**** BEGIN LOGGING AT Sat Dec 11 02:59:57 2010 Dec 11 09:59:39 JaMa: did you rebuild ffphonelog? Dec 11 10:27:29 y Dec 11 11:00:43 alexxy: y Dec 11 11:21:34 JaMa: where? :) Dec 11 11:21:55 ahh ok Dec 11 11:22:02 found it in feed Dec 11 11:31:41 Has someone asked E about the autostart issue? Dec 11 11:35:49 * JaMa not Dec 11 12:02:47 JaMa: you've already synced shr-t... Dec 11 12:03:02 my plan was to wait until it is working Dec 11 12:05:45 hmm Dec 11 12:05:55 any feedback wrt. the hanging bug? Dec 11 12:06:23 mickeyl: it seems that it is working Dec 11 12:06:33 excellent. then please close it Dec 11 12:06:56 ok Dec 11 12:11:56 here comes the autoconf noob... Dec 11 12:12:09 how can i check whether a certain function is present in a certain header? Dec 11 12:15:49 Heinervdm: sorry about that and it isn't working? Dec 11 12:16:05 JaMa: not fully Dec 11 12:16:15 because of this startup thing Dec 11 12:16:22 it works only sometimes :) Dec 11 12:16:42 this boot it worked Dec 11 12:16:59 I will write an announcement to the ML Dec 11 12:17:09 Heinervdm: old images and feeds are available in shr-testing.old Dec 11 12:17:15 ok Dec 11 12:17:18 I can move it back if you want Dec 11 12:17:37 i'm not sure Dec 11 12:17:45 your call.. Dec 11 12:17:52 now some have already upgraded Dec 11 12:18:12 but perhaps it's better to move it back Dec 11 12:18:29 if it's only about that statup thing then you can try to put short sleep there Dec 11 12:18:36 as we had for starting from xsession Dec 11 12:18:54 that should make "sometimes working" much more often Dec 11 12:19:09 yes Dec 11 12:20:08 My plan for shr-t is now: ask for testers then release a rc2 and perhaps a rc3 and then in january 2011 release the final version Dec 11 12:20:39 but from svn log I haven't noticed changes related to auto start.. so if it's some kind of rece than maybe we're just starting faster due to efreet changes Dec 11 12:21:03 yes i looked at them too Dec 11 12:21:22 last real change to startup is 4 months ago Dec 11 12:24:07 mickeyl, hi Dec 11 12:24:12 good morning GNUtoo|laptop Dec 11 12:24:32 mickeyl, why did you implement battery reading trough i2c? Dec 11 12:24:44 why not trying or rebasing my branch Dec 11 12:24:48 it was tested a bit Dec 11 12:24:54 but not extensively Dec 11 12:25:06 and modifying the charging script Dec 11 12:25:12 to read the value via sysfs Dec 11 12:25:34 so you have a kernel driver? Dec 11 12:25:40 of course Dec 11 12:25:47 I just improved the existing one Dec 11 12:25:49 i was not really aware of that Dec 11 12:25:51 look at the branches Dec 11 12:26:44 http://git.freesmartphone.org/?p=linux-2.6.git;a=shortlog;h=refs/heads/nokia900/battery-ultra_work_in_progress Dec 11 12:27:16 maybe the patches apply cleanly Dec 11 12:27:27 on top of master Dec 11 12:27:31 it's worth a try Dec 11 12:28:18 is this also enabling charging or just for reading? Dec 11 12:28:32 just readings Dec 11 12:28:56 freesmartphone.org: 03morphis 07msmcomm * r882eaadd5a42 10/ (8 files in 6 dirs): msmcomm: implement handling of network service status Dec 11 12:29:04 i see Dec 11 12:29:14 in that case we need a plugin anyways, right? Dec 11 12:29:14 charging is another chip Dec 11 12:29:26 yes unless we implement it in kernel Dec 11 12:29:31 right Dec 11 12:29:46 Heinervdm: so if you wrote the announcement to ML... I can move old back and you can ask those who upgraded and are happy with it to switch to tests/shr-testing and became testers Dec 11 12:29:55 freesmartphone.org: 03morphis 07cornucopia * r1555cbdf3b83 10/fsogsmd/src/plugins/modem_qualcomm_palm/ (data.vala mediators_network.vala unsolicited.vala utils.vala): fsogsmd: modem_qualcomm_palm: implement handling of network service/registration status the right way Dec 11 12:30:04 JaMa: ok will do so Dec 11 12:30:05 freesmartphone.org: 03morphis 07cornucopia * rd2b63abd5d1a 10/ (9 files in 7 dirs): Merge branch 'master' of ssh://git.freesmartphone.org/cornucopia Dec 11 12:30:29 ok then moving Dec 11 12:30:59 freesmartphone.org: 03morphis 07cornucopia * r738374943d27 10/fsodeviced/src/plugins/router_palmpre/plugin.vala: fsodeviced: router_palmpre: load even default scenario on boot Dec 11 12:32:29 so we are very close to a functional image for the palm pre which lets you make calls via the SHR UI !!! Dec 11 12:32:38 awesome progress Dec 11 12:33:46 :) Dec 11 12:33:58 also note that: Dec 11 12:34:02 *wifi works Dec 11 12:34:05 the problem with the service/registration status is even fixed Dec 11 12:34:07 at least can work Dec 11 12:34:14 I read the wrong value from the packet :D Dec 11 12:34:19 ah Dec 11 12:34:26 now the network registration status is always correct Dec 11 12:34:31 *palm-pre is avaliable in 3 different versions(pre,pre-plus,pre-2) Dec 11 12:34:53 Heinervdm: btw that shr_elm_softkey issue is probably a bit different than what we had before.. because before it was running and only not reacting to clicks, now I got it and it wasn't running at all Dec 11 12:34:58 which can make it cheaper(n900 isn't cheap) Dec 11 12:35:05 GNUtoo|laptop: and there will be more next year Dec 11 12:35:19 Palm annouced to release a new device every two months next year Dec 11 12:35:26 (for instance if you buy the original pre, it won't be that expensive) Dec 11 12:35:35 and maybe they are really close to what the pre is today Dec 11 12:35:49 JaMa: yes this time it's different to last time Dec 11 12:36:45 "which can make it cheaper(n900 isn't cheap)" correct! Dec 11 12:37:27 and we have now a simple utility to flash images/kernels to the device Dec 11 12:37:55 JaMa: will you build the latest full image too? And delete the older image from 7.12.? Dec 11 12:38:19 JaMa: ah it's already build Dec 11 12:38:30 but can you delete the older ones? Dec 11 12:44:41 yes Dec 11 12:48:32 Ok, have to go now and will try the startup workaround tomorow Dec 11 12:49:53 Heinervdm: "An upgrade from older SHR-testing seems to be broken" should be fixed Dec 11 12:51:18 JaMa: yes, i haven't tried that till now Dec 11 12:51:29 freesmartphone.org: 03mickey 07cornucopia * rbd06a2c7739a 10/fsodeviced/configure.ac: fsodeviced: we depend on i2c-tools now Dec 11 12:51:43 but it seems to be like that because of the reports in the last days Dec 11 12:53:06 those problems of GRiera in "[Shr-User] Problems upgrading testing" were 1) probably broken status file, 2) too old opkg (opkg update; opkg upgrade opkg; opkg upgrade; would be much better way to upgrade) 2) SRCPV - I've fixed this one Dec 11 12:53:56 I will look into this tomotrow :) Dec 11 12:54:04 ok thanks Dec 11 12:54:28 GNUtoo|laptop: could you get me your code as a new kernel module? Dec 11 12:54:35 or do we compile that built-in atm.? Dec 11 12:55:01 just add CONFIG_BQsomething as module Dec 11 12:55:16 hmm Dec 11 12:55:16 my code is just an improvement on an existing driver Dec 11 12:55:22 which is not enabled by defualt Dec 11 12:55:22 i don't have the environment setup Dec 11 12:55:30 ah you want the module? Dec 11 12:55:35 yeah Dec 11 12:55:36 for a quick test Dec 11 12:55:41 ok Dec 11 12:55:43 i don't want to reflash again Dec 11 12:56:02 so you don't have openembedded right now? Dec 11 12:56:15 yeah, not locally Dec 11 12:56:16 can build only remote Dec 11 12:56:20 and that always is a bit annoying Dec 11 12:56:23 ok Dec 11 12:56:31 I can try but I should be working my maths Dec 11 12:56:43 don't worry then Dec 11 12:56:47 perhaps tomorrow or so Dec 11 12:56:49 nothing urgent Dec 11 12:57:06 need to revive my pre as well ;) Dec 11 12:57:12 will you be there this afternoon or in some hours Dec 11 12:57:30 ya, that's likely Dec 11 12:57:41 not much on the agenda for today until 19:00 or so Dec 11 12:57:50 just a short visit for buying a new pullover Dec 11 12:57:50 ok Dec 11 12:58:55 mickeyl: if you are ready to get your pre running again, ping me Dec 11 12:59:33 morphis_: will do, thanks Dec 11 12:59:45 ok Dec 11 13:00:44 ha. my first i2c plugin works Dec 11 13:00:52 1970-01-05T14:04:28.000335Z [DEBUG] PowerSupplyN900 : Reading capacity from i2c... Dec 11 13:00:53 1970-01-05T14:04:28.002075Z [DEBUG] PowerSupplyN900 : i2c reports capacity as 88 Dec 11 13:01:37 battery is draining massively here Dec 11 13:02:01 it was about 30 minutes from 99% to 88% Dec 11 13:02:52 mickeyl, ouch check cpufreq Dec 11 13:02:59 and htop Dec 11 13:08:21 mickeyl, I'll compile it, you test and then I push the changes Dec 11 13:12:03 cool, thanks Dec 11 13:12:09 hmm, where is a microUSB cable when you need one Dec 11 13:12:10 *sigh* Dec 11 13:14:14 mickeyl, http://gnutoo.homelinux.org/mickeyl/modules/nokia900_all_modules_with_battery_reading_rebased.tar.bz2 Dec 11 13:14:24 thanks! Dec 11 13:14:30 mickeyl, huge: about 10M Dec 11 13:14:43 s/10M/17M Dec 11 13:14:59 but it contain where to put the modules and the possibles dependencies Dec 11 13:15:25 should I give you only some modules? Dec 11 13:15:34 like bq* and power_supply class? Dec 11 13:15:44 so you insmod them? Dec 11 13:16:55 mickeyl, ^^^ Dec 11 13:18:30 mickeyl, done I've put the 2 modules there Dec 11 13:18:45 106k Dec 11 13:18:53 and 195k Dec 11 13:19:04 ya, still downloading the huge thing Dec 11 13:19:13 try the 2 modules then Dec 11 13:19:16 let me get the smaller ones Dec 11 13:20:29 ok, insmodding Dec 11 13:20:38 I've put the patches there too Dec 11 13:20:42 so you can get source code Dec 11 13:20:49 awesome, looks good Dec 11 13:21:06 so the plan is to use kernel26_powersupply for reading battery Dec 11 13:21:14 yes Dec 11 13:21:17 sounds a lot better Dec 11 13:21:22 and add the powersupply_n900 plugin for controlling charging... until anyone wrote a kernel driver for that as well Dec 11 13:21:28 yes Dec 11 13:21:35 ok, good plan. we'll make it so Dec 11 13:21:56 btw did you check at what frequency was the n900 running Dec 11 13:22:01 is it running at 1.2GHz? Dec 11 13:22:02 500 Dec 11 13:22:04 ah ok Dec 11 13:22:15 mrmoku did an ondemand plugin Dec 11 13:22:20 which limits to 600 Dec 11 13:22:21 perhaps some RF is on-by-default Dec 11 13:22:29 I don't know Dec 11 13:22:36 battery is much longer for me Dec 11 13:22:45 but some sensors are on by default Dec 11 13:22:49 and the display is still on Dec 11 13:22:56 let me try FBIOBLANK Dec 11 13:23:05 :) Dec 11 13:24:19 ah, excellent Dec 11 13:24:19 works Dec 11 13:24:48 indeed the n900 is not so bad Dec 11 13:24:57 a lot of stuff works or can easily be fixed Dec 11 13:26:30 yep Dec 11 13:32:59 i don't understand why these crazy palm developers put the backlight on to 100% when you're a) charging and b) in recovery mode Dec 11 13:33:02 it hurts my eyes Dec 11 13:33:12 and i'm sure it's not good for the display either Dec 11 13:34:59 :) Dec 11 13:39:15 hmm Dec 11 13:39:31 root@saphir:~# novaterm Dec 11 13:39:31 unable to find device Dec 11 13:40:11 reconnect the device Dec 11 13:40:26 are you sure you are in recovery mode? Dec 11 13:40:27 already tried Dec 11 13:40:29 yes Dec 11 13:40:45 lsub shows Dec 11 13:40:51 Bus 001 Device 011: ID 0830:8001 Palm, Inc. Dec 11 13:40:56 and the device shows the USB icon Dec 11 13:41:04 and hurts my eyes... but we already mentioned that Dec 11 13:41:09 hehe Dec 11 13:41:18 novacomd is running in the background? Dec 11 13:41:19 novacomd runs Dec 11 13:41:23 ok Dec 11 13:41:29 let me restart it Dec 11 13:42:15 hmm, nothing Dec 11 13:42:27 does the OS on the device need to match the SDK? Dec 11 13:42:37 i just downloaded the latest SDK, but it's very possible the Palm is one or two versions behind Dec 11 13:43:49 no Dec 11 13:43:55 hmm, no go Dec 11 13:44:10 found a bug in ogsmd :) Dec 11 13:44:11 hm thats misterious Dec 11 13:44:16 let me try the laptop Dec 11 13:44:28 mickeyl: do you have the webos-doctor on your computer? Dec 11 13:44:37 no Dec 11 13:45:44 ok, works on the laptop Dec 11 13:45:47 let me try something Dec 11 13:48:52 why are we still having problems with reboot / shutdown, btw? Dec 11 13:48:53 what is missing? Dec 11 13:50:00 you mean on the pre? Dec 11 13:50:09 http://paste.debian.net/102131 Dec 11 13:50:11 interestingly i mean both the pre and the n900 Dec 11 13:50:15 it is broken for both Dec 11 13:50:28 ah ok Dec 11 13:50:38 on the pre it just needs some work Dec 11 13:51:02 (it is snowing heavily and i am stuck in traffic so i can not debug this right now) Dec 11 13:51:11 mickeyl: we already have tellbootie http://git.freesmartphone.org/?p=utilities.git;a=tree;f=palmpre/tellbootie;h=03fd7436185f8bb2eea05dfb6fa998789cad41f2;hb=HEAD Dec 11 13:51:21 which does the same as tellbootie on palm pre Dec 11 13:51:29 # tellbootie (reboot) Dec 11 13:51:41 # tellbootie recover (reboot in recover mode) Dec 11 13:52:01 it just have some bugs and is not integrated in the init-scripts Dec 11 13:52:10 and std. linux shutdown / reboot does not work Dec 11 13:52:11 on n900 shutdown works if you remove the usb cable Dec 11 13:52:14 why I don' know Dec 11 13:52:41 and tellbootie just calls reboot(...) Dec 11 13:53:47 GNUtoo|laptop: d'oh Dec 11 13:54:39 lindi-: no idea at first glance, it's been years ago i have written that :/ Dec 11 13:55:35 morphis_: should i unpack the latest image? Dec 11 13:55:49 does novaterm work now? Dec 11 13:56:11 on the latop yes, i resurrected my init process and can boot into webos again now Dec 11 13:56:18 ah ok Dec 11 13:56:24 then use the fso-installer Dec 11 13:56:30 but I have to update it first Dec 11 13:56:34 ok Dec 11 13:56:34 wait a moment Dec 11 13:56:38 yep Dec 11 13:57:24 freesmartphone.org: 03morphis 07utilities * r5345bd0dfdad 10/palmpre/fso-installer/Makefile: palmpre: fso-installer: cleanup and add flash-kernel target Dec 11 13:57:28 so Dec 11 13:57:45 do you already downloaded the webos-doctor? Dec 11 13:57:56 hmm Dec 11 13:58:02 let me check whether i have so much space available Dec 11 13:58:03 will give you the link Dec 11 13:58:06 mom Dec 11 13:58:27 its http://palm.cdnetworks.net/rom/pre/p145r0d06302010/eudep145rod/webosdoctorp100ueu-wr.jar Dec 11 13:58:50 http://palm.cdnetworks.net/rom/pre/p145r0d06302010/eudep145rod/webosdoctorp100ueu-wr.jar Dec 11 13:59:00 downloading... Dec 11 13:59:05 then you have to put it into downloads/webosdoctorp100ueu-wr.jar in the folder where the Makefile is Dec 11 13:59:29 so you have Dec 11 13:59:31 drwxrwxr-x 2 morphis morphis 4,0K 2010-12-11 13:44 downloads/ Dec 11 13:59:31 -rw-rw-r-- 1 morphis morphis 4,8K 2010-12-10 21:18 Makefile Dec 11 13:59:48 morphis@nirvana ...ce/fso/utilities/palmpre/fso-installer (git)-[master] % ls -alh downloads Dec 11 13:59:48 insgesamt 259M Dec 11 13:59:48 drwxrwxr-x 2 morphis morphis 4,0K 2010-12-11 13:44 ./ Dec 11 13:59:48 drwxrwxr-x 5 morphis morphis 4,0K 2010-12-11 14:58 ../ Dec 11 13:59:48 -rw-rw-r-- 1 morphis morphis 48M 2010-12-11 13:42 lite-palmpre.tar.gz Dec 11 13:59:50 -rw-rw-r-- 1 morphis morphis 2,3M 2010-12-02 09:45 uImage-2.6.24+gitr0+33239e4074f4716a674b450a443b64add201bf39-r91-palmpre.bin Dec 11 13:59:53 -rw-rw-r-- 1 morphis morphis 2,3M 2010-12-02 09:45 uImage-palmpre.bin Dec 11 13:59:55 -rw-rw-r-- 1 morphis morphis 207M 2010-12-09 09:49 webosdoctorp100ueu-wr.jar Dec 11 14:00:12 mickeyl: do already have a LVM partition for FSO? Dec 11 14:00:39 yes, slyon did that for me on the workshop Dec 11 14:00:43 i have SHR on the device Dec 11 14:00:45 and can boot into Dec 11 14:01:29 ah ok Dec 11 14:01:50 then FSO is on /dev/mapper/store-fso right? Dec 11 14:01:56 i hope so Dec 11 14:02:35 freesmartphone.org: 03morphis 07utilities * r9fd5ca98c4fc 10/palmpre/read_tokens/src/read_tokens.c: palmpre: read_tokens: disable debug mode Dec 11 14:02:43 ok Dec 11 14:03:22 we should teach the makefile download that on its own... Dec 11 14:03:40 jepp Dec 11 14:03:52 will do that as next step Dec 11 14:04:13 it even needs some more love Dec 11 14:04:22 + error handling Dec 11 14:04:27 but it works for now Dec 11 14:04:59 ok, i have downloaded everything to downloads Dec 11 14:05:00 what now? Dec 11 14:05:25 you are in recovery mode? Dec 11 14:05:26 i'm booted into webos with usb mass storage chosen Dec 11 14:05:33 and put the doctor in the download/ folder Dec 11 14:05:44 go into recovery mode Dec 11 14:05:45 one moment Dec 11 14:06:01 in recovery mode Dec 11 14:10:50 ik Dec 11 14:10:54 ah ok Dec 11 14:10:58 type 'make memload' Dec 11 14:11:16 it should now extract the doctor and boot the nova installer uImage with attached rootfs Dec 11 14:11:29 so it have a base env for any further commands Dec 11 14:11:47 ok, it#s operating... Dec 11 14:11:50 ok Dec 11 14:12:16 Wait for device to come back ... Dec 11 14:12:26 and i'm logged in Dec 11 14:13:46 log out Dec 11 14:13:55 k Dec 11 14:13:57 and type 'make mount-fso' Dec 11 14:14:10 done Dec 11 14:14:11 log in and check if something is mount on /tmp/fso Dec 11 14:14:35 s/mount/mounted/ Dec 11 14:14:35 morphis_ meant: log in and check if something is mounted on /tmp/fso Dec 11 14:14:41 yep, there's ext3 with the old SHR Dec 11 14:15:03 ok Dec 11 14:15:07 then logout Dec 11 14:15:14 and downlaod the new image from http://amethyst.openembedded.net/~morphis/oe/palmpre/deploy/images/palmpre/lite-palmpre.tar.gz Dec 11 14:15:20 put it into downloads/ Dec 11 14:15:38 and then type make IMAGE=downloads/lite-palmpre.tar.gz flash-image Dec 11 14:15:46 that will flash the image to the device Dec 11 14:16:01 and erase the one which was there before Dec 11 14:16:08 but maybe you want to do a Dec 11 14:16:13 make backup first Dec 11 14:16:18 which backups all webos stuff Dec 11 14:16:23 it does something Dec 11 14:16:24 with the tokens etc. Dec 11 14:16:26 ok Dec 11 14:16:26 however Dec 11 14:16:29 [: 3: downloads/lite-palmpre.tar.gz: unexpected operator Dec 11 14:17:48 how long should that take? Dec 11 14:17:54 hm Dec 11 14:17:59 it tooks some time Dec 11 14:18:16 is the image in the downloads folder? Dec 11 14:19:27 yes Dec 11 14:19:37 ok Dec 11 14:19:43 there's no prompt, so i guess it still unpacks Dec 11 14:19:59 can you give the whole output? Dec 11 14:20:46 http://pastebin.ca/2016459 Dec 11 14:20:49 and there it hangs atm. Dec 11 14:20:54 ok Dec 11 14:21:03 ok Dec 11 14:21:14 it unpacks the tar.gz on the device Dec 11 14:21:16 to check it Dec 11 14:21:21 you can login on another term Dec 11 14:21:26 and ls /tmp/fso Dec 11 14:21:32 if there is no /dev or /usr already Dec 11 14:21:32 ah, prompt is back again Dec 11 14:21:36 ah Dec 11 14:22:12 login and check /tmp/fso Dec 11 14:22:18 looks vaguely ok Dec 11 14:22:22 ok Dec 11 14:22:30 to proof that everything is right Dec 11 14:22:33 rm -rf /tmp/fso* Dec 11 14:22:50 and redo make IMAGE=downloads/lite-palmpre.tar.gz flash-image Dec 11 14:22:53 argh Dec 11 14:22:55 :D Dec 11 14:23:47 same output (even the error is the same) Dec 11 14:23:52 ok Dec 11 14:24:00 never saw that error before Dec 11 14:24:41 i'm using ubuntu with dash Dec 11 14:24:44 perhaps a dash/bash problem Dec 11 14:25:17 ok Dec 11 14:27:56 ok, 2nd run through Dec 11 14:28:02 tmp/fso looks vaguely ok Dec 11 14:28:38 ok Dec 11 14:28:49 then type make reboot Dec 11 14:29:00 it should shr now Dec 11 14:30:15 boots fine, but into webos Dec 11 14:30:24 hm Dec 11 14:30:39 if you are in webos Dec 11 14:30:50 type make recover Dec 11 14:30:59 then you are back in recovery mode Dec 11 14:31:01 or do Dec 11 14:31:05 make memboot Dec 11 14:31:16 which reboots and loads the webos installer Dec 11 14:31:33 flash-kernel is not necessary? Dec 11 14:31:49 how should the init process chose fso? Dec 11 14:31:50 depends in which state your device is Dec 11 14:32:04 i.e. what is the expected result of what i did? Dec 11 14:32:07 you booted webos with the webos kernel or the fso one? Dec 11 14:32:13 no idea Dec 11 14:32:15 to boot shr Dec 11 14:32:16 ok Dec 11 14:32:30 all i can tell is that i'm in webosnow Dec 11 14:32:31 it's possible I help you over ssh? Dec 11 14:33:00 anyway Dec 11 14:33:01 which part of the process should have chose SHR as the default boot? Dec 11 14:33:04 and how? Dec 11 14:33:21 the boot process looks like this: Dec 11 14:33:39 bootie choose the kernel from a symlink on /dev/mmcblk0p1 which is /boot in webos Dec 11 14:33:47 and boots /boot/uImage Dec 11 14:33:58 the uImage then executed /boot/sbin/init Dec 11 14:34:09 i see Dec 11 14:34:15 the symlink still points to the original kernel Dec 11 14:34:19 i see Uimage-palmpre.bin Dec 11 14:34:22 but it's not linked to Dec 11 14:35:11 that's why i suspected i need the flash-kernel target Dec 11 14:35:18 since this obviously creates a link Dec 11 14:35:26 no? Dec 11 14:35:32 jepp Dec 11 14:35:43 make IMAGE=uImage-yout-want flash-kernel Dec 11 14:36:05 then you have to fix the sbin/init on /dev/mmcblk0p1 Dec 11 14:36:11 äh mmcblk0p2 Dec 11 14:37:36 mom Dec 11 14:37:57 so to change between webos and shr i always need to boot into "the other one" and edit a) the link to the kernel and b) the init script? Dec 11 14:39:17 or into the memload image Dec 11 14:39:32 playya__: how's the state of the boot menu? Dec 11 14:39:37 to change between webos and shr you only have to fix the init script Dec 11 14:39:44 mickeyl: it depends on the non working kexec Dec 11 14:39:58 and as long as kexec is not working the boot menu does not work Dec 11 14:39:59 ok, so we can boot webos with the SHR kernel? Dec 11 14:40:03 jepp Dec 11 14:40:07 ah Dec 11 14:40:15 but then you for example don't have wifi Dec 11 14:40:15 ok, letme link the shr kernel Dec 11 14:41:12 http://pastie.org/1367909 Dec 11 14:41:22 this is how the sbin/init should look for SHR/FSO boot Dec 11 14:41:38 relevant are the following two lines: Dec 11 14:41:39 lvm.static lvchange -ay --ignorelockingfailure /dev/mapper/store-fso Dec 11 14:41:40 and Dec 11 14:41:49 mount -o rw /dev/mapper/store-fso /realroot Dec 11 14:41:56 which are line 50 and 54 Dec 11 14:42:17 maybe we should even find a automatic solution for this or disable multi boot solution as lon as kexec is not working Dec 11 14:43:57 hmm Dec 11 14:44:06 i'm a bit confused Dec 11 14:44:07 i have init Dec 11 14:44:09 init.old Dec 11 14:44:10 init.fso Dec 11 14:44:16 too manyinits Dec 11 14:44:25 hehe Dec 11 14:44:35 where in /boot/sbin/ ? Dec 11 14:44:40 yes Dec 11 14:44:48 then remove them and take mine Dec 11 14:44:57 as the only one Dec 11 14:45:07 as I think you don't need webos at moment, right? Dec 11 14:45:16 yes Dec 11 14:46:44 ok, so before i reboot lets check Dec 11 14:46:52 i have linked uImage to uImage-palmpre.bin Dec 11 14:47:00 and i have your init in /boot/sbin/init Dec 11 14:47:04 is that it? Dec 11 14:47:15 jepp Dec 11 14:47:18 ok, reboting now Dec 11 14:47:48 into SHR? Dec 11 14:48:23 no Dec 11 14:48:26 kernel hangs during bootr Dec 11 14:48:43 i got 6 times "modem wakeup timeout" Dec 11 14:48:49 and last message is Dec 11 14:48:55 ---------Full40 = 3600 Dec 11 14:49:00 stuck now Dec 11 14:49:41 i guess that kernel is too old Dec 11 14:50:09 you downloaded it form http://amethyst.openembedded.net/~morphis/oe/palmpre/deploy/images/palmpre/ ? Dec 11 14:50:27 no i'm afraid it was still the one slyon put me on Dec 11 14:50:27 the messages does not mean anything to you Dec 11 14:50:35 hm Dec 11 14:50:45 let me use make flash kernel now Dec 11 14:50:50 ok Dec 11 14:51:06 fucking shit this all Dec 11 14:51:15 hmm, make memload should check whether the jar is already unpacked Dec 11 14:51:22 I want to have real boot loader we can use right ... Dec 11 14:51:24 hehe, indeed. much too complicated, but what shall we do... Dec 11 14:51:39 (memload) jepp, thats a bug :) Dec 11 14:52:09 x-load is already loading on the device but we don't get a serial connection to it Dec 11 14:52:11 so make IMAGE=downloads/uImage-blabla flash-kenrel? Dec 11 14:52:51 jepp Dec 11 14:52:52 done Dec 11 14:52:55 ok Dec 11 14:52:58 next try :) Dec 11 14:53:28 I hope it works now Dec 11 14:53:39 2338936 Jan 1 00:01 uImage-palmpre.bin Dec 11 14:53:40 is that it? Dec 11 14:53:54 where? Dec 11 14:54:03 uImage-palmpre.bin 02-Dec-2010 19:45 2.2M Dec 11 14:54:11 http://amethyst.openembedded.net/~morphis/oe/palmpre/deploy/images/palmpre/ Dec 11 14:55:10 ok,rebooting Dec 11 14:55:27 nope Dec 11 14:55:30 hangs at the same point Dec 11 14:56:08 it mounts rootfs Dec 11 14:56:11 and launches init Dec 11 14:56:16 but not much more Dec 11 14:56:52 hm Dec 11 14:56:52 hello Dec 11 14:56:56 Hilfe anzeigen Dec 11 14:57:11 mickeyl: hm Dec 11 14:57:24 then something in the init process is broken Dec 11 14:57:27 yep Dec 11 14:57:52 I really should make an installer which does everything for you Dec 11 14:57:57 and removes webos Dec 11 14:58:10 so we have no side effects anymore Dec 11 14:58:11 morphis_ Hilfe anzeigen Dec 11 14:58:26 philippe: ? Dec 11 14:58:54 IRC helfen Dec 11 14:59:29 try http://www.macobserver.com/tip/2002/04/04.1.shtml Dec 11 14:59:42 helfen soll, die Pseudo-Lounge Dec 11 15:00:22 Ich bin mit einem französisch zeigen Dec 11 15:06:49 morphis_: fwiw, it stillboots into shr using the "manual" way Dec 11 15:06:51 setenv bootfile uImage-palmpre.bin Dec 11 15:06:51 setenv bootargs "root=/dev/mmcblk0p2 rootwait ro init=/sbin/init.fso" Dec 11 15:06:51 printenv Dec 11 15:06:51 fsboot Dec 11 15:07:08 ah ok Dec 11 15:07:18 so you are now in SHR? Dec 11 15:07:33 yes Dec 11 15:07:48 ok Dec 11 15:07:51 it hangs on default boot, but works via manual commands Dec 11 15:07:53 UI is running? Dec 11 15:07:55 yes Dec 11 15:08:00 PIN dialog is there? Dec 11 15:08:02 no Dec 11 15:08:16 will maybe take some seconds Dec 11 15:11:17 hmm no Dec 11 15:11:25 i'll relaunch fsogsmd Dec 11 15:13:07 hmm, multiple problems Dec 11 15:13:18 2010-12-11T13:48:55.884765Z [WARN] subsystem : Could not load plugin: Could not load /usr/lib/cornucopia/modules/fsogsm/lowlevel_palmpre.so: /usr/lib/cornucopia/modules/fsogsm/lowlevel_palmpre.so: cannot open shared object file: No such file or directory Dec 11 15:13:33 ah? Dec 11 15:13:36 indeed that file is missing Dec 11 15:13:37 ah ok Dec 11 15:13:43 it's not used anymore Dec 11 15:13:50 ok Dec 11 15:13:52 there is no need for a lowlevel plugin in fsogsmd Dec 11 15:13:58 just remove it from the config file Dec 11 15:14:02 I though I already did that Dec 11 15:14:03 but then Dec 11 15:14:05 2010-12-11T13:48:55.944885Z [CRITICAL] subsystem : Can't acquire service name org.freesmartphone.ogsmd; service already running or not allowed in dbus configuration. Dec 11 15:14:08 which is strang Dec 11 15:14:09 e Dec 11 15:15:12 hm Dec 11 15:15:49 probably shr autoloading or something Dec 11 15:15:52 killing X now Dec 11 15:16:06 hmm, no mdbus2 on rootfs Dec 11 15:16:08 tsss tss Dec 11 15:16:19 :) Dec 11 15:16:25 shr default ... Dec 11 15:17:23 I have the same image on my device and there I don't get such a message Dec 11 15:17:26 maybe shr-light-image default Dec 11 15:17:32 but shr-image should have it right? Dec 11 15:17:52 ah right Dec 11 15:17:56 it's shr-lite-image Dec 11 15:18:38 hmm Dec 11 15:18:44 2010-12-11T13:54:33.145660Z [DEBUG] QualcommPalmModem <>: triggerUpdateNetworkStatus() status = denied Dec 11 15:18:50 this SIM should still be active... Dec 11 15:18:51 thats how my fsogsmd.log looks like Dec 11 15:18:51 http://pastie.org/1367964 Dec 11 15:19:51 the msmcommd <-> fsogsmd handling does not handle all error cases right now (e.g. Modem resource not available anymore ... available again fsogsmd should connect again automatically) Dec 11 15:20:56 freesmartphone.org: 03morphis 07cornucopia * r51f8c401834b 10/fsogsmd/conf/palm_pre/fsogsmd.conf: fsogsmd: conf: palm_pre: remove unused lowlevel_palmpre plugin Dec 11 15:20:57 freesmartphone.org: 03morphis 07cornucopia * r4bf701d72e09 10/fsodeviced/configure.ac: Merge branch 'master' of ssh://git.freesmartphone.org/cornucopia Dec 11 15:21:19 hmm Dec 11 15:21:24 GetSimInfo makes the modem close Dec 11 15:21:35 mickeyl, afair puffthemagic is working on a boot menu, too Dec 11 15:22:09 mickeyl: jepp Dec 11 15:22:15 next commit will fix that for now Dec 11 15:22:20 k Dec 11 15:22:23 hmm Dec 11 15:22:28 playya__: jepp but he don't get kexec working until now Dec 11 15:22:31 now it hangs in http://pastie.org/1367964 Dec 11 15:22:33 err Dec 11 15:22:36 2010-12-11T13:58:19.007446Z [DEBUG] QualcommPalmModem <>: Modem agent is not ready yet ... waiting 5 seconds. Dec 11 15:23:07 ah, no automatic msmcommd start Dec 11 15:23:32 morphis_, he's not using kexec Dec 11 15:23:41 hmm Dec 11 15:23:51 doesn#t continue initializing though Dec 11 15:24:01 playya__: no? Dec 11 15:24:02 freesmartphone.org: 03morphis 07cornucopia * r9b4ae11665d8 10/fsogsmd/src/plugins/modem_qualcomm_palm/mediators_sim.vala: Dec 11 15:24:02 freesmartphone.org: fsogsmd: modem_qualcomm_palm: disable GetSimInfo mediator for now Dec 11 15:24:02 freesmartphone.org: When calling the GetSimInfo mediator the msmcomm daemon segfaults when handling the Dec 11 15:24:02 freesmartphone.org: response it gets from the modem. As I don't have the time right now to look into the Dec 11 15:24:03 freesmartphone.org: details the mediator is now deactivated. Dec 11 15:24:57 freesmartphone.org: 03morphis 07cornucopia * re37246f6301d 10/fsogsmd/src/plugins/modem_qualcomm_palm/mediators.vala: fsogsmd: modem_qualcomm_palm: only advance modem state when we have a real state change Dec 11 15:25:19 mickeyl: msmcommd is started via an init script Dec 11 15:25:23 /etc/init.d/msmcommd Dec 11 15:25:33 it should be started on boot before fsogsmd Dec 11 15:25:36 hmm Dec 11 15:25:39 why not via dbus? Dec 11 15:25:41 and it is on my device with the latest shr-lite-image Dec 11 15:25:50 time ... Dec 11 15:25:54 k Dec 11 15:26:04 would be nice to start it via dbus Dec 11 15:26:21 but I think there are currently other things which are more important to work on Dec 11 15:26:28 I even tried with shadow resources Dec 11 15:26:30 ok, can't camp the network Dec 11 15:26:37 but that didn't work as I expected Dec 11 15:26:44 perhaps my SIM is no longer valid Dec 11 15:26:51 :) Dec 11 15:27:14 I am recompiling the fsogsmd package if you want to get a more stable version Dec 11 15:27:26 well, my time is up for today anyways Dec 11 15:27:37 if the SIM is really invalid now, i can't do anything anyways Dec 11 15:28:02 it'd be nice to get automatic booting repaired Dec 11 15:28:06 but all that is not so important Dec 11 15:28:12 i can boot via recovery ;) Dec 11 15:29:10 it's cool that the device can run USB Dec 11 15:29:22 do you know whether the modem still works without a battery? Dec 11 15:34:00 mickeyl: puuh good question Dec 11 15:34:02 never tried Dec 11 15:37:01 hmm Dec 11 15:37:03 no service Dec 11 15:37:10 i guess the SIM is really dead then Dec 11 15:41:26 hm ok Dec 11 15:41:33 mickeyl, do you have another phone to try the sim? Dec 11 15:41:43 btw: I tried charge.sh.txt Dec 11 15:41:49 with the bq module Dec 11 15:41:50 it works Dec 11 15:42:13 nothing to change Dec 11 15:42:15 it's strange Dec 11 15:42:22 because I remember it not working before Dec 11 15:43:17 altough to make it work you need 0x10 or 0x90 + to prevent suspend Dec 11 15:45:53 what about this 15s timer? Dec 11 15:46:20 is it necessary to trigger the charging that frequently? Dec 11 15:47:12 mickeyl, I don't know...I just looked into readings and used the charging script which I didn't read attentively Dec 11 15:47:36 ok, so the expert is DocScrutinizer i guess Dec 11 15:47:47 mmm? Dec 11 15:47:49 yes and shadowjk in #meego-arm Dec 11 15:48:20 DocScrutinizer: i understand what the configuration is for, but what is the frequent timer reset for? Dec 11 15:48:23 i2cset -y -m 0x80 2 0x6b 0x00 0x80; # timer reset Dec 11 15:48:24 in particular Dec 11 15:48:38 mickeyl: yes, the times is a watchdog, that's 32s +-75% or somesuch Dec 11 15:48:54 ah, so the charger will automatically stop if you don't trigger? Dec 11 15:48:59 yep Dec 11 15:49:03 ok, good to know Dec 11 15:49:14 reset to power on defaults actually Dec 11 15:49:41 which are emergency charge to max 3.65V or something, for max 32min Dec 11 15:49:46 ok Dec 11 15:50:13 the wd is waking up as soon as you talk to the chip via I2C Dec 11 15:50:46 ok Dec 11 15:51:32 when the wallcharger is in, we can program 800mA? Dec 11 15:51:38 yes Dec 11 15:51:46 we can go max Dec 11 15:51:59 ok Dec 11 15:52:07 here, i never see FULL being anything else than 0 Dec 11 15:52:11 is that normal? Dec 11 15:52:24 with hosterr? Dec 11 15:52:33 err please rephrase Dec 11 15:53:30 is this about shadowjk's slightly oversophisticated script? Dec 11 15:53:37 probably Dec 11 15:53:40 http://enivax.net/jk/n900/charge.sh.txt Dec 11 15:54:29 this status 0xa0 is something i never get Dec 11 15:54:48 you mean $NAC is 0 ? Dec 11 15:55:39 no, 0x00 never contains 0xa0 Dec 11 15:56:20 although sometimes RSOC reaches 100% (this is percent, is it?) Dec 11 15:56:25 that's a mathematical axiom I guess Dec 11 15:56:49 aah, register 0 Dec 11 15:57:00 right? Dec 11 15:57:04 yep Dec 11 15:57:12 maybe oxA0 ? Dec 11 15:57:19 0xA0 even Dec 11 15:58:04 no, it always seem to stick in 0x10 here Dec 11 15:58:36 I've get 0x10 for usb cable Dec 11 15:58:40 and 0x90 for charget Dec 11 15:58:41 also this state only triggers when charging cureent as seen by bq24150 is < threshold. Which means a system that's busy is drawing so much power from charger so you never will reach that state Dec 11 15:58:44 *charger Dec 11 15:59:02 but I don't know if I'm speaking of the same thing Dec 11 16:00:47 DocScrutinizer: hmm, interesting, that could be Dec 11 16:00:57 mickeyl: that also happens with bme on maemo, if screen is kept lit for example Dec 11 16:01:01 can i observe what it currently draws? Dec 11 16:01:13 no, alas not Dec 11 16:01:16 ah, pity Dec 11 16:01:38 you can stop charging and see how much the system draws from battery then :-D Dec 11 16:01:49 by looking at bq27200 Dec 11 16:02:07 but bq24150 has no current gauge readout Dec 11 16:02:37 a smart charger (smarter than bme :-P) would handle things that way Dec 11 16:03:58 by determining the actual *system* current, and then adding that on top of charge-end current for cell, to set charge-end-threshold in bq24150 to that value Dec 11 16:04:30 hmm, ya. i don't see that happening soon though ;) Dec 11 16:05:02 bme however does the same shit OM did some 30 months ago: it restarts charger every 5 minutes Dec 11 16:05:19 heh Dec 11 16:05:21 deja vu Dec 11 16:05:22 wait a bit, it might come with jrbme v0.2 Dec 11 16:05:53 is that maemo-only or can we use it as well?= Dec 11 16:05:54 mickeyl: perhaps you can at least say that raise error.InternalException( "non-unique call state '%'" % status ) should be raise error.InternalException( "non-unique call state '%s" % status )? Dec 11 16:06:05 since %' is not valid python Dec 11 16:06:09 it's a script, like shadowjk's Dec 11 16:06:18 lindi-: yes, now that you say it, this is probably correct Dec 11 16:08:22 mickeyl: jrbme is supposed to handle charging, vboost, and hostmode activation. Plus some notification foo that's probably a bit maemo specific but can easily be adapted Dec 11 16:08:46 nice Dec 11 16:08:49 it's parsing +CSSU: 2,,"",128\r\n and dies since it finds more than one call Dec 11 16:08:56 it's primarily a cmdline tool Dec 11 16:09:07 but may also learn to talk dbus Dec 11 16:10:11 that sounds good Dec 11 16:10:24 mickeyl: I looked into the relevant fso python code, but failed to wrap my head around how to strip down things to get a cmdline binary `wait-for-knotify` Dec 11 16:10:33 mickeyl: does cornucopia have such a a check? Dec 11 16:10:49 mickeyl: I only found supplementary = new FsoFramework.Pair( direction, info ); Dec 11 16:11:35 lindi-: i'm not sure whether we do anything meaningful with CSSU atm. Dec 11 16:11:58 mickeyl: such a cmdline tool that waits until some kernel event is signalling a change in sysfs, that would be mad useful for jrbme, so I wouldn't have to poll all the time Dec 11 16:12:36 DocScrutinizer: have a look at http://git.freesmartphone.org/?p=cornucopia.git;a=tree;f=tools/mkdump;h=8e93093eb129d8d2309118d2b51826836832bfa9;hb=e37246f6301d71bbebe57d74a88274c96b9c7bbb Dec 11 16:12:51 something like this could probably easily be adapted Dec 11 16:13:55 heh, cool :-D Dec 11 16:13:56 mickeyl: so I could just remove it to make ogsmd more robust? Dec 11 16:14:28 looks way cleaner for me than oeventsd or whatever it was I looked at Dec 11 16:14:46 lindi-: this was referring to fsogsmd, i don't know what we are doing in frameworkd w/ CSSU Dec 11 16:15:10 lindi-: but it's probably only used to signalize whether you have been redirected or on hold or something Dec 11 16:15:16 so not entirely important Dec 11 16:16:29 lindi-: i'd really wish you'd take a look at fsogsmd, but i guess you don't want that :) Dec 11 16:16:45 did anyone try the kernel muxer? Dec 11 16:16:47 mickeyl: it does not impress me sorry no :/ Dec 11 16:16:50 hehe Dec 11 16:17:00 ok Dec 11 16:19:40 mickeyl: are lines 63 and 64 a duplicate (well 64 a dup of 63)? Dec 11 16:20:51 not that it matters much, I guess Dec 11 16:21:13 oh, good catch Dec 11 16:21:16 copy'n'paste Dec 11 16:21:26 yup, thought as much Dec 11 16:22:06 freesmartphone.org: 03mickey 07gdbus * r774a08dc9eb6 10cornucopia/tools/mkdump/src/main.vala: mkdump: remove cut'n'paste error as spotted by DocScrutinizer, thanks. Dec 11 16:24:39 mickeyl: also daemon.py has signal( signal, self.sigtermhandler ) which triggers TypeError: an integer is required apparently Dec 11 16:24:58 what means (properties) => { Dec 11 16:25:10 DocScrutinizer: that's a lambda function. Dec 11 16:25:17 WAAAAAHH Dec 11 16:25:33 can also use a callback instead Dec 11 16:25:45 nevermind Dec 11 16:26:08 lindi-: thanks, noted. Dec 11 16:26:18 mickeyl: that should be signal(SIGTERM, self.sigtermhandler) Dec 11 16:26:26 yeah Dec 11 16:28:38 ok, we got onKObjectEvent and onNetlinkEvent, and a number of (aiui) string-consts. Those could be passed to the binary via cmdline parameter as well, I guess Dec 11 16:29:13 to get a finegrained selection on what to wait for Dec 11 16:29:57 yeah, the kobject notification just picks up the key-value combinations from kobject and delivers them to the function Dec 11 16:31:16 why is the first one lowercase and second one uppercase, and is this always this way? Dec 11 16:31:47 as in "add" ... "ADD", "change" ... "CHANGE" etc Dec 11 16:32:21 I gather the Linux.Netlink.RtMessageType.NEWLINK are string constants? are they also the lowercase version of the second string? Dec 11 16:32:53 mickeyl: the 3gpp docs on +CSSU aren't terribly clear indeed Dec 11 16:33:07 mickeyl: i don't even know how to reproduce the use case where +CSSU is generated Dec 11 16:34:27 DocScrutinizer: the ADD in onKobject... is just for output, the real parameters to my generic watcher are "add", etc. for reference, the actual source of the KObjectNotifier is in libfsobasics: http://git.freesmartphone.org/?p=cornucopia.git;a=blob;f=libfsobasics/fsobasics/kobjectnotifier.vala;h=908024fe876f865cd5298a76bf7123b44d045e27;hb=e37246f6301d71bbebe57d74a88274c96b9c7bbb Dec 11 16:35:27 lindi-: call someone who has a redirection setup Dec 11 16:35:43 that should make it appear Dec 11 16:36:14 mickeyl: btw, the signal() bug has a funny side effect. sometimes frameworkd can not be killed with SIGTERM: http://paste.debian.net/102149/ Dec 11 16:36:58 *cough* looking at such sourcecode makes me ponder to learn for butcher or steelworker, and quit sw development Dec 11 16:37:40 quite a bit above my head Dec 11 16:37:52 freesmartphone.org: 03sre 07libisi * r5c0c3b88f4dd 10/src/Makefile: fix glib enum binding header generation Dec 11 16:38:02 well Dec 11 16:38:10 looking at an unknown language is always tough Dec 11 16:38:13 and it's not only because I never had a grip at vala Dec 11 16:38:18 ask me about my first months w/ objective-c ;) Dec 11 16:39:35 it starts with me forgetting about lambda functions as it was too complex to keep in my wetware storage Dec 11 16:40:15 :) just functions that are so unimportant that i didn't even bother to give them a name :) Dec 11 16:40:40 oh, here-functions Dec 11 16:41:35 yeah, now callback makes sense Dec 11 16:45:44 meh, it's embarrassingly simple - I just had to watch a bit closer. :-S Dec 11 16:46:48 missed to notice void onKObjectEvent( string action, HashTable properties ) and the next one Dec 11 16:47:11 :) Dec 11 16:47:56 and protected void _addMatch( string action, string subsystem, KObjectNotifierFunc callback ) is pretty clear about callback Dec 11 16:48:05 right Dec 11 16:48:45 I need breakfast. 4 latte M. are not sufficient it seems Dec 11 17:30:20 freesmartphone.org: 03sre 07libisi * r52929c21ce8b 10/src/Makefile: create installation directory if needed Dec 11 17:55:15 mickeyl, I've cleaned a bit the commit message, should I push to master my battery reading stuff(disabled by default in .config) Dec 11 17:59:01 I'll make a break Dec 11 18:01:18 yes, please Dec 11 18:01:44 i'll add a couple of more i2c convenience functions to i2c.vapi and then we're good to go - transferring the skript into the plugin Dec 11 18:02:52 are the i2c devices O_EXCL? Dec 11 18:03:40 (i.e. wondering whether i need to ioctl to chose the slave device everytime before sending a command) Dec 11 18:05:56 ok thanks Dec 11 18:10:09 also I must fix the Makefile of libisi, for gnu hash issue Dec 11 18:10:17 do I have access to fso libisi? Dec 11 18:10:27 sre said on xmpp: Dec 11 18:10:28 (07:06:42 PM) sre: you may either check it yourself or wait until I'm back home later Dec 11 18:11:31 anyway it's pushed for the kernel Dec 11 18:11:34 there's no fine-grained access control on git.fso atm. Dec 11 18:11:36 http://git.freesmartphone.org/?p=linux-2.6.git;a=shortlog;h=refs/heads/nokia900/master Dec 11 18:11:38 ok Dec 11 18:11:45 there is on SHR so I confused both Dec 11 18:11:48 so feel free to fix Dec 11 18:11:52 ok thanks Dec 11 18:11:58 should i autotoolize the project? Dec 11 18:13:03 | /tmp/ccRuWOgs.s:5441: Error: branch out of range Dec 11 18:13:03 | make[3]: *** [arithmetic.o] Error 1 Dec 11 18:13:11 R-project doesn't like arm hehe Dec 11 18:18:38 pespin: try adding -mlong-calls to CFLAGS :) Dec 11 18:19:42 pespin_: ^^ Dec 11 18:20:25 anarsoul, ok thanks, I'll try :) Dec 11 18:21:10 pespin_: but it's only a workaround. This error means that you're trying to jump out of 32mb segment Dec 11 18:22:12 anarsoul, could have somthing to do with SIZE_MAX define not being set to correct value? Dec 11 18:22:20 I've had to workaround that too Dec 11 18:22:40 pespin_: what are you trying to compile? Dec 11 18:22:49 something huge that uses boost? :) Dec 11 18:22:53 anarsoul, R-project for om-gta02 Dec 11 18:23:12 oh, I see Dec 11 18:23:20 anarsoul, it uses lots of differnet things as far as I've seen hehe Dec 11 18:23:26 c,fortran,c++, et.c Dec 11 18:23:49 DocScrutinizer: i2c -m 0xff is just ORing the old value with the new value, right? Dec 11 18:23:54 i2cset, that is Dec 11 18:24:06 (and then writing the resulting value back) Dec 11 18:25:07 mickeyl, how do I add something to CFLAGS in the OE recipe? Dec 11 18:25:24 pespin_: CFLAGS += somewhere in the recipe Dec 11 18:25:32 freesmartphone.org: 03GNUtoo 07libisi * rd25661e7b142 10/src/Makefile: (log message trimmed) Dec 11 18:25:32 freesmartphone.org: src/Makefile: permit passing LDFLAGS to the linker Dec 11 18:25:32 freesmartphone.org: Without that feature, the openembedded build system Dec 11 18:25:32 freesmartphone.org: fails the QA test like this(with the not yet pushed recipe of libisi): Dec 11 18:25:32 freesmartphone.org: ERROR: QA Issue with libisi: No GNU_HASH in the elf binary: '.../packages-split/libisi/usr/lib/libisi.so.0.0.0' Dec 11 18:25:32 freesmartphone.org: even when passing the LDFLAGS to the Makefile Dec 11 18:25:33 freesmartphone.org: Even if it can be resolved with a hack(using CFLAGS to pass the linker flags) Dec 11 18:25:40 oh I thought it would be more difficult, thanks! Dec 11 18:26:09 mickeyl, if you want Dec 11 18:26:35 mickeyl, but tell me because I'm fixing the Makefile Dec 11 18:26:50 maybe better autotoolizing Dec 11 18:30:16 yeah Dec 11 18:30:33 will do that tomorrow Dec 11 18:31:42 ok Dec 11 18:31:47 I'll eat then Dec 11 18:32:07 what should I do after and before eating? Dec 11 18:32:14 I commited battery Dec 11 18:32:18 s/commited/pushed Dec 11 18:32:34 ah right Dec 11 18:32:36 iphone patches Dec 11 18:32:44 I should look at them and talk to Lopi Dec 11 18:38:28 sounds good Dec 11 18:38:31 freesmartphone.org: 03mickey 07cornucopia * r96e8d1dd8680 10/fsodeviced/src/plugins/powersupply_n900/plugin.vala: fsodeviced: powersupply_n900: prepare for handling charging Dec 11 18:38:37 900 wifi is ok now? Dec 11 18:39:22 bbl Dec 11 18:40:58 mrmoku, is wifi ok now? Dec 11 18:42:12 mrmoku, btw did you permit the about 100Mhz on the nokia900 Dec 11 18:42:28 it's not permited on maemo because of some bugs in proprietary telephony apps Dec 11 18:42:33 but we don't have that stuff Dec 11 18:42:42 so we can permit 100Mhz Dec 11 18:42:53 GNUtoo|laptop: 100 MHz? Dec 11 18:43:31 I'll look the exact frequency Dec 11 18:44:04 125 Mhz Dec 11 18:44:06 to be exact Dec 11 18:44:19 it's from the frequency scalling stuff Dec 11 18:44:43 on the nokia900 Dec 11 18:46:44 GNUtoo|laptop: ok. please try to remember that it's MHz and not Mhz Dec 11 18:47:33 ah right I didn't pay attention when I was writing Dec 11 19:39:49 anarsoul, that worked for aritmethic.o but after that,i get same error on another file Dec 11 19:40:43 pespin_: maybe, CFLAGS are overrided somewhere somehow? Dec 11 19:42:10 anarsoul, nop, it's actually added to gcc http://paste.pocoo.org/show/303628/ Dec 11 19:43:51 pespin_: well, then I don't know how to help you, sorry Dec 11 20:09:40 hello! Dec 11 20:09:56 i still have the problem of speaker not working during calls Dec 11 20:10:07 has a fix been found? i don't remember... Dec 11 20:18:00 nvm gtg :) Dec 11 20:18:02 night Dec 11 20:53:05 GNUtoo|laptop: yup, 125MHz should be allowed Dec 11 20:53:30 indeed Dec 11 20:53:41 GNUtoo|laptop: and sorry... was off with family the whole day... did not further look at wifi Dec 11 20:53:44 yet Dec 11 20:53:59 ok np Dec 11 20:54:14 will you look now? Dec 11 20:54:38 let me try... already had some glass of wine ;) Dec 11 20:55:45 lol ok Dec 11 21:07:51 mrmoku: I've rebased those n900 u-boot patches to 2010.12 master and sofar it doesn't boot at all :/ Dec 11 21:09:15 JaMa: pity :/ Dec 11 21:09:26 JaMa: even with old toolchain? Dec 11 21:09:46 duhh... another toolchain update :P Dec 11 21:09:59 * mrmoku justed wanted to see if his udev changes build correctly :-) Dec 11 21:10:22 have to watch gcc-cross-initial building first.... Dec 11 21:21:47 mrmoku: yes even with old toolchain Dec 11 21:22:17 hmm Dec 11 21:23:26 ah I see I've forgot to migrate your n900 u-boot changes to 162 :/ Dec 11 21:41:26 GNUtoo|laptop: updated udev pushed Dec 11 21:42:08 ok Dec 11 21:47:22 mrmoku: see my message.. shr now defaults to udev-162 Dec 11 21:51:41 JaMa: ahh... :P Dec 11 21:52:03 did not get it, because you sed u-boot ;) Dec 11 21:52:09 s/sed/said/ Dec 11 21:52:09 mrmoku meant: did not get it, because you said u-boot ;) Dec 11 21:54:54 ah, sorry Dec 11 21:56:21 JaMa: I can do it... tomorrow though... if you don't beat me Dec 11 22:00:10 I probably won't beat you this time :) Dec 11 22:00:44 nah, you said that too the last time... and then you did it :) Dec 11 22:04:45 * mrmoku off to bed now :-) Dec 11 22:04:48 gnight Dec 11 22:12:18 mrmoku: gnight Dec 11 22:13:12 * JaMa also going to bed to repel temptation to beat mrmoku again :) Dec 11 22:20:34 JaMa: and then... I did it ;) Dec 11 23:12:40 mrmoku, beware when implementing bt on n900 Dec 11 23:12:57 launching iliwi before emtooth works Dec 11 23:13:07 but the coutrary makes iliwi segfault Dec 11 23:13:22 fortunately here I've an open wifi Dec 11 23:13:34 so I could cconnect from the commandline Dec 11 23:16:15 sigh sigh....spent a lot of time trying to make the keyboard and mouse accepted...just to see a midori crash Dec 11 23:17:13 midori crashes too often to be usable Dec 11 23:17:23 too bad because it's a really goodd browser Dec 11 23:21:13 GNUtoo: if you need some tips about using WPA from cli, feel free to ping me :) Dec 11 23:22:23 ok thanks a lot but I've no wpa here,altough some days ago I had wpa Dec 11 23:22:31 (different location) Dec 11 23:22:41 basically it require editing a config file Dec 11 23:22:52 which isn't present on the device Dec 11 23:22:57 with no manuals... Dec 11 23:23:16 and of course without internet Dec 11 23:23:29 altough it may have a dbus interface Dec 11 23:23:31 right? Dec 11 23:23:48 * GNUtoo will look with mdbus2 Dec 11 23:25:27 GNUtoo: nah, the easiest (suitable for WPA-PSK, and wpa2 too) is: Dec 11 23:25:30 out of memory Dec 11 23:25:39 GNUtoo: wpa_passphrase SSID >> /etc/wpa_supplicant.conf Dec 11 23:25:43 Then type your password Dec 11 23:26:07 wow Dec 11 23:26:09 ok Dec 11 23:26:12 Then if you do not have wpa_supplicant running, start it like "wpa_supplicant -Dnl80211 -iwlan0 -c/etc/wpa_supplicant.conf" Dec 11 23:26:50 If you do have it running already you can use "wpa_cli" utility which is handy enough, "reconfigure" there to force it reread your config and it'll most probably auto-select the network you've added. Dec 11 23:27:58 ok thanks a lot Dec 11 23:28:23 usually people use phone and the voicemail with Dec 11 23:28:27 you have a new message Dec 11 23:28:41 for me this night it's instead: Dec 11 23:28:45 you have a new bug Dec 11 23:28:53 mdbus2 failes with xchat Dec 11 23:29:16 fortunately bt stayed up Dec 11 23:29:36 (I though it wasn't because keyboard was frozen within mdbus2) Dec 11 23:29:42 after the crash Dec 11 23:31:08 GNUtoo: to be actually to use wpa_cli you probably have to add one/two lines to wpa_supplicant beforehand. But once they're there, you can use your non-root user even to scan networks and select which to connect to and see status etc. Dec 11 23:31:24 I'm root Dec 11 23:37:43 PaulFertser: and to get root? ;) Dec 11 23:38:24 lindi-: does it mean you've already found some interesting vulnerability how to get root over wpa_cli? ;) Dec 11 23:39:57 can't reveal anything of course :P Dec 11 23:51:31 mickey|away: (oring and writing) ack Dec 11 23:53:56 mickey|away: old value * !mask + new value * mask Dec 12 00:00:03 umm Dec 12 00:00:08 mickey|away: Dec 12 00:00:10 [2010-12-11 19:23:16] DocScrutinizer: i2c -m 0xff is just ORing the old value with the new value, right? Dec 12 00:00:44 no, -m 0xFF will ignore the old value aiui Dec 12 00:01:24 -m 0xFF is useless, you can discard it Dec 12 00:01:46 no idea who introduced it Dec 12 00:03:03 -m 0x80 for example means bit7 is new value, while bit[6:0] is old value, when writing to register Dec 12 00:05:36 mask-bit = 1 and value = 1 will set the according bit, maskbit = 1 and valuebit=0 will unset the according bit, and maskbit=0 value=x will keep the prev value of the bit Dec 12 00:09:07 as mentioned above, char newregvalue = (oldregvalue AND NOT mask) OR (newparametervalue AND mask) Dec 12 00:12:39 iirc I2Cset without -m does *no* register read prior to doing calculation above, but simply writes the newparametervalue to target register. This might sometimes make a difference Dec 12 00:13:25 also NB sometimes value on a register on read is different to what's been written to same register immediately before Dec 12 00:14:36 so e.g for reg0x00[7] which is wd reset on write, but some status info on read Dec 12 02:40:06 err... Dec 12 02:40:08 10:57 < DocScrutinizer> usually you say speed increasing measures are futile unless you gain a factor 2 at least Dec 12 02:40:15 sorry, that's just nonsense Dec 12 02:41:04 I have heard widely differing numbers regarding this (though certainly never "factor 2" so far); and I'm rather convinced that they are all totally made up Dec 12 02:41:19 I wonder whether anyone has actually conducted any serious studies regarding this... Dec 12 02:41:33 yes, Grman c't mag has, years ago Dec 12 02:41:52 I know for sure that when I upgraded from Pentium133 to Pentium166, I did notice it in *some* situations Dec 12 02:41:56 (not in many though) Dec 12 02:42:37 I have no doubt 600->800 is noticable in some cases Dec 12 02:43:35 when you used one setting for a long time, and suddenly a dialog only needs 600 ms to pop up instead of 800 ms, that's definitely noticable, and a considerable subjective improvement... Dec 12 02:44:00 cpu-freq != performance aka speed of system Dec 12 02:44:17 having said that, I'm not in favour of overclocking -- this usually stays below the perception level Dec 12 02:45:31 (in fact I'm a great fan of underclocking... one of my systems rated for 667 MHz @ 1.65 V runs with 585 @ 1.30 instead -- by my calculation halving the maximal power consumption, and consequently running almost passive :-) ) Dec 12 02:46:22 http://talk.maemo.org/showthread.php?t=66336 Dec 12 02:46:48 DocScrutinizer: not in average performance, no. but in specific situations there is actually an almost linear connection; and in these cases a 30% bump can certainly be noticable Dec 12 02:48:13 yeah, while a MTBF going down to 20% of the original time is usually attributed to "crappy device, crappy software, Nokia being idiots" Dec 12 02:48:49 while in facts it's the OCers that are the true idiots Dec 12 02:51:31 there's both an immediate impact on soft MTBF when bit glitches are caused by too high frequency, as well as a device wear due to electromigration which is irreversible and reduces device's MTBF at any clock freq Dec 12 02:52:27 BTW, I remember reading a 30% speedup figure as perception threshold in some c't article... don't know though whether that was before or after they conducted the study. I assumed the number to be made up -- especially because I did the 133->166 MHz upgrade at around that time, and as I said, the speedup was perceptible in a few situations Dec 12 02:52:56 OCers won't admit they fried they device, not before MTBF gone done to some seconds for any arbitrary app, due to severe EM damage Dec 12 02:54:44 the 30% number was for over all system performance, and a it's-worth-it figure, not a "3 out of 10 testers were able to tell it's there, on close inspection" Dec 12 02:54:50 well, the problem is that every singe device is quite different in this regard... desktop processors get binned after production according to their thermal performance in a test run -- but often they get binned down before sale for pure marketing reasons Dec 12 02:55:11 if you get such a binned down device, it can be overclocked without issues. but you never know... Dec 12 02:56:41 we are talking specifically about OMAP3430 here though, and there it's not thermal design issues but mere EM that is the major concern regarding max clock freq Dec 12 02:58:06 well, you said yourself EM is affected by heat... a chip that has hight hotspots will be affected more at the same fequency than a more lucky example with less severe hotspots... Dec 12 02:58:09 if it were thermal, you couldn't OC to 1000MHz without burning the CPU (plus related blocks like mem controller etc) in milliseconds **** ENDING LOGGING AT Sun Dec 12 02:59:58 2010