**** BEGIN LOGGING AT Tue Aug 28 03:00:03 2012 Aug 28 06:19:14 libpcap? Aug 28 06:22:23 IroN900:~# apt-cache policy libpcap0.8 Aug 28 06:22:25 libpcap0.8: Aug 28 06:22:26 Installed: (none) Aug 28 06:22:28 Candidate: 1.1.1-2maemo4 Aug 28 06:22:29 Version table: Aug 28 06:22:31 1.1.1-2maemo4 0 Aug 28 06:22:32 500 http://repository.maemo.org fremantle-1.3/free Packages Aug 28 06:22:45 NB: Installed: (none) Aug 28 06:23:15 tshark works nevertheless, afaik it has libpcap "built in" Aug 28 06:35:02 DocScrutinizer51: BTW copyright in the header applies to the source code, not to the underlying technology AFAIK, so it seemes javispedro's concern was about IDA Aug 28 06:35:49 nonono Aug 28 06:36:24 I'm pretty sure his whole concern been about GPLing some decompiler output Aug 28 06:36:50 DocScrutinizer06: http://gitorious.org/kaiser-linux/kernel/blobs/3834c3f227725e2395840aed82342bda4ee9d379/fs/ntfs/mft.c Aug 28 06:36:56 no matter if IDA or $your_fav_disassembler Aug 28 06:37:11 now, tell me how is that GPL-ed Aug 28 06:37:19 NTFS ;) Aug 28 06:38:30 every single file in NTFS driver is under GPL and is parto of the linux kernel Aug 28 06:38:39 well, looks to me like somebody has *written* that, and thus actually created a piece of intelectual property of his own Aug 28 06:39:06 DocScrutinizer06: you mean I didn't write what I published Aug 28 06:39:17 I did it, it is not a simple copy/paste Aug 28 06:39:26 it looks like IDA "written" it Aug 28 06:39:29 no Aug 28 06:39:51 lemme show you what ida produces as output to end that, just a minute Aug 28 06:40:44 well, then I don't know (and I believe in your words, no need to show me IDA output) Aug 28 06:41:13 ask javispedro if he as well had that misconception Aug 28 06:41:22 or he actually meant something else Aug 28 06:42:35 freemangordon: just use ida and throw the code out there! wooo!! Aug 28 06:42:39 hack the planet! Aug 28 06:43:20 DocScrutinizer06: it's not against licensing to decompile and rewrite code? ;) Aug 28 06:43:55 i always figured that was almost always part of a license with most closed software Aug 28 06:44:16 Macer: I missed your point(still having my coffee), but IDA/hexrays output cannot be compiled back, no matter what Aug 28 06:44:51 freemangordon: oh im just saying.. if you do decompile... and then write something from teh output... Aug 28 06:45:02 yes? Aug 28 06:45:11 i figure the decompiling itself is restricted by the software's license Aug 28 06:45:49 or is this the method typically used to "write" gpl'd src for hardware? Aug 28 06:45:58 well, take a look at NTFS example I gave above Aug 28 06:46:11 i did.. that is why i am asking Aug 28 06:47:26 the real question is how did they write it to begin with? Aug 28 06:47:38 did they take ntfs and decompile it, learn how it worked, then recompile? Aug 28 06:47:43 well.. write then recompile Aug 28 06:47:46 by doing RE Aug 28 06:47:55 if so how is that not a violation of the eula from ms? :) Aug 28 06:48:02 it is Aug 28 06:48:04 :D Aug 28 06:48:11 Macer: that's gray zone Aug 28 06:48:30 DocScrutinizer06: i doubt it is very gray. i'm sure MS has it vividly written in the eula Aug 28 06:49:00 most licences/eulas have a § that forbids decompiling but also in several jurisdictions that's just moot Aug 28 06:49:38 DocScrutinizer06: I am almost sure javispedro's concern is about me using IDA, which I don't have the right to :) Aug 28 06:49:41 DocScrutinizer06: meaning the eula doesn't hold any force of law? even tho the end user HAS to agre in order to use it? Aug 28 06:50:03 what's definitely forbidden, even under normal copyright, is decomoiling and then publishing Aug 28 06:50:31 yeah but what justifies a rewrite? Aug 28 06:50:43 what if there is only 1 way to do it? ;) Aug 28 06:51:10 i would assume there is probably no way to avoid using none of the decompiled code in order to get it to work Aug 28 06:51:12 meh, please consult wikipedia about it. IANAL Aug 28 06:51:14 DocScrutinizer06: but that us actually a rewrite, it is not the output of decompiler/disassembler Aug 28 06:51:43 no matter what tools i used during the rewrite Aug 28 06:51:55 i mean quite honestly... if i were a dev i'd probably just have my own special n900 Aug 28 06:52:08 running some other awesome os with stuff i decompiled and got working :) Aug 28 06:52:18 then throw a torrent on tpb .. but that's just me haha Aug 28 06:53:25 well, maybe it is better to wait for javispedro to appear and throw some lit on WTF "WTF" means Aug 28 06:53:35 s/lit/light/ Aug 28 06:53:35 freemangordon meant: well, maybe it is better to wait for javispedro to appear and throw some light on WTF "WTF" means Aug 28 07:36:52 Macer: _Legally_ (IANAL either) EULAs can't remove end user *rights* including the right to reverse engineer for interoperability etc. They can only provide *additional* rights to those normally reserved by copyright law. Aug 28 07:38:31 but plain (C) law is pretty simple, in that you mustn't redistribute intelectual property of others under your own name and licence. Aug 28 07:39:18 Except under "fair use" or with permission. Aug 28 07:42:28 doesn't matter if the original binary is represented as a binary file, a hex dump, or a disassembler output, if you copy complete or partial. You simply must not use the original code in your work if you want to distribute your work. You *may* use disassembler to understand what original code does, then write something that does about the same thing, since the algorithm isn't protected by law. Like with a book, you may write your won Aug 28 07:42:29 story that tells exacxtly same plot, but you mustn't copy a few lines of original book's text, no matter if you change layout or font or even translate it to another language Aug 28 07:43:29 s/won/own/. Aug 28 07:46:47 Exactly so, but EULAs claim to explicitly ban RE, which in most jurisdictions isn't legally enforcible (US law notwithstanding) Aug 28 07:47:18 I already mentioned that, yes Aug 28 07:47:27 yup Aug 28 07:47:56 I would be wary of doing just about anything if I lived in the US though... Aug 28 07:48:11 look at what happened to Samsung Aug 28 07:48:20 :nod: Aug 28 07:48:48 I've been reading the groklaw blog for years, it just seems to get crazier and crazier. Aug 28 07:48:59 well, google will probably sue apple now Aug 28 07:49:28 since we got "permission" from quim gill to redistribute Nokia(C) blobs, we probably are safe when we keep original Nokia (C) in header of disassembly of those blobs Aug 28 07:50:33 heh Aug 28 07:51:06 DocScrutinizer06: but that is not disassembly Aug 28 07:51:16 doesn't matter Aug 28 07:51:24 we just pretend it is Aug 28 07:51:32 US Pat. 5710987, Receiver having concealed external antenna Aug 28 07:51:51 Assignee: Motorola, Inc. Aug 28 07:51:57 zeq1: could you help on that? Aug 28 07:52:06 US Pat. $random, A multi-pronged utensil to eat food with. Aug 28 07:52:14 freemangordon: what is it exactly? Aug 28 07:52:25 the patent doesn't talk about a special technique to do so Aug 28 07:52:39 just... you know... put the antenna inside the body of the phone Aug 28 07:52:52 00:22 freemangordon: wtf are you doing Aug 28 07:52:54 freemangordon: http://gitorious.org/community-ssu/hildon-im-vkbrenderer3/blobs/master/src/hildon-im-vkbrenderer.c#line1854 Aug 28 07:52:56 freemangordon: you cannot just paste a IDA decompiler output and relicense it under GPL3 Aug 28 07:52:59 zeq1: ^^^ Aug 28 07:53:34 I still fail to grok the problem :( Aug 28 07:53:55 the problem is you put somebody else's code/property under a new licence Aug 28 07:54:03 you want my opinion on whether using the information gleaned from IDA is okay? Aug 28 07:54:19 DocScrutinizer06: it is not somebody else code Aug 28 07:54:24 zeq1: basically yes Aug 28 07:54:27 *sigh* Aug 28 07:55:09 it's the code to render vkb? Aug 28 07:55:10 what javispedro wrote actually *IS* assuming it's just that, somebody else's code Aug 28 07:55:21 zeq1: yes Aug 28 07:55:31 freemangordon: please dissect problem correctly Aug 28 07:56:23 >>javispedro: "you cannot just paste a IDA decompiler output and relicense it under GPL"<< this is absolutely correct Aug 28 07:56:24 it's slightly grey if you manually parsed the output from the decompiler and re-wrote it Aug 28 07:56:42 adn that is what I did Aug 28 07:56:45 if you wrote specs from the de-compiler then wrote new code from spec you're definitely ok Aug 28 07:57:15 if you can convince yourself that's what you did in your head... ;) Aug 28 07:57:56 my staement: "under qgil licence, you may paste a IDA decompiler output, givven you do NOT relicence it" Aug 28 07:58:57 to comply with the latter, it's _not_ mandatory to copy _complete_ orifinal code's decompiler output Aug 28 07:59:11 official* Aug 28 07:59:16 DocScrutinizer06: but it is not IDA decompiler output (well, some parts need cleanup, I agree) Aug 28 07:59:19 err original* Aug 28 07:59:28 but it is a matter of cleanup Aug 28 07:59:51 You could argue the IDA output is a machine generated specification Aug 28 08:00:02 whatever, either you grok the issue now or you need to find help from somebody else, I'm definitely short on time Aug 28 08:00:04 *maybe* Aug 28 08:00:43 nowhere in the IDA output there are structure/class definitions, function names, etc Aug 28 08:01:41 freemangordon: I would lean to your side, it's just that these things can be iffy Aug 28 08:01:47 DocScrutinizer06: np, thanks anyway, will wait for javispedro to appear and elaborate Aug 28 08:02:17 It's mostly comes down to whether you feel you can document your process Aug 28 08:02:51 zeq1: document? man, why using taboo words, I am developer after all :D Aug 28 08:04:04 well, I can explain, the you give me your toughts, ok? Aug 28 08:04:15 if you have tiem Aug 28 08:04:18 time* Aug 28 08:04:58 it's not likely there would be any issue. but if you got an email from Nokia you might want to be able to demonstrate you wrote your code from the functional specification rather than just manually parsing their "IP" Aug 28 08:06:03 IMHO, I really can't see it being a problem, we're probably "violating" patents with every line of code we write anyway... :P Aug 28 08:06:34 but as I said IANAL Aug 28 08:06:42 hmm, I got the point. The problem is that for the most SW i wrote (esp the FOSS one) there is no specification elsewhere but in my head Aug 28 08:07:20 not that I say I know how vkb works :) Aug 28 08:08:07 It might be worth you cleaning up your RE work into a specification document if you have any concerns. Aug 28 08:08:23 I know how that must make you feel though ;) Aug 28 08:09:04 or you consider my suggestion Aug 28 08:09:29 there's a reason commercial companies insist on lots of documentation Aug 28 08:09:42 and state "all disassembled parts in this document (C)Nokia, under licence applicable to maemo" Aug 28 08:10:04 definitely Aug 28 08:10:18 well, that is over any limits :D. So I will wait for javispedro to elaborate (as I still have the feeling he told me I am not allowed to use IDA) and after that will ask merlin1991 Aug 28 08:11:15 this is not over any limits, this is in line with qgil permission to redistribute (C)Nokia blobs in CSSU Aug 28 08:11:20 if it requires several months that source to be in a condition that will make everyone happy, i'll consider asking merlin to just remove it Aug 28 08:11:37 DocScrutinizer06: it was re documentation :) Aug 28 08:12:14 and probably the _only_ way to get this into official CSSU repos Aug 28 08:13:35 ok, I'll ask merlin1991 to put whatever license he finds appropriate, I told you, I don't care if my name is mentioned in the header Aug 28 08:15:20 though to be honest I don' think Nokia can care less :) Aug 28 08:17:26 freemangordon: I'm sure Nokia would rather we all just disappeared Aug 28 08:21:55 freemangordon: in http://www.snewbury.org.uk/maemo/sigmask there is the glibc source package Aug 28 08:23:06 zeq1: thanks Aug 28 08:23:41 it enables pselect by just bumping the min kernel version, and also the patch to allow packages using inlines to compile against it using current gcc versions Aug 28 08:24:08 zeq1: do you plan to compile and package newer gcc 4.7.2? Aug 28 08:24:11 I ported it from the glibc2.6 git branch Aug 28 08:24:23 yes, I will do it today if I get time Aug 28 08:25:05 BTW I think we should first asm merlin1991 to create a repo on gitorious, containing the original glibc code and apply those patches on top of it Aug 28 08:25:12 *ask Aug 28 08:26:15 it needs to depend on kernel "sigmask syscalls" feature, I didn't do that Aug 28 08:27:03 (and also remember it has to be built with the stock compiler) Aug 28 08:27:35 I never managed to find a way of not breaking the locales using 4.7 Aug 28 08:27:46 (or 4.6 for that matter) Aug 28 08:28:02 unfortunate. Aug 28 08:28:31 the libs are half the size compiled with thumb :( Aug 28 08:29:12 (no performance difference on benchmarks though) Aug 28 08:30:27 freemangordon: I have a patch for hildon-desktop to build using -ffast-math too, shall I put that on the server too? Aug 28 08:31:44 As I mentioned before my holiday it's necessary to change the finite math functions Aug 28 08:33:33 freemangordon: you already merged the float branch into your cssu-thumb version didn't you? Aug 28 09:14:18 zeq1: no, I didn't Aug 28 09:15:28 zeq1: andwe'd better use gitorious for patches, it does not make sense me to apply the patches you did :) Aug 28 09:15:48 i.e. cone, edit, request a merge is the correct way AIUI Aug 28 09:15:55 s/cone/clone/ Aug 28 09:15:56 freemangordon meant: i.e. clone, edit, request a merge is the correct way AIUI Aug 28 09:16:34 that way the others have the chance to review the patches too Aug 28 09:25:52 wtf Aug 28 09:26:02 my bupbat worked after more than a minute D: Aug 28 09:26:21 i don't even Aug 28 09:26:36 does this make sense? Aug 28 09:28:35 DocScrutinizer51: my bupbat resisted more than a minute of lack of battery Aug 28 09:36:13 DocScrutinizer51: are you sure that the bupbat doesn't power bq24k? Aug 28 09:36:34 it definetely powered bq27k Aug 28 10:22:02 freemangordon: ok will do Aug 28 10:32:26 kerio: (bupbat) yes, quite sure Aug 28 10:34:09 though I fail to understand how duration of bupbat supply would indicate it also supplying any bq chip Aug 28 10:35:38 anyway bupbat is connected to twl4030 and just supplies the CMOS clock and some registers in there Aug 28 10:36:26 DocScrutinizer51: i wanted to reset bq24k because it's still in a silly state because of Pali's module Aug 28 10:36:43 bq27k didn't lose the calibration - is it fair to think that bq24k wasn't reset either? Aug 28 10:36:54 NB twl4030 also is responsible for powering up device on VBUS detect Aug 28 10:37:29 there's basically nothing to reset in bq24150 Aug 28 10:39:14 i see Aug 28 10:39:22 so it's twl that could be in a weird state Aug 28 10:40:17 but bq24150.ko afaik also is messing with USB for charger detect and whatnot, and that's PHY 1707, and twl4030 Aug 28 10:41:19 we already had reports of allegedly permanently messed up twl4030 settings in early days of h-e-n Aug 28 10:42:01 and due to bupbat it *might* be tricky to reset twl4030 Aug 28 10:42:11 :c Aug 28 10:43:08 I for one never completely investigated that topic, since we made sure we don't do any evil Aug 28 10:43:23 bbl Aug 28 10:43:33 goddammit pali Aug 28 10:48:48 DocScrutinizer51: hm, where could i find a datasheet for twl4030 so i can look for a way to reset it? Aug 28 10:50:14 check wiki, there has to be a link Aug 28 10:50:39 tps65595 iirc Aug 28 10:51:12 tps65950 apparently Aug 28 10:52:25 oh god, 171 pages Aug 28 10:54:08 hm, the only mentions of "reset" talk about a power-on reset Aug 28 10:54:24 whoops Aug 28 10:54:38 so... idk Aug 28 10:59:52 kerio: your lucky day! I'm sitting at work, bored, with internet access on my fingertips Aug 28 11:03:07 yay! :D Aug 28 11:03:33 right now i'm just going to wait like an hour for bupbat to fucking fail Aug 28 11:03:46 i swear, those things do the exact opposite of what you want them to do Aug 28 11:06:17 DocScrutinizer51: assuming it's healthy, how much would i have to wait? Aug 28 11:10:28 read 8.1.2 Aug 28 11:25:25 ok, bupbat resisted 25 minutes Aug 28 11:25:29 :s Aug 28 11:47:31 kerio: what's up now? did it 'reset'? Aug 28 11:47:44 did you even flash sane kernel? Aug 28 11:48:04 of course not, still on kp51r1 :) Aug 28 11:48:25 i'd have to downgrade to kernel-cssu that matches kp50 Aug 28 11:49:40 anyway, the only problem i had was that triggering a watchdog reboot caused the phone to enter act-dead mode with no charger plugged in Aug 28 11:49:48 which was weird Aug 28 11:50:04 but once i properly boot, it works fine Aug 28 11:57:39 yep, i can replicate that Aug 28 11:58:05 watchdog reboot enters act-dead mode for a bit Aug 28 11:58:15 and then it boots Aug 28 11:58:27 oh pali, what did you do to my n900 ç_ç Aug 28 12:13:13 kerio: kernel-cssu3 is kp51r1 Aug 28 12:21:08 btw WTF are the register descriptions for tps65950? Aug 28 12:27:02 if *I* were TI designing twl4030 chip, I'd specify a generic reset procedure: hold poweron button while inserting battery, maybe keep it for a further 10s Aug 28 12:29:02 kerio: did you find further manuals than the one on TI webpage that has electrical specs and that bupbat stuff under 8.x.y as I said above? Aug 28 12:29:23 no, but i haven't actually searched Aug 28 12:29:29 i mean, that's the whole datasheet as provided by TI Aug 28 12:36:54 funny sidenote: for N9 Nokia obviously got a special custom build called TPS65951 Aug 28 12:44:03 hmm, I have an ex-TI 5m behind me Aug 28 12:44:15 what manuals are you lloking for? Aug 28 12:50:18 StyXman: twl Aug 28 12:50:29 aka tps65950 Aug 28 12:50:38 we want to know if there's a way to reset it without removing power to it Aug 28 12:50:46 because that would require removing the bupbat, if it still works (like mine) Aug 28 12:56:36 nm, all on one page, just need to scroll down Aug 28 12:58:47 www.ti.com/product/tps65950 and scroll down Aug 28 13:01:40 DocScrutinizer51: ack Aug 28 13:01:58 http://www.ti.com/lit/an/swca025/swca025.pdf Aug 28 13:03:51 StyXman: syn+ack Aug 28 13:04:12 HAHA!!! Aug 28 13:04:44 Ack, followed by syn+ack... must be some kind of a backwards handshake. Aug 28 13:04:56 Note: One exception: when the user presses the PWRON (power-on) button for 8 seconds and removes the battery in the next 8 seconds, the TPS65950 enters NO SUPPLY state instead of BACKUP state, even if a valid backup battery is present. In such a situation, the backup domain registers are also reset, along with the VRRTC domain registers. Aug 28 13:05:04 :D :D D: Aug 28 13:05:05 er Aug 28 13:05:06 :D :D :D Aug 28 13:05:55 gimme 5 Aug 28 13:06:06 if *I* were TI designing twl4030 chip, I'd specify a generic reset procedure: hold poweron button while inserting battery, maybe keep it for a further 10s Aug 28 13:06:17 the hw engineering is strong in this one Aug 28 13:06:51 hm, how would i use that thing? Aug 28 13:07:07 kerio: rst Aug 28 13:07:17 oh, i'll just shutdown with held PWRON and then remove the battery i guess Aug 28 13:08:31 would that work? Aug 28 13:09:19 ~n900-full-reset is when the user presses the PWRON (power-on) button for 8 seconds and removes the battery in the next 8 seconds, the TPS65950 enters NO SUPPLY state instead of BACKUP state, even if a valid backup battery is present. In such a situation, the backup domain registers are also reset, along with the VRRTC domain registers. Aug 28 13:09:20 DocScrutinzrWrk: okay Aug 28 13:09:44 DocScrutinzrWrk: would USB PHY be reset in the same way? otherwise it's not really a full reset :) Aug 28 13:10:19 PHY should get reset by removing battery (mind you, it's not backed up by bupbat) Aug 28 13:10:24 k Aug 28 13:10:33 removing battery + wait the usual minute or so? Aug 28 13:10:47 you however should remove bat for 120+ seconds Aug 28 13:11:23 hm, tps65951 has a way to reset its own USB PHY Aug 28 13:11:37 does the N9 have a separate chip for usb, like the n900? Aug 28 13:14:54 shorting N900 battery terminals of N900 might shorten capacitor discharge time Aug 28 13:15:08 does it work if i just touch them? :) Aug 28 13:15:23 N9 uses tps65951 PHY Aug 28 13:15:36 kerio: basically yes Aug 28 13:15:51 maybe with slighly humid fingertips Aug 28 13:16:12 Soon you'll have him spitting in his phone. Aug 28 13:16:19 hehe Aug 28 13:16:27 protip: dont pee on it Aug 28 13:16:59 i wasn't going to D: Aug 28 13:17:30 so Aug 28 13:17:35 bq27k calibration data is still there Aug 28 13:17:41 i'll try triggering a reboot again Aug 28 13:17:48 that seemed to showcase the problem before Aug 28 13:18:32 ok, normal booting :D Aug 28 13:18:41 Estel_: you might want to try this too Aug 28 13:18:54 (before it used to enter act-dead mode while charging for a bit) Aug 28 13:21:41 * DocScrutinzrWrk frowns on Note at bottom of §3 in http://www.ti.com/lit/an/swca025/swca025.pdf, p.6 Aug 28 13:22:20 * DocScrutinzrWrk wonders if maemo5 maybe never even activates bupbat charging Aug 28 13:33:13 DocScrutinzrWrk: hahahah oh wow Aug 28 13:33:17 maybe it never did Aug 28 13:33:37 although i thought that failures in bupbats were due to physical problems Aug 28 13:35:21 DocScrutinzrWrk: good finding, maybe we should ask Pali to include that in charger driver Aug 28 13:35:53 most probably that is the reason most of those batteries to be empty in an year or so Aug 28 13:37:02 we should also ask Pali to make the charger driver not completely fuck up twl4030 :) Aug 28 13:43:49 if bupbat never gets charged, it will break and go EOL after ~1 year or so Aug 28 13:43:57 due to deep discharge Aug 28 14:09:23 anyway, the n900 is surprisingly resilient Aug 28 14:09:49 i mean, the guy that nandtested the whole NAND has something that boots now Aug 28 14:11:39 nandtested? Aug 28 14:12:19 nandtest, from mtd-utils Aug 28 14:12:30 a destructive nand test Aug 28 14:12:44 dude overwrote most of CAL with random data Aug 28 14:12:48 well, all of CAL Aug 28 14:15:07 how did he fix it? Aug 28 14:18:54 Gh0sty: he didn't :) Aug 28 14:19:01 he coldflashed and then reflashed Aug 28 14:19:04 but CAL is still borked Aug 28 14:23:40 DocScrutinzrWrk: :nod: Aug 28 14:23:57 the question is if it can be revived Aug 28 14:25:02 kerio: does it boot now? Aug 28 14:28:34 jacekowski: he said it does Aug 28 14:29:02 freemangordon: a deep discharged LiIon is considered dead usually. Even a LiSiIon bupbat Aug 28 14:50:44 kerio: and how do you coldflash? Aug 28 14:51:00 ~coldflash Aug 28 14:51:01 i heard coldflash is http://mg.pov.lt/maemo-irclog/%23maemo.2010-10-31.log.html#t2010-10-31T23:09:54 and next ~200 posts, or http://wiki.maemo.org/Updating_the_firmware/Cold_Flashing Aug 28 14:55:10 kerio: what about rapuyama Aug 28 14:55:14 kerio: does rapuyama boot? Aug 28 14:55:28 DocScrutinizer05: well, depends on how deep Aug 28 14:55:37 DocScrutinizer05: i've discharged liion to 1.2V Aug 28 14:55:47 DocScrutinizer05: and recharged after that Aug 28 14:55:49 jacekowski: ask sambo7 Aug 28 14:55:52 DocScrutinizer05: and it recovered quite well Aug 28 15:09:00 You probably haven't. Aug 28 15:09:11 If you mean a nokia type battery Aug 28 15:09:33 once it hits 2.5V or so, the protection circuitry kicks in, and disconnects the cell Aug 28 15:14:48 indeed Aug 28 15:14:51 should Aug 28 15:28:11 hmm, but we are talking about cell battery, i don't think there is any protection circuitry built in Aug 28 15:30:11 ~seen javispedro Aug 28 15:30:24 javispedro <~javier@Maemo/community/contributor/javispedro> was last seen on IRC in channel #maemo-ssu, 18h 5m ago, saying: 'freemangordon: you cannot just paste a IDA decompiler output and relicense it under GPL3'. Aug 28 15:39:40 http://www.ti.com/lit/er/swcz001g/swcz001g.pdf SiErr on TPS6595x: 26 VDD1, VDD2, may have glitches when their output value is updated // The OMAP device may reboot if the VDD1 or VDD2 go below the core minimum operating voltage.// Negative glitches may occur on VDD1, VDD2 output. It occurs when the output voltage is changed and the DCDC are running on their internal oscillator. Aug 28 15:39:59 I read that as "don't use SmartReflex, it's instable" Aug 28 15:40:10 but... :c Aug 28 15:40:23 it worksforme! Aug 28 15:40:26 I might be wrong Aug 28 15:42:15 TI might be wrong! Aug 28 15:42:46 hm, is smartreflex even useful if you don't OC? Aug 28 15:43:12 also: 27 VDD1 and / or VDD2 DCDC clock may stop working when internal clock is switched Aug 28 15:43:42 Impact: VDD1 and/or VDD2 output voltages may collapse if clock stops. Aug 28 15:44:56 nice Aug 28 15:47:16 * kerio disables smartreflex Aug 28 15:48:48 i don't know how D: Aug 28 15:52:45 DocScrutinzrWrk: the same could happen when you change the OPP Aug 28 15:52:59 OPP? Aug 28 15:54:44 frequency Aug 28 15:55:07 hm, should i overclock this? Aug 28 15:55:21 DocScrutinzrWrk: also the note says "...DCDC are running on their internal oscillator." Aug 28 15:55:33 is that the case in n900? Aug 28 15:55:39 anyway, removing /etc/default/kernel-power and rebooting gave me the stock settings Aug 28 15:55:47 i'm probably going to *save* these to a file Aug 28 15:57:08 hmm, "default" differs from stock in more than one way Aug 28 15:57:47 the voltage for 250 is a bit lower, and the dsp frequency is 500 instead of 520 for a lot of high frequencies Aug 28 15:59:37 DocScrutinzrWrk: libhildon-im-vkbrenderer3_3.3.20-1+0m5_armel.deb (stock) does not contain license, what now? Aug 28 16:02:10 well, I guess then standard maemo licence applies Aug 28 16:02:24 which is? Aug 28 16:02:53 hmm, where is ivgalvez :( Aug 28 16:02:53 the eula you accept on firmware dl website Aug 28 16:06:12 and probably qgil's special 'licence' for CSSU overrides that EULA Aug 28 16:06:18 DocScrutinizer51: why is that eula applicable? I didn't get the binaries from the device. Maybe I should check what SDK license says. Aug 28 16:07:16 yep, that too Aug 28 16:07:58 Maemo SDK Virtual Image is part of the Maemo Eclipse Integration project. Aug 28 16:07:58 Contents released under the Eclipse Public License v1.0. Aug 28 16:08:19 http://www.eclipse.org/legal/epl-v10.html Aug 28 16:08:42 hehe "each Contributor hereby grants Recipient a non-exclusive, worldwide, royalty-free copyright license to reproduce, prepare derivative works of, publicly display, publicly perform, distribute and sublicense the Contribution of such Contributor, if any, and such derivative works, in source code and object code form." Aug 28 16:10:06 probably only applicable to SDK IDE itself Aug 28 16:10:45 sure, but if there is no license coming with .debs, then what? Aug 28 16:11:05 I am glad I am a developer, not a lawyer :D **** ENDING LOGGING AT Tue Aug 28 16:11:53 2012 **** BEGIN LOGGING AT Tue Aug 28 16:12:23 2012 Aug 28 16:13:18 hmm, there is an additional euale one should agree berfore download, lemme read it Aug 28 16:15:32 freemangordon: the nokia n900 battery has a little protection board in it Aug 28 16:15:36 it's not a bare cell Aug 28 16:15:51 SpeedEvil: we were talking about backup battery Aug 28 16:16:09 Oh - my mistake Aug 28 16:16:22 that does not have any protection at all. I should read further context Aug 28 16:16:33 :) Aug 28 16:22:09 DocScrutinizer51: well, according to that EULA (http://pastebin.com/4Sg73wZj) .debs we are talking about are not copyrighted, or at least that is my undertanding. So I will end that discussion until someone finds a reason why I should change the license Aug 28 16:22:36 DocScrutinzrWrk: ^^^ Aug 28 16:27:42 http://www.ti.com/lit/ug/swcu050g/swcu050g.pdf p.265 (SmartReflex) bottom: >>The registers are described in and .<< NIIIICE! Aug 28 16:27:54 but of course Aug 28 16:28:25 in 3430 TRM there is a note that this information is confidential :D:D:D Aug 28 16:30:01 "This information is not available in public domain" Aug 28 16:31:46 p.409/410 has SR regs Aug 28 16:33:22 does flasher do something more than nandwrite when flashing the kernel? Aug 28 16:34:59 DocScrutinzrWrk: no, those are voltage control registers ;) Aug 28 16:35:20 there are about 10-20 more Aug 28 16:37:45 DocScrutinzrWrk: exactly 13 more registers Aug 28 16:38:41 check in the trm, section 4.14.2.13 Global_Reg_PRM Registers Aug 28 16:44:38 TRM of what? OMAP? Aug 28 16:45:09 yes, OMAP34xx_ES3.1.x_PUBLIC_TRM_vZM.pdf Aug 28 16:46:32 I suppose that are OMAP registers then Aug 28 16:47:03 could be, but does not matter, as SR is actually controlled by those Aug 28 16:47:59 though iirc they are accessed through i2c Aug 28 16:55:07 http://www.ti.com/lit/ug/swcu050g/swcu050g.pdf p.441 ROTFL Aug 28 16:56:05 luckily we don't use any of that in N900 Aug 28 17:06:40 DocScrutinzrWrk: should i remind you that the N900 uses rx51_BME? Aug 28 17:20:12 kerio: how is that related? Aug 28 17:20:23 we're talking hw right here Aug 28 17:21:05 Hiyaa Aug 28 17:21:30 my issue is fixed by a reflash and have an awesome low number of wakeups now via powertop :) Aug 28 17:23:31 yay Aug 28 17:24:22 MrPingu: so what's the conclusion? Aug 28 17:24:35 kernel buggy? Aug 28 17:24:43 It's not the battery Aug 28 17:24:45 ;) Aug 28 17:24:59 aaah the battery issue Aug 28 17:26:00 Battery is just fine so it "must" have been something else... Aug 28 17:26:31 I asked "was it the kernel?" Aug 28 17:26:56 prolly you dunno yet Aug 28 17:27:32 Well mtd2 last oops was from kp47 :p Aug 28 17:27:45 seen that Aug 28 17:28:23 Don't know what to blame in my phone... Aug 28 17:28:29