**** BEGIN LOGGING AT Sat Aug 18 03:00:00 2012 Aug 18 06:50:16 what's the best way to run ssh-agent in maemo? Aug 18 08:34:20 freemangordon: I've been looking through romaxa's mercurial code repo on hg.mozilla.org. I thought I'd have a go at building the IPCLite based browser. Have you tried? Not sure which project to checkout: http://hg.mozilla.org/users/romaxa_gmail.com/ Aug 18 08:38:29 I'm guessing I need his xulrunner+qtmozembed..? Aug 18 12:47:11 Hi,nandtester here-I finally managed to get the battery charged...unfotunately I'm having some problems trying to cold flash Aug 18 12:47:24 I have followed these instructions: http://wiki.maemo.org/Updating_the_firmware/Cold_Flashing Aug 18 12:48:25 oh, hi Aug 18 12:48:31 i kinda forgot about you! Aug 18 12:48:32 :D Aug 18 12:48:41 but after the temporary bootloader has been loaded & I enter the flasher-3.5 -F image.bin --flash-only=nolo -f it gets stuck Aug 18 12:48:58 at this point Suitable USB device not found, waiting. Aug 18 12:49:08 at this point: Suitable USB device not found, waiting. Aug 18 12:49:20 well, did you connect your n900 at that point? Aug 18 12:49:42 also, is 2101 your hw revision? Aug 18 12:49:55 DocScrutinizer05: ping, i think you're interested too Aug 18 12:50:00 I first enter the boot-loader command & then connect the phone Aug 18 12:50:10 DocScrutinizer05: and besides, you can help him better than me Aug 18 12:50:19 sambo7: does the n900 do something at that point? Aug 18 12:50:20 after a while the Nokia logo appears Aug 18 12:50:34 hm, maybe you should hold "u" then? Aug 18 12:50:55 don't know which hw rev I actually have Aug 18 12:51:24 thought that I might find out with lsusb when phone is recognised but I didn't Aug 18 12:52:31 after the bootloader has been loaded this pops up in the syslog: hub 4-1:1.0: connect-debounce failed, port 3 disabled Aug 18 12:53:00 jacekowski: are you around? Aug 18 12:53:19 this comes up a few times & in the end I get this in the syslog: usb 4-1.3: USB disconnect, device number 7 Aug 18 12:53:45 so I guess that the connection is lost after the bootloader & is not established anymore Aug 18 12:54:24 the amber led is still on for a while but nothing happens anymore Aug 18 12:54:40 :( Aug 18 12:55:16 how many hw rev are there....maybe I try one after the other Aug 18 12:56:11 i have 2104 Aug 18 12:58:15 when I look at what pops up after the second flash command, there seems to be more that 30/40/50 hw revisions Aug 18 12:58:23 I will try with holding the u button Aug 18 12:58:44 sambo7: http://talk.maemo.org/showpost.php?p=776474&postcount=35 Aug 18 12:59:43 it could take a while to test all of them...if that is relevant at all Aug 18 13:00:21 anybody ever tried to flash with wrong hw rev. given?result? Aug 18 13:03:37 get the same result with the u button Aug 18 13:04:31 after the boot loader is loaded the usb connection is lost & the flash command stops at Suitable USB device not found, waiting. Aug 18 13:07:13 anybody ever succeeded to cold flash a device & is that just pure theory?:-( Aug 18 13:09:57 jacekowski DocScrutinizer05 DocScrutinizer DocScrutinizer51 Aug 18 13:11:42 kerio: ? Aug 18 13:12:17 sambo7: don't distonnect the phone Aug 18 13:12:21 sambo7: it has to stay connected Aug 18 13:12:37 had it connected for about 20min Aug 18 13:12:59 after the boot loader nothing happened...wait some longer? Aug 18 13:13:02 sambo7: you need switched off phone, run coldflashing command connect phone, insetrt the battery Aug 18 13:13:10 then run normal flashing command Aug 18 13:13:19 linux or windows? Aug 18 13:13:25 linux Aug 18 13:13:48 hmmm Aug 18 13:13:55 try different machine Aug 18 13:13:57 or different cable Aug 18 13:14:03 cannot switch on phone at all...nandtested it:-( Aug 18 13:14:10 jacekowski: this is the phone with the borked CAL Aug 18 13:14:14 could it be related? Aug 18 13:14:15 ahm Aug 18 13:14:19 broken cal Aug 18 13:14:36 after the boot loader I see usb diconnection in the syslog Aug 18 13:14:43 no way to recover that without phoenix Aug 18 13:14:54 what is that? Aug 18 13:15:26 jacekowski: hm, are you sure about that? shame :( Aug 18 13:15:36 nolo reads stuff from cal Aug 18 13:15:48 and probably if cal is gone then partition table is gone Aug 18 13:15:49 jacekowski: is there a way to rebuild a cal? Aug 18 13:15:52 and that's not recoverable as well Aug 18 13:15:56 kerio: theoretically Aug 18 13:15:59 after I enter the boot loader flash command & connect the phone the Nokia logo appears & boot loader is loaded...then disconnection Aug 18 13:16:14 sambo7: because bl is failing to read cal data Aug 18 13:16:24 it's looking for partition table Aug 18 13:16:27 and that's not there Aug 18 13:16:46 I nandtested mtd0-mtd5 accidentally Aug 18 13:16:54 "accidentally" Aug 18 13:17:03 sambo7: how, why? Aug 18 13:17:16 sambo7: you've done nolo nandtest? Aug 18 13:18:01 theoretically it is recoverable, but because we can't sign BL and therefore load any code onto n900 Aug 18 13:18:14 I was a little hasty...nandtested it before having read about that it doesn't test actually Aug 18 13:18:17 because nokia signed BL is crashing without valid cal Aug 18 13:18:26 it does test Aug 18 13:18:27 hey, it does test! Aug 18 13:18:32 it's destructive test though Aug 18 13:18:44 jacekowski: to be fair it should really print out more warnings Aug 18 13:18:54 yes,but without the -k not only test Aug 18 13:18:56 well, is that nolo nandtest we are talking about? Aug 18 13:19:11 or some other test? Aug 18 13:19:15 jacekowski: no, nandtest from mtdtools Aug 18 13:19:15 mtd-tools? Aug 18 13:19:19 I did enter nandtest mtd0 Aug 18 13:19:22 (destructive) Aug 18 13:19:25 the mtd1 Aug 18 13:19:29 till mtd5 Aug 18 13:19:48 well, problem is that all calibration data and imei was stored in cal Aug 18 13:20:31 so,no possibility to debrick it?what about the serial pins? Aug 18 13:20:40 nope Aug 18 13:21:05 nokia probably has copy of calibration data Aug 18 13:21:19 or at least could calibrate it again Aug 18 13:21:20 so it could be recovered without it going back to them if they were willing to help Aug 18 13:21:21 nandtest should be called nandfuck Aug 18 13:21:40 sambo7: nobody said it's not destructive Aug 18 13:21:50 but nevertheless it was my fault Aug 18 13:22:05 sambo7: send a letter to elop! Aug 18 13:22:13 nokia would have to supply signed bl that would write cal data Aug 18 13:22:27 so I might try to contact NOKIA & tell them what happened to my phone.... Aug 18 13:22:37 jacekowski: what's stored in CAL? Aug 18 13:22:38 or you have to send it back and they will redo it Aug 18 13:22:42 kerio: everything Aug 18 13:22:51 kerio: imei, radio calibration data, wifi mac Aug 18 13:23:10 * kerio kinda wants to backup his CAL Aug 18 13:23:22 jacekowski: ooh, so you can change your imei by writing to CAL? Aug 18 13:23:27 kerio: if you can sign it Aug 18 13:23:40 oh, it's in the signed part? :c Aug 18 13:23:47 spoilsports Aug 18 13:24:01 basically all calibration data and imei is stored in one block that is sent to rapuyama on boot Aug 18 13:24:06 and rapuyama verifies it's signature Aug 18 13:24:15 but hmmm Aug 18 13:24:17 fucking rapuyama :( Aug 18 13:24:24 you could clone another n900 probably Aug 18 13:24:40 unless rapuyama has it's own cal as well Aug 18 13:24:44 which it probably has Aug 18 13:25:00 I got one on ebay uk a few days ago & am waiting for this one to arrive Aug 18 13:25:25 sambo7: well question is exactly what have you done Aug 18 13:25:30 to be honest Aug 18 13:25:44 when i was looking at it Aug 18 13:25:48 I entered these commands Aug 18 13:25:56 nandtest mtd0 Aug 18 13:26:02 coldflash bl should be fine with messed up cal Aug 18 13:26:04 nandtest mtd1 Aug 18 13:26:06 nandtest mtd2 Aug 18 13:26:07 nandtest mtd3 Aug 18 13:26:09 nandtest mtd4 Aug 18 13:26:09 nandtest mtd5 Aug 18 13:26:25 it will not work without partition table Aug 18 13:26:25 after the last one the phone turned off Aug 18 13:26:31 hahahaha Aug 18 13:26:39 I bet that phone is busted Aug 18 13:26:47 sorry, but it is hilarious Aug 18 13:27:04 especially because i did the same thing on a sheevaplug Aug 18 13:27:07 with debian installed on the NAND Aug 18 13:27:07 there was great big warning when installing mtd-tools Aug 18 13:27:10 not so funny to me;-) Aug 18 13:27:27 and after i nandtested mtd2 shit started to blow up Aug 18 13:27:29 (rootfs) Aug 18 13:27:39 for me after mtd5 Aug 18 13:27:40 that's when i did "man nandtest" on my computer Aug 18 13:27:53 sambo7: yeah, the n900's rootfs is in mtd5 Aug 18 13:27:57 before all was fine Aug 18 13:28:23 sambo7: luckily for me, the sheevaplug has a low-level interface (jtag) to flash stuff to the NAND regardless of what you do with it Aug 18 13:28:29 what is the RM-number of N900? Aug 18 13:28:47 which is accessible via a mini-usb port Aug 18 13:28:54 flashing a n900 is... more complicated Aug 18 13:28:59 if it doesn't cooperate, i mean Aug 18 13:29:04 jacekowski: I knew that at one point. I forgot though. Aug 18 13:29:12 RM-number? Aug 18 13:29:23 yeah Aug 18 13:29:24 sorry Aug 18 13:29:24 from sticker Aug 18 13:29:30 there is rm-number Aug 18 13:29:36 i need that number Aug 18 13:29:41 sticker? Aug 18 13:29:55 rx-51 Aug 18 13:29:56 under tha battery? Aug 18 13:30:02 * SpeedEvil powers off. Aug 18 13:30:04 mine too Aug 18 13:30:25 rx51 Aug 18 13:30:36 Does anyone know how the web browser is able to dial phone numbers when you click on certain URLs? Aug 18 13:30:44 does the N900 have a jtag too? Aug 18 13:30:50 jonwil: "tel:xxx" URL scheme Aug 18 13:30:52 jonwil: I don't think it's that. Aug 18 13:30:55 oh Aug 18 13:31:00 yeah - that. Aug 18 13:31:15 I thought you meant the auto-dial thingy on any string of numbers in contacts Aug 18 13:31:17 what I mean is does anyone know what handles those URLs? Aug 18 13:31:30 I cant find the handler for it in microb-engine Aug 18 13:31:40 Not me, sorry. Aug 18 13:31:49 I want to do the same thing from my app :) Aug 18 13:31:50 on harm it's defined in a .desktop file Aug 18 13:32:06 jonwil: you can probably try to pass that url to hildon mime or similar lib Aug 18 13:32:30 sambo7: you can try flashing it with phoenix under windows Aug 18 13:33:07 where do I get that phoenix sw?not the nokia suite? Aug 18 13:34:03 ussualy some russian sites Aug 18 13:34:08 you have to find cracked version Aug 18 13:35:42 jacekowski: do you think that will restore CAL? Aug 18 13:35:59 well, phoenix is what nokia uses Aug 18 13:36:11 so i would expect that phoenix can recover completly dead phone Aug 18 13:36:51 looking forward to that then :P Aug 18 13:37:14 this is one of the reasons I'm personally quite happy with n9's openmode write-locking CAL Aug 18 13:37:44 well, you can lock nand blocks on n900 as ell Aug 18 13:37:45 well* Aug 18 13:37:48 found a cracked one...hopefully it runs on win7_64 or winxp Aug 18 13:38:07 jacekowski: have any hints on how that is done btw? I'm curious. It's a feature of the onenand chip? Aug 18 13:38:13 yes Aug 18 13:38:24 but nandtest unlocks everything beforehand Aug 18 13:38:34 hm. Aug 18 13:38:46 that's why he managed to fuck it up Aug 18 13:38:52 because normally it's locked Aug 18 13:41:21 @jacekowski dzienkuje...sobacze co z tego wyjdzie....:-) Aug 18 13:41:51 ok, it definatly looks like its done through libhildonmime Aug 18 13:42:59 now to find the code in microb that calls libhldonmime Aug 18 13:50:55 hmm, need to read the hildonmime documentation Aug 18 13:59:09 ok, better yet, I found a dbus interface I can trigger :) Aug 18 14:06:50 com.nokia.CallUI.ShowDialpad seems to work :) Aug 18 14:12:13 now I know how to trigger a call if I decide to write this app I am thinking about :) Aug 18 14:17:42 have to leave right now...will be back in the evening hopefully to give you an update if recovering with phoenix works Aug 18 14:17:45 its an app to make it easier if I want to call around to a bunch of different outlets of to figure out if any of them have in stock Aug 18 15:22:51 is it a good idea to use an ssh agent on the n900 Aug 18 15:22:53 ? Aug 18 15:26:30 per se it's neither good nor bad. Aug 18 15:26:47 yeah i ssh to my n900 all the time Aug 18 15:52:09 andre__: well, it means you have to keep your public keys unencrypted, at least for a bit, on a portable device that can be stolen kinda easily Aug 18 15:56:56 you mean private keys I guess, and "unencrypted" isn't true either Aug 18 15:58:36 er, yes Aug 18 15:58:53 don't you need the full plaintext of those to connect? Aug 18 15:59:24 sure, but they don't have to be stored unencrypted on the EMMC. Aug 18 16:02:24 I mean dm-crypt/cryptsetup, truecrypt. Aug 18 17:23:36 oh, he left Aug 18 17:23:44 NIN101 et al: i meant unencrypted in ram Aug 18 17:23:54 root can happily access /dev/ram and /dev/kram Aug 18 17:24:20 ssh-agent will have to keep the private key unencrypted to use it Aug 18 17:35:27 i mean, i use ssh-agent on my laptop but it requires a password to access Aug 18 17:35:45 so you can't get to the plaintext without freezing the ram Aug 18 17:35:50 and it's a laptop, so it's quite hard Aug 18 17:36:41 the n900, on the other hand, is absolute crap wrt security Aug 18 17:58:11 I vaguely recall someone saying they;'d gotten ffull dick encryptipn working. Aug 18 17:58:17 Full disc encryption Aug 18 17:59:17 LOL Aug 18 18:02:32 SpeedEvil: full rootfs compression is slow enough already Aug 18 18:02:44 and it's not a matter of encryption Aug 18 18:02:55 it's a matter of ssh-agent requiring the *decrypted plaintext* in ram Aug 18 18:03:37 DocScrutinizer05: should i file a bug against rootsh and sudser for allowing passwordless root execution? it's a pretty big security flaw, after all Aug 18 18:03:47 i mean, HAM itself is fucking awful Aug 18 18:05:44 I got /home/ and MyDocs encryption at least. No point in encrypted /, and it would be a big pain. Aug 18 18:13:42 Hello #maemo! Aug 18 18:14:26 * Estel_ is back to "civilization", after almost 3 weeks vacations - now only another day of travelling, and "sweet home" Aug 18 18:15:43 just little question - have anyone experienced (or is experiencing) strange behavior of N900, that - after turning off - have problems with turning on again, due to starting in act-dead mode *without* charger connected? Aug 18 18:16:06 otherwise, it works 100% ok - including USB data connection (incl. hostmode) and charging Aug 18 18:16:06 i always wait for 60 sec to restart Aug 18 18:16:18 mine never starts immediatly after i shut it down Aug 18 18:16:36 Gh0sty, my N900, after power off and usual little waiting time, starts immediately on pressing power buton, but, in act-dead Aug 18 18:16:47 hm no Aug 18 18:16:49 (framebuffer messages indicates that it's act-dead) Aug 18 18:16:54 if I shut it down Aug 18 18:16:57 press power again Aug 18 18:16:59 does nothing Aug 18 18:17:04 wait 10 - 20 sec Aug 18 18:17:04 and after a while it turns itself off, due to no charger connected Aug 18 18:17:08 press power Aug 18 18:17:10 it starts ... Aug 18 18:17:32 well, mine is in such loop no way to turn it on, until I take battery off for few minutes Aug 18 18:18:07 then, upon inserting battery, I can turn it on normally. Situation repeats on next power off Aug 18 18:19:22 but, if I insert battery too early - for example, 30 seconds after removal, instead of waiting 2 minutes - it turns itself on instantly, when I put battery in Aug 18 18:19:30 in act-dead, again Aug 18 18:19:59 when it happens (vibration on battery inserted, and act-dead) I know that I can't turn it on for a life of me - need to take battery away and wait few minutes Aug 18 18:20:19 it looks like it need to wait until goldcap discharge, no idea why Aug 18 18:21:13 I would say something about shorted data+ abd - pins, so it think that it's connected to charger and enters act-dead, *but* they're not shorted, and when working, it detects data + and - correctly Aug 18 18:21:21 no problem with USB data and charging Aug 18 18:21:49 as said, 0 problems, until I need to restart or turn it off and on Aug 18 18:22:10 quite irritating, though - every reboot = taking battery out for few mins (or no turning on) Aug 18 18:22:26 any ideas, anyone? DocScrutinizer? Aug 18 18:23:01 it would look hardware'ish, if not for fact that hardware seems to be 100% OK via other tests (like usb connection) and it "smells" software'ish :P Aug 18 18:33:00 Estel_: was there no such thing in some early power kernel .. which was related to boost current/voltage after using hostomode :/ Aug 18 18:36:45 5~/win 22 Aug 18 18:42:44 Sicelo, no idea, but it happens also when I haven't used hostmode (since last "reboot") Aug 18 18:46:10 Estel_: i had that once, phone thinking the charger was plugged in, and entered act_dead mode on shutdown Aug 18 18:46:24 but i could turn it on afterwards, and it fixed itself Aug 18 18:51:25 kerio, hm, mine doesn't want to fix itself for a week now Aug 18 18:52:11 Estel_: reflash and restore Aug 18 18:52:18 good joke :P Aug 18 18:52:38 I haven't done full reflash on this device since one after buying it Aug 18 18:52:41 no srsly Aug 18 18:53:20 I haven't did it even when faced Most Irritating Issue (tm) with broken .ogg/theora playback, so I'm not going to do it now :P Aug 18 18:53:24 btw, ask ShadowJK how to properly reset the battery/charging chips Aug 18 18:53:59 kerio, I've already chatted with him about it. Reseting bq2x00 doesn't help here Aug 18 18:54:14 sucks to be you, i guess Aug 18 18:54:21 check the connector Aug 18 18:54:28 it looks like for some twisted reason it doesn't shut down properly, and think that it's in constant act-dead Aug 18 18:54:52 conenctor is ok, and as said, usb data and charging works fine in all humanly possible combinations Aug 18 18:54:55 0 issues Aug 18 18:55:06 detecting wall-charger too Aug 18 18:55:17 (i.e. D+/D- short detection) Aug 18 18:55:24 dick around with r&d mode maybe Aug 18 18:55:37 it screws up act_dead mode, doesn't it Aug 18 18:55:39 maybe, but R& doesn't touch act-dead afaik :P Aug 18 18:55:49 hm Aug 18 18:55:51 I may end up like reinob kicking this sucker out of device Aug 18 18:56:02 act-dead, not reinob (kicking) Aug 18 18:56:13 try turning it off from bootmenu Aug 18 18:56:41 fun fact - when this bug happens, bootmenu/backupmenu doesn't show up, as it doesn't in act-dead :P Aug 18 18:57:20 lol! Aug 18 18:57:45 DocScrutinizer: your best friend's phone is stuck in act_dead mode, wat do? Aug 18 18:58:20 Estel_: did you use bq24x00_charger? Aug 18 18:58:49 kerio, sure, but it happens even if I don't use it since power cycle Aug 18 18:58:51 or not? Aug 18 18:58:54 need to check it Aug 18 18:59:17 but it definitely happens, when I have BME enabled and bq2415 rmod'ed Aug 18 18:59:31 i had it happen when i used pali's module Aug 18 18:59:39 and i think bme was still running Aug 18 18:59:48 but I haven5 checked if it happen, if I don't load it at all during runtime Aug 18 19:00:10 kerio, I never use Pali's module with bme Aug 18 19:00:16 but, it's good guess anyway Aug 18 19:00:24 Estel_: does it think the charger is plugged in when the system is up? Aug 18 19:00:30 I though about that, but was curious why no one else reports it Aug 18 19:00:34 no Aug 18 19:00:44 only when disabled Aug 18 19:00:51 You know what? I'll test it now :P Aug 18 19:04:08 estel set r&d flag power-button Aug 18 19:04:41 if n900 STILL act dead then it is time for a reflash. Aug 18 19:04:59 vi__: too late Aug 18 19:06:09 WTS 1*N950, 1000 euro. Aug 18 19:07:08 WTB 1*N950, 3 euro. Aug 18 19:19:00 wtb Nokia, 8 euro. Aug 18 19:20:35 fuck nokia Aug 18 19:21:00 what happened? Aug 18 19:29:05 8 euro? Aug 18 19:29:19 Venemo_N9: the elopcalypse or however it's spelled Aug 18 19:30:06 but it's kind of old Aug 18 19:30:27 i'd totally buy Nokia for 19.95 € Aug 18 19:30:53 especially if billy mays made an infomercial about it Aug 18 19:32:21 kerio, so nothing new Aug 18 19:32:26 nope Aug 18 19:32:34 but vi__ is selling his n950 for 1000 EUR Aug 18 19:32:40 i offered 3 Aug 18 19:33:40 3 what? Aug 18 19:33:46 3€ Aug 18 19:33:51 he didn't accept :( Aug 18 19:39:01 lol Aug 18 19:50:18 kerio, confirming, this bug seems to have something to do with latest bq2415 things in KP Aug 18 19:50:40 Sicelo's note about similaritiy to ancient bug with 5V is also in line with it Aug 18 19:51:25 strange obserfvation - if I kinsert battery without bug manifesting itself (no vibration = no instant-turning-on in act-dead), I have something like 5 seconds to turn device on normally Aug 18 19:51:41 if I do it later, it boots up in act-dead again Aug 18 19:52:21 Estel_: well, at least it can be kicked into shape by bme Aug 18 19:52:29 by the way, act_dead mode is stupid Aug 18 19:52:30 this, + fact, that leaving device without battery for few minutes temporary fixes it (*if*I turn device on fast enough after iknserting battery) would suggest some capacitors getting charged on 5V Aug 18 19:52:40 and device thinking it's in act-dead Aug 18 19:52:44 kerio, yea, it is Aug 18 19:53:05 but bme doesn't kick it in shape 100%. It just allow normal turn off/on cycle Aug 18 19:53:27 but pulling battery out and inserting again + turning device on more than 5 seconds after it = bug kicks again Aug 18 19:54:09 no matter if bq2415 was used, or not. I suspect, that it have something to do with changes in KP itself, that allow Pali's charging module to work Aug 18 19:54:13 need to report it to him Aug 18 19:54:31 thanks for good advices, both of You, it was shot for ten points Aug 18 19:55:17 Estel_: oh btw vi__ said something too, not sure if you read the logs Aug 18 19:57:02 kerio: my suggestion was a bodge. Aug 18 19:57:07 k Aug 18 20:03:41 vi__, thanks anyway, will check logs Aug 18 20:34:37 I can't find it on Google: Does anybody know what it means when the N900 makes that deep "ping" sound on shutdown? I think it's some error indicator and I had it when the internal backup battery was broken. Is there any place where I can get more information? Aug 18 20:41:11 i just had a discussion with an android fanboy. he pointed out the benefits of LBE Privacy Guard app, in, for example, showing which app is accessing IMEI info, etc. Aug 18 20:41:45 btw, is there a way for us to monitor such stuff on maemo? Aug 18 20:48:01 Sicelo: the source code Aug 18 20:48:01 >:D Aug 18 20:48:54 :P Aug 18 20:50:03 what about dbus thingies maybe :-/ Aug 18 20:51:24 strace Aug 18 20:52:11 besides, wtf is an "app"? Aug 18 20:53:09 lulz, it's what makes android tick.. they always say we have thousand of apps :P Aug 18 20:53:35 yeah but Aug 18 20:53:38 like Aug 18 20:53:41 how many apps is aircrack-ng? Aug 18 20:53:48 it's at least 5 or 6 binaries Aug 18 20:58:33 lulz..just now i burst out laughing.. so this app of his says Jorte is accessing IMEI info. Jorte is a calendar. heh Aug 18 20:58:47 wft is it doing with IMEI then. Aug 18 20:59:07 * Sicelo gives his N900 some love :D Aug 18 21:00:11 maybe it just lists the permissions used Aug 18 21:03:29 do you want to know what the "app" is sending/receiving? just use wireshark or tshark Aug 18 21:04:41 or, again, just look at the fucking source code Aug 18 21:05:01 every package in extras* has the source code Aug 18 21:09:13 Estel_: ( but, if I insert battery too early - for example, 30 seconds after removal, instead of waiting 2 minutes - it turns itself on instantly, when I put battery in) bq24150 and/or 1707 PHY misconfigured. Probably a kernel problem Aug 18 21:10:08 actually it needs a few minutes battery removal to reset those chips to proper state. see: Aug 18 21:10:13 ~flatbatrecover Aug 18 21:10:13 Remove battery for 1 minute. Insert battery. Plug powered Nokia wallcharger to device. Watch steady amber. Let sit and charge. Do NOT try to boot. After 30 min, you got either a) a booted up N900, b) flashing amber which means you can boot, c) steady amber going off - in this case start over again with ~flatbatrecover Aug 18 21:18:01 DocScrutinizer05: a few minutes... what if one still has a good bupbat? Aug 18 21:18:34 bupbat doesn't power bq24150 nor 1707 Aug 18 21:18:43 just small capacitors Aug 18 21:22:36 c1113 150uF, c1146 10uF... Aug 18 21:23:32 you may short battery connector pins (in N900) with some conducting object to reduce time for discharging those Aug 18 21:26:39 but actually the faulty config *might* as well happen in TWL4030 GAIA, which afaik detects vbus, and that chip indeed is powered by bupbat Aug 18 21:27:25 and it's known you might configure it in a way it will forbid powerup of device Aug 18 22:09:55 btw all this (and more, worse) may happen if you accept "patches" to kernel like the I2C-locking disable patch introduced to make bq27200.ko work Aug 18 22:10:41 DocScrutinizer, this "i2c-locking disable" patch is a history since looong time ago, i.e. it doesn't exist in KP for ages Aug 18 22:11:00 that's my rationale to oppose any unreviewed untested patches to the lately discussed "cssu kernel" Aug 18 22:11:26 personally, I don't use bq27200 at all - reading all things from i2c - but afaik bq27200 may be used properly if loaded before or without bme Aug 18 22:11:29 Estel_: yes, I know. I spotted the problem and I made it vanish Aug 18 22:11:39 and bless You :) Aug 18 22:12:00 btw, it's sane to keep heavy experimental things like bq2415 charging outside cssu-kernel Aug 18 22:12:11 and afaik it was the plan from beginning? Aug 18 22:12:48 anyway, thanks for Your pointers where problem may lie in kernel, I'll forward it to Pali, too Aug 18 22:13:12 the severity/danger of undetected lateral effects of any patches can't be deduced from the primary nature of the patch Aug 18 22:13:17 as I'm 99,99% sure that it started with latest KP and my experiments with bq2415 module charging Aug 18 22:14:33 sure, sure, that's what KP is for (such experimental things like modules for bme-less charging) until they're rock solid. No one opposed way of thinking about kernel-power as extras-experimental :) Aug 18 22:15:03 and "kernel" for cssu as place for patches and things stable for ages Aug 18 22:15:29 what makes me wonder about this particular trouble, is that no one else reported it. I'm the only one using bq2415 for charging?... Aug 18 22:15:50 + reinob, but he purged half of blobs, including act-dead itself :P Aug 18 22:16:09 I2C-fuckup been considered "stable for ages" since nobody associated the problems it definitely introduced to the nature of the patch Aug 18 22:16:17 personally, I use bq2415.ko for charging exclusively Aug 18 22:16:43 DocScrutinizer, afaik it was never stable - just added in times of Titan, when he applied every path without review... Aug 18 22:17:01 and before people like You, freemangordon, Pali etc. reviewed whole KP Aug 18 22:18:12 next thing like that - some good in nature, but unfinished, unmaintained, and kinda random addon - highpass filter in kernel, submitted by luke-jr after new KP duo got estabilished, was detected in minutes Aug 18 22:18:17 well, before Titan "reviewed" whole KP. A single developer looking cursory across the source is not any sane review, no matter if done by me, Pali, FMG, or Titan Aug 18 22:18:32 Titan never reviewed it :P Aug 18 22:18:51 then, left without single note, which made it even harder to properly review Aug 18 22:18:59 ~seen a pair of boobs in my life Aug 18 22:19:03 it took some time for Pali, fmg and others to clear mess Aug 18 22:19:03 vi__: i haven't seen 'a pair of boobs in my life' Aug 18 22:19:52 DocScrutinizer, btw, afaik this highpass thing go detected by kp users not from everyday usage, but from reviewing code (!) and then quickly confirmed by fmg Aug 18 22:20:04 wasn't it vi_ who detected it? Aug 18 22:20:23 nfc what yiu're talking about Aug 18 22:20:41 kinda a feat, as path was described wrong way, and brief look at it wasn't resulting in suspicious observations Aug 18 22:20:54 there was some good idea path by... Nox? I don't remember Aug 18 22:21:04 kernel controlled filters, ages ago Aug 18 22:21:13 DocScrutinizer05: afaik you can implement an FIR filter on the sound chip...or something. Aug 18 22:21:21 main goal was to inroduce hardware equalizer, but idea got abandoned Aug 18 22:21:39 Estel_: the hardware already supports it. Aug 18 22:21:49 vi__: I know, as it's again been me who spotted that (well together with HNZ or what's been his name) Aug 18 22:21:54 then, luke-jr got old code from this abandoned project, messed it a little, and submitted to KP as ultimate speaker protection Aug 18 22:22:25 the problem is, that it was filtering low frequencies no matter if You use headphones or speakers or fmtx or... Aug 18 22:22:56 +, code was not maintained, no one wazs interested in maintaining it, and it wasn't submitted by original author :P Aug 18 22:23:10 i.e. standard for KP in Titan's times, with all due respect Aug 18 22:23:36 ~seen titan Aug 18 22:23:39 titan <~zero@pdpc/supporter/student/titan> was last seen on IRC in channel #debian, 862d 12h 36m 55s ago, saying: 'extra ips on the default interface'. Aug 18 22:23:57 that's why things like i2c-fuckup were included, then. But, hardware highpass got worse luck, as it was submitted in times of real maintainership over KP Aug 18 22:24:28 they re-made total recall. Aug 18 22:24:33 without arnie. Aug 18 22:24:34 well, our whole discussion about act-dead fuckup proves, that many people, even far frommbeing kernel hackers, are interested in hammering bugs :P Aug 18 22:24:44 blasphemy Aug 18 22:24:58 vi__, You're using thumb and KP Aug 18 22:25:07 havemYou tried bq2415 module? Aug 18 22:25:07 yes. Aug 18 22:25:10 for charging? Aug 18 22:25:14 nope Aug 18 22:25:50 well, I use it exclusively and it works great, especially for charging 3Ah behemot battery with 1250 mAh instead of 950 ;) Aug 18 22:25:56 but, introduced strange bug Aug 18 22:26:16 turn device off, and on every ON attempt, it thinks it's in act-dead Aug 18 22:26:28 act dead mode can suck my balls. Aug 18 22:26:35 or "it thinks it *should* go into act-dead" Aug 18 22:26:41 got rid of it? :) Aug 18 22:26:42 just set power_button r&d flag. Aug 18 22:26:49 problem solved. Aug 18 22:27:06 no more act dead Aug 18 22:27:08 ever. Aug 18 22:27:20 it solves charging in act-dead when connecting to charger? I've forget about it, totally Aug 18 22:28:37 well, honestly, act_dead is a biatch, no bootmenu, no backupmenu, no u-boot... I wonder if I should be happy that butbat is flat Aug 18 22:28:50 it depends, if it's gaia problem or this 1st one Aug 18 22:28:59 just set power_button r&d flag. Aug 18 22:29:02 no more act dead Aug 18 22:29:05 problem solved. Aug 18 22:29:19 yea, just after I'll help in tracoing down issue, as it should be fixed in KP nevertheless Aug 18 22:29:26 it is a kludge. Aug 18 22:29:30 idgaf. Aug 18 22:29:56 If my 5V capacitors are charged when device is off, it is a problem, no matter of act_dead Aug 18 22:30:00 I'd say the only thing this "discussion" shows is that patches going to KP do not see even the minimum of review and QA needed for kernel patches Aug 18 22:30:19 DocScrutinizer, every thing we discuss about is in -devel Aug 18 22:30:28 DocScrutinizer05: feel free. Aug 18 22:31:08 Estel_: and you think it will get review before moving from -devel to KP ? Aug 18 22:31:14 Well, replacing bme is such a thing, that I don't mind hammering a bug or two in the process :) Aug 18 22:31:25 sure, it's what I'm participating now, in a small part :) Aug 18 22:32:00 you obviously have no concept of a proper review process Aug 18 22:32:01 name a single bug in KP in extras (not fixed later in further releases)) Aug 18 22:32:29 ^^^ another evidence that you don't Aug 18 22:32:39 Estel_: poweroff on reboot bug. Aug 18 22:32:48 fortunately, I'm not talking about new android toy for masses, so I don't need to give a shit about corporate review standards, and can follow regular FOSS way of people contributing via using -devel... Aug 18 22:33:02 vi__, there are reports of it happening without KP Aug 18 22:33:12 i.e. on stock Aug 18 22:33:28 this one is a real biatch, probably result of many different modifications Aug 18 22:34:14 idgaf anymore Aug 18 22:34:32 If poweroff on reboot ever get fixed, I'm going to open '36 bootle of wine Aug 18 22:34:34 vine Aug 18 22:34:38 even Aug 18 22:35:15 (and hope that it's not acid yet) Aug 18 22:35:25 Estel_: that's bullshit. The regular FOSS way to contribute kernel patches is exactly what I refer to, demand, and see is not folllowed a singe bit for KP Aug 18 22:35:43 erm, like? Aug 18 22:36:22 rewgular (FOSS) way for review of kernel patches: publish patch on LKML, so several hundered *really* experienced experts for that particular topic can review and comment Aug 18 22:36:46 ough, several hundred kernel hackers? out of our all 2 of them? Aug 18 22:36:51 well, maybe 3 or 4 Aug 18 22:37:18 I think You've missed scale, better compare us to things like openwrt, than to upstream kernel patches Aug 18 22:37:32 you can't substitute that by "review" done by Pali, estel, DocS, FMG and declaring patch "rock solid" after a few months in some marginal test kernel used by a few hundered users Aug 18 22:37:35 not to mention that some things from KP already got upstreamed, IIRC Aug 18 22:38:01 theres no difference between KP and any other kernel Aug 18 22:38:05 few hundred users would be a heaven, as we have few hundred power users as whole Maemo base Aug 18 22:38:41 this day's lesson finished, afk Aug 18 22:38:56 well, there is difference - KP is targeted for single device users. It's even less than openwrt, which work on few dozens of devices and really got few thousand of active power users Aug 18 22:39:32 even they don't follow Your way, as it's not scalable, unfortunately, for less than few ten thousands active users Aug 18 22:39:51 what we have in KP looks *very* sane for resources that we have in hand, IMO Aug 18 22:40:28 today's polemic finished, going back to hammering out bq2415 bugs by extensive testing in all possible scenarios, before reporting :P Aug 18 22:58:39 sure, KP looks *very* sane for what KP is. KP is not the default kernel though, and no way to move KP to CSSU-kernel, even when we kick out some 10% of most suspicious patches, since the remaining 90% are as well without any sane review that's mandatory for a solid kernel meant for regular users. Your own words: "few hundred users would be a heaven, as we have few hundred power users as whole Maemo base". suggests this KP thing by no Aug 18 22:58:40 means is tested for CSSU-grade standards applicable to kernel Aug 19 01:33:18 Estel_: meh, I never said it was ultimate speaker protection; I said it was better than nothing at all, which it still is **** ENDING LOGGING AT Sun Aug 19 03:00:01 2012