**** BEGIN LOGGING AT Tue Mar 13 02:59:58 2012 Mar 13 03:42:19 ad-hoc wifi consumes 180mA here Mar 13 03:43:44 it consumes 5mA for powersave managed mode Mar 13 03:44:04 hald the system power Mar 13 03:44:07 half Mar 13 03:45:07 is there possibility to add powersave to adhoc? alternatively to add managed AP mode to driver? Mar 13 03:45:50 No, and dunno Mar 13 03:46:30 Adhoc means that you jsut put the interface in promisc mode, and listen for packets, and hope that tcp does the retransmits. Mar 13 03:46:43 There is no scheduling which is required for powersave. Mar 13 03:47:45 In principle you could roll your own protocol. Mar 13 03:47:52 But it's not there by default Mar 13 03:48:29 yes for w n900s Mar 13 03:51:10 2 Mar 13 03:52:42 very high power usage.. i do not see how listening consumes so much Mar 13 03:53:20 Because it does. Mar 13 03:53:29 Keeping the radio awake is expensive. Mar 13 03:53:54 Use a normal AP, and turn powersave mode off Mar 13 03:57:21 so i want a fully managed mode on n900 Mar 13 03:58:34 Err - what? Mar 13 03:59:59 for mobile hotspot Mar 13 04:01:28 managed is some time division multiplexing? Mar 13 04:02:05 is special/different wifi raio needed? Mar 13 04:02:13 radio Mar 13 04:18:06 thank you for the information SpeedEvil Mar 13 04:26:25 You ... Mar 13 04:26:25 ah Mar 13 08:47:20 talk is offline - again. Mar 13 08:47:31 this sucks without end the third time this month **** BEGIN LOGGING AT Tue Mar 13 09:48:19 2012 Mar 13 09:50:40 Anyone here know what's up with tmo? Mar 13 09:50:53 talk is offline - again. Mar 13 09:51:28 Ok, so this is not scheduled maintainance or something like that? Mar 13 09:52:31 probably not Mar 13 09:52:48 fyi that was an annoyed user's complaint Mar 13 09:54:47 Thanks for the info. I guess since noone here knows it's probably not scheduled outage ;) Mar 13 09:55:58 I sort of have a belief that if there were any scheduled outage, it would be noted in the /topic. Mar 13 09:57:32 Hi guys, I'm looking to overclock my old N900 and was wandering if you could help me with an up-to-date guide? :) Mar 13 09:58:26 have you done the necessary hardware modifications to ensure that your N900 won't fry up in the process? Mar 13 09:59:23 I have done nothing - Was just looking at these old guides on talk.maemo.com (which seems to be down?) and they all seemed like it was a software hack only that wouldn't take more than 2 minutes Mar 13 10:00:32 -Then i saw the name on the package they said to download "Enchanced Linux kernal for power users (obsolete)" with "Please REMOVE this package" in the desc. Mar 13 10:00:33 its not exactly a software hack, it's more like replacing the kernel so that such options can be unlocked at user's expense Mar 13 10:01:25 By user's expense you mean? Mar 13 10:07:47 dixi Mar 13 10:10:09 psycho_oreos - Do you have any guides or advice or anything? :) Mar 13 10:10:43 heh Mar 13 10:10:54 is nokia closing down maemo.org? Mar 13 10:12:26 yes, in 2013 if I remember correctly Mar 13 10:16:06 So.. Anyone else knows anything about overclocking of N900? Mar 13 10:21:44 Lykkepillen: it will kill the device. Mar 13 10:21:56 Lykkepillen: the clock speed is not set for thermal or battery reasons. Mar 13 10:22:19 It will? Thats why the kernel is being removed? Mar 13 10:22:54 well, it may not kill it quickly every time Mar 13 10:23:01 also, what kernel is being removed? Mar 13 10:23:21 BCMM: please explain "kill" meaning :) Mar 13 10:24:43 brick, destroy, eventually-cause-to-become-non-functional Mar 13 10:25:14 if it doesn't kill, it will un-necessarily shortens the life of the device sporadically and exponentially depending on how high the user/owner sets Mar 13 10:29:34 BCMM From my post earlier: -Then i saw the name on the package they said to download "Enchanced Linux kernal for power users (obsolete)" with "Please REMOVE this package" in the desc. Mar 13 10:29:45 -That's what the guide i found told me to download Mar 13 10:29:51 huh, kernel-power isn't gone is it? Mar 13 10:29:55 i thought it was part of cssu now Mar 13 10:30:36 it has all sorts of non-OC uses like iptables Mar 13 10:30:42 No idea what cssu is - My phone didn't work for about a year and when i tried turning it on a few days ago it suddently powered up without problems - I'm outdated on all points Mar 13 10:30:56 ~cssu Mar 13 10:30:58 it has been said that cssu is http://wiki.maemo.org/Community_SSU Mar 13 10:32:18 kernel-provides more than just that, it also incorporates class 10 microsdhc fixes, other improvements within memory management, etc Mar 13 10:34:08 So i should or shouln't overclock my device? It sounded almost like it was allmost risk-free when i read the guide, but that was from 2009 Mar 13 10:35:25 you should question yourself is there ever a need to, and should also read up on the technical information in regards to the very same ARM Cortex A8 chipset inside that N900 Mar 13 10:36:02 ~oc Mar 13 10:36:02 extra, extra, read all about it, oc is an Optical Carrier, An OC1 has 672 channels or 44.736 Mbit/s Mar 13 10:36:08 ~overclocking Mar 13 10:36:08 "OK, listen up. This is your CPU." apt drops the CPU into a hot frying pan. "This is your CPU on overclocking. Any questions?" Mar 13 10:38:23 Theres no need to overclock it. I'm simply doing this because It's possible. When the phone stopped working i bought a Samsung Galaxy S2 that I'm using ATM so the N900 is really just a toy Mar 13 10:39:49 what do you mean stopped working? Mar 13 10:39:57 It just didn't boot Mar 13 10:40:13 Could have been some sort of water damage since it started working again Mar 13 10:41:08 poor old device, soon it'll probably gain rust before it dies.. and to overclock it will quickly send it on the way to hell Mar 13 10:43:05 Aw i hope not.. Was thinking of using the FM transmitter on the beach when summer hits Mar 13 10:44:18 you hope not.. its not like that's going to help anything.. when it was subjected to water, what were your first reactions apart from going out and buy samsung galaxy s2? Mar 13 10:45:51 FM transmitters with minijacks are sold quite cheap everywhere Mar 13 10:46:52 Haha the water damage part was just a guess, It was just laying on my desk. But i spent hours trying to figure out why it stopped working.. Didn't come up with a solution tho'.. When i think about it, same thing happened with my first iPod Touch - Might just be something in the air Mar 13 10:47:14 -But why would i buy a FM transmitter when i have a perfectly fine N900 with Youtube on it? Mar 13 10:50:11 I've noticed that the SIM holder is quite prone to humidity damage (mine seems to be losing its metallic shine); no clue about the insides of the N900 Mar 13 10:53:16 this is strange since I have a (much older) N95 too which has been exposed to the exact same conditions and seems brand new both inside and out Mar 13 10:54:10 the N900, in the other hand, is showing signs of wear in the corners, despite only having been extensively used for less than a year (and being only a little over 2 years old) Mar 13 11:10:13 I have no idea what you guys do to your phones... my beloved n900 survived a wheatbeer attack and is looking like week 2 after unboxing ever since Mar 13 11:10:53 wheatbeer coke mix to be precise Mar 13 11:11:09 wheatbeer and coke.. sounds pretty foamy Mar 13 11:38:32 Will talk.maemo.org be up again sometime soon? Mar 13 11:43:04 chem|st: do you carry it in a pocket? Mar 13 11:52:43 Firefox can't establish a connection to the server at talk.maemo.org. Mar 13 11:52:47 today is a good day Mar 13 12:14:55 yup tmo down for me too Mar 13 12:20:33 concept: you mean display foil, phonepocket stuff: no none, just in my pants Mar 13 13:00:29 BCMM: i see. i used to downclock it sometimes, to save battery..but it doesn't worth it Mar 13 13:00:58 guly: surely it has some sort of idle power saving anyway? Mar 13 13:01:26 hope so :) Mar 13 13:01:33 i have no idea if it's still true on ARM, but on modern x86, it basically works out as better to have the maximum non-idle clock speed available Mar 13 13:02:02 since it gets its work finished and gets back to the low-power idle state faster Mar 13 13:02:17 bear in mind that this was about frequency scaling rather than ordinary over/underclocking Mar 13 13:02:51 the conclusion was that it's not a good idea to automatically scale the freq down on an x86 laptop when on low battery as it will make it run out faster Mar 13 13:03:16 adding undervolting to the equation could well shift the balance - does the n900 let you adjust voltages? Mar 13 13:03:44 that's probably a silly question; i would hope a mobile power system is simpler than that Mar 13 13:04:10 I think it does Mar 13 13:04:34 i think so but i think it can be worst before it helps to save battery time Mar 13 13:04:37 You can do that with a custom kernel, I don't think you can in the standard one Mar 13 13:08:58 http://webcache.googleusercontent.com/search?q=cache:zIIkrpMdDpAJ:talk.maemo.org/showpost.php%3Fp%3D595582+&cd=1&hl=da&ct=clnk&gl=dk Mar 13 13:09:21 -When talk.maemo dosen't work Mar 13 13:14:24 ow your N900 is overclocked to 850Mhz using slightly less voltage than stock!! Mar 13 13:14:25 interesting Mar 13 13:35:06 bah, TMO is down :( Mar 13 13:36:05 it's been down for at least 5 hours Mar 13 13:55:08 guly: rush-to-idle is a well known optimal strategy usually Mar 13 13:55:38 OC though is a pretty bad idea Mar 13 13:55:53 even and esp with undervolting Mar 13 14:34:18 sure Mar 13 14:37:49 overclocking isnt gonna kill your device this week Mar 13 14:38:06 and chances are by the time overclocking would have killed it, you would be ready to move on anyway Mar 13 14:47:44 Could you give me a timespan? I wouldn't push it to the limits, just make it a bit faster Mar 13 14:48:16 Nobody can. Mar 13 14:48:38 I it likely to kill it in the time the device still works - possibly not. Mar 13 14:49:02 So.. 6 months? 1 year? Mar 13 14:50:37 Sazpaimon: within a yeaR? Mar 13 14:52:54 luke-jr, if you run it at 100% cpu for a year straight, then maybe Mar 13 14:53:15 Sazpaimon: you wouldn't get to a year at 100% CPU, at default clocks Mar 13 14:53:35 yeah probably not Mar 13 14:53:54 still most of the time you're overclocked, the cpu wont even be running at that clock Mar 13 14:54:46 there are people that have been overclocking for a year already, I havent seen any complaints thus far Mar 13 15:00:16 I am OC since aplil/may 2010, till kp49 up to 900, when kp49 came out up to 805 Mar 13 15:00:59 Well good to know.. I will check back tomorrow to ask a few more questions before i decide lol.. :) Mar 13 15:07:47 well, when your car's engine has a red zone and limiter @ 6000 RPM, is it sane to removeavtune the limiter so you can run the engine @ e.g. 8000 RPM? how long can you do? Mar 13 15:09:49 ah once again, the old car/computer analogy Mar 13 15:10:23 DocScrutinizer, I drive my BMW OC from 136 HP to about 170 HP for 5 or so years :p. No problem at all. Mar 13 15:11:37 and that analogy doesnt really hold up Mar 13 15:12:02 I disagree. Mar 13 15:12:08 if you remove the rev limiter, you would also need to replace the exhaust for it to do anything Mar 13 15:12:12 In both cases, you can't tell what will fail. Mar 13 15:13:29 it simply doesn't worth it Mar 13 15:13:35 the risk is to blow up a device Mar 13 15:13:44 the pro is to speed up just a bit Mar 13 15:37:28 A device that is not used because it is slow has effectively blown up. Mar 13 15:37:37 [10:52:54] luke-jr, if you run it at 100% cpu for a year straight, then maybe Mar 13 15:37:50 if you run 100% CPU at 600MHz it'll like fail Mar 13 15:38:10 MohammadAG: that's what I said Mar 13 15:38:23 SpeedEvil: 600 MHz isn't slow. the RAM is where N900 is lacking Mar 13 15:38:33 luke-jr: often Mar 13 15:38:36 luke-jr, just noticed Mar 13 15:39:05 [11:07:47] well, when your car's engine has a red zone and limiter @ 6000 RPM, is it sane to removeavtune the limiter so you can run the engine @ e.g. 8000 RPM? how long can you do? Mar 13 15:39:14 I just stick a paper with a drawn extended gauge Mar 13 15:39:24 then broke the needle that holds it at maximum Mar 13 16:13:19 all decent scientific studies show that user can't even notice 25% faster CPU, esp when other componenys (RAM etc) are still same speed, and when slow storage (HDD, flash) kicks in, clockspeed of CPU is negligible Mar 13 16:15:27 so except for that rage case where you get to 24fps but you'd need gt 25fps, OC is TOTALLY useless Mar 13 16:15:38 rare cases* Mar 13 16:47:57 Is it easy to make N900 to dual boot? If i have one OS in the phone and the other in memory-card of course. Mar 13 16:51:11 yes Mar 13 16:51:54 bootmenu, u-boot, flasher -- those are your options Dynamit, depending on how the second OS is configured Mar 13 16:53:26 u-boot works better for nemo/mer and multiboot for nitdroid. Mar 13 16:53:58 okey i don't have my N900 yet but i hope i have it on Thursday for hope for the Swedish post office to get it from a place near Göteborg to a place quiet near Örnskuldsvik Mar 13 16:54:14 ok thx for the info Sicelo and beford Mar 13 17:00:17 Are Maemo under development or have Nokia abandoned it? I have read that Nokia planning/have abandoned MeeGo. Mar 13 17:00:52 nitdroid/multiboot aren't advised. that's why i left them out in my list Mar 13 17:18:53 ~maemo-multiboot Mar 13 17:18:54 methinks maemo-multiboot is deprecated, and a horrible hack. PROBLEMS WITH NITDROID/MULTIBOOT? reflash rootfs&kernel aka COMBINED Mar 13 17:43:10 Is't any support for Ubuntu One sync in Maemo? Mar 13 17:49:22 Dynamit, https://bugs.launchpad.net/ubuntuone-client/+bug/512052 Mar 13 17:59:16 reading that link makes me laugh :D Mar 13 18:00:54 Marktrix (markgerlach) wrote on 2010-10-24: I am not a Developer and no Part of Canonical. But I think, the focus is iOS and Android, Maemo go out of date and Canonical needs cost-effectiveness. So first the Bigplayers and than maybe (hopefully) Meego Mar 13 18:02:28 I maybe will use Dual boot then it appearing to be MeeGo and Android or Maemo and Android Mar 13 18:03:36 The only reason i want Ubuntu One in the phone when i get it is because then i can writing in the calender in the mobile and get it sync on many computer's i use Mar 13 18:04:18 That would make my school studding more effective then it's today Mar 13 18:04:58 that's not what i was laughing about though Dynamit. i just find it dumb that the post is clearly about N900, and someone already made attempts to get One on N900. then some apple fanboy pops up and say drop it Mar 13 18:06:20 anyway, looks like canonical did 'follow' that piece of 'dis-advice' - http://voices.canonical.com/ubuntuone/2011/05/09/some-changes-to-contacts-2/ Mar 13 18:07:20 stupid Apple G** person then Mar 13 18:09:31 btw, lol, i don't even know what this One is, although it seems to be an Ubuntu-only feature.. /me has Debian Mar 13 18:10:39 you can use Ubuntu One on Windows too Mar 13 18:10:44 Android i Know Mar 13 18:11:35 ah, news to me. windows & ubuntu one :P Mar 13 18:12:53 https://one.ubuntu.com/downloads/windows/ Mar 13 18:19:26 cooool shit Mar 13 18:19:43 * DocScrutinizer is temped to continue where he left yesterday Mar 13 18:20:29 I.E. rant about canonical and buntkuh, and the idiotic product philosophy behind it, and the poor maintainability Mar 13 18:22:07 ubuntu == "linux" for bored windows users to feel "smart" Mar 13 18:25:49 /kick DocScrutinizer !! Mar 13 18:25:50 :D Mar 13 18:39:05 well, N900 made me start to use Linux... and somehow 'dictated' that i should choose Debian. i didn't like Ubuntu for the very same reason DocScrutinizer mentions. guess one future day i'll look into other linuxes (Arch, Gentoo, and such) :/ Mar 13 18:45:55 dunno if i'll really do Gentoo .. compiling everything is nice for learning -- but i'm not so sure doing it all the time is the best idea for me and types of computers i use. therefore, i think Arch looks like my next one. Mar 13 19:01:51 ping freemangordon Mar 13 19:19:08 so i have n900 now Mar 13 19:19:47 the usb port didn't wobble, don't know what to do with it next Mar 13 19:20:46 ketas: install cssu ;) Mar 13 19:26:13 merlin1991: i would like to have cshu Mar 13 19:26:34 ketas: http://wiki.maemo.org/Community_SSU Mar 13 19:28:07 merlin1991: "Update available for your MicroUSB port" Mar 13 19:28:07 :P Mar 13 19:33:19 lol Mar 13 19:40:14 round-trip min/avg/max/stddev = 119.768/2357.527/10736.500/2708.779 ms Mar 13 19:40:19 i love 3g Mar 13 19:40:21 :) Mar 13 19:41:26 no, it was 100.224/5976.568/27335.730/6531.975 ms actually Mar 13 19:42:16 what i find is that packets get queued somewhere, and then eventually arrive with a delay up to multiple minutes Mar 13 19:42:32 which is ridiculous, as any internet protocol will have timed out by then Mar 13 19:43:47 for example, a dns lookup... a few attempts are made to sed the query udp packet, but no response arrives... a minute later, the response packet to each request arrives... but the source port is closed already, and so for each one, an icmp destination unreachable is sent back to the dns server Mar 13 19:44:46 yea, sucks Mar 13 19:45:43 I've seen pings over 2 minutes -.- Mar 13 19:47:54 now, imagine you want to use ssh Mar 13 19:47:59 :P Mar 13 19:48:22 packet arrives 2 minutes later Mar 13 20:15:31 changed router at home and when I try to use apt-get on wlan i get "could not contact dns servers". browser works without problems. Mar 13 20:19:32 hmm Mar 13 20:19:41 how so Mar 13 20:20:12 shanttu: can you resolve a host? Mar 13 20:26:36 ketas, nope Mar 13 20:27:12 tried google dns/ opendns ? Mar 13 20:29:21 well, use CSD, it is almost realtime. Wait... N900 modem doesn't support CSD, not fashionable anymore :-S Mar 13 20:31:59 I used gui to set opendns and thought it was set. i changed it on /etc/resolv.conf and voila Mar 13 20:42:25 fremantle is not using /etc/resolv.conf, or rather: it must not get altered, as on fremantle you got dnsmasq Mar 13 20:43:55 so changing /etc/resolv.conf to anything other than 127.0.0.1 will break things Mar 13 20:44:18 device can't associate to new networks Mar 13 20:47:50 DocScrutinizer, the truth is now I can use apt-get. i'm extremely noob regarding networking Mar 13 20:48:33 the truth is this will change eventually Mar 13 20:49:01 will change it back then Mar 13 20:49:14 another truth is you haven't got any DNS cache now anymore, as far as I can tell Mar 13 20:49:39 so *each* DNS query now goes via network Mar 13 20:50:21 in short: you fsckdup your DHCP Mar 13 20:52:00 next hotspot you will run into trouble, as they usually rely on their own DNS server getting used at very least for authentication via browser frontpage Mar 13 20:52:26 mmh dnsmasq is writing a resolv.conf foreach different connection Mar 13 20:52:33 yep Mar 13 20:52:35 and nameserver used is the one got via dhcp Mar 13 20:52:37 so Mar 13 20:52:52 ->127.0.0.1->dnsmasq->192.168.1.1 Mar 13 20:52:54 exactly what I tried to say Mar 13 20:53:22 and also dnsmasq caches Mar 13 20:53:36 to speed up repeated queries *a lot* Mar 13 20:53:44 sure Mar 13 20:58:44 reading about opendns now and will try that. thanks for your help Mar 13 21:03:32 ummm Mar 13 21:07:56 maybe you simply should delete your home wlan form list of networks, then create it new - obviously after you first switched /etc/resolv.conf back to default 127.0.0.1 Mar 13 21:08:47 checking the config of your new AP also might help a lot Mar 13 21:09:37 and of course restarting the phone also may help to recover from hidden massive changes in WLAN network environment Mar 13 21:11:51 DocScrutinizer: ping Mar 13 21:12:53 DocScrutinizer: i've disassembled my n900, usb port "jumped out" a bit from board on one side Mar 13 21:14:15 when pressing it a bit from top to make it really touch the board, it works Mar 13 21:16:25 yeah Mar 13 21:16:32 semi-off Mar 13 21:16:40 you opened it just in time Mar 13 21:17:01 dos1: so, what's the question? :-D Mar 13 21:17:08 DocScrutinizer: just reporting :) Mar 13 21:17:29 solder it down, touch the connetion pins shortly with soldering iron, good you are Mar 13 21:17:57 don't apply tin to the contact pins when soldering them Mar 13 21:18:28 good flux maybe, but no tin beyond what's inevitably on your soldering tip already Mar 13 21:18:48 so basicaly nothing Mar 13 21:19:20 ok Mar 13 21:20:11 save all the lots of solder for the mach reinforcement blobs at left and right of component ;-) Mar 13 21:20:19 mech* Mar 13 21:20:30 and as always Mar 13 21:20:32 NO EPOXY Mar 13 21:20:40 YUP! :-D Mar 13 21:21:12 thanks for tips, i'll report in few days how it went (i don't have soldering iron here right now) Mar 13 21:21:18 hm I wonder how my usb port in my n900s is doing Mar 13 21:21:57 dos1: I guess it's redundant, and already known, but anyway: Mar 13 21:22:02 ~usbfix Mar 13 21:22:02 well, usbfix is http://talk.maemo.org/showthread.php?t=75920 - and **NEVER** use epoxy (unless you want to seal your device for underwater) Mar 13 21:22:37 hm I gues it's impossible to tell if they are broken from the outside Mar 13 21:22:48 yes, mostly Mar 13 21:23:16 you might tell when it moves 0.2mm when you press it "down" with some toothpick Mar 13 21:24:08 actually you can see this very well, as there's a 0.5mm gap from USB to plastic of case Mar 13 21:24:52 down means towards the battery side of the device? Mar 13 21:25:10 yep Mar 13 21:25:33 i tried to push it with stylus when checking for my n900 Mar 13 21:25:38 well, basically it mustn't move in either sirection Mar 13 21:25:43 wasn't wobbly Mar 13 21:26:02 they all seem to be rock solid Mar 13 21:26:19 stylus is a good idea Mar 13 21:26:25 yet, do i need to harden it with solder? Mar 13 21:26:28 still Mar 13 21:27:01 depends on your sleep, are you suffering insomnia when you think about N900 USB_? Mar 13 21:27:34 i don't really want it to break off in future Mar 13 21:27:42 are you a very gentle guy or rather a blacksnith? Mar 13 21:27:46 smith* Mar 13 21:28:22 i don't like to twist cables around really Mar 13 21:28:23 wouldn't the safest method be to remove the connector and solder a pigtail cable instead? ;) Mar 13 21:28:23 hm I wonder how blacksmith soldering would look like :P Mar 13 21:28:28 hm Mar 13 21:28:36 what you meant? Mar 13 21:28:53 r00t-eeepc: well... Mar 13 21:29:04 r00t-eeepc: i thought that, but that's stupid Mar 13 21:29:09 it's really hard to tell what's more risk - disassembling and trying to solder it while no appropriate expertise, or hoping for your luck not to trip over the cable Mar 13 21:29:41 2nd case at least gives you the chance that you don't try todo somehting that might break it Mar 13 21:30:39 ketas: it _looks_ stpid, but it's the best way to prevent destruction of the port ;) Mar 13 21:31:12 of course there are characters that can't build card houses without folding the cards. Those should try to find somebody to do the soldering for them Mar 13 21:33:09 I got no problems with 2 devices with all together >3 years of frequent USB plug-rounds Mar 13 21:33:26 both still fine Mar 13 21:33:37 no preemptive fixing Mar 13 21:33:49 hmm Mar 13 21:34:03 a method i find reasonable is to remove the locking hooks from the micro-usb plug, so it will slip out rather than rip out the port Mar 13 21:34:08 so i leave it without fix for now? Mar 13 21:34:09 but then again, I'm an EE who's supposed to know how to handle that kinda delicate stuff Mar 13 21:34:46 ketas: yours is broken, no? so fixing is obviously needed Mar 13 21:35:04 r00t-eeepc: that's basically BS, you never 'rip out' the port, it always gets loose by levering/bending the plug Mar 13 21:35:12 r00t-eeepc: not yet :P Mar 13 21:35:31 DocScrutinizer: i've seen phones with ripped out port on ebay ;) Mar 13 21:35:51 so be carefull on pulling the plug, don't bend it, and you're most likely fine Mar 13 21:36:13 someone claimed that it just came out Mar 13 21:36:22 DocScrutinizer: and it will also reduce stress on the port by making the plug fit less tight Mar 13 21:36:22 r00t-eeepc: so what? sure you can rip out a loose receptacle component Mar 13 21:37:14 hmm Mar 13 21:37:24 do they make usb plugs without hooks Mar 13 21:37:46 well, it's your cable. easy to replace once you ruined it. I never found it's needed as all my cables go in and out just fine, after ~200 rounds Mar 13 21:37:52 ketas: rip out with appropriate pliers Mar 13 21:38:10 i'll see Mar 13 21:38:16 now that's complete BS Mar 13 21:38:45 DocScrutinizer: rip the hooks out of the plug i mean, to make a hook-less plug Mar 13 21:38:46 go in and out just fine... Mar 13 21:39:06 yes, and I answered on that Mar 13 21:39:27 how else would you remove them (if you wanted to) Mar 13 21:39:40 with teeth Mar 13 21:39:50 meh, how else would you make a pig fly Mar 13 21:40:01 with wings of course Mar 13 21:40:02 ketas: it's unlikely your teeth have that kind of grip ;) Mar 13 21:40:11 simple answer, the whole idea is BS Mar 13 21:40:19 DocScrutinizer: possible Mar 13 21:40:21 also Mar 13 21:40:47 some usb cables are fucking thick Mar 13 21:40:56 which might be better for signal but... well Mar 13 21:41:04 DocScrutinizer: i'm not necessarily saying it's a great idea or helps much - but it won't hurt (the usb port) either Mar 13 21:41:17 will cause plug to fall out when no latches at all Mar 13 21:41:55 yes, but having the plug fall out is sure easier on the port than pulling the phone around by it Mar 13 21:42:20 ketas: when your plug is a bit tight at beginning, paint the outside (!) with a pencil Mar 13 21:42:46 hmm Mar 13 21:42:51 interesting idea Mar 13 21:42:56 graphite lubrication? Mar 13 21:43:00 yep Mar 13 21:43:00 yes Mar 13 21:43:13 what out inside :P Mar 13 21:43:18 about Mar 13 21:43:24 And still even without the hooks if you try on the cable you might be subjecting hte port to bending forces still, before the velocity vector of cable and attitude of n900 reorients to allow for plug to come out Mar 13 21:43:54 thanks ShadowJK, was too lazy to explain it Mar 13 21:43:58 ShadowJK: now that's some serious connector physics ;) Mar 13 21:44:16 i've seen massive microusb plugs with massive cables Mar 13 21:44:18 try=trip Mar 13 21:44:23 and hard to bend cables Mar 13 21:44:46 yes, that's why you want hooks on those Mar 13 21:44:48 those cables nokia provide are somewhat easier to work with Mar 13 21:45:52 USB *always* breaks on you pulling on the plug not in line with insertion direction Mar 13 21:45:58 aka bending it Mar 13 21:46:05 on pulling out Mar 13 21:46:18 hm? Mar 13 21:46:31 breaks what, how? Mar 13 21:46:53 your USB never will come off on a proper pull out in straight line, not even with hardest latches on plug Mar 13 21:47:04 now, what we REALLY need is induction coupled phone charging... Mar 13 21:47:20 if you need it, you should get it Mar 13 21:47:36 it's been done before Mar 13 21:47:49 helsinki airport has café with induction chargers under the tables Mar 13 21:47:50 well induction charing on phone is kind of hard to design, no? Mar 13 21:47:58 charging Mar 13 21:48:05 it's available somewhere... but probably too expensive to be worth it Mar 13 21:48:23 ~80..100 bucks for material IIRC Mar 13 21:48:36 with a spare for palm pre Mar 13 21:49:09 some cute hacker already did it Mar 13 21:49:13 i just bought an n900 for 100 bucks... Mar 13 21:49:16 on N900 Mar 13 21:54:15 wb Mar 13 21:54:36 oops, much late, nvm that Mar 13 21:56:35 bucks and boosts Mar 13 21:56:52 pali: when you mean 2.6.35 needs proc/bootreason and gpio keys to work... could you point in the right direction for further reading? Mar 13 21:58:07 Hurrian, szopin, if you want I can send my patch for bootreason and gpio to work Mar 13 21:58:30 patch is for meego n900 kernel from git Mar 13 21:58:47 and it create compatible interface for maemo 5 fremantle Mar 13 22:00:27 i'd like that Mar 13 22:00:36 let's see what can be built **** BEGIN LOGGING AT Tue Mar 13 22:04:38 2012 Mar 13 22:05:13 wow, I found my "book to read" for this night: /sys/kernel/debug/kprobes Mar 13 22:05:22 err, /sys/kernel/debug/* Mar 13 22:06:23 hmm, still quiet? Mar 13 22:06:35 yay works... thanks Pali Mar 13 22:07:08 disco_stu_droid: blasphemy Mar 13 22:08:00 cool, going to get debian 6 and scrathbox then Mar 13 22:08:16 hah, /sys/kernel/debug/wl12xx/fw-statistics/* wow Mar 13 22:08:25 sopin, Hurrian, patch is here: http://atrey.karlin.mff.cuni.cz/~pali/meego_n900_kernel_fremantle.patch Mar 13 22:08:48 against this git repo: http://meego.gitorious.org/meego-device-adaptation/n900_kernel Mar 13 22:09:03 please note that you need GCC 4.6 Mar 13 22:09:07 pali: is it possible to compile >.28 in SB, using gcc 4.6 for example? Mar 13 22:09:38 I think scratchobox does not compiling it correctly Mar 13 22:09:48 better install ARM GCC cross compiler Mar 13 22:10:05 oh Mar 13 22:10:06 and also, this patch is very hacky and is not complete!!!! Mar 13 22:10:09 so to get 35 running one can scratch scratchbox? Mar 13 22:10:10 no need for debian then Mar 13 22:10:17 brb, booting ubuntu vm Mar 13 22:10:22 :) Mar 13 22:10:39 ubuntu has package gcc-arm-linux-gnueabi Mar 13 22:10:56 and with it you can easy cross compile n900 kernel without problem Mar 13 22:10:57 @Pali, do I need to build newer binutils for fremantle to boot this kernel? Mar 13 22:11:04 no Mar 13 22:11:10 nice Mar 13 22:11:18 also, i see that you build k-pwr with gcc 4.2.1 Mar 13 22:11:23 libc6/libglib? Mar 13 22:11:26 would that also be necessary? Mar 13 22:11:26 also please note that I did not tested it on real n900 yet!!!! Mar 13 22:11:40 I can crash test it for you Mar 13 22:11:47 in qemu it has problem with BME Mar 13 22:11:49 got a freshly reflashed one Mar 13 22:11:55 you can compile the meego 2.6.37 in the harmattan sdk, worked fine for me. Mar 13 22:11:58 I needed to remove /etc/event.d/bme Mar 13 22:12:00 so just say a word Mar 13 22:12:26 @Pali, wouldn't that means that rd mode should be on? Mar 13 22:12:37 wait, kernel dev, riiight. turning rd on. Mar 13 22:12:40 in qemu is R&D mode on by default Mar 13 22:13:19 are you guys preparing 37 thumb2 release by any chance??? Mar 13 22:13:38 @sopin: /nemomobile/ Mar 13 22:13:48 I have no time to experiment with this now Mar 13 22:13:56 :/ Mar 13 22:14:05 but you got my last patch which has fixed problem with NAND Mar 13 22:14:13 and also with /proc/bootreason Mar 13 22:14:35 if you have things to try out, throw em at me Mar 13 22:14:41 so you can play with that code Mar 13 22:14:54 time to test and report is << learning kernel compilation Mar 13 22:15:04 will do, thanks Mar 13 22:15:07 but fremantle BME must be removed, because it working only with 2.6.28 Mar 13 22:15:14 @sopin - make rx51_defconfig Mar 13 22:15:19 cross compilation is easy Mar 13 22:15:25 uhh Mar 13 22:15:34 export ARCH=arm; make rx51_defconfig; make -j12 Mar 13 22:15:56 and export CROSS_COMPILE=arm-linux-gnueabi- before all Mar 13 22:15:57 @Pali, you don't need to specify the compiler? Mar 13 22:16:03 ah there. Mar 13 22:16:06 does meego provide replacement for BME that could be butchered in with newer kernel? Mar 13 22:16:13 no idea Mar 13 22:16:20 @sopin, yes, but it may not play with fremantle systems Mar 13 22:16:24 but BME can be replaced by my kernel module Mar 13 22:16:29 when it will be finished Mar 13 22:16:35 :D Mar 13 22:16:41 looking forward to it Mar 13 22:16:58 last version was announced on TMO in usb host mode thread Mar 13 22:17:25 there are many 100 pages+ threads there I need to dig through Mar 13 22:17:51 will get there eventually Mar 13 22:18:00 but be carefull what you will do!! I really really did not tested that kernel in real n900 Mar 13 22:18:21 hmm... flashing works always as long as hw is ok, no? Mar 13 22:18:31 your n900 might burst into flames Mar 13 22:18:37 is it really possible to brick it with just sw? Mar 13 22:19:11 the u pressing boot is outside of any possible sw infringment I though Mar 13 22:19:16 t Mar 13 22:19:39 DocScrutinizer, see ^^^ what do you think? it is possible to brick n900 with sw (bad kernel)? Mar 13 22:20:08 like rom Mar 13 22:20:14 no way Mar 13 22:20:18 i don't think bad kernel causes bricking Mar 13 22:20:44 but is there a way to make it unbootable and unflashable? Mar 13 22:20:48 apart from hammer? Mar 13 22:20:56 you can damage some peripherals with really nasty kernels (like bq27x00 I2Cbus "fix" ;-D) Mar 13 22:21:20 hmm what that did? Mar 13 22:21:29 you mean to force start BME and bq27x00 battery? Mar 13 22:21:36 both at same time? Mar 13 22:21:55 it "might" trigger a bad i2c condition Mar 13 22:22:15 and that damages what and how Mar 13 22:22:22 that kernet does not support fremantle BME, so this problem is solved :D Mar 13 22:22:22 charger chip iirc Mar 13 22:22:46 in qemu when BME tried to start it send some signal to DSME and DSME rebooted system Mar 13 22:22:46 I mean collision on I2C, due to some genius patched out locking Mar 13 22:23:08 not applicable to current PK versions afaik Mar 13 22:23:16 well, got my N900 unchargeable in many ways, but flashing always work (in the end) Mar 13 22:23:28 you need to full battery for flashing Mar 13 22:23:35 unless you mean hw dmg Mar 13 22:23:37 @sopin, just make sure you load this using u-boot or flasher -l Mar 13 22:23:41 there are other way too to "brick" N900 via sw, but for obvious reasons I'll not elaborate Mar 13 22:24:19 I know only one theoretical way how to brick n900 via sw (without damaging HW) Mar 13 22:24:22 so there is no ROM for handling powerup with 'u' pressed? Mar 13 22:24:39 @sopin, it's seen by nolo, which is not in rom Mar 13 22:24:41 Pali: cosnsists of A, L, and C Mar 13 22:25:02 always thought this was what saved me Mar 13 22:25:11 again and again recently Mar 13 22:25:19 Did they ever get flash playeer working on maemo? Mar 13 22:25:25 @sopin, plugging usb while off into flasher automatically gets you into flasher mode Mar 13 22:25:36 so without UV light we're safe, bad news I guess Mar 13 22:25:40 there is no need to press "U" Mar 13 22:25:59 DocScrutinizer, did somebody already try to delete that? Mar 13 22:26:05 Hurrian: TIL, wow Mar 13 22:26:12 not afaik Mar 13 22:27:13 and what is different between to n900's cal? Mar 13 22:27:26 only BT & WIFI addr? Mar 13 22:27:42 ser nr, localization Mar 13 22:28:01 some other mysterious crap Mar 13 22:28:13 on my server you find a dump of my CAL Mar 13 22:28:25 actually a number of dumps Mar 13 22:28:36 so if somebody erase it and we found some magic way how to flash other cal dump, will be n900 working again? Mar 13 22:28:41 filenames tell how they were done Mar 13 22:28:57 Pali: prolly yes Mar 13 22:29:17 that's why those dumps are there Mar 13 22:29:56 coldflashing into a rescue-initrd might actually work. Then use mtd_utils to restore Mar 13 22:30:20 DocScrutinizer, and will NOLO start with bad cal? Mar 13 22:30:30 that's the big question Mar 13 22:30:37 coldflashed NOLO should Mar 13 22:31:08 * ketas thinks where to start with his n900 Mar 13 22:31:16 it's still in the box :P Mar 13 22:31:35 maybe it feels lonely there Mar 13 22:31:44 ketas: start by *using* it without *any* crap and 'mods' installed Mar 13 22:31:45 I played with U-Boot SPL build and I forced X-loader to load U-Boot SPL instead NOLO Mar 13 22:31:48 if nokia made their software badly, it's an insta-brick Mar 13 22:32:13 ketas: ...to get a feeling how a working maemo fremantle feels like Mar 13 22:32:14 so maybe one day I can replace NOLO with U-Boot SPL Mar 13 22:32:17 @Pali, how did you get around signing issue? Mar 13 22:32:27 NOLO is not signed Mar 13 22:32:29 only X-loader Mar 13 22:32:33 ooh Mar 13 22:32:36 DocScrutinizer: of course Mar 13 22:32:38 and I decrypted X-Loader protocol Mar 13 22:32:54 its easy (I posted info to TMO thread) Mar 13 22:33:03 Pali: hey! :-D Mar 13 22:33:22 I only tested it in qemu + gdb Mar 13 22:33:29 @Pali, but you still don't know what exactly NOLO does, right? Mar 13 22:33:33 but it freeze at serial console instructions Mar 13 22:33:37 I've been treating it very badly, changing .so to ubuntu versions and back, never bricked totally, even if required few pushes to tart flashing Mar 13 22:33:39 yep, Hurrian Mar 13 22:33:43 yes Mar 13 22:33:44 that's the main problem Mar 13 22:33:48 DocScrutinizer: first i need to backup it, just in case and then look if previous owner did clean it up or not Mar 13 22:33:53 thought it was unbrickabe by sw Mar 13 22:34:02 @sopin, it technically is Mar 13 22:34:10 time to change game plan and use the spare for crash testing it seems Mar 13 22:34:18 if you have all possible service tools, it's unbrickable Mar 13 22:34:21 ketas: aaah, so you want to install sth actually: backupmenu Mar 13 22:34:24 DocScrutinizer, where do you have cal dumps? Mar 13 22:34:27 was using it as phone to limit wear on the newer one Mar 13 22:34:34 Pali: on my server Mar 13 22:34:39 I'd like to look at in via calvaria Mar 13 22:34:43 address? Mar 13 22:34:46 you'll not miss them when you look there Mar 13 22:35:06 maemo.cloud-7.de Mar 13 22:35:31 you should know this URL already Mar 13 22:35:33 Hurrian: service tools as in special sw/cables and screwdrivers? or available for dl progs? Mar 13 22:35:57 the latter Mar 13 22:36:09 OMAP downloads code from usb on boot Mar 13 22:36:13 http://maemo.cloud-7.de/ Mar 13 22:36:18 make it clickable :/ Mar 13 22:36:21 so no problem wrt hw flashers etc Mar 13 22:36:41 @DocScrutinizer, IMEI is not stored in CAL, but in BB5, right? Mar 13 22:36:42 no problem Mar 13 22:36:50 Hurrian: right Mar 13 22:36:58 actually, both Mar 13 22:37:05 iirc Mar 13 22:38:09 hmm, haven't tried, are n900 speakers as good as 5800's? :) Mar 13 22:38:43 DocScrutinizer, in your dump are a lot of sections "wlan-tx-cost3_0" Mar 13 22:38:57 nfc Mar 13 22:39:57 ketas: I dunno 5800, but N900 speakers not top notch Mar 13 22:40:23 5 Mar 13 22:40:27 I mean they're not exactly bad, but e.g. N810 had better ones Mar 13 22:40:36 DocScrutinizer: I got a 7" tablet - as a gift. Mar 13 22:40:38 @ketas, significantly worse than a N95's Mar 13 22:40:51 @SpeedEvil, running android, i presume Mar 13 22:40:54 DocScrutinizer: The audio is so badly designed that if you cover the speaker holes, the sound gets better Mar 13 22:41:04 ouch XP Mar 13 22:41:18 Wopad I7 Mar 13 22:41:33 DocScrutinizer: well 5800 speakers are quite amazing Mar 13 22:41:34 I use it for casual browsing - I haven't really investigated it deeply Mar 13 22:41:40 proper name for that Mar 13 22:41:59 >RK2918 - rockchip Mar 13 22:42:02 damn that's cheap Mar 13 22:42:07 doc: smscb.py - command line sms interface or for netwok messages? Mar 13 22:42:14 DocScrutinizer: lower frequencies are quite well represented Mar 13 22:42:19 It's a nice tablet - but for the audio. Mar 13 22:42:26 Well - for the price. Mar 13 22:42:28 btw take a bit of care about your N900 speakers - you know you potentially could blow them with the power of amp Mar 13 22:42:44 ketas: ^^^ Mar 13 22:42:56 hah Mar 13 22:43:02 so better no indiscriminate messing with ALSA's guts Mar 13 22:43:13 anyone burned out camera led? Mar 13 22:43:17 :) Mar 13 22:43:24 no such reports Mar 13 22:43:26 @ketas - nope Mar 13 22:43:29 well you can Mar 13 22:43:36 yes, we now Mar 13 22:44:00 SpeedEvil and me (iirc) found out about that, and how to avoid it Mar 13 22:44:02 i rather love devices without restrictions Mar 13 22:44:37 meh, restrictions conflictions Mar 13 22:44:48 what restrictions? Mar 13 22:45:28 the restriction that the amp is strong enough to blow the speakers, while on other devices you got the restriction of amp NOT being strong enough for that? Mar 13 22:45:40 so you can overvolt battery for it to explode on a plane, take that tsa Mar 13 22:45:50 flashmode @ 500mA and not turn it off in time? loel Mar 13 22:45:58 BS Mar 13 22:46:11 you're free to burn your pocket Mar 13 22:46:19 doesnt the bl-5j have built-in chip? Mar 13 22:46:27 sopin: don't tell nonsense here Mar 13 22:46:32 jk Mar 13 22:46:44 but no restrictions would be just that Mar 13 22:46:53 free to kill your device as you want Mar 13 22:46:57 a) hurrian is right, b) bq24150 charger chip limits at 4300mV Mar 13 22:47:12 stallman doesn't use N900 prolly because of it Mar 13 22:47:40 open: not really, free: far from it Mar 13 22:47:54 no, stallman would not use N900 because of sgx, bb5, bcmfw etc. Mar 13 22:48:03 It is in many places illegal to sell an open cellphone. Mar 13 22:48:06 with meego on top? Mar 13 22:48:16 stallman won't use any HDD either Mar 13 22:48:18 @sopin, meego-n900 comes with bcmfw Mar 13 22:48:23 Open in the sense that the user can get to the modem furmware. Mar 13 22:48:28 DocScrutinizer, these sections are in your cal dump (removed duplicates): als_calib bme cert-ccc cert-hwc cert-npc content-ver fmtx_pwl kernel-ver lock_code nolo-ver part_table phone-info root_device sw-release-ver wlan-tx-cost3_0 Mar 13 22:48:29 oh... Mar 13 22:48:30 And do stuff. Mar 13 22:48:49 is there some info what which section means? Mar 13 22:48:51 he does have an open laptop though last I heard, from skorea Mar 13 22:48:54 @SpeedEvil, byt hat definition any tegra phone should be illegal Mar 13 22:49:00 all free drivers etc Mar 13 22:49:01 @sopin, it's chinese Mar 13 22:49:08 sure he has a hard disk there Mar 13 22:49:16 could be Mar 13 22:49:17 By open, I mean that all the code is available. Mar 13 22:49:23 Including that running on the modem. Mar 13 22:49:34 that probably violates patents Mar 13 22:49:37 gsm patents Mar 13 22:49:47 Pali: nope Mar 13 22:49:51 ofono violates? Mar 13 22:50:00 DocScrutinizer: can n900 also blow your mind? Mar 13 22:50:00 ofono is a RIL Mar 13 22:50:20 ofono is NOT the firmware that runs on the baseband Mar 13 22:50:22 Patents aren't teh point. In the UK, once the first person abuses the openness, and it's reported to you, you have to fix it, or any further phones you sell are 'hacking tools' Mar 13 22:50:43 and i assume this is regarding to IMEI's? Mar 13 22:50:50 Hurrian: please stop using "@" Mar 13 22:50:53 This is not about 'AT' access to the modem. But for example, using it in GSM 'promiscuous mode' or changing IMEI Mar 13 22:50:56 whoa, that explains maemo fate Mar 13 22:51:06 SpeedEvil: and hacking tools are bad how? Mar 13 22:51:17 ketas: They put you in prison. Mar 13 22:51:20 DocScrutinizer: sorry, force of habit from another IRC net Mar 13 22:51:35 SpeedEvil: by just having them? Mar 13 22:51:45 ketas: For selling phone hacking tools. Mar 13 22:51:48 xchat highlights those correctly with or without @ Mar 13 22:51:49 Or phone network. Mar 13 22:51:54 not sure about irssi Mar 13 22:52:01 SpeedEvil, check "efs" partition on tegra devices, or any cheap chinese crap Mar 13 22:52:13 the cheap chinese crap has those "debug" features often enabled Mar 13 22:52:39 Hurrian: And yes. I'm not saying it makes sense, or it has been actionable against anyone. Mar 13 22:52:48 sopin, nokia has done a lot in the years and money it has spent on BB5 just ot not be able to do this stuff Mar 13 22:53:13 that law is fking ridiculous imo Mar 13 22:53:25 which law? Mar 13 22:53:28 Hurrian: they could get on terrorist list for sharing hacking tools/devices when opening further Mar 13 22:53:57 "anti h4xx0r t00ls law" Mar 13 22:54:00 at least the uk example here is woooshy Mar 13 22:54:27 sopa/acta got nothing on it Mar 13 22:54:30 hacking tools law is pretty stupid Mar 13 22:54:31 * SpeedEvil finds the 'phone hacking' scandal has made searches _much_ harder. Mar 13 22:54:49 by itself Mar 13 22:55:14 In short - if you make available devices which can change the IMEI, or do certain other things - and they are abused, and you do not fix it, any later sales are prosecutable. Mar 13 22:55:40 * SpeedEvil tries to remember the bill. Mar 13 22:56:12 i think it's illegal to fake imei here too Mar 13 22:56:29 opening any related stuff could be used as: they allow terrorists Mar 13 22:56:36 Aha! Mar 13 22:56:55 sopin: terrorists already have whatever they want Mar 13 22:57:05 :) Mar 13 22:57:07 but not open source Mar 13 22:57:17 hmm - not finding. Mar 13 22:57:18 faking imei is illegal almost everywhere in the world Mar 13 22:57:22 SpeedEvil: a russian hacker will send me some modules with FULL %EM and other nice stuff, fully documented. Like AT§CSIM Mar 13 22:57:26 revenging gsm is a HUGE task Mar 13 22:57:43 AT§CHGIMEI Mar 13 22:57:50 hahaha Mar 13 22:57:53 nice Mar 13 22:58:01 he is fbi Mar 13 22:58:06 don't pay him Mar 13 22:58:19 share your results :P Mar 13 22:58:30 Argh - legislation.gov.uk search just went down. Mar 13 22:58:39 who dossed that Mar 13 22:58:41 go 4chan Mar 13 22:59:02 uk is lawless now :) Mar 13 22:59:04 sopin, 4chan hasnt done any raids for YEARS Mar 13 22:59:21 was a joke, nvm Mar 13 22:59:43 prolly some high and mighty "we gotta avenge lulsec" wannabe HOIC-touters Mar 13 22:59:50 coincidence with the subject of discussion Mar 13 23:00:05 fire LOIC Mar 13 23:00:21 the lulsec fbi snitch was a huge news though Mar 13 23:00:42 hope his family will be ok Mar 13 23:00:58 what you mean HIS Mar 13 23:01:14 wife and 2 children iirc? Mar 13 23:01:29 Hurrian, sopin, if you will get some new kernel working with fremantle or you will play with it let me know Mar 13 23:01:34 now I'm going offline Mar 13 23:01:40 sure thing mate Mar 13 23:01:43 gn Mar 13 23:01:46 * SpeedEvil waves. Mar 13 23:01:47 k, downloading the patches one by one from obs actually Mar 13 23:01:54 bye, pali Mar 13 23:01:58 * DocScrutinizer waves back Mar 13 23:03:04 brb Mar 13 23:06:36 does anybody know what is stopping fremantle booting on N800 (stripping gsm related stuff)? Mar 13 23:07:26 is it ram limitation? Mar 13 23:07:45 screen, cpu etc should be all upstream Mar 13 23:07:49 lack of a working gfx accelerator, ram, a lot of assumptions about the kernel ... Mar 13 23:08:15 non-accelerated version could work then? Mar 13 23:08:32 no such thing Mar 13 23:08:37 looking at diablo state today one has to wonder if N900 is not going to end up in the same way Mar 13 23:09:00 unless kernel/libc/libglib get updated Mar 13 23:09:26 and non-compatible things replaced with meego pieces Mar 13 23:09:43 sad view Mar 13 23:09:46 meh, no thanks Mar 13 23:10:02 Why is that? Mar 13 23:10:18 Not really a competitor on the horizon Mar 13 23:10:27 szoppin: Cordia HD is always a possible upgrade path. Hildon UI on Mer. Mar 13 23:11:05 sounds good, but what's with proper power management in meego/mer? Mar 13 23:11:16 wmarone: updating seems like a more viable option, the things you mention are far from working solutions Mar 13 23:11:47 updating isn't an option due to all the closed bits Mar 13 23:11:57 they fell for the age old nasty misconception/fault of "let's deal with that later, for now let's just make it work" Mar 13 23:12:16 critical components are usually backward compatible to an extent Mar 13 23:12:21 you can not do power conserving design 'later' Mar 13 23:12:57 so small jumps in kernel should be possible, if maemo critical component is too hackish, get meego Mar 13 23:13:11 it works as ubuntu 12.04 to an extent on .35 Mar 13 23:13:12 forget meego kernel Mar 13 23:13:24 also forget meego middleware Mar 13 23:13:25 how then? Mar 13 23:13:37 DocScrutinizer: this way to doing things will always fail Mar 13 23:13:38 actually Mar 13 23:13:42 meego already boots on the n8x0 Mar 13 23:13:49 fremantle will NEVER work on the n8x0, so forget. Mar 13 23:14:36 I'm thinking about N900, precise runs so latest libc etc is possible (buggy I admit, but worth trying) Mar 13 23:14:41 I think OTOH that newer kernel, newer glibc, newer glib are certainly possible Mar 13 23:15:22 would grant N900 few extra years of meaningful existance Mar 13 23:15:23 well, with porting the Nokia patches, yes Mar 13 23:15:50 not many nokia patches to glibc or glib ;P Mar 13 23:15:56 kernel Mar 13 23:16:25 there's a whole truckload of .ko that never made it upstream Mar 13 23:16:40 and now their missing spoils meego standby time Mar 13 23:17:15 not even mentioning things like ICD2 Mar 13 23:17:20 (I think that you are making this problem bigger than it is) Mar 13 23:17:26 (standby time) Mar 13 23:17:27 but icd2 is maemo middleware Mar 13 23:17:37 sure, I know Mar 13 23:17:49 I mentioned middleware above Mar 13 23:18:34 anyway... n8 Mar 13 23:20:45 ? Mar 13 23:21:19 or was it n8(ight)? Mar 13 23:21:24 gn just in case Mar 13 23:22:10 if not, are the truckload of .ko you mentioned somewhere available? Mar 13 23:23:46 javispedro: would you have a plan that could possibly work (updating crucial libs/kernel)? Mar 13 23:24:22 I can volunteer to try out possible solutions, bruteforcing till we get something working Mar 13 23:26:17 going to try maverick's .so's next Mar 13 23:27:07 not sure if pre50 will be good enough for them, but unlike precise they lack symlinks, so all need to be renamed to actual .so's from maemo Mar 13 23:27:17 plan? :) Mar 13 23:27:21 hope maverick is not too big of a step Mar 13 23:27:38 javispedro: things to try out that could maybe hint at workable solution Mar 13 23:27:56 my attempts are like bringing a battleaxe to a barber Mar 13 23:28:37 pointless, harmfull, but if it would work... Mar 13 23:28:47 no harm in few flashes Mar 13 23:29:01 maybe error messages will change and give clues Mar 13 23:29:24 moving in the dark totally Mar 13 23:29:38 but maybe smaller jump will yield some result Mar 13 23:30:32 would love to try .29 kernel Mar 13 23:31:01 would be willing to bet 1$ fremantle closed source packages would still function Mar 13 23:31:35 and the ones that crash could give insight into what needs patching to try .30... rinse and repeat Mar 13 23:32:20 fortunately the board files for the N900 are upstream :) Mar 13 23:34:19 so we're looking at closed parts only non-functioning Mar 13 23:34:56 iirc there's something in the boot process that depends on something that was removed from the kernel after 2.6.28 but the last time I looked was over a year ago Mar 13 23:35:04 which for gsm is a big deal, but assuming (we all dream) they will keep working 2.6.3x would maybe be possible? Mar 13 23:35:20 proc/bootreason Mar 13 23:36:17 pali got patch to bring that back Mar 13 23:37:32 getting .35 or such to the level of kp49/50 will take a lot, but still newer libc/glib would be awesome Mar 13 23:38:34 btw, what happened to the guys that brought us fremantle? Mar 13 23:39:05 nokia fired most maemo devs, would they be willing to share some of their exp? Mar 13 23:39:22 As far as I know there are no closed source kernel pieces in the N900 so it should be possible to forward port any N900-specific patches to a newer kernel Mar 13 23:39:29 or is there NDAA in place even if we hunted them? Mar 13 23:40:26 considering licensing kernel should be open, yes, but hacks specific to .28 could be in place in all closed components Mar 13 23:40:49 so even move to 29 could turn out to be fatal Mar 13 23:41:06 hoping it is not the case, but until tested, it is unknown Mar 13 23:42:04 which is why I installed sb yesterday and going to read into kernel compiling/installing, want to try this Mar 13 23:42:16 but will take ages Mar 13 23:42:55 if someone can provide 29 for test, pls share Mar 13 23:48:48 pali's patch for meego kernel is 160kb :D need a few months vacation Mar 13 23:49:45 why not look at the patch set for Mer? Mar 13 23:50:00 https://build.pub.meego.com/package/show?package=kernel-adaptation-n900&project=CE%3AAdaptation%3AN900 Mar 13 23:51:03 pali's patch should allow newer kernels running with fremantle, so you get phone etc functionality and can look at updating libs (my limited understanding), does it all work on mer? Mar 13 23:51:23 nope, mer is a fundamentally different platform Mar 13 23:51:25 but Mar 13 23:51:54 the core patches to make it work on newer kernels are there individually Mar 13 23:51:56 not sure if that's what you need Mar 13 23:52:35 no idea yet, will know in few weeks/months (maybe) Mar 13 23:53:05 looking for simplest way of allowing newer libc etc requiring apps to compile run on N900 though Mar 13 23:53:22 without losing functionality as much as possible Mar 13 23:53:56 nemo effort I applaud, but is a bit different way of looking at the platform/device Mar 13 23:54:39 I am looking for a way to keep fremantle while also allowing newer stuff to run on it Mar 13 23:55:06 possibly totally unrealistic Mar 13 23:57:45 hmmm... wait Mar 13 23:58:12 compile new stuff with LD_LIBRARY_PATH=/opt/newlibglib/ Mar 13 23:58:21 aapo was thinking of porting gcc 4.6 and using this to update all other components, but autobuilder blocks this approach Mar 13 23:58:41 is the maemo repo code open? Mar 13 23:58:52 could someone clone it and remove this limit? Mar 13 23:59:05 we could possibly then move step by step?? Mar 13 23:59:06 hint hint: we break the working-well power management code with newer kernels and different bme Mar 13 23:59:16 so what we do, is step by step Mar 13 23:59:36 pls elaborate? Mar 13 23:59:44 recompiling whatever we can to newst libglib, and library patch is at /opt/newlibglib Mar 14 00:00:08 ok... Mar 14 00:00:16 once we confirm that all stuff works with new libglib, we can patch libglib to expose old interfaces to old, closed programs Mar 14 00:00:16 would this pproach work? Mar 14 00:00:35 then once that works, totally replace old libglib with new one Mar 14 00:00:42 we could use marmistrz repo for all things maemo repo blocks Mar 14 00:01:03 not sure if autobuilder can be added to it for ease of use Mar 14 00:01:51 but would make this almost transparent for participants Mar 14 00:02:04 just change sources list Mar 14 00:02:37 actually, i'd use linaro's gcc for this Mar 14 00:02:59 we need to get aapo, his experience and fight with maemo autobuilder would come in handy Mar 14 00:03:19 he is the guy behind idea update this shit, just how Mar 14 00:04:40 we would still probably face some closed thingie complaining but ldlibrarypath to older version could possibly alleviate that Mar 14 00:05:11 then again libc6/glib are quite connected... worth a try though definitely Mar 14 00:06:13 Hurrian: could you think of a plan or something like it, steps to check if assertion works Mar 14 00:06:38 few tests to tell if this doesn't fail by unexpected reasons Mar 14 00:06:56 I would be more than happy to carry those out and report Mar 14 00:08:03 if first results come back positive, this definitely deserves an epic thread in TMO Mar 14 00:08:27 and hope people will join to test all case scenarios Mar 14 00:10:54 having new libs for new stuff while keeping old versions for the ncompatible... sounds like few mb of opt to sacrifice for having up to date device Mar 14 00:11:13 would be revolution Mar 14 00:12:59 one test I can come up is... get a newer libc requiring app. use latest precise or so to ldlibrarypath... oh wait Mar 14 00:13:17 you'd need the full package with it Mar 14 00:13:37 c6/glib and will cry kernel too old, no? Mar 14 00:14:36 kernel still seems like 'required' part Mar 14 00:15:34 getting current libs to work on newer one and this approach should be good to go Mar 14 00:23:46 gn Mar 14 00:45:49 szop|zzz: patched the 2.6.37 kernel with pali's fremantle support patch Mar 14 00:45:59 building now. **** BEGIN LOGGING AT Wed Mar 14 00:54:05 2012 Mar 14 01:48:32 we have a problem here Mar 14 01:48:45 building 2.6.37 produces 100+ MB of modules Mar 14 01:48:46 ugh **** ENDING LOGGING AT Wed Mar 14 02:59:58 2012