**** BEGIN LOGGING AT Tue Jun 14 02:59:57 2011 Jun 14 07:07:31 morphis is not there...hmmm Jun 14 09:50:36 anybody could do me a favour about kernel stuff? Jun 14 09:51:32 check bq27000 kernel driver of OM, and how far it propagated upstream, and why it didn't make it further if applicable and answerable Jun 14 09:52:03 larsc: ^^^? Jun 14 09:52:54 everything is upstream Jun 14 09:53:04 oh Jun 14 09:53:39 what's the exact def of "everything"? relating bq27000, or everything OM? Jun 14 09:53:40 or at least something feature equivalent Jun 14 09:53:46 bq27000 Jun 14 09:54:13 aiui the upstream thing *might* be incomplete fsckdup crap Jun 14 09:55:00 reporting weird uncooked internal values that need to get multiplied by resistor value of shunt, and also not providing all the sysnodes OM driver had Jun 14 09:55:11 you probably understand it wrong Jun 14 09:55:16 thats all got fixed Jun 14 09:55:39 mhm, so the upstream driver actually is basically 'ours'? Jun 14 09:56:20 good to know, if that's the case Jun 14 09:57:03 larsc: do you have a html URL to the mainline source for a noob like me? Jun 14 09:57:38 just if you got it at hand Jun 14 09:58:31 http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=tree;f=drivers/power/bq27x00_battery.c Jun 14 09:58:56 thanks a lot! Jun 14 09:59:17 hmm 404 Jun 14 10:00:00 http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=drivers/power/bq27x00_battery.c;h=bb16f5b7e167490519d2793cbaa1546192e447f5;hb=HEAD Jun 14 10:00:08 seems like you need the hash Jun 14 10:01:17 nm, clicking my way down arbitrary branch Jun 14 10:07:33 PaulFertser: \o/ Jun 14 10:07:53 DocScrutinizer: hey man :) Jun 14 10:08:16 DocScrutinizer: i'm back from the trip finally, it was cool :) Jun 14 10:08:23 DocScrutinizer: how's it going? Jun 14 10:08:24 spent all my morning ranting about bq27k driver and bme ;-) Jun 14 10:09:07 Have you managed to convince some poor souls that are still not aware of the "shortcomings" of the way nokia did it? ;) Jun 14 10:09:14 seems the driver (like all sane drivers) occupies the I2C direct access to the chip when modprobed Jun 14 10:09:44 Sounds sane indeed Jun 14 10:10:04 meego is using this driver, but they still use some flavour of bme blob cooked in sulphuric acid Jun 14 10:10:22 * PaulFertser likes Pulster a lot, should certainly buy something from him Jun 14 10:10:33 so once the driver got loaded, bme will fail to access the chip XP Jun 14 10:12:27 meego folks are weird, sometimes it seems like they got no clew whatsoever but they seem to show at least some positive results. Jun 14 10:17:41 they have incredible endurance and faith in their project - that's for sure Jun 14 10:18:25 using bme instead of some KISS approach for battery charging is one of the weirder things they did Jun 14 10:29:55 PaulFertser: "As to the flash-based/using services in general (even as a PoC) i can Jun 14 10:29:55 only echo it once again: it'd be nicer if those devs just went and Jun 14 10:29:55 hugged their girls instead as it would have at least some potential at Jun 14 10:29:55 making this world better, imho.". Now that's an _amazing_ quote, thanks :D Jun 14 10:30:15 wowowow MICKEY Jun 14 10:30:19 hi DocScrutinizer Jun 14 10:30:51 mickey|office: hey! i'm rather glad to see you again :) Jun 14 10:31:46 PaulFertser: :) Jun 14 10:31:57 * DocScrutinizer suggests to freenode to allow 40char nick for mickey|office|bigdaddy Jun 14 10:32:03 hehehe Jun 14 10:32:09 ya, nick length is ridicolously short Jun 14 10:32:18 ~spell ridicolously Jun 14 10:32:20 possible spellings for ridicolously: ridiculously ridiculous meticulously ridicules ridicule's Jun 14 10:32:20 Oh, big things happend while i was absent, congratulations! Jun 14 10:32:25 thanks :) Jun 14 10:46:11 hello everyone Jun 14 10:46:26 does someone has error with bitbake parsing ? http://pastebin.com/cL5FLA0N Jun 14 10:50:59 never seen that one, try removing the cache (rm -rf tmp/cache) and let it reparse again. if it doesn't stop, take a look at the recipe in question, perhaps someone committed a broken line, it looks like it wants to parse the DEPENDS, but has a null object instead of a string Jun 14 10:51:05 bbl, lunch Jun 14 10:51:13 s/stop/work/ Jun 14 11:34:58 larsc: sorry for pestering :-/ Jun 14 11:34:59 mickey|office: I did remove tmp/cache but still the same error and what is weird is that the last recipe commit is 3 month old so ... Jun 14 11:35:07 * DocScrutinizer <---fool Jun 14 11:35:17 1) Upstream merges Jun 14 11:35:19 The bq27x00 and wm8753 branches have been merged upstream. Jun 14 11:40:57 larsc: WTF? our write performance to MMC is better than our read performance? that's for sure unusual Jun 14 11:48:09 what about that eMMC regression and the state of new kerenl now? Jun 14 11:49:04 Von: Martin Jansa Jun 14 11:49:06 Datum: 2011-05-26 21:52 Jun 14 11:49:08 uSD works fine after reverting Per Forlin's MMC patches and glamo-mci modifications to use them. Jun 14 11:50:52 * DocScrutinizer vaguely recalls recent user complaints about uSD not working on GTA02... probably over at #openmoko Jun 14 12:33:29 anyone around with experience running SHR on an HTC Magic phone? Jun 14 12:34:00 (basically an HTC Dream without the keyboard as I understand it) Jun 14 12:36:09 I can't seem to get any further than the carrier logo screen Jun 14 12:47:21 ex-parrot: I guess you should ask GNUtoo Jun 14 12:57:10 hi Jun 14 12:57:52 can I ask to you for gsmd? Jun 14 13:11:07 SystemR89, hi Jun 14 13:11:21 hi Jun 14 13:11:47 I've a teltonika rut-104 Jun 14 13:11:58 SystemR89, how did you format the microsd card? Jun 14 13:12:16 in proprietary firmware they use a modified version of gsmd Jun 14 13:12:52 specs? Jun 14 13:12:57 ex-parrot, hi Jun 14 13:13:10 oops I meant: Jun 14 13:13:16 ex-parrot, how did you format the microsd card? Jun 14 13:13:21 SystemR89: are you sure? Are you talking about the old gtk stack version of gsmd? Jun 14 13:14:20 yeah they use gsmd: # gsmd -h Jun 14 13:14:20 gsmd - (C) 2006-2007 by OpenMoko, Inc. and contributors Jun 14 13:14:20 This program is FREE SOFTWARE under the terms of GNU GPL Jun 14 13:15:43 is it a phone, or is it a router? Jun 14 13:15:57 they developed a vendor plugin for hso interfaces (I think that is their work, right?) Jun 14 13:16:01 router Jun 14 13:16:04 ok Jun 14 13:16:35 but I can't find the sources, this a gpl violation? Jun 14 13:16:41 possible Jun 14 13:16:47 did you receive an offer for source Jun 14 13:16:56 no.. Jun 14 13:17:38 where's the latest gsmd source repo? Jun 14 13:17:54 then I guess it's a gpl violation Jun 14 13:18:16 maybe hackable-1 has gsmd too Jun 14 13:18:41 old gsmd was at svn.openmoko.org Jun 14 13:19:03 we should contact sflc or gpl-violations.org Jun 14 13:19:17 http://svn.openmoko.org/trunk/src/target/gsm/ Jun 14 13:19:33 but I think h:1 forked it Jun 14 13:20:54 SystemR89: anyway, they probably used Openmoko's version from SVN above Jun 14 13:21:47 svn access is restricted? Jun 14 13:22:15 svn co svn://svn.openmoko.org/trunk/src/target;module=gsm doesn't work... Jun 14 13:22:43 dunno, it's old, unused repo Jun 14 13:22:48 maybe try via http Jun 14 13:26:46 ok, i found in sources h1 but hso module doesn't exist Jun 14 13:26:53 https://trac.hackable1.org/trac/browser/trunk/src/core/gsm/src/libgsmd Jun 14 13:27:19 so I think that this is really a gpl violation.. Jun 14 13:30:55 hi GNUtoo: I formatted it in an SD card adapter on my PC, used fdisk to make a new partition table with a "Linux" type partition, then mkfs -t ext3'd it Jun 14 13:32:14 ex-parrot, did you compile the kernel yourself? Jun 14 13:32:26 no GNUtoo, I'm using the latest unstable binary from the SHR website Jun 14 13:32:34 ah that won't work Jun 14 13:32:43 I forgott to add magic support in it Jun 14 13:32:44 aha :) Jun 14 13:32:50 I tried revision 20 and revision 17 Jun 14 13:32:53 with the same results Jun 14 13:34:25 what's needed to add magic support when building the kernel? I'm not averse to building my own but is it a patch, a configure option or what? Jun 14 13:35:35 how I should proceed? Jun 14 13:44:21 ah there we go, got the kernel going Jun 14 13:44:56 would be good to know what's req'd to add magic support while building my own though Jun 14 13:45:52 ex-parrot, configure option Jun 14 13:46:02 GNUtoo: too easy :) Jun 14 13:46:14 iirc it is the "sapphire" configure option? Jun 14 13:46:30 you'll also need some tweaks in the fso config Jun 14 13:46:36 yes I think so Jun 14 13:47:08 which part is fso? apologies if this question is answered elsehwere but I find the sheer number of forum threads all of this info is hidden in to be overhwleming Jun 14 13:50:25 SystemR89, I've some infos for you on the fact that they used at least the linux kernel Jun 14 13:52:18 http://www.pastie.org/private/dguicuhb5hivfyyjj3ykyg Jun 14 13:52:26 note that I hand-arranged the xml Jun 14 13:53:01 gpltool is at: http://www.binaryanalysis.org/trac/changeset/339/trunk?old_path=%2F&format=zip Jun 14 13:53:23 ex-parrot, will look Jun 14 13:53:30 GNUtoo: thanks! no rush :) Jun 14 13:53:48 yes, I know.. I've access to serial terminal Jun 14 13:54:15 I dumped the firmware! Jun 14 13:54:50 ex-parrot, just add saphire for it, do you want me to compile a kernel for you? Jun 14 13:55:04 the same binary in openwrt doesn't work, preloading teltonika libraries I get segmentation fault Jun 14 13:55:07 GNUtoo: no that's fine, I need to set up a build environment here anyway if I'm going to do much hacking Jun 14 13:55:21 SystemR89, I just did that to easier the process of reporting a gpl violation Jun 14 13:55:27 since everyone can: Jun 14 13:55:32 *download that firmware Jun 14 13:55:47 *reproduce it with this binary analysis tool Jun 14 13:56:02 but everyone doesn't have serial and access to the device Jun 14 13:56:51 ok, sorry. I'll try to use this tool now Jun 14 13:56:53 ex-parrot, yes,at least you will need to add freesmartphone configs for your device,else nothing would work Jun 14 13:57:14 GNUtoo: yeah, I figured as much. thanks for your help :) Jun 14 13:57:19 SystemR89, I guess sflc already knows how to use that tool Jun 14 13:57:32 SystemR89, does the firmware has busybox in it? Jun 14 13:57:34 sigh, my SD card is giving me IO errors already. Jun 14 13:58:09 yes Jun 14 13:58:31 are you in europe? Jun 14 13:58:39 do they use netfilter ? Jun 14 14:02:28 Yes, I'm in europe and kernel has netfilter built-in Jun 14 14:02:59 ok Jun 14 14:03:10 then I guess the best way is gpl-violations.org Jun 14 14:04:37 are they responding nowadays? Jun 14 14:05:46 I think after a year of waiting I contacted sflc and they handled the court case in months Jun 14 14:09:58 what was beeing violated? Jun 14 14:10:19 busybox Jun 14 14:10:24 gdb Jun 14 14:10:30 ok Jun 14 14:10:36 which device Jun 14 14:17:12 tahnk you very much! Jun 14 14:17:25 I'll write an email to gpl-violation today Jun 14 14:36:45 mickey|office: ping Jun 14 14:36:47 GNUtoo: ping Jun 14 14:37:08 morphis, pong Jun 14 14:38:06 morphis, I waited the morning but you didn't show up, but you're there now, I hope you have the time to talk right now Jun 14 14:39:16 GNUtoo: sorry, but I was more busy than I expected Jun 14 14:39:29 no problem Jun 14 14:40:16 GNUtoo: I have twenty minutes from now Jun 14 14:40:17 and my connection is not very stable .... damn this shit eduroam wlan .... Jun 14 14:40:20 ok Jun 14 14:40:34 ok, so let's resume what has already been said: Jun 14 14:40:42 ok Jun 14 14:40:47 *palm pre transport layer is high speed serial Jun 14 14:40:58 on top of that there is an msmcommd protocol Jun 14 14:41:03 msmcomm Jun 14 14:41:06 *msmcomm Jun 14 14:41:10 SystemR89: they don't offer sources for this hso module, or for anything? Jun 14 14:41:35 for geeksphone, I wonder: Jun 14 14:42:05 which you have to split in two parts Jun 14 14:42:05 1. Link Layer (mostly the same as found in the bluetooth HCI uart protocol) Jun 14 14:42:05 2. the HCI protocol (not the same like the bluetooth HCI protocol) Jun 14 14:42:24 *what is the transport layer(I guess rpc over shared memory since nothing shows up on the serial port, I added printk on dmesg and nothing shows up with the proprietary ril) Jun 14 14:42:38 ok Jun 14 14:43:02 there is a way of initializing it tough Jun 14 14:43:07 to AT commands Jun 14 14:43:13 it's the shared memory + rpc Jun 14 14:43:17 ok Jun 14 14:43:22 you can do mostly anything with this as base layer Jun 14 14:43:40 and thats why I think Palm does on the Pixi/Veer/Pre 3 devices Jun 14 14:43:52 you can do mostly anything with this as base layer Jun 14 14:43:58 what do you mean? Jun 14 14:44:29 they use shmem+rpc and then speak their HCI protocol as found in the Pre/Pre Plus/Pre 2 devices over this Jun 14 14:44:43 with shmem and rpc Jun 14 14:44:54 you can run different things on top of that Jun 14 14:45:01 depends on which payload you use Jun 14 14:45:15 ah ok Jun 14 14:45:45 so the payload depends on what's implemented in the modem right Jun 14 14:45:45 as the packages you have for gps on the htc dream are not the same as the ones I discovered for the pre Jun 14 14:45:48 some modem have AT Jun 14 14:45:49 yes Jun 14 14:45:50 some don't Jun 14 14:45:52 yes Jun 14 14:45:54 some are incomplete AT Jun 14 14:46:00 and you can even use AT with shmem+rp Jun 14 14:46:09 rp? Jun 14 14:46:13 rpc Jun 14 14:46:13 rpc? Jun 14 14:46:15 ok Jun 14 14:46:15 yes Jun 14 14:46:41 but the geeksphone appear to be able to talk msmcomm Jun 14 14:47:11 basically the problem we have now is that we don't know how to initialize it to AT Jun 14 14:47:27 but running the proprietary ril initialize it to AT Jun 14 14:47:58 ok, there are differences Jun 14 14:48:04 also the Palm Pre can speak At Jun 14 14:48:10 but a incomplete variant Jun 14 14:48:15 yes I know Jun 14 14:48:19 CRING doesn't work Jun 14 14:48:23 *+CRING Jun 14 14:48:23 and other things Jun 14 14:48:26 ok Jun 14 14:48:33 and the geeksphone can speak full AT with CRING included? Jun 14 14:48:39 no idea Jun 14 14:48:52 I think there are issues with CRING too if I remember well Jun 14 14:48:54 what is it using mainly for communication, the properitary protocol or AT? Jun 14 14:48:58 o Jun 14 14:48:59 k Jun 14 14:49:08 proprietary protocol Jun 14 14:49:39 (in the proprietary ril) Jun 14 14:50:19 ok Jun 14 14:50:26 do we have dumps of that? Jun 14 14:50:37 no Jun 14 14:50:42 how could I dump? Jun 14 14:51:00 since it uses rpc it's a bit hard to dump for me Jun 14 14:51:06 it's not simple printk in the kernel Jun 14 14:51:21 ok Jun 14 14:51:26 do have the kernel sources? Jun 14 14:51:30 and do have the device? Jun 14 14:51:34 yes Jun 14 14:51:35 no Jun 14 14:51:39 ok Jun 14 14:51:43 leviathan, hi Jun 14 14:51:54 leviathan, do you still have the geeksphone one? Jun 14 14:52:01 yes Jun 14 14:52:03 and hi Jun 14 14:52:04 :-) Jun 14 14:52:24 leviathan, maybe I should get your geeksphone one since you don't plan to use it or do any work on it Jun 14 14:52:27 leviathan has the device? Jun 14 14:52:32 yes Jun 14 14:52:47 and VQuickSilver in #replicant Jun 14 14:52:48 GNUtoo: give me the address Jun 14 14:52:53 then I'll send it to you Jun 14 14:52:54 :-D Jun 14 14:52:58 :) Jun 14 14:53:05 GNUtoo: where are the kernel sources? Jun 14 14:53:08 ok I'll give it to you in PM when I finish talking with morphis Jun 14 14:53:16 at github, let me find them Jun 14 14:53:37 https://github.com/rmcc/gp_one_kernel/tree/rmcc-2.6.35 Jun 14 14:53:48 GNUtoo: ok, then I will take a look at the sources and try to find a way to dump the payload Jun 14 14:54:19 also for nexus S what should we do if I get one? Jun 14 14:54:38 or if paulk(the other replicant developer) got one Jun 14 14:54:42 *get Jun 14 14:55:05 the galaxy S modem is a msm6xxx if I remember well Jun 14 14:55:30 yes it is according to http://laforge.gnumonks.org/weblog/2010/08/21/#20100821-playing_with_galaxy_s Jun 14 14:55:53 ok Jun 14 14:56:12 I talked with harald some weeks ago about the galaxy S Jun 14 14:56:17 ok Jun 14 14:56:26 he wanted to send me some dumps but never did that Jun 14 14:56:36 ok Jun 14 14:56:50 so the galaxy S modem is an msm too? Jun 14 14:57:30 I think so Jun 14 14:57:36 ok Jun 14 14:57:49 GNUtoo: btw I created a private repository very I collect most of my research stuff Jun 14 14:57:58 ok Jun 14 14:57:59 I will upload it somewhere and give you access Jun 14 14:58:14 ah? why would I need access? Jun 14 14:58:30 will it stay private? Jun 14 14:58:35 what's inside? Jun 14 14:59:16 yes Jun 14 14:59:23 ok Jun 14 14:59:50 morphis: pong, i'm here, but i need to finish something within the next couple of hours for a meeting tomorrow... Jun 14 15:00:09 mickey|office: no problem I have to leave anyway Jun 14 15:00:12 GNUtoo: I have to leave Jun 14 15:00:15 lets talk tomorrow Jun 14 15:00:18 bye Jun 14 15:00:20 morphis, ok bye Jun 14 15:01:01 antrik, they didn't provide any sources Jun 14 15:01:47 hi leviathan Jun 14 15:02:00 you don't plan to do any work at all on the geeksphone one? Jun 14 15:02:03 right Jun 14 15:02:30 GNUtoo: I'm working on the new OpenMoko now Jun 14 15:02:31 ;-) Jun 14 15:02:38 ok Jun 14 15:02:52 I can send it to you Jun 14 15:02:53 would it cost you a lot to send it to me Jun 14 15:02:59 no Jun 14 15:03:04 give me the address Jun 14 15:03:07 by email Jun 14 15:03:07 because I would have difficulties to pay back since I've no bank account Jun 14 15:03:08 ok Jun 14 15:03:11 or trough PM? Jun 14 15:03:19 doesn't matter Jun 14 15:03:23 I just need the address Jun 14 15:03:25 :-D Jun 14 15:05:23 I pm-ed the address to you Jun 14 17:48:01 does someone has error with bitbake parsing ? http://pastebin.com/cL5FLA0N Jun 14 19:11:37 GarthPS: you still have that problem? Jun 14 19:11:41 let me update and check here Jun 14 19:11:47 which tree/branch are you useing? Jun 14 19:12:22 mickey|daddy: ok thx daddy! congrats abain by the way!! i use the shr's makefile and trying to build shr-u for pre2 Jun 14 19:18:41 can you remember having changed anything or did it just broke by updating to a newer version= Jun 14 19:18:41 ? Jun 14 20:02:20 mickey|daddy: just broke up with a "make update" Jun 14 22:24:20 wiki.openmoko.org certificate expired again :( Jun 15 02:25:37 thanks to infobot I wonder if... Jun 15 02:25:38 DocScrutinizer: de nada Jun 15 02:25:43 yo Jun 15 02:25:50 thought as much **** ENDING LOGGING AT Wed Jun 15 02:59:57 2011