**** BEGIN LOGGING AT Wed Jul 28 02:59:56 2010 Jul 28 08:03:43 morgen Jul 28 08:04:01 anybody tried chromium on the free runner? Jul 28 08:05:27 their V8 JS engine seems noticably faster than the webkit's default SquirrelFish, at least on an omap zoom2 Jul 28 08:08:04 which is cortex, no idea if it will even work on armv4 Jul 28 08:09:18 hm, yup, that's what I was afraid of Jul 28 08:11:37 zub: i think lindi- actively tests it. Jul 28 08:12:06 zub: yes please see my bug report in upstream and bugs.debian.org (email timo.lindfors@iki.fi) Jul 28 08:12:14 zub: and iki.fi/lindi/chromium/ Jul 28 08:12:30 s/bug report/bug reports/ Jul 28 08:12:33 zub: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=584946 e.g. Jul 28 08:12:49 thanks Jul 28 08:17:00 so, it seems some version run, but then v8 started to rely on >v4 instructions Jul 28 08:17:37 zub: well there's only one instruction to patch iirc Jul 28 08:19:09 and what about memory usage? Jul 28 08:19:20 On my desktop, chromium is a real RAM eater... Jul 28 08:20:11 nice I didn't know that Jul 28 08:20:27 it's also possible you could build just webcore with v8 and not chromium Jul 28 08:20:39 though chromium is getting optimzied for arm as part of the chromiumos project Jul 28 08:20:54 tmzt: v8 w/o chromium backend? Jul 28 08:21:07 I think so, you'd have to patch it up Jul 28 08:21:55 imho would need some work. maybe not so much for webkit-efl, as that doesn't yet export any JS functionality, but webkit-gtk already includes some JavaScriptCore stuff (that I believe == SquirrelFish) Jul 28 08:22:28 the stuff they use for gnome3? or will Jul 28 08:22:57 ah, no idea about gnome3 Jul 28 08:25:13 or, maybe now I got it - you, one could builw webkit chromium port, w/o the browser itself Jul 28 08:25:25 yeah that's what I'm saying Jul 28 08:25:46 I'm actually now trying to build a hello world browser with wenkit-chromium, but not much of a success so far Jul 28 08:26:52 ok, I'm interested to know how it goes Jul 28 08:27:18 so far just on pc Jul 28 08:44:03 hi guys Jul 28 09:12:50 hi guys (two :) ) : In which file I need to put my PIN code to have the gsm work ? Jul 28 09:12:58 (i don't remember) Jul 28 09:16:23 founded /etc/phonefsod.conf (to set the pin code) Jul 28 09:35:06 FSO/SHR guys: du you know why I don't have gsm? here is a pice of fsogsmd.log the http://pastebin.com/0krwstFU Jul 28 09:41:43 GarthPS|office: seems like you need to unlock your SIM with the PUK... Jul 28 09:43:44 wonka : Wo Jul 28 09:44:37 what? Jul 28 09:45:07 wonka : lol eh in fact not very funny but It is because I enter a wrong pin and use the switch to turn on gms 2 or 3 times.. Jul 28 09:46:48 wonka : ok confirmed.. Jul 28 09:47:15 all good now? Jul 28 09:48:43 wonka : well.. not really because I don't have my puck code here.... Jul 28 09:49:02 wonka : so phon broken until this evening... :p Jul 28 09:49:16 too bad Jul 28 09:49:24 wonka : next time I will verify twice my pin Jul 28 09:49:33 wonka : but thx Jul 28 09:49:57 (I did not read the goog line beacase it was nearly the same than my previous test Jul 28 10:02:09 mrmoku: pign Jul 28 10:02:11 err Jul 28 10:02:13 ping, even Jul 28 10:21:31 just flashed latest shr-u. the finsh button in the firstrun wizard on the password screen is not working. how can i get past it? Jul 28 10:27:01 hi guys one question: is xrender support by the glamo? Jul 28 10:28:42 http://www.turran.org/2009/03/glamo-xrender-benchmark-with-expedite.html Jul 28 10:29:39 mickey|office: pong Jul 28 10:41:13 mickey|office: lunch now... bbiab Jul 28 11:06:37 mickey|office: re Jul 28 11:49:10 rohezal: yep, supported, but only for limited bpp and with the same overheads as blitting Jul 28 11:49:27 rohezal: hmm actually.. I think it's not accelerated in hardware Jul 28 12:06:24 Kensan: hey. are you around ? Jul 28 12:50:08 Hi all Jul 28 12:51:07 Can SB help me? I am trying to install Debian on my OM Freerunner Jul 28 12:53:41 Kronecker82, there's openmoko-debian channel or you can just ask Jul 28 12:55:03 ah i just found that in the message above thanks, ill ask in that channel Jul 28 13:39:27 Weiss: damn i need a way for transparent blitting... Jul 28 13:40:22 Weiss: do you think gtk and its functions are accelerated? Jul 28 13:40:48 Weiss: or qt and its functions? Jul 28 13:42:55 damn i must go :( Jul 28 13:43:02 im asking later again^^ Jul 28 14:28:15 hey guys, anyone can test shr WS fix? Jul 28 14:28:46 may be not really fix but idea. Jul 28 14:29:38 it works here of course. Jul 28 14:30:59 ehlo Jul 28 14:31:05 trying to use gprs via command line on shr-t, the script on the wiki reports: Jul 28 14:31:09 No method org.freesmartphone.GSM.PDP.SetCredentials Jul 28 14:31:19 http://wiki.openmoko.org/wiki/GPRS_FSO#Using_scripts Jul 28 14:32:20 b0ef: PDP.SetCredentials is no more Jul 28 14:32:35 err Jul 28 14:32:38 no Jul 28 14:32:42 it still is Jul 28 14:33:35 here's the full error: Jul 28 14:33:36 [ERR]: No method org.freesmartphone.GSM.PDP.SetCredentials found at /org/freesmartphone/GSM/Device for org.freesmartphone.ogsmd Jul 28 14:33:36 [ERR]: Need 3 params for signature 'sss', supplied 0 Jul 28 14:36:09 b0ef: well, you have to pass it apn, user and pw Jul 28 14:36:29 mrmoku: yes, I use the script on the link I just gave you Jul 28 14:38:46 does anyone interested in proper working 2-4-2 glamo timings? Jul 28 14:40:24 hm... Jul 28 14:40:47 gena2x, the one you posted on list was wrong? Jul 28 14:41:10 shr kernel has many wses then using this uboot Jul 28 14:41:39 i just found place which seem fix this. Jul 28 14:42:28 but this place in glamo, so i can't do proper fix anyway. Jul 28 14:43:46 but weiss seen not interested in this, and now i really don't know how to proceed with this issue. Jul 28 14:53:25 so, now i have kernel patch with added delay. Jul 28 14:56:17 but i don't really understand why nobody want to test it. this should give shr people working 2-4-2 timings. Jul 28 14:57:17 gena2x, maybe not many people are compiling their own kernel Jul 28 14:57:29 i have a kernel compiled Jul 28 14:57:31 I don't even know what you mean by 2-4-2 timings... Jul 28 14:57:49 if you just want just a tester, then why not, though Jul 28 14:57:53 zub: have you read my mail about glamo in community ml? Jul 28 14:58:05 nope, I'm not following that ML :-( Jul 28 14:58:25 I feel I'm already on too may MLs Jul 28 14:58:28 zub: http://lists.openmoko.org/pipermail/community/2010-July/062495.html Jul 28 14:58:33 ok, I'll have a look Jul 28 14:59:17 * zub is reading Jul 28 15:02:14 i hope if someone else can confirm that ws is gone with this patch, then i'll be able to convience weiss to take look to this patch. Jul 28 15:02:50 my lack of having done anything isn't due to non-interest in the results, I assure you.. Jul 28 15:03:25 just that work doesn't currently leave me with many brain cells to use for non-work (particularly the programming kind) Jul 28 15:04:11 gena2x: (just halfway through; multitasking with day job) seems you invested quite some work in this Jul 28 15:04:25 and it sure sounds interesting Jul 28 15:05:07 btw. I've been having issues with the kernel ever since it was upgraded to .32 :-/ Jul 28 15:05:13 Weiss: eh, so pity to hear that from fr point of view. but i hope work brings something great for you. Jul 28 15:05:18 but I didn't have time to look into it firther Jul 28 15:06:40 Weiss: so, it's better to wait for weekend for this issue? Jul 28 15:08:08 zub: ok, i'll proceed to next issue. Jul 28 15:09:21 Weiss: all that were just unclean for me - i posted patch yesterday, but seem you didn't look into it, so i've left in doubt. Jul 28 15:12:06 gena2x: finished the mail Jul 28 15:13:00 zub: ok, as a minimum, you know now what is meaning of glamo 2-4-2 :) Jul 28 15:13:16 gena2x: if you want a tester and are willing to explain it to a noob, then you can count me in :) Jul 28 15:13:32 gena2x: it's on my desktop to play with the next time I get enough time and can face a text editor.. Jul 28 15:14:59 Weiss: bsdmn.com/openmoko/glamo/gcc44wsfix.diff? Jul 28 15:15:07 yep, that one Jul 28 15:15:35 ok, nice. Jul 28 15:18:08 zub: ok. tell me then you have some time to spend on testing. i think this can take ~30mins. Jul 28 15:19:02 in general I'm available evening/night GMT+2, I'll ping you then Jul 28 15:19:20 zub: nice, thanks. Jul 28 15:19:47 n/p Jul 28 16:11:30 SHR: 03seba.dos1 07opimd-utils * r20e2bc2f6372 10/opimd-contacts: contacts: display user definied photos for contacts Jul 28 17:02:49 hi found it. you can draw transparent pixels in X via clipping: http://www.dis.uniroma1.it/~liberato/screensaver/trans.html Jul 28 17:03:02 just want to tell it if someone was searching Jul 28 17:20:06 mrmoku, Heinervdm, someone: Please bump EFL, cool stuff there :P Jul 28 17:20:36 Did anyone encounter this one already? "openmoko/shr-unstable/openembedded/recipes/mpfr/mpfr_3.0.0.bb failed" Jul 28 17:20:50 (And yes, I installed 'xz-utils' ;-) Jul 28 17:21:25 mrmoku, Heinervdm, someone: cool stuff = only rtl fixes that you probably don't care about. Jul 28 17:23:37 Here's (part of) the error log (I couldn't find a reference to the .log file in my backscroll...): http://pastebin.com/hF2PaqKu Jul 28 17:24:08 TAsn: i think it's better if mrmoku bumps it, because he can build it too :) Jul 28 17:24:38 Heinervdm, mrmoku: It's probably better to wait for all the bugs to float :P I huge commit I finished today :P Jul 28 17:25:11 :) Jul 28 17:28:55 :P Jul 28 17:31:09 its a * Jul 28 17:31:11 ffs my english. Jul 28 17:32:52 TAsn: right in time be fore the coding style script hits svn :P Jul 28 17:33:26 the coding style script is there Jul 28 17:33:29 they just didn't use it. Jul 28 17:33:38 Heinervdm: want me to add your key? Jul 28 17:33:41 they used it in some places. Jul 28 17:34:01 mrmoku: no, i don't have that much time currently Jul 28 17:34:11 yeah... know that deadline is sun asia time :P Jul 28 17:34:16 Heinervdm: k Jul 28 17:37:31 gena2x: ping Jul 28 17:37:45 pong Jul 28 17:38:20 ChristW: are you building the full image? Jul 28 17:44:26 ChristW: i think your error was fixed 36min ago: http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=138c033ea5fc63e0ad33d363180e40b9ab492e63 Jul 28 17:49:55 ChristW: ok, it isn't solved Jul 28 17:50:13 ChristW: it seems that this version isn't compatible with armv4t anymore Jul 28 17:51:06 heh :/ Jul 28 17:54:21 fun. Jul 28 17:58:20 but why do we need that lib? Jul 28 17:59:52 hmm, cpp depends on it... Jul 28 18:08:02 openbmap-logger seems to have disappeared from the shr-u repo Jul 28 18:26:25 I just tried building Qi for GTA02 (head), built with CROSS_COMPILE=arm-angstrom-linux-gnueabi- make CPU=s3c2442 (using angstrom toolchain, gcc 4.4.3) Jul 28 18:26:34 it built ok, but the resulting Qi didn't work, or at least didn't load kernel/no led was lit Jul 28 18:26:38 any idea what's wrong? Jul 28 18:32:59 hello is anybody repairing the shr-unstable build ? Jul 28 18:35:59 mrmoku: who can repair the shr-unstable build ? Jul 28 18:41:37 #shr Jul 28 18:47:19 nschle85: on the buildhost? Jul 28 18:47:58 mrmoku: hey. i found out we have some kind of voicemail support on the calypso Jul 28 18:48:08 mickeyl: huh :) Jul 28 18:48:19 i could synthesize a signal for when there is a new voicemail waiting for you Jul 28 18:48:34 always good for a surprise that calypso :) Jul 28 18:48:39 indeed Jul 28 18:48:45 and another thing Jul 28 18:48:57 i have repaired the missing 'issuer' field in GetSimInfo Jul 28 18:49:01 i have thought a bit about it Jul 28 18:49:10 but came to the conclusion that it's better handled in app space than in fsogsmd Jul 28 18:49:24 i.e. substituting 'display' string with 'issuer' when we are in home network Jul 28 18:49:50 (e.g. Alice vs. o-2.de) Jul 28 18:50:00 hmm Jul 28 18:50:13 you don't think so? Jul 28 18:50:17 hi Jul 28 18:50:22 don't understand fully yet :) Jul 28 18:50:23 ah Jul 28 18:50:24 ok Jul 28 18:50:28 example Jul 28 18:50:29 NetworkStatus will have only one then? Jul 28 18:50:53 anyone knows why I can't write a sms with SHR? Jul 28 18:51:04 no, we will still support 'display' (short name) and 'operator' (long name) Jul 28 18:51:05 however Jul 28 18:51:17 the virtual operator name will not be autosubstituted by us Jul 28 18:51:36 since that sounds like a layering violation to me Jul 28 18:51:59 it might be nice to expose both the virtual operator name and the actual operator name Jul 28 18:52:00 ahh, ok, you were doing that o-2.de --> Alice replacement by taking issuer of SIM when in home? Jul 28 18:52:01 (for power users ;) Jul 28 18:52:15 i thought about doing that, yes Jul 28 18:52:29 but i think that should be rather done by the phone app, if it wants to Jul 28 18:52:58 so, right now my phone shows M-Net instead of o2... where does the M-Net come from then? Jul 28 18:53:17 mrmoku: no i am using make file make image Jul 28 18:53:26 nschle85: and what fails? Jul 28 18:53:56 mrmoku: package mpfr 3.0.0 Jul 28 18:54:05 nschle85: update :) Jul 28 18:54:54 mickeyl: trying to understand what is happening now and what will change... Jul 28 18:55:21 nschle85: ahh... see you're discussing it in the correct channel :) Jul 28 18:55:46 hope it will help Jul 28 18:57:41 mrmoku: ok, the interesting thing is: Jul 28 18:57:45 mrmoku: updateb did not help Jul 28 18:57:47 some SIM cards have broken EONS tables Jul 28 18:58:04 if calypso can't read the EON table, it won't be able to substitute that on its own Jul 28 18:58:23 however you have little means to detect when this happens :) Jul 28 18:58:35 so the 'safe' way is to always use the SIM issuer if your are camped to home Jul 28 18:59:05 ok Jul 28 18:59:09 understand now Jul 28 18:59:15 and you were thinking about doing that in fsogsmd Jul 28 18:59:20 exactly Jul 28 18:59:23 and now I agree that you should not :) Jul 28 18:59:27 ah, good :) Jul 28 18:59:47 ok, wrt. voicemail, i'll create a signal that informs you about that Jul 28 19:00:02 we would need a Get thing too though Jul 28 19:00:03 org.freesmartphone.GSM.VoiceMail.Incoming or so Jul 28 19:00:13 yes, that's going to be available as well Jul 28 19:00:17 good Jul 28 19:00:19 as well as get the number Jul 28 19:00:25 which is conveniently stored on the SIM as well Jul 28 19:00:28 wow :) Jul 28 19:00:30 so you can do a poor man's "visual voicemail" Jul 28 19:00:34 (iphone thingy) Jul 28 19:00:56 i overlooked this part of the spec for many months Jul 28 19:01:03 but it's actually all documented Jul 28 19:01:09 although in the NDAed part Jul 28 19:01:13 since it's a GSM extension Jul 28 19:01:22 very secret thing that voicemail ;) Jul 28 19:01:30 so there are no standard commands for it, but just vendor specific Jul 28 19:01:31 hehe Jul 28 19:03:13 i'll document all those "new" commands in the code Jul 28 19:03:32 not before weekend though Jul 28 19:03:46 ok, we ill have to think about how to use that UI wise anyway :P Jul 28 19:03:52 +w Jul 28 19:04:49 *nod* Jul 28 19:05:20 i'm not sure whether there is a standard way to communicate with those boxes Jul 28 19:05:36 it would be nice if you could provide an answering machine interface Jul 28 19:05:41 but that might be out of scope Jul 28 19:08:17 to call the voice box and get the selected message? Jul 28 19:08:17 mrmoku: btw, does SHR inform the user somehow when the battery is low enough for calypso to not work? I guess it's mad confusing to have a working device with non-working modem. Jul 28 19:08:21 hehe, would be interesting Jul 28 19:08:41 PaulFertser: no, because I don't know how to Jul 28 19:09:09 mrmoku: (voice box) yes, Jul 28 19:09:31 an interface where you transparently dial the number of the voice mail server and present the messages Jul 28 19:10:47 (battery) only thing we can do is add a watchdog hook and notify the phone app when the modem doesn't answer anymore. whether this is due to a hangup or a low battery fsogsmd will not be able to find out Jul 28 19:11:36 mickeyl: that's the reason i talked to mrmoku who can measure the voltage pereodically and inform the user when it's too low. Jul 28 19:12:01 hmm Jul 28 19:12:19 i'm not sure whether there is a unique threshold we can find Jul 28 19:12:24 but it may be worth a try Jul 28 19:13:08 mickeyl: there's datasheet, i think it's safe to inform the user when the datasheet threshold is hit. Jul 28 19:14:07 the modem's datasheet tell for which voltage ranges it's specified? Jul 28 19:14:19 mickeyl: i guess it should. Jul 28 19:14:36 good Jul 28 19:20:02 mrmoku: would you be interested in SIM status countersß Jul 28 19:20:20 e.g. how many false attempts left Jul 28 19:20:24 (to unlock) Jul 28 19:20:43 mickeyl: that would be interesting too Jul 28 19:20:53 to show in the PIN dialog Jul 28 19:21:05 hi Deubeuliou Jul 28 19:21:16 is software keyboard usable on nexusone? Jul 28 19:21:18 hmm Jul 28 19:21:25 let me try to see whether this info is correct Jul 28 19:21:27 * mickeyl adds false SIM PIN Jul 28 19:23:30 right, that works Jul 28 19:24:28 mickeyl: hmm... iphone has that voicemail thing? Jul 28 19:24:36 http://www.anandtech.com/show/2571/2 Jul 28 19:25:03 yep, that's it Jul 28 19:25:04 there must be a way then ;) Jul 28 19:25:05 it's pretty slick Jul 28 19:25:22 note that it only works with selected operators though Jul 28 19:25:35 you can't just stick any SIM card into and expect it to work Jul 28 19:26:51 does it just telecommand the voicebox menu? Jul 28 19:28:16 Die Visual Voicemail ist ein Apple-Produkt, das auf einem iPhone mit einer O2-Karte nicht funktioniert. Jul 28 19:28:19 freesmartphone.org: 03mickey 07specs * r4a26955b481b 10/ (4 files in 4 dirs): org.freesmartphone.GSM.SIM: add GetUnlockCounters() -> a{sv} Jul 28 19:28:22 (http://www.o2online.de/nw/support/mobilfunk/smartphones/iphone/allgemeines/allgemeines.html) Jul 28 19:28:45 mrmoku: yes, i think it "just" telecommands the voicebox menu Jul 28 19:28:54 but obviously it needs to know the protocol Jul 28 19:29:03 which - of course - is different for every provider Jul 28 19:29:43 but we could have a configuration for that... Jul 28 19:30:56 freesmartphone.org: 03mickey 07libfso-glib * r9ab41751525d 10/configure.ac: bump Jul 28 19:31:00 yep Jul 28 19:31:14 it may be ambitioned and probably relatively unimportant, but still nice Jul 28 19:31:49 yeah, those iphone screens look nice :/ Jul 28 19:32:24 mickeyl: (OT) do you have experiences with ipsec on iphone against linux servers? Jul 28 19:32:44 no, never tried ipsec anywhere Jul 28 19:32:48 k Jul 28 19:32:52 hi Jul 28 19:33:00 security unfortunately always went somewhat under my radard... Jul 28 19:33:03 radar, that is Jul 28 19:33:04 yo dcordes_ Jul 28 19:33:05 * mrmoku has a stupid client that got iphones and now wants to access the exchange server :( Jul 28 19:33:12 heh Jul 28 19:33:43 mrmoku: I like my 4,3" hd2 screen better than any iphone Jul 28 19:34:13 * mickeyl mentions it's the combination of hardware and software that makes the iphone unrivaled... Jul 28 19:34:27 dcordes_: well... I was (unfortunatelly) not involved in that decision... Jul 28 19:35:09 yes the iphone experience is really nice Jul 28 19:35:15 in terms of usability Jul 28 19:35:19 and as admin of their linux server now have to make that thing just work Jul 28 19:35:31 and of course it can never be the fault of the iphone :( Jul 28 19:35:39 if it does not work Jul 28 19:36:12 mickeyl, hi PM me when you'll have a bit of time(it's not so important for you) Jul 28 19:37:41 freesmartphone.org: 03mickey 07cornucopia * r766536c968b9 10/fsogsmd/ (configure.ac src/plugins/dbus_service/plugin.vala): fsogsmd: add stub implementation for org.freesmartphone.GSM.SIM.GetUnlockCounters() Jul 28 19:37:42 gena2x: do you have kernel image with 2-4-2 and fixed WS somewhere? Jul 28 19:37:46 hm, seeing the issue again, I guess it's known - the Finish button in the last screen of the 1st time wizzard doesn't work Jul 28 19:37:48 GNUtoo|laptop: i'm just about to go to bed, can we talk tomorrow? Jul 28 19:37:58 anybody looked into it? Jul 28 19:37:58 yes we can Jul 28 19:37:58 * mickeyl ultratired today Jul 28 19:38:01 k Jul 28 19:38:05 g'night then Jul 28 19:38:08 (seem so obama) Jul 28 19:38:14 mickeyl: gnight Jul 28 19:38:14 ;) Jul 28 19:38:19 good night Jul 28 19:38:45 GNUtoo|laptop: hi my friend Jul 28 19:38:50 Martix: hi, 2-4-2 timing is set in bootloader. Jul 28 19:38:51 hi Jul 28 19:38:59 GNUtoo|laptop: some people asked me about the dream SHR :( Jul 28 19:39:02 Martix: so you need both bootloader and kernel. Jul 28 19:39:06 ok Jul 28 19:39:12 dcordes_, I started to make an install guide Jul 28 19:39:14 but.... Jul 28 19:39:15 GNUtoo|laptop: I told them the openmoko devs are working hard to make a beta relase soon :D Jul 28 19:39:21 GNUtoo|laptop: nice :) Jul 28 19:39:21 gena2x: ok, where is kernel? Jul 28 19:39:25 I was also busy with oe toolchains issues + wifi Jul 28 19:39:46 if I unlock wifi it would be so great Jul 28 19:40:06 GNUtoo|laptop: on my htc hd2, wifi works perfect in non-android. Jul 28 19:40:20 dcordes_, wl1251_sdio? Jul 28 19:40:22 GNUtoo|laptop: even configurable with nice wicd gui so it's user friendly Jul 28 19:40:30 oh no 4829 Jul 28 19:40:34 dcordes_, non-android means ? Jul 28 19:40:37 43 Jul 28 19:40:39 ubuntu Jul 28 19:40:40 ahh ok Jul 28 19:40:42 nice Jul 28 19:40:44 wow Jul 28 19:40:47 but it's not relveant. I thought you had same Jul 28 19:40:48 I'll get a nexusone Jul 28 19:40:57 so I'll have the same soon Jul 28 19:41:09 people love ubuntu on the hd2 :) Jul 28 19:41:18 in fact I like it a lot too Jul 28 19:41:19 ok Jul 28 19:41:27 but is it phone-friendly? Jul 28 19:41:35 I mean: Jul 28 19:41:38 *is there some phone apps Jul 28 19:41:44 *is it touchscreen friendly Jul 28 19:41:45 etc... Jul 28 19:41:57 I tuned it a lot to work well with touch only Jul 28 19:42:10 big scrollbars, kinetic scrolling plguin in browser, etc Jul 28 19:42:18 wow Jul 28 19:42:20 phone is very basic: no audio routing yet Jul 28 19:42:25 lol ok Jul 28 19:42:31 alsa? Jul 28 19:42:35 ;( Jul 28 19:42:37 it's qsd right Jul 28 19:42:40 right Jul 28 19:42:40 ah ok no alsa Jul 28 19:42:51 no qsd8xxx-WinCE alsa yet Jul 28 19:42:57 ok Jul 28 19:43:05 oh wait Jul 28 19:43:08 I've rebased my changes on top of msm official android kenrel Jul 28 19:43:14 there is non WinCE AMSS alsa at all - not even for MSM Jul 28 19:43:49 in alsa for dream: is it kernel only thing or you need some userspace workarounds ? Jul 28 19:43:53 basically alsa is just a wrapper arround the ADSP protocol Jul 28 19:44:02 so if you've got some sort of audio.... Jul 28 19:44:04 you can get alsa Jul 28 19:44:11 we have new adsp verison Jul 28 19:44:17 qdsp6 Jul 28 19:44:35 in android we have it working already Jul 28 19:44:40 nice Jul 28 19:44:44 I'll check something Jul 28 19:44:55 Martix: shr-like kernel is at hmmm, i'll place it in few moments. if you want to test - this is not needed now (zub already tested patch). Jul 28 19:45:09 where can I find that rebase kernel ? Jul 28 19:45:25 is kernel htc-msm-2.6.32 ? Jul 28 19:45:25 dcordes_, that'll work Jul 28 19:45:30 I've qsd alsa too Jul 28 19:45:39 qsd alsa ? Jul 28 19:45:40 where Jul 28 19:45:58 in my kenrel Jul 28 19:46:00 gena2x: I like to test it Jul 28 19:46:03 but it's for android Jul 28 19:46:09 alsa for android ? Jul 28 19:46:11 the rebase one Jul 28 19:46:14 yes Jul 28 19:46:22 the current oe one Jul 28 19:46:24 rebase branch Jul 28 19:46:31 gena2x: 242 with WSs is annoying for daily use Jul 28 19:46:35 I've alsa on top of a compilable+bootable+working kernel Jul 28 19:46:38 Martix: ok. Jul 28 19:46:41 Martix: wait a bit Jul 28 19:46:44 ok Jul 28 19:47:00 I've alsa for: Jul 28 19:47:03 msm7k Jul 28 19:47:05 msm7kv2 Jul 28 19:47:08 msm8k Jul 28 19:47:16 I've only tested msm7k tough Jul 28 19:47:18 I must check this right now Jul 28 19:47:25 I'll paste the address Jul 28 19:47:31 thx Jul 28 19:48:15 and it's even in oe Jul 28 19:48:17 Martix: www.bsdmn.com/openmoko/glamo/242/shrkernel_nows/ Jul 28 19:48:56 Martix: i'm not so sure about some aspects of my build env, so please tell me also will sound work for you with this kernel. Jul 28 19:48:58 gena2x: one thing I noticed: now I can boot with usb plugged in; before it would fail to mount root (mmc) Jul 28 19:48:58 http://gitorious.org/htc-msm-2-6-32/leviathan-incoming/trees/android-msm-2.6.32-rebase/sound/soc/msm Jul 28 19:49:15 basically it's the leviathan-incomming Jul 28 19:49:16 but... Jul 28 19:49:17 zub: hm. nice :) Jul 28 19:49:18 although maybe there was some update between that fixed it (?) Jul 28 19:49:27 android-msm-2.6.32-rebase branch Jul 28 19:49:37 zub: you may rebuild without patch and check. Jul 28 19:49:43 gena2x: yup, will try Jul 28 19:50:42 Martix: testing if sound is working is very important for me too. Jul 28 19:50:47 gena2x: ok, thanks Jul 28 19:51:17 Martix: but this is not related to 242 anyhow. Jul 28 19:51:59 Martix: and kernel in link has no patches excpect glamo patch. Jul 28 19:52:47 Weiss: so, just for info to sort out later: zub tried this patch (thanks to him) and found out that ws is gone. Jul 28 19:53:00 good Jul 28 19:53:00 dcordes_, I heard that it would be possible to share kernel between wince and android, at what point is that? Jul 28 19:53:01 gena2x: is this kernel with SHR patches? Jul 28 19:53:07 also on blank/unblank? Jul 28 19:53:40 Weiss: i have no idea how test this in shr. Jul 28 19:54:31 ? Jul 28 19:54:37 Weiss: in debian, i can do for ((;;)); do om screen power 1; sleep 1; om screen power 0; done Jul 28 19:54:48 or even without delay. Jul 28 19:54:49 ok, ignore the ? :) Jul 28 19:55:10 and this will test unblank in very good way. Jul 28 19:56:14 do the shr has lindi's 'omhacks'? Jul 28 19:56:42 gena2x: it's enrico's as well. and yes, it does Jul 28 19:57:00 zub: so, can you test blank/unblank? Jul 28 19:57:23 zub: may be install omhacks and do 'do for ((;;)); do om screen power 1; sleep 1; om screen power 0; done' in bash Jul 28 19:58:06 ok, just now usb networking died somehow... probably my fault Jul 28 19:58:22 i tested it here, and my unblank were ok. Jul 28 19:58:27 zub: depmod? Jul 28 20:01:31 odd, it was present, but not configured, ifdown usb0; ifup usb0 fixed it Jul 28 20:02:32 heh Jul 28 20:02:36 no "omhacks" in the shr feeds Jul 28 20:02:50 but it must be controllable via some /sys/... file, nope? Jul 28 20:02:56 lindi-: how did you named it? :) Jul 28 20:03:01 zub: ioctl :( Jul 28 20:03:16 hm, ok... if there any simple piece of C code? Jul 28 20:03:21 gena2x: collection of openmoko hacks = omhacks? Jul 28 20:03:35 lindi-: i mean for shr :) Jul 28 20:03:40 gena2x: I didn't Jul 28 20:03:44 gena2x: somebody else packaged it Jul 28 20:04:08 lindi-: how to find it? Jul 28 20:04:20 lindi-: i mean package? Jul 28 20:04:22 gena2x: not sure, opkg search|grep omhacks? Jul 28 20:04:31 lindi-: negative Jul 28 20:04:31 I haven't used opkg for ages Jul 28 20:04:40 it doesn't have a manual page and doesn't run as non-root Jul 28 20:04:41 opkg list should list available packages Jul 28 20:04:42 opkg list iirc Jul 28 20:04:44 b0ef: ping Jul 28 20:04:46 but no hacks Jul 28 20:05:13 hm. Jul 28 20:05:25 it's possible to do something similar with some X tool... Jul 28 20:05:46 can't I just grab omhacks srec and compile myself? Jul 28 20:05:53 http://git.debian.org/?p=pkg-fso/omhacks.git;a=tree this? Jul 28 20:06:00 sure Jul 28 20:06:08 but I saw BB recipe Jul 28 20:06:09 maybe I could create OE recipe for that... if that would be useful Jul 28 20:06:11 aha, ok Jul 28 20:06:25 I just don't use BB myself so I don't know if SHR includes it really Jul 28 20:06:26 we have omhacks in our feed Jul 28 20:06:35 lindi-: sure we do :) Jul 28 20:06:41 mrmoku: who coordinates which revision you include? Jul 28 20:06:49 though some things do not work with the new kernel Jul 28 20:06:59 mrmoku: we'd like to support stable releases, now it just seems to get the latest git commit? Jul 28 20:07:12 mrmoku: yes I've fixed things with new kernel Jul 28 20:07:21 so I'd like to know what determines which version you package Jul 28 20:07:28 zub: aha, you can do same thing with 'xset dpms force on' Jul 28 20:07:32 lindi-: let me check what the recipe does... Jul 28 20:07:35 zub: and 'xset dpms force off' Jul 28 20:07:46 openembedded/recipes/openmoko-3rdparty/omhacks_git.bb Jul 28 20:08:01 lindi-: it has e1e41c002148369a5ef61227dadc2cfd42a8491e as fixed rev Jul 28 20:08:19 lindi-: and whenever you ping one of us we can bump it Jul 28 20:08:33 gena2x: so just repeat those two with 1 sec sleep? Jul 28 20:08:39 zub: yeah. Jul 28 20:08:40 mrmoku: you might want to bumb it to 7f316203486acd3ae5902c02f20e59c805f019ab that has 2.6.32 support Jul 28 20:08:49 lindi-: ok, good, thanks :) Jul 28 20:09:00 zub: in old kernel you should sooner of later get ws Jul 28 20:09:01 mrmoku: it's version 0.11 + two commits Jul 28 20:09:17 zub: you may also try without pause. Jul 28 20:09:25 * mrmoku bumps PV too then Jul 28 20:09:42 mrmoku: however, there are still some things that don't work with 2.6.32. for example wifi power control Jul 28 20:10:15 gena2x: should I see the display blank/unblank? as it's not happening :-/ Jul 28 20:10:30 and what is happening? Jul 28 20:10:34 nothing :) Jul 28 20:10:38 gena2x: X won Jul 28 20:10:48 did you do export DISPLAY? ;) Jul 28 20:11:03 export DISPLAY=:0 Jul 28 20:11:11 /etc/profile did that for me :-P Jul 28 20:11:13 it's correct Jul 28 20:11:30 localhost:0 Jul 28 20:11:39 hm... Jul 28 20:11:43 anyway I'll try omhacks Jul 28 20:11:45 localhost:0 is same to :0 Jul 28 20:11:48 ??? Jul 28 20:11:53 should be Jul 28 20:12:06 anyway, if it's in OE, I can easily build omhacks Jul 28 20:12:13 localhost:0 would imply tcp? Jul 28 20:12:14 hm. just never knew that :) Jul 28 20:12:42 zub: just pushed the bump of the rev Jul 28 20:13:00 if [ "$DISPLAY" = "" ] ; then export DISPLAY=localhost:0; fi ... from SHR /etc/profile Jul 28 20:13:10 mrmoku: ok Jul 28 20:13:33 lindi-: seem yes Jul 28 20:14:50 zub: localhost:0 seem wrong. Jul 28 20:15:16 well, it certainly works for other X apps Jul 28 20:15:40 zub: this is wrong in other way Jul 28 20:15:46 yup? Jul 28 20:15:49 zub: it connects via tcp Jul 28 20:16:00 it forces tcp? Jul 28 20:16:05 :0 Jul 28 20:16:08 zub: while :0 (normal) X connection is via unix socket Jul 28 20:16:12 is the normal DISPLAY you want Jul 28 20:16:13 ok Jul 28 20:16:14 GNUtoo|laptop: sorry I got side tracked. Jul 28 20:16:23 np at all Jul 28 20:16:24 me too Jul 28 20:16:33 with a ril and a bug app Jul 28 20:16:42 zub: via tcp may be terrible slower. especially on freerunner. Jul 28 20:18:01 GNUtoo|laptop: all of the required code is in that directory ? Jul 28 20:18:21 no Jul 28 20:18:34 zub: but this not related to blank/unblank. Jul 28 20:18:36 you also need Makefile,Kconfig Jul 28 20:18:58 c297f5c5 (Thomas Zimmermann 2010-04-01 10:18:52 +0200 40) export DISPLAY=localhost:0 Jul 28 20:19:01 don't blmae me :-P Jul 28 20:19:09 blame Jul 28 20:19:29 and you need to have qdsp6 includes in mach-msm but I bet you already have them Jul 28 20:19:31 zub: no blame needed. fix needed :) Jul 28 20:19:36 mrmoku: maybe ^ should be changed to :0 ? Jul 28 20:19:59 well it's openembedded/base-files/base-files/shr/profile, one-liner... mrmoku? :) Jul 28 20:21:14 hmm Jul 28 20:21:43 not sure if that will break stuff via ssh... Jul 28 20:21:48 let me try Jul 28 20:21:54 brb Jul 28 20:22:20 anyway I don't think so... if DISPLAY is set by ssh + x forwarding, then it's not touched Jul 28 20:22:25 this is only if it's unset Jul 28 20:22:40 and then ssh is not involved, even if you log via ssh, it's local display Jul 28 20:23:48 gena2x: building omhacks, finally Jul 28 20:24:52 GNUtoo|laptop: yes sure. I'm compiling entire qdsp6 with every kernel build. Jul 28 20:25:33 GNUtoo|laptop: so sound/soc/msm is entirely created by leviathan and you ? Jul 28 20:27:05 no Jul 28 20:27:05 we just imported and bugfixed codeaurora's alsa Jul 28 20:27:05 note that it uses only the includes Jul 28 20:27:42 mrmoku: tested, ssh w/o forwarding, ssh + X forwarding, reboot so be sure it doesn't break local apps... seems ok Jul 28 20:27:51 gena2x: no WSs and sound is working Jul 28 20:28:04 Martix: very nice. thank you much. Jul 28 20:28:27 blah, I git pull'd and now quite some stuff is being rebuilt :-/ Jul 28 20:28:28 gena2x: good work Jul 28 20:28:32 lindi-: ^^^ sound is working. same kernel. Jul 28 20:28:45 lindi-: just a note. Jul 28 20:29:00 gena2x: which kernel? Jul 28 20:29:55 lindi-: does it matter? here, i have -110 write error, in shr it's working. Jul 28 20:30:07 lindi-: ok, still working on it :) Jul 28 20:30:58 Martix: this is not finished. just pointer for weiss. Jul 28 20:30:59 gena2x, lindi-: aplay works for me (on loudspeaker) Jul 28 20:35:17 same here, sound works with aplay Jul 28 20:36:56 zub: test blanks, your kernel should be ok anyway. Jul 28 20:37:13 Heinervdm: do you remember the DISPLAY problems we once had? and why it is localhost:0 instead of :0 ? Jul 28 20:37:40 zub: ok, thanks for testing :) Jul 28 20:37:47 :0 is standard. Jul 28 20:37:55 GNUtoo|laptop: ok so actually the entire qsd/msm alsa is not very device dependant Jul 28 20:37:57 mrmoku: no, i don't remember Jul 28 20:38:09 Heinervdm: ok, will change to :0 then Jul 28 20:38:09 who invent localhost:0? Jul 28 20:38:34 dcordes_, it's SOC dependant,but it's available for many socs Jul 28 20:38:50 like Jul 28 20:38:55 msm7kv2 Jul 28 20:39:07 msm8k == qsd ==snapdragoon Jul 28 20:39:10 and msm7k Jul 28 20:39:17 may be localhost:0 may be useful in some nfs cases, but on fr? Jul 28 20:41:21 fixed and pushed Jul 28 20:41:26 gena2x: testing, got my screen blinking... let's see what happens Jul 28 20:41:52 zub: main thing - you should always get normal screen at the end. Jul 28 20:44:11 and off to bed... gnight all Jul 28 20:44:52 night Jul 28 20:45:02 gena2x: so far no problems Jul 28 20:45:22 testing without any sleep is ok? Jul 28 20:45:46 hm.. better to try default kernel & timings. Jul 28 20:45:58 repoduce it and then retest with patch. Jul 28 20:46:07 i already did this. Jul 28 20:46:41 ok, then I'd have to revert qi + kernel Jul 28 20:46:46 ok. Jul 28 20:55:37 this time got working qi... it was all by the misguided idea to source environment-setup from toolchain Jul 28 20:56:05 heh Jul 28 20:56:59 gena2x: when WS appears, the display won't blank? Jul 28 20:57:17 zub: no, sometimes it can't restore Jul 28 20:57:40 zub: sometimes it just blank for one iteration. Jul 28 20:57:59 i can't recall exactly current problem. Jul 28 20:58:08 so far nothing; checked kernel, should be the old one Jul 28 20:58:30 check compilation date in dmesg is in doubt Jul 28 20:58:41 hm. may be xrandr? Jul 28 20:58:47 Linux om-gta02 2.6.32.13 #1 PREEMPT Tue Jul 20 13:15:11 CEST 2010 armv4tl GNU/Linux Jul 28 20:58:57 that's the old one Jul 28 20:59:23 true old. even with PREEMPT. Jul 28 20:59:36 why was preempt dropped btw? Jul 28 21:00:24 because dos1 added his opinion to mine and that made jama to agree. Jul 28 21:00:53 that explanation helped :-P Jul 28 21:01:06 may I inquire as for what the opinions were? :) Jul 28 21:01:29 zub: cause i think it is only causing overhead and no benefit. Jul 28 21:01:35 ah, ok Jul 28 21:01:47 cache flush on every context switch Jul 28 21:02:24 but it can be fixed by ARM FCSE Jul 28 21:02:35 btw, i reviewed FCSE Jul 28 21:02:42 FCSE? Jul 28 21:02:44 i mean read more about it. Jul 28 21:02:45 me too Jul 28 21:03:13 and seem opinion that it is something like win31 (where memory is shared) seemed wrong. Jul 28 21:03:49 i didn't came to conclusion, but as a minimum i'm not sure now at all about it. Jul 28 21:03:51 gena2x: that not right, imho Jul 28 21:04:22 Martix: emmm? what not right? Jul 28 21:04:29 gena2x: http://www.google.cz/url?sa=t&source=web&cd=5&ved=0CDUQFjAE&url=http%3A%2F%2Flwn.net%2Fimages%2Fconf%2Frtlws11%2Fpapers%2Fproc%2Fp01.pdf&ei=RZtQTJ6lOMKiOJ_JkfQH&usg=AFQjCNEPVgWOofDB3JHyzgwGliI-9s5j_w&sig2=CaUgamCbbL7K5uPYyYllQw Jul 28 21:04:38 * http://lwn.net/images/conf/rtlws11/papers/proc/p01.pdf Jul 28 21:05:14 yeah, i read that at the time it were published in lwn for the first time. Jul 28 21:05:54 but people here told it's something like making win31 (loosing memory protection) Jul 28 21:06:36 and i thought so too, until i reread info. Jul 28 21:06:47 I dont think so Jul 28 21:07:42 'don't think _so_' means 'don't think memmory protection will be lost' or 'i'm sure it'll be lost'? Jul 28 21:08:30 gena2x: first Jul 28 21:09:27 Martix: yeah. also i thought, why not to do for example 8 (not 92) recent tasks with 128Mb memory. (not 32Mb) Jul 28 21:09:37 and have fast switch between them Jul 28 21:10:15 but without limits. Jul 28 21:10:23 this will aid dbus lovers Jul 28 21:10:50 gena2x: thought? what changed your opinion? Jul 28 21:11:21 wrong word. i meant 'thought' while i were reading :) Jul 28 21:11:27 does other similar arm platforms enable preempt by default in kernel config? Jul 28 21:11:34 gena2x: ok Jul 28 21:12:00 Martix: also HZ. 200 sounds too much. Jul 28 21:12:17 gena2x: no luck with WS so far; how easy/difficult was it for you? Jul 28 21:13:01 zub: hm... i really tested 242. Jul 28 21:13:28 zub, ok. lets finish with it. Jul 28 21:13:32 xrmx: cache flush on every context switch is problem only on armv4 and v5 because they have VIPT caches Jul 28 21:14:03 Weiss: can say nothing about blank/unblack - cant reprobuce unblank problem on 4-4-4. Jul 28 21:14:46 xrmx: sorry, its VIVT Jul 28 21:15:08 zub: thank you much. Jul 28 21:15:29 gena2x: n/p Jul 28 21:15:39 if it makes sense, I can keep it running for few hrs Jul 28 21:15:51 zub: why not. Jul 28 21:15:54 ok Jul 28 21:16:47 Martix: and i have one worst problem in fact :) Jul 28 21:17:05 Martix: i mean 'puzzle' :) Jul 28 21:17:16 Martix: it's about memory read/write speed. Jul 28 21:17:28 Martix: it is very very strange for me on fr. Jul 28 21:18:08 Martix: without cache, read is 80Mb/s write 170Mb/s, memcpy 60Mb/s. Jul 28 21:18:47 xrmx: VIPT is on armv6 and v7, so no need for cache flush after each contex switch Jul 28 21:19:26 gena2x: and with cache? Jul 28 21:19:35 gena2x: but its strange, anyway Jul 28 21:19:52 Martix, ok, but does other armv4 and armv5 platforms enable preempt in defconfig? Jul 28 21:20:02 gena2x: bad SDRAM timings somewhere? Jul 28 21:20:18 Martix: with cache, read is fast (XXX, can't recall exactly), but write still 170Mb/s. but worst thing is on my last tests reading memory (in cache) then writeing something else to same place in cache is faster than just writing. Jul 28 21:20:34 xrmx: dont know, but we can fix it with FCSE Jul 28 21:20:55 Martix: i reviewed all timings 2 times. seem ok. Jul 28 21:21:26 Martix, is fcse uptream? Jul 28 21:21:38 xrmx: Martix: doesn't that require large userspace changes? Jul 28 21:21:38 Martix: so if this is not hw 'feature', this is somewhere else. Jul 28 21:21:57 lindi-: no changes at all. Jul 28 21:23:43 xrmx: no only in Android on FR, but I have contact to one developer, which is working on that, I tested it in SHR and reported some stability issues, but I didnt tested new bugfix version, yet Jul 28 21:23:46 lindi-: nope Jul 28 21:24:25 gena2x: FCSE is HW feature: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0151c/I47491.html Jul 28 21:24:56 gena2x: or you meant RAM reading? Jul 28 21:24:57 Martix: yeah, i've read that. Jul 28 21:25:01 ok Jul 28 21:25:17 Martix: yeah, ram puzzle is different puzzle. Jul 28 21:26:39 xrmx: lindi- I can take a look on FCSE again Jul 28 21:27:04 would be nice i guess. Jul 28 21:27:10 to have it Jul 28 21:27:11 :) Jul 28 21:27:36 but in adjusted version. Jul 28 21:27:39 now when is .32 in distribution, its easier to test Jul 28 21:28:05 heh, i wanted to test it again, but found no patch for my .34 Jul 28 21:28:19 ok. Jul 28 21:28:26 gena2x: imo, FCSE could be in distribution kernel, after some testing Jul 28 21:29:52 it's nice to hear that i undertand correctly about memory protection. Jul 28 21:46:51 which media player do you guys use in openmoko shr ? Jul 28 21:48:51 afaik there's a mplayer module for hw acceleration Jul 28 21:49:15 playya_: ah ok Jul 28 21:49:24 dcordes_: Intone Jul 28 21:49:25 I'm experimenting with mplayer on my device Jul 28 21:49:37 starting it from X. and the problem is only fast way to use it is fbdev Jul 28 21:49:50 playya_: but its very old and non DRM/DRI Jul 28 21:49:52 now when I start video with fbdev from X and the screen is refreshed below it looks ugly Jul 28 21:50:18 any idea to work around it ? Jul 28 21:56:29 no idea Jul 28 22:05:51 hmm. yes. it's pre drm/dri Jul 28 22:26:59 gena2x: build env ready, building .32 with FCSE patch now Jul 28 22:28:28 Martix: heh, i'm going to get my dreams now :) Jul 28 22:28:41 i mean to sleep. Jul 28 22:29:13 Martix: but how you planning to compare? Jul 28 22:29:18 Martix: lmbench? Jul 28 22:29:58 Martix: i can recommend lmbench - it measures context swiches latency. Jul 28 22:37:37 ok Jul 28 22:52:03 gena2x: isn't lmbench like non-free? **** ENDING LOGGING AT Thu Jul 29 02:59:57 2010