**** BEGIN LOGGING AT Tue Jan 08 02:59:58 2013 Jan 08 03:19:32 is there no documentation at all for ISI? Jan 08 03:22:06 no documentation anywhere for the stuff I am interested in Jan 08 03:22:08 no Jan 08 03:31:33 jonwil: I've said in the past that the whole infrastructure was designed to be as confusing and hard to piecemeal replace as possible. :'( Jan 08 03:33:07 I think when it came to MeeGo-on-N900 they said "We want to do the bare minimum required to support telephony in MeeGo-on-N900 and we want to give away as little of the modem secrets as we can" Jan 08 03:35:45 who benefits from ISI being kept secret still? did Nokia specify it or the modem from elsewhere? Jan 08 03:39:25 ISI is all but secret, it's just a PITA to study the docs formerly to be found on Wirelss Modem API Jan 08 03:39:44 http://www.wirelessmodemapi.com/ Jan 08 03:40:58 another of those orphaned servers Nokia pays for but forgot about them and now they're bitrotting and in a zombie state Jan 08 03:43:04 google for "Wireless Modem API G2 V2 11w05.zip" Jan 08 03:48:29 well, googling seems doesn't help much, you might want to try your luck in maemo.cloud-7.de/hidden/ Jan 08 04:00:11 Does anyone in here still use a Nokia N900? I'm wondering if Easy Debian can use the DSP, camera and microphone or not. Jan 08 04:11:36 thanks for the link DocScrutinizer Jan 08 04:11:47 yw Jan 08 04:13:10 don't post the full URL anywhere, when spiders come crawling I'll remove that file Jan 08 04:13:40 yes, I realised since you didn't post the full link :) Jan 08 04:15:35 I actually HAVE "Wireless Modem API G2 V2 11w05.zip" and it doesn't contain the information I need (the information on the SIM access) Jan 08 04:16:11 in fact, heck, I have copies of several versions of those docs and none contain the info I seek :( Jan 08 04:17:07 yeah, I think it's not comprehensive/complete Jan 08 04:17:29 maybe even ISI is incomplete Jan 08 04:17:46 (the implementation inside BB5) Jan 08 04:18:29 All that is public about the SIM access is a file pn_sim_isi.h from the cellmo-headers package (with NO documentation anywhere) plus the few bits of ofono code that deal with the SIM access Jan 08 04:19:05 I wonder WTF ofono guys got the details from Jan 08 04:19:40 well the checkins to ofono that related to the bit I care about are from @nokia people Jan 08 04:20:06 #@!!%&ðæ€€ Jan 08 04:20:56 so the Nokia guys had access to the modem specs and stuff internally (including presumably the source to the Fremantle Cellular Services Daemon) and got to use that info and code to produce the ofono N900 modem support Jan 08 04:21:07 presumably with the goal being to reveal as little info as possible Jan 08 04:21:22 ...as usual Jan 08 04:21:24 btw the only reference to pn_sim_isi.h in Google is in a file listing of a Nokia QT SDK that contains that file Jan 08 04:24:52 so unless some info appears out of the woodwork somewhere or someone at Nokia is willing to share, the chances that I will ever truly understand connui_cell_sim_get_service_provider/connui_cell_sim_is_network_in_service_provider_info or the underlying dbus and ISI calls that those functions end up calling to is pretty much zilch Jan 08 04:25:52 :-S Jan 08 04:26:52 lost info Jan 08 04:27:22 those who could share are not at Nokia anymore Jan 08 04:27:40 pretty certainly Jan 08 04:36:16 yeah the work was done by "Jukka Saunam�ki" whoever they are Jan 08 04:37:17 and that person hasn't been active on gitorious since last january Jan 08 04:37:49 ok, now I found a linkedin page that confirms they arent at Nokia anymore Jan 08 04:39:30 so yeah unless hell freezes over and someone decides to release source for cellular services daemon (which will never happen) we will never really understand the workings of the modem interface Jan 08 04:39:40 and will never be able to replace the modem in cssu with ofono Jan 08 04:41:01 hah, the latter is a sure bet Jan 08 04:41:19 no matter what happens Jan 08 04:41:35 Quite a few people have suggested " Jan 08 04:41:45 suggested "lets replace CSD with ofono for Fremantle" Jan 08 04:41:56 Those people who have suggested that clearly know nothing about ofono Jan 08 04:41:59 or about Fremantle Jan 08 04:42:03 yep Jan 08 04:42:11 was about to type exactly this Jan 08 04:49:30 I suspect that the internet connectivity daemon is the most likely piece of closed-source system level software to be replaced by CSSU Jan 08 04:51:58 and yes I still have plans to continue my reverse engineering work on icd :) Jan 08 04:55:06 unless there are other system components that are more in need of replacement than ICD in which case I will consider investigating those... :P Jan 08 04:57:07 stskeeps was sad that ICD never gone OSS, since he considered it way better than all the OSS alternatives. So you'd probably do something really good and heroic when you RE ICD Jan 08 04:57:45 well I wouldn't be cloning ICD, just identifying what interfaces matter when it comes to replacing it Jan 08 04:57:49 in the end your RE work could go into nemo and sailfish Jan 08 05:02:06 so you are saying that working on ICD would be more use than working on other things? Jan 08 05:02:19 quite probably Jan 08 05:05:56 ok Jan 08 05:06:16 well I will continue reverse engineering the dbus, gconf and other interfaces exposed by ICD and its related plugins then Jan 08 05:08:11 a proper howto on the configuration of that annoying blow was already somewhing I'd appreciate Jan 08 05:08:18 blob Jan 08 05:10:28 * DocScrutinizer05 idly wonders what's on MWKN Jan 08 05:38:25 If hell froze over and I could have the source code to any 10 closed N900 packages it would be connui-home-cellular, libconnui, libconnui-cellular, libgles2-sgx-img, libicd-network-wlan, libisi1, mce, opengles-sgx-img-common, osso-wlan-security and icd2 Jan 08 05:39:49 how do you make it auto switch back to a real connection after switching to the MMS apn? Jan 08 05:40:48 jonwil: I think a petition to open source it could be successful Jan 08 05:41:02 We could put some stuff about 'good PR' in there. Jan 08 05:41:07 I'll write it if you'd like Jan 08 05:42:17 which thing? Jan 08 05:43:10 * jonwil suspects internetishard doesn't understand the realities of the situation :P Jan 08 05:44:22 Indeed. And my lack of understanding could have a good outcome :) Jan 08 05:44:53 People who accomplish great things have to be ignorant enough to not realize they're not supposed to :P Jan 08 05:45:14 I seriously doubt Nokia will open source anything further no matter what the community does or petitions Jan 08 05:45:30 Especially now that they are so in bed with Micro$osft :P Jan 08 05:45:40 Micro$oft :P Jan 08 05:46:09 Yeah, but how would they re-use any of that in their M$ efforts? Jan 08 05:46:14 Also, they need to lay people off probably Jan 08 05:47:22 point is that open sourcing stuff costs money for no benefit to Nokia Jan 08 05:47:25 hence why they wont do it Jan 08 05:47:36 but if you want to send Nokia a useless petition or request feel free to do it... Jan 08 05:47:49 internetishard, they rather destroy stuff than open sourcing it. see Meltemi Jan 08 06:08:36 if we want to replace ICD we have to do it ourselves :) Jan 08 06:14:18 not to mention, they were probably shredding stuff since Elopocalypse Jan 08 06:43:00 Anyone interested in the Ubuntu Phone? They are planning to release in Feb for Galaxy Nexus owners. Jan 08 06:43:40 to a degree Jan 08 06:43:56 more interested in some ways in jollamobile Jan 08 06:44:02 yep!! Jan 08 06:44:05 jollamobile Jan 08 11:00:49 http://www.theregister.co.uk/2013/01/08/elop_anything_goes/ Jan 08 11:01:42 hnngh Jan 08 11:01:46 ruskie: it was lost in translation Jan 08 11:01:49 nokia issued an update: Jan 08 11:01:52 “Do you rule out 100% launching a smartphone based on Android in 2013? Jan 08 11:01:54 Elop: So, the way I think about it is, in the current war on ecosystems, we are fighting with Windows Phone. That’s what we’re doing. Now, what we’re always doing is asking, how does that evolve? What’s next? What role does HTML5 play? What role does Android or other things play in the future? We’re looking further into the future, but it terms of what we’re bringing to... Jan 08 11:01:55 ...market, and what... Jan 08 11:01:57 ...we’re immediately focused on, we’re focused on Windows Phone.” Jan 08 11:03:36 thedead1440, that still sounds the same frankly Jan 08 11:03:44 nothing really different Jan 08 11:04:00 "we're with ms for now bet we're not excluding anything in the future" Jan 08 11:09:10 ruskie: actually he said the same thing around a year ago Jan 08 11:09:16 nobody made a fuss of it then Jan 08 11:09:29 as CEO he *cannot* say I'm excluding anything Jan 08 11:09:42 so he just says we are always open to what's in the market-place Jan 08 11:35:52 some more dis-heartening news for Nokia: http://timesofindia.indiatimes.com/tech/tech-news/telecom/NokiaIndiafactory-raided-over-Rs-3K-cr-tax-evasion/articleshow/17938056.cms around $0.5billion in tax evasion Jan 08 12:08:47 that's called "going down with style" :) Jan 08 14:07:41 As easy as the fans applaud succes, just as easy do they hate when things dont go as well Jan 08 16:20:18 Since northing has been happening for more than an hour and since I've noticed a certain... distaste for Android here, let me ask you this: Are there parts of it that are, in fact, hidden? As in that it's not as open as Google might want to claim? Jan 08 16:20:36 Because I remember reading about that on Register, but I can't find it... Jan 08 16:31:22 hidden - no Jan 08 16:31:25 well. Jan 08 16:31:37 you can build android from source. Jan 08 16:31:51 however, you don't get the Google apps Jan 08 16:31:58 those are binary only Jan 08 16:32:05 Yeah, so I hear. Jan 08 16:32:09 Well, that's better than I thought. Jan 08 16:32:22 Thank you for clarifying that to me :) Jan 08 16:32:23 there aremseperate issues around hardware support Jan 08 16:32:35 binary blobs are regrettably common Jan 08 18:20:30 * MrPingu has come to say hello Jan 08 18:25:37 * Apic greets you. Jan 08 18:35:01 hm, i have connected a usb keyboard to my n900 Jan 08 18:35:22 i have noticed that keys such as Home, End, etc dont work Jan 08 18:35:38 is there a solution for that somewhere? Jan 08 18:36:43 Close the N900 keyboard ;) Jan 08 18:37:17 And select the right keyboard from the ext-kdb menu in settings, if installed ;) Jan 08 18:38:10 ill try that thanks Jan 08 18:39:47 How's migration going, btw? Jan 08 18:39:47 always a delight to hear how natural users use h-e-n Jan 08 18:40:49 MrPingu: well, it's hard work for council to decide on the right way to go. For now most things are already migrated to a set of (possibly interim) vservers Jan 08 18:41:42 I see :) Jan 08 18:41:59 see http://mwkn.net/2013/01/community.html Jan 08 18:43:04 * MrPingu is reading mwkn right away Jan 08 18:43:13 first headline "maemo.org migration issues" Jan 08 18:43:50 I provided that etc/hosts file if you want to use the currently alive servers Jan 08 18:44:21 MrPingu nice, both mouse and keyboard are working great Jan 08 18:44:21 MrPingu: i don't think you really need to close the n900 keyboard Jan 08 18:44:22 we're about to eventually migrate tmo Jan 08 18:44:24 thanks for the tip Jan 08 18:44:29 at least, not with a bluetooth keyboard Jan 08 18:44:36 trx: extkbd and extmou are kinda great Jan 08 18:44:49 yeah Jan 08 18:45:04 extkbd will overwrite your xkb data though Jan 08 18:45:16 so you could lose your own personalizations Jan 08 18:45:24 kerio: I had to, to get it working right ;) Jan 08 18:45:59 Atleast when using it in conjuction with dosbox and H-E-N Jan 08 18:46:24 BTkbd might be quite different to USB-kbd Jan 08 18:46:51 Don't have BTkdb lying around, sadly Jan 08 18:49:14 I got a MX5000 logitech BT kbd and BT mouse here, but I don't think it's particularly easy to make them work with *any* ordinary BT under linux Jan 08 18:50:00 DocScrutinizer05: if they're HID, then it should automagically Just Work™ Jan 08 18:50:30 if they are ordinary HId at all, then crippled Jan 08 18:50:47 huh? Jan 08 18:50:52 About that, I am planning to buy Raspberry together with BT kbd + mouse + BT-adapter Jan 08 18:51:15 is there something I need to watch out for specific chipsets? Jan 08 18:51:42 To get BT working nicely? Jan 08 18:51:58 Now we are talking about keyboards Jan 08 18:52:11 qt does listen to xkb, partially Jan 08 18:52:21 BT keyboards are kinda tricky always Jan 08 18:52:49 plug and pray Jan 08 18:53:00 lol Jan 08 18:53:22 It listens to my dead-characters I mapped on fourth level Jan 08 18:53:49 plug? Jan 08 18:53:56 you're doing bluetooth *wrong* Jan 08 18:54:32 ignores the rest though :( Jan 08 18:54:34 * DocScrutinizer05 glares at his BT USB dobgle and frowns Jan 08 18:54:46 dongle* Jan 08 18:55:16 * merlin1991 glares at his integrated bt chip and cries for proper drivers Jan 08 18:55:17 You said keyboard, but I already guessed you meant dongle Jan 08 18:55:22 * DocScrutinizer05 glares at Logitech USB "BT" dongle and frowns even more Jan 08 18:55:40 Logitech has it own special magic Jan 08 18:56:01 also I'd like a proper keyboard with cherry mx red switches that is not in the 150€ and up pricerange Jan 08 18:56:32 reed, but yeah Jan 08 18:59:26 reed? Jan 08 19:01:01 merlin1991: the singer from The Velvet Underground Jan 08 19:01:22 The Nascom-II kbd I had in early 80s was awesome. It used small ferrite cores (like you maybe know them from ancient core memory) with two wire frames as half winding of a coil. each key had a magnet that moved away from its core when key got pressed, so removing magnetic saturation from the ferrite core/ring. The unsaturated ring transferred a sharp current spike in col wire frames to the row wireframe where it got detected by an amp Jan 08 19:01:23 and thus sensed the row and col for a pressed key Jan 08 19:01:47 merlin1991: ooh I thought you meant reed-switches Jan 08 19:02:02 DocScrutinizer05: how would it detect multiple keys? Jan 08 19:02:17 like any other kbd Jan 08 19:02:21 badly? Jan 08 19:02:30 optimal Jan 08 19:03:00 the operation priciple has implicit "n-key rollover" Jan 08 19:03:25 I need 5 key rollover at max :D Jan 08 19:03:27 in that it doesn't feed back from row_a via col to row_b Jan 08 19:11:11 They make awesome keyboard these days Jan 08 19:11:19 with proper resistance and CLICK action Jan 08 19:11:30 they come with earplugs for when you get tired of the clackity-clack Jan 08 19:15:05 (kbd) tried to spot the schematics, but couldn't find in which of the documents it hides, so if you're interested go and have a look at http://www.nascomhomepage.com/ Jan 08 19:16:25 here's the kbd from outside: http://www.old-computers.com/museum/computer.asp?c=815&st=1 Jan 08 19:17:07 (been fun to assemble that PCB shown there next to kbd :-D) Jan 08 19:18:01 ooh, I just read it been late 70s, not early 80s Jan 08 19:21:17 luckily mine came without MS Basic ROM, so I had to use assembler (in hex) to do any programming. That way I learnt a bit about computers Jan 08 19:24:23 hey, QBASIC was awesome Jan 08 19:25:29 gwbasic was the shit when i was younger Jan 08 19:25:40 computer class was so awesome in the early 90s, zero security Jan 08 19:27:32 well getting things done was #1 prority, nothing about web security and friends Jan 08 19:28:34 http://www.nascomhomepage.com/pdf/NASCOM1.PDF p.5-6 is that kbd Jan 08 20:26:14 ivgalvez back and well! :-) \o/ Jan 08 20:26:30 "still a bit dizzy" Jan 08 20:26:59 hello guys, do you have any idea how to prevent phone application on N900 to mess with cpufreq settings? Jan 08 20:27:37 you don't want to Jan 08 20:28:10 phone application sets CPU freq to some value that guarantees proper audio for your calls Jan 08 20:28:19 and only during calls Jan 08 20:28:59 but it sets it to 125 to whatever was there Jan 08 20:29:18 for sure not, 125 is forbidden freq Jan 08 20:29:42 for good reason, it's instable Jan 08 20:29:57 WizardNumberNext: it sets minimum to 125, however 125 is avoided and still 250 will be minimum Jan 08 20:30:47 I do not want anything lower then 500, as I am running on conservative instead of ondemand and down_threshold is 70, so any frequency below 500 is pointless, as 70% of 500 is higher load on 250MHz then 95% (which up threshold) Jan 08 20:31:02 and there is no avoid frequencies on conservative Jan 08 20:31:27 :shrug: Jan 08 20:31:38 never liked ondemand as it doesn't go up step by step - it is going directly to top sped Jan 08 20:32:05 that's called speed-to-idle and a sane policy Jan 08 20:32:13 usually Jan 08 20:32:44 WizardNumberNext: so... put 500 as the minimum Jan 08 20:32:46 you mean the parameter for phone aplication? Jan 08 20:32:58 *race-to-idle Jan 08 20:32:59 I do put minimum 500 or 600 Jan 08 20:33:06 WizardNumberNext: then your minimum will be 500 Jan 08 20:33:12 or what kerio said ^^^ Jan 08 20:33:25 phone app is changing it and leaving it on 125 Jan 08 20:33:31 ...no it's not Jan 08 20:33:43 it tries to Jan 08 20:33:46 but 125 isn't allowed Jan 08 20:33:53 make it so that 250 isn't allowed either, if you want Jan 08 20:34:04 you cannot prevent any frequency on conservative! Jan 08 20:34:17 I am not using ondemand - I hate this crap! Jan 08 20:35:00 so what? ask us to make pigs fly? Jan 08 20:35:40 I prefer flying penguins... Jan 08 20:36:01 Just want to know, if you know how to prevent phone aplicantion to change lower frequency, not an instruction how to setup ondemand governor Jan 08 20:36:37 answer: no we don't since that's what phone app is supposed to do Jan 08 20:37:06 as soon as I would get cross compiler working I am putting 3.8 kernel on it without ondemand governor Jan 08 20:37:21 good luck with that Jan 08 20:37:36 WizardNumberNext: yeah, good luck Jan 08 20:37:59 first I would use kexec to boot this kernel Jan 08 20:38:06 WizardNumberNext: wanna link to 4.7.2 for scratchbox? Jan 08 20:38:13 if it was that easy to put 3.8 on it, PK wasn't 2.6 still Jan 08 20:38:15 see how is it behaving, then I would make pernament decission Jan 08 20:38:36 3.8 is still rc2 Jan 08 20:38:52 so obviously not many people are interested Jan 08 20:39:04 yeah sure Jan 08 20:39:07 WizardNumberNext: which SGX driver are you going to use? just out of curiosity. Jan 08 20:39:10 bbl Jan 08 20:39:29 I had a good look on it and it does support most of devices, but I still would like to have et8ek for it Jan 08 20:39:29 DocScrutinizer51: best wishes to ivgalvez Jan 08 20:39:40 will do Jan 08 20:39:45 https://lesswatts.org/projects/applications-power-management/race-to-idle.php Jan 08 20:39:47 ^^ Jan 08 20:39:56 power-kernel doesn't have any SGX driver as far as I can see Jan 08 20:40:10 WizardNumberNext: look again Jan 08 20:41:18 freemangordon, look yourself in both configuration and kernel sources itself, then start to talk about it - quite a lot of devices have drivers Jan 08 20:41:21 *something* is bound to have them Jan 08 20:41:52 WizardNumberNext: get your peers classified before getting offensive Jan 08 20:42:06 WizardNumberNext: I don;t know if you are aware, but I am one of KP developers, so I am pretty aware of what is included there ;) Jan 08 20:42:33 that is why I told you to look again, there is SGX (aka display) driver Jan 08 20:42:42 if you are one them, then could you give me device listing please Jan 08 20:43:06 device listing? try lshal Jan 08 20:43:08 and if something isn't in vanilia kernel, then source of those drivers? Jan 08 20:43:32 WizardNumberNext: what? Jan 08 20:43:35 does he come to irc? Jan 08 20:43:44 KP sources are in maemo repos Jan 08 20:43:54 I mean plain patches? Jan 08 20:44:00 on garage Jan 08 20:44:11 garge.maemo.org Jan 08 20:44:19 search for kernel Jan 08 20:44:34 thats something - I would have good look there Jan 08 20:44:38 an KP is 2.6.28 + lots of stuff in debian/patches Jan 08 20:45:18 so you get clean .patch files if you download tar.gz from maemo repos Jan 08 20:45:53 WizardNumberNext: BTW lshal is not he Jan 08 20:46:09 freemangordon, ups Jan 08 20:46:13 it is a tool to executo on your device to get a fair amount of devices Jan 08 20:46:17 *execute Jan 08 20:48:35 WizardNumberNext: BTW here http://merlin1991.at/~freemangordon/kernel3.5/ you will find a couple of patches which allow vanilla 3.5 to boot soick fremantle Jan 08 20:48:38 *stock Jan 08 20:49:08 wish you luck forward-porting them to 3.8 :) Jan 08 20:49:51 I wasn't aware they already have changed API Jan 08 20:50:29 between 2.6.28 an 3.8? you can bet the did. Jan 08 20:50:35 *they Jan 08 20:50:37 freemangordon, where do I get ishal - I seam to not have it in my repos Jan 08 20:50:47 you have it on your n900 Jan 08 20:50:55 between those - yes few times Jan 08 20:51:04 pre-installed, run it in xterm Jan 08 20:51:07 I was thinking between 3.5 and 3.8 Jan 08 20:51:44 Actually I don't know how much is API changed between 3.5 and 3.8 Jan 08 20:52:00 I just wished you luck ;) Jan 08 20:52:01 me neither Jan 08 20:52:15 I do not seam to have Ishal Jan 08 20:52:19 just tried it Jan 08 20:52:22 lshal Jan 08 20:52:32 that is small latin letter L Jan 08 20:52:56 something like list hardware abstraction layer :D Jan 08 20:53:05 you mean LSHAL? Jan 08 20:53:20 yep , but not in capital letters Jan 08 20:53:44 here are some other patches for 3.5: https://atrey.karlin.mff.cuni.cz/~pali/linux-n900/ Jan 08 20:53:55 untested, but freemangordon patches included Jan 08 20:54:08 Pali, thats usefull Jan 08 20:54:36 Pali, anyway I do nt care, if those are tested or not, because I would run kernel from kexec Jan 08 20:54:42 Pali: aah, you ported BT patches? good boy :) Jan 08 20:54:49 WizardNumberNext, ask more on #maemo-alternatives Jan 08 20:54:53 WizardNumberNext: kexec does not work in 2.6.28 afaik Jan 08 20:55:03 there are more peoples trying to boot new kernel Jan 08 20:55:08 why kexec? :s Jan 08 20:55:13 freemangordon, I got it - got listing, looks like listing of know devices (didn't have good look on it yet) Jan 08 20:55:15 there's uboot Jan 08 20:55:17 Pali: but not with fremantle on top ;) Jan 08 20:56:13 WizardNumberNext, kexec not working on n900 correctly Jan 08 20:56:22 there is problem wih SDHC card reader Jan 08 20:56:49 Pali: afaik I am the only one so far successfuly booted fremantle with 3.5. except if you did that too Jan 08 20:56:56 and maemo 2.6.28 kernel can kexec only new kernels (cannot kexec maemo 2.6.28 kernel...) Jan 08 20:57:00 Pali, could you precise it? Jan 08 20:57:08 freemangordon, I did it in qemu Jan 08 20:57:16 hehe, qemu :P Jan 08 20:57:19 not yet on real HW... Jan 08 20:57:40 Pali, qemu doesn't count Jan 08 20:57:45 well, so I am the man :D:D:D Jan 08 20:57:49 WizardNumberNext: it counts Jan 08 20:58:18 I created fremantle patches for meego kernel when I played with qemu... Jan 08 20:58:31 was it 2.6.37? Jan 08 20:58:32 WizardNumberNext, there are kexec problems: http://lists.meego.com/pipermail/meego-dev/2010-May/002277.html Jan 08 20:58:39 freemangordon, yes Jan 08 20:58:43 I had 3.8-rc2 working on kvm on my amd phenom, but I do not consider it working success, as it didn't boot bare metal Jan 08 20:59:17 WizardNumberNext: btw Pali upstreamed a couple of n900 patches in 3.8 ;) Jan 08 20:59:19 WizardNumberNext: qemu to emulate a n900 Jan 08 21:00:17 wait qemu emulating N900? That would would quite long time to implement such amount of hardfware, not to mention DSP and such stuff Jan 08 21:00:24 Did one of you try to backport the newer SGX drivers to 2.6.28? Jan 08 21:00:42 MrPingu: what for? Jan 08 21:00:45 vsync? Jan 08 21:00:48 MrPingu, where to get new SGX drivers? Jan 08 21:01:01 WizardNumberNext: meego (nemo) Jan 08 21:01:16 WizardNumberNext, not everything is emulated, but it working Jan 08 21:01:16 SGX driver is one of hardest to get for me yet Jan 08 21:01:21 you need maemo MADDE Jan 08 21:01:35 part of maemo MADDE is qemu n900 image with patches qemu Jan 08 21:01:59 Pali: can one attach a debugger to running qemu? Jan 08 21:02:03 yes Jan 08 21:02:06 that's actually a lot more, then I would expect Jan 08 21:02:08 good Jan 08 21:02:10 also to kernel Jan 08 21:02:21 yeah, that's what I was asking Jan 08 21:02:26 From my googling I just did there are two parts of SGX 1: the open source modules and 2: the closed libraries Jan 08 21:02:31 this is only good way how to debug kernel... Jan 08 21:02:40 well, i'll stick to printk's for now :P Jan 08 21:02:47 hehe Jan 08 21:02:49 looking on internet it seams n900 is dead as development hardware - looking here is looks like it is very much alive Jan 08 21:03:03 qemu debugging is needed when printk is not enought... Jan 08 21:03:03 WizardNumberNext: which internet do you look at? Jan 08 21:03:03 internet lies Jan 08 21:03:16 Pali: sure, I was joking Jan 08 21:03:20 where to get them, is another story Jan 08 21:03:27 ~cssu Jan 08 21:03:28 well, cssu is http://wiki.maemo.org/Community_SSU Jan 08 21:03:38 DocScrutinizer51: that one lies too ;) Jan 08 21:03:44 freemangordon, do you know how hard is to force n900 kernel to panic? Jan 08 21:04:00 I already have CSSU - t is nice replacement for dead SSU Jan 08 21:04:00 *(char*) = 0? Jan 08 21:04:00 Pali: cat /dev/kmem Jan 08 21:04:14 I wrote for panic my kernel module :D Jan 08 21:04:15 *(char*)0 = 0? Jan 08 21:04:32 I called kernel panic() function Jan 08 21:04:32 freemangordon, I would not try it even on my desktop/server Jan 08 21:04:41 ^^^ should do the job :D Jan 08 21:05:09 on normal comuper you can crash kernel by echo 'c' to sysrq Jan 08 21:05:10 WizardNumberNext: why not, you get oops, so what? Jan 08 21:05:20 but for unknown reason that not working on n900 Jan 08 21:05:31 kernel wrote that crashed, but nothing happened Jan 08 21:05:35 still working :D Jan 08 21:05:55 hehe, isn;t it better that way? :D Jan 08 21:05:56 Pali: don't you have to enable sysrq? Jan 08 21:06:04 that one too Jan 08 21:06:24 I enabled it Jan 08 21:06:31 Pali: try cat /dev/kmem Jan 08 21:06:38 Pali: aah, I have very good idea - undervol Jan 08 21:06:41 and echo > sysrq working if is disabled too Jan 08 21:06:43 *undervolt Jan 08 21:06:50 :D Jan 08 21:06:58 and I wrote that kernel written message to dmesg that crashed :D Jan 08 21:07:13 freemangordon: don't be a pussy, lock it to 1150 Jan 08 21:07:14 so sysrq called that method Jan 08 21:07:24 freemangordon, undervolt works very well, if you want to crash - I was getting loads of crashes on 38 for 600 Jan 08 21:07:34 kerio: no, it locks then, does not opps ;) Jan 08 21:07:41 heh Jan 08 21:07:42 now I am on 42 and it is rock-stable Jan 08 21:08:02 WizardNumberNext: enable SR and forget about manual undervolting Jan 08 21:08:10 until you visit arctic or sahara Jan 08 21:08:43 SR doesn't work too good for me on frequencies above 720 Jan 08 21:08:55 hehe Jan 08 21:08:59 hmm, yah, you might have problems in sahara. but the problem with n900 not stable will be tha last in the list :D:D:D Jan 08 21:09:00 also enable cssu-thumb and forget about overclocking :D Jan 08 21:09:02 been running 900, 950 and 1000 stable Jan 08 21:09:29 OC is nonsense anyway Jan 08 21:09:42 WizardNumberNext: those are insane frequencies Jan 08 21:09:49 esp 950 and 1000 Jan 08 21:10:06 kerio, CPU won't be faster by itself, because of thumb and still apps have to be compiled with thumb Jan 08 21:10:19 but hey, it might make sense when actually locking CPU clock speed to 1000, without any ondemand ever throttling it down X-P Jan 08 21:10:31 CPU isnt the bottleneck, IO is Jan 08 21:10:37 indeed Jan 08 21:10:42 DocScrutinizer05: yeah, lot of sense :D Jan 08 21:11:10 freemangordon, yes those are insane, so I normaly run up to 805 sometimes 850, but is it pointless, as there won't be any real improvement and battery would go slightly quiecker Jan 08 21:11:24 WizardNumberNext: CPU won;t be faster, it just will execute better optimized code much better fit in caches Jan 08 21:11:51 However, in dosbox 900mhz is alot more playable than 600mhz ;) Jan 08 21:12:11 MrPingu: aah, maybe it is time to thumb-compile dosbox Jan 08 21:12:17 who was the maintainer? Jan 08 21:12:23 lemme check Jan 08 21:12:42 freemangordon, I do realize what thumb actually does - it is nice thing, but still apps have to be comppiled to thumb to get any improvement Jan 08 21:12:45 I believe it was Javier Pedro Jan 08 21:12:53 Yes, confirmed Jan 08 21:13:02 * freemangordon has some cloudy memories that once we argued with him hether thumb-compiled dosbox will be stable Jan 08 21:13:10 WizardNumberNext: meh, the huge improvement is from gcc4.7-linaro rather than the different ABI Jan 08 21:13:33 kerio: both thumb2 and gcc4.7.2-linaro Jan 08 21:13:48 kerio, and that is why I am trying to compile cross-compiler Jan 08 21:13:54 gcc-4.7.2 Jan 08 21:13:58 freemangordon: did you receive latest bb-p sources for thumb? Jan 08 21:14:27 For some time my indicator is flashing that bb-p can be updated, however non-thumb Jan 08 21:14:35 MrPingu: the thumby version matches the non-thumby one for me Jan 08 21:14:36