**** BEGIN LOGGING AT Mon Jun 21 02:59:57 2010 Jun 21 04:44:24 GNUtoo|laptop: http://ftp.o2s.ch/log.do_compile.3206 Jun 21 04:45:10 leviathan, file /usr/lib/libgcc_s.so.1 Jun 21 04:45:18 ah Jun 21 04:45:23 ohh Jun 21 04:45:25 /usr Jun 21 04:45:26 yes Jun 21 04:45:29 I guess I know what Jun 21 04:45:32 CROSS_COMPILE_BADNESS Jun 21 04:45:32 :-) Jun 21 04:45:34 mom Jun 21 04:46:13 rm /usr/lib/libgcc_s.so.1 Jun 21 04:46:17 does the trick Jun 21 04:46:26 you're crazy Jun 21 04:46:34 won't that break your system Jun 21 04:46:38 I already thid this several time Jun 21 04:46:46 I remeber now Jun 21 04:46:52 yes but it breaks a lots of programs Jun 21 04:47:09 :-) Jun 21 04:48:05 its gentoo Jun 21 04:48:07 will mickeyl 's suspend work be suspended during the week and continue next week-end? or will it continue? Jun 21 04:48:09 I make the link Jun 21 04:48:14 afterwords again Jun 21 04:48:14 ok Jun 21 04:48:43 GNUtoo|laptop: I do not know if it will continue Jun 21 04:48:54 but I guess Jun 21 04:48:56 it will Jun 21 04:48:57 ok Jun 21 04:48:57 hmm Jun 21 04:49:02 no farnell yet Jun 21 04:49:03 =) Jun 21 04:49:13 like christmas! Jun 21 04:49:26 oh how its written Jun 21 04:49:28 xD Jun 21 04:49:40 there is something strange: Jun 21 04:49:58 public void suspend(){ Jun 21 04:50:10 [...] Jun 21 04:50:24 while (true){ Jun 21 04:50:55 /check for device node Jun 21 04:51:07 if ( ev.code == powerkeycode ){ Jun 21 04:51:17 asert ( debugging message) Jun 21 04:51:29 else{asert(other debugging message)} Jun 21 04:51:38 any of theses 2 debugging messages are printed Jun 21 04:51:57 (I do not see any of them) Jun 21 04:53:55 GNUtoo|laptop: it seems linux-next would have a runnable graphics for dream perhaps Jun 21 04:54:01 as daniel walker wrote me Jun 21 04:54:05 right now Jun 21 04:54:07 wow nice Jun 21 04:54:12 I'm cloning Jun 21 04:54:15 ohh Jun 21 04:54:16 ok Jun 21 04:54:19 and he suggested Jun 21 04:54:21 guess what Jun 21 04:54:29 that we got a tree? Jun 21 04:54:30 tip: I pasted a link Jun 21 04:54:37 http://www.instructables.com/id/Android-G1-Serial-Cable/ Jun 21 04:54:41 that was the link Jun 21 04:54:45 its what he is using Jun 21 04:54:46 ok Jun 21 04:54:46 xD Jun 21 04:55:01 if we go upstream I need this cable too Jun 21 04:55:03 but... Jun 21 04:55:09 I want something that works first Jun 21 04:55:14 upstream is not better yet Jun 21 04:55:25 I know Jun 21 04:55:40 our drivers are so crapy that I'm undecided if I should switch to palm pre Jun 21 04:55:48 =) Jun 21 04:56:03 I had samsung in mind Jun 21 04:56:14 that is to say I wouldn't stop doing dream work Jun 21 04:56:21 but at least I would get an usable phone Jun 21 04:56:26 yes Jun 21 04:56:43 I hope with a clean codebase everything gets better Jun 21 04:57:11 basically it need some work Jun 21 04:57:17 and I'm a bit impacient Jun 21 05:00:37 i'll check linux-next Jun 21 05:00:51 when we get a bootable upstream Jun 21 05:01:01 our parallel development would be nonsense Jun 21 05:01:07 ok Jun 21 05:01:16 but we would still need some things like: Jun 21 05:01:18 wakelocks Jun 21 05:01:20 or similar Jun 21 05:01:22 for now Jun 21 05:01:31 http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=summary Jun 21 05:01:42 in upstream they do things differently Jun 21 05:02:16 just keep that in mind, when designing the software Jun 21 05:02:30 make it good abstracted Jun 21 05:02:49 yes they have pm_qos Jun 21 05:03:00 but I don't think it's in dream's drivers Jun 21 05:03:04 isn't it? Jun 21 05:03:18 I think they simply removed wakelocks Jun 21 05:19:14 yes Jun 21 05:19:19 its not booting Jun 21 05:19:20 either Jun 21 05:19:34 I guess I'll now lie down again Jun 21 05:19:45 after having seen, that the farnel state hasnt changed yet Jun 21 05:19:48 =) Jun 21 08:24:23 JaMa|Wrk: Antonio's ncurses recipe was missing BBCLASSEXTEND :P Jun 21 08:25:18 JaMa|Wrk: still have to verify if nano builds now though Jun 21 08:47:59 mrmoku: ah :/ Jun 21 08:55:12 GNUtoo|laptop: re Jun 21 08:55:24 I fear its taking longer than a day until it comes Jun 21 08:55:32 so I will go now to Stäfa Jun 21 08:55:42 for my ferrit cores Jun 21 08:55:46 and to pusterla Jun 21 08:56:03 buy'n additional parts Jun 21 13:28:31 mickeyl: what does: 'TiCalypsoModem <4C>: Did not receive OK (instead 'NO CARRIER') for FsoGsmV250A' mean? Jun 21 13:38:40 mickeyl, hi Jun 21 13:39:14 mrmoku: that's harmless. NO CARRIER is also an allowed answer to A Jun 21 13:39:17 hi gnu Jun 21 13:40:07 mickey|office, so, is the android usage plugin ready? should I start to test it,because I've a hard time making it work Jun 21 13:40:49 it's ready for testing, no idea whether it's fully correct :) Jun 21 13:41:07 debugging suspend/resume issues is always annoying without a serial cable Jun 21 13:41:20 bbs Jun 21 13:44:06 mickey|office: ok Jun 21 13:44:28 mickey|office: had that in the log after answering a call which vanished then... might just have hung up the other side... Jun 21 13:44:44 in what conditions does it work? Jun 21 13:46:58 basically I can get it working only in that condition: Jun 21 13:47:05 I let it suspend with timeout Jun 21 13:47:11 sometime it works Jun 21 13:47:23 and I can't resume after because fsousaged died Jun 21 13:47:31 that is to say: Jun 21 13:47:33 I press power Jun 21 13:47:40 the display comes back Jun 21 13:47:44 but not the touchscreen Jun 21 13:47:57 I'm using shr-image with xorg Jun 21 13:48:14 given the nightmares you had with x Jun 21 13:48:19 i'd recommend trying without first Jun 21 13:48:23 ok Jun 21 13:48:37 if fsousaged dies, i'd welcome a backtrace Jun 21 13:48:37 I suspend with apm -s then? Jun 21 13:48:39 or at least a log Jun 21 13:48:47 it dies because of assert Jun 21 13:48:52 bbs Jun 21 13:48:54 if you have fso-apm, then yes Jun 21 13:48:58 if not, rather call Jun 21 13:49:03 org.freesmartphone.Usage.Suspend Jun 21 13:49:29 I've it Jun 21 14:01:55 mickey|office, http://pastebin.com/b9Q9ayRf Jun 21 14:02:13 and I think after that it finally suspend Jun 21 14:03:34 I'm back btw Jun 21 14:24:33 did you leave USB in? Jun 21 14:24:40 usually it suspends right after detaching USB Jun 21 14:24:48 it should not try 10 seconds without being able to tsuspend Jun 21 14:25:13 yes usb was in Jun 21 14:25:18 well Jun 21 14:25:24 it does not work that way thanks to android wakelocks Jun 21 14:25:36 I know but what about that use case: Jun 21 14:25:45 fsoraw -r CPU "intone" Jun 21 14:25:54 that would also prevent suspend Jun 21 14:26:03 but the person may want to have fake suspend Jun 21 14:26:12 to save battery life when listenning music Jun 21 14:30:38 mickeyl, well fresh boot + timeout = screen's brightness =0 + shr-lock app Jun 21 14:30:47 I'll wait longer Jun 21 14:31:47 still waiting Jun 21 14:31:48 ... Jun 21 14:32:59 mickeyl, strange strange....I touch the screen and it doesn't resume Jun 21 14:33:11 so maybe it's suspended but with screen on Jun 21 14:36:01 wow I think it's the first time that I got apm -s working.... Jun 21 14:36:11 still under X Jun 21 14:37:15 but resuming has the same issue with fsousaged dying Jun 21 14:37:27 gdb trace would be useless Jun 21 14:37:36 the code is obvous Jun 21 14:37:40 *obvious Jun 21 14:37:54 it's the line with the corresponding error message Jun 21 14:39:41 ha waititng a bit and the mouse is ok again Jun 21 14:39:49 it must have been restarted by something Jun 21 14:46:24 for when it comes back the screen get black/back/black/back and I've the mouse again Jun 21 15:03:16 we're not supporting fake suspend. we're doing the same in userland by disabling the resources and the display Jun 21 15:03:17 mickey|office, I've found a trick for resuming: remove and replug the earphones make the mouse usable again Jun 21 15:03:23 heh Jun 21 15:03:26 sounds like a kernel problem Jun 21 15:03:40 not shure Jun 21 15:03:52 echo mem > /sys/power/state and echo on > /sys/power state Jun 21 15:04:00 works well when I do: Jun 21 15:04:04 /etc/init.d/dbus1 stop Jun 21 15:04:14 s/well/perfectly Jun 21 15:33:43 mickey|office, so what should we do? Jun 21 15:35:08 more testing Jun 21 15:35:41 i have very little time in general Jun 21 15:35:50 only weekends or wednesday Jun 21 16:05:27 ok Jun 21 16:05:42 but as nearly nothing work I don't know where to start Jun 21 16:15:42 mickey|office, zhone2 depends on libelementary-ver-pre-svn-05-0 Jun 21 16:15:48 which is outdated Jun 21 16:15:54 there is the 06 version now Jun 21 16:20:15 automatic dependencies Jun 21 16:20:19 a rebuild will update them Jun 21 16:28:35 mickey|office, so I clean and rebuild zhone2? Jun 21 16:28:50 must be autorev etc...I think Jun 21 16:28:58 yep Jun 21 16:29:06 atorev may have not bumped pr... Jun 21 16:29:12 dependencies that a binary links against are computed automatically Jun 21 16:29:23 I know how it works Jun 21 16:29:36 but I wonder why it linked to 05 instead of 06 Jun 21 16:29:46 it's even in the manual I think Jun 21 16:58:43 <[Rui]> hi Jun 21 16:59:15 <[Rui]> larsc: so, what's new on the latest linux 2.6.34-oe4+gitr4+5c40321e6e6a8ca8552c5c69d041d7f7f3205318-r0-om-gta02 ? :) Jun 21 17:02:08 <[Rui]> anyone else knows? Jun 21 17:15:28 mickeyl, hi Jun 21 17:15:34 I think I'll start with zhone2 Jun 21 17:15:37 and test that Jun 21 17:15:55 larsc, works fine (although glamo-gpio and jbt6k74 have to be built-in in order to be useful, but I'll see how to avoid that later), thanks! Jun 21 17:16:34 ThibG: 2.6.34 has code in the mach file to avoid that Jun 21 17:17:22 I have a few questions, though: in glamo-core, there is disabled code, about power/resume, what should we do with that? Jun 21 17:17:28 hm, ok, will do a diff, then Jun 21 17:17:55 but I should have mostly the same mach file, with the exception of bluetooth, wifi, etc. support Jun 21 17:18:04 hm. that code reinitalises the glamo completely Jun 21 17:18:07 on resume Jun 21 17:18:16 which should imo no necessary Jun 21 17:18:24 not Jun 21 17:18:47 well, the current code reinitialises the glamo completly, and the disabled code does not? Jun 21 17:18:59 wait? Jun 21 17:20:42 the non-disabled code calls external_reset, then, runs the init script Jun 21 17:21:29 hm, right Jun 21 17:22:29 * ThibG is gonna eat, brb Jun 21 17:22:42 larsc: gena@work:~/prg/openmoko/kernel5/linux-2.6$ git checkout origin/om-gta02-2.6.34 Jun 21 17:22:42 M drivers/input/touchscreen/s3c2410_ts.c Jun 21 17:22:42 M drivers/mfd/glamo/glamo-mci.c Jun 21 17:22:53 build_mine GTA02 |tee log.txt Jun 21 17:23:14 /home/gena/prg/openmoko/kernel5/linux-2.6/arch/arm/mach-s3c2440/mach-gta02.c:107:35: error: linux/bq27x00_battery.h: no such file or directory Jun 21 17:23:25 make: *** [sub-make] Error 2 Jun 21 17:23:44 gena@work:~/prg/openmoko/kernel5/linux-2.6$ find .|grep bq27x00_battery.h Jun 21 17:23:44 gena@work:~/prg/openmoko/kernel5/linux-2.6$ Jun 21 17:23:49 ? Jun 21 17:24:10 what did i miss? Jun 21 17:24:40 (this is after fetching latest git) Jun 21 17:26:27 fixed Jun 21 17:27:02 i pushed a commit which should not have been pushed yet Jun 21 17:28:01 thanks, checking.... Jun 21 17:28:05 gena2x: btw. have tried to conntact the upstream authors about the ts issue? Jun 21 17:34:26 larsc: i started preparing long-and-good description, but found mistakes in my logs (i cut them from huge /var/log/messages, were a bit in hurry), so only first log is good, but third is dublicate to 1st and 2nd is impossible to explain, so it's totally possible i fixed some other issue, despite of testing. do i decided to redo investigation to produce better logs and to recheck my understanding of issue, but temporary switched to other task (want to c Jun 21 17:34:26 ompile qt/x11), which i expected to be relatively easy but in fact still not finished :) Jun 21 17:36:43 so, all this is my list. as reproducing issue is not easy, this will take some (more) time... Jun 21 17:39:41 and more testing is not bad anyway :) Jun 21 17:39:42 ok Jun 21 17:40:41 so far i hear no comments about issue. don't know if it is because it is fixed or because noone is using .34 :) Jun 21 17:41:46 (with touchsreen) Jun 21 17:43:00 * ThibG is back Jun 21 17:43:41 so, the commented code was an attempt to reset only a few things on the glamo? Jun 21 17:43:56 ThibG: is think the best would be to try to make it run without reinialising the whole chip Jun 21 17:44:16 ok Jun 21 17:44:37 according to the comment, it might be tricky, though Jun 21 17:45:05 "glamo state on resume is nondeterministic in some fundamental way, it has also been observed that the Glamo reset pin can get asserted by, eg, touching it with a scope probe. So the only answer is to roll with it and force an external reset on the Glamo during resume." Jun 21 17:45:58 * Weiss finds such weirdness with modesetting also.. Jun 21 17:46:20 larsc: kernel build is ok now. Jun 21 17:46:20 currently, with KMS, the first rotation always seems fine, and the rotation back to normal, but then the second rotation breaks it Jun 21 17:48:16 I had a question on glamo-fb, too: in reg_read and reg_write, there are calls to nop(), why that? Jun 21 17:49:18 delay-drived syncronisation :) Jun 21 17:49:26 *driven Jun 21 17:53:18 mickeyl, I've not the code under my eyes but I wonder if you do that: Jun 21 17:53:21 /wait for suspend Jun 21 17:53:30 //suspend Jun 21 17:53:31 well, both glamo-core and glamo-mci read/write without waiting Jun 21 17:53:36 //look for the nodes Jun 21 17:53:51 or do you abort suspend when the nodes like gpio_kp are pressed? Jun 21 17:53:55 I'll look Jun 21 17:53:58 I think I took those nops out for KMS, but I must have put them back in at some point.. probably for a reason Jun 21 17:55:21 ok, thanks Jun 21 17:55:28 imo they can go away Jun 21 17:55:49 well, best way to figure out is to try :) Jun 21 17:56:08 GNUtoo|laptop: i don't really handle the fact when the suspend does not happen at all within 10s Jun 21 17:56:26 mickeyl, ah ok Jun 21 17:56:32 i guess it will trigger resume with resume-reason unknown Jun 21 17:56:35 ThibG: with your define i have 1.3Mb/s Jun 21 17:56:47 ThibG: hm... Jun 21 17:56:48 mickeyl, I'm not shure Jun 21 17:56:55 gena2x, odd Jun 21 17:56:58 btw mickeyl did you saw some screen corruption at resume? Jun 21 17:57:08 ThibG: may be kernel parameter... Jun 21 17:57:33 GNUtoo|laptop: no, never noticed that Jun 21 17:57:34 ? Jun 21 17:57:49 mickeyl, ouch I've it Jun 21 17:57:57 sounds like X Jun 21 17:58:06 mickeyl, I've it under fso2-demo-image Jun 21 17:58:16 ah Jun 21 17:58:21 I've captured the log Jun 21 17:58:32 in that case it could just be a missing full redraw Jun 21 17:58:32 GNUtoo|laptop: did you recieve your cable? Jun 21 17:58:37 I'm still waiting Jun 21 17:58:41 ThibG: (since Glamo will just slow the bus down some more if it needs to. we know it does that.. I don't think it'd even be possible to talk to it too fast) Jun 21 17:58:45 GNUtoo|laptop: btw, hi Jun 21 17:58:46 =) Jun 21 17:58:54 leviathan, my cable? Jun 21 17:59:01 yeah Jun 21 17:59:03 adapter Jun 21 17:59:04 serial Jun 21 17:59:04 I've not ordered yet any cable? Jun 21 17:59:12 oh Jun 21 17:59:16 Weiss, ok, thanks :) Jun 21 17:59:17 I'll do it later Jun 21 17:59:18 then it was dcordes only? Jun 21 17:59:27 I think so Jun 21 17:59:37 dcordes: did YOU get your cable? =) Jun 21 17:59:51 http://pastebin.com/TenJQ0Xt Jun 21 18:00:15 mdp irq already on already on 4 4 + mdp_dma_to_mddi: busy Jun 21 18:00:29 but after when plugging the usb cable it succeed to re-init the screen Jun 21 18:00:49 leviathan, did you test fso2-demo-image? Jun 21 18:00:57 uhm Jun 21 18:01:00 leviathan, do you have framebuffer issues sometimes? Jun 21 18:01:07 uhm, yes Jun 21 18:01:10 ah ok Jun 21 18:01:12 such as? Jun 21 18:01:13 wah, that's a race condition in the display driver Jun 21 18:01:19 indeed Jun 21 18:01:29 I hope, that some of this issues may fall away in linux-next Jun 21 18:01:31 but how to handle that? Jun 21 18:01:42 btw. Jun 21 18:01:49 throw your userland wakelocks away Jun 21 18:01:51 they're not helping :) Jun 21 18:01:52 I think I'll fix that Jun 21 18:01:55 ah? Jun 21 18:02:15 mickeyl, what do you mean? I remove CONFIG_USERLAND_WAKELOCKS? Jun 21 18:02:23 yeah Jun 21 18:02:31 in case you didn't already Jun 21 18:02:31 and everything still works? Jun 21 18:02:37 no I didn't already Jun 21 18:02:40 ah Jun 21 18:02:45 i thought you were finally using the OE configuration Jun 21 18:02:57 I'm using it with small modifications Jun 21 18:03:02 so maybe it's away Jun 21 18:03:05 mickeyl: is it possible to build an initrd with basic shell and dropbear? Jun 21 18:03:42 we do not have nand or working sd support yet in official kernel, as it seems Jun 21 18:03:50 leviathan: i'd think so. needs a custom OE recipe Jun 21 18:03:58 hmm Jun 21 18:04:07 that was my fearing Jun 21 18:04:15 mickeyl, no in oe there is user wakelocks Jun 21 18:04:56 I'll remove it Jun 21 18:05:02 locally Jun 21 18:10:36 i guess the serial port is not attached to the custom USB connector, right? Jun 21 18:47:41 hi Jun 21 18:48:39 i have problems linking a program for the freerunner using shr-unstable bitbaking system Jun 21 18:49:19 I simply added a function to a file contained in a library (my program depends on that library and is in the same project source) Jun 21 18:49:36 and only for that function it gives "undefined reference" Jun 21 18:50:48 mmm... interesting Jun 21 18:50:55 any function I add since now doesn't work Jun 21 18:51:25 doesn't make sense :| Jun 21 18:53:25 ok I'm an idiot :S Jun 21 19:00:58 daniele_athome, use the devshell Jun 21 19:01:24 GNUtoo|laptop, don't worry... just an old library binary inside the shr-unstable staging area Jun 21 19:01:26 :) Jun 21 19:01:33 ok Jun 21 19:01:38 use packaging staging then Jun 21 19:01:40 seems that the library path had the priority Jun 21 19:01:53 *packaged Jun 21 19:01:58 the library is mine GNUtoo|laptop Jun 21 19:02:03 I know Jun 21 19:02:04 is within the same project Jun 21 19:02:10 i use a custom script for speed Jun 21 19:02:19 but I have a recipe too Jun 21 19:02:20 but the old library is the one you're doing ? Jun 21 19:02:32 ok Jun 21 19:02:38 GNUtoo|laptop, yes Jun 21 19:02:41 bitbake -b can also be used for speed Jun 21 19:02:43 ok Jun 21 19:04:00 i know GNUtoo|laptop thanks Jun 21 19:05:01 ok Jun 21 19:06:33 there is also srctree,but maybe you know that too Jun 21 19:06:45 there is also gitver but I didn't got it to work Jun 21 19:09:15 leviathan: regarding the cable: I didn't buy it from farnel. 40 euros is just too much Jun 21 19:09:23 leviathan: what about you ? Jun 21 19:09:44 hi btw Jun 21 19:11:09 dcordes: hi Jun 21 19:11:12 I ordered it Jun 21 19:11:24 my nerves are blank Jun 21 19:11:33 wanna see what linux-next is doing on my dream Jun 21 19:11:34 didn't find an option ? Jun 21 19:11:35 :-) Jun 21 19:11:42 the cable Jun 21 19:11:46 on farnell Jun 21 19:11:53 bought on company's bill Jun 21 19:11:57 seems well enough Jun 21 19:12:02 lol Jun 21 19:12:15 can you buy another one on company's bill ? Jun 21 19:12:27 freesmartphone.org: 03morphis 07msmcomm * rf50e4160422f 10/msmcommd2/ (10 files in 3 dirs): msmcommd2: add first (not compilable) version of msmcommd reimplementation in vala Jun 21 19:12:31 I will send you a postcard with linux-next booting on leo if you will Jun 21 19:12:37 :-) Jun 21 19:13:03 as soon as we got samsung on board for out new project Jun 21 19:13:10 we will need developers anyway Jun 21 19:13:16 so then you can get debug stuff Jun 21 19:13:20 and deployment Jun 21 19:13:37 but we first need too collect all stuff Jun 21 19:13:40 project Jun 21 19:13:43 commitment Jun 21 19:13:51 business case Jun 21 19:14:01 all the stuff needed for a project Jun 21 19:14:14 you could help me get you a deployment? Jun 21 19:14:15 :-) Jun 21 19:14:35 the plan Jun 21 19:14:37 sorry, I don't think I'm following 100% Jun 21 19:14:44 hmm Jun 21 19:14:48 its that Jun 21 19:15:02 we like to do a own platform in cooperation with samsung Jun 21 19:15:07 like the freerunner Jun 21 19:15:11 but better design Jun 21 19:15:17 good drivers Jun 21 19:15:22 support by official kernel Jun 21 19:15:26 SHR on it Jun 21 19:15:31 sounds good Jun 21 19:15:33 ^^^ cool Jun 21 19:15:57 samsung...will you choose a chip with free software 3d? Jun 21 19:15:59 but we need a lot of stuff for this purpose Jun 21 19:16:05 you mentioned on #htc-linux you will propose a project like that to your company. so they liked it ? Jun 21 19:16:06 GNUtoo|laptop: does that exist ? Jun 21 19:16:14 it does Jun 21 19:16:18 ati Jun 21 19:16:21 is on my list Jun 21 19:16:24 cr2 did something Jun 21 19:16:25 free 3d? Jun 21 19:16:35 thats the point Jun 21 19:16:42 I do not have exact information Jun 21 19:16:48 we need to collect a block diagramm Jun 21 19:17:03 all known the be supported by official kernel and available for mobile devices Jun 21 19:17:10 http://gitorious.org/xf86-video-s3c64xx Jun 21 19:17:13 then we need to fill a smartphone block diagramm Jun 21 19:17:54 I'll look if there is free 3d Jun 21 19:17:54 so graphics, wifi, gps, sd, nand Jun 21 19:17:59 thx a lot Jun 21 19:18:16 next step is a nice presentation, thats my part xD Jun 21 19:18:30 * leviathan is joking Jun 21 19:18:37 we can do all together Jun 21 19:18:49 then we sell it to them under o2s-flag Jun 21 19:18:55 and can then deploy developers Jun 21 19:19:13 initial part will be free development Jun 21 19:19:19 misc, ah sorry the driver doesn't have 3d Jun 21 19:19:21 as soon as the first series is out Jun 21 19:19:30 we can pay Jun 21 19:19:31 =) Jun 21 19:19:33 misc, I wonder where I read that there is freee 3d Jun 21 19:19:41 GNUtoo|laptop: dunno Jun 21 19:19:44 but you see Jun 21 19:19:53 its difficult to do such stuff even for us Jun 21 19:19:59 finding free mobile chips Jun 21 19:20:00 leviathan: have fun then Jun 21 19:20:13 dcordes: it would be a oportunity for us all Jun 21 19:20:21 yes HARDware is the HARDest part Jun 21 19:20:25 yes Jun 21 19:20:33 first we need a good part list Jun 21 19:20:44 not too overwealming Jun 21 19:20:49 we need to stay realistic Jun 21 19:21:02 it shall be a business capable smartphone at the end Jun 21 19:21:17 what's better? omap or samsung? Jun 21 19:21:17 thought for managment stuff but also usable for childs Jun 21 19:21:24 duuno Jun 21 19:21:32 mickey||WM: what do you think? Jun 21 19:21:40 you were involved into OM Jun 21 19:21:45 so you have some knowledge Jun 21 19:23:27 * leviathan => DoctorWho: and silence follows Jun 21 19:23:32 ... Jun 21 19:23:42 I'm starting to hate football Jun 21 19:23:43 leviathan: let me know if they have any job openings :) Jun 21 19:23:44 really Jun 21 19:23:53 dcordes: :-) Jun 21 19:24:17 I'll send you a development phone, if it works Jun 21 19:24:18 ok? Jun 21 19:24:19 =) Jun 21 19:24:46 GNUtoo|laptop: ok Jun 21 19:24:53 whats around on supported 3d? Jun 21 19:25:00 ok Jun 21 19:25:01 cr2 did something I guess Jun 21 19:25:11 then next thing would be wifi Jun 21 19:25:21 what wifi chips are available and supported? Jun 21 19:25:24 gps Jun 21 19:25:35 does it the well known chip in OM? Jun 21 19:25:41 UMTS Jun 21 19:26:03 theres no chance to get a free one Jun 21 19:26:05 for sure Jun 21 19:26:07 ? Jun 21 19:26:11 leviathan, hardware costs a lot to produce,specially phones Jun 21 19:26:32 leviathan, did you find some investors? or where the money comes from? Jun 21 19:26:38 O2S Jun 21 19:26:56 is O2S a big company? Jun 21 19:26:59 my dad gets more then a million frank per year Jun 21 19:27:05 he's the owner Jun 21 19:27:07 ok Jun 21 19:27:08 :-) Jun 21 19:27:27 lets say, he doesnt waste a lot Jun 21 19:27:32 if you earn much Jun 21 19:27:34 ok Jun 21 19:27:38 you learn something Jun 21 19:27:42 how to keep the money Jun 21 19:27:43 =) Jun 21 19:27:54 dream session? Jun 21 19:28:10 dream? Jun 21 19:28:15 leviathan, you'll need good hardware engineer Jun 21 19:28:24 yes Jun 21 19:28:27 leviathan: or your serious? Jun 21 19:28:33 maybe talk to qi-hardware guys Jun 21 19:28:35 gena2x: I'm serious Jun 21 19:28:40 I wanna do this Jun 21 19:28:57 gena2x: do you know "the first 20 millions" Jun 21 19:28:59 ? Jun 21 19:29:00 ;-D Jun 21 19:29:17 GNUtoo|laptop: ok Jun 21 19:29:28 * GNUtoo|laptop is thinking about the other unknown free phone under android Jun 21 19:29:42 thinking in the sense of hmmmm Jun 21 19:29:53 hmm? Jun 21 19:30:01 the gizmo4you Jun 21 19:30:12 they were able to produce a working phone Jun 21 19:30:19 geek's phone ? Jun 21 19:30:24 no Jun 21 19:30:52 ah, I see Jun 21 19:31:07 s/4/for Jun 21 19:31:43 I think they are not well known for several reasons: Jun 21 19:31:48 *their phone cost a lot Jun 21 19:31:56 *it runs only android Jun 21 19:32:03 I wanna do it as project on samsung Jun 21 19:32:08 it's FLOW Jun 21 19:32:16 *it's the Jun 21 19:32:26 because they had already the idea to use enlightenment and so on Jun 21 19:32:34 so why not do a right phone for them? Jun 21 19:32:55 we could give them a realistic whis list Jun 21 19:33:10 where already a codebase exists in official branches Jun 21 19:33:27 so we could manage to make it work cleanly and without big hacking Jun 21 19:33:29 http://www.gizmoforyou.net/site/en/shop/gizmos/computing/flow/flow-g1-5.html Jun 21 19:35:49 hmm Jun 21 19:36:21 is samsung ok for power management? Jun 21 19:36:38 btw: I ment the company samsung Jun 21 19:36:40 :-) Jun 21 19:36:44 they also use OMAP Jun 21 19:36:50 but as I know Jun 21 19:36:57 theire smartphones sell Jun 21 19:37:06 ok Jun 21 19:37:11 the issue with omap is 3d Jun 21 19:37:18 yes Jun 21 19:37:19 beside that it seem ok Jun 21 19:38:09 also I talked with morphis Jun 21 19:38:16 ok Jun 21 19:38:19 palm pre will be 100% usable soon Jun 21 19:38:26 very well Jun 21 19:38:32 when mickeyl will implement all in fso Jun 21 19:38:34 I can't wait for my cable =) Jun 21 19:38:36 becuase for now Jun 21 19:38:46 there is only msmcommd Jun 21 19:38:55 basically now there is touchscreen Jun 21 19:39:03 there is phone calls send/receive Jun 21 19:39:14 GNUtoo|laptop: btw touchscreen Jun 21 19:39:30 GNUtoo|laptop: you've written a tslib module, haven't you? Jun 21 19:39:55 no Jun 21 19:40:12 I've hacked the touchscreen driver of the htcdream to make it like tslib Jun 21 19:40:15 in the kenrel Jun 21 19:40:19 not in userspace Jun 21 19:40:21 ah ok Jun 21 19:40:38 basically it was only commenting stuff,changing a value for 0,1 Jun 21 19:40:39 etc... Jun 21 19:40:44 something easy Jun 21 19:40:47 because i've a problem with the tslib module and there is no answer on the tslib ml Jun 21 19:40:59 tslib of the palm pre? Jun 21 19:41:04 yes Jun 21 19:41:06 what's the problem? Jun 21 19:41:17 it doesn't click ;) Jun 21 19:41:23 yes, you can make calls, received calls, and received sms Jun 21 19:41:30 sending sms is already researched, needs to be implemented Jun 21 19:41:31 ok Jun 21 19:41:41 Heinervdm, I know kernel stuff only but I can help Jun 21 19:41:54 what does the python or C program that print event says? Jun 21 19:42:21 don't know as i have no fso image on my pre... Jun 21 19:42:30 Heinervdm, also most important, I've no palm pre yet Jun 21 19:42:43 so it would be difficult to help Jun 21 19:43:01 yes Jun 21 19:43:01 Heinervdm, ok so what do you have? Jun 21 19:43:36 more precisely: Jun 21 19:43:52 what's your setup for talib development? Jun 21 19:44:08 do you have python on the palm pre? Jun 21 19:44:23 no i don't think so Jun 21 19:44:34 GNUtoo|laptop: raster says there's no option to have free 3D drivers for any decent embedded graphics :( Jun 21 19:44:37 ok so I'll look for the C program Jun 21 19:44:58 PaulFertser, I was thinking about sub-decent 3d Jun 21 19:45:15 the problem is that we have the coordinates, and i thought that tslib will do a click automatically when there is a short event Jun 21 19:45:22 but it doesn't do it. Jun 21 19:45:27 http://github.com/playya/tslib/blob/master/plugins/cy8mrln-palmpre.c is the code Jun 21 19:45:40 Heinervdm,ok one sec I look for the test program I used Jun 21 19:45:55 I think it was evtest.c Jun 21 19:47:17 it's even in oe...wow Jun 21 19:48:13 samp->pressure Jun 21 19:48:19 what does it print Jun 21 19:48:24 something like 0,1 Jun 21 19:48:31 or something like 0->255 Jun 21 19:48:49 I was shortly afk Jun 21 19:48:51 Heinervdm, you should really test xorg + illume I think Jun 21 19:48:55 now zenob is gone again... Jun 21 19:49:07 zenob? Jun 21 19:49:12 no patience these people nowadays Jun 21 19:49:14 yeah Jun 21 19:49:44 ah someone in #htc-linux? Jun 21 19:49:52 queried me Jun 21 19:49:54 hmm Jun 21 19:50:00 I'll place my jabber into wiki Jun 21 19:50:07 for future questions Jun 21 19:50:28 GNUtoo|laptop: we have values 0...70 for samp->pressure Jun 21 19:50:55 ok Jun 21 19:51:11 Heinervdm, how are you testing exactly? Jun 21 19:51:33 currently i don't test at all Jun 21 19:51:47 Heinervdm, maybe I should ask my mother to bring me back a palm pre from france? Jun 21 19:51:53 i let others do the tests Jun 21 19:52:00 Heinervdm, ok who is testing then? Jun 21 19:52:13 playya and morphis Jun 21 19:52:19 ok Jun 21 19:52:21 playya_, hi Jun 21 19:52:53 I need to talk to playya_ then Jun 21 19:53:01 and then we merge the changes in your code Jun 21 19:53:05 if we succeed Jun 21 19:53:27 because I had nearly the same issue I think Jun 21 19:53:31 it's playya's repo :) Jun 21 19:53:37 ok Jun 21 19:55:31 Event: time 3088.452287, type 3 (Absolute), code 24 (Pressure), value 0 Jun 21 19:55:35 that's evtest Jun 21 19:55:44 is playya_ here? Jun 21 19:56:22 i have to go now Jun 21 19:56:34 Heinervdm, also I heard that the hardware is fragile, is it true? Jun 21 19:56:34 ok Jun 21 19:57:32 i heard that too, but mine is ok and i already droped it from the table by accident multiple times Jun 21 19:57:55 GNUtoo|laptop, pong Jun 21 20:04:37 GNUtoo|laptop: I'm building a sinus generator right now Jun 21 20:04:41 ahh, and btw Jun 21 20:04:46 I got again about the calcs Jun 21 20:04:50 like I'm doing it Jun 21 20:04:57 I can't blow up my house Jun 21 20:04:57 xD Jun 21 20:05:12 only my room Jun 21 20:07:51 playya_, hi Jun 21 20:07:55 hi Jun 21 20:08:22 playya_, so what's your setup for testing the tslib plugin code? Jun 21 20:08:29 playya_, what evtest says? Jun 21 20:09:06 i don't have a working preboot atm. never used evtest Jun 21 20:09:12 ok Jun 21 20:09:22 I'm stupid,evtest won't work Jun 21 20:09:32 because you have raw values from the dev node Jun 21 20:09:33 is evtest to test the tslib plugin or evdev? Jun 21 20:09:37 yes Jun 21 20:09:40 evdev Jun 21 20:09:53 ok Jun 21 20:10:04 so what does that pring: Jun 21 20:10:07 *print Jun 21 20:10:26 fprintf(stderr,"RAW---------------------------> %d %d %d\n", Jun 21 20:10:26 samp->x, samp->y, samp->pressure) Jun 21 20:10:27 we want to move to a new kernel later, which should have mor default interfaces Jun 21 20:10:33 ok Jun 21 20:10:40 atm my code segfaults :( Jun 21 20:10:50 somewhere in ecore Jun 21 20:10:59 playya_, ??? ecore??? Jun 21 20:11:12 ok, guys. i have a big question about glamo. Jun 21 20:11:15 and mmc Jun 21 20:11:18 playya_, in http://github.com/playya/tslib/blob/master/plugins/cy8mrln-palmpre.c Jun 21 20:11:19 yes Jun 21 20:11:33 i want it not be lost in chat :) Jun 21 20:11:35 what does fprintf(stderr,"RAW---------------------------> %d %d %d\n",samp->x, samp->y, samp->pressure); print Jun 21 20:11:49 specially for pressure Jun 21 20:12:01 basically the following infos are needed: Jun 21 20:12:08 which KEY is pressed Jun 21 20:12:14 by key I mean : Jun 21 20:12:26 so. now. i checked speed of my mmc. i wanted to tune it. Jun 21 20:12:32 and what i see? Jun 21 20:12:54 gena2x: what? Jun 21 20:13:01 i see, that my mmc speed is 1.2 Mb/s Jun 21 20:13:02 BTN_TOUCH ? Jun 21 20:13:08 i can't tune it. Jun 21 20:13:23 PaulFertser: wait a bit, i'll describe it now. Jun 21 20:13:47 ABS_PRESSURE ? Jun 21 20:13:54 basically I report ABS_PRESSURE Jun 21 20:13:59 here's my code: Jun 21 20:14:05 input_report_abs(ts->input_dev, ABS_PRESSURE, (finger?255:0) ); Jun 21 20:14:34 finger was int finger = buf[0] & 7; Jun 21 20:14:46 try that Jun 21 20:14:56 that made it detected by tslib and work under illume Jun 21 20:15:05 but port it to userspace Jun 21 20:15:18 initally i wanted to test ThiG suggestion about setting #define GLAMO_MCI_WORKER in drivers/mfd/glamo/glamo-mci.c Jun 21 20:15:23 as you use userspace plugins Jun 21 20:15:57 so, i expect some diff with MCI_WORKER on and off Jun 21 20:16:08 but i got 1.1 Mb/s in any case Jun 21 20:16:41 then, i did om screen power 0 Jun 21 20:17:05 and what i see? 2.6Mb/s without MCI_WORKER and 2.9Mb/s with. Jun 21 20:17:19 i have no X running. Jun 21 20:17:33 so how did that possible? Jun 21 20:18:12 also, then om screen power 0, i can see that no touchscreen buzz happens. at all. Jun 21 20:18:15 gena2x: with screen blanked glamo mci is more performant Jun 21 20:18:28 As it shares some internal resources Jun 21 20:18:45 touchscreen is not connected to glamo, so why buzz disappears? Jun 21 20:19:02 i think may this things somehow be connected? Jun 21 20:19:07 playya_, ping? Jun 21 20:19:19 buzz? Jun 21 20:19:25 yes. Jun 21 20:19:30 GNUtoo|laptop, i think the pressure isn't correct in the plugin Jun 21 20:19:42 playya_, ok what does it print then? Jun 21 20:19:46 we should set it to a fixed value Jun 21 20:19:55 we have special 'filters' on our touchscreen driver, unique to openmoko to fix that 'buzz' Jun 21 20:20:00 playya_, ok Jun 21 20:20:16 playya_, set it to the max then Jun 21 20:20:26 and 0 otherwise Jun 21 20:20:41 ThibG: random wrong values. more random on long axis. Jun 21 20:20:44 i'm just applying the latest patch to the kernel and it doesn't compile anymore Jun 21 20:20:48 yes Jun 21 20:21:09 gena2x, is that new? Jun 21 20:21:14 but touchscreen is basically ADC, so it just measure voltage Jun 21 20:21:18 gena2x: hm, interference from PWM backlight? Jun 21 20:21:34 gena2x: you can try disabling backlight only, without screen blanking. Jun 21 20:21:45 ThibG: the 'buzz' fact is not new Jun 21 20:21:58 playya_, I'm lost are you modifying the kernel too, or just making an userapace tslib plugin? Jun 21 20:21:59 ok Jun 21 20:22:08 ThibG: i don't know if the fact about it is related to screen state is new Jun 21 20:22:15 PaulFertser: how? Jun 21 20:22:35 gena2x: oh, some pcf50633 foo... let me look it up Jun 21 20:22:37 ATM I use the 1.3.1 patch for the kernel, just try to update it to 1.4.1 Jun 21 20:22:49 that's not affiliated to the touchscreen Jun 21 20:22:51 GNUtoo|laptop: because I'm the only one, who has really ordered such a debug serial Jun 21 20:22:57 i need a working kexec Jun 21 20:22:58 PaulFertser: my very wild guess - may be all our problem somehow related to power source? Jun 21 20:23:07 playya_, ok I have no pre yet so I don't know what the patchset mean Jun 21 20:23:11 GNUtoo|laptop: I'll be the one, who has need to make it usable for development for you Jun 21 20:23:16 right? Jun 21 20:23:24 I mean linux-next Jun 21 20:23:25 leviathan, ? Jun 21 20:23:40 linux-next is unusable for us atm Jun 21 20:23:42 leviathan, I need to talk to mickeyl Jun 21 20:23:46 because not even fbcon starts Jun 21 20:23:52 of course Jun 21 20:23:56 for our plan Jun 21 20:23:59 PaulFertser: i'm on .34 Jun 21 20:24:01 but he ist watching tv Jun 21 20:24:02 you mickeyl and me Jun 21 20:24:05 yes Jun 21 20:24:06 we need to talk Jun 21 20:24:14 we should do something like a video conference Jun 21 20:24:20 do you have skype? ;-D Jun 21 20:24:22 irc is fine Jun 21 20:24:23 no Jun 21 20:24:25 leviathan, GNUtoo|laptop: are you guys trying to get fbconsole on dream going? Jun 21 20:24:26 ok Jun 21 20:24:26 I don't want skype Jun 21 20:24:34 I've ekiga and an asterisk server tough? Jun 21 20:24:34 GNUtoo|laptop: kk Jun 21 20:24:37 Kensan: yes Jun 21 20:24:44 under linux-next Jun 21 20:24:49 means official kernel Jun 21 20:24:54 gena2x: pmu 29=0 to turn it off, =2 to turn it on. Jun 21 20:24:55 one of the projects Jun 21 20:24:59 leviathan: does it work with your own tree? Jun 21 20:25:03 yes Jun 21 20:25:04 gena2x: not kernel dependent. Jun 21 20:25:07 but not with official Jun 21 20:25:16 Kensan: and we have too much divergencies Jun 21 20:25:16 leviathan: so you want to port msm_fb to mainline? Jun 21 20:25:22 there it is Jun 21 20:25:24 already Jun 21 20:25:30 but it doesnt get initialised Jun 21 20:25:32 PaulFertser: sorry, pmu? Jun 21 20:25:38 leviathan, I've an idea: we do a phone conference...like wait a min I'm booting...oops it crashed...let me look in the kernel source I'm here soon....oops what's this codeaurora mess....give me a min.... Jun 21 20:25:46 PaulFertser: how to change it? Jun 21 20:26:00 GNUtoo|laptop: lol Jun 21 20:26:00 ok Jun 21 20:26:01 I see Jun 21 20:26:04 you're right Jun 21 20:26:08 PaulFertser: ( i only viewed pcf doc and kernel code briefly) Jun 21 20:26:19 GNUtoo|laptop: then no conference ;-) Jun 21 20:26:33 leviathan: what are you basing your work on? Jun 21 20:26:36 leviathan: I mean which tree? Jun 21 20:26:42 htc-msm-2.6.32 Jun 21 20:26:47 leviathan, I was just joking on how 3 phone developers make a conference with their unstable phone Jun 21 20:26:47 what is not very cool Jun 21 20:26:49 leviathan: no I mean upstream Jun 21 20:26:54 leviathan, irc is fine and has logs Jun 21 20:26:55 because next has already .35 Jun 21 20:26:59 k Jun 21 20:27:17 mickey||WM, hi ping me when you're back Jun 21 20:27:22 leviathan: linux-next? Jun 21 20:27:28 Kensan: http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=summary Jun 21 20:27:29 yes Jun 21 20:27:33 gena2x: pmu is the Werner's utility Jun 21 20:27:35 hm Jun 21 20:27:42 its the branch which gets merged into torvalds Jun 21 20:27:52 why not walker's msm tree? Jun 21 20:27:55 where all the nice linux-2.*.tar,bz2 come from Jun 21 20:27:59 gena2x: http://svn.openmoko.org/developers/werner/pmu Jun 21 20:28:10 because walkers msm-tree is outdated Jun 21 20:28:14 I assume patches would go via linux-arm-msm end then via daniel to linus Jun 21 20:28:15 and its codeaurora Jun 21 20:28:22 I think Kensan and alain2210 should come to the irc conference too Jun 21 20:28:32 yes Jun 21 20:28:34 how far outdated? He had patches for 2.6.35-rc2 Jun 21 20:28:36 basically everyone who has some interest in the dream Jun 21 20:28:45 k Jun 21 20:28:51 that got pulled by linus Jun 21 20:29:37 GNUtoo|laptop: when I get a couple hours in one piece I will solder my serial cable for the nexus and see how far I get with UART for nexus one. Jun 21 20:29:52 playya_, btw also be shure that you send the correct button Jun 21 20:29:59 GNUtoo|laptop: but I guess we are all trying to push into the same direction. Jun 21 20:30:01 playya_, maybe xev can help to debug? Jun 21 20:30:17 Kensan, ok Jun 21 20:30:45 Kensan, the problem is that I'm starting to become impacient Jun 21 20:31:00 suspend progress is so slow Jun 21 20:31:14 alsa buffer underruns Jun 21 20:31:16 GNUtoo|laptop, nope. we get the raw data from the kernel Jun 21 20:31:21 and wifi doesn't do PSM well Jun 21 20:31:28 playya_, ok Jun 21 20:31:38 then we have no xorg rotation and no xorg keyboard mapping Jun 21 20:31:46 gps could possibily crash the phone Jun 21 20:32:14 bluetooth doesn't work yet Jun 21 20:32:18 GNUtoo|laptop: yes, the whole picture is not pretty. Jun 21 20:32:31 in another hand we've got nearly everything Jun 21 20:32:33 GNUtoo|laptop: so Jun 21 20:32:39 but the nearly everything doesn't work well Jun 21 20:32:42 GNUtoo|laptop: how about a concentrated effort? Jun 21 20:32:46 the 3 main things are: Jun 21 20:32:48 *suspend Jun 21 20:32:50 Kensan: I'll check out codeauroras for-next Jun 21 20:32:50 *wifi Jun 21 20:32:53 *audio Jun 21 20:32:55 perhaps it initialises Jun 21 20:33:18 GNUtoo|laptop: get all devs to have a little codesprint e.g. over a weekend or so or even just one sunday or so Jun 21 20:33:29 PaulFertser: backlight via pmu have no effect on both issues Jun 21 20:33:35 audio: I'm still waiting for Thingol in #alsa Jun 21 20:33:50 GNUtoo|laptop: prioritize issues and everybody works together to solve 1 or 2 issues together Jun 21 20:34:04 wifi: I retried recently => no changes => mail sent to the wifi mailing list Jun 21 20:34:19 and suspend doesn't work as expected Jun 21 20:34:25 moreover I've display issues Jun 21 20:34:30 hmm Jun 21 20:34:33 ok Jun 21 20:34:34 thats the reason Jun 21 20:34:44 why I wanna port our drivers to a clean codebase Jun 21 20:34:46 PaulFertser: i guess touchscreen may measure something important for us Jun 21 20:34:49 gena2x: of course it can have no influence on sd card speed. And well, you basically confirm that jbt refreshing the screen does produce interference on TS. I guess DocScrutinizer can say more. Jun 21 20:34:53 leviathan: https://www.codeaurora.org/gitweb/quic/kernel/?p=dwalker/linux-msm.git;a=summary Jun 21 20:35:00 yes Jun 21 20:35:01 leviathan: this is daniel walker's stuff Jun 21 20:35:14 https://www.codeaurora.org/gitweb/quic/kernel/?p=dwalker/linux-msm.git;a=tree;h=refs/heads/for-next;hb=for-next Jun 21 20:35:39 leviathan: if you do any work you should base it on one of these branches Jun 21 20:35:39 PaulFertser: i tuned your register. i see no effect on both issues. Jun 21 20:35:42 imho Jun 21 20:35:55 PaulFertser: but screen turned black Jun 21 20:36:36 gena2x: thanks for testing Jun 21 20:36:38 oh I got an question from zenob too Jun 21 20:36:58 GNUtoo|laptop: if you can get all the interested people to agree on a plan of action you might get some progress but afaict people are still pulling into different directions Jun 21 20:37:00 PaulFertser: thanks you for ideas... hope this is not last one... Jun 21 20:38:07 PaulFertser: but ok, how glamo can influence mmc speed while X is not running? Jun 21 20:38:19 PaulFertser: bus should be free, yes? Jun 21 20:38:21 gena2x: long anticipated behaviour of r-ts. Jitter may come from both LCD refresh as well as backlight LED stepup converter. That's why Isuggested to do antijitter in kernel rather than in tslib Jun 21 20:38:34 Kensan, ok Jun 21 20:38:47 are we all suscribed to a common mailing list? Jun 21 20:39:12 gena2x: larsc found that some time ago and he's got an explanation :) Jun 21 20:39:19 GNUtoo|laptop: I am probably not subscribed but feel free to add me as cc Jun 21 20:39:21 gena2x: basically that's refresh Jun 21 20:39:30 ok Jun 21 20:39:31 gena2x: to refresh the screen glamo needs to read its internal memory. Jun 21 20:39:39 gena2x: every n ms. Jun 21 20:39:49 gena2x: so it has less bandwidth for other communication Jun 21 20:40:30 DocScrutinizer51: we have backlight off. Jun 21 20:40:35 PaulFertser, it doesn't share the bandwith used for CPU communication, does it? Jun 21 20:41:14 GNUtoo|laptop, 255 is max? Jun 21 20:41:21 DocScrutinizer51: so it is LCD refresh? Jun 21 20:41:29 playya_, on my driver yes but on yours it may be different Jun 21 20:41:38 ThibG: probably it first dumps data from mmc to its own RAM? Jun 21 20:41:53 PaulFertser, yes Jun 21 20:41:55 possible. simply power down the jbt and test Jun 21 20:42:07 DocScrutinizer51: we did this test Jun 21 20:42:18 but in this case, the worker should greatly improve the perfs Jun 21 20:42:21 gena2x: no Jun 21 20:42:27 GNUtoo|laptop: did you give zenob answer? Jun 21 20:42:28 gena2x: we powered down just the backlight Jun 21 20:42:31 gena2x: not jbt Jun 21 20:42:32 as it copies data from glamo to RAM as it comes Jun 21 20:42:39 PaulFertser: oh. Jun 21 20:43:35 and when i do om screen power 0 Jun 21 20:43:46 i turn jbt off? Jun 21 20:43:47 leviathan, he was already gone Jun 21 20:43:57 * gena2x stracing Jun 21 20:44:15 Kensan, leviathan maybe we all suscribe to smartphone-userland? Jun 21 20:44:19 gena2x: you do screen blanking Jun 21 20:44:27 gena2x: that in turn does it, inside the kernel Jun 21 20:44:57 do this turns glamo off? Jun 21 20:45:04 gena2x: it does ret = ioctl(fd, FBIOBLANK, val ? FB_BLANK_UNBLANK : FB_BLANK_POWERDOWN); Jun 21 20:45:18 gena2x: not the whole glamo Jun 21 20:45:40 'LCD refresh'? Jun 21 20:45:50 thatMd be bad - for uSD Jun 21 20:46:09 GNUtoo|laptop: well I'd rather not subscribe to yet another ML but oh well :) Jun 21 20:46:30 Kensan, which mailing list are we all suscribed? Jun 21 20:46:42 anyway for suspend => smartphone-userland Jun 21 20:46:52 ok, i'll try to grep kernel... Jun 21 20:47:10 GNUtoo|laptop: ok, I will subscribe smartphone-userland then. Jun 21 20:47:24 Kensan, I've not posted any post yet there Jun 21 20:47:35 but as mickeyl is always working I think I'll do Jun 21 20:48:05 GNUtoo|laptop: maybe you can do a writeup à la "State of HTC Dream" and propose next steps Jun 21 20:48:08 feeling great today, as another time my contributions were appreciated by a manufacturer sending me a free device :-D Jun 21 20:48:13 ok Jun 21 20:48:15 I'll do it Jun 21 20:48:27 gena2x: to turn off jbt only you need to LCM_3V... Jun 21 20:48:30 DocScrutinizer51: nice... Jun 21 20:48:35 LDO6OUT Jun 21 20:48:48 DocScrutinizer51: wow, what that is? Jun 21 20:49:03 PaulFertser: may you... be more preciese? Jun 21 20:49:07 gena2x, on FB_BLANK_POWERDOWN, glamo output is disabled Jun 21 20:49:42 N900 Jun 21 20:49:47 GNUtoo|laptop: is leviathan subscribed? Jun 21 20:50:08 I don't know Jun 21 20:50:10 but I am Jun 21 20:50:21 ThibG: need to fix comments there both same Jun 21 20:50:25 I'll do Jun 21 20:50:35 how is the formular? Jun 21 20:50:39 where is it? Jun 21 20:50:44 so that I can subscribe Jun 21 20:50:55 http://lists.linuxtogo.org/pipermail/smartphones-userland/ Jun 21 20:50:57 gena2x, hm? Jun 21 20:50:58 thx Jun 21 20:51:29 ThibG: glamo-fb:73 and :88 Jun 21 20:51:51 leviathan: here's the direct link to the form: http://lists.linuxtogo.org/mailman/listinfo/smartphones-userland Jun 21 20:51:54 oh, yeah :) Jun 21 20:52:06 ok Jun 21 20:52:13 did subscribe Jun 21 20:52:21 PaulFertser: are you speaking about pcf regulators? Jun 21 20:52:26 need to smoke,brb Jun 21 20:52:36 leviathan: "need" ;) Jun 21 20:52:43 PaulFertser: what to do to disable jbt? Jun 21 20:52:43 yeah Jun 21 20:52:47 gena2x: yes... i just tried pmu 38=0 Jun 21 20:52:55 gena2x: gives hell of a noise on the screen. Jun 21 20:53:06 gena2x: i guess it's still powered by the other lines. Jun 21 20:53:22 gena2x: so probably disabling it is not as easy to do. Jun 21 20:53:28 wsod :) Jun 21 20:53:33 DocScrutinizer51: from the vendor themselves? :-O Jun 21 20:53:40 no, modern wsol Jun 21 20:53:52 PaulFertser: yup Jun 21 20:53:59 well I am off Jun 21 20:54:17 PaulFertser: i got wsol this way! Jun 21 20:54:18 GNUtoo|laptop: will answer to your mail once it's on the ML Jun 21 20:54:21 DocScrutinizer51: congratulations, that's really cool :) means at least someone is seeing all that hard work there. Jun 21 20:54:29 gena2x: i did =2 and it was back to normal Jun 21 20:54:41 PaulFertser: no, i can't Jun 21 20:54:50 ok Jun 21 20:54:57 GNUtoo|laptop: I think the only way to move forward is to have a concerted effort of everybody involved to fix the few main issues. Jun 21 20:54:59 I'll try something for suspend Jun 21 20:55:03 PaulFertser: may be this is our famous wsol? Jun 21 20:55:13 PaulFertser: well, community council had to push hard Jun 21 20:55:28 talk to you later Jun 21 20:56:07 but the package came from HEL, quim gil's secretary Jun 21 20:56:13 gena2x: unlikely... i'm not sure how to test what causes the interference. Anyway, if it's the jbt refresh, so what? Needs filtering anyway. Jun 21 20:56:54 DocScrutinizer51: i wonder if anybody's doing anything to properly solve that privacy breach issue. Jun 21 20:56:59 Not following it lately. Jun 21 20:57:26 PaulFertser: sync with vsync then Jun 21 20:58:06 DocScrutinizer51: does glamo even have vsync interrupt? i'm afraid not :( Jun 21 20:58:38 (priv breach) prolly not until high court :-/ Jun 21 20:58:38 (priv breach) prolly not until high court :-/ Jun 21 20:59:00 hm, there is some code to sync with vsync... Jun 21 20:59:13 I thought we have Jun 21 20:59:24 but it does polling Jun 21 20:59:40 Hm, so i was right there's no irq ;) Jun 21 20:59:44 larsc: np here Jun 21 20:59:49 yeah, it polls a register which tells what line is displayed Jun 21 20:59:57 needed only after ts event detected Jun 21 21:00:39 PaulFertser: with pmu 38=0, on screen seem shifted but still displays everything Jun 21 21:00:50 PaulFertser: restore with 38=2 Jun 21 21:00:52 detect push on ts, sync to vsync, probe Jun 21 21:01:08 gena2x: ts still has jitter Jun 21 21:01:08 gena2x: ^^ seems to be a damn nice idea Jun 21 21:01:49 gena2x: (screen displays) parasitic currents i guess, so it's still powered. Not any good for a test... Jun 21 21:03:07 PaulFertser: i thought backlight need much current Jun 21 21:03:29 gena2x: i didn't turn off the backlight with that Jun 21 21:03:38 gena2x: those circuits are independent. Jun 21 21:07:54 PaulFertser: you need to disable glamo interface to jbt. otherwos reverse feed Jun 21 21:08:10 ask weiss Jun 21 21:08:22 or larsc Jun 21 21:09:13 DocScrutinizer51: sure Jun 21 21:09:39 mjor cause of WSOD Jun 21 21:10:10 no, it seem restores on loading of X11 Jun 21 21:12:04 * PaulFertser autohibernates. Have a nice evening folks :) Jun 21 21:12:24 oh... Jun 21 21:12:31 don't leave... Jun 21 21:13:38 with userspace wakelocks removed I do not see any screen coruption.... Jun 21 21:17:16 re, btw Jun 21 21:17:23 needed something with sugar Jun 21 21:17:30 so longer afk Jun 21 21:17:35 searching the kitchen Jun 21 21:17:36 =) Jun 21 21:17:42 ehlo Jun 21 21:17:55 DocScrutinizer51: ehlo leviathan.localdomain Jun 21 21:18:04 =) Jun 21 21:20:07 any usb crack here? Jun 21 21:20:39 Husb kernel modules mainly) Jun 21 21:20:45 hm... did om screen 0 , then ./pmu 38=2 and ./pmu 29=2 (got white screen), then some dd and then om screen 1. not i hace glamo-mci-read process eating 99% cpu Jun 21 21:20:53 s/H/(/ Jun 21 21:20:53 DocScrutinizer51 meant: (usb kernel modules mainly) Jun 21 21:21:02 s/hace/have/ Jun 21 21:21:03 gena2x meant: hm... did om screen 0 , then ./pmu 38=2 and ./pmu 29=2 (got white screen), then some dd and then om screen 1. not i have glamo-mci-read process eating 99% cpu Jun 21 21:21:49 gena2x, glamo-mci-read is the worker Jun 21 21:21:58 DocScrutinizer51: hmm Jun 21 21:22:01 it copies from the glamo to RAM in a busy loop Jun 21 21:22:09 (the same is done later, without the worker) Jun 21 21:22:13 DocScrutinizer51: which platform? Jun 21 21:22:30 GNUtoo|laptop: seems, as mickey will not come back this evening Jun 21 21:22:46 perhaps his group won Jun 21 21:22:50 and hes making party Jun 21 21:22:52 ah Jun 21 21:22:57 ThibG: lets try reproduce this... Jun 21 21:22:57 doesn't matter first instance Jun 21 21:23:01 musb-usb.c Jun 21 21:23:09 * GNUtoo|laptop would like partying when the htcdream would work Jun 21 21:23:15 ThibG: need reboot as system is complete unresponsible Jun 21 21:23:37 GNUtoo|laptop, tslib_test reports some good results: Jun 21 21:23:39 946686125.641021: 188 427 38 Jun 21 21:23:39 946686125.657928: 198 392 49 Jun 21 21:23:39 946686125.674835: 203 377 45 Jun 21 21:23:39 946686125.691772: 207 337 51 Jun 21 21:23:39 946686125.708679: 213 287 39 Jun 21 21:23:42 leviathan: won? Jun 21 21:23:46 without worker: glamo reads from µSD, then sends an IRQ saying the data is available, the IRQ handler copies the data Jun 21 21:23:53 playya_, ok Jun 21 21:23:58 DocScrutinizer51: football Jun 21 21:24:03 with worker: the worker polls the glamo and copies the data as it comes Jun 21 21:24:34 playya_, does it work under Xorg? Jun 21 21:24:39 leviathan: I need somebody who knows his way around usb kernel drivers Jun 21 21:24:52 I don't use X Jun 21 21:24:59 leviathan: I know the particular details of platform Jun 21 21:25:19 ThibG: irq sounds better Jun 21 21:25:40 playya_, ok Jun 21 21:25:50 so, the fact that the worker eats a lot of CPU isn't a problem Jun 21 21:25:59 but it shouldn't run for ages... Jun 21 21:26:27 ThibG: saw this :) Jun 21 21:26:40 ThibG: now i'll try to reproduce Jun 21 21:26:45 ok Jun 21 21:27:52 damn fsck... Jun 21 21:29:58 playya_, try that: Jun 21 21:30:11 ts_test: click on draw Jun 21 21:30:15 and see if it takes the click Jun 21 21:30:35 leviathan: maybe you want to visit h-e-n.garage.maemo.org Jun 21 21:31:01 for the details tty in a minute Jun 21 21:35:50 DocScrutinizer51: interested to know about the N900 by the way... I'm thinking of getting one.. Jun 21 21:36:41 ThibG: all ok, can't reproduce Jun 21 21:37:11 hm ok Jun 21 21:37:45 well, the worker isn't enabled by default, after all :) Jun 21 21:38:03 ThibG: may be problem were not with worker Jun 21 21:38:39 GNUtoo|laptop, it seems that my screen isn't working atm Jun 21 21:38:49 DocScrutinizer51: hmm, ok Jun 21 21:38:58 playya_, what do you mean by isn't working? Jun 21 21:39:11 unless you've corrupted kernel memory or made the glamo crash Jun 21 21:39:17 I've got a cosinus function generator in front of me Jun 21 21:39:18 it stays black Jun 21 21:39:26 thats the reason why I do not answer that much Jun 21 21:39:33 but cat /dev/urandom > /dev/fb0 works Jun 21 21:39:34 ah ok Jun 21 21:39:38 breadboard and keyboard a concurating Jun 21 21:39:39 :-) Jun 21 21:39:48 I don't know omapfb well Jun 21 21:39:53 cat /dev/urandom > /dev/fb0 need to try that \o/ Jun 21 21:40:23 http://www.opentom.org/Tslib got all the variables Jun 21 21:40:29 if you need to change variables Jun 21 21:42:53 ThibG: ueah, measures bandwidth :) Jun 21 21:43:01 Weiss: you'll love it Jun 21 21:43:18 the last few closed blobs are the ones I'm working on Jun 21 21:43:44 stskeeps is working on the aequiv of what you did for glamo Jun 21 21:44:02 gena2x? Jun 21 21:44:17 DocScrutinizer: DRI2/KMS? It's an evil imgtech chip IIRC? Jun 21 21:44:23 ThibG: redraw speed :) Jun 21 21:44:26 leviathan: lemme quickly toss you a ref to a src Jun 21 21:44:37 gena2x, oh, that could be interesting :p Jun 21 21:44:43 Weiss: it's OMAP3430 Jun 21 21:45:20 ah yeah, SGX Jun 21 21:46:34 interesting, how framebuffer contents delivered to glamo? Jun 21 21:47:36 FYI: I'm planning to submit the glamo and jbt drivers by the end of the week Jun 21 21:48:06 guys, how? Jun 21 21:48:25 ThibG: submit to debian ? Jun 21 21:48:26 leviathan: finally got the order done. I found it in some special context. They are also used for ardupilot programming Jun 21 21:48:44 misc, no, to kernel.org Jun 21 21:49:16 ThibG: wow, nice Jun 21 21:51:52 <[Rui]> hi all Jun 21 21:51:58 <[Rui]> larsc: ping? Jun 21 21:52:41 hmm at first boot it seem that I've the thomas white screen of death Jun 21 21:53:01 with 2.6.32-oe3 Jun 21 21:53:07 meaning, the white screen of life? Jun 21 21:53:20 yes Jun 21 21:53:35 ah ok it panicked Jun 21 21:53:37 I'll look Jun 21 21:53:38 what're those strange names? Jun 21 21:53:55 [Rui]: pong Jun 21 21:54:09 <[Rui]> larsc: hi, saw my question from earlier today? Jun 21 21:54:11 leviathan: http://mxr.maemo.org/fremantle/source/kernel/drivers/usb/musb/musb_core.c and relatives, http://h-e-n.garage.maemo.org, and emantle/source/kernel/drivers/usb/musb/musb_core.c and relatives, http://h-e-n.garage.maemo.org, and http://talk.maemo.org/showthread.php?p=696115#post696115 http://talk.maemo.org/showthread.php?p=685367#post685367 http://talk.maemo.org/showthread.php?p=685914#post685914 Jun 21 21:54:14 it says "# All kernels here (except 2.6.34-oe1 now) have DRM patches from Thomas White (How DRI and DRM Work)" Jun 21 21:54:21 so I made a joke about it Jun 21 21:54:27 :) Jun 21 21:54:31 damn, sorry Jun 21 21:55:13 <[Rui]> larsc: «so, what's new on the latest linux 2.6.34-oe4+gitr4+5c40321e6e6a8ca8552c5c69d041d7f7f3205318-r0-om-gta02 ? :)» Jun 21 21:55:23 anyway, away for another drink Jun 21 21:55:34 [Rui]: nothing I guess Jun 21 21:56:35 <[Rui]> larsc: oh :( here I was a-hoping things like the White Screen on idle blank got fixed :( Jun 21 21:57:29 GNUtoo|laptop: I hope tomorrow is solding time Jun 21 21:57:31 :-) Jun 21 21:57:36 ok Jun 21 21:57:45 anyway, my wafeform generator now works properly Jun 21 21:57:45 ah this time it booted fine Jun 21 21:58:12 tomorrow I'll try to combine it with the OpAmp and drive my testing generator Jun 21 21:58:13 and no space left on device Jun 21 21:58:27 "kein weltraum links auf gerät" Jun 21 21:58:30 *gg* Jun 21 21:58:33 ok Jun 21 21:58:37 I should sleep Jun 21 21:59:06 need to go to a costumer tomorrow afternoon Jun 21 21:59:10 ok Jun 21 21:59:16 today afternoon Jun 21 21:59:17 -_- Jun 21 21:59:18 gn8 Jun 21 21:59:24 * leviathan goes sleeping Jun 21 22:00:49 night Jun 21 22:06:13 larsc, hm, one of my patch is not good Jun 21 22:06:36 which one? Jun 21 22:06:50 the last one Jun 21 22:07:04 ok Jun 21 22:07:37 that's not "iounmap(fbinfo->screen_base);" but "iounmap(glamofb->fb->screen_base);" Jun 21 22:08:58 will resend a patch for that Jun 21 23:43:39 freesmartphone.org: 03mickey 07cornucopia * r6a3d8755af8f 10/fsousaged/src/plugins/lowlevel_android/plugin.vala: fsousaged: don't call glib's error(), but just emit a warning, if we can't suspend Jun 22 00:55:40 leviathan: what is it you're after?? o.O **** ENDING LOGGING AT Tue Jun 22 02:59:56 2010