**** BEGIN LOGGING AT Tue Oct 28 02:59:57 2008 Oct 28 03:00:21 use led framework Oct 28 03:01:29 ok Oct 28 03:25:49 http://www.startribune.com/lifestyle/33338904.html?elr=KArksUUUU :b Oct 28 03:41:53 Sleep_Walker, shouldnt you be already asleep Oct 28 03:41:54 ? Oct 28 03:42:01 why so? Oct 28 03:42:17 it's 5am Oct 28 03:43:03 not yet :b Oct 28 03:43:24 and what about you? ;) Oct 28 03:49:38 Sleep_Walker, Im sleeping already Oct 28 03:49:51 I see Oct 28 03:50:18 right, I can do more things at once Oct 28 03:51:27 ag Oct 28 03:51:36 Sleep_Walker: sorry, it's that CIA thing being sucky again Oct 28 03:51:47 * BobOfDoom goes to disable it Oct 28 03:52:07 it's really weird how that happens Oct 28 03:52:26 probably has something to do with Farcaller's ACL security stuff Oct 28 03:52:36 I dunno how that stuff all works, I've never played with it much Oct 28 03:53:16 Sleep_Walker: ok you should be able to commit now Oct 28 03:53:35 Sleep_Walker: by Y axis... are you takling about the touchscreen? Oct 28 03:53:38 pushed :) Oct 28 03:53:41 BobOfDoom: yes Oct 28 03:54:08 Sleep_Walker: you'll probably need to figure out the calibration for 680 Oct 28 03:54:26 BobOfDoom: ts_test works perfectly Oct 28 03:54:44 only tslib driver in X behaves this way Oct 28 03:55:06 ah Oct 28 03:55:08 yaeh Oct 28 03:55:13 I had issues like that Oct 28 03:55:20 it behaved weirdly in X Oct 28 03:55:30 when ts_calibrate, ts_test were ok Oct 28 03:55:40 and have you solved that? Oct 28 03:55:44 I ended up just tweaking pointercal manually until I was happy with it Oct 28 03:55:58 maybe there's some other calibration program that could be used Oct 28 03:59:10 oh.. and about ato-linux-2.6.. I was going to use that for anything experimental/untested Oct 28 03:59:42 it should be based off stable Oct 28 04:00:18 BobOfDoom: I understand that Oct 28 04:00:31 git remote add ato hackndev.com:/srv/git/ato/linux-2.6-arm Oct 28 04:00:40 git remote add stable hackndev.com:/srv/git/linux-2.6 Oct 28 04:00:49 I do that Oct 28 04:00:54 then push to whichever one Oct 28 04:00:57 just - you comitted treo680.c directly into linux-2.6 so it wasn't in ato-... Oct 28 04:01:03 right Oct 28 04:01:06 cause I'd tested it :p Oct 28 04:01:36 ok, I probably should have left out the backlight stuff and put that in ato Oct 28 04:01:48 since that was broken Oct 28 04:02:13 but yeah I dunno Oct 28 04:02:24 maybe it's not necessary to have two branches who knows Oct 28 04:02:50 it doesn't matter much, I'd just rather be synced with you since we're working on similar things Oct 28 04:02:56 :) Oct 28 04:03:17 btw. keyboard backlight is probably attached to another PWM, isn't it? Oct 28 04:03:51 GPIO 24 Oct 28 04:03:57 right? Oct 28 04:04:00 that is just on/off Oct 28 04:04:02 but right Oct 28 04:04:05 yeah Oct 28 04:04:14 I think they've done something evil Oct 28 04:04:28 and this explains why the keypad backlight goes off or full brightness when you boot linux Oct 28 04:04:36 I think they've just connected it to some random GPIO Oct 28 04:04:42 and then Palm OS flicks it on and off Oct 28 04:04:45 I'm not sure though Oct 28 04:04:49 heh Oct 28 04:05:07 why not PWM? it will blink better and more precisly :b Oct 28 04:05:24 yes... seems crazy Oct 28 04:05:35 saves CPU time too Oct 28 04:07:33 PWM1 = GPIO 17 Oct 28 04:07:41 whcih they use for camera Oct 28 04:07:58 PWM2 = 11, PWM3 = 12 Oct 28 04:08:09 12 is also used for camera Oct 28 04:08:19 but 11 is some input that they've got an interrupt on Oct 28 04:08:25 they could have used that one at least Oct 28 04:12:04 I copy & pasted backlight code, changed 0 into 2, added GPIO11_PWM2_OUT, MFP_CFG_OUT(GPIO24, AF0, DRIVE_HIGH) Oct 28 04:12:12 but I can't boot Oct 28 04:12:35 (it won't get to framebuffer Oct 28 04:13:01 I'm off Oct 28 04:13:08 gn Oct 28 04:13:13 good luck :) Oct 28 04:16:30 err what Oct 28 04:16:33 don't do that! Oct 28 04:16:40 11 is an input Oct 28 04:16:53 never set an input as an output Oct 28 04:17:00 you could damage the harwdare Oct 28 04:17:23 ;) Oct 28 09:35:49 BobOfDoom: and what about linux-2.6 kernel, which has not even defined all GPIOs? is that danger too? Oct 28 09:38:23 s/danger/dangerous/ Oct 28 10:46:14 no Oct 28 10:46:20 not defined is fine Oct 28 10:46:38 Sleep_Walker: since then it'll default to whatever palmos has set Oct 28 10:46:48 also setting stuff as an input is fine Oct 28 10:46:52 since inputs have high resistance Oct 28 10:47:00 it's just outputs you have to be careful with Oct 28 10:47:26 ok Oct 28 10:47:33 if you have an low output on one chip connected directly to a high output on another chip, there will be a surge of current from one to the other Oct 28 10:48:06 if there's a resistor in between you're probably ok Oct 28 10:48:25 in many cases the chip might detect that there's too much current and just reset Oct 28 10:48:40 but in some cases somthing might be damaged, something overheats or a fuse is blown Oct 28 12:48:40 opie player playing viedeo very slow. MS Mpeg4 Oct 28 12:48:56 fdv_: at least it plays Oct 28 12:49:38 fdv_: where? Oct 28 12:50:01 palm tx Oct 28 12:50:27 шьфпу акщь ьшылф Oct 28 12:50:27 hmm Oct 28 12:50:33 i guess its decoding in software Oct 28 12:50:41 ほんと? Oct 28 12:50:42 image from miska. sorry Oct 28 12:50:45 すごいい!! Oct 28 12:51:06 臭い魚ですよ! Oct 28 12:51:08 :) Oct 28 12:51:18 so, very small picture Oct 28 12:51:25 not real size Oct 28 12:51:48 if its in software - not surprising Oct 28 12:51:52 is the original video high resolution? Oct 28 12:51:56 it take s afair bit of power to decode an mp4 Oct 28 12:52:03 if you transcode it to something smaller it'll probably be faster Oct 28 12:52:13 so it doesn't have to scale it Oct 28 12:52:42 oh, fullsized video about 1-2 fps Oct 28 12:52:46 * |miska| tried it a little bit too, mplayer looks better Oct 28 12:53:36 original video - from webcam: 480x640 15fpsa Oct 28 12:53:38 original video - from webcam: 480x640 15fps Oct 28 12:54:19 oh god Oct 28 12:54:21 no wonder Oct 28 12:54:24 640x480 Oct 28 12:54:29 transcode it to 160x120 Oct 28 12:54:36 and u'll see acceptable framereate Oct 28 12:55:09 160x120? I have screen higher resolution Oct 28 12:58:27 but u dont have the processing power to decode Oct 28 12:58:39 what player are u using? Oct 28 12:59:33 opie player Oct 28 12:59:43 hmm Oct 28 12:59:50 dont know what it uses for its decoder Oct 28 12:59:50 <|miska|> You thing that TX is not enough to decode 480x640@15fps ? Oct 28 12:59:56 likely something not optimal Oct 28 13:00:04 but u pretty much either Oct 28 13:00:07 <|miska|> raster: It uses xinelib ;-) Oct 28 13:00:11 so, I can not stop video Oct 28 13:00:21 1. use another player (mplayer, xine come ot mind as being better optimised - maybe a gstreamer one) Oct 28 13:00:29 and/or u reduce the video qualiyt Oct 28 13:00:33 also avoid scaling Oct 28 13:00:38 |miska|: aaah Oct 28 13:00:52 |miska|: tx is 312mhz right? Oct 28 13:00:56 (Same as treo650) Oct 28 13:00:57 ? Oct 28 13:01:08 <|miska|> I think mplayer seems to be better in performance, but it has some issues with opie Oct 28 13:01:24 <|miska|> raster: I think yes Oct 28 13:01:47 tx has iwmmxt, mplayer is able to utilize that Oct 28 13:01:58 dunno if the codec as well, doubt so Oct 28 13:02:27 thats not a lot Oct 28 13:03:18 try mplayer in console with pxafb overlaying and see the difference Oct 28 13:13:15 mplayer working fast, but I see distortion video Oct 28 13:13:26 with same file Oct 28 13:13:59 and mplayer not switch off screensaver Oct 28 13:15:25 and when I press central button(joystick), it crashed(?) And screen not refreshed Oct 28 13:19:39 so, when palm go to sleep by self, it's hard to wake up it Oct 28 13:21:06 it switch off the screen immediate Oct 28 13:21:26 when I pres @power@ Oct 28 13:22:18 so, bluetooth not configured... Oct 28 13:22:36 so, wifi not working... Oct 28 13:22:41 so ... Oct 28 13:31:11 <|miska|> maybe that mplayer isues can be fixed only with some configuracion... Oct 28 13:32:59 I try Oct 28 13:47:51 opieplayer not undersnding MJPEG Oct 28 13:50:59 and mplayer too with problems Oct 28 13:51:23 What the option "ass" in mplayer config? Oct 28 13:51:29 :) Oct 28 13:57:28 <|miska|> I think it's something related to subtitles Oct 28 13:58:54 config for Zare72 Oct 28 13:58:59 Zire Oct 28 14:22:47 marex: there is some typo in wm97xx_battery.c Oct 28 14:22:52 already known? Oct 28 14:22:54 drivers/power/wm97xx_battery.c: In function ‘wm97xx_bat_probe’: Oct 28 14:22:56 drivers/power/wm97xx_battery.c:182: error: invalid application of ‘sizeof’ to incomplete type ‘struct ucb1400_batt_data’ Oct 28 14:22:57 drivers/power/wm97xx_battery.c:198: error: expected ‘;’ before ‘}’ token Oct 28 14:22:58 make[2]: *** [drivers/power/wm97xx_battery.o] Error 1 Oct 28 14:23:35 (build as module) Oct 28 14:24:23 so, how to stop mplayer during plaing? Oct 28 14:24:39 q? Oct 28 14:25:03 on palmtx Oct 28 14:25:08 :b Oct 28 14:25:11 <|miska|> q Oct 28 14:25:13 :) Oct 28 14:25:13 graffiti? Oct 28 14:25:29 I can not see the keyb Oct 28 14:25:39 * |miska| used onscreen keyboard to stop it Oct 28 14:25:53 Sleep_Walker, that's a driver from mainline ? Oct 28 14:25:59 fdv_: assume qwerty onscreen keyboard Oct 28 14:26:01 marex: http://git.hackndev.com/?p=linux-2.6;a=summary Oct 28 14:26:12 I'm not sure :/ Oct 28 14:27:46 onscreen keyboard not visible at full screen mode and not selectible Oct 28 14:28:22 well if it's not from mainline, I have nothing to do with it Oct 28 14:28:27 use the one from mainline Oct 28 14:29:41 so I have to confirm that is the same file? Oct 28 14:29:51 mmm, ok Oct 28 14:40:29 hmmm, Treo680 is working on 416 MHz :) Oct 28 14:41:35 hm Oct 28 14:46:42 can someone update our linux-2.6 repo to be up-to-date? Oct 28 14:46:52 marex: it differs, probably outdated :/ Oct 28 14:47:35 your problem then Oct 28 14:50:41 marex: is possible to have OUR repo, which will be up-to-date and we can just work against and create patches? Oct 28 14:52:15 <|miska|> Sleep_Walker: I think it is possible, but it will require some cron job to update our kernel... Oct 28 14:52:33 ok, I'll ask farcaller then Oct 28 14:55:45 funny, as others - he's pretty sure that is is possible, but has no experiences Oct 28 15:22:24 http://hackndev.com/node/225 :-) Oct 28 15:29:55 Sleep_Walker: linux-2.6 has 2.6.27 (final) Oct 28 15:30:11 BobOfDoom: and that should be base for us? Oct 28 15:31:04 Sleep_Walker: well.. newer is 2.6.28 (rc-2) but I was gonna to wait for a bit to let things settle down upstream Oct 28 15:31:20 Sleep_Walker: for stable branch, I wanted to base off a stable upstream Oct 28 15:31:26 ook Oct 28 15:31:44 marex: you broke stable! ugly marex Oct 28 15:32:00 broke? Oct 28 15:32:12 oh you mean the battery driver? Oct 28 15:32:18 that might be a bad merge Oct 28 15:32:25 I never tried that driver Oct 28 15:32:27 Sleep_Walker, fuck you ... that wasnt ever in stable Oct 28 15:32:33 ok, taking back for a while :b Oct 28 15:32:39 fine Oct 28 15:32:49 I just applied all marex's patches, so there's probably some conflict with what's upstream Oct 28 15:33:09 BobOfDoom, those arent up-to-date Oct 28 15:33:37 mm Oct 28 15:34:02 we probably should rebase off upstream now that some more of marex's patches have been merged since when I first applied them Oct 28 15:34:30 BobOfDoom, I have some more pending ... I might end up submitting them soon Oct 28 15:35:17 well, I guess merge window is closed, so they won't get applied upstream until 2.6.29? Oct 28 15:35:31 * BobOfDoom is not 100% on how that merge window stuff works though Oct 28 15:38:23 BobOfDoom, under normal curcumstances ... yes Oct 28 15:38:50 alsa is somehow special about this - that fucked up suse guy, Takashi - wont merge it till .30 Oct 28 15:39:10 but that concerns only alsa drivers Oct 28 15:40:15 wtf.. Oct 28 15:40:26 why is it referring to ucv1400 at all Oct 28 15:40:28 that's weird Oct 28 15:40:30 hehe Oct 28 15:42:44 Sleep_Walker: do a s/ucb1400_batt_data/wm97xx_batt_info/ on drivers/power/wm97xx_battery.c Oct 28 15:43:18 that makes sense :b Oct 28 15:44:04 Sleep_Walker: I've never tried that driver though, it couldt be broken in more ways than that Oct 28 15:44:37 I think I'll wait - I consider that as more general problem Oct 28 15:44:47 also, we need to figure out how to talk to the "smart battery" Oct 28 15:44:54 that should get us much better battery measurements Oct 28 15:46:00 can measure the current and temperature with it :D Oct 28 15:46:47 there is a Linux driver, but we need to find out which GPIO it is, which could be hard Oct 28 15:46:56 BobOfDoom: btw. I copy & paste much of your code, will you mind if I let there your credits, or you'd rather me to update it so your name is not near my buggy code? Oct 28 15:47:19 XD Oct 28 15:47:35 I don't really mind too much either way Oct 28 15:47:47 ok Oct 28 15:48:48 680 is looking great. :) Oct 28 15:49:15 more RAM, no stupid ASIC, bluetooth on USB Oct 28 15:49:37 it's my new favourite palm Oct 28 15:51:59 BobOfDoom: you said something like that you made some patched gsmd, do you have it somewhere? Oct 28 15:53:10 it's in openmoko SVN Oct 28 15:53:12 http://svn.openmoko.org/trunk/src/target/gsm/src/gsmd/vendor_bcm.c Oct 28 15:53:20 but they dropped that gsmd project Oct 28 15:53:25 and are working on a different one Oct 28 15:53:44 so I don't think there's much software that works with that Oct 28 15:54:13 what we should start looking at is gsmd2: http://git.freesmartphone.org/?p=gsmd2.git;a=summary Oct 28 15:54:56 there's also frameworkd which is used on openmoko currently: http://git.freesmartphone.org/?p=framework.git;a=summary Oct 28 15:55:16 I hope there is at least something in debian armel Oct 28 15:55:17 which is more fully-featured, but it's writeen in python and uses tons of RAM Oct 28 15:55:43 I've got plenty of RAM and big swap :b Oct 28 15:56:19 well raster and I want to support the 650 too, so we've been looking at gsmd2 Oct 28 15:56:42 no patches yet though Oct 28 15:57:02 and none of them are going to be in Debian Oct 28 15:57:06 they're too early Oct 28 15:57:09 well... Oct 28 15:57:31 actually frameworkd is in Debian's pkg-fso repository: http://pkg-fso.alioth.debian.org/ Oct 28 15:57:36 but not mainline debian yet Oct 28 15:57:55 ooh Oct 28 15:57:59 the other thing I had patches for Oct 28 15:58:03 is gsm0710d Oct 28 15:58:23 which is needed by frameworkd Oct 28 15:58:33 it's a GSM multiplexer Oct 28 15:58:59 (lets multiple communication channels between GSM and CPU.. so one for commands, one for GPRS etc) Oct 28 15:59:10 I didn't get around to patching frameworkd though Oct 28 16:00:51 Sleep_Walker: I've just forwarded you my email to the freesmartphone.org standards mailing list about that Oct 28 16:17:43 ack Oct 28 16:25:22 BobOfDoom: it would be really ugly as part of FFUART initialization code send "AT+CFUN=1", right? :b Oct 28 16:25:35 got treo680_pm_gsm Oct 28 16:41:33 yeah Oct 28 16:41:40 that'd be nasty Oct 28 16:41:46 and there's no hooks to do that in upstream Oct 28 16:41:51 well no obvious ones Oct 28 16:42:50 but I don't think it'll be a problem, both frameowrkd and gsmd2 want to support many different types of modems Oct 28 16:43:02 and that issue is going to affect any device with a BCM21xx Oct 28 16:43:11 (for example there are some USB modems based on it) Oct 28 16:43:31 anyway, I'm off now Oct 28 16:43:32 gn Oct 28 16:43:36 gn :) Oct 28 16:43:40 see you today Oct 28 16:43:46 :) Oct 28 16:47:46 is it clear to add pullup in UDC setup code? Oct 28 16:49:44 hmmm :) GSM works great :) Oct 28 17:26:48 I tried using cocoboot-0.5.2.prc , zImage-k27x-sw1 and debian-lenny-armel-20081004.rootfs.tar.bz2 on my treo 680 and just get it hanging failing to find the rootfs /dev/mmcblk0p2 (which is there and working fine on my linux desktop) I'm guessing the sd drivers aren't in the kernel so I should compile my own? if so is linux-2.6 sane yet for 680 or should stick to linux-hnd? Oct 28 17:28:20 bfree: do you have rootdelay=2 on kernel command line? Oct 28 17:28:48 I had rootdelay=1 initially but tried rootdelay=5 just in case with same result :-( Oct 28 17:28:54 it takes a while SD/MMC to initialize Oct 28 17:28:56 hmm Oct 28 17:29:09 I can give you my hot kernel if interested Oct 28 17:29:50 sure! (nice to have something working to compare against before hacking) Oct 28 17:29:59 fine, gimme a minute Oct 28 17:30:23 no big rush, thanks! Oct 28 17:34:11 http://sleepwalker.hackndev.com/kernel-2.6 Oct 28 17:34:21 enjoy not commited yet GSM Oct 28 17:34:31 hmmm, damn slow upload Oct 28 17:34:39 now it should be whole Oct 28 17:36:17 btw. you were using my kernel before? Oct 28 17:36:20 2.6.21? Oct 28 17:36:32 it's weird that you have problems... Oct 28 17:39:40 bfree: ah! Oct 28 17:39:45 you didn't use inird Oct 28 17:40:04 that was heritage of kEdAR's scripts Oct 28 17:40:05 correct, I didn't use an initrd Oct 28 17:40:15 you probably didn't have ext2 module ;D Oct 28 17:40:45 this should be ok with new kernel I uploaded Oct 28 17:42:03 I was using http://sleepwalker.hackndev.com/kernel/zImage-k27x-sw1 Oct 28 17:42:14 I understand it now Oct 28 17:42:32 sw1 means sleep_walker ^_^ Oct 28 17:45:18 it's certainly getting further with your new kernel :-D Oct 28 17:46:02 well, that is called development :b Oct 28 17:46:09 once you maybe get to command line :b Oct 28 17:46:53 I'm logged in and password changed Oct 28 17:47:10 I thought that! :) Oct 28 17:47:46 last kernel before (m)eat Oct 28 17:49:43 I guess I can't type symbols e.g. "/" with the built in keyboard though Oct 28 17:51:39 bfree: there is no keymap yet Oct 28 17:51:51 but if you'll create one, you will help much Oct 28 17:56:39 gotta go eat somethin, bbl Oct 28 17:59:28 enjoy the food and thanks Oct 28 18:01:41 IrDA just resists for now :( Oct 28 18:30:24 trying to get a network connection to the treo I've connected the usb cable to my (linux) laptop but I just get lots of "usb 4-1: device not accepting address 13, error -71" and can't spot any usb0 device so "ip addr add 192.168.0.200/24 dev usb0" just gives 'cannot find device "usb0"' :-/ do I need to install the modules on the palm/sd or am I missing something else? Oct 28 18:39:23 guess not, the modules.tar.bz has 2.6.21-hnd3k27x-sw1 :-/ Oct 28 19:19:33 I feel much better now :b Oct 28 19:33:40 does anyone know if is Treo680 capable of FIR mode? Oct 28 19:40:27 something is really weir on my system now :( Oct 28 19:41:02 s/wier/wierd/ Oct 28 19:44:25 I found a suggestion that it doesn't do fir (http://www.junefabrics.com/palmnet/faq.php#ir) but not exactly a very reliable source :-/ Oct 28 19:45:00 it can be PalmOS limitation (as usual) Oct 28 19:47:18 omg, mc won't run because it runs `id' and it timeouts :E Oct 28 19:47:24 (on my notebook) Oct 28 19:48:33 omg, LDAP client :E Oct 28 19:50:40 /etc/nsswitch.conf inspired madness perhaps? Oct 28 19:58:37 * |miska| doesn't like LDAP but he likes it more then NIS Oct 29 02:48:21 hi. Oct 29 02:48:35 what wifi chip is used on palmtx? Oct 29 02:59:47 SyChip 6064... which is based on the Marvell 88W8381 **** ENDING LOGGING AT Wed Oct 29 02:59:57 2008