**** BEGIN LOGGING AT Mon Jun 15 02:59:57 2009 Jun 15 03:19:37 Hello Jun 15 03:19:55 need some guidance on checking out a specific kernel Jun 15 03:20:15 not well versed with git Jun 15 04:11:17 Anybody available to guide me please Jun 15 04:18:19 :( Jun 15 04:18:31 hi i read about a buzz hardware issue on the openmoko phone and was wondering how do i find out which version i have? (a5 a6 etc) Jun 15 04:18:54 is this bug on all the versions? Jun 15 06:05:23 good morning Jun 15 06:29:40 is it monday yet? Jun 15 08:35:06 Hi All! Jun 15 09:10:43 Who was the guy i was discussing gta01 Qi SDHC support with? Jun 15 10:23:52 dddd Jun 15 12:07:02 hi guys, what i need to install to compile on openmoko on om2009? i need cmake :/ Jun 15 12:07:46 I think you need cmake, gcc, may be g++ Jun 15 12:09:37 max_posedon: but how can i install it? cmake isnt on repository Jun 15 12:10:55 you should compile it with OE Jun 15 12:11:11 oe? Jun 15 12:14:14 OpenEmbedded Jun 15 12:14:26 it short, 1. it isn't easy Jun 15 12:14:36 2. I don't suggest compile anything on neo Jun 15 12:18:09 max_posedon: ok...i'll take a look on it Jun 15 12:18:44 mihairu: why don't you compile cmake on FR itself? Jun 15 12:19:10 max_posedon: i'd suggest compiling on FR ;) Jun 15 12:19:20 max_posedon: it's much easier for small programs.h Jun 15 12:19:39 slackware?! no-go for me Jun 15 12:20:06 max_posedon: for testing purpuses it's ok Jun 15 12:20:58 for testing - yes. Jun 15 12:21:11 PaulFertser: i don't know how :) Jun 15 12:23:50 mihairu: what's the difference? FR is just a gnu/linux computer, do the same as on any pc. Jun 15 12:30:26 FR is a screaming fast server - of a decade ago. Jun 15 12:30:32 It's not actually that slow. Jun 15 12:32:29 If one isn't trying to do floating point on it. Jun 15 12:32:50 yeah Jun 15 12:33:02 Also probably SD performance suck comparing to the HDD of those servers. Jun 15 12:33:06 True Jun 15 12:33:20 Hmm, I cannot connect to my FR from my computer Jun 15 12:33:23 (given the slow glamo bus especially) Jun 15 12:33:24 Im running SHR Jun 15 12:33:47 http://pastebin.com/d3d4740de Jun 15 12:34:03 Etu: read dmesg Jun 15 12:34:48 http://pastebin.com/d1f249ac8 Jun 15 12:35:13 Etu: obviously Jun 15 12:35:58 Ah, I see Jun 15 12:36:03 It's eth1 Jun 15 12:38:50 witch configure cmake Cannot find appropriate C compiler on this system. Jun 15 12:39:04 g++ and gcc i have installed Jun 15 12:39:52 mihairu: read config.log Jun 15 12:43:57 mihairu, gcc, g++, gcc-symlinks, g++-symlinks installed? Jun 15 12:46:01 max_posedon: yeah, i installed it Jun 15 12:50:00 in log is this when it tries gcc: cmake_bootstrap_2181.test.c:5:18: error: stdio.h: No such file or directory Jun 15 12:51:00 mihairu: install libc-dev Jun 15 12:51:19 or glibc-dev, i don't know how it's called in OE. Jun 15 13:01:49 libc-dev nor glibc-dev is not in repositories Jun 15 13:02:13 only glibc-dbg and doc Jun 15 13:35:52 oh ive got it...but now log says: gcc: error trying to exec 'as': execvp: No such file or directory Jun 15 13:50:48 "good morning" Jun 15 14:05:32 How would one best go about writing elementary apps for the FR? I would think you would get elementary on your normal computer, write the app, then cross-compile for FR, but I can't get elementary to build on my laptop so I wonder if there's another way... :P Jun 15 14:11:29 Digital_Pioneer, what about install e17-dev packages for your disto Jun 15 14:11:34 and just start developing? Jun 15 14:12:25 nobody knows where i can grab 'as' program? :/ Jun 15 14:14:27 mihairu, usually its from binutils Jun 15 14:15:11 mihairu, just opkg install binutils binutils-symlinks Jun 15 14:15:22 max_posedon: thx :) i searching it, googling it, nothing...thx Jun 15 14:21:39 grr now error with CXX :D Jun 15 14:30:32 mihairu, are you sure that you have g++ and g++-symlinks? Jun 15 14:31:59 max_posedon: I am...error is: cmake_bootstrap_3034.test.cxx:2:21: error: iostream: No such file or directory...what library I missed? Jun 15 14:32:30 libstdc++-dev Jun 15 14:32:59 max_posedon: great :) Jun 15 14:33:08 max_posedon: sry for dumb questions Jun 15 14:45:10 max_posedon: and i hope last question, 'make' is in which package? Jun 15 14:46:19 mihairu, guess) Jun 15 14:47:43 max_posedon: i don't know...make package doesn't exists :D Jun 15 14:48:12 really? Jun 15 14:48:20 because for me it exists Jun 15 14:48:33 opkg install make Installing make (3.81-r0) to root... Jun 15 14:49:46 * Cannot find package make.\ Jun 15 14:50:28 I use shr-unstable Jun 15 14:54:11 max_posedon: thx now it works :) Jun 15 15:55:05 hi Jun 15 16:00:52 hi Jun 15 17:02:06 shr-release? Jun 15 17:02:32 u guys need a knife? Jun 15 18:27:08 * DocScrutinizer wonders what dehqan is doing right now o.O Jun 15 19:30:35 raster: I have my own dictionary; but after a reboot I always have to select a dictionary to be used (clicking upperleft of the keyboard), is there a way to use one by default? Jun 15 20:30:20 I'm struggling with Navit. Is anyone able to enter a route using the "internal" gui? I can't work out how to tell it what country I'm in Jun 15 20:51:02 how come the gps time is sometimes 15 seconds or so off? Jun 15 20:51:37 i understand if it cant find the time, but being that much off seems strange to me Jun 15 20:52:10 Maybe a firmware bug calculating the offset between "GPS time" and UTC (due to leap seconds) Jun 15 20:53:13 would that explain why it sometimes takes a long time to get a fix? Jun 15 20:53:22 (when being outside) Jun 15 20:55:34 I wouldn't expect there to be a connection, but I don't know anything about the internal algorithms in the GPS firmware Jun 15 20:56:01 henkie_: please run iki.fi/lindi/openmoko/compare-clock-sources.pl Jun 15 21:00:02 lindi-, which perl packages do i need? Jun 15 21:00:45 henkie_: those that have HiRes.pm and Local.pm Jun 15 22:17:38 henkie_: SHR e.g. shows local time until GPS gets a fix. So if your RTC/systime is 15sec off, you would see this 15sec off time until it's adjusted by GPS fix Jun 15 22:20:14 this is probably a flaw created by shr-settings-gps showing the time that's actually present in GPS chipset - irespective of any flag indicating it got a fix - while agps is storing systime to gps chipset on power-up to speed up the TTF Jun 15 22:25:00 henkie_: mmontour's explanation also might apply for a short time (though we have only 12 leapsec afaik) Jun 15 22:25:42 short time here means max 1min Jun 15 22:48:36 does anyone here tried out to use an external usb wlan antenna with a freerunner? Jun 15 22:48:57 agg1n|freerunner: usb wlan antenna? Jun 15 22:49:10 agg1n|freerunner: wtf? Jun 15 22:51:38 wtf indeed :) Jun 15 22:53:59 PaulFertser: just tested cpufreq patches btw. they seem to work, i wonder why they haven't got the deserved attention Jun 15 22:54:25 lindi-: what about power consumption benefits? Jun 15 22:55:07 lindi-: you mean the latest patches by Rask? They look cool. Is that latest patch still supports only 100 and 400 MHz for main SoC clock? Jun 15 22:55:09 PaulFertser: ~30 mA: http://lists.openmoko.org/pipermail/openmoko-kernel/2009-June/010248.html Jun 15 22:55:16 PaulFertser: yes, those Jun 15 22:56:28 lindi-: really nice and important! Jun 15 22:57:24 lindi-: one more question: does xset s blank really call FBIOBLANK ioctl? Can you see anything on the display or is it really turned off during that time? Jun 15 22:57:28 i dont know how to call it correctly :/ i mean those usb wifi adapters Jun 15 22:57:42 PaulFertser: bl_power == 0 at least Jun 15 22:58:07 lindi-: no doubt, but what about lcm itself? Do you have a werner's pmu utility? Jun 15 22:58:19 * lindi- searches svn Jun 15 22:58:37 agg1n|freerunner: if such an adapter works with your GNU/linux pc/laptop, it will most probably work with FR too. Jun 15 22:58:54 now I do Jun 15 23:00:37 PaulFertser: ok so I need to lookup the PMU register to read? Jun 15 23:00:50 PaulFertser: ok, thx, thats all i wanted to know :) Jun 15 23:01:10 lindi-: no, i'd just enable backlight directly and look at the screen. Jun 15 23:01:15 lindi-: pmu 28=50 Jun 15 23:01:18 lindi-: pmu 29=1 Jun 15 23:01:41 lindi-: This will enable backlight not touching anything else so you'll see if LCM is actually running. Jun 15 23:02:11 PaulFertser: I see screen contents after 29=1 Jun 15 23:02:50 lindi-: that's because 28 is set to 2 already, it shouldn't be set to 0 or else one will need to fully disable backlight and enable it again. Jun 15 23:03:03 lindi-: (though probably it's ok kernel refrains from doing that) Jun 15 23:03:25 lindi-: so it'd be interesting to know how much can you save by turning off lcm too. Jun 15 23:03:53 PaulFertser: i need to write some app to throw that ioctl? Jun 15 23:04:26 lindi-: hm, i think you can just do $ python Jun 15 23:04:29 PaulFertser: (is pmu 29=0, pmu 28=2 the way to undo what I just did?) Jun 15 23:04:32 framebuffer = open( "/dev/fb0" ) Jun 15 23:04:41 lindi-: (yes, it is) Jun 15 23:04:53 import fcntl Jun 15 23:05:29 fcntl.ioctl( framebuffer, 0x4611, 4 ) Jun 15 23:06:32 lindi-: btw, now that you're after those freq scaling patches i believe they'll receive the stability testing they deserve and i think nobody will object to merging them after that. :) Jun 15 23:07:37 lindi-: 4 is to blank, 0 is to unblank, i just checked on my device Jun 15 23:08:15 measuring... Jun 15 23:08:48 lindi-: btw, blanking forces backlight off but xset s blank doesn't seem to force blanking. Jun 15 23:09:23 PaulFertser: "blanking" as in FBIOBLANK? Jun 15 23:09:54 PaulFertser: hmm, I can't believe my readings. consumption just dropped to 55 mA Jun 15 23:10:18 (i'll rerun the test) Jun 15 23:11:28 yeah, 55 mA :) Jun 15 23:11:34 lindi-: for real? Jun 15 23:11:42 lindi-: running system just 55mA? Jun 15 23:11:44 PaulFertser: yep Jun 15 23:11:50 if I can trust current_now Jun 15 23:11:51 100MHz? Jun 15 23:11:54 yep Jun 15 23:12:08 Do you mean glamo LCM consumes 50mA Jun 15 23:12:22 i have some wifi stuff here too Jun 15 23:12:31 it makes comparison bit hard Jun 15 23:12:52 (if the phone is printing results over wifi it affects the results.) Jun 15 23:13:19 lindi-: what about using usb cable and force usb curlim to 0? Jun 15 23:13:38 PaulFertser: the effect of FBIOBLANK seems to be 104 mA -> 89 mA Jun 15 23:13:58 not printing one line per second over wifi drops it to that 55 mA Jun 15 23:14:08 lindi-: it is still a considerable saving, more than a potential benefit of fixing #1024. Jun 15 23:14:14 yep Jun 15 23:14:28 PaulFertser: so xserver-xorg-video-fbdev should throw that ioctl? Jun 15 23:14:47 lindi-: no idea, sorry Jun 15 23:15:06 lindi-: currently, framework's odeviced can do that (and i think it does). Jun 15 23:16:04 * lindi- does apt-get source xserver-xorg-video-fbdev Jun 15 23:16:39 Yep, on my system frameworkd issues the ioctl, it's enabled by default. Jun 15 23:16:48 PaulFertser: btw, I did not know about that "force usblim to 0" trick, thanks a million! Jun 15 23:16:58 lindi-: but shouldn't video-glamo driver do the same? Jun 15 23:17:08 lindi-: (trick) does it work? Jun 15 23:17:54 PaulFertser: probably, i'm trying to make xserver-xorg-video-fbdev a stable fallback while xserver-xorg-video-glamo undergoes major KMS rewrite Jun 15 23:17:54 lindi-: as i don't know pmu operation in details, it is just a guess that it should be possible. Jun 15 23:18:34 i'll see in 20 seconds Jun 15 23:19:05 lindi-: if the trick doesn't work i think i can come up with a dirty "pmu ??=??" solution. Jun 15 23:19:28 55mA with not suspended neo?! Jun 15 23:19:33 PaulFertser: works! Jun 15 23:19:43 this will make my tests a lot easier Jun 15 23:19:46 * max_posedon 3rd time re-reading backlog Jun 15 23:19:46 max_posedon: i can't believe it too, but as lindi- says it, he can't be wrong. Jun 15 23:19:54 I can be wrong :) Jun 15 23:20:12 just try the cpufreq patches to verify the results Jun 15 23:20:21 I can belive that its theoretically possible Jun 15 23:20:24 lindi-: but you can't be wrong for too long as you always do enough tests to get a statistically correct value, afaict :) Jun 15 23:21:04 21h not suspended neo... my mind... Jun 15 23:21:26 max_posedon: you're dreaming, it's 2:20 already for you, i guess. Jun 15 23:21:36 PaulFertser: I was writing 500 to a non-existent file for three months when force_usb_limit_dangerous got renamed Jun 15 23:21:55 PaulFertser, yep, could be). Jun 15 23:22:07 max_posedon: note, without gsm, with wifi Jun 15 23:22:09 lindi-: another argument to the point that every operation's return code should be checked ;) Jun 15 23:22:47 PaulFertser: well it wrote an error message to .xsession-errors but glamo had spammed that too much Jun 15 23:22:49 lindi-: i think gsm with deep sleep adds about 4mA minimal and no more than 10mA on average? Jun 15 23:22:53 usually I want gsm, and dont want wifi Jun 15 23:23:00 max_posedon: i can test gsm too Jun 15 23:23:08 lindi-: with deep sleep? Jun 15 23:23:54 PaulFertser: deep sleep recamps here all the time Jun 15 23:24:04 lindi-: why haven't you fixed it yet? ;) Jun 15 23:24:10 PaulFertser: i'm waiting for debconf Jun 15 23:24:30 lindi-: ah, ok, have a good time there :) Jun 15 23:25:14 PaulFertser: thanks Jun 15 23:25:38 lindi-: it'd be nice if you posted your measurements with gsm on, wlan off, LCM blanked to the ML, probably it'll boost interest in Rask's patches. Jun 15 23:25:54 lindi-: and of course any stress-testing of those is much appreciated. Jun 15 23:26:28 it takes a long time to start frameworkd when cpu is running at 100 MHz :-) Jun 15 23:26:56 lindi-: you can switch it on the fly, can't you? Start it with 400MHz, then switch to 100. Jun 15 23:28:20 lindi-: it looks like with that stuff gta02 will be able to live in idle mode as long as gta01 can in suspend ;) (though i very much hope the latter will be fixed soon, at least i know some guys who have motivation to do a bit of testing and hardware hacking and i'm ready to assist with the kernel). Jun 15 23:30:46 lindi-: and given we hope that we won't need python for the core functionality soon (like before NY), startup time will be less of a problem. Jun 15 23:33:56 PaulFertser: yep we need some sort of automatic control for the freq Jun 15 23:34:25 lindi-: if the transition time is not prohibitively long Jun 15 23:35:11 lindi-: there's ondemand gouvernoeur (let french write their words correctly, i don't care) Jun 15 23:37:11 PaulFertser: yep but i understood from Rask that it doesn't work at the moment for some unknown reason Jun 15 23:37:54 lindi-: no, he posted a followup. Jun 15 23:38:00 max_posedon: seems to be around 59 mA with gsm on Jun 15 23:38:13 I need to verify that wifi really got switched off Jun 15 23:38:47 lindi-: http://markmail.org/message/qnvnqaralyd6acbt Jun 15 23:39:22 ah Jun 15 23:40:22 lindi-: that latest patch has even 50 availabl Jun 15 23:40:25 it seems Jun 15 23:40:36 that'll have to wait till tomorrow :) Jun 15 23:40:38 I mean you have a broader freq selection to choose from. Jun 15 23:40:47 50, 100, 200, 300, 400 **** ENDING LOGGING AT Tue Jun 16 02:59:57 2009