**** BEGIN LOGGING AT Sun Dec 25 02:59:57 2011 Dec 25 04:17:17 Good day all , my SHR was ok for about year or more but have not used fr for a while after that SHR stops at booting page , why ? Dec 25 11:12:26 hi JaMa|Off Dec 25 11:12:45 multi-machine builds are still broken in SHR? Dec 25 11:27:37 GNUtoo: sort of Dec 25 11:30:09 without rm_work it works, but it's not as effective as it was with old OE-classic, because it needs populate sysroot for each machine and IMHO incorrectly does do_package for each machine too Dec 25 12:24:57 ahh ok Dec 25 12:25:03 so I should remove rm_old_work? Dec 25 12:58:56 JaMa|Off, btw what to bump when adding kenrel modules install to the machine config? Dec 25 12:59:07 I rebuilt the kernel and task-something and it didn't add it Dec 25 12:59:09 in the image Dec 25 12:59:53 *task-base Dec 25 13:07:02 iirc task-base but better git grep for that variable to be sure Dec 25 13:10:21 ok Dec 25 13:15:00 hmmm task-basic uses it too Dec 25 13:15:03 which one we use Dec 25 13:15:08 task-basic or task-base Dec 25 13:15:10 let me look Dec 25 13:18:14 tasks/task-shr-minimal.bb: task-base Dec 25 13:18:44 then I'll retry cleaning the kernel and task-base but the last time I did it it failled to add the module in the image Dec 25 14:52:37 SHR: 03Martin.Jansa 07meta-smartphone * r89ffdb6d2428 10/meta-fso/recipes-freesmartphone/meego-cellular/libcmtspeechdata_git.bb: libcmtspeechdata: it checks for dbus in configure Dec 25 14:52:37 SHR: 03Martin.Jansa 07meta-smartphone * ra17ab4f91816 10/meta-shr/recipes-shr/3rdparty/vagalume_0.8.3.bb: vagalume: fix LICENSE metadata and gst plugins package names Dec 25 14:52:37 SHR: 03Martin.Jansa 07meta-smartphone * rfbc2df6fe03d 10/meta-shr/recipes-shr/3rdparty/numptyphysics_svn.bb: numptyphysics: fix LICENSE metadata Dec 25 14:52:37 SHR: 03Martin.Jansa 07meta-smartphone * r9326fd4d2da3 10/meta-shr/recipes-shr/tasks/task-shr.bb: task-shr: uncomment apps we had included by default Dec 25 15:00:29 JaMa|Off, I don't understand....once again the module is not included Dec 25 15:01:41 I cleaned both the kernel and task-base Dec 25 15:02:02 GNUtoo: you have added kernel module names in MACHINE_EXTRA_RRECOMMENDS, right? Dec 25 15:02:38 yes I guess so Dec 25 15:02:39 let me look Dec 25 15:03:03 and reinstalled task-machine-base on target Dec 25 15:03:20 lol Dec 25 15:03:21 kernel-module-module-omap2430 \ Dec 25 15:03:25 that's wrong Dec 25 15:03:27 I'll correct it Dec 25 15:03:45 I must have been very tired not to see it Dec 25 15:04:52 thanks a lot Dec 25 15:05:19 yw :) Dec 25 15:17:04 JaMa|Off, btw I'll try to work on fixing midori for work Dec 25 15:17:09 because it segfault there too Dec 25 15:17:39 have you seen my backtrace on ML? Dec 25 15:17:54 it seems like gtk bug Dec 25 15:18:04 ah no Dec 25 15:18:20 now on angstrom it fails too Dec 25 15:18:28 maybe it's a dependency problem? Dec 25 15:18:35 see [OE-core] [PATCH] webkit-gtk svn: update to r101488 Dec 25 15:18:45 ok Dec 25 15:18:46 why dependency? Dec 25 15:19:15 because it worked on my bug device Dec 25 15:19:24 under angstrom Dec 25 15:19:26 under oe-core Dec 25 15:19:32 when it segfaulted on shr Dec 25 15:19:38 but maybe the toolchain changed since then Dec 25 15:19:45 and maybe that broke it Dec 25 15:20:08 it also complain abuot some deps before segfaulting Dec 25 15:31:25 is angstrom still using gcc-4.5? Dec 25 15:31:56 because if it's still segfaultin with angstrom/gcc-4.5 then it's less likely toolchin issue Dec 25 15:34:26 midori 0.2.9 Dec 25 15:34:41 I guess that's way too old Dec 25 15:36:20 after building for gta04 and after pushing I'll build for bug 2.0 Dec 25 15:38:25 iirc it was latest when I was sending it Dec 25 15:38:55 ah I have 0.4.3 Dec 25 15:39:07 yes Dec 25 16:01:22 JaMa|Off, I've a huge problem: Dec 25 16:01:37 I saw that the defconfig was not a saved-defconfig for gta04 Dec 25 16:01:42 and contained dirty things Dec 25 16:01:46 so I tried to update it Dec 25 16:02:08 but because of a bug in oe a config that was needed for usbnet switched from =y to =m Dec 25 16:02:26 then I tried to find why in the kenrel instead of oe and at the end I was fed-up Dec 25 16:02:38 so I fixed the compilation error and tough: let's use that as module Dec 25 16:02:51 then I adapted oe to use that as module Dec 25 16:02:54 but at the end I see Dec 25 16:03:13 g_ether doesn't load Dec 25 16:03:16 and wasn't loaded Dec 25 16:03:20 because "no such device" Dec 25 16:03:31 even if omap2430 module is loaded Dec 25 16:03:37 I'm out of idea and really fed-up Dec 25 16:03:50 what should I do to get a correct defconfig? Dec 25 16:04:23 I spent more than one day on it Dec 25 16:04:34 like the day of yesterday + now Dec 25 16:04:38 it's way too much Dec 25 16:04:42 just for ajusting a config Dec 25 16:08:19 JaMa|Off, what should I do for that? Dec 25 16:11:07 just add =y for USB controller? Dec 25 16:11:32 doesn't work Dec 25 16:11:37 it changes the =y in =m Dec 25 16:11:52 then add =y for all dependencies of that controller Dec 25 16:12:10 CONFIG_USB_MUSB_OMAP2PLUS=y becomes CONFIG_USB_MUSB_OMAP2PLUS=m Dec 25 16:12:17 after do_configure process Dec 25 16:12:29 I've done that too Dec 25 16:12:34 there is a bug Dec 25 16:12:52 loading the defconfig with make ARCH=arm xconfig and saving produce the correct output Dec 25 16:12:57 I tried yesterday Dec 25 16:13:43 and if you use defconfig altered by linux.inc sed calls? Dec 25 16:14:20 let me look Dec 25 16:14:37 before do_configure no config is present Dec 25 16:14:53 but right after do_configure the config has been seded or something like that Dec 25 16:14:56 and loaded Dec 25 16:15:01 I've to trace it then Dec 25 16:15:06 let me look Dec 25 16:15:47 that's done in linux.inc Dec 25 16:15:49 cmdline also change btw Dec 25 16:15:50 ok Dec 25 16:16:02 I was unsure between linux.inc and the kernel bbclass Dec 25 16:16:16 so I picked the kenrel bbclass and started to debug from there Dec 25 16:17:14 linux.inc is changing defconfig Dec 25 16:17:18 ok Dec 25 16:17:30 and after changes defconfig is used to create .config with std kernel tools Dec 25 16:17:45 loglevel=3 is added I think Dec 25 16:17:57 so if it changes from =y to =m then it's defconfig or Kbuild issue, not OE Dec 25 16:18:11 but I verified Kconfig Dec 25 16:20:25 do if you copy your defconfig as kernel/.config and do 'yes || ARCH=arm make oldconfig' then it works? Dec 25 16:20:47 let me verify Dec 25 16:20:55 I was loading with xconfig and saving Dec 25 16:23:53 so it's a problem with Kconfig Dec 25 16:24:01 with make ARCH=arm xconfig it works Dec 25 16:24:09 but not anymore with: Dec 25 16:24:16 yes '' | make ARCH=arm oldconfig Dec 25 16:26:31 USB_GADGET_MUSB_HDRC) [M/n/?] m Dec 25 16:37:37 SHR: 03Martin.Jansa 07meta-smartphone * rd78a05da4ec1 10/meta-shr/recipes-shr/3rdparty/vagalume_0.8.3.bb: vagalume: fix gst plugins package names Dec 25 16:42:18 JaMa|Off, Platform Glue Layer [M/y] (NEW) Dec 25 16:42:28 does it means that it didn't choose? Dec 25 16:43:11 because that's a choice which makes USB_MUSB_OMAP2PLUS m or y Dec 25 16:44:08 I means that it wasn't included in defconfig (it's new) Dec 25 16:44:21 it cannot Dec 25 16:44:24 how to include it? Dec 25 16:44:35 choice Dec 25 16:44:36 prompt "Platform Glue Layer" Dec 25 16:44:36 depends on USB_MUSB_HDRC Dec 25 16:44:40 that's the choice Dec 25 16:44:49 It doesn't have a value Dec 25 16:44:57 just save it to defconfig Dec 25 16:45:01 I can't Dec 25 16:45:03 how? Dec 25 16:45:35 save with it set and check resulting .config and add that line to defconfig Dec 25 16:45:48 It won't work Dec 25 16:45:53 let me explain Dec 25 16:46:04 this choice has no value associated with it Dec 25 16:46:16 instead the value got written on USB_MUSB_OMAP2PLUS Dec 25 16:46:38 but this yes '' | make... create an issue Dec 25 16:46:46 since it evaluate platform glue layer first Dec 25 16:46:54 I don't know what it does Dec 25 16:46:59 but maybe it doesn't set it Dec 25 16:47:12 which makes USB_MUSB_OMAP2PLUS have only the M/n possible values Dec 25 16:47:15 which breaks usb Dec 25 16:51:36 so I guess I'm out of luck Dec 25 16:51:48 I've no serial console Dec 25 16:51:53 so I cannot fix that stuff Dec 25 16:52:02 I've fixed the compilation Dec 25 16:52:06 of that =m option Dec 25 16:52:20 but it doesn't work at runtime Dec 25 16:52:30 and it's a pain working without ssh/serial-console Dec 25 16:55:50 great I'm nervous aagain Dec 25 16:59:59 JaMa|Off, what should I do now? Dec 25 17:02:37 when I do modprobe g_ether it still says no such device Dec 25 17:02:47 but it doesn't give *any* indications on dmesg Dec 25 17:02:49 dmesg -c Dec 25 17:02:54 modprobe g_ether Dec 25 17:02:55 dmesg Dec 25 17:03:01 doesn't give any indications Dec 25 17:03:11 no message get printed in the second dmesg Dec 25 17:13:22 JaMa|Off, ping should I abandon? Dec 25 17:26:06 JaMa|Off, else what can I do? Dec 25 17:26:12 I mean multi-machine is broken Dec 25 17:26:35 I cannot do oe because morphis didn't push his devshell Dec 25 17:26:43 I cannot work on another machine Dec 25 17:27:06 s/do oe/do cornucopia work/ Dec 25 17:27:56 and the touchscreen is somewhat problematic on 3.2 Dec 25 19:27:26 JaMa|Off, hi Dec 25 19:27:32 fatal: loose object d78a05da4ec180939facf17d9e3934355f451663 (stored in .git/objects/d7/8a05da4ec180939facf17d9e3934355f451663) is corrupt Dec 25 19:29:00 I'll re-clone it Dec 25 19:29:57 GNUtoo: Once I had to change the defconfig in oe and the easier way I came up with is following... Dec 25 19:32:49 I took codesorcery, and started linux souce with "make menuconfig" loaded the config from OE, made changes I wanted, saved this as alternate config and changed defconfig in OE with the new alternate config. Dec 25 19:33:00 GNUtoo: marry xmas... Dec 25 19:33:54 GNUtoo: might not be compatible with linux.inc changes. Dec 25 19:34:18 nschle85, thanks to you too Dec 25 19:34:34 nschle85, oops Dec 25 19:34:48 nschle85, thanks. merry christmaths to you too Dec 25 19:34:59 ilold9, the problem is the savedefconfig format Dec 25 19:37:41 GNUtoo: well, It is a while since then... but that time a .config renamed as defconfig worked quite well for me in OE. Dec 25 19:39:44 savedefconfig is something else Dec 25 19:39:50 it's a compressed form Dec 25 20:05:31 GNUtoo: hello, i would like to update http://shr-project.org/trac/wiki/Devices/NokiaN900/InstallGuide , so i will mark the current information as deprecated and renew the instructions for newcomers installing shr-core having currently standard maemo kernel installed. do you agree ? Dec 25 20:06:13 nschle85, yes if you want, as long as the new info is better than the old one Dec 25 20:06:36 GNUtoo: yes, but i wont loose any information Dec 25 20:07:04 nschle85, yes go on, make a deprecated section Dec 25 20:07:07 and renew the info Dec 25 20:07:14 that page is indeed deprecated, old etc.... Dec 25 20:07:31 GNUtoo: ups is there any newer ? Dec 25 20:07:39 or more relevant ? Dec 25 20:07:42 yes Dec 25 20:07:54 http://shr-project.org/trac/wiki/Devices/NokiaN900/Booting Dec 25 20:10:07 GNUtoo: where is it linked from ? Dec 25 20:10:46 ok found Dec 25 20:11:05 the new page (booting ) has a better structure Dec 25 20:12:04 GNUtoo: i know you are a very good developer... but both pages are confusing newcomers to install shr on n900 Dec 25 20:12:36 yes Dec 25 20:12:59 then rename that page or make it clear that it's a developer resource Dec 25 20:13:20 or do something like info for developers for the parts that are for developers Dec 25 20:14:09 GNUtoo: hmm first ill collect the relevant information, but which page should i change ? boot or install ? Dec 25 20:14:36 I don't know, do the thing you think is the best Dec 25 20:14:49 personally I would think changing Boot Dec 25 20:14:54 moo Dec 25 20:14:59 it has way more infos Dec 25 20:14:59 ill change install :-) Dec 25 20:15:08 murry xmess Dec 25 20:15:19 install is linket from shr top page Dec 25 20:15:25 merry Xorg Dec 25 20:15:39 or Xmass Dec 25 20:15:50 and installing is what newcomers wanto to do Dec 25 20:15:53 nah I stay single, no marriage fo rme ;-D Dec 25 20:16:47 DocScrutinizer: now you are merried with clear case :-) Dec 25 20:16:57 WAAAAH Dec 25 20:17:08 DocScrutinizer: forever :-) Dec 25 20:17:18 clear case is not very clear Dec 25 20:17:47 GNUtoo: the devel is talking in you :-) Dec 25 20:17:49 hey, http://www.netzwelt.de/news/89305-nokia-dual-core-windows-phones-arbeit.html Dec 25 20:18:52 NovaThor ;-D - guess what's my clearcase views are named like. Thorium_1, Thorium_2 etc Dec 25 20:19:41 i will ask GNUtoo: which daily phonei should buy if my N900 dies... Dec 25 20:19:56 do they have ensteinium and stronzium too? Dec 25 20:20:14 nschle85, hmmm Dec 25 20:20:26 nschle85, I think you should wait for the gta04 to be ready Dec 25 20:21:13 GNUtoo: no, the gta04 is no option for me currently Dec 25 20:21:31 nschle85, why? Dec 25 20:21:45 GNUtoo: beveled case Dec 25 20:21:55 GNUtoo: no keyboard Dec 25 20:21:59 beveled??? what does that means? Dec 25 20:22:05 zub, hey, I've been checking/updating elementary.vapi against new Elementary.h during last days (I'm at about line 6k from 30k). It would be nice if you could check that your apps are working ok in case I made some typo or you need to upgrade it :) Dec 25 20:22:12 yes but there is not that much as an option Dec 25 20:22:44 GNUtoo: no samsung phone will be supported ? Dec 25 20:22:51 nschle85, you were at FSOSHRCON and I guess you remember that we currently support the following phones: Dec 25 20:22:53 *nexus S Dec 25 20:22:56 *n900 Dec 25 20:22:59 *gta04 Dec 25 20:23:02 *om-gta03 Dec 25 20:23:04 *om-gta02 Dec 25 20:23:10 s/*om-gta03// Dec 25 20:23:23 and palm-pre which I wouldn't advise Dec 25 20:23:52 merry xmas to everybody btw :) Dec 25 20:24:19 GNUtoo: it was the FSO list Dec 25 20:24:55 so slyon was telling us samsung phones are supportable right ? Dec 25 20:26:07 old Samsung Corby S3650 can boot linux? i can buy one if so Dec 25 20:26:54 GNUtoo: ups it was not slyon it was morphis Dec 25 20:27:02 pespin: thnx; merry xmas to you too :D Dec 25 20:27:12 nschle85, samsung phones have no keyboard either Dec 25 20:27:23 and the port is not finished yet Dec 25 20:27:35 for instance alsa works fine Dec 25 20:27:38 GNUtoo: i know, and this i would accept Dec 25 20:27:41 but there are issues with wifi currently Dec 25 20:27:53 issues means it doesn't work at all Dec 25 20:28:04 but that's the only problematic point Dec 25 20:28:08 the rest is pretty good Dec 25 20:28:09 GNUtoo: so the only alternative is GTA04 at all ? Dec 25 20:28:17 no you have the nexus S too Dec 25 20:28:22 I forgott about it that's all Dec 25 20:28:31 I've many devices, see the people page Dec 25 20:29:02 GNUtoo: what about huwei devices ? Dec 25 20:29:16 GNUtoo: like X3 ? Dec 25 20:29:23 no idea Dec 25 20:29:28 what's the SOC Dec 25 20:29:28 ? Dec 25 20:29:31 JaMa|Off, hi Dec 25 20:29:41 JaMa|Off, do you know how to fix stuff like that locally: Dec 25 20:29:42 | ERROR: Package version for package task-base went backwards which would break package feeds from (0:1.0-r75 to 0:1.0-r73) Dec 25 20:29:57 without diffing with the layers Dec 25 20:30:06 I mean I want to go backward with cleansstate Dec 25 20:30:11 without changing a thing Dec 25 20:30:15 in the metadata Dec 25 20:30:19 just with cleansstate Dec 25 20:51:25 GNUtoo: do you know someboy who can provide combined shr uboot-maemo installation packages ? ? Dec 25 20:52:27 no I don't Dec 25 20:52:43 maybe we should provide it Dec 25 20:53:05 but maemo users usually run different kernels Dec 25 20:53:29 GNUtoo: do you know someting about the laws ? can we provide maemo kernel binaries in any form ? Dec 25 20:54:36 GNUtoo: so providing firmware packages for n900 would also be nice Dec 25 21:07:38 GNUtoo: no einsteinium, but hassium Dec 25 21:07:52 strontium too iirc Dec 25 21:08:03 next chip is radium Dec 25 21:09:04 hassium is actively marketed atm afaik Dec 25 21:09:34 thorium not yet, it seems Dec 25 21:10:16 GNUtoo: how should i partition my sd card to have my own shr boot (opened keyboard) and boot shr staging ? Dec 25 21:10:40 current boot options are sub-optimal Dec 25 21:10:51 nschle85, you need a boot script then Dec 25 21:14:35 run p2 boots from 2dn and takes kerl from 3rd. so why not take both from 2dn ? Dec 25 21:14:44 GNUtoo: that's pretty nonsensical a statement, as the kernel loaded by uBoot is determined by the uBoot maintainer, and usually lives on either uSD or eMMC for N900 Dec 25 21:15:54 nschle85: maemo kernel is FOSS, so yes you of course can pack the maemo kernel with uBoot and provide it Dec 25 21:16:11 ? Dec 25 21:16:20 (about the boot script) Dec 25 21:17:28 GNUtoo: where can i upload a picuture to show my problem ? Dec 25 21:17:43 the boot scheme is as follows: NOLO (2nd stage BL) load-runs whatever is found on "kernel" MTD partition. We place a combined uBoot plus maemokernel there, so NOLO chainloads uBoot. uBoot then either chainloads maemo kernel from same partition, or any kernel on either uSD or eMMC Dec 25 21:18:01 the latter one usually SHR kernel then (or meego, or...) Dec 25 21:21:05 IIRC uBoot looks for some plaintext parameter file on uSD 1st VFAT(?) partition, in /boot/uBoot.menu or sth like that Dec 25 21:21:30 at least the uBoot version we got in "maemo" Dec 25 21:21:41 or meego - whatever Dec 25 21:22:06 I can't see why this uBoot can'T boot SHR incl SHR kernel as well Dec 25 21:22:46 maemo users managed to set up triple-boot with stock-maemo, maemo-powerkernel, and meego Dec 25 21:25:35 DocScrutinizer: current shr-uboot (not maemo-uboot, which is buggy) has a strange boot layout, so if you are interested in i would like to send a screenshot Dec 25 21:25:57 maemo uBoot is buggy? Dec 25 21:25:59 uhuh Dec 25 21:26:18 yes, it cannot boot fron ext3 Dec 25 21:26:31 meh Dec 25 21:27:27 DocScrutinizer: are you drunken ? the package provided byofficial maemo is really buggy Dec 25 21:27:44 that's probably not buggy but an inevitable limitation introduced by restrictions on available size on kernel partition, with maemo kernel living in same partition Dec 25 21:28:21 seems you're drunken, there's no such thing like an "official maemo uBoot image" Dec 25 21:28:23 DocScrutinizer: no its a bug, the limitation does not exist Dec 25 21:28:52 DocScrutinizer: yes there is Dec 25 21:28:59 btw OM uBoot had same "bug" since like ever Dec 25 21:29:13 nschle85: BS Dec 25 21:29:23 DocScrutinizer: BS ? Dec 25 21:29:31 Official would mean Nokia provided uBoot for maemo Dec 25 21:29:38 and they definitely didn't Dec 25 21:29:53 DocScrutinizer: yes its in official repos Dec 25 21:30:05 pfff Dec 25 21:30:27 maybe in meego repos Dec 25 21:30:33 DocScrutinizer: noooo Dec 25 21:30:59 DocScrutinizer: ill show you, please wait Dec 25 21:31:34 http://maemo.org/packages/view/uboot-pr13/ Dec 25 21:31:40 there's just one (set of) uBoot pkg I know of, and all are in maemo-extra Dec 25 21:32:12 and maintained by community, so no way "official" - official repo is Nokia repo, not maemo repo Dec 25 21:32:48 but this uboot is buggy, it cannot boot ext3 Dec 25 21:33:12 in shr uboot is fixed to boot ext3 Dec 25 21:33:38 http://maemo.org/packages/search/?org_maemo_packages_search[1][property]=name&org_maemo_packages_search[1][constraint]=LIKE&org_maemo_packages_search[1][value]=uboot&org_maemo_packages_search[2][property]=title&org_maemo_packages_search[2][constraint]=LIKE&org_maemo_packages_search[2][value]=same Dec 25 21:33:40 but boot layout is a litte bit strange Dec 25 21:34:09 nschle85: so I don't give a shit about your "my uboot - your uboot" Dec 25 21:35:22 if you're not happy with either of both, fix it. preferably the in maemo-extra as it needs to go together with stock maemo kernel into same partition anyway Dec 25 21:35:30 DocScrutinizer: your fucking links dont gave me more informaton :-) Dec 25 21:36:20 DocScrutinizer: uboot is still fixed Dec 25 21:36:56 DocScrutinizer: but configuration for it lacks dual booting Dec 25 21:37:21 whatever you say darling Dec 25 21:37:50 I'm not interested in debates about the topic, as I'm not gaining anything with it Dec 25 21:37:57 DocScrutinizer: hope Clear Case will fuck you deeply :-) Dec 25 21:40:27 DocScrutinizer: thank you :-) Dec 25 21:49:22 so i am sorry, but i was telleing you that current shr uboot configuration does not support siply partitioning. and that that current uboot package available in maemo does not work for us Dec 25 21:51:09 so we have to change current (shr) uboot configuration (maemo packaged one does not wor for us) Dec 25 21:51:44 DocScrutinizer: do you understand me ? Dec 25 21:56:07 my 2 cent. I took a uboot image somewhere from shr, and flashed it with maemo kernel. works for me. can boot shr, meego, nemo and even ubuntu. Dec 25 21:57:05 of course maemo boots too. Dec 25 21:57:58 ilold9: you combined it with maemo kernel ? Dec 25 21:58:59 and booted any other distribution from ext3 ? Dec 25 21:59:07 yes, like it was desribed here, i think: http://elektranox.org/n900/kernel/uboot.html Dec 25 21:59:54 yes ext3 worked, ext4... not sure but there is fat partition support for that. Dec 25 22:00:16 so thats a feature of shr uboot Dec 25 22:00:50 ilold9: standard uboot (maemo) cannot boot from ext3 Dec 25 22:02:12 ilold9: so my problem at the moment is to have dual boot from sdcard Dec 25 22:02:34 yes, that one from maemo didn't boot anything for me. like iI said, I took the uboot provided by shr. has option to boot shr, shr2 and p2 Dec 25 22:03:06 ilold9: yes its without bugs Dec 25 22:04:21 but standard boot sceme is not very friendly for dual boot from sdcard :-) Dec 25 22:06:07 ilold9: so rootfs and kernel has to be on different partitions Dec 25 22:06:47 thats i wanted to explayn DocScrutinizer: but he did not understand Dec 25 22:07:12 I see. yes thats a bit difficult right now, Dec 25 22:08:08 you'd be surprised what I understand and you don't even notice I answered Dec 25 22:08:23 the thing is, I'd like to make one myself, but I read things like "uboot on n900 doesn't boot newer kernels" and stuff, so I keep me hands of for now. Dec 25 22:09:05 ilold9: size limitations Dec 25 22:09:22 kernel gets fatter and fatter Dec 25 22:09:40 space left over in kernel partition for uBoot gets lower and lower Dec 25 22:09:52 ah, I see. Dec 25 22:10:08 that'S why meego uBoot already had to bin support for ext3 Dec 25 22:10:47 dual boot might work with flasher3.5 tool, or is there also a limit if its just loaded and not flashed? Dec 25 22:10:57 which isn't a big issue as you only need a small VFAT first partition to keep your own kernel, and that kernel of course has ext3 support for rootfs then Dec 25 22:11:39 with flasher you can load virtually everything to everywhere Dec 25 22:12:25 btw same setup been used for GTA02 to boot from uSD Dec 25 22:12:45 kernel on small VFAT, then ext3 rootfs partition Dec 25 22:12:50 ilold9: so most N900 users shr uboot works, but for testers it should be configured a little bit Dec 25 22:13:56 back when the problem was an actual "bug" in uBoot though, it didn't support 256byte inode size Dec 25 22:14:29 and when default inode size of mkfs.ext3 changed from 128 to 256, uBoot "magically broke" Dec 25 22:14:37 nschle85: flasher loading the kernel might do what you need, It's not a portable soulution but for testing it should work. Dec 25 22:16:02 DocScrutinizer: thanks for clearing that up for me. Dec 25 22:16:34 so probably with gta02 ubot we could have used kernel-on-ext3 setup if only we'd initialized ext3 with 128 inode size Dec 25 22:17:13 I've always been to busy or too lazy to test if it actually works Dec 25 22:17:19 too* Dec 25 22:19:22 DocScrutinizer: gta02 I don't know, but n900 boots ext3 for me with the shr provided uboot bin. Dec 25 22:19:58 sure, but probably it won't fit into kernel partition together with e.g. powerkernel Dec 25 22:20:21 yes, I do use stock maemo kernel. Dec 25 22:20:30 I'm not even sure it fits with stock maemo kernel Dec 25 22:20:40 ilold9: as i say shr-uboot works and maemo-uboot does not boot ext3 Dec 25 22:21:29 "does not boot ext3" is a rather fuzzy term that actually says nothing Dec 25 22:21:31 I use shr-uboot combined with maemo stock kernel. this boots ext4 Dec 25 22:21:34 ext3* Dec 25 22:21:41 not sure about ext4 Dec 25 22:22:19 DocScrutinizer: ok boot from ext3 partition Dec 25 22:23:06 DocScrutinizer: but i dont know why talking whith you about obsolete stuff Dec 25 22:23:06 the uboot form maemo repositories uboot-pr13 didn't work for me at all. Dec 25 22:23:56 oh, and I can boot ext2 too :) Dec 25 22:24:04 ilold9: may be you can explain DocScrutinizer: why the maemo-uboot does not work Dec 25 22:24:07 it's not uBoot but the kernel that needs support for $random-fs on / built in. UBoot only load a kernel image from a defined location, this location can be a dedicated partition (mtd1 "kernel") or a VFAT partition anywhere large enough to hold kernel bzimage, or ext3 only when uBoot has support for that fs Dec 25 22:25:53 and please stop using the nonsensical term "booting ext" as it'S the kernel that does this, not uBoot Dec 25 22:26:05 DocScrutinizer: and what is this for ? http://www.mail-archive.com/u-boot@lists.denx.de/msg33058.html Dec 25 22:26:08 uBoot is only loading linux kernel, it's not "booting" Dec 25 22:27:12 DocScrutinizer: loading from what ? Dec 25 22:27:16 uBoot loads kernel, kernel boots an ext3 rootfs Dec 25 22:27:56 that I understand. sorry for saying wrong things out of limited knowledge. let's say, uboot-pr13 package from maemo repositories does not what I'd expect it to do. :-) Dec 25 22:28:15 if yu're too lame to partiton your uSD accordingly, you better don't blame uBoot for missing features Dec 25 22:29:41 ilold9: uboot-pr13 cannot boot from ext2/3 partitions but the shr one can do Dec 25 22:30:02 uboot DOES NOT BOOT Dec 25 22:30:12 uboot LOAD A KERNEL Dec 25 22:30:38 DocScrutinizer: ok it cannot load a kernel from ext3 partition Dec 25 22:30:58 DocScrutinizer: but what is the difference ? Dec 25 22:31:00 nschle85: yes. the shr u-boot b... loads the kernel from ext2/ext3. Dec 25 22:31:08 and who gives a fuck if it needs a 4MB VFAT partition to load kernel from? Dec 25 22:31:36 DocScrutinizer: me Dec 25 22:31:44 your problem Dec 25 22:32:24