoothd activated by dbus. Sep 10 11:53:25 mmm Sep 10 11:53:31 daniele_athome: or even that's not needed as the first bluetooth-using app will do that. Sep 10 11:53:43 daniele_athome: yeah, why not use dbus activation for bluetoothd? Sep 10 11:54:05 i'll try something Sep 10 11:54:36 until it is done in FSO, it wouldn't be too hard to write a small app that listens for the bluetooth resource being enabled and then checkes to see if bluetooth is running; if not, start it Sep 10 11:56:54 PaulF: I'm using the latest unstable out of NAND; I've been getting lots of kernel panics; they usually go in spurts; could that be an indication of bad blocks in NAND? Sep 10 11:57:47 net_tux: no idea tbh, and also i'm still not sure bad block handling is done right. I'd just use SD. Sep 10 11:58:08 yesterday, it kept panicing shortly after getting booted up; I finally gave up and left it off for several hours; then next time I booted it, it was fine Sep 10 11:58:10 ok Sep 10 11:59:43 it's really strange, sometimes it is random; other times it seems to be related to doing something like plugging in headphones Sep 10 12:57:49 gena2x: iteresting question - how comes that i have 30mA in suspend when i test it from console but only 18mA when i am running suspend from qtmoko Sep 10 13:00:08 radekp: gsm? Sep 10 13:00:22 that's my idea too Sep 10 13:00:47 i have powered it of with echo 0 > /sys/.../gsm/power_on Sep 10 13:00:56 but maybe it still stays on? Sep 10 13:02:50 lindi-: here is pcf regs dump http://pastebin.com/bcSCXv0y Sep 10 13:03:23 GNUtoo|laptop: is it possible to create a dev node for a touchscreen on palm pre without udev Sep 10 13:03:39 lindi-: i cant add comments to bug tracker, tried also creating new account but no luck - i tried even with 3 different web browser Sep 10 13:04:08 GNUtoo|laptop unfortunately i'm not familiar with mknod or makedev Sep 10 13:04:10 papa-piet, I don't know, I've devtmpfs Sep 10 13:04:18 on the htcdream Sep 10 13:04:20 arg Sep 10 13:04:22 and n900 uses udev I bet Sep 10 13:04:29 so just install udev in the image right? Sep 10 13:04:36 pre also but not shr Sep 10 13:04:44 ah ok Sep 10 13:04:47 so what then? Sep 10 13:04:52 debian? Sep 10 13:05:00 fso-demo2-image? Sep 10 13:05:10 i work via NFS-rootfs Sep 10 13:05:41 GNUtoo|laptop: after installing udev nfs-bbot fails.... Sep 10 13:06:02 papa-piet, that's easy Sep 10 13:06:08 you need to rm something Sep 10 13:06:19 and to boot with rw args Sep 10 13:06:41 rm something? Sep 10 13:07:42 papa-piet, try that on the target filesystem: Sep 10 13:07:52 rm target_fs/etc/init.d/devices Sep 10 13:09:57 papa-piet_, rm target_fs/etc/init.d/devices Sep 10 13:10:09 target_fs is where your microsd is mounted Sep 10 13:10:35 GNUtoo|laptop, no microsd on pre :/ Sep 10 13:11:02 ah right Sep 10 13:11:10 so how do you install an image? Sep 10 13:11:26 but there is no devices in etc/init.d/ Sep 10 13:11:35 via nfs Sep 10 13:11:39 then try to put rw as kernel argument Sep 10 13:11:51 okay Sep 10 13:12:26 GNUtoo|laptop, thank you, will try that Sep 10 13:12:32 ok Sep 10 13:12:39 also try debug configuration Sep 10 13:12:41 that is to say Sep 10 13:12:47 print the messages on the framebuffer Sep 10 13:12:56 � Sep 10 13:13:08 radekp: really? Sep 10 13:13:36 lindi-: yup, it seems that openmoko bug tracker does not like me Sep 10 13:15:23 PaulF: but with 2.6.29 i can suspend from console and have just 9mA current Sep 10 13:21:13 radekp: that is not a good idea Sep 10 13:21:20 radekp: i mean, we need a bug tracker Sep 10 13:21:42 radekp: I have 3mA current in suspend with 2.6.29 Sep 10 13:23:37 lindi-: interesting, how do you suspend it? Are your doing it from console? Sep 10 13:39:41 radekp: console? Sep 10 13:39:51 radekp: I use the standard pm-suspend command of course Sep 10 13:40:08 i tried AT@POFF and now i have 16..18mA with 2.6.34 Sep 10 13:40:27 radekp: I use "om gsm power 0" Sep 10 13:41:16 lindi-: do you know if there is difference between pm-suspend x echo mem > /sys/power/state? Sep 10 13:41:41 and between om gsm power 0 and echo 0 > /sys/path/to/gsm/power_on ? Sep 10 13:55:33 radekp: no diff Sep 10 13:55:41 radekp: for gsm there can be diff Sep 10 13:58:46 lindi-: hmm i have 9.1 mA with 2.6.29 Sep 10 14:00:41 radekp: when/where do you dump the pcf registers? Sep 10 14:01:10 larsc: hi, dumped them in pcf-core device suspend Sep 10 14:02:08 larsc: patched pcf driver according to http://paste.debian.net/89071/ Sep 10 14:05:58 radekp: some regulators might be disbled later Sep 10 14:06:32 larsc: can i make dump later? Sep 10 14:08:42 larsc: btw i dont think it's backlight driver which is wrong... i backported the old generic-bl driver to 2.6.34 and it had no effect on power in suspend Sep 10 14:13:05 GNUtoo|laptop, print messages say: http://shr.pastebin.com/QwXaaj9j i guess it's somewhere here Sep 10 14:13:23 ok Sep 10 14:13:24 maybe Sep 10 14:13:32 try a blank /etc/network/interfaces Sep 10 14:26:04 radekp: you can use dev_pm_ops and the suspend_noirq callback Sep 10 14:30:39 larsc: oki, give me a few minutes Sep 10 14:35:26 larsc: do you need it also for 2.6.29 kernel? Sep 10 14:36:24 GNUtoo|laptop, no difference :/ Sep 10 14:36:34 papa-piet_, hmmm Sep 10 14:36:45 maybe ask #oe? Sep 10 14:37:44 GNUtoo|laptop, i mean kernel mounts the rootfs via nfs but after udev is started... Sep 10 14:37:48 radekp: well, since the mmc regulator is enabled for 2.6.29 in your dump i would say yes Sep 10 14:37:58 papa-piet_, and without NFS? Sep 10 14:38:01 ahh I get it Sep 10 14:38:12 will test that today, Sep 10 14:38:12 NFS....that's the cause Sep 10 14:38:17 I'll explain Sep 10 14:38:26 basically it mount NFS Sep 10 14:38:29 and then block there Sep 10 14:38:38 because it does something with the network interface Sep 10 14:38:46 GNUtoo|laptop, yes i guess so too Sep 10 14:38:51 tell /etc/network/interfaces not to do anything Sep 10 14:38:56 because if it reset the ip Sep 10 14:39:03 or ifconfig down Sep 10 14:39:08 you loose the NFS connection Sep 10 14:39:19 btw audio is so great on n900 Sep 10 14:39:25 larsc: hmm i am not sure how to implement it for 2.6.29 Sep 10 14:39:27 i've done echo "" > etc/network/interfaces Sep 10 14:40:00 papa-piet_, hmmm try to understnad the syntax Sep 10 14:40:18 or remove the init script that brings up the interfaces [HACK] Sep 10 14:41:05 radekp: should there be a late_suspend or something like that callback in the platform_driver structure? Sep 10 14:42:55 GNUtoo|laptop, sorry, what syntax? of udev? Sep 10 14:44:04 papa-piet_, /etc/network/interfaces Sep 10 14:45:26 larsc: in 2.6.29 we are fillin .driver which is struct device_driver Sep 10 14:45:37 larsc: it has only suspend and resume Sep 10 14:46:02 larsc: maybe if i filled driver.pm which is dev_pm_ops it would work Sep 10 14:54:39 it compiled ok, but i have screwed my rootfs.. will take a few minutes to get it back Sep 10 15:13:13 larsc: i think suspend_noirq is not getting called - i see no dump in dmesg now Sep 10 15:13:47 larsc: hmm maybe i am returning wrong value Sep 10 15:16:01 return 0 seems ok, but i see no dump in log Sep 10 15:19:35 larsc: http://activationrecord.net/radekp/pub/dump_regs_noirq_29.patch Sep 10 15:19:41 larsc: http://activationrecord.net/radekp/pub/dump_regs_noirq_34.patch Sep 10 15:19:51 maybe you can see something wrong Sep 10 15:23:54 looks good to me Sep 10 15:26:33 * lindi- added some info to bug #2349 Sep 10 15:29:00 radekp: i'll have to go now. will be back on monday Sep 10 15:29:14 larsc: oki, see you Sep 10 15:43:53 freesmartphone.org: 03Frederik.Sdun 07cornucopia * r5fb6ba8bdc11 10/tools/mdbus2/data/mdbus2: mdbus2: update bash completion Sep 10 16:02:35 freesmartphone.org: 03Frederik.Sdun 07cornucopia * r7c54e0fa49d5 10/tools/mdbus2/data/mdbus2: Sep 10 16:02:35 freesmartphone.org: mdbus2: remove properties in bash completion Sep 10 16:02:35 freesmartphone.org: fix a small bug introduced by last commit Sep 10 16:23:53 mickeyl, hi Sep 10 16:24:01 hi dos1 Sep 10 16:24:17 dos1, xorg.conf is broken, try that: Sep 10 16:24:22 hi Sep 10 16:24:28 Option "InvertY" "True" Sep 10 16:24:29 Option "Calibration" "159 3880 354 3839" Sep 10 16:24:36 in touchscreen input device Sep 10 16:24:55 GNUtoo|laptop: calibrating with xinput calibrator on first run worked for me Sep 10 16:24:55 else big issue: sound, yes again!!! Sep 10 16:24:59 ok Sep 10 16:25:17 and -ao sdl didn't find the sound card this time Sep 10 16:25:22 basically for sound: Sep 10 16:25:27 play some music with mplayer Sep 10 16:25:31 go in settings Sep 10 16:25:36 of enligtenment Sep 10 16:25:44 go to extensions Sep 10 16:25:46 go to modules Sep 10 16:25:48 and scroll Sep 10 16:25:50 and boom Sep 10 16:26:05 sounds is cut a lot Sep 10 16:26:06 GNUtoo|laptop: but you have to have libstdc++6 installed - there is (or was?) some issue with it not being pulled into image Sep 10 16:26:15 no I opkg installed it Sep 10 16:28:29