**** BEGIN LOGGING AT Tue Mar 26 02:59:58 2013 Mar 26 05:28:24 Anyone using a Linux tablet? like a jailbroken Nook, Kindle, or the Samsung something-tablet II? Mar 26 05:31:20 kwtm, I'm using a Samsung Galaxy Tab 8.9. Mar 26 05:33:10 kwtm: It runs most "apps" fine, but any more than 10 tabs in Chrome and it starts to poo itself. Mar 26 05:33:45 Which I guess would be alright on a phone-class device, but isn't quite so on a tablet. Mar 26 05:36:41 tl;dr Avoid Tegra 2 at all costs Mar 26 05:54:28 tegra 2 is creaky and old Mar 26 05:54:38 and some early revs had critical errata Mar 26 09:02:03 hi. i have a weird problem on a friend's N900. It tends to close applications (most notably Modest), whenever the space key is pressed. So, he can't write emails containing spaces, etc Mar 26 09:42:26 Sicelo: check PR version Mar 26 09:43:27 I'd not be surprised to hear you found PR1.0.1 on that device Mar 26 09:44:40 it's 1.3. i had flashed it for him about a year or two ago Mar 26 09:44:41 you also could disable spellchecking and autocompletion in settings-inputmode Mar 26 09:45:10 (or somesuch) Mar 26 09:45:31 anyway I'd suggest a reflash Mar 26 09:46:54 we did :-/ Mar 26 09:47:15 although now i think he restored a backup Mar 26 09:47:48 and maybe mv ~user/.osso/dictionaries/.personal.dictionary and .used.dictionary to somewhere else Mar 26 09:48:45 if that still doesn't help, do a full reflash and check on a plain PR1.3.1 without any apps or settings restored Mar 26 09:50:31 I'm rather sure I heard about a similar bug in PR1.0.x with spellcheck or autocompletion (thus I never encountered it since I disabled both on day 1) Mar 26 09:52:05 ok. thanks for the pointer. will check Mar 26 09:54:47 Sicelo: of course enabling coredumps might also help to track down the issue Mar 26 09:55:15 how do i enable that? Mar 26 09:56:03 I think a package called rich-coredomps or similar should accomplish that Mar 26 09:56:15 s/o/u/ Mar 26 09:56:17 DocScrutinizer05 meant: I think a package called rich-curedomps or similar should accomplish that Mar 26 09:56:21 dang Mar 26 09:56:26 thanks. i'll check Mar 26 09:56:59 if he wasn't my friend, lol, i'd just let him be, and make an offer to buy it :p Mar 26 09:58:34 sp-rich-core ? Mar 26 09:59:35 http://maemo.org/packages/view/sp-rich-core/ Mar 26 10:00:35 http://maemo.org/packages/view/sp-rich-core-postproc/ Mar 26 10:44:30 Hurrian: Your Galaxy Tab is jailbroken? Mar 26 10:46:14 To be clear: I'm looking for something that has a shell with scripting (preferably bash) and some general programming language, preferably Python. Mar 26 10:49:23 I'm looking for educational kids' games; and I thought I'd get nice ones with MS Win or iOS, but I find that the apps there generally aren't what I want or pop up with ads to buy extensions to the games or just aren't customizable to the age group that I want. Mar 26 10:50:14 KDE games, on the other hand, have what I want so I'm trying to see if they can run on a tablet platform so I don't have to use the laptop for them. Mar 26 11:52:49 kwtm: it's rooted, and has a custom Android 4.2.2 ROM on it. Mar 26 11:53:23 bash/sh/ksh/csh work just fine on it without a GNU user land chroot, but it helps. Mar 26 11:53:24 Hurrian: ok. shell scripting? general programming language? I'm guessing you're using Cyanogen, but I'm not familiar enough with it to know its features.. Mar 26 11:53:32 Hurrian: Ah, thx. Mar 26 11:54:10 Yes, it's Cyanogenmod 10.1. CM adds neat features that aren't a massive departure from stock Android, but you'll eventually learn to love them. Mar 26 11:54:48 Take for example the editable power widget. It gives you the same flexibility the status area on Maemo has. Mar 26 11:55:28 After having to use it a few times in a pinch, having a flashlight toggle available is quite neat. Mar 26 11:55:53 kwtm: GCC works fine in a GNU chroot. Mar 26 11:56:36 Mind the free RAM though. You will eventually need to use a swapfile. Mar 26 11:57:29 …And yes, KDE SC runs fine under a loopback VNC connection. Do note that compositing is /not/ available. Mar 26 11:57:51 There is zero hardware acceleration for X11 programs running under the Android stack. Mar 26 11:58:06 Hurrian: Ah, having KDE is useful. At least they have nice dialog boxes callable from script. Mar 26 11:58:16 Keep that in mind if the kids' apps, or the kids require a low-latency solution. Mar 26 11:58:51 There's not a whole lot of animation. My goal is to have the usual apps, but launched by some quiz mechanism. Mar 26 11:59:11 E.g. "Want to play this Reversi game? First you have to answer this question: 3+5= ? " or something like that. Mar 26 11:59:32 ah. yup, it'll work out just fine, then. Mar 26 11:59:48 I thought you were trying to run Tux Racer under X11 on Android ;0 Mar 26 11:59:51 Right now he just whines, "I wanna play the game!" but if he sees that he has to answer questions (not just because daddy says so but because the computer won't let him otherwise), he might be motivated more. Mar 26 12:00:41 kwtm: do you need this process to be automated? Mar 26 12:01:36 I've toyed with launching programs under X11 before, and the simplest I've got it down to was running a shell script in the terminal emulator (under Android), which launches a chroot, starts X11, and then starts a game. User intervention is required to start the VNC client. Mar 26 12:03:20 Hurrian: automation preferred but not absolutely necessary. Depends on how much time I'd have to spend on it. Mar 26 12:04:06 Around 10 seconds at most every time you want to launch the X11 app. Mar 26 12:04:54 I simply launch the terminal emulator, run "sh /data/local/start-game", then switch to the VNC client. It's smooth sailing from there. Mar 26 12:51:09 Hello . is it possible to run maemo on other phones? Mar 26 12:51:14 i mean other nokia phones Mar 26 13:09:36 TheBlindOwl: no (well, usually anything is possible, but ...) Mar 26 13:11:35 Sicelo: Hello how are things going on in swaziland? Mar 26 13:14:19 wtf!! Mar 26 13:14:35 then it mean my initial suspicion was right Mar 26 13:14:37 :P Mar 26 13:14:49 there is only one blind owl, lol Mar 26 13:14:58 Swaziland is good :) Mar 26 13:15:09 what brings you to this channel? Mar 26 13:16:42 Sicelo: I'm ok Mar 26 13:16:47 Long story Mar 26 13:17:01 some kind of gang stole my money and phone in street Mar 26 13:17:04 i had SE p1 Mar 26 13:17:08 you remember? Mar 26 13:17:52 After that i bought a used n97 Mar 26 13:17:58 for 70$ Mar 26 13:18:12 i thought i can install maemo on it and get rid of its stupid os Mar 26 13:19:14 ok. well, quick answer is no. Mar 26 13:19:20 good to see you again though :) Mar 26 13:19:44 Sicelo: Thanks.i'm happy to see you again too.skype no longer works Mar 26 13:20:10 so n97 is crap? Mar 26 13:20:26 TheBlindOwl, you haven't used the N97 for long, have you Mar 26 13:20:43 N97 is crap, yes. Mar 26 13:22:18 Hurrian: I'm using it for about 3 weeks now Mar 26 13:22:26 dealing mostly with java apps Mar 26 13:22:35 ouch. Mar 26 13:22:48 you didn't opt to get an E7 or N8? Mar 26 13:23:11 Or N95 >.> Mar 26 13:23:24 Well.I was looking for a cheap device with QWERTY and wifi Mar 26 13:23:31 Aoyagi_joytop, battery life on WiFi isn't exactly great on the N95 Mar 26 13:23:32 like my beloved SE p1 Mar 26 13:23:49 e7 and n8 are expensive Mar 26 13:24:09 Hurrian: Battery is better than any modern "smartphones" ! :D I don't know about WiFi though, I don't use it. Mar 26 13:24:55 DocScrutinizer05: thanks a million about your dictionary suggestion.. looks like my friend is a happy man now, much to my sadness, lol :P Mar 26 13:25:00 so you mean i wasted some 70$? Mar 26 13:25:11 It's not -that- bad phone. Mar 26 13:25:15 Aoyagi_joytop: I'd still be using mine for phone calls, MP3s and photos if the flex didn't give up. Mar 26 13:25:20 Sicelo: :-D Mar 26 13:25:42 TheBlindOwl: you could've bought a cheap China android phone. Mar 26 13:25:46 Hurrian: I separated mine into two parts while in amok. Mar 26 13:26:19 I had only one nokia phone before. which is a 3310 that still works Mar 26 13:27:18 http://img.susepaste.org/images/c79dd983.jpg Mar 26 13:27:19 this Mar 26 13:27:37 meh, my 3310's battery has bloated beyond repair. Mar 26 13:28:57 My battery works for about 8 hours standby and 1 hour talktime Mar 26 13:29:01 it was stronger in past Mar 26 13:29:12 i'm having this guy for some 11 years Mar 26 13:31:06 3210 and 3310 are almost immortal. Mar 26 13:33:22 I enjoy using it in public :D Mar 26 13:33:24 not sure why Mar 26 13:35:59 need to find a 105 somewhere so i can use it after the n900 dies <.< Mar 26 13:36:30 jaska, I bet shipping from Asia would be more expensive than the phone. Mar 26 13:37:07 i dont want the one-up model they sell here that has the shitty proxy web browser, doubt ill buy another smartphone as there arent any on the market <.< Mar 26 13:37:16 nah, China Post makes it cheap Mar 26 13:37:29 I'm still hopeful for a Jolla device so that I can really, really retire my N900 Mar 26 13:37:48 give it to me :P Mar 26 13:38:01 im fairly doubtful of jolla ever generating a replacement Mar 26 13:38:04 the n900 i mean Mar 26 13:38:10 Or maybe not retire it at all :P Mar 26 13:39:13 buy an n9 Mar 26 13:39:15 its nice Mar 26 13:39:23 linux powered Mar 26 13:39:33 Probably I should learn more about electronics and find a way to manufacture a more modern board into this n900's chassis :P Mar 26 13:40:17 it'd take years to fix N9's OS though, and aegis is in the way blocking things? Mar 26 13:42:21 FlameReaper-N900, components for that are $$$. Mar 26 13:42:39 although you could probably just buy a pandaboard and 3G air card though Mar 26 13:42:55 Yeah... Mar 26 14:58:08 or get a TI blaze ;-P Mar 26 14:58:33 whatever you do, for sure there won't run maemo on it ;-) Mar 26 15:00:15 so if you just want a device with some bitter flavour of *nix on it, get an arbitrary android or fruitphone Mar 26 15:01:48 if you want the most secure smartphone available to mere mortals (US president is using a special dualboot variant of winphone iirc), then get a N9 Mar 26 15:04:38 sailfish kinda looks like re-inventing the wheel but without spokes - I don't see the apps nor the userbase for it yet Mar 26 15:04:53 im more worried about hardware :) Mar 26 15:05:52 it probably won't be worse than N9 Mar 26 15:08:26 and the GUI is HARM v3.0 - extremely touch oriented, with no indications in sight for any decent hw-kbd support or stylus-level accuracy of pointer device Mar 26 15:10:04 guess i'll have to keep on waiting. Mar 26 15:12:50 I suspect many users in here will be waiting forever Mar 26 15:16:45 prolly, back to dumbphones :) Mar 26 15:17:39 that's just being whiny and opting for something completely closed because you can't get your perfect open device Mar 26 15:32:03 well, a sane concept: the best or the cheapest. Always worked for me Mar 26 15:32:57 for now I feel like I found the best with N900 Mar 26 15:39:51 paying more for a lot less would make me feel like I got fooled Mar 26 15:42:14 DocScrutinizer05: doesn't the US president use a Blackberry 9900 or am I missing the joke? :D Mar 26 15:44:06 the joke about him being black? Mar 26 15:44:37 o_O Mar 26 15:44:53 i was referring to >> if you want the most secure smartphone available to mere mortals (US president is using a special dualboot variant of winphone iirc), then get a N9<< Mar 26 15:45:01 oops Mar 26 15:45:36 the presidents get all the special stuff :( Mar 26 15:45:59 the most secure smartphone available to mere mortals is a n900 with full disk encryption! Mar 26 15:47:45 arguably yes Mar 26 15:48:02 and encrypted SIP for calls i guess Mar 26 15:48:16 just iptables the shit out of everything except for an encrypted tunnel Mar 26 15:48:27 but only when you take own care about both the disk-enc *and* security audit of all apps you install Mar 26 15:49:55 I'd still love to see a ZRTP implementation on maemo5 Mar 26 15:51:04 though with a VPN to an asterisk server under your control you might get away with "normal" SIP/RTP Mar 26 15:51:12 The most secure smartphone is the one you do not lose. Mar 26 15:52:24 http://bit.ly/TN9Hiy <- that is the most secure smartphone Mar 26 16:04:21 Estel_, kerio, I edited bme wiki page, look at it and comment what is/is_not OK Mar 26 16:04:39 Pali: wrong channel :) Mar 26 16:07:36 ~seen M4artink Mar 26 16:07:47 Woody14619: i haven't seen 'm4artink' Mar 26 16:07:53 grr Mar 26 16:07:59 ~seem m4rtink Mar 26 16:08:08 ~seen m4rtink Mar 26 16:08:08 ~seen M4rtinK Mar 26 16:08:09 m4rtink is currently on #maemo (1h 46m 27s) #harmattan (1h 46m 27s) #meego (1h 46m 27s) #maemo-ssu (1h 46m 27s), last said: 'also, all 0.04 - 0.40 changelogs are here: http://modrana.org/trac/wiki/ChangelogEN :)'. Mar 26 16:08:10 m4rtink is currently on #maemo (1h 46m 28s) #harmattan (1h 46m 28s) #meego (1h 46m 28s) #maemo-ssu (1h 46m 28s), last said: 'also, all 0.04 - 0.40 changelogs are here: http://modrana.org/trac/wiki/ChangelogEN :)'. Mar 26 16:08:22 yo, vi__ ^^ Mar 26 16:08:43 .oO(If I could type today, I'd be less dangerious.) Mar 26 16:13:40 Pali: i don't think that shutting down immediately below a voltage would work properly Mar 26 16:14:12 kerio, this is estel_ request when he want to disable EDV1 shutdown Mar 26 16:14:27 yeah, but *one* measurement below a threshold can happen for various reasons Mar 26 16:15:33 kerio, if voltage will be (by default) below EDV1 voltage (or will be set to 0) this option will never cause shutdown Mar 26 16:15:36 Pali: the EDV1 flag is set when the voltage is below edv1 for some seconds Mar 26 16:15:38 device will be olready off Mar 26 16:15:55 like, maybe make it 10 or 15 seconds below the threshold? Mar 26 16:16:01 kerio, I can set default value to 0mV Mar 26 16:16:15 no, i mean for people who *want* to use that option :) Mar 26 16:16:32 and soembody can set that option to what thet want Mar 26 16:16:50 yeah but a sudden load can cause a voltage spike drop Mar 26 16:17:50 ShadowJK and DocScrutinizer05: help? how do we make sure that the voltage is actually below a certain threshold and it's not just a spike? Mar 26 16:18:09 kerio now (in my git tree) hald-addon-bme waiting 10s after sending dbus empty signal and before sending dsme empty signal (which causing turn off) Mar 26 16:18:29 Pali: while checking at every poll if it goes back up the threshold? Mar 26 16:18:34 10s is ok i guess Mar 26 16:18:53 maybe make it 15, for consistency with the edv1 flag? idk Mar 26 16:19:20 Pali: why does hald-addon-bme have to specify a value for design capacity? Mar 26 16:19:41 and why do you keep talking about the design capacity like it's something that matters, as opposed to things that actually matter like last full charge? Mar 26 16:20:05 kerio, now when hald-addon-bme decide to power of device it do: send dbus signal, update hal, add g_timeout which will call dsme shutdown funcion after 10s Mar 26 16:20:12 using I2C_SLAVE_FORCE can make the magic smoke come out Mar 26 16:20:26 Pali: yeah but does it cancel the shutdown if the situation changes? Mar 26 16:20:32 like, voltage goes up or charger is connected Mar 26 16:20:45 kerio, because 1) bme hal api needs it 2) because applet using that value Mar 26 16:21:02 kerio, or propose what applet should show for 2) Mar 26 16:21:08 the applet doesn't need that value, all the applet needs is the percentage from bq27200 Mar 26 16:21:29 also, let's move this to #maemo-ssu so E_ can read it too Mar 26 16:21:47 kerio, for design capacity now I mean maximal (is is written on wiki) this value is shown in applet in form A/B (A current, B maximal) Mar 26 16:22:26 kerio, I2C_SLAVE_FORCE: I think that no, but ask that developer on LKML which suggest to use that flag Mar 26 16:23:17 kerio, hald-addon-bme does not cancel shutdown situation because it already sent dbus shutdown signal and some apps using them to "clean" shutdown Mar 26 16:23:29 so it doesn't wait 10 seconds Mar 26 16:23:51 i asked you to require like 10 measurements below the threshold to consider the measurement valid Mar 26 16:24:19 going to #maemo-ssu Mar 26 16:27:57 kerio: sample voltage for 60s, every second Mar 26 16:28:16 60 seconds? we might not have that much time when it's about to shut down Mar 26 16:32:07 Woody14619: ? :) Mar 26 16:32:21 kerio: for sure we have Mar 26 16:41:46 look here: http://linux-kernel.2935.n7.nabble.com/Driver-lis3lv02d-i2c-not-working-on-Nokia-RX-51-td602214.html Mar 26 16:41:55 and on last post :-) ^^ Mar 26 16:42:20 Pali: doc says we need 60s of measurement to know for sure if we're below the threshold Mar 26 16:42:36 ok Mar 26 16:43:39 Pali: linus ^_^ Mar 26 16:44:16 M4rtinK: vi__ was looking for you earlier. :) Mar 26 16:44:29 But missed the a=>4 conversion. ;) Mar 26 16:45:15 oh, that was probably because of the UK Ordnance Survey layer Mar 26 16:45:25 I'Ve already added it in the meantime :) Mar 26 16:45:43 if the threshold in question is 3248mV, I'd just edv1 flag, bq does magic to differentiate between brief excursion below it at high load, and slowly approaching it Mar 26 16:46:24 Likely. Btw: astounding job on Mondrana. I've not used it as much recently since I got a nav with the new head unit in the car, but I have been following the TMO thread. Mar 26 16:46:49 hm. no gpg in symbian :-/ does maemo have it in its stock client? Mar 26 16:47:24 Woody14619: thanks ! :) Mar 26 16:47:51 Macer: client of what? Mar 26 16:48:21 I finally got back to working on it after getting the BB10 & Android Python & PySide ports to an usable state Mar 26 16:48:46 Macer, gpg binary is included in stock maemo5 (because is needed for apt-get) Mar 26 16:50:22 BTW, you can probably soon expect modRana packages fro Android and later BB10 (as that requires making the codebase Python 3.2 compatible - WHILE keeping it also 2.5 compatible) :) Mar 26 16:50:34 *for Mar 26 16:51:35 having at least optified Python 2.6 in Extras would help quite a lot in this regard :) Mar 26 16:51:55 even the Neo FreeRunner has Python 26 or even 2.7 :) Mar 26 16:52:59 M4rtinK: having python 2.7 as the official python would be good Mar 26 16:53:21 at this point i think we can ignore the nokia politics with qt and pyside Mar 26 16:53:28 kerio: even 2.6 would do Mar 26 16:53:37 yeah but let's go for 2.7 :) Mar 26 16:53:42 kerio: the main concern is exception handling Mar 26 16:53:49 i like set comprehensions :) Mar 26 16:54:19 in 2.6 you can do "except Exception as e" Mar 26 16:54:34 which works even in Python 3 Mar 26 16:54:59 up to Python 2.5 only "except Exception, e" is available Mar 26 16:55:38 but sure, there are also other goodies :) Mar 26 17:05:47 "even the Neo FreeRunner has [newer Python]" Ouch. :P :) Mar 26 17:20:32 Pali: I think it is VAUX1.OUT on GAIA Mar 26 17:22:00 Pali, accel is on the first page of schematic, on the right sidee of omap Mar 26 17:22:05 *side Mar 26 17:22:35 aiui it is using 2 power supplies, V28 and VIO 18 Mar 26 17:27:36 Pali: yep, seems like VDD is VAUX1 LDO Mar 26 17:29:12 Pali: the other supply (VIO 18) is TWL VIO dc-dc Regulator Mar 26 17:32:00 Don't you hate it when you sell all your motorbikes, and your crash the last one? Mar 26 17:36:31 Pali: here? Mar 26 18:02:02 <_PanzerSajt> I have installed N12 with vakkov's modifications but it can't access "sdcard" I was playing around with vold.conf but I still can't get it working. I know the default would be MyDocs, but I have problems with that partitiun under maemo as well. (can't save pictures to it) and the currenty android is installed on a micro-sd card Mar 26 18:12:20 what's n12? Mar 26 18:14:16 nitdroid Mar 26 18:21:51 _PanzerSajt, perhaps something for the #nitdroid channel? Mar 26 18:22:27 <_PanzerSajt> yes I have posted to it also but they are not the fastest Mar 26 18:24:14 Ah :) Mar 26 18:25:08 one advice, _PanzerSajt: never set your speakers volume too loud. It will blow them Mar 26 18:25:16 on nitdroid* Mar 26 18:29:06 <_PanzerSajt> ok thanks RiD Mar 26 18:29:32 it all sounds better for a while.... then *pof* Mar 26 18:32:55 freemangordon: what's the problem with lis302? Mar 26 18:44:28 LOL, read the thread, <3 Linus Mar 26 18:49:49 hmmm.. I find threads like this http://talk.maemo.org/showthread.php?t=89616 rather insulting to community efforts. Mar 26 18:51:39 sixwheeledbeast: idiots/jokers are everywhere sadly... Mar 26 18:54:53 thedead1440: but it's a shame they don't see how insulting it is to the devs, I cba explaining about CSSU and thumb and everything else positive on tmo, again... Mar 26 18:55:18 indeed; such people are not worth wasting time on Mar 26 18:58:29 sucker Mar 26 18:59:06 I'd love to explain him how to install winXP using nandtest Mar 26 19:00:11 "how to install windows XP on it" --- WAAAAAH!!! >:-( Mar 26 19:00:25 or android 4.2 :D Mar 26 19:02:58 should we start a bet? about this luser, most likely OCed to 1100 and still it's eating too much battery and speedpatch doesn't run fast enough Mar 26 19:03:16 (SIC!) Mar 26 19:04:13 i'm ready to bet he used both *patches :D Mar 26 19:04:38 mmm what happened with the whole *patch thing? Mar 26 19:04:55 wasn't MT supposed to start a thread for it Mar 26 19:05:10 :shrugs: Mar 26 19:05:11 seems like he forgot... maybe sixwheeledbeast should send council another email reminding them :D Mar 26 19:06:08 IIRC it was mentioned again at last meeting, but hopefully someone will remember... Mar 26 19:13:24 DocScrutinizer51: yeah, Linus and accel :D Mar 26 19:14:14 DocScrutinizer05: Pali was unable to find which power regulator powers the accelerometer Mar 26 19:14:45 for all I know (too lazy to check it) lis302 has no VDD that's faintly optional or able to suspend Mar 26 19:15:25 it's probably on general 3V3 and 1V8_IO Mar 26 19:16:21 similar to Linus' SSD Mar 26 19:16:27 ;-D Mar 26 19:18:18 DocScrutinizer51: no, if I read the schematics correctly, VDD is attached to V28 (VAUX1 LDO) Mar 26 19:18:36 hmm Mar 26 19:19:04 you're urging me to do what I tried to avoid Mar 26 19:19:12 DocScrutinizer05: would you take a look, I may be wrong :P Mar 26 19:19:19 :-S Mar 26 19:24:25 yes, correct: VAUX1.OUT, and VIO_18 Mar 26 19:27:18 V28, just like main camera cover switch, uSDcard hotswap detection sensor, slide open sensor, FM transmitter Mar 26 19:27:32 and some stuff on flex board Mar 26 19:27:51 probably prox det, ALS, a.s.o. Mar 26 19:28:15 *very* unlikely that system would decide to power down that LDD Mar 26 19:29:50 sure you could push whatever cruft into kernel, to handle every microscopic aspect of a platform on the logical descriptive level Mar 26 19:30:50 for me as an EE it's clear to see that this is a power rail that must not get shut down ever Mar 26 19:32:43 while I boggle to have a closer look at the driver itself, last time I looked at upstream lis302 kernel driver, it took me 3h to stop puking Mar 26 19:34:37 DocScrutinizer05: but shutting down things is good! Mar 26 19:34:41 DocScrutinizer05, this looks like lis driver support more accelerometers, I see that on my notebook is driver lis3lv02d loaded and in use Mar 26 19:34:50 wtf ... my n900 doesnt connect to wlan anymore after it died on battery yesterday Mar 26 19:35:08 iirc they configued the lis302 to do paced reporting at 100Hz, and they spit that shit out via /dev/accel at same rate. This been only way to use that chip with this driver. IRQs not supported Mar 26 19:36:48 and then you wonder why upstream kernel is poor at power-management Mar 26 19:37:09 while devels bitch about regulators - LOL Mar 26 19:37:59 iirc the lis302 - in power-down mode (a internal command/mode of lis302) - eats ~0.1µW Mar 26 19:38:48 is that a lot? is it very little? Mar 26 19:38:59 verrry little Mar 26 19:39:57 a 1/200000 of a dim led Mar 26 19:40:55 or, less than your wrist watch Mar 26 19:41:16 daaamn, that's *very* little Mar 26 19:41:18 i don't have a wrist watch Mar 26 19:41:33 yes, it needs less than that Mar 26 19:41:47 so powering it down /would/ be good Mar 26 19:42:13 yes, by sending i2c command "power-down" to lis302 Mar 26 19:42:51 though even when on, it needs not really any noticeable energy unless it triggers an IRQ or transmits data Mar 26 19:46:39 seriously, the upstream lis302 driver was cheap uninspired crap written for a hands-free joystick type of HID Mar 26 19:47:08 the one I seen last time I checked. Mar 26 19:48:31 and it's been 100% incompatible with maemo's lis302 kernel driver (C) Nokia, and 10000% inferior Mar 26 19:50:26 while Nokia's driver only lacks support for differential mode aka highpass filter, and general kevents support, the upstream version lacks basically everything except what you'd need for a 3D-mouse or whatever, a nintendo-controller or younameit Mar 26 19:52:55 and qtm resp sensorfw are based on such crap and introduce even one more level of timer paced polling on top Mar 26 19:56:17 result: while with a slightly augmented/improved nokia lis302 driver you could build an app that makes N900 yell an alarm when you knock the table the N900 sits on, and that app could run for 10 days from one battery charge, the upstream driver with qtm allows you to build a similar app where you'd need to turn N900 face down for 3s before alarm starts, and that app would cut through your battery in one day or less Mar 26 20:04:09 DocScrutinizer05: lets upstream our driver then Mar 26 20:04:24 Pali: I hope you have your answer (which VDD) now Mar 26 20:04:52 yes, now needs to understand linux regulator api Mar 26 20:05:01 and add needed vdd Mar 26 20:06:40 Pali: you may send an email in the meanwhile, otherwise kernel maintainers war is imminent :D Mar 26 20:08:27 DocScrutinizer05: are you sure there are other consumers on the same rail? as there are V28_A, etc. Other 2.8V exist too Mar 26 20:09:29 DocScrutinizer05: asking just out of curiocity, just answer yes/no Mar 26 20:09:53 freemangordon: I listed them above Mar 26 20:10:04 DocScrutinizer05: ok Mar 26 20:10:25 aah, yes I see Mar 26 20:10:49 chances are zero this power supply to be ever switched off Mar 26 20:10:57 exactly Mar 26 20:11:04 just like IO_18 Mar 26 20:11:43 Pali: you should already have that VDD. Is there rx51_board.c code on github? Mar 26 20:11:50 so adding a regulator to lis302 is mere cruft and bloat Mar 26 20:12:07 DocScrutinizer05: but will make kernel maintainers happy :D Mar 26 20:12:49 freemangordon: https://gitorious.org/linux-n900/linux-n900 Mar 26 20:12:58 as already mentioned, I don't give much about that particular kernel driver upstream anyway Mar 26 20:13:03 Pali: ok, will look at it Mar 26 20:13:25 nobody sane will *ever* use it on N900 Mar 26 20:13:40 DocScrutinizer05: don;t bet your ass on that ;) Mar 26 20:14:02 I bet my ass on it, you always lose on definition of "sane" Mar 26 20:14:10 :) Mar 26 20:15:41 freemangordon: btw all thumbs up for upstreaming "our" lis302 driver, alas we need to either fight a patchwar or need to rename it to sth like lis302-irq-embedded Mar 26 20:16:37 I asked several times if those guys are serious about the crappy lis302 driver upsteam. Their answer: send patches Mar 26 20:16:49 see Mar 26 20:17:33 of course we could just patch out line 1-99999 of original, every file, and insert all lines of our driver ;-P Mar 26 20:18:12 Pali: I guess it is thet https://gitorious.org/linux-n900/linux-n900/blobs/v3.8-rc3-n900/arch/arm/mach-omap2/board-rx51-peripherals.c#line736 one Mar 26 20:18:29 DocScrutinizer05: you said FMTX is on the same lane? Mar 26 20:18:45 Pali: here you are https://gitorious.org/linux-n900/linux-n900/blobs/v3.8-rc3-n900/arch/arm/mach-omap2/board-rx51-peripherals.c#line739 Mar 26 20:18:46 err TX or RX, can't recall Mar 26 20:18:56 whatever I wrote been correct Mar 26 20:19:00 we already have it Mar 26 20:19:48 _supply rx51_vaux1 Mar 26 20:19:51 yep Mar 26 20:19:58 Pali: iiuc you should add lis302 to rx51_vaux1_consumers Mar 26 20:20:06 I see Mar 26 20:20:09 also to rx51_vio_supplies? Mar 26 20:20:30 Pali: I guess so Mar 26 20:21:01 there should be some VIO18 or something Mar 26 20:21:38 static struct regulator_init_data rx51_vaux1 = { Mar 26 20:21:42 is the LDD Mar 26 20:21:58 .always_on = true, /* due battery cover sensor */ Mar 26 20:22:01 X-P Mar 26 20:23:47 static struct regulator_init_data rx51_vio = { Mar 26 20:23:53 894 Mar 26 20:25:35 shit, closed the window by mistake :( Mar 26 20:25:42 ah, battery empty... what is i2c addr of lis device? Mar 26 20:26:07 mmmmfff Mar 26 20:26:37 Pali: I think you should add REGULATOR_SUPPLY("vio", "i2caddress") to rx51_vio_supplies and REGULATOR_SUPPLY("vdd", "i2caddress") to rx51_vaux1_consumers Mar 26 20:26:47 though I guess you already grok it Mar 26 20:26:55 freemangordon, not Mar 26 20:27:03 static const char reg_vdd[] = "Vdd"; Mar 26 20:27:03 static const char reg_vdd_io[] = "Vdd_IO"; Mar 26 20:27:09 http://wiki.maemo.org/N900_Hardware_Bus_I2c Mar 26 20:27:10 this is in lis driver ^^^ Mar 26 20:27:38 Pali: you are missing the board code Mar 26 20:27:45 change lis driver to: Mar 26 20:28:05 bus3 0x1D Mar 26 20:28:22 static const char reg_vdd[] = "vdd"; Mar 26 20:28:22 static const char reg_vdd_io[] = "vio"; Mar 26 20:28:40 freemangordon, not possible Mar 26 20:28:46 Pali: well, ok, you can leave as it is I guess Mar 26 20:28:48 it is in lis driver Mar 26 20:29:04 rather I will add names from lis driver to board code Mar 26 20:29:09 sure Mar 26 20:29:29 REGULATOR_SUPPLY("Vdd_IO", "i2caddress") to rx51_vio_supplies and Mar 26 20:29:29 REGULATOR_SUPPLY("Vdd", "i2caddress") to rx51_vaux1_consumers Mar 26 20:29:35 agree? Mar 26 20:29:43 ok Mar 26 20:29:46 :) Mar 26 20:30:23 Pali: and... ask next time :P Mar 26 20:30:36 :D Mar 26 20:30:51 now we need to test if that patch working Mar 26 20:31:34 well, I can't do it here, will leave it to you ;) Mar 26 20:31:59 I cannot do it too, battery is empty (now started charging)... Mar 26 20:35:02 NIN101, where I can send patches for rescueOS? Mar 26 20:40:29 https://github.com/NIN101/N900_RescueOS if you can bothered, or nin101 at lavabit dot com Mar 26 20:57:33 freemangordon, we are slow Mar 26 20:57:50 now same patch appeared on mailinglist Mar 26 20:59:35 freemangordon: before you upstream "our" driver, please fix negligences like http://mxr.maemo.org/fremantle/source/kernel/drivers/i2c/chips/lis302dl.c#164 Mar 26 21:00:43 http://linux-kernel.2935.n7.nabble.com/Driver-lis3lv02d-i2c-not-working-on-Nokia-RX-51-td602214.html#a623699 Mar 26 21:02:48 freemangordon: oooh, forget my rant about kevents missing: http://mxr.maemo.org/fremantle/source/kernel/drivers/i2c/chips/lis302dl.c#223 Mar 26 21:03:52 just musing why it's missing there: http://mxr.maemo.org/fremantle/source/kernel/drivers/i2c/chips/lis302dl.c#237 Mar 26 21:09:05 freemangordon: anyway, please ping me about missing sysnodes before upstreaming the lis302dl maemo driver Mar 26 21:09:29 we need ~2 more than currently are there Mar 26 21:09:36 one of them is filter Mar 26 21:09:51 and another one is IRQ config Mar 26 21:21:00 Pali: ooh, *WE* are slow. Ok, if you say so :P Mar 26 21:21:25