**** BEGIN LOGGING AT Thu Feb 17 02:59:58 2011 **** BEGIN LOGGING AT Thu Feb 17 07:32:21 2011 Feb 17 10:16:40 jluis, gmorning. Did the image built for you? Feb 17 10:24:31 pespin: no it didn't but I think that this evening could be working (it has some failure during kernel config) Feb 17 10:25:47 jluis, ok, I'll try building from scratch removing tmp dir Feb 17 10:25:52 this afternoon Feb 17 10:27:29 jluis, btw, meego is no longer booting for me after I reparted. There's a kernel panic at bootup related to VFS. Feb 17 10:27:52 remember to use shr-testing Feb 17 10:28:24 ok Feb 17 10:29:45 I installed archlinux with grub, but there's now way to switch between them if there's no hw kb connected :P Feb 17 10:30:30 pespin: I think the problem is that you touch de partition table but didn't sincronize it with efi partition table ( http://es.wikipedia.org/wiki/Extensible_Firmware_Interface) Feb 17 10:30:49 jluis, can be, I think it had some problem with some kind of sync Feb 17 10:46:32 jluis, hacking a mac? Feb 17 10:46:57 methril_work, we got exoPC slate tablets from intel yesterday :P Feb 17 10:47:09 pespin, nice!! Feb 17 10:47:14 yeah :D Feb 17 10:47:21 how are you? Feb 17 10:47:44 methril_work, in an economy class lol. I'm fine :) Feb 17 10:47:50 methril_work, how's life in brasil? Feb 17 10:47:56 hi methril_work Feb 17 10:47:59 what are you going to do with the exoPCs? Feb 17 10:48:19 methril_work, I'm planning to have meego, archlinux and SHR there. Feb 17 10:48:21 porting shr to it Feb 17 10:48:39 pespin, everything is ok, but i didn`t have so much time to hack devices Feb 17 10:48:46 uhm!! nice!! Feb 17 10:49:24 i need come back to shr/oe hacking ;) Feb 17 10:49:38 we're waiting for you! hehe Feb 17 10:49:42 hehehe Feb 17 10:50:00 i also have to find time to do some kernel dev & hw hacking too Feb 17 10:50:48 let me organize my time Feb 17 10:51:45 where did you bought the exoPCs? Feb 17 10:52:05 methril_work, we didn't bought them. Feb 17 10:52:21 pespin, this is better /) Feb 17 10:52:24 ;) Feb 17 10:52:37 methril_work, we went to a MeeGo conference / Intel AppUp Lab and they gave them to us :P Feb 17 10:52:49 oh!! :) Feb 17 10:53:02 pespin, you went to FOSDEM? Feb 17 10:53:08 methril_work, yes :) Feb 17 10:53:17 methril_work, hope to se you next year Feb 17 10:53:59 Barcelona must be the city with more exoPC in the world Feb 17 10:54:03 * methril_work is jealous Feb 17 10:54:18 jluis, lol!! Feb 17 10:54:40 i`ve to plan my next (or later) trip to Europe Feb 17 10:55:21 * GNUtoo|laptop is not jealous, he already has too much hardware Feb 17 11:18:28 hi mrmoku Feb 17 11:18:54 mrmoku, look at udev, for instance udev 151,162,165 Feb 17 11:19:02 there is a big issue Feb 17 11:19:33 in commi a8d15196 you forgott to MACHINE_ARCH your changes Feb 17 11:20:09 what should we do now? Feb 17 11:20:15 I machine arch everything? Feb 17 11:20:18 I mean Feb 17 11:20:22 udev 151,162,165 Feb 17 11:20:31 should I bump PR after that? Feb 17 11:49:02 hi GNUtoo|laptop Feb 17 11:49:48 * mrmoku called for lunch Feb 17 11:49:49 bbiab Feb 17 12:23:42 mickey|office, hi Feb 17 13:03:07 hi GNUtoo|laptop Feb 17 13:04:45 * mrmoku should take some time today to debug libphone-ui... segfaults everywhere :-/ Feb 17 13:05:11 mickey|office: hi, did you see my note regarding ppp? Feb 17 13:05:37 mrmoku: yes, although i have no idea what could be wrong. that code hasn't been touched for ages Feb 17 13:06:01 i don't have a SIM card with a data plan atm., so it will take a while until i can test that for myself Feb 17 13:06:06 yeah, maybe something needs just rebuilding Feb 17 13:27:32 mrmoku, could you fix udev? Feb 17 13:34:02 GNUtoo|laptop: let me see what I did break Feb 17 13:34:18 you didn't MACHINE_ARCH your changes Feb 17 13:34:20 for n900 Feb 17 13:35:13 you override stuff for nokia900 Feb 17 13:35:18 and don't machine arch the result Feb 17 13:35:57 I bet I don't have to explain the rationale behind machine arch to you Feb 17 13:36:09 in the countrary just ask Feb 17 13:37:52 yeah, will fix Feb 17 13:38:01 can't do it right now though... on my way to the office Feb 17 13:38:44 bbl Feb 17 13:39:00 ok Feb 17 13:39:02 thanks a lot Feb 17 13:39:39 mrmoku, btw udev is a crutial thing right, it's considered core right? Feb 17 13:39:59 so usually you send a patch and that requires 2 acks Feb 17 13:40:23 because else if you break something you could have issues Feb 17 13:40:31 with people like koen Feb 17 13:40:59 mrmoku: fwiw, this ppp problem might motivate me to look at ofono's ppp implementation... Feb 17 13:41:29 mrmoku: i have been looking for something to substitute external process control for a while Feb 17 13:42:16 mickey|office, you need more tests? Feb 17 13:49:27 for the modem bringup? not yet, i need to think about why it doesn't work without running testping Feb 17 13:50:28 ok Feb 17 13:50:34 I can test stuff Feb 17 13:50:46 such as bringing testping to the minimum Feb 17 13:50:57 such as bringing the modem up only Feb 17 13:52:12 GNUtoo|laptop: hi Feb 17 13:52:16 hi Feb 17 13:52:22 look at this : http://www.youtube.com/watch?v=9jcrSqCiNGo Feb 17 13:53:32 wow Feb 17 13:53:38 can I try it Feb 17 13:53:39 ? Feb 17 13:53:44 on a real phone? Feb 17 13:54:24 GNUtoo|laptop: let me finish some stuff and i will put the code on E svn Feb 17 13:54:31 ok nice Feb 17 13:54:37 thanks a lot!!!!! Feb 17 13:54:37 maybe next week Feb 17 13:54:39 ok Feb 17 13:54:40 :) Feb 17 13:54:55 code will be under BSD licence Feb 17 13:55:01 as it's a e module Feb 17 13:55:26 ok Feb 17 13:55:40 as long as it's free software it's fine for me Feb 17 13:55:45 for me too :) Feb 17 13:56:06 if e is under BSD it's the best thing to do Feb 17 13:56:19 basically using the same license as the project is a good strategy Feb 17 14:45:13 captainigloo, btw is your kenrel based on mainline? Feb 17 14:45:25 because I see a lot of activity on bcm80211 Feb 17 14:45:38 but I've no idea if it supports bcm4329 Feb 17 14:51:38 GNUtoo|laptop: my kernel is this one : https://github.com/teamdesire/cm-kernel Feb 17 14:51:53 no idea on wich branch/repo it is based Feb 17 14:51:58 ok Feb 17 14:51:59 it's my main problem .... Feb 17 14:52:40 I think it's based on cyanogen kenrel Feb 17 14:53:15 it's a 2.6.37 Feb 17 14:53:46 i'm using the unified branch Feb 17 14:53:53 2.6.37 unified branch Feb 17 14:53:53 what's that? Feb 17 14:54:17 ah ok Feb 17 14:54:18 the branch of the link i pasted Feb 17 14:54:21 it's the only branch I bet Feb 17 14:55:05 https://github.com/teamdesire/cm-kernel/tree/android-msm-2.6.37-unified/drivers/staging/brcm80211 Feb 17 15:10:31 captainigloo, if ((bus->sih->chip == BCM4329_CHIP_ID) in https://github.com/teamdesire/cm-kernel/blob/android-msm-2.6.37-unified/drivers/staging/brcm80211/brcmfmac/dhd_sdio.c Feb 17 15:17:36 GNUtoo|laptop: trying right now Feb 17 15:18:23 ok Feb 17 15:18:29 note that it's a staging driver Feb 17 15:18:32 GNUtoo|laptop: i enable Broadcom IEEE802.11n WLAN drivers Feb 17 15:18:42 and that there were a lot of commits after the 2.6.37 Feb 17 15:18:44 ok Feb 17 15:18:54 disable the old driver Feb 17 15:18:58 when you try Feb 17 15:19:55 you mean for 2.6.38-rcx ? Feb 17 15:27:47 for the kenrel you have Feb 17 15:27:58 you disable the CONFIG_ for bcm4328 Feb 17 15:28:02 *4329 Feb 17 15:28:11 and you add the config for brcm80211 Feb 17 16:22:49 GNUtoo|laptop: ok compilation done, i will try at it tonight Feb 17 16:23:04 ok Feb 17 16:23:08 nice Feb 17 16:23:20 tell me so I know if it's worth trying on nexusone Feb 17 16:23:38 I must still check the initramfs for cotulla Feb 17 16:23:47 I've tons of stuff to do Feb 17 16:23:55 I've also a kenrel patch to look after Feb 17 16:24:01 for mainline Feb 17 16:24:39 I've to do some stuff for replicant too Feb 17 16:24:48 like re-doing a ril tracing patch Feb 17 16:28:28 jluis, I'm back Feb 17 16:31:17 GNUtoo|laptop: btw i need to send my patches for the bravo platform to openembedded Feb 17 16:31:30 pespin: I'm goin home bbl Feb 17 16:32:21 GNUtoo|laptop: but i don't know if it's a good idea to provide oe and .bb patches against this cm-kernel ? what do you think ? Feb 17 16:33:35 hmmm if I want to connect the tablet with the PC through usb0, do I need to set one to device? I suppouse both are set as host.. Feb 17 16:34:08 as none of them shows usb0 when I connect them with a cable. Feb 17 16:34:21 (I'm using archlinux in both tablet and PC) Feb 17 16:36:23 captainigloo, hmm I did it that way: I forked an android kenrel and put it on gitorious Feb 17 16:36:29 s/I/we/ Feb 17 16:36:30 GNUtoo|laptop meant: captainigloo, hmm we did it that way: I forked an android kenrel and put it on gitorious Feb 17 16:36:40 s/I/we/g Feb 17 16:36:42 GNUtoo|laptop meant: captainigloo, hmm we did it that way: we forked an android kenrel and put it on gitorious Feb 17 16:38:17 and then we did an oe recipe Feb 17 16:38:42 http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/linux/linux-leviathan_git.bb Feb 17 16:39:13 ok so I guess i should do the same for bravo Feb 17 16:40:15 yes Feb 17 16:40:19 I need to pach some other files, those on wich there is occurence to nexusone Feb 17 16:40:28 and add | bravo Feb 17 16:40:33 patch where? Feb 17 16:40:47 can't remember Feb 17 16:40:52 but i will check Feb 17 16:40:53 openembedded? Feb 17 16:40:55 kernel? Feb 17 16:40:56 yes Feb 17 16:40:59 oepnembedded Feb 17 16:41:01 ah ok openembedded Feb 17 16:41:06 I'll explain Feb 17 16:41:17 basically we have overrides in openembedded Feb 17 16:41:29 that is to say you can have one kernel for multiples machines Feb 17 16:41:37 and only the defconfig is different Feb 17 16:41:51 here's what interesses you are the config files Feb 17 16:41:55 like xorg.conf Feb 17 16:42:07 to add an xorg.conf for your machine you need 2 things: Feb 17 16:42:08 yep but it was for other things like xorg Feb 17 16:42:16 *add the xorg.conf in: Feb 17 16:43:18 recipes/xorg-xserver/xserver-xorg-conf/htcbravo/ Feb 17 16:43:25 if your machine is named htcbravo Feb 17 16:43:38 the file should be named xorg.conf Feb 17 16:43:47 then you must increase the PR in: Feb 17 16:43:55 http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/xorg-xserver/xserver-xorg-conf Feb 17 16:44:31 hum ok Feb 17 16:44:58 seems easy when somebody explain ;) Feb 17 16:45:00 thanks a lot Feb 17 16:46:56 basicaly pr is what makes the package rebuild and what makes your htc device install it when you opkg upgrade Feb 17 16:48:39 PR for ? Feb 17 16:49:41 package revision Feb 17 16:49:51 there is one other thing that i would like to test, it's the msm xorg driver Feb 17 16:50:17 but it seems there is a pacth to remove neon optimisations in oe Feb 17 16:50:24 that's old Feb 17 16:50:32 basically forget that for now Feb 17 16:50:35 it can work Feb 17 16:50:37 why ? Feb 17 16:50:43 it need some modifications Feb 17 16:50:47 like kenrel modifications Feb 17 16:50:51 ah Feb 17 16:51:00 but it can give great speedups Feb 17 16:51:09 it is obtimized for NEON Feb 17 16:51:18 and has some hardware 2d acceleration too Feb 17 16:51:25 but it depend on certain syscalls Feb 17 16:51:30 s/syscalls/ioctls Feb 17 16:51:39 and a 32bits for e17 composite manager Feb 17 16:51:45 wow Feb 17 16:51:49 ah I understnad Feb 17 16:51:52 :) Feb 17 16:52:00 then you could invest time in it Feb 17 16:52:14 btw did you try 32bits framebuffer on the n900 ? Feb 17 16:52:22 I didn't succeed Feb 17 16:52:26 and had other stuff to do Feb 17 16:52:29 ok Feb 17 16:52:43 if give it a try on a beagleboard, and it was working out of the box Feb 17 16:52:45 basically when I changed the arguments of the omapfb kernel framebuffer driver Feb 17 16:52:51 xorg didn't start anymore Feb 17 16:53:01 I've no beagleboard Feb 17 16:53:05 but I can get a bug 2.0 Feb 17 16:53:19 oops Feb 17 16:53:23 sorry I missundersood Feb 17 16:53:32 what did you do on beagleboard? Feb 17 16:53:56 quick test at work Feb 17 16:54:02 for a new client Feb 17 16:54:03 ok Feb 17 16:54:11 what did you do exactly? Feb 17 16:54:20 and what was it running? Feb 17 16:54:21 i working for an automotive compagny Feb 17 16:54:22 angstrom? Feb 17 16:54:24 ok Feb 17 16:54:32 no I mean on the beagleboard Feb 17 16:54:39 you just opkg installed enlightenement Feb 17 16:54:50 and went into the settings Feb 17 16:54:54 and it worked out of the box? Feb 17 16:54:55 run angstrom + firefox + adobe flash Feb 17 16:55:06 just change uboot settings Feb 17 16:55:08 no I mean for getting composite Feb 17 16:55:14 ah ok Feb 17 16:55:24 and give the right arg to the kernel Feb 17 16:55:25 can you tell me what you changed in uboot settings? Feb 17 16:55:31 yesp Feb 17 16:55:35 what kenrel arguments? Feb 17 16:55:38 i'm seraching Feb 17 16:55:59 ok thanks a lot Feb 17 16:56:49 setenv dvimode 1280x720MR-32@60 Feb 17 16:56:58 ok thanks a lto Feb 17 16:57:00 *lot Feb 17 16:57:07 and vram settings Feb 17 16:57:17 ah right Feb 17 16:57:23 omapfb.vram=0:8M,1:4M,2:4M Feb 17 16:58:08 and that's all related to graphics mode Feb 17 16:58:30 ok thanks a lot, I'll give a second try Feb 17 16:58:31 and i use omapfb driver in the xorg config Feb 17 16:58:37 but my computer is already compiling..... Feb 17 16:58:38 ok Feb 17 16:58:42 I use that already Feb 17 16:58:55 so after that you should be able to run software composite manager Feb 17 16:59:06 but on 720p display is not really smooth Feb 17 16:59:15 it should be better on a 800x480 screen Feb 17 16:59:38 ok Feb 17 16:59:49 and harware accelerated composite works with closed sgx driver Feb 17 17:00:05 yes but as long as it's proprietary I won't use it Feb 17 17:00:08 :) Feb 17 17:11:44 PaulFertser, hi, do you know if it's posible to set usb of one of my PCs in "device" mode? I'm trying to connect with usb tablet+PC both using archlinux and I suppouse both are in host mode... Feb 17 17:14:30 pespin, I don't think it's possible, however there are other ways of networking, right? Feb 17 17:17:03 GNUtoo|laptop, problem is I installed archlinux and forgot to activate iw-tools while installing... so I only have wlan0 and no iwconfig u.u Feb 17 17:17:27 oh, but I have wpa_supplicant Feb 17 17:17:31 do you have wired networking? Feb 17 17:17:36 wpa_supplicant can work Feb 17 17:17:42 GNUtoo|laptop, tablet doesn't have eth Feb 17 17:17:52 even on open networks Feb 17 17:18:07 you have usb host on the tablet? Feb 17 17:18:13 if so connect an usb keyboard Feb 17 17:18:44 else with usb the possibility are nearly infinite: Feb 17 17:18:46 *usb key Feb 17 17:18:52 *usb wired card Feb 17 17:18:55 etc... Feb 17 17:19:00 GNUtoo|laptop, yeah, I'm using the kb+mouse Feb 17 17:19:07 ok nice Feb 17 17:19:16 I'll use wpa_supplicant Feb 17 17:19:30 but I'd like to be able to use usb0 there too Feb 17 17:44:00 * jluis reading the backlog Feb 17 18:00:03 PCs usually have no matching PHY for doing gadget mode on USB Feb 17 18:14:59 jluis, any news on OE built? Feb 17 18:17:41 looks like it failed in linux-meego fetch step Feb 17 18:18:02 it didn't get the full sources Feb 17 18:19:06 jluis, uh that's weird, maybe some problems with git server? Feb 17 18:19:44 I don't know I testit it now Feb 17 18:20:22 GNUtoo|laptop: wifi doesn't work Feb 17 18:20:39 captainigloo, ok Feb 17 18:20:45 do you have more details? Feb 17 18:21:14 firmware is not load Feb 17 18:21:19 ah Feb 17 18:21:25 maybe you need another firmware? Feb 17 18:21:26 but i don't say anything about it Feb 17 18:21:35 dmesg? Feb 17 18:21:56 http://pastebin.fr/10427 Feb 17 18:23:14 captainigloo, ok, how did you compile it, as module or in the kenrel Feb 17 18:23:18 maybe I know why Feb 17 18:23:21 in the kernel Feb 17 18:23:23 ok Feb 17 18:23:30 maybe you need the activator Feb 17 18:23:36 that is present in bcm4329 Feb 17 18:23:55 note that theses activator are non-standard Feb 17 18:24:01 activator ? Feb 17 18:24:02 and that the runtime PM api is the standrad Feb 17 18:24:05 yes like that: Feb 17 18:24:15 http://bobcopeland.com/srcs/android/msm_wifi.patch Feb 17 18:24:24 that's the one we use on htcdream Feb 17 18:25:07 ok i see Feb 17 18:26:39 pespin: http://meego.gitorious.org/meego-os-base/kernel-source ony has the patches and scripts to apply to a mailine kernel I'm reading them now Feb 17 18:27:27 GNUtoo|laptop: http://pastebin.fr/10428 I guess it's my wifi activator Feb 17 18:27:47 no Feb 17 18:27:52 it's part of it only Feb 17 18:28:12 ah Feb 17 18:28:16 basically that's what you have to use to make the activator Feb 17 18:28:23 static struct wifi_platform_data bravo_wifi_control = { Feb 17 18:28:28 that's the functions to call Feb 17 18:28:39 jluis, aah ok :) Feb 17 18:28:42 but WARNING it's a hack, it non-standard etc.... Feb 17 18:28:43 lol: http://jan.wildeboer.net/2011/02/microsoft-absolutely-no-free-software-for-windows-phone-and-xbox-apps/ Feb 17 18:30:48 TAsn, lol: Simon Phipps says: Presumably this also means that software under Microsoft’s own open source licenses is excluded. Feb 17 18:30:57 GNUtoo|laptop: http://gitorious.org/htc-msm-2-6-32 it's the kernel your are using on nexus one ? Feb 17 18:31:41 GNUtoo|laptop, lol. Feb 17 18:32:12 I use the ubuntu sound branch here: Feb 17 18:32:32 TAsn: what's so funny :P ? Feb 17 18:33:13 http://gitorious.org/htc-msm-2-6-32/leviathan-incoming/trees/nexusone+ubuntu_sound+WIP Feb 17 18:33:22 which is based on cyanogenmod Feb 17 18:33:31 + a patch for alsa Feb 17 18:33:35 yep ok Feb 17 18:34:10 I don't understant why it's working of you out of the box Feb 17 18:34:15 does someone know if meego uses Xorg? Feb 17 18:34:19 captainigloo, dunno :) Feb 17 18:34:21 what is working for me? Feb 17 18:34:24 wifi? Feb 17 18:34:25 pespin: yes Feb 17 18:34:27 it's not Feb 17 18:34:30 GNUtoo|laptop: yes Feb 17 18:34:32 I can't scan either Feb 17 18:34:35 captainigloo, ok :) Feb 17 18:34:39 ah ! Feb 17 18:34:42 I just did that Feb 17 18:34:47 nano /etc/network/interfaces Feb 17 18:34:52 or vi rather Feb 17 18:34:56 and added my essid Feb 17 18:35:01 and I ifup eth0 Feb 17 18:35:08 ok understant now :) Feb 17 18:35:24 but I could look for the activator if you want Feb 17 18:35:30 but it'll go in the todo list Feb 17 18:35:36 and has no guarantee to be done Feb 17 18:35:42 ok np Feb 17 18:35:45 I've too much stuff to do at the moment.... Feb 17 18:36:01 I've a kernel patch Feb 17 18:36:07 and to repair something I broke in oe Feb 17 18:36:18 or that is broken rather Feb 17 18:36:35 i will return to my launcher then ;) Feb 17 18:37:11 ahhh I didn't see a line in the recipe Feb 17 18:45:29 pespin: the juice is MeeGo-1.0 branch it has sources of 2.6.35 and 37 as tar.bz2 (mainline?), the machines configs patches and scripts Feb 17 18:48:10 jluis, ok Feb 17 18:48:20 jluis, uhm but I thought they were in meego 1.1 now Feb 17 18:52:32 jluis, pass me the modified recipe once you have something working please :) Feb 17 18:56:03 captainigloo, you could try wifi it should not be too difficult Feb 17 18:56:06 just add the patch Feb 17 18:56:11 and then change the .name to Feb 17 18:56:12 bcm4329_wlan Feb 17 18:56:30 maybe it will just work Feb 17 18:56:33 maybe not Feb 17 19:06:06 pespin: I'm just reading the meego kernel documents Feb 17 19:07:02 http://wiki.meego.com/Tips_on_compiling_and_packaging_kernels_in_MeeGo Feb 17 19:07:20 jluis, ok, I'm currently fighting agaisnt Xorg config in archlinux :) Feb 17 19:07:40 pespin, what's wrong? :P Feb 17 19:08:24 TAsn, have to configure xorg to work well with exoPC slate table :) Feb 17 19:08:39 *tablet Feb 17 19:08:41 fun :) Feb 17 19:09:01 how's the tablet? Feb 17 19:09:10 ok, now xorg works well, only thing missing is TS Feb 17 19:09:49 i.c Feb 17 19:09:50 pespin: do you wish the meego X config? Feb 17 19:09:50 TAsn, great :) http://www.exopc.com/en/exopc-slate.php Feb 17 19:09:51 any drivers? Feb 17 19:09:59 pespin, yeah, I saw the website Feb 17 19:10:01 jluis, I'm already using it :) Feb 17 19:10:07 also says Feb 17 19:10:09 but had to install/compile some drivers from AUR Feb 17 19:10:14 it's a windows 7 tablet Feb 17 19:10:35 TAsn, we got them at IntelAppup Lab, and they gave them to us with meego in it :) Feb 17 19:10:56 no win in our tablets Feb 17 19:10:58 :) Feb 17 19:11:50 I see a label on the back that remeber us that they had it >:) Feb 17 19:12:02 yep xD Feb 17 19:54:24 ah mickey|zzZZzz is not there Feb 17 19:56:58 btw guys, do you know of some bootloader which has touchscreen support? Feb 17 19:59:16 pespin, kexec ? Feb 17 20:02:51 GNUtoo|laptop, uhm well that would be more to boot from a running linux system with touchscreen to another system right? Feb 17 20:03:03 GNUtoo|laptop, I'd like something similar to grub but with TS support Feb 17 20:03:10 hmmm no idea Feb 17 20:03:16 kexec is the easiest way Feb 17 20:06:58 ok Feb 17 20:27:27 pespin: you can put the system on an sd and use he bios bootmenu that is touchscreen sensitive Feb 17 20:33:38 hi Feb 17 20:34:06 did you publish the draft? Feb 17 21:25:02 to whom it may concern (kernel guys?): N900: some hw facts: there are at very least 2 independent ways to change backlight brightness: direct programming of LCM via I2C (that'S what probably is done now via /sys), plus a BL inhibit GPIO line from SoC that probably can get pulsed via hw PWM (I suspect that one currently unused) Feb 17 21:25:45 (GPIO line) GAIA N4200:M4 GPIO6/PWM0 - the EE @ Nokia is really decent, most of the time Feb 17 21:26:03 btw, probably via GPIO7/PWM1 IHF_EN you might be able to output audio beep without any higher level DA data streaming interface or codec set up Feb 17 21:27:01 (the inputs at speaker amp are labeled _SDR and _SDL, whatever that means) Feb 17 21:27:01 interresting for the audio bip Feb 17 21:27:24 specially for bootloaders Feb 17 21:28:20 the more interesting part is about BL brightness. Seems the LCM controller has only 7 usable different brightness levels, via direct control of the LCM controller via I2C Feb 17 21:28:39 many users think the lowest level still way too bright Feb 17 21:28:40 ok Feb 17 21:28:52 wow ok Feb 17 21:29:21 with driving GPIO6 PWM you can arbitrarily reduce lower limit brightness Feb 17 21:29:30 ok Feb 17 21:29:47 kernel dudes seem didn't care about it Feb 17 21:29:57 can that be done with sysfs? Feb 17 21:30:08 like /sys/class/gpio Feb 17 21:30:17 I don't think the GPIO6/PWM0 has a sysnode Feb 17 21:30:21 ok Feb 17 21:30:40 haven't checked kernel sources Feb 17 21:31:05 * GNUtoo|htcdream has really a hard time trying to stay awake Feb 17 21:31:16 so that's all handwaving, both for how kernel works now, as well as what hw can do. Just an educated guess Feb 17 21:31:20 but...what bother me more is pm Feb 17 21:32:06 we have nearly no PM at all on nokia900 kernel > 2.6.28 Feb 17 21:32:13 like 2.6.35 Feb 17 21:32:23 or 2.6.37 Feb 17 21:32:42 even in the PM branch... Feb 17 21:33:17 pespin: hey. I see you've got plenty of answers already :) Feb 17 21:34:55 DocScrutinizer, btw with the media framwork...will the front camera work? Feb 17 21:36:27 and how loc can backloght go.....sounds very interesting... Feb 17 21:36:43 PaulFertser, yep :) Feb 17 21:36:52 hardware encryption, Feb 17 21:36:59 wifi improvements Feb 17 21:37:04 etc.... Feb 17 21:37:15 lots of interresting stuff... Feb 17 21:39:18 there is even ad-hoc... Feb 17 21:39:21 I was at my friend's and configured bluetooth networking between his PC and n900 (to get internet to n900 via NAP). Surprisingly, default maemo config has "network" bluez plugin disabled, but even after enabling it, the NAP doesn't work well, it connects and all but the download speed is ~500 B/s. Strange stuff. Feb 17 21:39:36 with batmand....that could be great Feb 17 21:39:48 GNUtoo|htcdream: BL can go down to 0, I guess. The question is granularity. humans need a factor 2 for one step of brightness, so with 8bit for 0.255 you got 7 brightness levels, plus off. PWM-dimming level 1 might add another maybe 8 or 10 levels down to off Feb 17 21:39:49 there is even batmanv2 in staging Feb 17 21:40:05 ah ok Feb 17 21:40:50 GNUtoo|htcdream: what's interesting is that afaict all human senses work the same way: logarithmic sensitivity. Feb 17 21:41:02 yep Feb 17 21:41:07 10 => 1024 values... Feb 17 21:41:57 PaulFertser: a warmup task for you ;-D Is GPIO6/PWM0 referenced anywhere in maemo kernel code (or mainstream or meego, for this topic) Feb 17 21:42:10 DocScrutinizer: i'll look it up now Feb 17 21:42:22 really? you're incredible Feb 17 21:42:34 Actually, warm down, i'll go sleeping after that :) Feb 17 21:42:41 :-) Feb 17 21:45:55 the obvious use case is to use the n900 in the dark without light... Feb 17 21:46:37 PaulFertser: if you're near it: GPIO7/PWM1 is also interesting (see backscroll some 50 lines) Feb 17 21:47:18 might be actually mad handy for POST in kernel/bootloader Feb 17 21:48:14 there are also some gpios for boot config Feb 17 21:48:25 call me Schliemann - ancient Nokians invented it, I dug it out Feb 17 21:48:30 that can be set with devmem2 or 0xFFFF Feb 17 21:48:39 not really Feb 17 21:48:43 ah? Feb 17 21:48:57 what are they then? Feb 17 21:48:58 the GPIO for bootcontrol are hardwired Feb 17 21:49:08 really? Feb 17 21:49:26 I mean the stuff that put the bootrom order Feb 17 21:49:46 like microsd, flash,usb Feb 17 21:50:05 that is looked for by the bootrom Feb 17 21:50:36 * GNUtoo|htcdream is really physically tired...and might say confusing stuff Feb 17 21:52:36 I actually can't even find the pins right now Feb 17 21:53:01 anyway setting these GPIO wouldn't make any sense, as they are sampled on power-on Feb 17 21:54:08 and what you're talking about are flags in CAL that xloader is looking at to determine where to look for NOLO, afaik Feb 17 21:55:00 what about "bootloader project" page at omappedia? Feb 17 21:55:12 rom BL is checking all soures in sequence, with a rather short timeout Feb 17 21:55:19 yes Feb 17 21:55:31 but the sequence can be changed Feb 17 21:55:47 yes, by hardwired GPIO Feb 17 21:55:55 ok Feb 17 21:56:19 that can be set? Feb 17 21:56:24 or not? Feb 17 21:56:29 by devmem2? Feb 17 21:56:38 like in that wiki page, Feb 17 21:56:41 it's hardwired, I.E. you need soldering iron to change it Feb 17 21:56:46 ah ok Feb 17 21:56:52 strange... Feb 17 21:57:38 GNUtoo|htcdream: mail with udev fix sent Feb 17 21:57:40 why that devmem location changed then when I changed the order with 0xFFFF Feb 17 21:57:49 mrmoku,ok Feb 17 21:57:58 wtf is devmem? Feb 17 21:58:28 a tool to change memory content...from userspace Feb 17 21:58:43 DocScrutinizer: maemo kernel seem to be not using pwm devices at all, checking upstream now. Feb 17 21:59:08 NFC, but what you can change with flasher and 0xffff are the CAL records Feb 17 21:59:14 ok Feb 17 21:59:24 what are theses? Feb 17 21:59:28 PaulFertser: expected. Opens option to improve things Feb 17 21:59:47 DocScrutinizer: btw, there's a ready-to-use driver to control backlight with any pwm device. Feb 17 22:00:06 What is CAL http://talk.maemo.org/showthread.php?t=20465 Feb 17 22:00:40 PaulFertser: \o/ I guess it's not used in maemo either Feb 17 22:01:07 DocScrutinizer: not used for anything, but present in the kernel sources. Feb 17 22:01:15 :-) Feb 17 22:01:24 is it configured? Feb 17 22:01:27 M? Feb 17 22:01:56 No, and it needs platform data anyway which is lacking. Feb 17 22:02:05 I see Feb 17 22:03:26 ahhh ok /dev/mtd1 Feb 17 22:05:06 ooooh OOOOPSIE, N4200 is GAIA as I said above. So quite probably those PWM stuff and GPIO6 aren't to be found in kernel directly Feb 17 22:05:46 I'm also not really fresh anymore Feb 17 22:05:48 DocScrutinizer: upstream kernel has arch/arm/mach-omap2/board-rx51-peripherals.c: gpio_request(gpio + 6, "backlight_pwm"); but it doesn't seem to be used anyhow. Feb 17 22:06:31 sounds like it's exactly our GPIO6/PWM0 Feb 17 22:06:44 of twl5030 GAIA Feb 17 22:07:16 GNUtoo|htcdream: duuh... forgot that it is subscriber only... and sent with the wrong address :/ Feb 17 22:07:19 (schematics p4 if you wonder) Feb 17 22:14:24 mrmoku? Feb 17 22:14:48 you're not suscribed? Feb 17 22:14:59 to oe-dev? Feb 17 22:16:14 DocScrutinizer: looks like even upstream lacks proper pwm support Feb 17 22:17:20 meh, I closed the 15MB pdf for twl5030 some 90min ago :-S Feb 17 22:18:40 DocScrutinizer: i mean it lacks pwm support for omap altogether, so to utilize it one has to write the pwm driver first it seems. Feb 17 22:18:52 ah ok Feb 17 22:19:00 PaulFertser: yes, got that Feb 17 22:19:26 NB this is a function of twl5030 Feb 17 22:19:32 what else is connected to PWM? Feb 17 22:19:40 lemme check Feb 17 22:19:59 Ah, yes, you've told it already, connected to GAIA Feb 17 22:20:18 IR led is on CPU PWM Feb 17 22:20:27 what is GAIA already, I saw it many times Feb 17 22:20:37 that'S it Feb 17 22:20:45 the PMU Feb 17 22:20:59 and what abuot twl4030 then? Feb 17 22:21:00 multipurpose, Feb 17 22:21:10 5030==4030 Feb 17 22:21:11 there are 2 pmu? Feb 17 22:21:23 ok Feb 17 22:27:18 [2011-02-17 23:26:54] pandora uses gaia pwm for leds, so there's plenty of kernel code avail Feb 17 22:27:36 Well, probably i've missed something. Feb 17 22:27:43 Or looking at the wrong tree. Feb 17 22:28:03 Anyway, time to sleep, have a nice evening, see you folks. Feb 17 22:28:34 Too many trees around... Feb 17 22:34:24 :-D Feb 17 22:34:52 night paul, dream nice. Here's a candy: leds-twl4030-pwm.c Feb 17 22:36:01 http://bugs.openpandora.org/index.php?getfile=43 Feb 17 22:37:29 beep won't fly, that's actually _ShutDownLeft & _ShutDownRight inputs of amp Feb 17 22:37:37 (GPIO7) Feb 17 22:38:13 well maybe it works nevertheless Feb 17 22:38:24 if input has some DC component Feb 17 22:38:40 after all that'S a class D amp Feb 17 22:39:49 meh, but with 100nF series in input lines Feb 17 22:42:06 Nokia EE is too good, no glitch here. We would love to see one ;-D Feb 17 22:45:19 absolutely correct design - DC-blocked RF-filtered inputs, class-D amp, no output capacitors (esp no 1uF ;-P) Feb 17 22:45:34 yet they managed to smoke up their speakers X-D Feb 17 22:46:28 probably the PCB circuit wasn't designed for the ID which had to select crappy transducers :-) Feb 17 22:48:19 or maybe even transducers would be ok with that input power, if only they had a proper case design around them (closed volume resonance, or transmission line, or exponential acoustic transformer) Feb 17 22:49:04 so it's once more the case design screwing up everything - LOL Feb 18 02:29:32 * pespin writing from archlinux on exopc slate, touchscreen is running and illume works quite well :D **** ENDING LOGGING AT Fri Feb 18 02:59:57 2011