**** BEGIN LOGGING AT Sun Sep 23 02:59:59 2012 Sep 23 03:50:42 for me neither Sep 23 03:51:20 and what does "not working at all" mean?` Sep 23 03:51:58 working fine for me too Sep 23 03:53:21 I admit occasionally a "killall browserd" can help to free up considerable amounts of RAM, despite browserd gets immedeately restartded by dsme Sep 23 14:12:25 hello!can't mount rootfs.jffs2 under debian."no space left" Sep 23 14:13:19 searched a lot ,didn't find any solution,just complains Sep 23 14:14:04 i hope,that somebody can help me here Sep 23 14:48:29 should maemo have hw accelerated mp3 decoding? Sep 23 14:48:48 Skry: eeeh Sep 23 14:48:54 it definetely could, the hardware is there Sep 23 14:49:04 but i think that the dsp is only used for videos, by default, in mafw Sep 23 14:49:31 I mean, just checking /lib/dsp and there are nothing for audio Sep 23 14:50:04 gst-dsp has at least aac :) Sep 23 14:50:43 not that it matters, just asking for curiosity Sep 23 14:55:43 Skry: a couple of dudes tried to figure that out Sep 23 14:56:20 it turns out that decoding mp3s with the dsp would be about the same power draw as decoding them with the cpu Sep 23 14:57:22 yeah, I suspected the same Sep 23 14:57:36 note that it should still be done, imo Sep 23 14:57:59 so you can use the cpu for something else if you want to Sep 23 14:59:25 agree with that one. hmm, gst-dsp does aac and gstreamer-ti does both aac and mp3. Sep 23 15:00:22 oh god do i have to download yet another TI sdk :S Sep 23 15:01:27 hahahaha Sep 23 15:01:39 Skry: freemangordon probably knows about this Sep 23 15:05:48 yeah he does. i already have latest gst-dsp and tidsp-binaries working on other os without any fuzz, was just wondering if mpeg1 audio decoding existed in maemo. Sep 23 15:07:11 for instance, in my case i really, really want either the mp3 decoding or the a2dp encoding done by the dsp Sep 23 15:09:41 didn't a2dp have the option for supporting mpeg4 audio? can't remember but if so, then it can be done already Sep 23 15:10:27 if gst-dsp has aac encoding Sep 23 15:10:55 also, i wonder Sep 23 15:10:59 can i preencode my songs? Sep 23 15:11:18 it would be *neat* Sep 23 15:11:37 nah, decoding only :\ Sep 23 15:11:53 make it happen, #maemo Sep 23 15:12:10 i want cpuless songs-to-btheadset Sep 23 15:18:21 guess i'll take a look at that gstreamer-ti if it has some uber secret mpeg1 audio decoding bits Sep 23 15:18:46 hopefully it is not such an abomination as their graphics sdk Sep 23 15:19:06 first it makes you laugh, then it makes you cry Sep 23 15:44:48 Pali, Hi, i'm flashing my n900 with libusb-sniff, i flash only mmc, only root, root+mmc.... you need another log ? Sep 23 15:45:37 rlinfati, now I implemented initial flashing support, but logs are still usefull for me :-) Sep 23 15:46:33 mmc flashing is done via mk II protocol (not nolo) and I did not looked at it yet Sep 23 15:48:02 freemangordon: ping Sep 23 16:06:01 how risky is it to switch to the linux kernel for power users? Sep 23 16:06:14 i just want to be able to run mobile hotspot Sep 23 16:06:42 wasn't a risk for me. risk comes with what you do with it afterwards Sep 23 16:07:42 yeah i dont plan to do anything with it that im aware of Sep 23 16:07:52 i just have a need for mobile hotspot Sep 23 16:10:56 then you'll be ok Sep 23 16:32:32 wow.. took five attempts before it booted Sep 23 16:32:40 seems fine though Sep 23 16:34:13 i guess i should avoid looking into all the things i could do now... Sep 23 16:34:25 hehe Sep 23 16:34:51 though you may want to turn SmartReflex on Sep 23 16:35:22 vi__: pong Sep 23 16:36:45 freemangordon: what is that Sep 23 16:38:21 ludens: http://wiki.maemo.org/Smartreflex Sep 23 16:43:52 I'd feel reluctant to suggest SmartReflex to anybody, without even any proper evaluation of the expected benefit, compared to the very real concerns about Nokia telling it's not stable on hw level Sep 23 16:44:35 DocScrutinizer05: there was an evaluation (burried somewhere on TMO) Sep 23 16:45:12 Nokia has never said it is unstable at HW level afaik, if you have some reference, please share it Sep 23 16:47:23 they pulled it in the kernel Sep 23 16:47:43 also, there was a silicon errata on the part Sep 23 16:47:50 well, "Nokia" said exactly that, back in times where titan (iirc) found the SR sysnodes, and we asked here why they aren't used. Nokia even eliminated them in later kernels (again iirc, with PR1.1 kernel) Sep 23 16:48:05 SpeedEvil: for SmartReflex? Sep 23 16:48:14 do yo have some link Sep 23 16:48:18 *you Sep 23 16:49:11 check last four weeks chanlog, I posted something with comment "in my book that reads like 'don't use SR', or at least it could" Sep 23 16:50:34 and there might also be circuit errata, not everything got screwed by TI Sep 23 16:50:37 ;-) Sep 23 16:50:38 DocScrutinizer05: come on, I explained a couple of times why SR in stock kernel is beoken. Sep 23 16:50:51 *broken Sep 23 16:51:36 you explained why it's definitely broken. You *didn't* explain that's the root and only cause it got deprecated by Nokia Sep 23 16:52:17 maybe it's other way round? Nokia simply didn't care about those efuse settings or whatever it was, since they already knew it been fsckd up on several other levels Sep 23 16:52:51 Didn't I? fcked efuse calibrations are not the root cause SR to be unusable if using default algotithm? Sep 23 16:53:14 ANd you are well aware Nokia never cared that much for n900 Sep 23 16:53:19 it's not even been investigated yet if every power domain is connected to the LDO it is meant to be Sep 23 16:53:36 USB host mode is a good example of that Sep 23 16:53:50 you're spinning away Sep 23 16:53:54 we've been told every now and then it is impossible Sep 23 16:54:01 no, whY? Sep 23 16:54:18 thumb2 is another good example Sep 23 16:54:45 you don't get it, eh? Sep 23 16:55:07 DocScrutinizer05: I refuse to get anything in IIRC basis Sep 23 16:55:12 *on Sep 23 16:55:31 for thumb we have a pretty clear SiErr report, and there's little reason to assume there's any other good rationale why thumb wasn't usable (except oooops compiler...) Sep 23 16:55:52 DocScrutinizer05: remember my conversations with mra? Sep 23 16:56:05 "conversations" even Sep 23 16:56:22 for SR you found *some* "broken" part, that definitely isn't proven to be the root cause Sep 23 16:56:59 DocScrutinizer05: and you fail to provide any evidence something is broken besides "IIRC Nokia said" Sep 23 16:57:27 if we all remember correctly Sep 23 16:57:32 and for me that's enough Sep 23 16:57:40 I'm not accused here Sep 23 16:58:03 but not for me, I need something written by either TI or Nokia Sep 23 16:58:21 so go search for it, maybe you find it in 2008 chanlogs Sep 23 16:58:41 DocScrutinizer51: noo, it is you who insist "there must be something else" no me Sep 23 16:58:46 *not Sep 23 16:58:57 so, go search for it Sep 23 16:59:03 and I already told you I posted URL and quote of some SiErr I found some few weeks ago Sep 23 16:59:36 and I look at them, we're not hit AIUI Sep 23 16:59:43 looked even Sep 23 16:59:53 the difference is I *know* there's something else, particularly that satement from "those who should know" Sep 23 17:00:27 DocScrutinizer05: come on, those who know told us USB HM is impossible, thumb2 is impossible, etc Sep 23 17:00:52 I don;t need any other evidence that tyhose who know in practice just don't care Sep 23 17:00:53 and honestly I couldn't bother less to defend my memory against your silly accusations Sep 23 17:01:03 silly? Sep 23 17:01:05 ok Sep 23 17:01:27 yes, silly. If "iirc" is enough for you to start arguing and accusing me Sep 23 17:01:48 DocScrutinizer05: I m not accusing you, sorry if I sound offensive Sep 23 17:01:51 go bitch at somebody else Sep 23 17:02:13 sure Sep 23 17:03:50 while I can't see the monster benefit of hw driven LDO config over sw controlled LDO config which is working since beginning of N900 Sep 23 17:04:23 probably something like 300 machine cycles "wasted" on every C-state change Sep 23 17:04:30 then you have NFC what SR is Sep 23 17:04:41 uhu Sep 23 17:06:18 SW programs LDO with a particular voltage which is guaranteed to be stable for every device/temperature/load. SR lowers that voltage to one that is stable for the current device/temperature/load Sep 23 17:07:25 this voltage is usually 20-30 % lower than SW controlled. draw your conclusions on battery life Sep 23 17:07:52 mhm, and what makes you think cpufreq doesn't reprogram LDOs dynamically? Sep 23 17:08:02 because I know it Sep 23 17:08:27 cpufreq uses OPP table and does one time programming on OPP switch Sep 23 17:09:12 one time? on the one time OPP is switched, or what? Sep 23 17:09:35 yes, when OPP is changed voltage is changed too Sep 23 17:09:45 TADAAA! Sep 23 17:09:54 but voltagee remains constant as long as OPP is not changed Sep 23 17:10:30 DocScrutinizer05: ? Sep 23 17:10:33 and WTF do you think can SR do better in that regard? Sep 23 17:10:44 hehe Sep 23 17:11:24 as long as opp switch is complete and new voltage is programmed, SR lowers that by 20-30% as I already told you Sep 23 17:11:34 actually now it seems it's you who has NFC what's SR Sep 23 17:11:37 based on temp/load/device calibrations Sep 23 17:11:47 sure, I have NFC Sep 23 17:12:07 Does somebody know a pdf viewer/ebook reader for fremantle with word wrapping on zoom? Sep 23 17:12:15 DocScrutinizer05: go read SR driver, you will get it (i hope) Sep 23 17:12:43 thanks, I already have read about SR in TPS65050 and OMAP UM Sep 23 17:14:33 and I've read the list of ll known SiErr, and found some that sounded suspicious Sep 23 17:14:37 re SR Sep 23 17:14:57 and now I'm fed up with this non-discussion Sep 23 17:15:38 the one you found would've already hit us if applicable, as the same happens when OPP is changed Sep 23 17:15:59 twl is re-programmed via i2c ;) Sep 23 17:18:28 everything that goes into the repository has to be OpenSource? Sep 23 17:20:40 Sc0rpius: basically yes Sep 23 17:22:09 NIN101: could pdf do reformatting even in theory? AIUI the format is inherent component of pdf Sep 23 17:25:10 I know(read) that some readers can do it... e.g. adobe's or the one in android... Sep 23 17:28:05 dont know what you guys are arguing about but at least in upstream cpufreq switches freqs and voltages by opp and handles voltage control to sr which does what it should. In upstream, sr is broken for some 3430s, sr in kernel-power works rock solid and does what it should. Sep 23 17:28:41 but not going to argue as you both know the hardware better Sep 23 17:29:56 Skry: exactly the same happens in KP ( cpufreq switches freqs and voltages by opp and handles voltage control to sr which does what it should ) Sep 23 17:32:04 yup Sep 23 17:33:17 my thought about this being: if it works, then it works and it has benefits so should be used. Sep 23 17:33:40 and if it does not work, fix it :) Sep 23 17:33:46 yup :D Sep 23 17:33:57 Skry: sorry, I have to ask about your definition of 2works" Sep 23 17:34:53 is one kernel oops per day acceptable. one per week? 10% of devices not stable with it, or would 5% be tolerable? Sep 23 17:35:34 DocScrutinizer05: so far there was not a single report of broken SR with KP >= 50 and stock frequencies Sep 23 17:35:47 pfff Sep 23 17:35:49 aha Sep 23 17:35:52 and in my book that means "it works" Sep 23 17:36:02 but it is actually quite confusing when engineers from ti, nokia and people in general all think different about this. Sep 23 17:36:15 sure, since there's no flag rising "THIS KERNEL OOPS TRIGGERED BY SR!" Sep 23 17:36:40 so users shrug and reboot Sep 23 17:37:00 and you say "fine, no ticked against SR opened" Sep 23 17:37:03 DocScrutinizer05: I am not aware of kernel oops in KP since KP50 got out. Besides bq_ thingie, but that is a different story Sep 23 17:37:05 ticket* Sep 23 17:39:13 my definition of "it works" in short is that _I_ don't get any oopses or lockups with sr and kp, and for _me_ sr improves battery life. For upstream kernel it is quite the opposite. Sep 23 17:39:47 and all this still doesn't yield a single evidence about supposed gigantic savings of SR over software doing exactly same Sep 23 17:41:08 DocScrutinizer05: come on Sep 23 17:41:10 the quoted 30% lower voltage on SR are bogus, since that's nothing software couldn't do as well )and probably even does) Sep 23 17:41:20 did not say anything about gigantic Sep 23 17:41:58 :not: Sep 23 17:42:16 DocScrutinizer05: on average it should extend battery life by 10-15% Sep 23 17:42:30 depends on usage pattern of course Sep 23 17:42:32 but this whole discussion is like "hahaha you say 'iirc', so I call you a liar until you bring me a tape record of that talk", and "WFM, so it's proven" Sep 23 17:42:56 never said "hahaha" Sep 23 17:43:08 and I'm out Sep 23 17:43:19 and yes, I need some evidence to support your statement "there are SiErrs" Sep 23 17:43:59 pease stop spreading FUD, it been SpeedEvil who first said that Sep 23 17:44:22 and he didn't say "iirc" Sep 23 17:45:18 and in my book that's 2 to 1 for those who know what they know, while the one person just prefers to think they're both lying or telling shit Sep 23 17:45:42 DocScrutinizer05: well, lets agree to disagree and stop that fight, ok :) ? Sep 23 17:45:46 and I'm not interetsed in discussions on that level Sep 23 17:46:14 DocScrutinizer05: my level? come on Sep 23 17:46:20 I'd feel reluctant to suggest SmartReflex to anybody, without even any proper evaluation of the expected benefit, compared to the very real concerns about Nokia telling it's not stable on hw level Sep 23 17:46:22 stands Sep 23 17:47:56 DocScrutinizer05: test it on a couple of n900s! Sep 23 17:48:12 MEEE?? dream on Sep 23 17:48:26 kerio: it is still WFM basis, thus not enough Sep 23 17:49:27 damn, you're exceptionally simple today, are you? evaluation savings is *totally* unrelated to friggin WFM, if done correctly Sep 23 17:50:22 DocScrutinizer05: I was talking about stability Sep 23 17:50:34 you were talking... about... aha Sep 23 17:50:47 as if it is not stable, it makes no sense to measure the savings Sep 23 17:51:01 you're not talking at all, you're living your bitchy mood Sep 23 17:51:18 actually my mood is pretty ok ;) Sep 23 17:51:38 yeah, I bet. Seems it's your preferred mood anyway Sep 23 17:51:51 for sure you must love it to argue Sep 23 17:52:20 no matter if there's any rationale in it ot not Sep 23 17:52:42 and honestly, I'm already more than fed up with it Sep 23 17:52:49 so for good now, cya Sep 23 17:59:35 oh why cant the sgx driver be reverse engineered :S Sep 23 18:00:09 Skry: sgx driver is OSS :P Sep 23 18:00:09 who says it can't? Sep 23 18:01:37 yeah the kernel part Sep 23 18:03:17 ok, i can handle those closed libraries, but there really should be working (modern) driver for xorg Sep 23 18:04:20 Skry: but,but... I thought it is upstreamed. Isn't it? Sep 23 18:04:37 aah, no Sep 23 18:04:44 no, andnever will be Sep 23 18:04:56 unless someone does some serious cleanup Sep 23 18:05:17 BTW did you manage to compile it? Sep 23 18:05:24 (xorg driver) Sep 23 18:06:06 freemangordon: xorg driver is closed Sep 23 18:06:49 i can compile kernel drivers ok if disabling xorg support from it, yes, they are linked Sep 23 18:08:02 i stopped trying to compile it with x support when I started getting unknown symbols from drm kmod Sep 23 18:09:08 hi there.I extracted the rootfs from maemo image,tared,then untared in mmcblk0p5 modified rcS-late,uncommented fstab generation.it hangs with the maemo loader Sep 23 18:09:14 line Sep 23 18:09:36 now fstab is static Sep 23 18:10:20 use bootloader.didn't touch to the original fs Sep 23 18:10:30 need some advice please Sep 23 18:13:08 my goal is to boot maemo from sd card Sep 23 18:25:56 I think it's been accomplished before, maybe NIN101 knows details Sep 23 18:27:27 dunno, never did it myself. Sep 23 18:32:19 http://pastebin.com/raw.php?i=QH2EfTux parts of this config might help, basically you need the file system of sd card partition compiled into the kernel, not as a module... Sep 23 18:32:36 and modify the comamnd line of the kernel to actually boot from sdcard then Sep 23 18:32:43 (or do it by using flasher) Sep 23 18:33:06 yep, main obstacle being the fs driver Sep 23 18:33:27 KP should be fine Sep 23 18:34:29 just make uBoot load KP image from mmcblk0pN, and pass it kernel cmdline to have rootfs also on mmcblk0pN Sep 23 18:35:30 you can also modify the built-in command line when recompilng the kernel(this is what you have to do anyway) Sep 23 18:35:34 keep in mind during early boot uSD is mmcblk0 and eMMc is mmcblk1. It only gets renamed later Sep 23 18:37:15 the nifty rationale of that probably being an option to make ROMBL load xloader/NOLO/$n1st-stage-BL from a swap-able uSD Sep 23 18:38:18 the ugly part of it: if there's no uSD plugged in, the eMMC will become mmcblk0 even during early boot/kernel-init Sep 23 18:39:04 sounds like windows drive letter hell, doesn't it? ;-P Sep 23 18:49:23 http://i.imgur.com/FdGXY.jpg Sep 23 18:53:31 hehe Sep 23 18:54:24 NIN101: why do you think zsolt needs a custom kernel? Sep 23 18:56:11 shouldn't uBoot boot a proper KP plus maemo5 rootfs from uSD or eMMC, just like it did (and still does?) for meego? Sep 23 18:58:18 I'd probably install meego dualboot setup on uSD, then replace meego kernel by KP, and copy fremantle rootfs over meego rootfs. Then just check uBoot config file for possible differences needed in kernel cmdline parameters Sep 23 18:59:16 hmm ok, didn't assume uboot wa sused Sep 23 18:59:41 well, sounds like the easiest approach Sep 23 19:00:02 plus keeps original NAND maemo intact Sep 23 19:00:56 actually I placed exactly this task on my ToDo some 2 years ago. Eventually it fell off the bottom end Sep 23 19:01:31 i installed u-boot Sep 23 19:02:33 thank you for help Sep 23 19:05:19 it is funny,i told for u-boot to boot from mmcbl0p5 and put the zImage to /boot,what happens?it boots up from nand Sep 23 19:07:02 could you give some details about how you told u-boot to do that? Sep 23 19:09:11 nah, gtg Sep 23 19:18:22 i seriously need some kind of bluetooth remote control for the n900 :( Sep 23 19:18:31 that is, something on my laptop that lets me do stuff with the n900 Sep 23 19:18:53 at the very least connect to a wifi network Sep 23 19:19:18 hey, why does mine connect to WLAN automatically? ;-) Sep 23 19:20:34 DocScrutinizer05: because you like to waste power looking for a wlan >:C Sep 23 19:21:10 anyway, i do realize that it's kind of a niche usecase Sep 23 19:21:19 kerio: incorrect, I have mine setup to do autoconnect every 5 minutes Sep 23 19:22:14 kerio: I once had a QB desktop button to toggle the "connect to WLAN if available" setting Sep 23 19:22:26 DocScrutinizer05: the problem is accessing the desktop Sep 23 19:22:36 I needed that when my DSL was flaky Sep 23 19:22:55 the usecase is this: i'm pooping, and my phone is in my room Sep 23 19:23:10 and it's not connected to the wifi network Sep 23 19:23:33 :P Sep 23 19:23:39 why? Sep 23 19:23:50 DocScrutinizer05: because i don't have autoconnect on Sep 23 19:23:59 yeah, why? Sep 23 19:24:07 because battery! :( Sep 23 19:24:21 actually being connected to WLAN is pretty lightweight Sep 23 19:25:33 so it seems your problem isn't to make N900 connect to WLAN at arbitrary time by user interaction, but rather make it connect to WLAN every time you enter the room Sep 23 19:25:49 or rather, come home Sep 23 19:28:38 DocScrutinizer05: actually, i know what i want Sep 23 19:28:43 ssh-through-bluetooth Sep 23 19:29:12 well, if you're not interested in alternatives, go and set it up Sep 23 19:29:46 i wonder how Sep 23 19:29:50 PAN Sep 23 19:29:53 last i heard, bluez was a giant mess Sep 23 19:30:06 nah, PAN on fremantle is weird as fuck Sep 23 19:30:13 i need a serial connection, bluetooth allows that Sep 23 19:30:18 you asked for it. Explicitly Sep 23 19:31:06 actually I'd be surprised to find ssh work via plain serial Sep 23 19:31:20 i said ssh-over-bluetooth, not ssh-over-tcp-over-bluetooth :) Sep 23 19:31:28 it's rather network based, no? Sep 23 19:31:31 DocScrutinizer05: are you kidding? ssh over pipes is insanely useful Sep 23 19:31:39 ProxyCommand Sep 23 19:32:24 mhm, and how you're handling the URLs etc, the server name based auth Sep 23 19:32:27 hmm, is there a command (apart for screen) to get input-output to and from a device file? Sep 23 19:33:04 no, known_hosts isn't checked if you use ProxyCommand Sep 23 19:33:15 because the hostname could be something weird Sep 23 19:33:20 :shrug: Sep 23 19:33:24 too weird for me Sep 23 19:34:01 get WLAN working. If you for F*s sake can't do that, get PAN working Sep 23 19:34:11 the ssh client has an option to specify an executable to use as a proxy pipe Sep 23 19:34:28 input and output eventually have to reach a ssh server, in some way Sep 23 19:34:52 if you can't even do that, send a semaphore file to some well known location on N900, via BT OBEX Sep 23 19:35:01 oooh, that's simple Sep 23 19:35:03 neat Sep 23 19:35:32 but you said you're not interested in alternatives Sep 23 19:35:35 DocScrutinizer05: is there a command to do bidirectional transfers with a device file? Sep 23 19:36:00 something like "screen /dev/ttywhatever" but that only does simple piping Sep 23 19:36:03 err, isn't that called minicom? Sep 23 19:36:09 minicom is interactive too Sep 23 19:36:54 cat foo >/dev/bar Sep 23 19:37:34 that's one-way only, though Sep 23 19:37:43 cat /dev/bar& Sep 23 19:37:53 heh Sep 23 19:38:33 dunno what you're planning to do Sep 23 19:39:08 i need the stdin of the command to be sent to the device file, and the output read from the device file to be outputted in the stdout Sep 23 19:39:41 what's the use of that? Sep 23 19:40:11 something like >> date|cat|cat|cat < how's stdout of any such tool different from piping the original device somewhere? Sep 23 19:40:58 ssh wants an executable Sep 23 19:42:14 for good reason Sep 23 19:42:30 oh nvm, you actually need to set the device file up in some way if it's a tty Sep 23 19:43:28 anyway, i only need this on my computer, on the n900's side its' just *getty that calls sshd -i or nc localhost 22 Sep 23 19:43:40 don't you think getting PAN to purpose would be more sane? Sep 23 19:43:46 nope! Sep 23 19:44:21 :shrug: Sep 23 19:44:27 hmm, what's a good time for the automatic WLAN search? is 5 minutes too short? Sep 23 19:44:43 you might win some funny award with this solution, if you ever make it work Sep 23 19:45:04 I've put it to 10min Sep 23 19:45:17 set it to* Sep 23 19:46:56 way smarter of course would be to set it to 'off' and only trigger an immediate scan or set autoscan to 5min when you see some special trigger. GPS is probably no brilliant idea, but a certain cell-id on cellmo might work nicely Sep 23 19:48:01 your beloved SMSCB opername widget already dumps cellids to the log files Sep 23 19:48:24 DocScrutinizer05: a special trigger such as my laptop opening a bluetooth connection of some sorts, maybe... Sep 23 19:48:35 i'm still going through with this idea of ssh-over-serial, btw Sep 23 19:49:31 "opening BT connection"? why not just trigger on associate Sep 23 19:49:44 yeah, that one Sep 23 19:50:01 (although os x doesn't have a way to associate without doing something with the association) Sep 23 19:50:15 there's a tool on desktops that locks screens automatically when you leave BT range with your paired device, and unlocks when you come back Sep 23 19:50:42 standard bluez afaik Sep 23 19:51:29 you probably could plumb sth together on maemo that does pretty much the same Sep 23 19:51:57 I mean, association is a bilateral thing Sep 23 19:52:12 or rather, connection Sep 23 19:52:22 and loss thereof Sep 23 19:52:58 agetty -8 -i -n -L -l "/usr/bin/nc localhost 22" /dev/ttyB0 Sep 23 19:53:01 or something like this :D Sep 23 19:53:54 so... /etc/bluetooth/main.conf looks unused Sep 23 19:59:51 I'd be honestly surprised to see this work Sep 23 20:18:07 hrmpf, i can't find documentation on how to open a serial bluetooth port Sep 23 20:18:59 kerio: wait a minute, I wrote something an year ago, I think it will help Sep 23 20:19:06 freemangordon: i need the "server" part Sep 23 20:21:45 you are looking for "rfcomm", i think Sep 23 20:21:50 kerio: don't know what do you mean by "server part",but you need to call rfcomm Sep 23 20:22:05 Sicelo: aah, you beat me :D Sep 23 20:22:17 freemangordon: i meant that i want the n900 to wait for a serial connection from my laptop Sep 23 20:22:45 kerio: though if you wish I can give you my btsettingsqt :D Sep 23 20:23:34 lemme try if it still works (and how) Sep 23 20:23:35 oh ok, rfcomm listen Sep 23 20:27:25 kerio: it is still ok, bt helper in settings (if you want it) Sep 23 20:27:41 you can connect com port with it Sep 23 20:27:58 it creates serial device in /dev and goodies like that Sep 23 20:29:15 kerio: http://dox.bg/files/dw?a=a957243826 Sep 23 20:29:57 ty Sep 23 20:31:29 hrmpf, "serial port" connects to the bluetooth DUN interface Sep 23 20:31:55 you have DUN on your laptop? Sep 23 20:32:07 service that is Sep 23 20:32:53 kerio: DUN is a trick Sep 23 20:33:24 afaik, it's just using pnatd to pipe between serial BT and cellmo Sep 23 20:33:39 kerio: you tried rfcomm or my app? Sep 23 20:33:40 done in some scriptie or event file Sep 23 20:34:17 kerio: you probably want to replace pnatd by your cxontraption to interface to ssh Sep 23 20:34:24 freemangordon: not yet Sep 23 20:34:47 kerio: try it, it connects the correct thing (IIRC) Sep 23 20:34:56 not DUN Sep 23 20:35:02 hm Sep 23 20:35:23 well apparently os x doesn't support RFCOMM out of the box anyway Sep 23 20:35:27 so first i need a RFCOMM utility Sep 23 20:36:16 less /etc/event.d/phonet-at Sep 23 20:37:10 DocScrutinizer05: hmm, that's the gadget serial though Sep 23 20:37:19 sure Sep 23 20:38:24 i see Sep 23 20:38:31 kerio: I just made a connection between /dev/rfcomm0 and COM4 on my PC Sep 23 20:38:45 DocScrutinizer05: it's /etc/event.d/bluetooth-dun :) Sep 23 20:38:48 and it's about as simple Sep 23 20:38:58 i need to register a SPP instead of a DUN, but it's really the same thing Sep 23 20:39:01 tyvm y'all Sep 23 20:45:01 hmm, maybe i should finish that application some day Sep 23 20:46:41 no, you should start coding me sgx xorg driver first ;> Sep 23 20:46:49 hehe Sep 23 20:47:11 no, you should boot maemo5 with kernel 3.6 first :D Sep 23 20:47:30 actually I'm about to try :) Sep 23 20:47:34 well, 3.2 will do the job too :P Sep 23 20:47:56 Skry: ask Pali for help, he has some bits needed to do that Sep 23 20:48:18 Skry: you said thumb2+hardfp compiled stuff will run on n900 in a chroot? Sep 23 20:48:26 vi_: pong Sep 23 20:48:32 vi_: yes Sep 23 20:48:35 (very late one) Sep 23 20:48:35 freemangordon: hey. Sep 23 20:48:50 freemangordon: I kludged through my issue. Sep 23 20:48:59 thankyou anyways. Sep 23 20:49:01 Skry: only with KP or kernel-cssu Sep 23 20:49:10 yep Sep 23 20:49:16 because of thumb Sep 23 20:49:32 vi_: ok Sep 23 20:49:38 np Sep 23 20:49:46 Skry: does strait up debian have a hardfp+thumb2 branch? Sep 23 20:49:53 Skry: I have patch for meego (2.6.37+) kernel which add maemo5 specific interface Sep 23 20:50:11 but not everything Sep 23 20:50:19 Pali: hi Sep 23 20:50:22 hi Sep 23 20:50:38 I still wait an answer re libbmeipc :P Sep 23 20:50:54 why you don;t like my code :D Sep 23 20:51:09 vi_: I dont know, they had it in planning state though at some point, if memory serves Sep 23 20:51:13 for unknown reason qemu cannot use emmc Sep 23 20:51:24 so I was not able to test that patch in qemu Sep 23 20:51:25 Pali: could you give me a link for that patch? Sep 23 20:51:30 yes, wait Sep 23 20:52:06 neon optimised bzip2 would be sweet. Sep 23 20:52:20 Skry: http://atrey.karlin.mff.cuni.cz/~pali/meego_n900_kernel_fremantle.patch Sep 23 20:52:24 vi_: who uses it? apt? Sep 23 20:52:34 patch is for meego n900 kernel (from gitorious branch) Sep 23 20:53:02 freemangordon: but seriously, someone should start doing that xorg driver, I think one could get xf86-video-modesetting and start including sgx stuff ala fbdev-sgx Sep 23 20:53:13