**** BEGIN LOGGING AT Sun Jun 14 02:59:57 2009 Jun 14 07:16:12 rtp: interesting thing about gta01's fb setup is that it seems that gpio's are never explicitly initialized (gpdcon and friends) are not specified unlike on other devices (h1940, jive etc). Jun 14 07:32:52 In fact i'm now sure that for enlightment battery gadget to work properly, one needs to fix its code, not hal. Because it doesn't make any sense to average percentages. So if batgat sees several hal "batteries" it should take into account only those that provide actual capacity values in mAh units (and both apm battery and usb "battery" would be omitted from calculation because of these constraints). Jun 14 09:28:53 should the partition on sh card be mounted async, or sync? Jun 14 09:28:59 s/sh/sd/ Jun 14 09:28:59 onen|openBmap meant: sdould the partition on sh card be mounted async, or sync? Jun 14 09:30:41 I really suggest async Jun 14 09:42:04 max_posedon: thanks! (I don't remember the answer I have been given a couple of days ago. And I forgot to look at it before reflashing my FR...) Jun 14 10:08:58 PaulFertser: morning. imho they don't setup gpio in gta01 fb because they assume the bootloader did it Jun 14 10:24:38 rtp: hm, that's a wrong assumptions. Qi doesn't. Jun 14 10:24:54 rtp: but i know that it works nevertheless, i saw gta01 with Qi. Jun 14 10:25:49 will launcher (there is a thread on om-community) be included in shr? it looks quite nice Jun 14 10:27:44 what software is suggested now to configure wifi(shr-unstable)? Jun 14 10:28:37 max_posedon: i'd still recommend vim ;) Jun 14 10:29:07 hm... vim ? Jun 14 10:29:20 vim /etc/wpa_supplicant/wpa_supplicant.conf of course. Jun 14 10:29:40 Or /etc/network/interfaces for Debian-way. Jun 14 10:29:43 I use wep) Jun 14 10:29:54 max_posedon: wpa_supplicant can work with wep too. Jun 14 10:32:25 Weiss: alive? Jun 14 10:32:29 yep :D Jun 14 10:32:33 ha Jun 14 10:32:35 6 hours into my KMS hackathon :O Jun 14 10:32:36 And kicking! Jun 14 10:32:51 i have spent some time adding the gem unref code Jun 14 10:33:08 and now the wait rendering ioctl Jun 14 10:33:40 finally I have problems with the interrupt again Jun 14 10:34:09 i only receive it once Jun 14 10:34:34 hmm Jun 14 10:35:14 @shr devs: my moko muted without my intervention - no sound whatever i do (calling, getting called,...) a reboot solved it Jun 14 10:36:49 Weiss: how is kms going? Jun 14 10:37:35 PaulFertser: well, someone has to setup the gpio. doing it the fb driver is the right place for that not any other random place Jun 14 10:37:44 DieMumie1: better than expected. my FR hasn't melted yet.. Jun 14 10:37:54 ah you are using the debug board Jun 14 10:38:19 rtp: and it already does that for other machines. One just has to supply mask and values for port C and D. Jun 14 10:38:35 Weiss: you can change that easily - just make a call with a duration >30min Jun 14 10:38:43 hehe Jun 14 10:38:57 the neo gets really hot Jun 14 10:39:39 the hardest thing is the reversal of the normal heirarchy of things with KMS... /dev/fb0 becomes a "second class citizen", set up by DRM just to give the kernel something to put a console onto. that means moving a lot of code from "old" glamofb.c Jun 14 10:39:41 PaulFertser: yeah Jun 14 10:40:04 Weiss: how about just dropping console support temporarily? Jun 14 10:40:06 yeah! Neo uses audio video vibration and temprature to inform the user about the system state! Jun 14 10:40:53 Weiss: just discovered something strange Jun 14 10:41:06 ... cant believe it Jun 14 10:41:08 lindi-: option 1 is to do that, then go and get the KMS DDX working to see something on the screen. option 2 is to get the kernel console working first. both are equally difficult really, i think. Jun 14 10:41:35 Weiss: looks i used the wrong registor offset all the time, to setup the irq Jun 14 10:41:41 -o +e Jun 14 10:41:54 interesting..? Jun 14 10:42:44 hmm just silly Jun 14 10:42:58 which offset? Jun 14 10:43:00 it seems that either the kernels are getting unstabler or the buzzfix broke some things in a really nasty way Jun 14 10:43:21 Weiss: ok. i'd be happy to test option 1 since i don't use console for anything really. With option 2 I could do a short test but would then be forced to reboot back to a working kernel to actually use the phone Jun 14 10:43:41 bumbl: how do they crash? Jun 14 10:43:54 Weiss: the local reg_write and reg_read Jun 14 10:44:06 they only refer to the cmdq registers Jun 14 10:44:26 DieMumie1: ah, yes.. Jun 14 10:44:41 i did not notice Jun 14 10:44:41 there's a specific reg_write and reg_read for almost every source file.. it's quite icky Jun 14 10:44:43 untill now.. Jun 14 10:45:01 i have to use gdrm->glamo_core->base Jun 14 10:45:30 hence i never enabled or disabled the irq but poked the cmdq_ctl register :( Jun 14 10:45:48 aha.. Jun 14 10:46:18 or other registers.. Jun 14 10:47:54 lindi-: kind of nsod on resume (again), random muting, random extrem bad voice quality on the other side,... Jun 14 10:48:23 bumbl: does jtag work in that nsod state? Jun 14 10:48:34 jtag? Jun 14 10:48:38 bumbl: debug board Jun 14 10:48:46 i don't have one Jun 14 10:49:01 bumbl: if you setup a led to blink, will it blink on resume? Jun 14 10:49:24 i'll try Jun 14 10:49:55 + distro hopping (including a kind of kernel hopping) seems to draw all the nasty bugs to my device Jun 14 10:49:56 Weiss: can one configure settings like double buffered or single buffered with kms - and control the buffer switch? Jun 14 10:51:49 lindi-: i got the nsod and random extremly bad call quality with paroli unstable - to the nsod issue - s&r worked perfectly the whole day and night but after i woke up the other day each resume brought up a nsod with that kernel (resume; hard reset; resume; hard reset) Jun 14 10:51:52 DieMumie1: i haven't looked in detail yet, but there is pageflipping support which we can use Jun 14 10:52:25 bumbl: i'll wait for the led blink results :) Jun 14 10:53:53 lindi-: well tell me how to tell the led to blink on resume + tell my neo that it should again produce the nsod series Jun 14 10:57:07 bumbl: I have a script like http://paste.debian.net/39066/ Jun 14 10:57:12 bumbl: which I call with e.g. "100 200" Jun 14 10:57:32 ok Jun 14 10:58:05 bumbl: I also verified that if I configure leds to blink they will blink even after suspend/resume Jun 14 10:59:54 Weiss: interrupt works now :( Jun 14 11:00:10 DieMumie1: :D surely? Jun 14 11:00:20 yup Jun 14 11:04:20 now i try to find out how much i can remove to break it again :) Jun 14 11:05:02 hehe Jun 14 11:05:09 well spotted though Jun 14 11:06:26 bumbl: do you use nsod workaround (lowering glamo mem clock frequency?) Jun 14 11:06:53 PaulFertser: afaik this fix is provided by default Jun 14 11:07:02 bumbl: why do you think that? Jun 14 11:07:33 because the "nsod" stopped from one kernel version to another in the past Jun 14 11:08:20 + i don't have the original nsod (xrandr -> nsod) but s&r -> nsod and that only sometimes with special kernels Jun 14 11:10:19 bumbl: https://docs.openmoko.org/trac/ticket/2217 Jun 14 11:11:01 ah Jun 14 11:11:09 confused that with another *sod Jun 14 11:11:37 bumbl: so it looks like those with nsod still should activate workaround explicitly. Jun 14 11:12:22 PaulFertser: well for shr it is default afaik Jun 14 11:12:51 as i bugged mrmoku|away really hard Jun 14 11:12:53 bumbl: in fact i'm sure, git log shows 7f859d161097631a3c76ed1dbb1cfcb08ebe9759 as the last commit touching glamo-core. Jun 14 11:13:03 bumbl: cat /proc/cmdline to prove that! Jun 14 11:13:13 and never had this problem again Jun 14 11:13:19 *starting up the neo* Jun 14 11:15:18 nsod? Jun 14 11:15:41 khiraly1: noise screen of death, read the ticket mentioned Jun 14 11:15:50 ahh Jun 14 11:16:04 noise screen of death (here: a white screen fills slowly with black noise until it stays black, backlight on) Jun 14 11:16:19 bumbl: are you using Qi? Jun 14 11:16:51 And SD? Jun 14 11:16:57 ah yes Jun 14 11:17:06 problem only occurs when using qi Jun 14 11:17:09 rootfs is on nand Jun 14 11:17:39 bumbl: how on earth can SHR than make that workaround default? Jun 14 11:18:00 bumbl: SHR shouldn't patch upstream kernel! Jun 14 11:18:38 PaulFertser: afaik shr always had had their cherrypicked kernel Jun 14 11:18:44 mrmoku|away: ping Jun 14 11:18:45 bumbl: no Jun 14 11:18:50 bumbl: at least not shr unstable Jun 14 11:18:57 Ainulindale: ping Jun 14 11:21:46 bumbl: why do you ping them instead of looking at SHR sources via webgit interface? Jun 14 11:26:31 upgrading navit becomes harder and harder... i killed the xserver, have 512 MB swap space, and still it just stops... Jun 14 11:26:43 when opkg upgrading navit everything hangs Jun 14 11:27:05 when downloading it first and opkg upgrade navit-current.opk, it just quits after a while (no error message...) Jun 14 11:27:13 is there any way to force opkg? Jun 14 11:27:49 Zork|AwaY: probably you need to strace opkg? Jun 14 11:29:14 good idea Jun 14 11:31:22 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 Jun 14 11:31:22 --- SIGCHLD (Child exited) @ 0 (0) --- Jun 14 11:31:22 rmdir("/tmp/opkg-WFmLls") = 0 Jun 14 11:31:22 exit_group(0) = ? Jun 14 11:31:22 Process 1753 detached Jun 14 11:32:32 Zork|AwaY: -f Jun 14 11:57:44 pastebin is being a dick... Jun 14 12:02:46 PaulFertser (lindi-): http://users.telenet.be/404/navit_strace.txt Jun 14 12:05:48 Zork|AwaY: i can't guess what's wrong :-/ Jun 14 12:06:22 thank anyway for taking a look at it; it's all chinese to me Jun 14 12:07:03 PaulFertser: it's a rather big opk (5MB) maybe that's causing problems... Jun 14 12:07:54 Zork|AwaY: is it the log when it "just quits after a while"? Jun 14 12:08:28 ho Jun 14 12:09:00 DocScrutinizer: playing santaclause? Jun 14 12:09:18 ho ho ho Jun 14 12:09:41 isn't it about time to start? Easer is gone quite a while now Jun 14 12:09:54 easter Jun 14 12:10:49 Zork|AwaY: probably it just decides that it shouldn't upgrade it for some reason? Jun 14 12:11:02 Zork|AwaY: you can try to remove already installed navit first. Jun 14 12:11:04 all easter choc bunnies are reworked to santas now Jun 14 12:11:45 you can try to tell opkg -vvv ? Jun 14 12:11:56 what does that do? Jun 14 12:12:14 no idea Jun 14 12:12:31 seems there's been some "-verbouse" flag Jun 14 12:12:45 root@om-gta02 ~ $ opkg -vvv upgrade navit-svn-2315_armv4t.opk Jun 14 12:12:45 opkg version 0.1.6 Jun 14 12:13:02 very useful :) Jun 14 12:13:28 Zorkman: you silly ass, try opkg --help Jun 14 12:13:58 and let me drink my coffee and wake up first ;-D Jun 14 12:15:14 opkg --help | grep verbose Jun 14 12:15:16 already doing it "right" :) and take your time, we don't wanta kranky doc in here :) Jun 14 12:15:23 yeah did -V 3 Jun 14 12:16:08 it ends with Nothing to be done Jun 14 12:16:09 hash_table[pkg-hash] n_buckets=0 n_elements=10008 max_conflicts=0 n_conflicts=0 Jun 14 12:16:09 hash_table[file-hash] n_buckets=0 n_elements=9071 max_conflicts=0 n_conflicts=0 Jun 14 12:16:09 hash_table[obs-file-hash] n_buckets=0 n_elements=0 max_conflicts=0 n_conflicts=0 Jun 14 12:17:40 maybe version not incremented? Jun 14 12:20:13 --force-FOO (where foo like install etc) might help Jun 14 12:21:02 opkg --help | grep force Jun 14 12:43:39 Weiss: if i dont do a proper disable the interrupt will not be triggered again.. so an initial cmdq setup with irq and irq_enable when required, and irq_disable during the interrupt handler is sufficient Jun 14 13:03:06 Zorkman: 5mb opkg shoouldn't be a problem Jun 14 13:04:49 stefen_schmidt: you here? Jun 14 13:05:38 spaetz: [14:16] it ends with Nothing to be done Jun 14 13:13:46 Zorkman: --force-reinstall seems the one Jun 14 13:47:54 raster: Hey :) So do you agree that batget hal code should be fixed to calculate the percentage only for those devices that have a maximum and current capacity expressed in some sane units (mAh) and that will ignore both apm emulated battery (it provides only percentage) and usb (that has no capacity attribute at all)? Jun 14 13:49:58 raster: hey! :-) long time no see Jun 14 14:27:11 http://feeds.digg.com/~r/digg/container/technology/popular/~3/4DrLOrz4nFY/Apple_Patent_Could_Ease_911_Cellphone_Calls < Jun 14 14:27:20 many of those features are already included in the moko :) Jun 14 14:36:46 TAsn: another case of patenting "bread with butter" Jun 14 14:37:35 aye. Jun 14 14:37:49 did you see the feature concerning power reduction Jun 14 14:37:58 by shutting off devices? Jun 14 14:38:06 (gsm/bt/wifi etc) Jun 14 14:38:12 a big lol. Jun 14 14:38:20 yup. Seems FR's "suspend during call" is a superior concept ;-) Jun 14 14:38:45 saves more battery... Jun 14 14:38:50 :) Jun 14 14:39:42 well, iPhone can't do this, as they might need SoC's audio-PCM anyway Jun 14 14:40:23 I remember you told me the iphone doesn't have a sound card :) Jun 14 14:41:11 at least AUX-button's initial name "911" seems a clear case of prior art Jun 14 14:41:21 freesmartphone.org: 03mickey 07specs * r8c8f665dd480 10/ (4 files in 3 dirs): org.freesmartphone.Usage: create enumeration for SystemAction Jun 14 14:43:19 freesmartphone.org: 03mickey 07libfso-glib * r2dd1e7dbcd34 10/.gitignore: refresh; add .gitignore Jun 14 14:45:01 TAsn: (doesn't have soundcard). I remember that's been pure guess. IIRC iPhone is using WM8753 mixer as well, but probably in a different way than FR Jun 14 14:45:37 freesmartphone.org: 03mickey 07libfso-glib * r61884626f6c8 10/src/freesmartphone.vala: refresh for UsageSystemAction Jun 14 14:46:37 mickeyl: how was deoeche mode ;p Jun 14 14:47:26 freesmartphone.org: 03mickey 07cornucopia * ra7fa0e47eeef 10/fsousaged/src/plugins/controller/plugin.vala: fsousaged: controller: use FreeSmartphone.UsageSystemAction enum Jun 14 14:47:27 great stuff Jun 14 14:47:39 they played a lot of their older pieces Jun 14 14:47:44 s/eo/ep/ Jun 14 14:47:44 bumbl meant: mickeyl: how was depeche mode ;p Jun 14 14:51:18 hmm so most of them were before i was born ;) Jun 14 14:51:41 I'm going to a dream theater concert on the day after tomorrow :) Jun 14 14:53:02 TAsn: to where: Jun 14 14:53:03 :? Jun 14 14:53:24 khiraly1, ? Jun 14 14:53:28 dream theater concert Jun 14 14:53:31 don't you know them? Jun 14 14:53:32 yepp Jun 14 14:53:39 oh you mean where in the world? Jun 14 14:53:41 Israel Jun 14 14:53:41 I know them, but where is the concert? Jun 14 14:53:46 hmm Jun 14 14:53:54 because they will play in Budapest too Jun 14 14:54:05 just cant remember exactly when Jun 14 14:54:08 july 1? Jun 14 14:54:10 I'm talking about 16/6 Tel-Aviv Israel :) Jun 14 14:54:14 khiraly1, no idea. Jun 14 15:31:44 PaulFertser: >> >> + .input_key = KEY_PLAY, <<; could you tell me where to find the numeric keycode, as well as a list of all keycodes probably same place? Jun 14 15:35:03 /usr/include/linux/input.h Jun 14 15:35:25 PaulFertser: independent of Apps using keycodes to control them probably need a way to config which key shuld do what anyway to be cross platform compatible, we should not use any of the "standard" 104-key kbd keycodes for HOLD button Jun 14 15:35:39 mickeyl: :-) thanx Jun 14 15:40:56 PaulFertser: I "solved" my battery charging issue I hit yesterday. see openmoko-kernel :-) Jun 14 15:40:59 hmm, KEY_PLAY seems a good choice Jun 14 15:56:38 Ainulindale: you around? Jun 14 16:00:51 TAsn: You know anything about the ipk creation/install process? I'd like to add files to shr-settings, but I'm not sure how. Jun 14 16:03:12 Toaster`: edit shr-settings setup.py Jun 14 16:03:30 or edit shr-settings bitbake repipe Jun 14 16:04:20 freesmartphone.org: 03mickey 07cornucopia * rd4755ce1396d 10/fsodeviced/src/lib/ (Makefile.am resource.vala): Jun 14 16:04:20 freesmartphone.org: fsodeviced: lib: add AbstractSimpleResource and BasePowerControlResource. Jun 14 16:04:20 freesmartphone.org: The former encapsulating the connection to org.freesmartphone.usaged, Jun 14 16:04:20 freesmartphone.org: the latter exports a BasePowerControl as a FSO Resource. Jun 14 16:04:22 freesmartphone.org: 03mickey 07cornucopia * r923aaf6195be 10/fsodeviced/src/plugins/openmoko_powercontrol/plugin.vala: fsodeviced: openmoko_powercontrol: register Bluetooth and WiFi as FSO Resources Jun 14 16:05:25 Toaster`, I agree with bumbl :) Jun 14 16:05:45 bumbl: ok, I can see somethings that looks appropriate in setup.py, the data files list, but I'm confused as to the root of the tuple values. (target_dir, source_dir)? Jun 14 16:05:59 how would I place files in /etc? Jun 14 16:06:43 if I use the full path for the target_dir will it use that instead of going from /usr/share/ ? Jun 14 16:07:31 Toaster`: ../etc/ Jun 14 16:07:43 ../../etc Jun 14 16:07:44 sry Jun 14 16:08:00 ok, so it's fully relative from /usr/share then. Jun 14 16:08:05 would be the easiest solution Jun 14 16:08:12 lindi-: i thought you found why udev consumed 10% cpu :) Jun 14 16:08:14 ok, thanks. I'll try not to break the package Jun 14 16:08:32 setuptools might provide a var for config files too Jun 14 16:08:38 i don't know Jun 14 16:09:24 but yes you have to specify them relative to /usr/share Jun 14 16:09:29 PaulFertser: well, maybe that's the real bug? Jun 14 16:09:44 PaulFertser: it does not consume 10% if I force charging Jun 14 16:10:45 freesmartphone.org: 03mickey 07cornucopia * r02caa203ef80 10/libfsoframework/fsoframework/ (fsoframework-2.0.vapi interfaces.vala): fsoframework: interfaces: add org.freesmartphone.Resource Jun 14 16:11:22 lindi-: is it supposed to consume 10% continiously ever? BTW, looking at commit diff i'd say that it really does the same as the old option. Jun 14 16:12:36 a system service consuming any % constantly is supposed to be broken anyway Jun 14 16:13:15 lindi-: and yes, the commit message should have been more clear, it's a rebase on .29 actually. Jun 14 16:13:19 bumbl: last question, are parent directories to files created if they don't exist? Jun 14 16:14:27 freesmartphone.org: 03mickey 07cornucopia * r7cf39d7891d9 10/fsodeviced/src/lib/resource.vala: fsodeviced: lib/resource: use constants from FsoFramework Jun 14 16:14:37 afaik yes Jun 14 16:19:21 SHR: 03frazier.cameron 07shr-settings * r8c6186e6ab2a 10/setup.py: [Setup] Added backup.* files to packaging/distribution scipt Jun 14 16:23:22 mrmoku|away: Can you rebuild the shr-settings package? My last commit should sort out the config file issues with the backup module Jun 14 16:24:03 DocScrutinizer: heh, probably it got triggered by a kernel bug (like continious stream of uevents or smth like that) :) Jun 14 16:24:37 k, so this is a nasty kernel bug for sure then Jun 14 16:46:59 freesmartphone.org: 03mickey 07cornucopia * r2552e9a6f233 10/fsodeviced/src/plugins/kernel_idle/ (Makefile.am plugin.vala): fsousaged: kernel_idle: add Cpu resource and Display resource Jun 14 16:49:06 * mickeyl checks todo list Jun 14 16:49:43 ok, it looks as after converting the alsa scenarios to the new format, fsodeviced is done. Jun 14 16:55:54 nice Jun 14 17:00:25 fsousage and fsodevice alone should give quite a nice speedup, i.e. when accepting calls Jun 14 17:01:28 very nice ;) Jun 14 17:01:46 how much work is converting the alsa scenarios? Jun 14 17:02:20 if the tools work, 10 minutes Jun 14 17:02:22 :) Jun 14 17:02:30 just loading all scenarios and dumping them with the new tool Jun 14 17:04:01 freesmartphone.org: 03mickey 07cornucopia * r11ea6675c286 10/fsodeviced/src/plugins/kernel_idle/Makefile.am: fsodevice: autofoo magic to get everything to build in one run Jun 14 17:07:47 PaulFertser: udev consumes cpu since it gets power supply events from kernel at fast rate Jun 14 17:08:18 bumbl, pong Jun 14 17:09:37 mickey|dinner: new alsa scenario fileformat? Jun 14 17:14:36 mrmoku: do you provide the nsod fix by default? Jun 14 17:23:34 lindi-: i knew that ;) Jun 14 17:23:56 lindi-: and how is usb_curlim renaming related to those power supply events? Jun 14 17:24:12 bumbl, hmm Jun 14 17:24:22 bumbl, I don't know Jun 14 17:24:41 as far as I know we do not have any additional kernel patches Jun 14 17:24:57 because shr is the only distro with a working kernel by default Jun 14 17:25:32 then the fix must be in the om kernel repo already Jun 14 17:25:42 maybe we just use a newer version than others Jun 14 17:26:17 yeah... actually we use AUTOREV for the kernel in unstable Jun 14 17:26:23 i have not got a nsod for a long time now but with the paroli unstable image i flashed friday i got none during the heavy use on friday and friday night but on saturday morning (no reboot) Jun 14 17:26:55 om kernel does not include the fix by default (it has to be explicitly set by kernel command line) Jun 14 17:27:20 what do you have to set on the cmdline? Jun 14 17:28:31 bumbl: and why dont you add to the append-GTA02 file? Jun 14 17:29:31 mrmoku: glamo3362.slow_memory=1 Jun 14 17:30:16 mrmoku: or 2-7, if one is adventerous Jun 14 17:30:17 khiraly1: because no append-GTA02 file is possible for nand kernels + with shr the problem is nonexistant and i thought it was with paroli too (it was/is) Jun 14 17:31:43 btw, fso will be rewrite from vala to plain C ?;-P Jun 14 17:31:44 bumbl: i don't see a reason to set this by default as the issue is with the minority of devices and the workaround hurts performance. Jun 14 17:31:48 rewritten Jun 14 17:31:56 khiraly1: What for? Jun 14 17:32:01 ok Jun 14 17:32:07 khiraly1: or is it a joke? Jun 14 17:32:40 PaulFertser: bad joke partly. If it still left some perfomance issues Jun 14 17:32:44 bumbl: otoh you might want to ask for a proper method to supply this kernel arg when one is using Qi and wants to boot from NAND. Jun 14 17:32:46 PaulFertser: anyway it is interesting why it started to appear out of nowhere Jun 14 17:32:47 bumbl, is that ticket #2255 ? Jun 14 17:33:14 khiraly1: i mean i hope you know that Vala is compiled to C code ;) Jun 14 17:33:33 bumbl: looks like hardware issues. And they are very often "random". Jun 14 17:33:48 mrmoku: 2217 Jun 14 17:33:50 tracfeed: Ticket #518 (Webserver) created Jun 14 17:33:50 yeah, And it is as good as the vala compiler. So maybe a rewrite to plain C helps in the most critical parts Jun 14 17:33:53 mrmoku: i think it is #2217 Jun 14 17:34:01 PaulFertser: and after plain C, there is only assembly left Jun 14 17:34:03 ;-P Jun 14 17:34:09 (that was the joke part;) Jun 14 17:34:11 khiraly1: machine code Jun 14 17:34:29 assembly is already machine code;) Jun 14 17:35:03 PaulFertser: well something has to trigger it because after the image worked for 1,5 days and numerous s&r it stopped - rebooting didn't help - flashing shr did Jun 14 17:36:24 khiraly1: nope, it a mnemonic form to write the machine code. And sometimes you have to use machine code directly, esp. if you're doing some self-modifying stuff or other dirty tricks. Read the story about Mel, the real programmer. Now :) Jun 14 17:37:16 PaulFertser: I already did it Jun 14 17:37:49 yeah, a mnemonic. Its easier to remember as hexa numbers. But dont will you give any perfomance benefits Jun 14 17:39:37 PaulFertser: well, I only hit this since i was not forcing 500 mA :) Jun 14 17:40:36 someone here with regex powers: what do i have to do if i want to get all ["number1""number2"] expressions? \[*\] seems not to work Jun 14 17:41:44 lindi-: :) so what the events were about? Jun 14 17:44:08 bumbl: \[.*\] Jun 14 17:44:09 first Jun 14 17:44:26 bumbl, hmm... to be honest I can't see why we don't suffer from it :P Jun 14 17:44:43 bumbl: only numbers: \[[0-9]*\] Jun 14 17:44:53 it defaults to 0 in current kernel and we don't override that as far as I can tell... Jun 14 17:45:04 bumbl: would be nice to past some example, to see what we want to match Jun 14 17:45:24 mrmoku: most people have no hardware problems with that, that's why. Jun 14 17:46:03 khiraly1: thank you Jun 14 17:46:22 mrmoku: strange Jun 14 17:47:01 well i don't bother as long as everything stays as it is with shr Jun 14 17:49:24 PaulFertser, I know... I never suffered from it. Just wanted to tell that I don't think we have it enabled by default Jun 14 17:49:43 anyway... dinner time Jun 14 17:49:44 bbl Jun 14 17:57:34 PaulFertser: will just strace do? Jun 14 18:06:35 lindi-: i'm sure you know how to use strace a lot better than me. Jun 14 18:09:01 PaulFertser: it alternatves between recvmsg(... POWER_SUPPLY_ONLINE=1 and POWER_SUPPLY_ONLINE=0 Jun 14 18:09:46 PaulFertser: hw-issue? backscroll puzzled me Jun 14 18:11:52 * DocScrutinizer really wonders which weirdo invented names like online for powersupply Jun 14 18:12:29 DocScrutinizer: talking about NSOD Jun 14 18:12:37 aah Jun 14 18:12:59 DocScrutinizer: bumbl thought that shr includes the glamo memory clock workaround activated by default. Jun 14 18:13:36 lindi-: hm, it shouldn't be like that, should it? To be honest, i've already forgot in which circumstances you see it. Jun 14 18:13:37 glamo memclock? duh, even went unnoticed by me Jun 14 18:14:20 SHR: 03seba.dos1 07shr-settings * ra55e92d93e80 10/ (setup.py shr_settings_modules/shr_backup.py): [backup] [setup.py] change /etc/shr/settings/ to /etc/shr-settings/ Jun 14 18:14:28 SHR: 03seba.dos1 07shr-settings * r00a54faf1861 10/shr_settings_modules/shr_gprs.py: [gprs] store pickle file in /etc/shr-settings/ Jun 14 18:14:28 SHR: 03seba.dos1 07shr-settings * r36651e6eaa53 10/setup.py: [setup.py] install conf files correctly Jun 14 18:14:34 DocScrutinizer: nice thing: paroli image and kernel works for 1,5 days then starts to nsod on resume - reboot does not solve the problem - flashing of shr + according kernel solved it Jun 14 18:14:42 PaulFertser: 1) attach usb battery, put screen brightness to 127 Jun 14 18:14:50 immediately Jun 14 18:16:20 lindi-: would be nice to have some helper app. Which would popup, and ask for feedback Jun 14 18:16:25 well, all those fsckng *SOD was supposed to be fixed for good by powering down LCM during suspend, thus mimicking a clean power-on aka boot, where we never seen ysod before. Now that paroli thing is absolutely nasty Jun 14 18:16:33 khiraly1: helper for what? Jun 14 18:17:09 like you push in the usb, the helper up popups, and asks: [disable suspend][dont change suspend behaviour] Jun 14 18:17:18 s/up/app/ Jun 14 18:17:19 khiraly1 meant: like you push in the usb, the helper app popups, and asks: [disable suspend][dont change suspend behaviour] Jun 14 18:17:37 the same for birghtness, etc Jun 14 18:18:04 DocScrutinizer: https://docs.openmoko.org/trac/ticket/2217 Jun 14 18:18:17 you push in a charger, the helper app asks: [100mA][500mA][1000mA] Jun 14 18:18:21 etc Jun 14 18:18:37 DocScrutinizer: in how far is paroli nasty? Jun 14 18:19:11 bumbl: ask PaulFertser or see backscroll Jun 14 18:19:11 PaulFertser: i have WSOD now only with Paroli Jun 14 18:19:27 dos1|neo: it sounds stupid Jun 14 18:19:30 dos1|neo: :p Jun 14 18:19:49 * DocScrutinizer away for H2G2 ON Pro7 Jun 14 18:19:58 DocScrutinizer: how much back? Jun 14 18:20:27 PaulFertser: heh ;D i mean kernel from Om2009 of course Jun 14 18:20:46 dos1|neo: sure ;) Jun 14 18:20:52 PaulFertser: i didn't have WSOD even with 2007.2 Jun 14 18:21:04 PaulFertser: i had with ASU Jun 14 18:21:17 PaulFertser: and now with om2009 Jun 14 18:21:35 PaulFertser: on SHR, I don't have WSOD at all Jun 14 18:21:39 dos1|neo: WSOD!=NSOD Jun 14 18:22:02 PaulFertser: yep. i mean WSOD on resume Jun 14 18:22:18 dos1|neo: anyway, that's strange, have you tried om2009 kernel with SHR to prove it's om2009 kernel issue? Jun 14 18:22:18 PaulFertser: that depends - here only white and black are included Jun 14 18:22:20 ;) Jun 14 18:22:22 NSOD i have even on SHR when playing with xrandr Jun 14 18:22:56 anybody in here tried canola? it's really nice, if i could reach shr-project.org I would add a ticket to add it to the feed Jun 14 18:23:24 PaulFertser: no. i just try whole distro once. and it WSODed Jun 14 18:23:27 i only had once a nsod when xrandring but that included colours Jun 14 18:25:06 dos1|neo: so far that can't be considered a valid bugreport ;) Jun 14 18:27:14 PaulFertser: i know, it isn't bugreport ;) Jun 14 18:27:34 PaulFertser: just adding to discussion Jun 14 18:27:49 dos1|neo: adding fud to discussion, i see ;) Jun 14 18:28:41 oh, that's not FUD. that's fact, when reffering to whole distro. Jun 14 18:29:14 DON'T PANIC Jun 14 18:29:15 dos1|neo: people tend to misintepret facts like that, you know. That's basically how rumors spread. Jun 14 18:29:45 in fact SHR doesn't blank framebuffer, it only turns off backlight Jun 14 18:30:19 dos1|neo: hm, interesting. That makes sense. Jun 14 18:30:22 aaah, and of course I forgot to mention chapter "Useless Info" Jun 14 18:30:57 dos1|neo: what's the power consumption difference? Who was in charge of that decision? Jun 14 18:31:12 PaulFertser: I assume people here are inteligent enough to think before they spread rumors - it's -cdevel chan ;) Jun 14 18:31:18 actually the movie is blending "DON'T PANIC" from yellow to red with yellow outline ;-) Jun 14 18:31:53 PaulFertser: that's not decision. we are using xset blank Jun 14 18:32:16 dos1|neo: "just because you're paranoid doesn't mean they're not after you!" Jun 14 18:32:17 s/xset blank/xset s blank/ Jun 14 18:32:18 dos1|neo meant: PaulFertser: that's not decision. we are using xset s blank Jun 14 18:33:31 dos1|neo: hm, i thought blanking on FR is equivalent to "dpms force off". Jun 14 18:34:10 PaulFertser: FSO is blanking fb when backlight is set to 0 through dbus call Jun 14 18:34:45 PaulFertser: we're not using dbus call for that, as it overrides current backlight value on blanking Jun 14 18:38:42 dos1|neo: you're right, FSO is using ioctl FBIOBLANK. Jun 14 18:41:17 dos1|neo: why do you think xset s blank doesn't? Jun 14 18:41:52 PaulFertser: cause in sun i can see content of screen when dimmed Jun 14 18:44:00 dos1|neo: PaulFertser: confirmed here Jun 14 18:45:25 Wow Jun 14 18:45:33 isn't there a ticket for that since roundabout 4 weeks? Jun 14 18:45:50 Probably SHR and i don't follow SHR bugtracker. Jun 14 18:47:22