**** BEGIN LOGGING AT Sat Feb 21 02:59:57 2009 Feb 21 03:22:54 hehe Feb 21 03:23:31 SuperDave Feb 21 03:28:22 davey Feb 21 03:37:10 http://www.daveyandgoliath.org/snowboard/index.html Feb 21 03:51:27 :D Feb 21 04:02:07 http://en.flossmanuals.net/bin/view/NeoFreeRunner/WebHome Feb 21 04:03:01 fso milestone 5.5 Feb 21 04:03:11 where is fso 5.5? Feb 21 04:10:12 5.5? Feb 21 04:15:39 http://wiki.openmoko.org/wiki/Community_Updates/February_20%2C_2009 Feb 21 04:16:19 FSO meets paroli-for a fully functional paroli, FSO team came up with fso milestone 5.5. FSO MS 5.5 will contains everything needed by paroli. Feb 21 04:16:44 so its not ready yet i guess Feb 21 04:57:56 ok I'm back Feb 21 04:58:01 boot_usb built Feb 21 05:01:19 ./boot_usb read 0x60800 2048 kct Feb 21 05:01:39 got something Feb 21 05:03:29 oops wrong channel :-) Feb 21 06:47:08 /exec ifconfig Feb 21 06:52:40 :P Feb 21 06:54:13 I have tested koolus beta 3 and OM testing Feb 21 06:54:30 the major issue for me is the "never resume" with WSOD Feb 21 06:54:48 WSOD, later, LCD led turn on and nothing appear Feb 21 06:55:02 tks Feb 21 06:55:32 lewis: have you tried any of the images on http://panicking.kicks-ass.org/download/ Feb 21 06:55:52 they are working on the android on freerunner, and seem to have quite a bit of success Feb 21 06:56:28 people report all sorts of cool things on their mailing list, too bad I haven't had time to check it yet Feb 21 06:58:37 lewis: here's the forum that has most of the info about their things: http://forum.koolu.org/viewforum.php?f=10 Feb 21 06:59:41 yep, I know, thank you. but i dont like web forum format Feb 21 07:00:22 ok, thanks for your efforts Feb 21 07:00:33 just to report, maybe Feb 21 07:00:41 :) Feb 21 07:01:44 I will try those paniicking.kiick-ass image Feb 21 07:02:08 but i think that is not only an Android issue Feb 21 07:02:46 it feels like it is very interconnected Feb 21 07:03:11 like Qi issues that don't show up in OM images cause trouble in android, and sometimes vice-versa Feb 21 07:03:18 maybe some firmware... I would like to buy a debug board Feb 21 07:03:45 i mentioned it to you, because they do successful suspense - resume, and sort... Feb 21 07:03:57 yeah, debug board would help :) Feb 21 07:04:06 :) Feb 21 07:04:11 depending on how deep one wants to dive in... Feb 21 07:05:25 yep Feb 21 07:07:17 sometimes WSOD bug disappear, but it looks like recurrent... I hope that it is a different cause each time that it appear Feb 21 07:43:03 Hey :) can at least two people with GTA02 and at least one with GTA01 provide me with their "hciconfig hci0 revision |grep SCO\ mapping" (with bluetooth enabled). Feb 21 07:52:53 Anyone tried netsurf yet? Feb 21 07:55:01 Defiant-: can you please help me? Feb 21 07:55:09 PaulFertser_: still booting... Feb 21 07:55:28 Ok, i didn't try netsurf btw. Feb 21 07:56:34 PaulFertser_: SCO mapping: PCM Feb 21 07:56:37 GTA02 Feb 21 07:57:15 Defiant-: cool thanks. I think that's the same for everybody. Now i wonder about GTA01. Feb 21 08:05:54 hello guys Feb 21 08:06:47 anyone thought of using real time os ? Feb 21 08:07:53 I wonder if anybody has ever routed SCO (not A2DP) over hci usb on FreeRunner. Feb 21 08:15:59 ok, netsurf doesn't feel bad Feb 21 08:20:15 PaulFertser_: root@om-gta02 ~ $ hciconfig hci0 revision |grep SCO\ mapping Feb 21 08:20:15 SCO mapping: PCM Feb 21 08:20:15 root@om-gta02 ~ $ Feb 21 08:20:28 billk: thanks :) Feb 21 08:27:18 morning Feb 21 08:49:02 lpotter: ping Feb 21 10:11:14 I am really concerned, I have a FR running qi-s3c2442, but each time I try install Om2008.12 I get an error about AS0c drviers only working on a GTA01 device????? Feb 21 10:11:36 normal afaict Feb 21 10:12:19 ? Feb 21 10:13:13 does it stop booting? or does it continue and boot normal? Feb 21 10:13:15 it's ASoC, nor? "alsa system on chip" Feb 21 10:14:42 it stops booting, I am really interestied in trying Om2008.12 FDOM is too bloated for normal use Feb 21 10:15:24 jrodger: OM2008.12 has no future, do you know about that? Feb 21 10:16:07 jrodger: PaulFertser_ is right rather try an om unstable or FSO MS5 Feb 21 10:16:18 jrodger: or SHR Feb 21 10:16:29 PaulFertser_: hehe :) true or that one Feb 21 10:17:15 mirko-paroli: btw, it'd be cool if you helped Debian folks to package Paroli. ;) Feb 21 10:17:42 PaulFertser_: I would love to and also to get the remaining e packages just not sure if I can contribute anything, got a name? Feb 21 10:18:52 PaulFertser_: or a channel? Feb 21 10:18:54 I tried SHR, I liked it but was concerned about the lack of sound..... Feb 21 10:19:46 mirko-paroli: i think you can contact nomeata here or write a mail to pkg-fso mailing list and if you mention all the components needed to run paroli i hope you'll get an answer. I'm not using E yet, so can't say much more about it. I think you can find them also at #openmoko-debian Feb 21 10:20:08 jrodger: lack of sound? It works for everybody with SHR. Get shr unstable, get corresponding kernel. Feb 21 10:20:18 PaulFertser_: cool thx will try :) Feb 21 10:20:44 mirko-paroli: it will really widen the target audience because not everybody likes running OE-based distro, you know :) Feb 21 10:21:14 SHR rocks. mostly. Feb 21 10:21:17 PaulFertser_: yepp i do know :) didn't give it a shot yet (debian that is) because of the lack of packages Feb 21 10:21:25 debian is not bad too Feb 21 10:21:43 in some sorts more usable actually Feb 21 10:22:04 i have debian running most of the time, actually Feb 21 10:22:23 for IRC over SSH over GPRS and some GPS Feb 21 10:22:34 how do I identify the unstable release? Feb 21 10:24:06 jrodger: it's called unstable :) Feb 21 10:29:10 PaulFertser_:I could not see the bin file in Unstable......am I blind? Feb 21 10:30:02 jrodger: http://build.shr-project.org/shr-unstable/images/ Feb 21 10:30:50 PaulFertser_: just using the "latest" bin, correct? Feb 21 10:31:07 jrodger: ah, you're talking about the kernel. Feb 21 10:31:34 yes Feb 21 10:31:35 jrodger: then yes, there's only one there anyway :) "latest" is most probably a symlink. Feb 21 10:32:27 hrm. Feb 21 10:32:58 is there any way to start qi without flashing it to the u-boot partition? Feb 21 10:33:13 (without using debugboard) Feb 21 10:33:59 and: does the u-boot in NOR use the u-boot_env partition? Feb 21 10:34:10 Wonka: no Feb 21 10:34:19 Wonka: otherwise you could brick it? Feb 21 10:34:54 lindi-: ok, yes. but what config does it use? Feb 21 10:35:50 Wonka: one in the NOR Feb 21 10:35:59 lindi-: i think you can start qi after downloading it to RAM via dfu. Feb 21 10:36:13 Wonka: but i don't see a reason why that can be needed. Feb 21 10:36:19 lindi-: any way to read that config Feb 21 10:36:29 Wonka: sure Feb 21 10:36:44 PaulFertser_: i'd like to play with qi and a small kernel designed to choose another kernel to kexec Feb 21 10:36:55 $ grep 0000 *wiki|grep -i nor Feb 21 10:36:56 Neo_FreeRunner_Memory_Mapping.wiki:nCS0: 00000000 07FFFFFF 4K steppingstone or NOR (Aux held down)
Feb 21 10:36:59 Neo_FreeRunner_Memory_Mapping.wiki:nCS3: 18000000 1FFFFFFF NOR
Feb 21 10:37:19 Wonka: u-boot in NOR always uses the same config and some steps on boot. I think i can find it in the sources if you really need it. Feb 21 10:37:41 Wonka: but flashing Qi to your u-boot partition doesn't affect u-boot environment anyway. Feb 21 10:38:00 PaulFertser_: i just want to be sure i can somehow boot my debian even if i damaged my NAND bootloader Feb 21 10:38:05 Wonka: so at least in kernel land you can ioremap(0x18000000, 0x1FFFFFFF) and read it Feb 21 10:38:11 Wonka: but you can always flash u-boot back Feb 21 10:38:16 anyway, i could just reflash u-boot Feb 21 10:38:17 yes Feb 21 10:38:19 Wonka: not sure if /dev/mem suffices Feb 21 10:38:22 Wonka: using NOR that you can't damage Feb 21 10:38:23 via nor u-boot Feb 21 10:38:42 lindi-: one can read nor u-boot via /dev/mtd0 Feb 21 10:39:05 ah, yes. but not write. Feb 21 10:39:07 PaulFertser_: really? Feb 21 10:40:00 hmm, by holding aux on bootup? Feb 21 10:40:14 lindi-: no, just anytime Feb 21 10:40:22 hexdump -C /dev/mtdblock0 | less, /boot Feb 21 10:40:24 works Feb 21 10:40:35 finds stuff looking like u-boot internals Feb 21 10:41:03 Wonka: but that can be the nand uboot Feb 21 10:41:33 lindi-: no Feb 21 10:41:46 to which one? ;) Feb 21 10:41:57 was there a GTA02v6 in July 2008? Feb 21 10:42:01 lindi-: that can't be nand u-boot, as it's read from NOR :) Feb 21 10:42:10 PaulFertser_: really? Feb 21 10:42:40 there's "Defaulting to GTA02v6" in mtdblock0... Feb 21 10:42:44 PaulFertser_: but i thought you could upgrade nand uboot by writing to mtdblock0 Feb 21 10:43:03 Wonka: what's the version number? Feb 21 10:43:21 lindi-: of what? Feb 21 10:43:27 lindi-: to mtdblock1 Feb 21 10:43:28 I do sudo strings /dev/mtdblock0 | grep ^U-Boot | head -n1 Feb 21 10:43:34 to get my u-boot version number Feb 21 10:43:47 it says 'U-Boot 1.3.2-moko12 (May 9 2008 - 10:28:48)' Feb 21 10:43:51 U-Boot 1.3.2-moko12 (May 9 2008 - 10:28:48) Feb 21 10:44:15 mtdblock1: Feb 21 10:44:16 U-Boot 1.3.2-moko12 (Sep 16 2008 - 09:53:49) Feb 21 10:44:19 different. Feb 21 10:44:36 so, mtdblock0 clearly is NOR, and not NAND. Feb 21 10:45:38 i want to have some qi that does _nothing_ than loading and booting a kernel image from nand flash. Feb 21 10:46:00 (and the stuff absolutely necessary before that) Feb 21 10:46:26 Wonka: what do you read from mtdblock1? Feb 21 10:46:34 lindi-: see above Feb 21 10:47:14 Wonka: but you read from mtdblock0? Feb 21 10:47:16 ah Feb 21 10:47:49 that kernel should take care of initializing the rest, search for bootable configs, maybe have the user choose one, and kexec the appropriate kernel Feb 21 10:47:50 Wonka: i think you'll need to modify Qi for that because it will try to boot from uSD before NAND. Feb 21 10:48:02 PaulFertser_: then i will Feb 21 10:48:13 PaulFertser_: qi should not even try to access uSD, IMO. Feb 21 10:48:26 Wonka: why don't you like that it tries to access uSD? Feb 21 10:48:45 PaulFertser_: "linux can do it better". Feb 21 10:48:59 Wonka: better access uSD? Feb 21 10:49:04 PaulFertser_: boot a small linux to display a boot menu and kexec the kernel Feb 21 10:49:20 PaulFertser_: keep qi simple and stupid Feb 21 10:49:30 Wonka: i understand the idea and like it. But there's one thing you need to know. Feb 21 10:49:32 PaulFertser_: keep complexity that linux needs anyway only in linux Feb 21 10:49:46 Wonka: GTA03 will have only uSD boot. Feb 21 10:49:58 PaulFertser_: aaargh. the stupidity. Feb 21 10:50:13 PaulFertser_: how much flash does the GTA03 SOC have? Feb 21 10:50:15 Wonka: not the stupidity. Maintainability and predictability. Feb 21 10:50:27 Wonka: do not know yet. Feb 21 10:50:48 Wonka: it's quite easy to brick Qi if it only accesses NAND, right? Feb 21 10:50:52 PaulFertser_: if it's 4MB, its enough for qi and that stepstone kernel Feb 21 10:51:07 lindi-: that's what we have NOR for, nor? Feb 21 10:51:19 Wonka: so you want it to only access NOR? Feb 21 10:51:20 Wonka: the thing is: they are going to even start Qi from uSD. SoC will fetch Qi from uSD automatically afaik. Feb 21 10:51:31 Wonka: or you want that u-boot will be shipped too? Feb 21 10:51:33 lindi-: via nor you can repair qi and its kernel Feb 21 10:51:51 lindi-: not necessarily u-boot but an dfu-thingie Feb 21 10:52:15 Wonka: but with GTA03 you don't even need a NOR. Just plug uSD in any cardreader and do what you want. Feb 21 10:52:22 PaulFertser_: whee. will it have one or two usd slots? Feb 21 10:52:27 Wonka: no :( Feb 21 10:53:09 PaulFertser_: meh. separation of system and data impossible? Feb 21 10:53:26 ok, now i have debian on usd too, next to my data... Feb 21 10:53:34 Wonka: why? Partition your uSD however you like it. Feb 21 10:54:06 how much trouble putting qi the right way on a uSD? Feb 21 10:54:19 will, e.g., windows damage it? Feb 21 10:54:32 (end users...) Feb 21 10:54:58 Wonka: i don't think windows will damage it, though i don't care Feb 21 10:55:39 i soooo like this concept... undamageable NOR dfu update util, boot loader and small but usable system for gsm, gps, gprs, ssh in nand, debian on uSD... Feb 21 10:56:11 PaulFertser_: I for me don't care either. but End Users[tm] should not easily damage it, or they will not want to use it Feb 21 10:56:35 a phone that does not work after you removed the SD card is... somehow stupid, IMO. Feb 21 10:56:36 Wonka: i'm sick of end users. I'd better pay more for hardware. Feb 21 10:57:00 the more End Users are interested in the same hardware as you, the less you pay Feb 21 10:57:24 and i am a) interested in paying less, and b) interested in having more people pay less for better stuff Feb 21 10:57:41 Wonka: no, NAND is expensive and SD not. Think of it. I've recently removed internal GPS antennae and forgot to plug it back. Wondered why gps is not working :) Who's at fault? ;) Feb 21 10:58:22 _that_ is obviously your fault Feb 21 10:59:06 i want an uSD slot for system, and a SDIO capable externally accessible SDHC slot for data. Feb 21 10:59:14 can't be that hard, can it? Feb 21 10:59:20 Wonka: external usb sd card reader? ;) Feb 21 10:59:44 lindi-: can't put that in my pocket together Feb 21 10:59:56 lindi-: don't want any stuff clinging to my mobile Feb 21 11:00:09 Wonka: i think it's hard because S3C has only one SD interface probably. Feb 21 11:00:25 i'd be totally happy with only one uSD Feb 21 11:00:47 but i'd like to have some extra AVR chip inside the case :) Feb 21 11:01:01 Wonka: having usb host provides you with much more than SDIO Feb 21 11:01:03 and a second usb port for charging would be cool too Feb 21 11:01:05 lindi-: i'd be happy with only one uSD, and happier with an additional SD slot Feb 21 11:01:29 PaulFertser_: i don't necessarily need SDIO, but it opens possibilities. Feb 21 11:01:32 Looks like GTA03 is more consumer-oriented therefore we won't be happy with it anyway Feb 21 11:01:40 data SD access via USB would be fine too Feb 21 11:02:09 and having some bus to connect stuff... Feb 21 11:02:35 at least one free internal usb port ... Feb 21 11:02:39 s2c Feb 21 11:02:44 er. Feb 21 11:02:46 i2c Feb 21 11:03:32 Wonka: leaving empty space inside the case if of course not something that usually is wanted Feb 21 11:03:55 alternate back plate? Feb 21 11:04:10 maybe Feb 21 11:04:13 and in gta02 there seems to be enough space for an internal debug board... Feb 21 11:04:43 you need to be able to attach stuff somewhere first... Feb 21 11:05:19 hm, watertight moko with enough air in it to swim Feb 21 11:07:49 but still, if qi just booted a small bootloader linux, then that one could do so much stuff so much better Feb 21 11:08:13 you could even edit kernel commandlines via on-screen keyboard Feb 21 11:08:32 Wonka: to my understanding GTA03 will look for bootloader at the end of SD card, that should be left unpartitioned. So i see no problems with any os here. Feb 21 11:08:40 (end should be unpartitioned) Feb 21 11:08:50 mhmh Feb 21 11:09:41 well, most of my genious plan is still possible :) Feb 21 11:10:21 hrm, how big has grub2 gotten, by the way? Feb 21 11:10:22 windows users will likely damage that... doesn't qi fit in the mbr? Feb 21 11:10:50 Wonka: yes, but i see no reason to trim Qi to disallow reading from uSD. Moreover you can use it for debugging your initramfs by placing the kernel with initramfs to uSD. Feb 21 11:10:51 windows users will possibe damage anything they can write to Feb 21 11:11:18 henk: i've just told you where the Qi is written according to 6410-partition-sd.sh Feb 21 11:11:31 PaulFertser_: there is no reason to disallow qi reading from usd if gta03 will require that Feb 21 11:11:44 Wonka: :) Feb 21 11:12:09 PaulFertser_: i just think the NOR / NAND / uSD stuff was a very good and very stable idea Feb 21 11:12:29 Wonka: probably yes... I won't argue on that one. Feb 21 11:12:38 PaulFertser_: don't start again, will you? ;) Feb 21 11:12:56 but then, if SoC directly reads bootloader from end of uSD, that's also quite unbreakable Feb 21 11:13:04 henk: why not? :p Feb 21 11:13:08 and if uSD is defective, exchange it. Feb 21 11:13:20 henk: but Qi is not placed in MBR according to the script Feb 21 11:13:35 Wonka: it's unbreakable? how abot someone repartitioning the usd? Feb 21 11:13:36 Wonka: until SoC sucks at reading shitty uSDs Feb 21 11:14:00 henk: repartition it again and that's all :) You don't break hardware by partitioning. Feb 21 11:14:12 PaulFertser_: i know, you said that you said so ;) i just wonder why it's not in the mbr... Feb 21 11:14:18 henk: not permanently breakable. you will always get it back into a working state without special hardware tools. Feb 21 11:14:27 henk: without JTAG and stuff Feb 21 11:14:35 yeah thats true Feb 21 11:14:47 you don't even need DFU Feb 21 11:14:49 henk: MBR is 512 bytes only, you know Feb 21 11:14:50 not bricked but broken Feb 21 11:15:06 PaulFertser_: yes. but grub fits doesn't it? Feb 21 11:15:06 ah, ok. that's the word i missed using :) Feb 21 11:15:14 henk: haha. stage1 fits. Feb 21 11:15:29 henk: then loads stage1.5 from sectors 1 to n Feb 21 11:15:36 henk: then loads stage2 from fs Feb 21 11:16:22 MBR is not the safest place anyway, too many programs like to touch it unlike the end of uSD :) Feb 21 11:16:45 (or does stage1 load sectorlist from sector written into stage1, and then stage1.5 from sectorlist, and then stage2 from fs? something like that...) Feb 21 11:16:47 Wonka: ah, right... Feb 21 11:17:37 PaulFertser_: but repartitioning the usd would include the end of usd in a partition, which breaks the bootloader. apart from that i'd say that keeping stuff in unpartitioned space is broken by design... Feb 21 11:18:13 henk: i can't imagine a better scheme though. Feb 21 11:18:53 henk: the idea is that you can easily fix the boot without hardware and software. Without NOR in every device, etc. Feb 21 11:18:55 PaulFertser_: use primary partition #4 instead of unpartitioned space? Feb 21 11:19:15 henk: what if a user repartition that too? ;) Feb 21 11:19:33 PaulFertser_: bad user. Feb 21 11:19:44 include the space into any partition Feb 21 11:19:48 at least Feb 21 11:19:59 should be doable independently Feb 21 11:20:15 iirc windows shifted around partitions on me once Feb 21 11:20:28 so that primary 3 suddenly was primary 2 and 2 was 3 Feb 21 11:20:31 PaulFertser_: if he doesn't know what it's for he shouldn't delete it. but if he doesn't know something's there at all, he is likely to just maximize his partition. Feb 21 11:21:23 henk: ok, you can partition it like that if you like, SoC won't care anyway ;) Feb 21 11:21:43 PaulFertser_: ah good. Feb 21 11:23:20 SoC reading directly from uSD sounds overly complex to me anyway Feb 21 11:24:01 Wonka: i don't trust it's ROM bootloader because it can just randomly not like some cards. But anyway... Ask Andy ;) Feb 21 11:24:15 i'd accept having some (updatable, even if over jtag) code in some flash to read bootloader more easily Feb 21 11:24:33 same reasons Feb 21 11:24:42 and maybe some more, don't know Feb 21 12:02:23 i just installed SHR but i've got no keyboard, in settings i just get none and default so i select default but nothing shows up Feb 21 12:53:15 With the contacts on SHR - how do you new contacts? - I cant move Feb 21 12:53:37 to the second window to enter the actual number! Feb 21 13:23:13 Hi, what is now the best distribution for just calling and sms? The official 2008.12? Feb 21 13:24:03 probably qtopia Feb 21 13:25:41 if I have a dist on my sd-card and using Qi, there is no way to switch to the installed dist on the phones memory? Feb 21 13:26:53 using qtopia on my sd-card but want to test other dists on the phones memory Feb 21 13:27:01 and using Qi Feb 21 13:27:09 is ot possible? Feb 21 13:43:51 ahoi Feb 21 13:44:28 Daverix: chroot? Feb 21 13:48:06 have you ever experienced illume on debian? keyboard seems half opened Feb 21 13:49:08 how do I use chroot? :) Feb 21 13:49:45 Daverix: you can skip boot possibilities with aux button in qi Feb 21 13:50:50 Daverix: not in qry please Feb 21 13:51:42 qry? Feb 21 13:53:01 query Feb 21 13:53:01 i read that you must implement your own boot loader if you want to use Qi? Then it's only possible to boot from one dist at the moment if not one dist has a boot option? Feb 21 13:53:16 no. qi tries sd partition 3, then 2, then 1, then nand Feb 21 13:53:25 read the wikipage about qi Feb 21 13:53:50 but if I have a dist on my sd-card and want to boot from nand? Feb 21 13:54:00 i have Feb 21 13:54:20 then read it again Feb 21 13:54:25 there is no boot loader Feb 21 13:54:44 Daverix: boot using NOR Feb 21 13:55:32 Daverix: http://wiki.openmoko.org/wiki/Qi#Choosing_a_Kernel Feb 21 13:57:40 aha! now i get it :) Feb 21 14:04:49 Daverix: :) Feb 21 14:07:14 uhm wtf? 'stop and exit zhone' shuts the device down? Feb 21 14:35:56 spaetz: what FSO version is the current unstable release of SHR using? Feb 21 14:37:37 stra: its based on the fso/milestone5 branch of openembedded... for frameworkd SHR is using AUTOREV aka current HEAD Feb 21 14:38:37 stra: you can check the installed packages in http://build.shr-project.org/shr-unstable/images/om-gta02/openmoko-shr-lite-image-glibc-ipk--20090218-om-gta02-testlab/installed-packages.txt Feb 21 14:40:30 mrkoku, thank you for the info Feb 21 14:40:41 mkmoku: what exactly is your role in SHR? Feb 21 14:41:13 stra: don't know :-) doing this and that... Feb 21 14:41:31 looking after the buildhost is one thing of that Feb 21 14:43:17 i think I will give shr a try Feb 21 14:44:39 hope to get a chat with spaetz, he lives just 20min from me Feb 21 15:03:04 0ffrf/exi Feb 21 15:29:21 hi, I have a pygame issue, anybody could help me please? Feb 21 15:31:03 lpotter: ping Feb 21 15:35:23 <|Marco|> sylar91: what's the problem ? Feb 21 15:36:28 |Marco|: when I create a window it fills with the black color the part of the screen unused by pygame Feb 21 16:10:53 feb 18 andys tracking Feb 21 16:11:00 on shr Feb 21 16:11:57 i guess i have damaged Qi Feb 21 16:22:23 Hi, Ive got a question to the openmoko toolchain Feb 21 16:22:54 Ive downloaded the most recent one and followed the steps from the wiki page about the toolchain Feb 21 16:23:22 but there is no file opkg-target in the archive I downloaded Feb 21 16:23:35 only opkg-key and opkg-cl Feb 21 16:25:57 run . ./setup-env Feb 21 16:26:06 i did Feb 21 16:26:21 I also run fine /usr/local/openmoko -name opkg-target Feb 21 16:26:28 and add path Feb 21 16:26:34 I did Feb 21 16:26:43 find didnt return any results Feb 21 16:26:53 so opkg-target doesnt exist Feb 21 16:27:20 should exist Feb 21 16:27:25 re check all Feb 21 16:27:42 it doesnt exist Ive tried it several times now Feb 21 16:28:10 sorry, try that . /usr/local/openmoko/arm/environment-setup Feb 21 16:28:44 you dont seem to understand. find doesnt find any file or directory named opkg-target Feb 21 16:29:07 so whatever I would add to my path, it doesnt come into existence Feb 21 16:32:03 dunno, haven't thought about that before :) Feb 21 16:32:47 after environment-setup, opk-target appears at my Feb 21 16:32:57 hm Feb 21 16:33:09 what version of the toolchain do you have Feb 21 16:33:42 http://downloads.openmoko.org/developer/toolchains/openmoko-i686-arm-linux-gnueabi-toolchain.tar.bz2 Feb 21 16:34:27 yeah, thats the one I have Feb 21 16:34:38 so youvve got the latest Feb 21 16:34:44 yep Feb 21 16:34:49 because this file is actually no file but a link Feb 21 16:35:24 can you give me the md5 sum Feb 21 16:35:47 mine is 85b479850768a851847f33cee4584d79 Feb 21 16:35:56 of the bzipped archieve Feb 21 16:38:01 i've deleted it :) Feb 21 16:38:12 hm, k Feb 21 16:39:04 when did you download it Feb 21 16:39:26 two days ago Feb 21 16:39:31 oh Feb 21 16:39:44 ill give it another try Feb 21 16:44:29 baba_melone_ : opkg-target is alias Feb 21 16:44:31 alias opkg-target='LD_LIBRARY_PATH=/usr/local/openmoko/arm/lib /usr/local/openmoko/arm/bin/opk Feb 21 16:44:41 look at /usr/local/openmoko/arm/setup-env Feb 21 16:48:51 hmm, is it possible that ram contents is preserved over poweroff? Feb 21 16:52:25 lindi-: yes, for some time Feb 21 16:52:28 There is a near-zero power mode which will. Feb 21 16:52:33 Self-refresh of RAM. Feb 21 16:52:39 PaulFertser_: this is biting me. i have a circular log buffer Feb 21 16:52:41 Even without that, data may persist for some seconds. Feb 21 16:53:10 lindi-: even without power SDRAM stores its contents for some time. Especially when it's cold. Feb 21 16:53:12 PaulFertser_: it is recognized by some magic numbers. those magic numbers seemed to survive poweroff but the contents of the log is full of garbage Feb 21 16:53:36 PaulFertser_: now it takes 5 minutes for it to dump this random garbage to my screen :) Feb 21 16:53:38 lindi-: how long do you keep it in poweroff? Feb 21 16:53:47 probably around 5 seconds Feb 21 16:54:03 lindi-: then preserving some sdram contents is normal Feb 21 16:54:13 kind of interesting Feb 21 16:54:31 lindi-: there's a security research paper about, made a lot of fuss. Feb 21 16:54:42 PaulFertser_: so if the phone does not come out of suspend i can take the battery out for a really short while and then be able to read the kernel log buffer Feb 21 16:54:46 PaulFertser_: yeah, cold boot attacks Feb 21 16:55:13 (watchdog is disabled during suspend so it does not help) Feb 21 16:55:25 lindi-: i think so, yes. Feb 21 16:55:26 lindi-: reliability really depends on lots of 'stuff'. Feb 21 16:55:50 SpeedEvil: yep Feb 21 16:55:57 A good rule of thumb is that it only works if you don't want it to. Feb 21 16:56:02 :) Feb 21 16:56:11 sdram content may survive like 90% after 10min Feb 21 16:56:29 DocScrutinizer2: that has interesting power saving opportunities :) Feb 21 16:56:43 DocScrutinizer2: just put disk cache to ram and CRC each page Feb 21 16:56:46 lindi-: btw, that patch of yours is cool :) Feb 21 16:56:55 lindi-: self-refresh is very low power. ~3mW IIRC. Feb 21 16:56:56 DocScrutinizer2: then powerdown the other 64M ram chip Feb 21 16:57:05 PaulFertser_: which one? Feb 21 16:57:08 lindi-: you could in principle last months. Feb 21 16:57:12 PaulFertser_: ah, ramconsole Feb 21 16:57:33 alas power consumption to regenerate the lost bits will be higher than possible savings Feb 21 16:59:27 often after boot you also see a short flash of last framebuffer content, even after minutes Feb 21 16:59:46 Isn't that internal glamo? Feb 21 16:59:49 (on FR) Feb 21 16:59:54 yup Feb 21 17:00:16 anyway it's also dram Feb 21 17:00:44 baically lots of small capacitors Feb 21 17:00:45 But implementing suspend to self-refreshed RAM, and going cold otherwise would in principle save oodles of power. Feb 21 17:00:59 yup Feb 21 17:01:07 (though maybe not really so startling if you've got GSM on) Feb 21 17:01:12 suspend to ram Feb 21 17:02:54 Good day, Joerg :) Feb 21 17:03:02 iirc we're already switching ram to lowpower selfrefresh mode on suspend. we don't real suspend to ram though, keeping register content in cpu static Feb 21 17:03:12 hi dave Feb 21 17:03:57 doc|home: yeah - from memory the amount of power the CPU uses is really quite high, even in its lowest power mode (though less so if clocked down). But ~48h is fine. Feb 21 17:04:00 DocScrutinizer2 Feb 21 17:04:49 powering down SoC is a quite complex task Feb 21 17:04:52 (lowest power modes that don't require essentially a reboot on startup) Feb 21 17:04:56 yes. Feb 21 17:05:32 You've got to teach the bootloader stuff, get it all coming up really rapidly. Feb 21 17:05:59 And if it lasts 48h without bothering, it's a hell of a lot of work to go from 48-70h (with GSM on) Feb 21 17:06:02 nasty is samsung doesn't provide decent power consumption figures in their doc Feb 21 17:06:12 for power down states Feb 21 17:06:26 of different parts of SoC Feb 21 17:06:35 They somewhat break it down. Feb 21 17:07:10 (from memory of the 2440) Feb 21 17:07:25 But that, and 'real' measurements are very lacking. Feb 21 17:08:16 Well - rds(on) will lead to _much_ higher power dissipation if you replace a 0.5R fet with a 2R one. So you can't just assume that it'll work. Feb 21 17:08:18 oosp Feb 21 17:12:16 SpeedEvil: would be cool if somebody uses the device as a PDA and does not have GSM on Feb 21 17:12:41 yes. Feb 21 17:12:57 also we should get gsm deep sleep work but that is another story Feb 21 17:13:05 Potentially also if you're in quiet mode, you want to turn GSM off. Feb 21 17:13:27 Or if you've got under 5% of battery power, suspend to disk even. Feb 21 17:14:03 sure Feb 21 17:14:14 (though with glamo-SD, that's hideously slow) Feb 21 17:14:28 (or with GTA01-SD too.) Feb 21 17:14:58 If you can hit the SD's 'real' speed, then it may not be bad at all with fast SD. Feb 21 17:18:15 btw, just a poll, how much do your phones consume during suspend when gsm is on? Feb 21 17:18:50 * SpeedEvil hasn't measured with anything close to a recent kernel. Feb 21 17:18:54 i just replied to andy that here it is around 33-35 mA Feb 21 17:19:12 With very ancient GSM firmware on GTA01 with everything else shutdown 30mA Feb 21 17:19:22 that's shutdown - not suspended. Feb 21 17:19:43 if I can trust 'current_now' immediately after resume Feb 21 17:19:53 SpeedEvil: interesting Feb 21 17:20:01 i think i should measure this with GSM off Feb 21 17:20:07 I dunno. I'd trust the integrated charge over an hour more. Feb 21 17:20:15 SpeedEvil: is the AT@POFF required to realy power of GSM? Feb 21 17:20:21 lindi-: it was. Feb 21 17:20:27 but isn't? Feb 21 17:20:32 lindi-: I don't know if it is with current firmware. Feb 21 17:20:42 lindi-: GTA02v5 has a hardware power switch Feb 21 17:20:47 ok Feb 21 17:21:02 lindi-: on gta02 nope Feb 21 17:21:26 * SpeedEvil hasn't been keeping up on GTA02. :/ Feb 21 17:21:39 ok, i'll check what this consumes when gsm is off Feb 21 17:21:52 yeah Feb 21 17:22:26 * DocScrutinizer2 assumes that's the strange difference between "gsm-antenna power" and "gsm enable" Feb 21 17:24:30 PaulFertser_: what's stae of integrating switch and pushbutton operation in /sysfs gsm node? Feb 21 17:24:47 s/stae/state/ Feb 21 17:24:47 DocScrutinizer2 meant: PaulFertser_: what's state of integrating switch and pushbutton operation in /sysfs gsm node? Feb 21 17:25:54 DocScrutinizer2: to be honest, i was distracted by other funny things :( Feb 21 17:26:36 PaulFertser_: I stopped Werner and Andy from looking into it... Feb 21 17:26:51 SpeedEvil: 26 mA during suspend when GSM is off Feb 21 17:27:02 i am not sure if that one data point can be trusted though Feb 21 17:27:41 SpeedEvil: that would mean that gsm consumes only 7-9 mA Feb 21 17:27:47 SpeedEvil: and the other parts consume most of it Feb 21 17:29:06 lindi-: there's just one sensible way to answer your question: suspend for one hour, on time with gsm and another time without, then each time readout charge state of C-counter before and after Feb 21 17:29:35 7-9mA looks about right though Feb 21 17:30:50 DocScrutinizer2: i was all enthusiastic about that and then suddenly i thought "we lived so long with it working improperly, why hurry now"... And somehow that put this idea to the back of my mind. And, btw, for GSM flashing process i'd use pmu and gpio anyway to be on the safe side and kernel-independent. I can make it my first priority if you like (the other tasks i have in mind is routing diagram and looking more into alsa driv Feb 21 17:33:00 would be great to come up with a fix, if it was only for keeping credibility ;-) We told "will be fixed" Feb 21 17:34:00 Choose: routing diagram first or the fix first? ;) Feb 21 17:34:06 and like with alsa noone really understands how to operate the current nonsense Feb 21 17:34:31 sysfs gsm that is Feb 21 17:35:16 lindi-: and yes, I agree 7-9mA seems about right Feb 21 17:36:32 should be 4-5mA iirc, but so hey what Feb 21 17:37:43 even on 10mA gsm can operate 120h from a fullcharged bat Feb 21 17:38:26 for bat live we need to find powerhogs that really do nasty Feb 21 17:39:21 gsm could be one of them, when actively reregistering every minute Feb 21 17:39:37 (1024) Feb 21 17:40:12 7-9mA might be accurate for sleep=2 Feb 21 17:40:58 sleep=4 might give you 4mA and constant recamping Feb 21 17:41:50 ...pushing average consumption to several tens of mA Feb 21 17:42:52 put FR on your active noiseqube to check ;-) Feb 21 17:43:51 each register-action gives U wellknown dr-drr-dr-drr-dr buzz Feb 21 17:52:59 PaulFertser_: (know operate gsm nonsense) even Werner comes up with paradox results about "switching off gsm" doesn't reduce or even increases consumption. My only explanation for this is he didn't realize switching sysfs is *not* powering down gsm reliably Feb 21 17:54:44 DocScrutinizer2: i don't say it's not necesarry. I should have done that long time ago. :( Feb 21 17:55:25 DocScrutinizer2: as to his results, probably he used Qi before we found that its PMU setup prevented controlling power switch. Feb 21 17:56:17 basically not gsm is like motor bike. switch on ignition (powerswitch) isn't actually starting motor. releasing kickstart (sysfs) doen't stop the motor Feb 21 17:56:48 s/not/now/ Feb 21 17:56:48 DocScrutinizer2 meant: basically now gsm is like motor bike. switch on ignition (powerswitch) isn't actually starting motor. releasing kickstart (sysfs) doen't stop the motor Feb 21 17:58:09 unlike motorbike on gsm user has no real feedback about actual state of "motor" Feb 21 17:58:18 Though sysfs performs PMU action to switch GPIO that controls power anyway. My guess is that he used Qi in that testing and that PMU GPIO issue wasn't yet known. Feb 21 17:58:46 yup, maybe Feb 21 17:59:11 Turning off ignition will stop the motor no matter which state kickstart is :) Feb 21 17:59:46 anyway I consider sysfs status flaky at least, and needs a decent fix for sure. The sooner the better Feb 21 18:00:35 (kickstart) like does switch, no matter what stae of pushbutton is Feb 21 18:01:27 I understand :( So, what goes first, wm8753 routing diagram with alsa control names or this sysfs issue (which can require some time to test how much time pushbutton requires to act reliably). Feb 21 18:02:00 freesmartphone.org: 03mickey 07fso-abyss * rf9e59f7f5fe0 10/src/serial.vala: io_add_watch_full patch has been commited upstream with changes; catch up Feb 21 18:02:29 power up (=sysfs on) should be: enable powerswitch. wait 100mS, push button. Feb 21 18:02:47 I know, i saved your detailed explanation Feb 21 18:03:11 button released by first "hello" from gsm uart Feb 21 18:03:34 (nasty for different-driver-issue?) Feb 21 18:04:28 power off via sysfs should be: send AT@POFF, wait 1sec, disable powerswitch Feb 21 18:04:50 s/1/2/ Feb 21 18:04:51 DocScrutinizer2 meant: power off via sysfs should be: send AT@POFF, wait 2sec, disable powerswitch Feb 21 18:05:10 Why send AT@POFF? Feb 21 18:05:25 (gsm modem likes tto say bye to BTS) Feb 21 18:05:47 How they lived when this command was not yet implemented? Feb 21 18:05:58 PaulFertser_: I'd prefer fixing sysfs first Feb 21 18:06:48 DocScrutinizer2: sure, i'll start working on this right now then Feb 21 18:06:59 (lived) doesn't really hurt. BTS will recover eventually - in the end you could have went downstairs to parking deck as well Feb 21 18:07:26 So, probably we can avoid using it to not mess with different driver. Feb 21 18:07:42 it's just polite Feb 21 18:07:58 PaulFertser_: (right now) great Feb 21 18:09:11 You previous description of powering included asserting reset too. Do you now think that applying power without touching reset is enough? Feb 21 18:09:37 yup Feb 21 18:10:08 reset isn't applicable for gta02 anyway, and seems we don't need it Feb 21 18:11:14 PaulFertser_: calypso does poweron-reset autonomously it seems. Feb 21 18:12:30 in gta01 reset didn't do any good besides "cycling" for gsmflashing. it even hurts. so it's NC for gta02 Feb 21 18:12:46 we don't care about it Feb 21 18:16:53 ahoi Feb 21 18:17:03 DocScrutinizer2: yes, i see NC now. Sure, doesn't seem to do any good. BTW, Andy allowed to do a 1-2 sec msleep there on power on, so i guess implementing some special measures to depress pushbutton only on first bytes from calypso. Feb 21 18:17:31 ack Feb 21 18:17:52 first shot it's perfect anyway I guess Feb 21 18:18:41 Because it looks like the serial driver is generic (that flowcontrolled node is implemented independently), so messing with can be too hard Feb 21 18:19:31 maybe we find we can reduce to sth like 50mS so we're fine with 100mS here as well. that's chrome though, shugar on top Feb 21 18:20:22 just go ahaead implement a decent wS delay to start with Feb 21 18:20:31 err 2S Feb 21 18:21:02 mickey|cinema: enjoy! :-) Feb 21 18:21:08 thanks, cu Feb 21 18:21:17 cu Feb 21 18:36:32 <|Marco|> PaulFertser_: hey, I've got a few questions arround modifying qi if possible Feb 21 18:37:51 |Marco|: go on :) Feb 21 19:04:55 <|Marco|> Wonka: are you accessable a.t.m. ? Feb 21 19:05:12 |Marco|: what for? Feb 21 19:05:25 aka "depends" Feb 21 19:05:40 <|Marco|> initramfs questions (if I've bugged the right person) Feb 21 19:06:36 <|Marco|> you've been working on some form of bootmenu for initramfs? or have I poked the wrong person ? Feb 21 19:09:03 |Marco|: i intend to be some time, but atm i am just moving some thoughts around the idea. Feb 21 19:09:27 <|Marco|> can I poke you in pm ? Feb 21 19:12:04 <|Marco|> well, how much access do you get to the host device from initramfs, to begin with Feb 21 20:08:33 |Marco|: host device? Feb 21 20:29:32 is a higher class of micro sdhc better than lower classes? Feb 21 20:29:45 s/classes/class Feb 21 20:42:54 which kernel is best suited for shrlatest .jffs2 file feb 18 Feb 21 20:43:20 the latest uimage? Feb 21 20:45:03 i have andy more drivers tracking -8601 Feb 21 20:54:13 inconsistencies in flashing are killing me Feb 21 20:55:18 ok dave u win im dead Feb 21 20:55:25 Hello everyone. Feb 21 20:55:49 howdy Feb 21 20:56:00 I'm trying to make an illume keyboard for Greek. Feb 21 20:56:19 yea Feb 21 20:56:38 Does anyone know about which keysyms openmoko uses? Feb 21 21:42:47 well it finally does work Feb 21 21:43:10 raster: mofo, e is on my phone again Feb 21 21:43:40 * rooly will bbl Feb 21 21:45:04 eeeek Feb 21 21:45:28 is it she feb 18 an latest.bin Feb 21 21:45:30 ? Feb 21 21:45:35 shr Feb 21 21:46:11 i have had trouble getting sound to work before now Feb 21 21:46:45 When is the next phone expected to be released? (ie, the one after the Freerunner) Feb 21 21:47:15 afaik not specified yet Feb 21 21:48:00 mmm Feb 21 21:48:05 indeed, not specified yet Feb 21 21:48:13 Does the freerunner have touchscreen? Feb 21 21:48:13 "normally" it would be "this year" Feb 21 21:48:31 but we all know about the FR, the deadline just kept moving :) Feb 21 21:48:57 i like my freerunner Feb 21 21:49:00 Ah yes, well I always buy devices shortly before the next one is announced Feb 21 21:49:22 So my plan is to get the GTA03 at release time Feb 21 21:50:16 check the recent archives for some recent dialog about that Feb 21 21:53:56 ziarkaen: maybe you can obtain a defected freerunner and try to repair it, like i did. got a FR this way for ~50 euros.. but i had to resolder the AUX button, which turned out to be broken (which was still better than what i thought first as i couldn't enter NOR flash bootloader and thought i had to get a debug board Feb 21 21:55:15 yea Feb 21 21:55:16 That seems very reasonable Feb 21 21:55:24 Although there are no guarantees :p Feb 21 21:57:58 ziarkaen: sure. but when i grabbed mine at a well known internet auctionist, i thought that it was pretty likely to get rid of it for about the same amount of money or at least without much loss Feb 21 22:01:08 How many have been produced? Feb 21 22:02:06 <20k afaik Feb 21 22:02:23 Wow. Feb 21 22:02:32 it's a collectors item :p Feb 21 23:47:11 <[Rui]> gah my new sim card doesn't register! Feb 21 23:47:30 <[Rui]> I've upgraded the GSM firmware to moko10 and it doesn't register, sadly :( Feb 21 23:50:52 <[Rui]> echo -en 'AT+CGMR\r' >/dev/ttySAC0 is not giving me the gsm version! Feb 21 23:51:26 * [Rui] phews... it does now... Feb 21 23:51:26 [Rui]: if you just upgraded to the version and fluid didn't give any errors, most probably everything is ok. Try your telephony distro. Feb 21 23:51:38 <[Rui]> at least it doesn't seem to be borked Feb 21 23:51:59 <[Rui]> PaulFertser_: I did that, and it still doesn't register the sim card Feb 21 23:52:25 [Rui]: probably 3G-only network? Feb 21 23:52:51 <[Rui]> PaulFertser_: not likely, since its plan allows gprs Feb 21 23:53:37 [Rui]: there were no reports about 2G-compatible sims not working after upgrade to moko10, afaik. Check electrical contacts probably. Feb 21 23:54:00 <[Rui]> PaulFertser_: I'm plugging my old sim Feb 21 23:54:15 <[Rui]> PaulFertser_: the problem is that in a few days it won't work anymore :( Feb 21 23:54:29 [Rui]: what if you have grease on your sim or pins? Feb 21 23:55:08 <[Rui]> PaulFertser_: doesn't look like it Feb 21 23:55:47 [Rui]: It'd be good if you try your new sim in another 2G-only phone. Feb 21 23:56:04 <[Rui]> PaulFertser_: I don't have any Feb 21 23:56:14 <[Rui]> I think Feb 21 23:57:16 [Rui]: what distro are you using? Can it be that it's not registering problem itself, but something else? Feb 21 23:57:24 <[Rui]> PaulFertser_: 2008.12 Feb 21 23:57:53 <[Rui]> PaulFertser_: I'm using it as my day to day phone, and it's the best one there is, for the moment, and to me. Feb 21 23:58:49 <[Rui]> now i have rebooted with my old SIM, and it worked purrfectly Feb 21 23:59:23 [Rui]: it might worth trying to register manually using instructions from http://docs.openmoko.org/trac/ticket//2231 and look what it gives. Feb 21 23:59:30 [Rui]: otherwise, i'm out of ideas. Feb 22 00:00:20 <[Rui]> PaulFertser_: well, thanks for the link, it's one idea more I haven' tried :) Feb 22 00:00:44 <[Rui]> maybe it needs to load up once on a more conventional phone Feb 22 00:01:00 <[Rui]> aha! his other phone doesn't register either Feb 22 00:01:02 <[Rui]> that's good Feb 22 00:01:14 <[Rui]> this way I have a quite normal phone to complain about Feb 22 00:01:39 <[Rui]> it's a N O K I A so it should work, shouldn't it? (hope the vendor balks) Feb 22 00:02:07 [Rui]: Good. BTW, maybe you need some additional "first registration" steps? Feb 22 00:02:38 [Rui]: or it doesn't register to network at all? Feb 22 00:02:59 <[Rui]> PaulFertser_: doesn't register. Feb 22 00:03:10 <[Rui]> message was: "failed to register sim card" Feb 22 00:03:25 [Rui]: cool, now you have a decent way to complain :) Feb 22 00:03:33 <[Rui]> hell yeah! Feb 22 01:20:42 <|Marco|> is there a trick to storing changes to u-boot boot-menu ? Feb 22 01:57:28 my shr has vertical lines Feb 22 01:57:38 after suspend Feb 22 01:58:13 have to reboot **** ENDING LOGGING AT Sun Feb 22 02:59:57 2009