**** BEGIN LOGGING AT Fri Nov 19 02:59:58 2010 Nov 19 06:52:36 morning all Nov 19 06:57:08 morning Nov 19 07:18:42 morning Nov 19 07:25:36 DocScrutinizer: you're still sure you won't come to the hw thing in munich in december? Nov 19 07:26:10 hmm, I'm not sure if I'm sure Nov 19 07:26:20 :)( Nov 19 07:26:29 for now I just know I'm broke Nov 19 07:26:37 ohh Nov 19 07:28:02 DocScrutinizer: that does not sound like a good thing Nov 19 07:28:56 yeah, and somewhat forbids such enterprises like traveling to M for a meeting Nov 19 07:29:44 guess we will have to visit Christkindlmarkt this year then Nov 19 07:30:01 though I really should come, talk to Nikolaus - Paul asked me to have a look at "gta03" Nov 19 07:30:09 04? Nov 19 07:30:16 yeah Nov 19 07:30:24 hmmm Nov 19 07:30:25 weird name Nov 19 07:31:23 it has none of the lases and gyros I once specified for gta04 :-P Nov 19 07:32:44 it's just 80min with train, so not really a big thing Nov 19 08:26:41 seems like my n900 suffers from https://bugs.maemo.org/show_bug.cgi?id=6063 :/ Nov 19 09:09:12 I'm pondering phone numbers in opimd. At present there are five fields which specify phone numbers. Hold on I should go and refresh my knowledge of the docs. back later... Nov 19 09:15:23 Nope can't really find what I'm looking for. I'm sort of unsure about the flexability of phone numbers being stored in PIM but I may not really understand. At present there are 5 fields for phone number, one of which is a fax number. Is this a bit limiting? Nov 19 09:16:53 Arigead: this is just the default set... add as many phone number fields as you wish Nov 19 09:21:51 mrmoku: Ah so the fact that I can't add a new field into the current contacts don't mean that I can't actually add a new field. And if I could add a new field how does the system know that it's a phone number. Say my sister is in Germany and has a German phone which I call Germany? Nov 19 09:23:44 Arigead: you have fields... and fields have types Nov 19 09:23:46 http://git.freesmartphone.org/?p=specs.git;a=blob_plain;f=html/org.freesmartphone.PIM.Types.html;hb=HEAD Nov 19 09:24:10 if you add a new field you specify the type Nov 19 09:24:11 http://git.freesmartphone.org/?p=specs.git;a=blob_plain;f=html/org.freesmartphone.PIM.Fields.html;hb=HEAD#AddField Nov 19 09:24:18 and one type is phone number :-) Nov 19 09:26:07 Arigead: just the UI part in libphone-ui-shr is missing for that Nov 19 09:32:04 mrmoku: Thanks for that. Sorry X crashed and locked my machine. Don't want to bother you but if I add that new field for my sister "Germany" type "phonenumber" Then every contact I add now has the ability to insert a field "Germany" which don't make much sense to me. Nov 19 09:32:31 Fields becomes cluttered with guff Nov 19 09:33:38 Arigead: no, you _can_ add that field to any contact... but not every contact has that field by default, just because you added that sort of field to opimd Nov 19 09:34:50 sure, your list of possible fields might grow big Nov 19 09:35:06 mrmoku: Ahh perfect. Nov 19 09:36:54 mrmoku: my next question would be on phone capabilities for a number. If a contact only has a landline then I should not be able to select "SMS" and that should be greyed out. On the other hand in some countries, or some operators allow SMS to a landline. Nov 19 09:49:17 Arigead: yeah, difficult thing Nov 19 09:49:38 bbl Nov 19 09:52:52 mrmoku: I'm browsing at the SQL level in sqlite3 just strikes me as odd the five labels for phones. I guess we'd all do things differently. I guess I'd say a phone number is just a phone number and it has a label and caps. Nov 19 09:55:03 Aragh.. Just had a full build in shr-unstable going on and eglibc-initial_2.12.bb fails on me. Wonder why it's even trying to remove: "cannot remove `/usr/include/gnu/option-groups.h': Permission denied" Nov 19 10:36:15 mrmoku: where does freesmartphone.h come from in libphone-ui ? Nov 19 10:36:26 libfso-glib? Nov 19 10:38:27 hmm, none of libfso-glib, libframework-glib and freesamrtphone-glib supports dates domain... Nov 19 11:28:43 Heinervdm: freesmartphone.h is from libfso-glib, yes (IIRC) Nov 19 11:49:04 yep Nov 19 11:52:42 morning... Nov 19 12:01:16 SHR: 03Martin.Jansa 07shr-makefile * r1b5f57e1b5e0 10/Makefile: Makefile: add require for shr-autorev.inc in shr-unstable config Nov 19 12:01:27 SHR: 03Martin.Jansa 07shr-makefile * r3bd68e117c5a 10/Makefile: Makefile: change order of updated components, from more often changed to less often Nov 19 12:01:28 SHR: 03Martin.Jansa 07shr-makefile * r9c00ce82cf1b 10/Makefile: Makefile: update local-builds.inc example with srctree and gitpkgv Nov 19 12:03:21 daniele_athome: moin, there are some patches for mokosuite on shr ML Nov 19 12:03:36 JaMa|Wrk, yeah i saw, mokohome right? Nov 19 12:03:38 daniele_athome: can you check those and apply in your repo? Nov 19 12:03:48 yup Nov 19 12:04:18 JaMa|Wrk, but those will get lost with next mokosuite release Nov 19 12:04:21 because of the split... Nov 19 12:04:39 but better if you manage tham then if they rot in OE recipe Nov 19 12:04:41 JaMa|Wrk, you'd better apply it in OE Nov 19 12:05:02 why not apply those in current branch as well as splitted? Nov 19 12:05:07 JaMa|Wrk, yeah but i will apply those in the new release Nov 19 12:05:17 because codebase is different :) Nov 19 12:05:30 it's all new repo and code Nov 19 12:05:45 i mean in a few weeks appbunch will be deprecated Nov 19 12:06:09 please JaMa|Wrk apply to OE :) Nov 19 12:06:12 I've notice it in gitorious but still dont understand why not push those also to old branch? Nov 19 12:06:29 mmm ok i will nvm Nov 19 12:06:51 and I'll just bump SRCREV :) Nov 19 12:08:21 mmm Nov 19 12:08:28 JaMa|Wrk, git am isn't working... Nov 19 12:08:37 should i use just git am on patch, right? Nov 19 12:08:53 mmm Nov 19 12:09:02 patch is also for oe :S Nov 19 12:09:04 ok sec. Nov 19 12:09:06 yes but you have to extract only mokowm change Nov 19 12:09:27 yes Nov 19 12:10:44 mrmoku: ok, but there is no gernerated on in git, so i have to look at my builddir? Nov 19 12:11:40 mrmoku: do you want to switch to freesmartphone-glib or do you want to stay with libfso-glib? Nov 19 12:15:48 JaMa|Wrk, done. 4df87c5066e260ec7a155554e3e55201cd08d206 Nov 19 12:18:26 thx Nov 19 12:18:43 mrmoku: it's because i tried to start shr-calendar, but for this i have to add Dates support to the lib, and i don't wan't to do the work twice :) Nov 19 12:21:17 hi, is there the source of open pilot ( http://scap.linuxtogo.org/files/d8845b19098305f3bf4a25e10ca22979.png ) somewhere ? Nov 19 12:21:32 Heinervdm: yeah, looking at the builddir is the way to go Nov 19 12:24:13 Heinervdm: and I would like to switch to libfso-glib with gdbus :) Nov 19 12:24:40 mrmoku: so i have to wait for gdbus? ;) Nov 19 12:24:52 no, don't think so Nov 19 12:25:04 should make not much of a difference Nov 19 12:27:10 hmm, how can i add a xml file libfso-glib? Nov 19 12:27:20 there arn't any sources... Nov 19 12:32:18 oh, it's already there Nov 19 12:36:19 Heinervdm: will be your shr-t branch based on release-2010.12 tag? Nov 19 12:37:30 JaMa|Wrk: with your shr.conf patch, it's current master + no-drm kernel Nov 19 12:37:49 So i think i will base it on release 2010.12 tag Nov 19 12:39:16 so better to not pull master now and release-2010.12 branch instead to it Nov 19 12:39:48 JaMa|Wrk: i'm not sure Nov 19 12:39:55 ie those libxml changes were a bit broken (but I hope I've fixed it now) Nov 19 12:39:55 moin Nov 19 12:40:11 moni Nov 19 12:40:13 moin :) Nov 19 12:40:13 JaMa|Wrk: as it seems that nobody knows how to do that release ;) Nov 19 12:40:26 you mean shr-t release or oe :) Nov 19 12:40:38 JaMa|Wrk: oe Nov 19 12:40:40 because about both there were more e-mails than actuall work :) Nov 19 12:41:10 perhaps that branch will be abonded tomorow or so :) Nov 19 12:41:51 I think it's just misunderstanding.. Nov 19 12:42:00 current shr-u is in very good shape Nov 19 12:42:03 but branch will be abanded for sure Nov 19 12:42:20 only tag from it will remain later (but that's good starting point for shr-t) Nov 19 12:42:45 for now i will only build testimages for me Nov 19 12:42:53 ok Nov 19 12:43:01 thanks for caring about shr-t Nov 19 12:43:19 i want to have that shutdown bug fixed in FSO :) Nov 19 12:43:59 * JaMa|Wrk is always shutting down from terminal .. but I know it's stupid use case for phone :) Nov 19 12:44:20 but also using phone mostly over ssh is also pretty strange use case :) Nov 19 12:44:36 yes :D Nov 19 12:44:59 * JaMa|Wrk is strange user indeed :) Nov 19 12:45:19 at least without piping the voice over ssh too it's less phone ;) Nov 19 12:45:54 but the problem is not just the shutdown, fsogsmd hangs if you try to disable GSM Nov 19 12:46:11 mmhhh lunch is ready :) Nov 19 12:46:14 bbiab Nov 19 12:46:18 so i think airplane mode will fail too Nov 19 12:46:35 * JaMa|Wrk should try to fix n900 speaker to use voice on it instead of piping over ssh to neo :) Nov 19 12:47:02 btw that hanging shr-wizard is only when you have PIN enabled, right? Nov 19 12:47:32 because I never seen it (and newer had PIN enabled) on neo, but had it everytime on n900 with PIN enalbed sim Nov 19 12:48:10 hmm, n900 should be faster Nov 19 12:49:22 the problem is, that fso is still starting and you can continue only when it has finisched starting Nov 19 12:49:39 so if you wait until the first suspend it works Nov 19 12:49:46 but that needs phoneui to send PIN right? Nov 19 12:49:55 no Nov 19 12:50:12 pin dialog comes after the wizard Nov 19 12:50:52 are you sure (maybe it's because n900 is so fast, but iirc I got it before 2nd step of wizard) Nov 19 12:52:09 hmm, i usally get it after the wizard, so i thought that pin dialog is under the wizard and you can see it only after wizard has finished Nov 19 12:54:59 hmm cannot reboot n900 now to SHR :/ Nov 19 12:55:22 will try when I arrive home Nov 19 13:09:23 Hm, anyone who wants to guide me through the first arm disassembly steps? I feel totally stupid :( Nov 19 13:24:15 PaulFertser, which tool do you use? Nov 19 13:31:11 playya: objdump Nov 19 13:31:22 playya: and i feel totally clueless. Nov 19 13:32:32 playya: http://paste.debian.net/100202/ and 94ec is in the middle of the .plt section, i can't understand a thing. Nov 19 13:39:04 PaulFertser, hi Nov 19 13:39:10 what's the issue? Nov 19 13:39:31 mickey|office, did we fix all the spam or is there some left? Nov 19 13:40:11 * JaMa|Wrk knows probably less then Paul Nov 19 13:40:26 * GNUtoo|laptop already looked and did some small arm assembly Nov 19 13:40:43 the first thing is to get the ISA manual Nov 19 13:41:31 I'll explain step by step from what I remember or what seem right Nov 19 13:41:41 first line: Nov 19 13:42:00 ldr r3, [pc #20] Nov 19 13:42:19 that loads in r3 the content of pc + 20 if I remember well Nov 19 13:42:22 I must check Nov 19 13:42:29 add is obvious Nov 19 13:42:39 GNUtoo|laptop: i can read manuals about what instruction means what. Nov 19 13:42:44 ok Nov 19 13:42:50 note the eq Nov 19 13:42:54 that's a conditions Nov 19 13:42:58 GNUtoo|laptop: i can't start to make sense of this executable. Nov 19 13:43:00 that is set by cmp Nov 19 13:43:02 GNUtoo|laptop: yep, i know. Nov 19 13:43:06 ok Nov 19 13:43:15 ah so if I understnad well Nov 19 13:43:23 you understand the steps and thre instructions Nov 19 13:43:28 but not the higher level Nov 19 13:43:33 not what it's meant to do Nov 19 13:43:40 like what it corresponds in C code? Nov 19 13:46:08 GNUtoo|laptop: i basically start from the .init and then this strange unconditional b to plt (that doesn't correspond to any dynamic symbol) is what happens. And if bxeq happens, that means in turn means exit from the program. Nov 19 13:49:12 PaulFertser, could it be thumb code? Nov 19 13:49:37 GNUtoo|laptop: unlikely, i assume GNU always starts in arm mode. Nov 19 13:49:38 PLT seem to be something Nov 19 13:49:52 try to search it in a search engine Nov 19 13:49:55 or in the ISA Nov 19 13:50:13 GNUtoo|laptop: oh come on. :) Nov 19 13:50:36 ??? Nov 19 13:51:26 do you have the complete program disassembly? Nov 19 13:51:58 PaulFertser: single step it with gdb Nov 19 13:52:30 ah .plt section Nov 19 13:52:31 sorry Nov 19 13:52:36 GNUtoo|laptop: sure, grab from https://paulfertser.is-a-geek.org/files/n900/fiasco-updater/flasher Nov 19 13:52:49 GNUtoo|laptop: armv4tl-softfloat-linux-gnueabi-objdump -xTd flasher Nov 19 13:52:58 PaulFertser, ok Nov 19 13:53:12 I mean, by instruction... nexti Nov 19 13:53:17 zub: well, it's an option sometimes, but i do not have the device handy, and no virtual machine etc and i want to learn to disassemble manually. Nov 19 13:53:20 zub: i know ni Nov 19 13:53:29 ah, ok Nov 19 13:53:37 I thought it's jsut an executable for the FR :) Nov 19 13:56:43 PaulFertser, plt seem to be the table for loading libraries Nov 19 13:57:11 PaulFertser, is strace an option? Nov 19 14:00:17 just a random search result with some GOT and PLT info: http://www.symantec.com/connect/articles/dynamic-linking-linux-and-windows-part-one - maybe it helps Nov 19 14:00:40 there is a red-hat pdf about libs Nov 19 14:01:50 Linkers and Loaders... an interesting book; too bad I have only 3 bytes of memory, so don't remember... :-/ Nov 19 14:04:32 hey wait a sec Nov 19 14:04:50 ah no Nov 19 14:05:35 I know how to google, i can even read elf specs etc. But i'd like to receive a more direct help this time. Nov 19 14:05:46 ok Nov 19 14:05:54 ok, sorry Nov 19 14:06:30 I've not great experience in disassembly Nov 19 14:06:38 s/great/big Nov 19 14:06:49 usually I do it for illegal instructions stuff Nov 19 14:06:58 I ask on irc what's the illegal instruction Nov 19 14:07:16 when it's not in the ISA Nov 19 14:09:11 PaulFertser, maybe try khem in #oe Nov 19 14:13:05 PaulFertser, I'm disassembling a hello world and it seem to do something similar Nov 19 14:14:36 * GNUtoo|laptop repasses a bit arm assembly and be back for helping Nov 19 14:15:45 for instance: Nov 19 14:15:46 8310: e3530000 cmp r3, #0 Nov 19 14:15:55 Just a compare... Nov 19 14:15:55 8314: 012fff1e bxeq lr Nov 19 14:15:55 8318: eaffffe3 b 0x82ac Nov 19 14:16:13 GNUtoo|laptop: looks like it just returns from main if that matches, no idea why do that. Nov 19 14:16:14 with 0x82ac beeing in .plt Nov 19 14:16:27 ok Nov 19 14:16:29 I'll look Nov 19 14:17:11 GNUtoo|laptop: yes, but my plt seem to lack dynamic symbol import for that address. Nov 19 14:18:03 Currently building glibc for arm to try hello world. Nov 19 14:21:17 http://pastebin.com/wrtiHsAf Nov 19 14:21:54 the strange thing with your binary is that it file says stripped Nov 19 14:22:05 but you have some symbols like <_init+0x44> Nov 19 14:22:13 that disappear when I strip hello Nov 19 14:23:21 GNUtoo|laptop: nah, init is a section, it can't be stripped. Nov 19 14:23:36 ok Nov 19 14:23:43 GNUtoo|laptop: hello world is nice in proving i do not know shit about elfs :/ Nov 19 14:24:01 I'd like to learn too Nov 19 14:26:25 PaulFertser, I didn't mean the section but the comments Nov 19 14:27:00 GNUtoo|laptop: objdump knows the section names so it adds comments, at least the version i use here. Nov 19 14:27:06 ah ok Nov 19 14:27:40 btw what are you trying to achieve? Nov 19 14:27:58 GNUtoo|laptop: to understand how exactly the kernel on n900 is stored. Nov 19 14:28:02 ah Nov 19 14:28:08 simply read 0xFFFF ? Nov 19 14:28:09 GNUtoo|laptop: to be able to back it up and flash from a running system. Nov 19 14:28:17 GNUtoo|laptop: if it worked, i'd do. Nov 19 14:28:21 ok Nov 19 14:28:55 0xFFFF can flash pieces Nov 19 14:29:02 like a zImage Nov 19 14:29:11 did you try by other means Nov 19 14:29:16 GNUtoo|laptop: i'd also like to understand how to "refactor" (at least rename labels) during the investigation of a disassembly listing. Nov 19 14:29:19 like flashing a zImage with 0xFFFF Nov 19 14:29:25 and looking into the MTD Nov 19 14:29:30 GNUtoo|laptop: it works nicely to flash a zImage over USB. Nov 19 14:29:37 GNUtoo|laptop: but it doesn't work locally. Nov 19 14:29:50 yes but it contains infos on fiasco format Nov 19 14:30:01 GNUtoo|laptop: also i'd like to learn about disassembling anyway. And that's clearly a more direct way to get a "flasher" replacement. Nov 19 14:30:10 yes Nov 19 14:30:13 GNUtoo|laptop: that's for sure, i was going to use it as a fiasco format reference. Nov 19 14:30:14 I'd like too Nov 19 14:30:51 but my goal of learning is quite different Nov 19 14:31:01 I don't plan to disassemble proprietary stuff for now Nov 19 14:31:09 only learn to see compiler errors etc... Nov 19 14:31:19 to be able to debug better stuff Nov 19 14:37:29 PaulFertser, that could be interesting: http://www.linuxjournal.com/article/1060 Nov 19 14:39:31 GNUtoo|laptop: yep, useful to understand the broader picture about plt/got stuff. Nov 19 14:39:35 Thanks. Nov 19 14:39:40 :) Nov 19 14:45:12 GNUtoo|laptop: i realised i should be looking at the start of the .text section actually. Nov 19 14:45:24 yes Nov 19 14:45:32 that's what I did for hello Nov 19 14:45:43 also note that gcc include some stuff Nov 19 14:45:52 ,ctrl0 or something like that Nov 19 14:46:05 it's used for starting the binary Nov 19 14:47:57 http://www.linuxjournal.com/article/1059 was the previous article btw Nov 19 14:51:50 PaulFertser, ^^^ Nov 19 14:54:01 GNUtoo|laptop: ok, with armv4tl-softfloat-linux-gnueabi-objdump -xdT hello i now clearly see how the hello world works. Nov 19 14:54:10 ok Nov 19 14:54:15 PaulFertser: isn't powerkernel-flasher (or how it's called) flashing it too? Nov 19 14:54:15 how does it work then? Nov 19 14:54:17 GNUtoo|laptop: i mean -xdsT <-- notice the s Nov 19 14:54:23 ok Nov 19 14:54:31 JaMa|Wrk: it uses the binary already installed on the device afaik. Nov 19 14:55:49 and what you meant by "but it doesn't work locally."? Nov 19 14:56:17 JaMa|Wrk: if i compile 0xFFFF on n900, i can't make a backup of my currently installed kernel. Nov 19 14:56:35 JaMa|Wrk: also if i want to flash it not via USB, i can't do it from the device itself if not using binary flasher. Nov 19 14:57:05 ah and you don't have sources for that binary flasher from meamo, right? Nov 19 14:58:01 JaMa|Wrk: exactly Nov 19 14:58:29 ok makes sense :) Nov 19 14:58:42 * JaMa|Wrk needs more coffee Nov 19 15:07:25 Hm, so the question remains: if i understand something, how to sanely make modifications to the listing? Nov 19 15:08:26 GNUtoo|laptop: fyi the first parameter for the libc_start_main is passed in r0 and so in my case main() is at 1145c Nov 19 15:11:00 Now i basically understand how to further disassemble, but i've no clue how to do that efficiently. Nov 19 15:11:13 GNUtoo|laptop: btw, thanks for the encouragment! Nov 19 15:12:27 I know there's "lida" but it seems abandonded. Nov 19 15:24:40 I feel like i really missing something. I wasn't able to find an emacs mode to aid disassembly or any other supported tools. Nov 19 15:24:44 And yes, i know of IDA. Nov 19 15:24:50 fuck, no clue what to do. Nov 19 15:30:03 What's the best way to get the timestamp of the first day of a month and the last day of a month in C? Nov 19 15:39:09 Heinervdm: mktime()? Nov 19 15:40:03 anarsoul_: mktime takes a struct tm, does it work, if i set sec, min and hour to 0 and mday to 1? or will it fail? Nov 19 15:41:28 JaMa|Wrk: I want to add a recipe for the nokia900 kernel at git.fso.org Nov 19 15:41:44 JaMa|Wrk: cp linux_git.bb ? Nov 19 15:41:46 anarsoul_: ok, found it, mktime ignores yday and wday Nov 19 15:49:51 mrmoku: yes, that's reasonable start, or are you asking for recipe name suggestions? Nov 19 15:50:04 JaMa|Wrk: no, for the start : Nov 19 15:50:05 ) Nov 19 15:50:17 name would be linux-nokia900_git.bb I guess Nov 19 15:52:13 yes Nov 19 15:52:54 ok, thanks Nov 19 15:56:27 mrmoku: where does FSO_FRAMEWORK_PIM_ServiceDBusName come from? Nov 19 15:59:33 Heinervdm: moment Nov 19 15:59:43 mrmoku: fsoframework.h Nov 19 16:00:35 Heinervdm: ahh, you mean where it is defined? Nov 19 16:01:02 mrmoku: and now to what package it belongs :) Nov 19 16:01:13 because there dates is missing Nov 19 16:02:42 Heinervdm: take a look at src/freesmartphone-pim.vala in libfso-glib Nov 19 16:03:33 mrmoku: where can i find those .vala files? http://git.freesmartphone.org/?p=libfso-glib.git;a=tree;f=src;h=3198577b2e195ba0ef026cc6268ba6f6b95fa8f6;hb=HEAD is empty Nov 19 16:04:16 Heinervdm: my local build has public interface Date : GLib.Object Nov 19 16:04:46 the vala files get generated from the spec Nov 19 16:05:32 Heinervdm: and I have Date, DateQuery and Dates Nov 19 16:06:10 mrmoku: yes, that's there but the define of FSO_FRAMEWORK_PIM_DatesServiceFace is missing Nov 19 16:06:35 ahh Nov 19 16:08:14 and FSO_FRAMEWORK_PIM_DatesServicePath too Nov 19 16:10:55 I added such thing somehwere sometime in the past for another interface... just don't remember where and how :/ Nov 19 16:12:00 for now i will define it in my source Nov 19 16:14:44 PaulFertser, try radare Nov 19 16:15:27 PaulFertser, http://radare.org/y/ Nov 19 16:16:29 PaulFertser, the person behind 0xffff is also behind radare if I understand well Nov 19 16:16:51 GNUtoo|laptop: this stuff looks rather promising Nov 19 16:17:17 GNUtoo|laptop: big thanks Nov 19 16:17:27 let me know how it goes :) Nov 19 16:17:28 GNUtoo|laptop: i wonder how i managed to not see it while extensive googling. Nov 19 16:17:40 I'm sure it won't go nice :| Nov 19 16:27:00 Heinervdm: found it :) Nov 19 16:27:08 it's in libfsoframework in cornucopia Nov 19 16:27:33 libfsoframework/fsoframework/interfaces.vala Nov 19 16:27:50 though defining it yourself is probably what you want if you don't need libfsoframework otherwise Nov 19 16:28:09 Radare seems to be damn nice and useful, exactly the stuff i wanted to get access to. GNUtoo|laptop, BIG thanks! Nov 19 16:28:23 nice!!! Nov 19 16:28:51 mrmoku: ok Nov 19 16:28:53 thx Nov 19 16:30:26 yw Nov 19 17:02:45 GNUtoo|laptop: kernel does not build :/ Nov 19 17:02:46 | /OE/tmp/work/nokia900-oe-linux-gnueabi/linux-nokia900-2.6.28.10+gitr0+e93534addfc33a2fefbca13cb86a8f787e4b650c-r54/git/drivers/input/keyboard/omap-twl4030keypad.c:334: undefined reference to `keyb_led_set' Nov 19 17:03:56 mrmoku, which one? Nov 19 17:04:17 just change defconfig Nov 19 17:04:19 to make it build Nov 19 17:04:34 grep for keyb_led_set Nov 19 17:04:40 ahh... me stupid, yes defconfig Nov 19 17:04:41 and include the right CONFIG_ Nov 19 17:04:42 http://paste.pocoo.org/show/293611/ Nov 19 17:04:58 it's a bug Nov 19 17:05:00 ledtrig-keyb Nov 19 17:05:09 but you can workarround with defconfig Nov 19 17:05:47 mrmoku, if you succeed we get usb host right? Nov 19 17:06:05 thanks a lot PaulFertser for usb host, because I will need it very soon Nov 19 17:07:17 LEDS_TRIGGER_KEYB Nov 19 17:07:20 GNUtoo|laptop: DocScrutinizer was the real fanatic of it, moving the project forward all the time, and it was he who got me a device to allow me to finally write the right patches. Nov 19 17:07:37 then big thanks to DocScrutinizer too Nov 19 17:07:53 TAsn: waiting for a response on shr-core :P Nov 19 17:33:39 PaulFertser: I wonder why my n900 updated kernel-power to 46 and I don't have it in git though :/ Nov 19 17:33:43 anyway... dinner time Nov 19 17:33:55 GNUtoo|laptop: kernel built btw. Nov 19 17:34:04 ok Nov 19 17:34:14 I did a mistake while trying to install the gnash plugin: Nov 19 17:34:23 mkdir ./usr/lib/mozilla/plugins/ -p Nov 19 17:34:31 rm -rf /usr/lib/ :( Nov 19 17:34:43 I'll reinstall an shr-image Nov 19 17:34:52 it's doing do_rootfs right now Nov 19 17:35:08 btw we must add an /etc/init.d/ondemand Nov 19 17:35:24 else it defaults at 500Mhz Nov 19 17:35:31 which is not optimal in any cases Nov 19 18:13:25 GNUtoo|laptop: ok Nov 19 18:14:45 what's blue in htop for the cpu? Nov 19 18:25:14 GNUtoo|laptop: hello, i have small question relating to GIT: if i clone a repository and commit change "a" and afterwards commit change "b" and the push it into its origin, how many commits are seen in the origin ? One which contains "a" and "b" or two single commits "a" and "b" ? Nov 19 18:25:46 nschle85, hi Nov 19 18:25:50 so you're doing: Nov 19 18:26:05 git add a Nov 19 18:26:07 git commit Nov 19 18:26:09 git add b Nov 19 18:26:10 git commit Nov 19 18:26:12 right? Nov 19 18:26:20 GNUtoo|laptop: yes Nov 19 18:26:23 you will have 2 commits Nov 19 18:26:26 a then b Nov 19 18:26:37 is that what you want? Nov 19 18:26:49 GNUtoo|laptop: yes Nov 19 18:26:52 ok Nov 19 18:26:56 what you push Nov 19 18:27:00 will be what you have locally Nov 19 18:27:06 so if it looks right locally Nov 19 18:27:09 you can push Nov 19 18:28:45 GNUtoo|laptop: but in the origin are the 2 commits seen ? or push will add A and B in one step ? Nov 19 18:30:13 GNUtoo|laptop: in detail: are the experiments in the local repository pushed or the last version ? Nov 19 18:31:01 GNUtoo|laptop: duuh... thought the kernel built... failed in modules :/ Nov 19 18:31:29 | ./include/linux/autoconf.h:556:1: fatal error: /drivers/gpu/pvr/pvrconfig.h: No such file or directory Nov 19 18:32:42 GNUtoo|laptop: you're welcome. I hope I manage to have my name show up in final patch somewhere. For now 99% seem to think MohammadAG has invented h-e-n hostmode Nov 19 18:48:45 nschle85, I'm eating Nov 19 18:48:47 bbl Nov 19 18:48:52 ok Nov 19 18:49:14 it seem that nobody want to take credit for the usb host mode...lol... Nov 19 18:53:18 GNUtoo|laptop: eh? Nov 19 18:55:25 credits for actual patches go to Paul, for kicking off the whole doubt of Nokias statement "it's impossible" and checking what we got on hardware and for reading 50.000p of datasheets to find details and to suggest the concept I'm not shy to claim all credit Nov 19 18:56:47 mrmoku, I watched harry potter 7 all day long :P Nov 19 18:56:55 (just for 3 hrs actually) Nov 19 18:57:04 mrmoku, haven't checked emails yet Nov 19 18:57:05 mmnt Nov 19 18:57:28 TAsn: heh... potter fever in Israel too? :P Nov 19 18:58:06 mohammad did the GUI thingy after I suggested exactly how to do it, and he for sure did lots of building, testing, and packaging Nov 19 18:59:07 mrmoku, no idea, that's weird actually Nov 19 18:59:14 I asked a friend of mine to order us tickets Nov 19 18:59:15 a week ago Nov 19 18:59:21 so we'll see it asap Nov 19 18:59:32 but there were no tickets available Nov 19 18:59:43 but since then I talked to other people Nov 19 18:59:54 who said they ordered tickets yesterday and the day before Nov 19 19:00:02 so I think my friend is just dumb :YP Nov 19 19:00:08 And there's no fever, but maybe I'm wrong. Nov 19 19:00:19 anyhow, replied. Nov 19 19:01:51 GNUtoo|laptop: radare2 is _DAMN_ nice. And btw, i really appreciate you trying to help. Was sorta pissed nobody around me (atm in RL) tried to do that... Nov 19 19:04:52 TAsn: ok, thanks :) Nov 19 19:06:12 mrmoku: phoneuid is really complicated, i tried to add sth for the calendar and failed... Nov 19 19:13:07 mrmoku: o/ how's home improvement? Nov 19 19:13:17 Heinervdm, it's not really complicated Nov 19 19:13:27 Heinervdm, libphone-ui-shr is what you need if you just want to tweak the ui Nov 19 19:13:34 and it's the same as any other application Nov 19 19:13:38 (elm/edje) Nov 19 19:13:58 TAsn: the first libphone-ui-shr part is finisched Nov 19 19:14:07 oh Nov 19 19:14:11 adding the dbus method? Nov 19 19:14:26 TAsn: that's the point where i failed Nov 19 19:14:40 i added some specs Nov 19 19:14:45 compiled them Nov 19 19:14:53 that's too crazy for me as well. Nov 19 19:14:54 :P Nov 19 19:15:02 :D Nov 19 19:15:52 have to go now Nov 19 19:16:10 will continue tomorow Nov 19 19:18:29 DocScrutinizer: successfully changed that damn thing :P Nov 19 19:20:52 DocScrutinizer: and wife is happy again ;) Nov 19 19:21:25 so let's call it a day Nov 19 19:26:42 DocScrutinizer: and then I got to know that today is world day of toilet :) Nov 19 19:27:10 NOOO! that's just too funny :-D Nov 19 19:28:47 PaulFertser, ok :) Nov 19 19:30:35 PaulFertser, do you know a bit the kernel btw? Nov 19 19:31:09 GNUtoo|laptop: hm, what do you mean? Nov 19 19:31:45 PaulFertser, do you know how to port forward a board file + serial? Nov 19 19:32:06 because I've really newbish questions Nov 19 19:32:07 GNUtoo|laptop: basically, yes. Nov 19 19:32:11 ok Nov 19 19:32:15 do you have time now? Nov 19 19:33:09 GNUtoo|laptop: sort of, feel free to ask. Nov 19 19:33:29 ok give me some sec Nov 19 19:35:15 be back in a moment Nov 19 19:35:17 GNUtoo|laptop: after disabling CONFIG_PVR the kernel finally built Nov 19 19:35:55 but the main question was in http://pastebin.com/hVK9dwTk what was really necessary? was .map_io = mx31bug_map_io, necessary ? Nov 19 19:36:19 mrmoku, ok Nov 19 19:37:33 GNUtoo|laptop: if the bootloader maps IO for you, then most probably no. Nov 19 19:37:47 I bet it doesn't Nov 19 19:37:57 as it was uncommented in the 2.6.30 BSP Nov 19 19:38:34 be back soon Nov 19 19:44:56 mrmoku: I do not understand what you are talking about but will shr-unstable build ? Nov 19 19:47:36 back Nov 19 19:55:55 nschle85: I successfully built shr-u overnight from scratc Nov 19 19:55:58 h Nov 19 19:55:59 so yes Nov 19 19:59:39 mrmoku: ok Nov 19 20:02:38 mrmoku: i noticed a new tag and branch release-2010.12 is that shr relevant ? what is that branch for ? Nov 19 20:04:51 nschle85: not directly shr relevant.... that tag tags a good state of OE Nov 19 20:05:05 which could be used to branch of for doing some stable works Nov 19 20:05:26 such a release will probably be tagged around 4 times a year Nov 19 20:05:29 at least that's the plan Nov 19 20:05:54 someday we could use such a tag to branch off for shr-stable :) Nov 19 20:06:01 not this time though Nov 19 20:06:52 TAsn: shit my arch linux is not multilib... pure x86_64 Nov 19 20:07:02 now the nokia flasher does not run anymore :/ Nov 19 20:07:06 haha Nov 19 20:07:09 just install multilib Nov 19 20:07:16 (get the multilib repos and install whatever you need) Nov 19 20:07:20 yeah Nov 19 20:07:28 mrmoku: just use 0xFFFF Nov 19 20:07:34 * mrmoku installed ArchLinux on his laptop too now :) Nov 19 20:07:35 mrmoku: why is such tag seen on Nov 19 20:07:41 PaulFertser: yeah, but 0xFFFF lacks one feature Nov 19 20:07:45 mrmoku: which one? Nov 19 20:07:46 to load a kernel without flashing Nov 19 20:08:13 mrmoku, yayz :P Nov 19 20:08:28 TAsn: guess what... it's fast :D Nov 19 20:08:28 * TAsn is a good missioner Nov 19 20:08:36 yeah, super fast. Nov 19 20:08:54 I installed it on slow computers in the past Nov 19 20:09:03 PaulFertser: we still lack a good solution to dualboot SHR Nov 19 20:09:16 mrmoku: isn't uboot working? Nov 19 20:09:55 hmm Nov 19 20:10:17 what disturbs me with that u-boot solution is that I have to take the SD out to boot maemo (IIUC) Nov 19 20:11:09 mrmoku: i thought with keyboard opened you can choose. Nov 19 20:11:18 really? Nov 19 20:11:45 * mrmoku reads up on meego/uboot Nov 19 20:12:46 PaulFertser: http://wiki.meego.com/ARM/N900/Install/U-Boot_from_scratch Nov 19 20:13:02 at first boot it looks for a bootable vfat partition & image on the SD card. If it finds one, it boots that (MeeGo), if it does not, it boots Maemo as normal Nov 19 20:16:30 PaulFertser: ahh Nov 19 20:16:32 To boot back in to Maemo turn off your phone, then either remove the microSD and then turn on your phone, or turn it on and interrupt the boot during the countdown by pressing a button and then type "run noloboot" Nov 19 20:16:40 so there is a way without pulling the SD Nov 19 20:17:57 mrmoku: yes Nov 19 20:20:07 PaulFertser: still we would have to build our own uboot as the meego one does not quite fit Nov 19 20:20:21 expecting a third vfat partition... Nov 19 20:20:39 hi uboot issues? Nov 19 20:20:47 I know which patch to use for ext2load Nov 19 20:21:34 GNUtoo|laptop: and for 'run noloboot' ? Nov 19 20:21:40 ext2load Nov 19 20:21:53 means beeing able to load the kernel from an ext2 partition Nov 19 20:21:58 instead of a vfat one Nov 19 20:22:00 I know Nov 19 20:22:03 ok Nov 19 20:22:17 are there the sources for the meego uboot somehwere? Nov 19 20:22:19 you also have to modify the args Nov 19 20:22:21 should be... Nov 19 20:22:21 yes Nov 19 20:22:33 http://al.robotfuzz.com/~al/maemo/u-boot/ Nov 19 20:22:37 that's the one I know Nov 19 20:22:44 maybe you are already aware of it Nov 19 20:22:54 for the patches: Nov 19 20:23:06 http://www.mail-archive.com/u-boot@lists.denx.de/msg33058.html Nov 19 20:23:11 and modify the args Nov 19 20:23:29 as they can't be saved/modified after the compilation Nov 19 20:23:48 we should add a recipe for that... or tweak the existing one for nokia900 Nov 19 20:24:45 there are lot's of u-boot recipes... so probably add a u-boot-nokia900 one Nov 19 20:30:02 GNUtoo|laptop: and then we have the 2MB limit Nov 19 20:30:45 2MB for uboot+kenrel aka combined.bin ? Nov 19 20:30:49 -rw-r--r-- 1 bitbake bitbake 1868764 Nov 19 20:01 zImage-2.6.28.10+gitr0+e93534addfc33a2fefbca13cb86a8f787e4b650c-r54-nokia900.bin Nov 19 20:30:52 yup for both Nov 19 20:31:05 ok Nov 19 20:31:10 uboot is small Nov 19 20:31:31 might fit... hmmm Nov 19 20:31:36 if it's necessary we could try again to use a smaller size for uboot Nov 19 20:31:40 I already tried Nov 19 20:32:12 basically keep uboot config as-is but modify the address where the kernel is stored Nov 19 20:32:27 I'll explain better Nov 19 20:32:42 yeah Nov 19 20:32:46 understand Nov 19 20:32:59 (the padding to 0x4000) Nov 19 20:33:03 uboot is suposed to be 0x40000 Nov 19 20:33:06 to know where the kernel starts Nov 19 20:33:06 pad to 0x30000 Nov 19 20:33:09 yes Nov 19 20:33:16 and you gain some ks Nov 19 20:33:43 uboot is smaller than 0x30000 Nov 19 20:33:52 you have to modify uboot source too Nov 19 20:34:02 the part where it fetches the address of stuff Nov 19 20:34:08 (atags,kenrel,etc...) Nov 19 20:35:02 http://al.robotfuzz.com/~al/maemo/u-boot/0004-Nokia-RX-51-aka-N900-support.patch Nov 19 20:35:13 in: Nov 19 20:35:15 +#ifdef CONFIG_CHAINLOADER Nov 19 20:35:24 you must also modify the python script Nov 19 20:35:33 but I'm not sure theses 2 things are enough Nov 19 20:35:53 should 80008000 be modified too? Nov 19 20:36:20 from the mkimage Nov 19 20:38:00 well, will first try if it fits Nov 19 22:01:05 GNUtoo|laptop: -rwxr-xr-x 1 bitbake bitbake 160908 Nov 19 22:50 tmp/deploy/images/nokia900/u-boot-nokia900-git-r0.bin Nov 19 22:01:25 ok Nov 19 22:01:38 wow Nov 19 22:01:45 it works? Nov 19 22:01:48 no idea :P Nov 19 22:01:52 just finished building Nov 19 22:02:02 ok Nov 19 22:03:00 btw. current u-boot has a ext2load patch Nov 19 22:03:07 a bit different... Nov 19 22:03:21 but the other patches don't apply to current u-boot Nov 19 22:08:24 ok Nov 19 22:08:43 mrmoku, I'll go maybe I'll connect with my n900 to irc Nov 19 22:31:18 mrmoku, does it boot? Nov 19 22:33:01 GNUtoo: still don't know yet Nov 19 22:33:27 ok Nov 19 22:33:29 * mrmoku got sidetracked by a crying baby :P Nov 19 22:33:36 ah? Nov 19 22:33:54 my small son Nov 19 22:33:58 ok Nov 19 22:34:41 guess I will see tomorrow if it boots Nov 19 22:34:45 going to bed soon now Nov 19 22:34:47 ok Nov 19 22:35:02 without n900 I guess Nov 19 22:35:10 yeah :P Nov 19 22:35:16 lol Nov 19 22:35:30 * mrmoku is not allowed to play with electronic devices in bed Nov 19 22:35:35 my wife would hate me ;) Nov 19 22:35:41 lol ok Nov 19 22:36:39 btw Nov 19 22:36:43 GNUtoo: where does u-boot know from where to boot? Nov 19 22:36:51 ? Nov 19 22:37:04 I mean... the meego u-boot is looking for a third fat partition Nov 19 22:37:12 i'm on n900 right now Nov 19 22:37:18 ah ok Nov 19 22:37:46 just grep for fatlad in rx51.h Nov 19 22:37:55 *fatload Nov 19 22:38:11 ok, we change that for the first ext2 partition then? Nov 19 22:38:17 yes Nov 19 22:38:20 ok Nov 19 22:39:42 loadbootscript=fatload mmc 0 ${loadaddr} boot.sc Nov 19 22:39:47 that is interesting Nov 19 22:39:50 boot script Nov 19 22:39:52 yes Nov 19 22:40:04 is that a u-boot feature? Nov 19 22:40:12 and contains u-boot commands? Nov 19 22:40:12 I think so Nov 19 22:40:16 yes Nov 19 22:40:18 good :) Nov 19 22:41:25 wifes are build system that create images? Nov 19 22:41:41 such as male or female images? Nov 19 22:42:08 hehe Nov 19 22:42:50 then the image grow while eating packages of food Nov 19 22:43:10 italian image then ;) Nov 19 22:43:15 lol Nov 19 22:44:13 maybe I'm a bloated image... Nov 19 22:44:30 I need a dietlibc Nov 19 22:44:52 (dietlibc really exists) Nov 19 22:45:31 :) Nov 19 22:47:57 anyway for uboot Nov 19 22:48:28 change the load address only when you're sure that the rest work Nov 19 22:49:00 yeah, will only change the fatload to ext2load Nov 19 22:49:10 ok Nov 19 22:49:16 and keep the fatload for boot.sc from first partition Nov 19 22:49:22 ok Nov 19 22:49:41 that way one can satisfy maemo with a first vat partition and put a boot.sc there to tell it where to boot the good stuff from :) Nov 19 22:50:14 or have an ext2 with uImage as first partition Nov 19 22:51:13 yes ext2 first partition with meego kernel Nov 19 22:51:49 then change to maemo with the atags passing...etc... Nov 19 22:52:09 can u-boot do menus? Nov 19 22:52:52 like in gta02 Nov 19 22:52:59 I guess so Nov 19 22:53:54 or do that Nov 19 22:53:58 fatload Nov 19 22:54:02 if fail Nov 19 22:54:11 then ext2load Nov 19 22:54:22 both on first partition Nov 19 22:54:41 first with boot.sc and then with uImage Nov 19 22:54:44 yeah, sounds good Nov 19 22:55:27 anyway... will get me some sleep before Nov 19 22:55:31 I wish you a good night :-) Nov 19 22:55:43 ok Nov 19 22:55:48 thanks Nov 19 22:56:14 I'm reading a funny article about const and volatile Nov 19 22:56:51 there is even an image of a volatile... Nov 19 22:57:07 a chicken **** ENDING LOGGING AT Sat Nov 20 02:59:57 2010