**** BEGIN LOGGING AT Fri Sep 04 02:59:57 2009 Sep 04 03:04:14 i just installed SHR and am having problems with suspend. sometimes it works fine, other times it leaves a pink line through the display Sep 04 03:04:34 kind of freezes the phone, ssh does not work Sep 04 03:04:41 buttons dont work etc Sep 04 03:06:06 is this normal? perhaps i should run without suspend? Sep 04 03:06:38 unstable? Sep 04 03:09:08 Ke: Are you sure Openmoko uses non-little endian? I thought it was little endian, same as x86. Sep 04 03:12:59 hmm Sep 04 03:13:06 not quite sure Sep 04 03:13:08 fredrin: yes Sep 04 03:13:34 could be that Sep 04 03:13:46 fredrin: i had the same problem months ago on another distro Sep 04 03:13:53 same pink line through the screen Sep 04 03:13:54 give the FR some love and it should work Sep 04 03:14:00 :) Sep 04 03:14:07 maybe it's hardware :( Sep 04 03:14:14 Qi? Sep 04 03:14:22 fredrin: how do you mean Sep 04 03:14:33 could be Qi too Sep 04 03:14:37 slaxxer: does Qi do anything when suspend starts? Sep 04 03:14:52 i dont know Sep 04 03:15:07 ok Sep 04 03:16:12 brolin: still I guess it's a thing one should not rely on when programming to arm (or non-native, or even otherwise) Sep 04 03:19:59 i updated qi Sep 04 03:20:21 lets see if it still happends Sep 04 03:22:10 fredrin: how did you mean it could be a hardware error though Sep 04 03:22:28 it's always one alternative.... Sep 04 03:22:36 or option, but i hope not Sep 04 03:22:39 new kernel too? Sep 04 03:23:06 but it sound strange when the hold things just hangs Sep 04 03:24:11 it only happends when it goes into suspend Sep 04 03:24:33 is your kernel up to date? Sep 04 03:24:38 yes Sep 04 03:25:08 still same error Sep 04 03:25:14 :/ Sep 04 03:25:25 its all dark Sep 04 03:25:30 but still lit up Sep 04 03:25:32 change to an older kernel an see Sep 04 03:25:36 and with a pink line through Sep 04 03:25:40 slaxxer: ok Sep 04 03:31:47 :/ still happends Sep 04 03:32:17 tried with the new kernel, a older one and a really old one Sep 04 03:32:30 sometimes the suspend works though Sep 04 03:37:34 Are the accelorometers represented in /dev/? I have a strange idea I want to try. >.> Sep 04 03:39:43 That was decidadly less fun than I hoped. >.> cat /dev/input/event3 | aplay Sep 04 03:40:13 only 1 thing left:) Sep 04 03:41:12 i guess u have the 090808 shr-u? Sep 04 03:43:07 slaxxer: you mean me? Sep 04 03:43:12 yyea Sep 04 03:43:37 090808 <-- kernel? Sep 04 03:44:04 i dont know what shr-u is Sep 04 03:44:05 rootfs Sep 04 03:44:21 oh, ive done a upgrade today Sep 04 03:44:21 shr-unstable Sep 04 03:44:26 cos it was not working Sep 04 03:44:30 but yeah i think it was that one Sep 04 03:44:34 did it work before? Sep 04 03:44:38 n Sep 04 03:44:39 no Sep 04 03:45:16 infact last time i tried shr (looong time ago) i had the same issue i think Sep 04 03:45:28 but i only used shr for a day Sep 04 03:45:41 have u tried it without a microsd in? Sep 04 03:45:41 atm its suspending without any problem Sep 04 03:46:04 no Sep 04 03:46:10 my OS is on my mircosd Sep 04 03:46:49 hmmm Sep 04 03:47:30 now it has suspended like 10 times ok Sep 04 03:47:39 before it did not suspend once for about 2 hours Sep 04 03:47:45 (withut freezing) Sep 04 03:49:30 its a very clear bug, pink line across a black screen Sep 04 03:50:56 and the only way to turn it off is removing the battery Sep 04 03:57:47 it stopped freezing, ive restarted and suspended a few times, ill try it over the weekend Sep 04 03:57:53 you're ill? Sep 04 03:58:17 Chocky: my FR is a bit ill Sep 04 04:40:28 burrp Sep 04 07:45:58 So, Wifi anyone?! The patch was reverted it should work again "properly" now! Please try it and report! Sep 04 08:12:36 i want to run the command opkg update on my FR. i have created a usb network connection between my netbook and the FR and can ssh into it. my netbook has an internet connection via wifi. should the FR use the netbook connection in order to execute the opkg update command? Sep 04 08:13:40 LiL_mr-t: yes, you're supposed to know how to configure networking Sep 04 08:14:12 LiL_mr-t: you can make your netbook act as bridge or you can make it act as NAT router. Sep 04 08:14:33 hi PaulFertser. phone is working well thanks to your help last night. made a call this morning. :) Sep 04 08:14:42 LiL_mr-t: nice Sep 04 08:31:21 I'm afraid i'm not so sure how to create the bridge/NAT solution. can anybody try to point me in the right direction? Sep 04 08:33:06 read qemu networkibg docs Sep 04 08:33:48 as for simple NAT - wiki.openmoko.org have docs about NAT Sep 04 08:34:54 thanks max_posedon1, i'll check the wiki entry Sep 04 08:41:07 PaulFertser: wifi patch seems to work at first glance Sep 04 08:41:33 will continue testing and then recommend it to others to test Sep 04 09:36:06 lindi-: tbh i've already pushed it to andy-tracking head because everybody agreed that the patch itself is wrong. Sep 04 10:13:04 morning Sep 04 10:14:18 DocScrutinizer: morning indeed Sep 04 10:18:56 hi all Sep 04 10:21:35 hello Sep 04 10:38:30 freesmartphone.org: 03betheg 07framework * rfb001f3ce766 10/framework/subsystems/ogsmd/modems/abstract/mediator.py: Sep 04 10:38:30 freesmartphone.org: ogsmd: substitute empty provider string with information from network.tab. Fixes FSO #466 Sep 04 10:38:30 freesmartphone.org: Some providers are too new for the modem, hence we get something like: Sep 04 10:38:30 freesmartphone.org: +COPS: (2," "," ","26201"),(1,"o2 - de","o2 - de","26207"),(1,"E-Plus","E-Plus","26203"),(1,"Vodafone.de","Vodafone","26202") Sep 04 10:38:34 freesmartphone.org: This fixes it by checking our (hopefully more current) provider database. Sep 04 11:51:37 PaulFertser: wlan still works :) Sep 04 11:52:37 enlightenment error while running opkg install task-shr-apps task-shr-games task-shr-gtk. install still downloading and installing. should i choose recover or exit on the FR screen? Sep 04 11:54:32 or just leave it until the download/install has finished? it is currently configuring. Sep 04 11:56:05 <_jan> i would wait, and next time run it in a screen session :) Sep 04 12:01:29 ok. i exited. had to remove battery to turn FR off. rebooted and all appears ok. desktop loaded... shows intone, pidgin and other new icons. Sep 04 12:01:38 hopefully all is ok Sep 04 12:02:47 <_jan> just rerun upgrade the youll probably see :) Sep 04 12:05:23 _jan. i just ran the following commands originally... opkg-cl update, opkg-cl upgrade and then opkg install task-shr-apps task-shr-games task-shr-gtk. so shall i run opkg upgrade now? Sep 04 12:11:50 hey PaulFertser Sep 04 12:11:58 naeg: hey :) Sep 04 12:12:21 PaulFertser: we have ubifs built in now, so how do i get this speedboost now? just mount as type ubifs? Sep 04 12:12:23 anyone know if lazarus works for creating apps for gta01/02? Sep 04 12:12:34 (ubifs-utils packaged too) Sep 04 12:12:59 naeg: you need to flash rootfs of type ubifs then. And use kernel parameter rootfs=ubifs iirc. Sep 04 12:13:19 PaulFertser: and mkfs.ubifs to make image? Sep 04 12:13:25 naeg: yes Sep 04 12:13:43 ah ok thanks Sep 04 12:13:56 naeg: it'd be interesting to hear the results ;) Sep 04 12:15:33 PaulFertser: i will make speedtest: only sd card vs booting from nand(using ubifs), union mounting the sd card in initrd over / Sep 04 12:18:12 forget that point with initrd Sep 04 12:18:41 also - ubifs on sd Sep 04 12:19:35 ubifs on sd is stupid and useless Sep 04 12:19:38 SpeedEvil: user can use whatever they want on sd, but we use ext2, bootloader requires that Sep 04 12:19:47 ah Sep 04 12:19:50 even if it wil work Sep 04 12:19:53 max_posedon: why? Sep 04 12:20:09 its fs for RAW flash Sep 04 12:20:18 max_posedon: random write of small files on SD can have a performance of several K/second Sep 04 12:20:39 think about sd internals Sep 04 12:21:02 max_posedon: 4K writes to a device with 128K erase blocksize = pain. Sep 04 12:21:02 any such flash already have all mehanizm Sep 04 12:21:38 max_posedon: they don't - small block writes make for _terrible_ performance on all SDs I have tested. Sep 04 12:21:42 but you HAVEN'T access to flash on sd Sep 04 12:21:47 Sure. Sep 04 12:21:56 its all done by firmware inside Sep 04 12:22:03 But a large block write will tend to equal a large block write. Sep 04 12:22:17 you never know Sep 04 12:22:31 also, it already done by fimware Sep 04 12:22:49 I was seeing speedups of hundreds of times doing large blocks - 256k - compared to 1K Sep 04 12:23:00 someone ever created ubifs images for his neo? do i have to set any device depending args? Sep 04 12:23:09 on sd? Sep 04 12:23:14 no, for NAND Sep 04 12:23:28 (did you mean me?) Sep 04 12:23:29 naeg: erase page size Sep 04 12:23:55 Hi Sep 04 12:24:18 I'm taking my moko for a backpacking trip to France (for 2 weeks) Sep 04 12:24:29 PaulFertser: no such option: http://pastebin.com/m1001adcd Sep 04 12:24:31 what image do you suggest and what apps? ;) Sep 04 12:24:54 (tangogps is a winner for sure ;-)) Sep 04 12:25:14 SHR i guess pieterc, most stable, most usable Sep 04 12:25:19 naeg: probable -e ? Sep 04 12:25:36 naeg: of course SHR, but what shr image :) Sep 04 12:25:45 pieterc: unstable Sep 04 12:25:50 4august? Sep 04 12:25:50 PaulFertser: -e 0 ? Sep 04 12:26:05 pieterc: dunno, not really into SHR Sep 04 12:27:02 is there another shr image place than http://build.shr-project.org/ <-- this one? Sep 04 12:27:13 because the latest image over there is 8th of august Sep 04 12:27:20 i'd like a somewhat more recent image? Sep 04 12:27:25 Anyone got a recent build? Sep 04 12:27:31 naeg: no idea if it's correct but OE uses this: EXTRA_IMAGECMD_ubifs = "-m 2048 -e 129024 -c 2047" Sep 04 12:28:17 i'll give it a try Sep 04 12:29:45 PaulFertser: well i gess 3,7mb is less Sep 04 12:31:54 ok, i did -d rootfs/* which seems to be wrong, it's -d rootfs/ Sep 04 12:35:54 PaulFertser: can i flash like this: dfu-util -a rootfs -R -D rootfs_filename.ubifs Sep 04 12:36:03 i don't want to destroy anything ;D Sep 04 12:38:40 naeg: you'll just reflash your NAND rootfs, no way to destroy anything Sep 04 12:38:47 naeg: does dfu-util accept that file? Sep 04 12:39:26 * PaulFertser was under the impression sometimes dfu-util needs to have a "DFU footer" in files it flashes. Sep 04 12:41:26 PaulFertser: yes, it's flashing Sep 04 12:41:34 we'll see if it boots... Sep 04 12:42:01 naeg: what bootloader are you using? Sep 04 12:42:01 it has 117mb, so it takes some time Sep 04 12:42:08 PaulFertser: qi Sep 04 12:42:34 naeg: i'm afraid it'll force the kernel to think it's jffs2 and it won't be able to mount it :( Sep 04 12:43:06 PaulFertser: uboot? Sep 04 12:43:08 naeg: qi has kernel command line hardcoded for NAND Sep 04 12:43:22 naeg: well, with uboot you can tweak it however you like Sep 04 12:43:34 PaulFertser: change command line and compile? Sep 04 12:43:42 open source ftw :D Sep 04 12:44:13 naeg: yes, i'd just change the line and Qi and recompile Sep 04 12:45:32 naeg: Open source ftw... okay, but a shitty proprietary software thing would not have hardcoded it in the first place ;-) Sep 04 12:47:29 naaah, open source ftw :D Sep 04 12:49:46 pieterc: you actually made me think... Qi already reads the factory partition to get the MAC. It can as well read kernel command line from there. Sep 04 12:51:14 PaulFertser: exactly what I was thinking Sep 04 12:53:03 pieterc: do you know what FS the factory partition uses? Sep 04 12:53:35 PaulFertser: hm... no idea Sep 04 12:54:02 pieterc: then you are not thinking the same i do ;) Sep 04 12:54:06 hi there :) Sep 04 12:54:39 <_phil> PaulFertser: should rootfstype=jffs2 show up /proc/cmdline when booting with Qi? Sep 04 12:55:06 _phil: in case you boot from NAND, yes Sep 04 12:55:31 <_phil> ahh, i was missing that :) Sep 04 12:59:54 ok, found commandline option rootfstype=jffs2, and changed to rootfstype=ubifs, now waiting til flashing of rootfs is finished, then flashing fresh compiled qi and hoping for good results ;) Sep 04 13:11:46 PaulFertser: it's not booting, i think it's the ubifs image, can't see anything special outer "Exeption Stack(0x781fed8 0x781ff20)" Sep 04 13:12:54 PaulFertser: no, qi debug mode confused me: kernel panic! Sep 04 13:13:22 not syncing: VFS: unable to mount root fs on unknown-block(31,6) Sep 04 13:14:12 naeg: how about trying to mount the partition first as non-root fs? Sep 04 13:15:53 lindi-: how? Sep 04 13:16:14 naeg: with the mount command? Sep 04 13:16:16 naeg: boot from SD Sep 04 13:16:21 naeg: the same kernel Sep 04 13:16:29 naeg: and try to mount /dev/mtdblock6 Sep 04 13:16:29 yes, this was _phil's idea too Sep 04 13:21:36 <_phil> PaulFertser: are there plans to support initrd on the GTA02 with Qi? Sep 04 13:23:21 _phil: it is supported already Sep 04 13:23:36 _phil: when booting from uSD, not NAND Sep 04 13:24:26 but we need it from NAND :/ Sep 04 13:24:51 <_phil> oh, alright, i've tried with /boot/initramfs.gz and with initrd=/boot/initramfs.gz which did not work for me Sep 04 13:26:21 <_phil> i've a small patch that did the trick, but it doesn't fell right: http://pastebin.com/m76ee4135 Sep 04 13:26:31 <_phil> -fell+feel :) Sep 04 13:31:28 PaulFertser: ping Sep 04 13:31:33 BillK_: pong Sep 04 13:31:50 PaulFertser: the link you gave in the email Sep 04 13:31:59 to the wifi fixed kernel looks Sep 04 13:32:11 wrong - 762 bytes in wget and firefox ... Sep 04 13:32:59 BillK_: well, you need in fact ar6000 driver module installed, not the kernel image. Sep 04 13:33:56 BillK_: and 762 bytes is correct, that's a meta-package Sep 04 13:34:37 PaulFertser: ahh, explains it, sorry :) Sep 04 13:35:13 <_phil> PaulFertser: is there any other way to pass the initramfs.gz file, which i am overlooking? Sep 04 13:35:31 _phil: the method you use seem to be correct... Sep 04 13:40:24 it's no more booting from SD, in verbose mode i get "Empty Flash at " (can't read) Sep 04 13:44:10 <_phil> PaulFertser: patch, http://pastebin.com/m76ee4135, does the trick for me Sep 04 13:45:57 _phil: as i said, there's support in Qi, just a little bit buggy it seems ;) Sep 04 13:47:23 <_phil> PaulFertser: this is more a patch submission than a question, thought you might be interested :) Sep 04 13:48:07 _phil: i am indeed, do you want to pick it up as is or are you going to do proper git-format-patch? Sep 04 13:48:39 _phil: do you have any idea why INITRD_OFFSET is wrong? Sep 04 13:49:52 <_phil> PaulFertser: will reformat it. No, without the kernel complains about overlapping memory on boot Sep 04 13:52:49 _phil: i wonder why Andy wrote it that way... Sep 04 13:53:14 _phil: you might want to do it in two separate commits even: one for bug-fixing, one to add gta02 configuration Sep 04 13:53:46 <_phil> PaulFertser: k, where to send'em? Sep 04 13:54:39 _phil: if you're subscribed to openmoko-kernel, then there. If not, mail me directly. fercerpav (at) gmail (dot) com Sep 04 14:02:37 ok, at least shr boots from nand, but qi ignores my sd -.- Sep 04 14:24:40 which irc client do zou use on your freerunner Sep 04 14:25:28 andi: I mostly ssh home to use irssi :( Sep 04 14:25:37 andi: xchat does not handle 30+ channels well Sep 04 14:27:41 the gtk1 version was faster Sep 04 14:29:39 running irsi in a screen session I guess Sep 04 14:37:59 pingdin can do irc too Sep 04 14:45:03 xchat here Sep 04 14:46:42 after adjusting font to 3pt and siabling menubar it's ok for me Sep 04 14:47:15 s/siab/disab/ Sep 04 14:47:16 DocScrutinizer meant: after adjusting font to 3pt and disabling menubar it's ok for me Sep 04 14:51:46 andi: please report on misguiding and "noise" on your compass hack. Is it linear wrt heading vs readout? does it jump on e.g. changing backlit brightness / tx action on modem? Sep 04 14:53:10 andi: as there's vibrator motor next to the place you put that compass chip I'd expect severy missguiding. can it be compensated? Sep 04 14:53:46 serverE Sep 04 15:04:27 andi: you also might want to shield/remove the speaker as its magnetic field is much stronger still than vibrator motor Sep 04 15:04:47 But should be constant Sep 04 15:05:07 vibrator motor will come to rest in various orientations, and probably significantly modulate the field. Sep 04 15:05:07 yup Sep 04 15:05:11 Even when stationary Sep 04 15:05:40 you can calibrate out stationary fields fairly well - require the user to turn in a circle three times forex Sep 04 15:05:41 that's why I asked if it can be compensated or simply drives compass into saturation Sep 04 15:05:50 ah Sep 04 15:08:38 as long as you get a non-ambiguous relation between heading and readout you always can compensate/calibrate Sep 04 15:09:57 but speaker magnetic field is so much stronger than earth field... I seriously doubt this will fly. So any comment of andi highly welcome Sep 04 15:12:44 in the end if andi is wildly determined to make that work, he can discard vibrator and replace speaker by a piezo Sep 04 15:12:55 i'm running the post-installation script for SHR-unstable. i need to change the VLANGUAGE entry from de to a british one. do i use uk or gb or en-gb? Sep 04 15:15:03 * DocScrutinizer seems to remeber some postinstall script from wiki somewhere, which was discussed and mostly considered obsolete or even harmful, a few days ago Sep 04 15:15:36 * DocScrutinizer might be wrong though Sep 04 15:16:24 DocScrutinizer, really? hmm, maybe a good thing i stopped it then. i found it here, http://wiki.openmoko.org/wiki/SHR_post-installation Sep 04 15:16:48 check from when it dates Sep 04 15:17:20 DocScrutinizer, 25-8-2009 Sep 04 15:17:20 probably old and bitrotten Sep 04 15:17:25 ooh Sep 04 15:17:35 no comment on that then Sep 04 15:17:46 nevermind my remarks Sep 04 15:18:31 DocScrutinizer, ok, np... thanks. you don't happen to know what the vLANGUAGE entry should be for the uk do you? Sep 04 15:18:52 not from top of my head Sep 04 15:18:53 DocScrutinizer, by default the scfript is set to de for germany. Sep 04 15:19:09 so try en I guess Sep 04 15:31:35 can anyone confirm that "en" is the correct code to put under vLANGUAGE? Sep 04 15:32:31 PaulFertser: could you join #arch-arm for the issue with ubifs? mounting doesn't work Sep 04 15:36:52 for anyone interested, it appears to be "en-gb" Sep 04 16:51:23 http://www.electronicsproductionworld.com/articleView~idArticle~71680_173413417372008.html Sep 04 16:51:25 oops Sep 04 17:18:24 DocScrutinizer,SpeedEvil: compass measurement errors seem to be compensable, there is a non-ambiguous relation between heading and atan(y,x) when I rotate the freerunner on the ground of a parking lot (away from metal things) Sep 04 17:18:52 what influence backlight, the vibration motor and similar stuff has, I have still to test Sep 04 17:20:04 * SpeedEvil tries to recall where he put that paper. Sep 04 17:21:31 Somewhere I have a nice paper on calibration of a magnetometer in proximity to soft and hard errors Sep 04 17:24:10 offtopic strawpoll: WinCE + Mozilla/Webkit + flash player? Sep 04 17:24:25 yes, an open an arbitrary question. Sep 04 17:28:52 syntax error Sep 04 17:29:44 but from that selection only webkit has some appeal to it Sep 04 17:36:28 is it correct that .bash_history is named .ash_history on the FR? Sep 04 17:50:18 is it possible to install apps to SD? Sep 04 17:50:32 yes (how depends on distro) Sep 04 17:50:51 you can use symlink comedy. Sep 04 17:50:55 mjr: that was for me? Sep 04 17:51:12 and there are various things like 'stow' which can do tricks in regular distros Sep 04 17:51:32 Chocky: :) i'm aware of the comedy...no official solutions or recommendations in this area? Sep 04 17:51:41 used to do that in Zaurus Sep 04 17:51:47 not off the top of my head, no. Sep 04 17:51:56 i guess that's why it's recommended ot install to SD and use qi Sep 04 17:52:11 yes, when I did OO.org for the NetBook, it had to be on a CF card, so I used the symlink comedy to run it Sep 04 17:52:43 or you could put /usr on SD; that would be a cheap work around. Sep 04 17:57:01 who recommends using Qi? Sep 04 17:57:14 its developer ;P Sep 04 17:57:43 DocScrutinizer: some of the wiki pages and some folks here...it's said that it helps increase suspend / resume times and boot times Sep 04 17:57:58 Undrwater: anybody can edit wiki :) Sep 04 17:58:00 that's a lie Sep 04 17:58:05 basically Sep 04 17:58:30 i never noticed increased anything...but i wasn't about to call anyone a liar ;) Sep 04 17:58:39 the Qi vs uBoot war is somewhat... err religious Sep 04 17:58:43 i prefer uboot 'cause i can choose the partition i boot Sep 04 17:58:53 ahh...like emacs vi? Sep 04 17:58:59 kde gnome Sep 04 17:59:00 yep Sep 04 17:59:03 yep Sep 04 17:59:11 christianity, islam? Sep 04 17:59:27 shall we bomb the Qi people? Sep 04 17:59:36 toilet paper them Sep 04 17:59:44 apparently they have longer noses Sep 04 17:59:56 well i bought into that... the wiki did tend to big up qi... ;) Sep 04 17:59:59 or crooked pinky toes Sep 04 18:01:01 qi was the solution to all openmokos problems - the bootloader that would finally make it user-ready. Sep 04 18:01:04 or something. Sep 04 18:01:22 i understand that...it hides everything "booty" away Sep 04 18:01:30 like a consumer devic Sep 04 18:01:32 e Sep 04 18:01:38 from the wiki... "Qi is the recommended bootloader for the Neo FreeRunner. Installing it will result in faster booting. If you want to have multiple operating systems on your phone you should still stick with U-Boot, which is the older bootloader." Sep 04 18:01:54 "older" :) Sep 04 18:02:04 isn't it Sep 04 18:02:44 alternate is a better word Sep 04 18:03:05 or even more feature full Sep 04 18:03:24 that it is Sep 04 18:07:51 featureful Sep 04 18:08:04 "feature full" would mean something else Sep 04 18:09:25 still only relevant bootloader feature is to get the OS running Sep 04 18:10:53 And to resume Sep 04 18:12:57 resume? Sep 04 18:13:09 suspend to disk? Sep 04 18:13:17 The processor is 'off' when suspended to RAM Sep 04 18:13:34 It's essentially a normal boot up until something notes the 'suspended' flag Sep 04 18:13:53 ok Sep 04 18:14:09 i've put a swapfile on my sdcard, when FR boots it reports it has holes in it. should i worry? :) Sep 04 18:14:48 LiL_mr-t: probably not since the media doesn't suffer from fragmentation Sep 04 18:15:26 It just means swapping in and out is slower afaik Sep 04 18:15:41 Ke, do you know why it occurs? Sep 04 18:15:57 swap file is fragmented Sep 04 18:15:58 Does it say sparse? Sep 04 18:16:02 brb... gotta get a cat out of a wardrobe... Sep 04 18:16:24 =D Sep 04 18:16:30 I suggest a flamethrower. Sep 04 18:16:41 will never see him again Sep 04 18:17:13 lol Sep 04 18:17:20 hmm or would that message include refusal to use swapfile at all Sep 04 18:17:35 SpeedEvil: don't believe it does (say sparse that is) Sep 04 18:17:47 used truncate to create it? Sep 04 18:18:23 swapon: swapfile has holes Sep 04 18:18:26 Ke: linux won't activate swap on a file that has holes. Sep 04 18:18:46 yes Sep 04 18:19:01 Sparse files mean a file that is created to a given size - but the storage is not allocated. Sep 04 18:19:15 For example - open a file - seek to a gigabyte - write one byte - close it Sep 04 18:19:27 :o Sep 04 18:19:30 On most unix filesystems, that will result in a disk usage of 4k Sep 04 18:19:30 i also get a fat: invalid media value at boot Sep 04 18:19:36 SparseEvil Sep 04 18:19:41 don't know if that relates Sep 04 18:19:47 (or whatever the block size is) Sep 04 18:20:05 The file only has storage allocated when a block is actually written Sep 04 18:20:14 This - for obvious reasons - sucks with swap Sep 04 18:20:46 how did you make the swapfile? Sep 04 18:21:24 SpeedEvil: dd if=/dev/zero of=/swapfile bs=1024 count=65536 Sep 04 18:21:43 that should not make a sparse file Sep 04 18:21:59 SpeedEvil: then added echo "/swapfile swap swap defaults 0 0">> /etc/fstab Sep 04 18:22:22 SpeedEvil: finally, mkswap /swapfile Sep 04 18:22:53 http://www.google.com/search?ie=UTF-8&oe=UTF-8&sourceid=navclient&gfns=1&q=%22swapfile+has+holes%22 Sep 04 18:23:00 i followed the instructions over at http://wiki.openmoko.org/wiki/SHR_User_Manual Sep 04 18:24:32 I think fragmentation can also create files not valid for swaping Sep 04 18:24:46 or remember reading Sep 04 18:24:52 indeed Sep 04 18:25:32 Ke, the sdcard is formatted as ext2 Sep 04 18:25:38 oops, I was in scrollback Sep 04 18:26:06 The kernel soruce says that that message is a result of the bmap call returning 0 when asked for the block address of the first block of the file. Sep 04 18:26:41 So it's being asked for the block address of the first block of the file and being unable to find it Sep 04 18:28:16 incidentally, I once wrote a convenience wrapper for dd for truncating files to a given length (useful for making sparse empty files or clipping large files to size, for instance if you want to delete a large file on a sucky filesystem such as ext3 and don't want the process to bring file operations to a screeching halt ;] Sep 04 18:28:20 SpeedEvil: hmm. above and beyond me... (like most things!) Sep 04 18:28:22 http://mjr.iki.fi/software/trunc Sep 04 18:29:03 mjr: not truncate? Sep 04 18:29:14 SpeedEvil: so an issue with my sdcard? Sep 04 18:29:18 (and of course the kludge for the delete is to trunc the file smaller and smaller in increments) Sep 04 18:29:19 LiL_mr-t: no. Sep 04 18:29:25 LiL_mr-t: probably not Sep 04 18:29:26 Ke, nah, that wouldn't be truncated Sep 04 18:29:42 LiL_mr-t: I would first try making a new swapfile and see if that works Sep 04 18:30:02 LiL_mr-t: I would also check I hadn't done something ssutpid like disabled swap in the kernel config Sep 04 18:31:00 SpeedEvil: so go through the procedure again? how do i check if swap is diabled in the kernel config? Sep 04 18:37:48 LiL_mr-t: if you haven't played with the kernel or built your own, you didn't make that change Sep 04 18:38:24 Undrwater, i sure haven't... ;) Sep 04 18:43:42 what about a ls -l and stats /swapfile ? Sep 04 18:43:55 yeah Sep 04 18:43:57 and e2fsck Sep 04 18:44:10 that as well Sep 04 18:47:30 hrmmmm Sep 04 18:48:07 is there something wrong with usb in linux 2.6.30.5? i don't see my FR in lsusb output Sep 04 18:48:42 but when I manually(!) modprobe cdc_ether, i can then ifup usb0 and ssh to my FR Sep 04 18:48:54 for some reason I prefer swap *partitions* Sep 04 18:49:10 Wonka: hmm, usually not Sep 04 18:49:33 i _can_ see my webcam and my keyboard in lsusb output, though Sep 04 18:49:55 /proc/bus/usb is not even there... that might be related to kernel changes. Sep 04 18:50:12 but why do I see the keyboard and not the FR? Sep 04 18:50:16 *confused* Sep 04 18:50:43 DocScrutinizer, SpeedEvil, ls -l /swapfile = -rw-r--r-- 1 root root 67108864 Sep 4 19:40 /swapfile Sep 04 18:51:51 yeah - does e2fsck say it's a clean fs? Sep 04 18:51:56 LiL_mr-t: stat /swapfile ? Sep 04 18:52:04 Wonka: odd Sep 04 18:53:15 DocScrutinizer, stat not found Sep 04 18:53:22 :-/ Sep 04 18:53:38 I *HATE* BUSYBOX Sep 04 18:53:56 uh? Sep 04 18:53:56 well stopit Sep 04 18:53:58 SpeedEvil: e2fsck 1.38 (30-Jun-2005) Sep 04 18:53:58 Couldn't find ext2 superblock, trying backup blocks... Sep 04 18:53:58 e2fsck: Bad magic number in super-block while trying to open /swapfile Sep 04 18:54:06 compiled 2.6.28.10, booted... Sep 04 18:54:11 err - no - the filesystem it resides on Sep 04 18:54:22 now I see a friend's FR with USB ID 0525:a4a2? Sep 04 18:54:29 SpeedEvil: sorry... Sep 04 18:54:47 my own one is 1457:5122... Sep 04 18:55:23 btw 67M seems a little stingy for a sap Sep 04 18:55:27 sWap Sep 04 18:55:56 DocScrutinizer: I dunno - without actual ordered swap - you're going to take several minutes to fill that Sep 04 18:56:39 then it's worth nottin at all. Consider FR has 128M ram iirc Sep 04 18:56:42 SpeedEvil: Couldn't find ext2 superblock, trying backup blocks... Sep 04 18:56:42 e2fsck: Bad magic number in super-block while trying to open /dev/mmcblk0 Sep 04 18:57:22 whThat is usually indicative that you've tried to r2fsck the wrong thing. Sep 04 18:57:26 heh Sep 04 18:57:30 DocScrutinizer, yeah, think that's a test example on the wiki Sep 04 18:57:32 Are you sure tha'ts the right filesystem? Sep 04 18:57:41 Also - this was e2fs/? Sep 04 18:57:46 is Sep 04 18:57:53 filesystem comedy Sep 04 18:58:03 SpeedEvil: LiL_mr-t: N900 maemo e.g. has 256M ram and 3 times that for swap Sep 04 18:58:17 SpeedEvil, i ran fdisk -l, is that correct? Sep 04 18:58:39 DocScrutinizer: yes - I remain to be convinced that without swap patches it's actually useful in the face of huge block erase latecnies. Sep 04 18:59:04 DocScrutinizer: I assume you don't have hardware? :) Sep 04 18:59:21 huh?? Sep 04 18:59:25 DocScrutinizer: did you see that swap thing I was posting the other day - the swap paper. Sep 04 18:59:34 n900 Sep 04 18:59:35 no Sep 04 18:59:56 don't have hardware??? Sep 04 18:59:57 http://www.usenix.org/event/hotos09/tech/full_papers/saxena/saxena_html/ Sep 04 19:00:10 DocScrutinizer: SpeedEvil: LiL_mr-t: N900 maemo e.g. has 256M ram and 3 times that for swap Sep 04 19:02:27 SpeedEvil: should the sdcard be listed by fdisk -l ? Sep 04 19:02:29 SpeedEvil: ""We find that software latencies for a page fault could be as high as the time taken to read a page from flash..."" WTF isn't that plain self-evident? Sep 04 19:02:41 DocScrutinizer: err - no Sep 04 19:03:35 DocScrutinizer: At least not to me. I'd assumed it would be substantially faster. Sep 04 19:04:30 SpeedEvil: when page fault == need to swap in a page from flash (what is the definition of page fault it seems to me) then how could it ever be faster than what it is defined to do? Sep 04 19:05:43 The total time before it begins to swap in from flash after the page fault occurs exceeds the total time to read one block from flash. Sep 04 19:05:51 Which was surprising to me. Sep 04 19:06:16 that's a completely different statement and it's true Sep 04 19:06:34 I disagree it's a different statement. Sep 04 19:06:55 software latencies do not include hardware read times Sep 04 19:07:10 DocScrutinizer: I hate busybox as well :) Sep 04 19:07:23 I keep typing stuff like ps aux on my OpenWRT ;) Sep 04 19:07:49 SpeedEvil: is fdisk -l the correct command to locate my sdcard? Sep 04 19:07:59 well maybe that's my poor English, but ""could be as high as the time"" suggests to me "we thought it's lower but actually it can go up to..." Sep 04 19:08:06 LiL_mr-t: maybe. Sep 04 19:08:24 lower==faster Sep 04 19:08:51 LiL_mr-t: try df -h Sep 04 19:09:10 DocScrutinizer: It is inexact - but clear in its meaning to this english speaker. Assuming you understand that 'software latencies' would never include hardware read times - as I did. Sep 04 19:09:13 or mount Sep 04 19:09:28 so argue. Sep 04 19:10:14 DocScrutinizer, mount shows, /dev/mmcblk0p1 on /media/card type ext2 (rw,errors=continue) Sep 04 19:10:32 SpeedEvil: you're right Sep 04 19:11:33 SpeedEvil: I missed the "*software* latencies" Sep 04 19:12:23 DocScrutinizer, df -h shows /dev/mmcblk0p1 972404 1272 921736 0% /media/card Sep 04 19:12:46 LiL_mr-t: so you know where your uSD is Sep 04 19:14:24 SpeedEvil wrote "whThat is usually indicative that you've tried to r2fsck the wrong thing." does this mean i've done something wrong then? Sep 04 19:14:28 seems to be a 1G card, with absolutely nothing on it! Sep 04 19:14:47 That is usually indicative that you've tried to e2fsck the wrong thing Sep 04 19:14:56 I was typing with my eyes shut. Sep 04 19:15:20 I should really look at the screen occasionally. Sep 04 19:15:27 e2fsck /dev/mmcblk0p1 Sep 04 19:15:29 also Sep 04 19:15:32 that's irrelevant Sep 04 19:15:35 SpeedEvil, i'm confused? Sep 04 19:15:48 you've created the 64M swapfile on the root filesystem not your SD card Sep 04 19:16:13 ah! Sep 04 19:16:17 i get that! Sep 04 19:16:21 do the dd if=/dev/zero of=/media/card/sw... thing Sep 04 19:17:22 and then do mkswap /dev/mmcblk0p1/swapfile Sep 04 19:17:38 yo Sep 04 19:17:54 lilmer no Sep 04 19:18:01 /media/card/swapfile Sep 04 19:18:11 /dev/mmcblk0p1/swapfile Sep 04 19:18:14 aah :-S Sep 04 19:18:21 will say no such file or directory I think Sep 04 19:19:00 sure will.. phy dev Sep 04 19:19:34 no fs on /dev Sep 04 19:21:08 still I suggest you repartition uSD to have a small swap partition at end and use that rather than swapfile Sep 04 19:22:03 SpeedEvil, so should i type /media/card/swapfile or /dev/mmcblk0p1/swapfile ? Sep 04 19:22:10 and still I suggest to make swap same size as ram at least Sep 04 19:22:19 Ok, I am going to attempt to make my own USB female to mini male adaptor, I have a mutilated wire with a mini end, and a usb port, What faces the most damage if I do this wrong, the moko or the device I test this with? And hints are much appreciated. Sep 04 19:22:40 /media/card Sep 04 19:22:48 there's no such thing like /dev/mmcblk0p1/swapfile Sep 04 19:23:46 ok Sep 04 19:24:30 SeligArkin: buy ready made adapters. It's simply not worth the effort. And for Ur Q: FR will suffer most, prolly will incinerate Sep 04 19:24:30 DocScrutinizer, so how should dd if=/dev/zero of=/swapfile bs=1024 count=65536 look to use the whole card? i gather i change count= ? Sep 04 19:25:06 LiL_mr-t: U don't want to use whole card Sep 04 19:26:07 and U need to use /nedia/card/swapfile (if you go for file rather than partition) Sep 04 19:26:29 double the count Sep 04 19:26:47 (or the bs ;-) Sep 04 19:53:11 <[Rui]> ah, the sun and beach was awesome today :) Sep 04 19:53:40 <[Rui]> and now I have a 3G card, I will be able to submit some 300 km or so of data for openbmap Sep 04 19:55:18 now just squeeze the card into the FR Sep 04 19:58:24 <[Rui]> Chocky: which won't do anything interesting with it? :) Sep 04 19:58:34 ;-) Sep 04 19:58:59 the internals of the UM175 are pretty small, as long as you don't mind being tied to Verizon Sep 04 20:03:54 still a touch big for existing space in FR, unless you had a smaller battery, and then of course, the extra drain would kill your logetivity Sep 04 20:11:28 3g cards are almost small enough. Sep 04 20:11:46 I saw one the other day that was ~22*48*4mm Sep 04 20:13:36 Though I ended up getting a SIM for my existing 3G modem which turns out to have been locked Sep 04 20:13:59 the UM175 was very close to that size Sep 04 20:14:22 the ones they put inside laptops are pretty similar too Sep 04 20:14:35 http://www.t-mobile.co.uk/shop/mobile-broadband/data-plans/pay-as-you-go/pay-as-you-go/daily-mobilebroadband-plus/t-mobile/usb-stick-150/overview/ Sep 04 20:16:24 Though I don't know the modelnumber Sep 04 20:16:33 <[Rui]> http://maps.google.com/maps/ms?ie=UTF8&hl=pt-PT&msa=0&msid=110365518025191930255.000472c5f214b8310964c&ll=37.102015,-8.208847&spn=0.122672,0.308647&t=h&z=12 Sep 04 20:17:58 http://cgi.ebay.co.uk/1000x-NEW-UNLOCKED-HUAWEI-E156-3G-HSDPA-3-6mb-USB-MODEM_W0QQitemZ370223194987QQcmdZViewItemQQptZUK_Computing_LaptopAccessories_PCMCIACards?hash=item563305a36b&_trksid=p3286.c0.m14 :) Sep 04 20:22:40 how do you view icons that are off the desktop with shr-unstable? Sep 04 20:23:18 SpeedEvil, by the way, got swapfile working. thanks for all your help. DocScrutinizer helped me get it to work Sep 04 20:24:00 SpeedEvil: it was basically down to my stupidity (as you had already worked out!) Sep 04 20:24:20 :) Sep 04 20:32:25 http://people.openmoko.org/joerg/unsorted/FR+UMTS.jpg Sep 04 20:34:11 :) Sep 04 20:34:43 SeligArkin: ^^ Sep 04 20:35:26 ))) Sep 04 20:35:43 yes, comedy Sep 04 20:36:22 6 x lifesize! Sep 04 20:39:51 SeligArkin: http://www.conrad.de/goto.php?artikel=971535 Sep 04 21:03:56 fdisk -l /media/card ?? Sep 04 21:05:08 slaxxer: ??? Sep 04 21:16:38 DocScrutinizer: That cable one - do ou know how much one of those would cost to get to the US? Thats the kind I was trying to make, like the idea of allowing some distance. :P Sep 04 21:17:24 SeligArkin: you should find those in US as well, for sure Sep 04 21:17:42 they all come from China :p Sep 04 21:17:54 Chocky: :P Sep 04 21:22:09 SeligArkin: also the female2female adapter you find on the photo is very handy, as you convert every simple standard USB-A-m--miniUSB-m cable to the f--mini-m type you want Sep 04 21:43:53 Hi! I installed the latest testing release of shr, started pidgin and got sound. Now I have installed the latest unstable, and got no sound. I would like to have sound. How? Sep 04 21:52:38 TeNoVi: hmmm Sep 04 21:52:42 TeNoVi: did you install navit? Sep 04 21:52:46 Did you reboot? Sep 04 21:53:25 Ainulindale, I haven't installed navit. Pidgin ;) Yes I have rebooted Sep 04 21:53:54 TeNoVi: I'm asking that because known issue wrt speech-dispatcher, a navit dependency Sep 04 21:53:58 You don't have pidgin sounds then? Sep 04 21:54:59 Ainulindale, I added some sounds from amsn and configured them in the prefs.xml since I couldn't get the regular sounds working Sep 04 21:55:25 amsn? Sep 04 21:55:36 Why don't you use msn-pecan with pidgin? Sep 04 21:57:24 Ainulindale, Hmm. Haven't heard about msn-pecan, or just a little maybe Sep 04 21:57:35 well Sep 04 21:57:39 I packaged it for SHR a while ago Sep 04 21:57:44 It's installed as a default on full image Sep 04 21:57:56 As an alternative to the default MSN provider in libpurple Sep 04 21:58:01 As it's way better... Sep 04 21:58:11 hmm Sep 04 22:01:55 Ainulindale, Could msn-pecan get the sound to work you think? Sep 04 22:02:13 I don't think it's related Sep 04 22:02:21 I don't use pidgin myself anyways Sep 04 22:02:29 okey Sep 04 22:03:44 cu@all Sep 04 22:47:39 openmoko: 03daveb * r5604 10/trunk/gta02-core/ (cpu.sch docs/ecn/ecn0030.txt gta02-core.sch modem.sch): modem component references, most things looking pretty good now. Sep 04 22:52:33 freesmartphone.org: 03mickey 07cornucopia * r44932ba654a8 10/fsousaged/ (3 files in 2 dirs): Sep 04 22:52:33 freesmartphone.org: fsousaged: add (optional) libpersistence support to improve restartability Sep 04 22:52:33 freesmartphone.org: NOTE: configure with --enable-persistence, if you want to enable this feature. Sep 05 00:30:10 hi. how do i achieve the following from the command line? "Now you can use network-manager with mac-address specific settings and get it to automatically connect. " Sep 05 00:50:32 freesmartphone.org: 03mickey 07cornucopia * rc45f7dc4c6d2 10/fsodeviced/src/plugins/kernel_input/plugin.vala: fsodeviced: kernel_input: send coldplug input notifications on startup. Fixes FSO #239 Sep 05 01:40:32 how do you get the toolbar back in midori? the one that shows file, tools etc Sep 05 01:45:42 how do you right click? Sep 05 02:01:33 does anyone know how to right-click with the FR? Sep 05 02:55:47 to partition the microsd card on my fr, can i do it without ssh'ing to it? **** ENDING LOGGING AT Sat Sep 05 02:59:58 2009