| 00:42.39 | lkcl | no, iclem - nobody here :) |
| 00:44.40 | iclem | Ok Mr nobody, i've some dumb questions if you had time |
| 00:46.04 | iclem | when a gpio is reported to be "O" output, is it intended to be written or read ? |
| 00:46.20 | lkcl | output indicates to be read. |
| 00:46.37 | lkcl | let me think some more... |
| 00:46.38 | lkcl | nope. |
| 00:47.00 | iclem | not obvious |
| 00:47.04 | lkcl | O indicates that it is an output, therefore you write to it to change the output. |
| 00:47.05 | lkcl | sorry. |
| 00:47.13 | iclem | ok thanks |
| 00:47.34 | lkcl | am still struggling through this damn gcc 4.0.2 build. |
| 00:47.59 | lkcl | i'm wondering if i should make sure gcc 3.4.x is used and trash the compile of the past DAY damnit. |
| 00:48.04 | lkcl | anyway. |
| 00:48.08 | lkcl | back to bed :) |
| 00:48.51 | lkcl | oh - btw: if you wanted something to experiment with: set .bus_shift = 1 in himalaya.c because the asic3 is used in the same way as the hx4700 |
| 00:49.10 | iclem | ok, i will try |
| 00:49.11 | lkcl | as things stand, it still doesn't work - the lcd goes completely awol. |
| 00:49.47 | lkcl | but both the hx4700 and the himalaya are "normal" asic3s (rather than 1/2-width register addressing) |
| 00:50.04 | lkcl | and the hx4700 has bus_shift = 1. |
| 00:50.42 | lkcl | asic3_mmc gets a bit unhappy - there's something still quite wrong, there :) |
| 00:51.20 | iclem | nothing sure about the irq to be used |
| 00:52.06 | iclem | could this interact with lcd or ts ? |
| 00:56.43 | iclem | blueangel.c has bus_shift=1 too, and uses packed addressing |
| 01:00.52 | iclem | going to bed, good night |
| 01:34.07 | lkcl | hmmm - i think you're right. then there's a bug in the hx4700 code. |
| 01:34.13 | lkcl | bus_shift = 0 it is. darn. |
| 02:44.10 | *** join/#htc-blueangel pof (n=pof@173.red-62-57-1.user.auna.net) |
| 02:58.46 | lkcl | what a waste of 24 hours. had to trash the gpe-image build after discovering that someone's added gcc 4.0.2 as a preferred "default" to the bitbake files. i've put gcc 3.4.4 into the blueangel.conf and hopefully that should stop kernel compile errors, c++ compile errors... dang, it's _awful_ :) |
| 03:06.43 | pof | today at 22c3 (berlin) a xanadux developer (can't remember the name) gave a short talk about linux on blueangel devices |
| 03:07.41 | pof | I have a PDA2k, and I want to give it a try... i'm reading the docs now, completly newbie :P |
| 03:12.29 | lkcl | heya pof. |
| 03:12.31 | lkcl | cool. |
| 03:13.21 | lkcl | well, it's 3:12am, GMT, here - there are only a few of us around. yes. it was martin (he's known as cp15 when he appears here). |
| 03:13.56 | lkcl | do take a look at the IRC logs, on http://ibot.rikers.org - #htc-blueangel. |
| 03:14.31 | lkcl | at least two or three people have got it going, recently, in the past couple months. |
| 03:14.57 | lkcl | if you find any mistakes in the wiki, or believe it should be better put, do just log in to the wiki and update it :) |
| 03:16.06 | lkcl | anyway: i'm just off back to sleep (again) but will likely be up again: i'm keeping my irc client running so do ask a question but don't be surprised if the answer comes back a few hours later :) or if someone else answers |
| 03:16.47 | rob_w | lkcl i will put back the code from before 10 days ... fed up with it |
| 03:18.53 | lkcl | rob_w - surprise :) |
| 03:19.20 | lkcl | cvs update -d -P -D "10 days ago" - but you knew that already |
| 03:19.33 | lkcl | at least you'll have an 18 hour battery life. |
| 03:20.21 | lkcl | anyway. bitbake rebuild underway, with damn damn damn gcc 4.0.2 using gcc 3.4.4 specified in the blueangel.conf file as the preferred c compiler. |
| 03:26.49 | rob_w | lkcl, it did bake for today with a pull from 2 days |
| 03:27.35 | rob_w | i wonder if it is enough to only put back the lcd code |
| 03:32.13 | pof | lkcl: thanks for the info... I was away... I'm going to sleep too! |
| 03:33.15 | rob_w | i am going to cebit this year ,,, maybe we get a chance to it there again with a working projector setup |
| 03:33.20 | rob_w | sleeping again |
| 06:12.13 | *** join/#htc-blueangel rob_w (n=bob@p85.212.148.147.tisdip.tiscali.de) |
| 06:12.13 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 07:47.13 | *** join/#htc-blueangel rob_w-cgi (n=txuser@dd1010.kasserver.com) |
| 07:47.13 | *** mode/#htc-blueangel [+o rob_w-cgi] by ChanServ |
| 08:51.11 | *** join/#htc-blueangel apt_ (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
| 08:51.11 | *** topic/#htc-blueangel is HTC BlueAngel is still under development ! Feel free to ask and check also -> http://www.handhelds.org/moin/moin.cgi/BlueAngel | http://wiki.xda-developers.com/index.php?pagename=BlueangelResearch | http://gnulinux.biz/files/blueangel/sd/linux/ channel chat logs at http://ibot.rikers.org/%23htc-blueangel/ **upcoming gpe release is in org.openembedded.oz354fam083 dev tree !!!!*** |
| 09:58.02 | rob_w-cgi | ok , so for another boot |
| 10:17.18 | rob_w-cgi | hmm now tux went away but screen stayed black |
| 10:34.41 | *** join/#htc-blueangel _bob_w (n=bob@p85.212.195.56.tisdip.tiscali.de) |
| 10:35.28 | *** join/#htc-blueangel pH5 (n=ph5@p5485F5E5.dip.t-dialin.net) |
| 10:56.04 | *** kick/#htc-blueangel [_bob_w!n=txuser@dd1010.kasserver.com] by rob_w-cgi (rob_w-cgi) |
| 11:17.57 | *** join/#htc-blueangel machinagod (n=jsad@195-23-156-188.net.novis.pt) |
| 11:18.11 | machinagod | hi to all |
| 11:20.58 | *** join/#htc-blueangel iclem (n=clement@iclem.net) |
| 11:22.15 | machinagod | anyone around here is researching Universal at the moment? |
| 11:22.55 | pH5 | hi machinagod |
| 11:23.01 | pH5 | i know at least cr2 is |
| 11:27.52 | *** join/#htc-blueangel makinadedios (n=jsad@195-23-156-188.net.novis.pt) |
| 12:18.55 | rob_w-cgi | ok now have all gpio`s in the pre-init i got a yellow looking screen .. still had 30 ms delay between each call .. will try next without delay |
| 12:19.01 | rob_w-cgi | as this was the origianl setup |
| 12:24.42 | lkcl | yeh, machinagod - i am. sort-of. |
| 12:26.46 | lkcl | DAMNIT now gcc 3.4.4-r3 fails to compile xorg with an internal compiler error |
| 12:30.50 | rob_w-cgi | ok having all gpio is a row 1-5 without delay also gives a blank screen |
| 12:31.15 | rob_w-cgi | *thinks that he waits for lkcl to fix that |
| 12:33.58 | rob_w-cgi | ur compiling xorg `? |
| 12:41.05 | lkcl | bitbake gpe-image is compiling xorg... *sigh* |
| 12:41.29 | lkcl | i've been trying for 2 days to get a damn gpe-image, rob! sorry! |
| 12:41.35 | lkcl | okay. |
| 12:42.08 | lkcl | so you switched off all four GPIOs, then switched them all back on again, and when you did, you got a blank screen? |
| 12:44.07 | rob_w-cgi | i did not use zero them ... |
| 12:44.18 | lkcl | and the "switch back on again" stuff is identical to what was there ten days ago? |
| 12:44.35 | rob_w-cgi | well lets see .. |
| 12:44.54 | lkcl | it might be worthwhile deliberately doing a test like that. |
| 12:45.15 | lkcl | switching off all four GPIOs, delay of 500ms, then go through the power-up sequence. |
| 12:45.27 | rob_w-cgi | ahm .. the crap is i am now unsure what kernel i was booting when it worked at least to use it without suspend |
| 12:45.45 | lkcl | if that works, then there's a reasonable chance that the problems being experienced on resume is NOT anything to do with the LCD/ATI initialisation sequence. |
| 12:46.15 | rob_w-cgi | lkcl : tell me : first lcd_init is called .. then lcd_set_power |
| 12:46.30 | rob_w-cgi | but there you commented out the lcd_init_pre |
| 12:47.10 | rob_w-cgi | i wanna find where the lcd_pre is called ?!? |
| 12:47.20 | lkcl | in drivers/video/w100fb.c |
| 12:47.40 | rob_w-cgi | oh |
| 12:47.41 | lkcl | it's a higher-order-function table. |
| 12:48.00 | lkcl | lcd_struct->resume_pre(dev) |
| 12:48.05 | lkcl | lcd_struct->resume_post(dev) |
| 12:48.12 | lkcl | something like that. |
| 12:48.45 | rob_w-cgi | ok, cool .. so now iam also sure that the code from Dec 10 was used on the gnulinuz.biz kernel from dec. 11 |
| 12:49.29 | rob_w-cgi | so it seems that that code works for normal boot up but "fails" on resume .. your split ..seems not working for me at all |
| 12:50.30 | rob_w-cgi | ok let me build a Dec10 code and see if that still works for me |
| 12:51.09 | lkcl | yes. that's why i suggested, with the "old" code, to put in zeroing-out of GPIO. *cut out* the resume issue COMPLETELY. |
| 12:51.56 | lkcl | don't bother with the resume thing right now: just get booting right, but with zeroing-out the four GPIO lines first plus a delay of 500ms before initialising them. |
| 12:52.15 | rob_w-cgi | so you say zero all .. then 500 then all on ? |
| 12:57.16 | lkcl | yep. |
| 12:57.36 | lkcl | and don't bother testing the resume thing. |
| 12:58.17 | lkcl | if that _works_ then we can reasonably conclude that the LCD/w100fb initialisation is _not_ at fault. |
| 12:58.35 | lkcl | and that there is something else - somewhere in the resume code or the resume process - that's the problem. |
| 12:59.00 | lkcl | e.g. the GPDR, GAFR or the ASIC3 or some PXA registers somewhere aren't being restored correctly. |
| 12:59.04 | rob_w-cgi | i also need to roll back the w100fb.c .h |
| 12:59.15 | lkcl | probably, yes. |
| 12:59.43 | lkcl | i wouldn't, personally. but then, i can't test this because i don't have a rev 5 board. |
| 13:00.16 | lkcl | if i were you, i would simply have nothing in the "post" side of the if (board_id != 6) |
| 13:00.25 | lkcl | and have everything in the "pre" side. |
| 13:01.33 | rob_w-cgi | well no i do that .. that makes me more confident .. |
| 13:05.54 | rob_w-cgi | ok screwd the files .. .will revert and use the post only way ;-) |
| 13:08.07 | lkcl | :) |
| 13:11.25 | rob_w-cgi | let me switch to my linux box to better paste code in here |
| 13:11.30 | *** part/#htc-blueangel rob_w-cgi (n=txuser@dd1010.kasserver.com) |
| 13:13.09 | *** join/#htc-blueangel rob_w-cgi (n=txuser@dd1010.kasserver.com) |
| 13:13.10 | *** mode/#htc-blueangel [+o rob_w-cgi] by ChanServ |
| 13:13.16 | rob_w-cgi | re |
| 13:13.58 | rob_w-cgi | http://pastebin.com/484093 |
| 13:14.02 | rob_w-cgi | is my next try |
| 13:21.06 | rob_w-cgi | hmm again this yellos sscreeenn |
| 13:30.01 | rob_w-cgi | blueangel_lcd.c:210: error: unknown field `resume' specified in initializer |
| 13:44.24 | rob_w-cgi | ok i am out of play time now .. need to run .. cu |
| 13:50.10 | lkcl | that's because (for when you check the logs, rob) you didn't copy over the header file. |
| 13:50.13 | lkcl | run ctags -R |
| 13:50.19 | lkcl | use vi to edit the file |
| 13:50.43 | lkcl | in blueangel_lcd.c, on the structure at about line 207, press "tab". |
| 13:51.05 | lkcl | that will take you to the definition of the structure that contains the field "resume". |
| 13:51.21 | lkcl | probably in blueangel_lcd.h |
| 13:51.41 | lkcl | and you will find _four_ functions in it, not three - one called resume_pre, the other resume_post. |
| 14:17.40 | lkcl | DAMNIT. |
| 14:17.50 | lkcl | gcc 3.4.3 fails on xorg as well |
| 14:18.55 | lkcl | yep - that's the one, rob. |
| 14:19.20 | lkcl | i feel that there's a GPIO missing, however. |
| 14:19.32 | lkcl | there really should be four separate and distinct GPIOs. |
| 14:19.38 | lkcl | oh well. |
| 15:05.26 | *** join/#htc-blueangel JuNuKN (n=JuNuKN@Mdcbe.m.pppool.de) |
| 15:05.48 | JuNuKN | ~botmail read |
| 15:08.28 | lkcl | heya JuNuKN |
| 15:08.51 | JuNuKN | Hey lkcl |
| 15:11.00 | JuNuKN | One short question for the night .... I have now downloaded the sorces, and the familar Cross-Toolchain. I know now, which path for the chain to be included as my path ,.... but which command do i need to build Himalaya stuff ? |
| 15:12.51 | iclem | "make" ? |
| 15:13.56 | lkcl | cp arch/arm/config/himalaya_config .config |
| 15:14.18 | lkcl | CC=/daskdhaskjhdaskdj/arm-linux-gcc |
| 15:14.20 | lkcl | export CC |
| 15:14.21 | lkcl | make |
| 15:14.25 | lkcl | i _think_. |
| 15:14.35 | lkcl | do a google search on how to cross-compile a kernel. |
| 15:14.52 | lkcl | you might have to set CROSS_CC=/asdkalsdakja/arm-linux- |
| 15:14.55 | lkcl | export CROSS_CC |
| 15:14.58 | lkcl | or something like that. |
| 15:17.01 | JuNuKN | Thanks for the tip ... now I start to understand ... because I wondered, that I didn't found a make in the cross-chain and never saw to give a pathto gcc or so for a compile command. Thanks ... I will google tonight ... I am a Windows MsDev-UI guy ;o) I have to find into commandline envirenment |
| 15:19.06 | lkcl | ah joy. you are in for a treat :) |
| 15:22.26 | JuNuKN | Mainly yes, I want to learn about linux and to be familar with it. At job I always have to handle with windows, ... but I am free in the choise of my PDA and Phone ... so I think thats a god starting point ... |
| 15:27.41 | lkcl | you picked a hard-core project to start on, dude |
| 15:30.31 | JuNuKN | Yes ... I think so, but in principle I started somewhere arround this many years ago in hardwaredeveloping with microprocessors, I2C Bus Peripheral and pascal ... developing own "OS" ... |
| 15:31.27 | JuNuKN | Not a real OS,- more a monolit, with uploadable extensions ... |
| 15:31.53 | cr2 | hi guys. i've returned back and slept a bit. |
| 15:32.28 | JuNuKN | Hello again |
| 15:32.55 | cr2 | hi JuNuKN. something new ? |
| 15:32.56 | lkcl | heya cr2 |
| 15:33.10 | cr2 | hi lkcl. |
| 15:34.08 | cr2 | some nice guys told be that the constantly switched on flashlight on universal will destroy the led... makes sense. |
| 15:35.03 | lkcl | oops... |
| 15:35.11 | lkcl | it _is_ a bit powerful... |
| 15:35.30 | cr2 | and its a _flash_ light |
| 15:38.09 | cr2 | several people asked about linux on himalaya :) |
| 15:38.16 | lkcl | surprise... |
| 15:38.50 | lkcl | *sigh* . need the asic3_mmc to work, really, for that. booting up over NFS into a ramfs - well... :) |
| 15:39.18 | cr2 | the beamer didn't want to work with the notebook of cp15, so i guess a lot of people simply did't get what it was about. |
| 15:39.42 | lkcl | darn. i don't mind doing talks without any notes (at all). |
| 15:39.58 | lkcl | so i do lots of "diagrams in the air" as visual cues. |
| 15:40.06 | lkcl | which people understand really well. |
| 15:40.13 | cr2 | i have learned something from two hone talks. |
| 15:40.19 | lkcl | ... but if you're not prepared for that sort of thing... |
| 15:41.19 | cr2 | yes. it looked like that. for example, you don 't use yellow colour, you don't use your notebook. the beginner mistakes... |
| 15:42.27 | cr2 | but still. we are actually a big success :) |
| 15:42.36 | lkcl | cool! |
| 15:43.28 | cr2 | took 1200 km to learn that. and terrible weather ;) |
| 15:44.02 | cr2 | a lot of interesting people there. |
| 15:44.54 | cr2 | lkcl: what's with asic3_mmc ? |
| 15:45.31 | lkcl | oh. *grump*. |
| 15:46.32 | lkcl | got more to deal with than damn asic3_mmc. |
| 15:46.33 | cr2 | can we have one more platform_device ? |
| 15:46.41 | cr2 | ? |
| 15:46.49 | lkcl | you can have _annnything_ you like, cr2 :) |
| 15:47.38 | lkcl | latest kernel code isn't booting properly for the blueangel. |
| 15:48.18 | cr2 | well. i don t get it. what went wrong ? do you have a hi0 branch or playing in ba0 ? |
| 15:48.51 | lkcl | ba0. |
| 15:49.12 | cr2 | hmm. not that good. |
| 15:49.22 | lkcl | yeh. |
| 15:50.07 | cr2 | did you change something in *asic3.h ? |
| 15:50.14 | lkcl | i have no idea what's happened. gonna have to revert back, find a zImage that works, etc. etc. |
| 15:50.43 | lkcl | i need to confirm that i _can_ boot linux on this device, first! |
| 15:51.00 | cr2 | ok. |
| 15:51.23 | lkcl | damn, why is doing this stuff always so time-consuming? :) |
| 15:52.05 | *** join/#htc-blueangel rob_w (n=bob@p85.212.195.56.tisdip.tiscali.de) |
| 15:52.05 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 15:52.08 | cr2 | because it is hacking, and not doing the right thing :) |
| 15:52.18 | lkcl | heya again rob |
| 15:52.24 | rob_w | heya |
| 15:52.40 | cr2 | hi rob_w |
| 15:53.24 | rob_w | so hi hope i find time tomorrow or late today night to get some more try`s ,, i was stuck on a compile error .. |
| 15:54.23 | lkcl | yes. i wrote some notes for you but you'd left |
| 15:54.36 | rob_w | sorry |
| 15:55.18 | cr2 | lkcl: write a Makefile , not notes :) |
| 15:55.23 | rob_w | i was thinking of a user space tool to address gpio`s ... |
| 15:56.02 | rob_w | but as u said , it all might be not related to the lcd .. nonetheless the latest code does nto work for me |
| 15:58.50 | lkcl | yeh, rob - i sort-of started on one. |
| 15:59.07 | lkcl | based on cp15's gsminit.c |
| 15:59.22 | lkcl | which mmaps /dev/mem and then directly accesses the memory. |
| 15:59.59 | lkcl | i got a segfault - so figured it might be better just to have some kernel function where you can call them and display the GPIO info - or do you mean actually _change_ the gpios? |
| 16:11.20 | rob_w | no i meant exaclty that to turn them on / off |
| 16:11.26 | rob_w | need to run again .. cu |
| 16:11.30 | lkcl | ah _ha_ i got a boot. |
| 16:11.32 | lkcl | okay rob! |
| 16:11.48 | lkcl | *whew*. i got a boot. |
| 16:17.38 | lkcl | boot yer bastard :) |
| 16:17.44 | lkcl | _eexcellent_ |
| 16:18.22 | lkcl | muhahahah |
| 16:21.19 | lkcl | _what_???? |
| 16:21.23 | lkcl | suspend WORKS????? |
| 16:21.26 | lkcl | i don't understand... |
| 16:21.34 | lkcl | there must be a mistake... |
| 16:21.50 | lkcl | screen goes to sleep.... |
| 16:22.14 | lkcl | wake it up... |
| 16:22.24 | lkcl | well bugger me. |
| 16:22.33 | lkcl | how the hell did _that_ happen??? |
| 16:23.06 | lkcl | the LCD _works_ after a resume. |
| 16:23.25 | lkcl | wtf is going on?? |
| 16:24.51 | lkcl | dang me. dmesg shows it's actually going to sleep. |
| 16:25.00 | lkcl | ahhh - i have a boardid 0x5. |
| 16:25.04 | lkcl | that'd be why it works. |
| 16:25.14 | lkcl | oh, bugger. |
| 16:25.43 | lkcl | rob_w - when you get back: you should find that you're okay. it's the boardid 0x6 stuff that needs fixing. |
| 16:25.44 | lkcl | oops. |
| 16:26.26 | lkcl | so, you _shouldn't_ need to make any modifications to the source code that's there, already. |
| 16:26.41 | lkcl | drat - i need an XDA IIs not an MDA III. drat. |
| 16:58.34 | cr2 | lkcl: do you know if can i pull rts and dtr with stty ? i recall, it is not "posixly correct" ;) |
| 17:01.57 | lkcl | rts and dtr... haven't a clue, dude |
| 17:02.34 | lkcl | ah hahahahahahhh i just unlocked the mdaIII with xdadev_allunlock v 1.6 |
| 17:06.41 | cr2 | look at that: http://tuxmobil.org/Infrared-HOWTO/infrared-howto-a-serial-infrared-port-sniffers.html |
| 17:17.41 | lkcl | oh, right, |
| 17:17.53 | lkcl | i know someone who will know, |
| 17:18.44 | lkcl | i'll just email him and ask if he can help. |
| 17:18.51 | lkcl | hey, phone calls work, btw. |
| 17:18.57 | lkcl | volume on the mic is far too low. |
| 17:19.56 | lkcl | damn - and incoming phone calls wake it up! |
| 17:22.31 | JuNuKN | Do you speak from Himalaya ? |
| 17:22.55 | lkcl | blueangel. mda III |
| 17:23.04 | lkcl | i'm not _that_ brave, JuNuKN ! |
| 17:24.23 | JuNuKN | Ahh OK ... I thought and hoped, because you worked last night on the SD stuff of himalaya ... |
| 17:24.48 | lkcl | naaah - that was a very frustrating and irritating time. |
| 17:25.40 | JuNuKN | why ... you found the BA hack / changes in the mmc driver, which made it impossible for Himalaya ,,, |
| 17:26.24 | lkcl | que,que? |
| 17:26.39 | lkcl | when i get bored of the blueangel (probably in about 10 mins, because it sort-of works) |
| 17:26.46 | lkcl | i'll try the himalaya again. |
| 17:27.24 | lkcl | hmmm.... |
| 17:29.14 | lkcl | i think there's some sort of memory corruption going on... this poor beastie is _very_ unhappy about restarting gpe.. |
| 17:29.20 | JuNuKN | Would be great. Can I also be from help somehow, with some Logging stuff ? I have a SDIO WLAN card in my MDA2, and be prepared to run the WGPIO with watching ASIC3 and COM Port output. !? |
| 17:30.10 | JuNuKN | Oh I see ,,, still other, bigger problems ,,, |
| 17:30.53 | lkcl | it's not the wgpio that we need to know |
| 17:31.19 | lkcl | it's the asic3 memory areas that need to be watched. |
| 17:31.35 | lkcl | i used to have a python program that did a telnet to haret.exe |
| 17:31.48 | lkcl | and it would repeatedly do a "pd NNNNNN 400" |
| 17:31.52 | lkcl | and display the results on-screen. |
| 17:32.01 | lkcl | that's the kind of thing that would be useful |
| 17:32.28 | lkcl | trouble is that you can't run haret _and_ have stuff happening in wince at the same time. |
| 17:32.32 | lkcl | haret takes over. |
| 17:32.49 | lkcl | it _might_ work... |
| 17:33.06 | JuNuKN | Ahh ok, so I am still not from help .... I thought also, why nobody has prepared some kind of interpreter, to play with the gpio lines without reinstalling everything .... maybe with Haret. |
| 17:34.06 | cr2 | JuNuKN: you can play with gpio with haret. |
| 17:34.20 | cr2 | can you make some "static" dumps ? |
| 17:34.30 | JuNuKN | I am not shure about, what you say. I tried, to let Haret watch gpios and Some banks, while I switched to the file explorer and browsed throught the directory ... |
| 17:34.41 | lkcl | okay. himalaya play time. |
| 17:35.14 | cr2 | dump the sd control and sdio control areas with and without the sd/sdio card. |
| 17:35.35 | lkcl | pd 0x0d000000 400 |
| 17:35.39 | cr2 | better: the supposed sd control and sdio control. |
| 17:35.45 | lkcl | and pd 0x0d800000 400 |
| 17:35.49 | lkcl | isn't it? |
| 17:35.54 | lkcl | something like that? |
| 17:35.56 | cr2 | no. its asic3 gpio & things. |
| 17:36.00 | lkcl | oh. |
| 17:36.18 | lkcl | <PROTECTED> |
| 17:36.23 | JuNuKN | Yes I can do explain me the steps or give me the Haret-Script for it. I will do it immidiatey.... Do I have also some boardID, which you have to know? Can I find out with HaRet ? |
| 17:36.25 | cr2 | look at the asic3 wiki page |
| 17:36.31 | lkcl | <PROTECTED> |
| 17:36.48 | cr2 | wait. |
| 17:37.48 | cr2 | SD_CTRL_Base +0x1000 |
| 17:37.57 | cr2 | SDIO_CTRL_Base +0x1200 |
| 17:38.32 | cr2 | just this page |
| 17:39.36 | lkcl | so... pd 0x0c801000 400 |
| 17:39.37 | cr2 | guys. don't be so chaotic. we know the board id determination process for himalaya, blueangel and universal. do it and put into wiki. |
| 17:40.06 | cr2 | lkcl: yes |
| 17:40.20 | lkcl | i _like_ chaos! |
| 17:40.28 | cr2 | :) |
| 17:40.44 | lkcl | cr2 - i'm a discoverer and chaos-creater, not a completer. |
| 17:41.16 | cr2 | ok. i've got it already :)))) |
| 17:41.27 | lkcl | you have? gosh, that didn't take long. |
| 17:42.36 | cr2 | well. without documenting things we confuse a lot of beginners... |
| 17:43.00 | lkcl | yep :) |
| 17:43.01 | lkcl | okay. so. |
| 17:43.11 | lkcl | what was it you were suggesting to put into the wiki? |
| 17:43.33 | lkcl | cos not even _i_ know how to determine board ids except by catching the thing at the beginning of dmesg |
| 17:43.34 | JuNuKN | Yes you do :o) Please with slow steps ... If wished, I will document it in Wiki , - If I can follow ... |
| 17:43.35 | cr2 | what is the board id of your hima, ba and universal ? |
| 17:43.39 | cr2 | ? , 5, 5 |
| 17:43.53 | lkcl | don't know, 5, and don't know. |
| 17:44.33 | cr2 | the hima id is in the asic3 gpio setup. look into wiki page for it. |
| 17:44.37 | JuNuKN | How to determine with Haret ? Which commands ? I am connected with haret ... |
| 17:44.39 | cr2 | the rest i have done for you. |
| 17:45.00 | lkcl | ooo. cool. i get it :) |
| 17:45.44 | lkcl | okayokaokay. |
| 17:45.47 | lkcl | whatamidoing. |
| 17:45.48 | lkcl | pd. |
| 17:45.48 | cr2 | JuNuKN: pd 0x0d800000 0x800 |
| 17:47.11 | cr2 | lkcl. it's all documented. |
| 17:47.16 | JuNuKN | Ok a lot of Output .... What to do with it? |
| 17:47.29 | cr2 | hima: dump asic3 gpio area. decode |
| 17:47.39 | cr2 | ba: dump pxa gpio. decode |
| 17:47.59 | cr2 | universal: tdump asic3 gpio area. decode. |
| 17:48.18 | cr2 | everything in *asic3 and *gpio wiki pages. |
| 17:49.17 | cr2 | JuNuKN: either decode it looking at the bit descriptions in HimalayaEGPIO, or send to me. |
| 17:50.16 | JuNuKN | OK, I will do both .... I will send it and try to interpret by myself .... |
| 17:50.26 | cr2 | JuNuKN: you can also do pd 0x0c800000 0x400 with and without the sd/sdio card. |
| 17:51.21 | JuNuKN | Maybe I be from more help, because I am connected via WLAN and Seriell-Port with the HaretVersion of llange with ASIC3 Buissy loop check and seriell output .... If somebody needs!? |
| 17:51.37 | JuNuKN | At the same time ;o) |
| 17:51.41 | lkcl | okay, got four dumps. |
| 17:52.07 | cr2 | lkcl: where ? |
| 17:53.24 | lkcl | -0c801000 | 3c123c12 02000200 08000800 00de00de | .<.<............ |
| 17:53.25 | lkcl | +0c801000 | 3c123c12 02000200 18001800 00de00de | .<.<............ |
| 17:53.33 | lkcl | -0c801060 | 002d002d 00000000 10001000 00000000 | -.-............. |
| 17:53.33 | lkcl | +0c801060 | 00720069 00000000 10001000 00000000 | i.r............. |
| 17:53.37 | JuNuKN | ~seen cr2 |
| 17:53.47 | apt | cr2 is currently on #htc-blueangel. Has said a total of 52 messages. Is idling for 1m 40s |
| 17:53.47 | lkcl | - line is with the SDIO card in |
| 17:53.52 | lkcl | + line is without |
| 17:54.50 | cr2 | we need a "dump pxa" and "dump asic3" commands in haret. |
| 17:55.13 | lkcl | 2nd diff: |
| 17:55.14 | lkcl | -0c801030 | 09000900 00000000 07000700 00800080 | ................ |
| 17:55.14 | lkcl | +0c801030 | 09000900 00000000 07000700 00a000a0 | ................ |
| 17:55.17 | cr2 | lkcl: look in the asic3_mmc register map. do this changes make sense ? |
| 17:55.24 | lkcl | -0c801060 | 414d495a 00000000 10001000 00000000 | ZIMA............ |
| 17:55.25 | lkcl | +0c801060 | 20204547 00000000 10001000 00000000 | GE ............ |
| 17:55.43 | cr2 | zimage ??? |
| 17:55.59 | JuNuKN | cr2 how to send you the log ? |
| 17:56.00 | cr2 | is it a sd data buffer ? |
| 17:56.11 | cr2 | JuNuKN: xda-dev. |
| 17:56.23 | lkcl | oh yeh. how odd |
| 17:57.04 | lkcl | it's possibly some characters from the serial buffer or something weird. |
| 17:57.17 | lkcl | or... nah. |
| 17:57.24 | lkcl | maybe even from the SDIO card. |
| 17:57.26 | lkcl | how weird! |
| 17:59.04 | cr2 | wait. the offsets in the asic3_mmc. are they for the "standard" or for the blueangel layout ? |
| 17:59.11 | lkcl | "standard". |
| 17:59.38 | JuNuKN | lkcl do you have also a sdio card ? Also a wlan card? |
| 17:59.57 | lkcl | if it was "packed" it would be pd 0x0c800800 i would have needed to dump. |
| 18:00.07 | lkcl | i have a 256mb SD card. |
| 18:00.44 | lkcl | okay. so SDIO_CTRL_Cmd is 0x00. |
| 18:01.03 | cr2 | at the dump time. |
| 18:01.37 | lkcl | 14 is 0x0c which is SDIO_CTRL_Response0 |
| 18:02.00 | lkcl | sorry. |
| 18:02.11 | lkcl | 14 is 0x0e which is Response1 |
| 18:02.15 | cr2 | 14 dec or hex ? |
| 18:02.16 | JuNuKN | I alway thought a SD-Memnory card is no sdio card at all ... |
| 18:02.27 | lkcl | 14 dec is 0x0e hex |
| 18:02.28 | cr2 | ok. |
| 18:02.51 | cr2 | makes sense. |
| 18:03.33 | cr2 | JuNuKN: sdio != sd card |
| 18:03.53 | cr2 | s/sd card/sd memory card/ |
| 18:04.55 | cr2 | sdio is more like pcmcia |
| 18:05.52 | cr2 | but the memory cards are a separate device class |
| 18:08.45 | cr2 | JuNuKN: you can test the sdio irq by inserting and taking out the sdio card, while running wgpio |
| 18:09.13 | lkcl | i _think_ i managed to get that to work: i saw one change on GPIO 8 |
| 18:10.10 | lkcl | GPLR 74 and 2 |
| 18:10.36 | cr2 | we can also check the SD clock settings in asic3. btw, what is the "SD Control Register Base Address" in H3900_ASIC3_SD_CONFIG ? |
| 18:10.51 | cr2 | lkcl: wiki ? |
| 18:10.53 | lkcl | GPLR[8] changed |
| 18:11.04 | lkcl | clock settings... |
| 18:11.08 | lkcl | one mo... |
| 18:11.58 | *** join/#htc-blueangel iclem (n=clement@iclem.net) |
| 18:12.05 | lkcl | hiya iclem |
| 18:12.24 | iclem | hi lkcl |
| 18:12.26 | iclem | hi all |
| 18:12.27 | cr2 | lkcl: i have only reformatted the available data. gpio8 can be mux gpio interrupt. |
| 18:12.35 | cr2 | hi iclem |
| 18:12.56 | lkcl | GPIO 2 is usb - power or something. |
| 18:13.10 | lkcl | GPIO 74 - seem to remember that's an SSP somewhere... |
| 18:13.19 | lkcl | GPIO 8 was the only other one i saw change. |
| 18:13.45 | cr2 | BP ?? bluetooth phone ? |
| 18:13.54 | lkcl | HX4700_ASIC3_SD_CONFIG_Base is #defined to 0x0200 |
| 18:14.13 | cr2 | http://wiki.xda-developers.com/index.php?pagename=HimalayaGPIO |
| 18:14.57 | lkcl | GPIO 8 - asic3 |
| 18:15.02 | cr2 | no. there is also #define _H3900_ASIC3_SD_CONFIG_Addr0 0x10 /* [9:31] SD Control Register Base Address */ |
| 18:15.10 | cr2 | after the Command |
| 18:15.30 | lkcl | ah _ha_. 74. "reacts to SD and also to bluetooth". |
| 18:16.14 | lkcl | #define _HX4700_ASIC3_SD_CONFIG_Addr0 0x10 /* [9:31] SD Control Register Base Address */ |
| 18:16.48 | lkcl | 1 mo... |
| 18:18.27 | lkcl | ah HA what the heck is this?? |
| 18:18.52 | lkcl | host->ctl_base = ioremap(0x0e000000, 0x2000) |
| 18:18.53 | lkcl | ???? |
| 18:19.05 | cr2 | SSP was gpio84. |
| 18:19.16 | lkcl | damn that's hard-coded for blueangel. |
| 18:19.19 | cr2 | SDIO_BASE. for hx4700 and blueangel |
| 18:19.26 | lkcl | damn damn that's why it doesn't work on the himalaya. |
| 18:19.47 | JuNuKN | I have to leave for half an hour or so ... familie wants to eat ...Back again |
| 18:19.54 | cr2 | look in the logs. i've said to change it :) |
| 18:19.56 | JuNuKN | back again |
| 18:19.57 | lkcl | it's because ctl_base is being mapped to 0x0e000000 which is the blueangel base. |
| 18:20.11 | lkcl | yeh, i am sure you did, cr2 - it went by in a hurry and i missed it :) |
| 18:21.17 | *** join/#htc-blueangel iclem (n=clement@iclem.net) |
| 18:21.22 | lkcl | FRAK :) |
| 18:21.37 | lkcl | *sigh* okay will try 0x0c800000.... |
| 18:22.45 | cr2 | we need a portable asic3-mmc platform_device :) |
| 18:22.51 | lkcl | tell me about it... |
| 18:28.33 | JuNuKN | I know I am to stupid, but can someone explain me how to use a seriall port as TCP/IP connection ? |
| 18:29.48 | JuNuKN | And I didn't see the boardid - location in Wiki stuff please point it to me ... |
| 18:30.24 | iclem | hi JuNuKN |
| 18:30.53 | iclem | I never used serial port i can't help you |
| 18:31.12 | iclem | I've you tried to boot the existent kernel ? |
| 18:32.48 | JuNuKN | OK, and with USB ? |
| 18:33.21 | JuNuKN | ... Oh no, sorry, USB cable is on the job desk .. |
| 18:34.42 | iclem | tell JuNuKN about himalaya boardid |
| 18:34.47 | iclem | ~tell JuNuKN about himalaya boardid |
| 18:35.54 | lkcl | you need synce |
| 18:36.23 | lkcl | apt-get install synce-serial and synce-dccm |
| 18:36.43 | lkcl | you then run dccm -p <yourpinnumber> as an ordinary user |
| 18:36.53 | lkcl | and then run synce-serial-start as root. |
| 18:37.04 | lkcl | oh. |
| 18:37.09 | JuNuKN | OK.... I need to boot linux ... and try to get the USB-seriell adapter to work first... |
| 18:37.11 | lkcl | that's with the USB cable :) |
| 18:37.26 | lkcl | sorry, i just described to you how to do it with the USB cable :) |
| 18:37.49 | lkcl | with the serial, it'd be similar-ish - but you're better off using it just with a console line. |
| 18:38.03 | lkcl | you'd need to set up pppd in the initrd of the himalaya |
| 18:38.24 | lkcl | and run pppd also on your linux box. |
| 18:38.43 | cr2 | hima boardid is in asic3 gpioc: http://wiki.xda-developers.com/index.php?pagename=HimalayaEGPIO |
| 18:38.46 | lkcl | you're better off just using the serial cable as a terminal, on the console (minicom) |
| 18:39.25 | JuNuKN | So I have to boot linux first. By the way WinCe gives me as Hardware Idetifier PH10A and not the Hex Values ... I don't know, where to look for within WinCE |
| 18:39.47 | lkcl | ah _ha_ my himalaya is BoardID 0x4. |
| 18:40.27 | iclem | board id 4 may be for PH10A, I've id 6 and PH10B |
| 18:41.07 | iclem | version differs by lcd brand |
| 18:41.39 | lkcl | GPIOD13 SD card inserted = 1 iclem RE |
| 18:43.02 | iclem | is it the SD irq needed by asic3_mmc ? |
| 18:43.39 | lkcl | i'm looking for the damn thing. |
| 18:43.56 | cr2 | iclem: do you know which LCD you have ? http://wiki.xda-developers.com/index.php?pagename=HTC_Partnumbers |
| 18:44.15 | cr2 | GPIOD13 is probably the SD detect thing. |
| 18:44.31 | cr2 | we are looking for the pxa sd irq. |
| 18:44.38 | lkcl | SD detect, yes - but is it the SD IRQ? blast. damn. |
| 18:44.56 | iclem | cr2: no i've not disassembled my device |
| 18:45.01 | lkcl | RIGHT. |
| 18:45.02 | lkcl | that's it. |
| 18:45.19 | iclem | how do you know ? |
| 18:45.20 | lkcl | i'm going through the whole lot until it works. |
| 18:47.17 | JuNuKN | iclem ... Yes I booted the kernel yesterday and send the log to cr2, but I didn't found the board-ID there as well ... Why did you ask? |
| 18:48.02 | iclem | cr2: think it's the Toppoly TD035STEB2 |
| 18:48.35 | lkcl | brb.. phone call... |
| 18:48.36 | iclem | JuNuKN: you can read in dmesg |
| 18:48.44 | cr2 | iclem: why do you think so ? |
| 18:49.33 | JuNuKN | iclem: Thanks .... Now it is to late for it ... I will look next time, when it is up. |
| 18:49.46 | cr2 | JuNuKN: pd 0x0d8000230 4 |
| 18:50.07 | cr2 | pd 0x0d800230 4 |
| 18:50.27 | iclem | I had screen offset with llange kernel |
| 18:51.40 | cr2 | intersting. these are different LCDs from the blueangel, so the ATI things must be different. Can you read the atihwtbl0 on wince ? |
| 18:52.02 | iclem | we use cp15's dumpati tool to get ati registries |
| 18:52.48 | cr2 | well. we can directly compare the "original wince" setup data, too. |
| 18:54.02 | iclem | i had this file saved somewhere but can find it anymore |
| 18:54.37 | iclem | found, do you want it ? |
| 18:55.06 | JuNuKN | cr2: 0d800230 | 3e64fe64 | d.d> |
| 18:57.14 | cr2 | can we put them into wiki ? |
| 18:57.24 | cr2 | JuNuKN: you have id=4 |
| 18:58.53 | iclem | cr2: you want me to put my atihwtbl0 into wiki ? |
| 18:59.19 | JuNuKN | Thanks ... Is it the last Byte ? I didn't found out with the big dump and the different GPIO Ports How to see the Offset to Port C ? |
| 19:00.39 | iclem | JuNuKN: it is the 3,4,5 (http://wiki.xda-developers.com/index.php?pagename=HimalayaEGPIO) |
| 19:00.41 | cr2 | look for "Read here" in http://wiki.xda-developers.com/index.php?pagename=ASIC3 |
| 19:01.56 | cr2 | iclem: yes. create the table for boardids. |
| 19:02.25 | cr2 | then we will push the BA people to do the same. |
| 19:02.50 | cr2 | there is no ATI on universal and magician. |
| 19:15.06 | JuNuKN | Sorry cr2 , I still didn't see my board ID .... 3e means 00111110 : Wiki sais Board ID are Bits 3, 4 and 5, which are all a 1 ? Gives 7 ???? Please ... I will dok it in Wiki ! |
| 19:16.47 | iclem | cr2: http://wiki.xda-developers.com/index.php?pagename=HimalayaBoardIDs |
| 19:22.52 | iclem | JuNuKN: if think bit 0 is the most right bit, gurus sould correct me if i'm wrong |
| 19:22.53 | JuNuKN | please don't let my die stupid .... |
| 19:23.57 | iclem | and the value you should watch is FE64 |
| 19:24.16 | cr2 | iclem: it is great. there are settings for both LCDs there. |
| 19:24.39 | cr2 | JuNuKN: take lower 16bit 0xfe64 |
| 19:25.10 | cr2 | write down the bits : 1111111001100100 |
| 19:25.34 | cr2 | take 5,4,3: 100 == 4 |
| 19:27.01 | cr2 | JuNuKN: can you put your hwtbl0 into wiki ? |
| 19:29.46 | JuNuKN | I see, and try to understand ,,,, We ordered 4 Bytes from Adress 0x0d800230 ... It returns us the 4 Bytes: 3e 64 fe 64 : $ bytes are 32 Bits ... but Port c has only 16 Bits !? I will do, if you explain me, what hwtb10 is , or where to find!? |
| 19:30.24 | iclem | you can find atihwtbl0.txt in \windows |
| 19:31.08 | iclem | I don't remember if you need to set a key in registry or not |
| 19:32.44 | cr2 | haret reads 4bytes at once (register size). the cpu is little endian. the asic3 registers are 16bit. its "ldr vs. ldrh" thing. |
| 19:32.45 | JuNuKN | OK, I found the file .... I will put it into the wiki ,,, few seconds please ... |
| 19:34.00 | JuNuKN | OK, I have to google for the little endian description .... |
| 19:35.31 | iclem | and what are the last 16 bits intended to ? |
| 19:35.44 | cr2 | little endian means that the bits in a multibyte are counted from left to right. |
| 19:35.47 | cr2 | junk. |
| 19:36.47 | cr2 | it is easier for a risc processor to grab a register than to implement a byte load. |
| 19:37.15 | iclem | really junk ? ram is so cheap nowadays |
| 19:38.45 | cr2 | the original dec alpha 21064 didn't have the byte and short load/stores. it was a great thing :) |
| 19:39.02 | JuNuKN | OK, it is in the wiki now, but don't know the LCD type ... maybe we have to add a comment about it to the link ? |
| 19:39.59 | iclem | if you have correct display with the kernel i provide you it should be the Sony one |
| 19:40.36 | cr2 | you can compare your LCD data with the table. |
| 19:42.17 | iclem | i have dumpati log from llange and mine, is it of interest to put it into wiki too ? |
| 19:44.49 | cr2 | why not ? |
| 19:46.28 | iclem | the two atihw.. files are the same |
| 19:47.23 | cr2 | imho, we have a much better documented project than any other on hh.org, maybe excluding those who got the data directly from hp :) |
| 19:56.07 | iclem | lunch time, i will be back later |
| 20:01.58 | JuNuKN | Lets resume the Board-ID calculation like I will describe it into Wiki as an example to Work with Haret and GPIOPorts ... : We ordered 4 Bytes at Adress 0xd800(000) Plus Offset of PortC Status register which makes 0x0d800230 ... command "pd 0x0d800230 4" ; we got the result as described above ... 3e 64 fe 64 where the lowest bytes are at the right ... so we translated the fe64 into 111111100110 |
| 20:01.58 | JuNuKN | 0100 Now we took the bits 3, 4 and 5 , which must be countet from the right as Bit Number 4, 5 and 6. These are 100 And this is dezimal value 4 (from right to left: 0 times a 1; 0 times a 2; 1 times a 4) |
| 20:14.25 | JuNuKN | OK, I updated the wiki, - please crosscheck ... |
| 20:14.35 | *** join/#htc-blueangel start_ (n=rl@nezumi.plus.com) |
| 20:18.05 | start_ | Hi Luke, its Richard. What was the problem with the IRDA sniffer? |
| 20:18.24 | JuNuKN | OK, now I think I understand a little bit more about the EGPIO's and how to analyse with HaRET. Probably I can create a loop script with Haret, which dumps the Adressframe , waits a few milliseconds and dumps it again! ? Correct way? |
| 20:26.12 | lkcl | heya richard |
| 20:26.29 | lkcl | cr2 - can i introduce you my friend, richard? |
| 20:27.20 | lkcl | richard - cr2 was trying to compile a program that contained some rts and cts stuff in it (or something) and he wasn't sure how to build the program. |
| 20:27.53 | lkcl | oh - btw: this is a low-traffic irc channel so people tend to go away and then come back again after a few mins or a few hours... |
| 20:28.35 | start_ | I used: gcc -o irdasniff -Wall -W irdasniff.c |
| 20:28.43 | lkcl | and that was it?? :) |
| 20:29.18 | start_ | The only hassle was converting < to >, and similar for > and &. |
| 20:29.25 | lkcl | ... i _fink_ he might have a bit more of a technical question than that :) |
| 20:31.09 | start_ | There are a pile of unused variables, there is not code to select the serial port, and at a guess, it will fail on some hardware because CD might be floating. If you know what the problem is, I will be able to help. |
| 20:31.27 | JuNuKN | lkcl can you please crosscheck my description in the wiki for getting the board ID ... I think still the thing with little endian is wrong!? |
| 20:32.03 | lkcl | ack.. 1mo... |
| 20:32.06 | lkcl | which page? |
| 20:32.17 | JuNuKN | Recent edits ,,,, |
| 20:32.24 | lkcl | oh yeh.. |
| 20:32.29 | JuNuKN | http://wiki.xda-developers.com/index.php?pagename=HowToWorkWithHaRET |
| 20:34.22 | lkcl | bits 345 - right to left... 100 equals 4 |
| 20:35.06 | lkcl | btw because you have BoardID in the text, wiki thinks that's a page |
| 20:35.13 | JuNuKN | What with my idea to make a loop in HaRET script over a Adressrange for the ports, while I am connceted with the WLAN-Card and transfering the logdata over it ... should show everything of SDIO-Interfacing ;o) |
| 20:35.43 | lkcl | i _really_ should find that python script i created. |
| 20:35.54 | JuNuKN | Is it correct? cr2 told us, that it is little endian, which should be counted in the other direction !? |
| 20:36.23 | lkcl | richard: well, when cr2 wakes up, we'll be able to ask him |
| 20:36.38 | cr2 | hi richard, i'm back. |
| 20:37.19 | cr2 | i have compiled the program, but don't see anything with it. |
| 20:37.23 | lkcl | pd will dump the data for you in the visual manner in which you'd expect to see. |
| 20:37.50 | lkcl | ah, right, cr2, so it's _usage_ not thingy. |
| 20:38.12 | lkcl | btw does anyone have a dmesg log of a blueangel booting up, conveniently to hand? |
| 20:38.14 | cr2 | on one side there is a serial port with an IR transmitter, on the other side the SIR receiver. |
| 20:38.37 | cr2 | the SIR receiver works with linux irda without any problems. |
| 20:38.56 | start_ | cr2: I will read the code more carefully, and see what I can find. |
| 20:39.09 | cr2 | thanks. |
| 20:39.09 | lkcl | 'sok i got one. |
| 20:43.28 | lkcl | hmmmm.... |
| 20:46.44 | start_ | cr2: Only a few terminal attributes are set. You must explicitly set every single one that is documented, otherwise, the results will depend on the previous ate of the serial port. I would guess there are problems with VMIN/VTIME. (I have done plenty of serial devices, but never IDRA) |
| 20:47.15 | start_ | s/ate/state/ |
| 20:47.23 | cr2 | can i do it with stty ? |
| 20:49.05 | cr2 | ok. i will try to do it properly without any hacks. |
| 20:50.39 | start_ | Just checked stty: you can set the oef char, which is sometimes the same as VMIN, but I did not see a guarenteed method of setting VMIN with stty. |
| 20:50.52 | start_ | s/oef/eof/ |
| 20:51.27 | cr2 | ok. i see. |
| 20:58.21 | JuNuKN | lkcl do you have the instruction for analysing the SD interface in mind or history? My client still has deleted it .... Should we watch the ports only or different adressspaces? |
| 20:59.18 | JuNuKN | iclem ? |
| 21:01.01 | cr2 | JuNuKN: i think the key question is how often will you "watch" , and if it will be triggered by a timer, interrupt, or just busy looping for a period of time. |
| 21:01.43 | JuNuKN | How did cp15 it for BA ? |
| 21:03.47 | JuNuKN | I think, If we only have to look on the 4 Ports, it should be possible to store this 4 times 2 Byte Values in memory in buissy loop. After a given Time write everything as Binary-Values to output, which will show the changes like an logic analyser |
| 21:03.54 | cr2 | he has some "cpu simulator" which is slow, and runs only the bootloader. i don't think the bootloader supports sdio. |
| 21:03.59 | lkcl | well at the moment, it looks like the commands aren't getting called |
| 21:04.15 | lkcl | ah... yes, i believe it does read the sd card. |
| 21:04.24 | lkcl | iirc correctly from a couple of years ago. |
| 21:04.42 | lkcl | if you issue the right command. |
| 21:04.43 | cr2 | and write. but not sdio :) |
| 21:05.21 | lkcl | hmm, it was a long time ago. maybe it was the DoC that w4xy was telling me about. oh well. |
| 21:05.45 | cr2 | the busy loop poses too much stress on the system. |
| 21:06.23 | cr2 | the tight busy loop. |
| 21:09.07 | JuNuKN | And if do everytime the same process with Haret ... maybe trigger a dir listing in a separate thread and doing the bussy loop multiple times ... we should see as a "merge" the cummulated handling on the ports ... ? maybe? Other idea? |
| 21:10.31 | JuNuKN | Maybe cp15 can also analyse our startup code with his simulator? |
| 21:10.38 | lkcl | a better way: use telnetlib.py to set up a connection, write("pd 0xxxxxxxx 16") |
| 21:10.49 | lkcl | then do a "read_until" and parse the results. |
| 21:10.53 | lkcl | display them on-screen. |
| 21:10.54 | lkcl | repeat. |
| 21:11.04 | cr2 | create a timer. 10ms < t < ? . but i don't know wince programming that good. |
| 21:11.42 | cr2 | lkcl: the latency is too high. it's easier to built sleep() into the current code. |
| 21:12.03 | lkcl | ah. worried about latency. you should have _said_ :) |
| 21:12.59 | cr2 | :) |
| 21:13.59 | cr2 | i don't know how m$ inmplemented sleep(). it may be very braindead :) |
| 21:16.14 | cr2 | afaik there is no arm cpu simulation in qemu. |
| 21:16.34 | iclem | isn't it what the egpio funcs added by llange intend to do ? |
| 21:19.02 | JuNuKN | Yes, thats what I told you at the beginning. I have running it ... |
| 21:20.12 | cr2 | iclem: do you have the source ? |
| 21:21.17 | iclem | let me look |
| 21:21.22 | iclem | http://himalaya.lange.nom.fr/HaRET-asic3.diff |
| 21:22.21 | iclem | http://himalaya.lange.nom.fr/asic3gpio.h (to add in include/) |
| 21:22.50 | iclem | http://himalaya.lange.nom.fr/asic3gpio.cpp (to add in src/) |
| 21:23.16 | iclem | binary: http://himalaya.lange.nom.fr/haret-0.3.6-com1+asic3.exe |
| 21:23.55 | iclem | got to go, bbl |
| 21:24.01 | lkcl | cr2: yes there is. |
| 21:25.51 | cr2 | huh? Dec 19, 2005 QEMU version 0.8.0 is out Support for ARM Integrator/CP board system emulation. |
| 21:26.02 | cr2 | lkcl: i didn't know :) |
| 21:26.34 | lkcl | <PROTECTED> |
| 21:26.34 | lkcl | <PROTECTED> |
| 21:26.34 | lkcl | <PROTECTED> |
| 21:26.34 | lkcl | <PROTECTED> |
| 21:27.03 | cr2 | will it run wince binaries ? :) |
| 21:27.10 | lkcl | ouch |
| 21:27.43 | cr2 | wine for arm. it should be much easier than for x86 ! |
| 21:27.55 | cr2 | no legacy crap. |
| 21:28.05 | lkcl | yeh, someone has already considered it.... mad. completely mad. |
| 21:28.12 | cr2 | :))) |
| 21:29.10 | cr2 | you only need the dynamic loader. the libraries are already there. |
| 21:31.00 | cr2 | "to adapt" the loader. |
| 21:34.07 | lkcl | the asic3_mmc command being issued is "zero". |
| 21:34.09 | lkcl | that's not right. |
| 21:34.19 | lkcl | 0 is "idle" |
| 21:35.00 | cr2 | iclem: ok. it's also busy loop. with the same funny things like asic3 = (uint16 *)memPhysMap (HIMALAYA_ASIC3_PHYS + _IPAQ_ASIC3_GPIO_A_Base ); in a time loop. |
| 21:35.51 | cr2 | lkcl: who is sending command0 ? |
| 21:36.26 | lkcl | ahhh.... yes, it's correct. |
| 21:36.34 | lkcl | mmc_idle_cards. |
| 21:36.36 | lkcl | hmmmm. |
| 21:37.03 | lkcl | and there's no response back. |
| 21:37.29 | cr2 | what for do you need a response ? |
| 21:37.38 | lkcl | 1 mo... |
| 21:37.52 | lkcl | probably an interrupt... :) |
| 21:39.26 | lkcl | schedule_work() will wake up some detection code... |
| 21:40.58 | cr2 | i remember that is was impossible to rmmod asic3_mmc without leaking memory. is it still the case ? |
| 21:41.03 | lkcl | do we know what part number the SD chip is? |
| 21:41.10 | cr2 | ? |
| 21:41.16 | lkcl | well at the moment it's not possible to rmmod asic3_mmc at all |
| 21:41.37 | cr2 | what part number ? |
| 21:41.39 | lkcl | because it's still waiting for the IRQ that will schedule the detection stuff. |
| 21:42.00 | lkcl | yeh. the SD host controller on the himalaya - from photos |
| 21:42.23 | cr2 | htc_partnumbers in wiki. |
| 21:43.06 | lkcl | ack. |
| 21:44.22 | lkcl | oh _right_ - it's really an HTC chip? oh bugger. |
| 21:46.13 | cr2 | there are some toshiba sd docs on hh.org site. |
| 21:47.55 | cr2 | isn't the detection based on its own irq ? |
| 21:50.20 | lkcl | *sigh*... don't know. |
| 21:52.24 | cr2 | there is a detection irq, but i'm not an expert in the mmc code. |
| 21:58.04 | JuNuKN | ... ok, and thanks for now .... I have to go to bed, ... we well have a long night tomorrow .... wish you all a nice switch into the new year ... perhaps we will met next year only ;o) bye |
| 22:01.02 | JuNuKN | bye explicitely cr2, lkcl and iclem , good night and much success this and next year. |
| 22:11.05 | *** join/#htc-blueangel _cronos_ (n=cronos@ron34-1-82-224-149-102.fbx.proxad.net) |
| 22:12.55 | lkcl | heya cronos |
| 22:13.00 | _cronos_ | hie |
| 22:16.48 | Funklord | happy new year in advance =P |
| 22:20.45 | *** join/#htc-blueangel lee (i=lee@i.loathe.ms) |
| 22:25.56 | lkcl | BAStard. i had the config_base and ctrl_base at the wrong sizes in the SDIO code. |
| 22:27.46 | lkcl | FRAAAKKK :) |
| 22:28.27 | lkcl | thinksthinksthinks... |
| 22:28.56 | lkcl | do i double the #defines or do it as a function? |
| 22:29.01 | lkcl | macro, i mean... |
| 22:30.03 | lkcl | ahhh.... the values are memory-mapped and doubled-up... so i think i cannn.... i can multiply by two in the macros. ARGH> |
| 22:46.24 | rob_w | re |
| 22:46.29 | lkcl | BASTARD! |
| 22:46.32 | lkcl | wooo-hoo! |
| 22:46.33 | lkcl | it worked! |
| 22:46.37 | lkcl | i got it, rob. |
| 22:46.43 | lkcl | the sdio. |
| 22:46.49 | lkcl | muhahahhah. |
| 22:50.15 | lkcl | FAKRRRRR muhahah. |
| 22:50.25 | lkcl | yessss that means i can work of the sd-card. |
| 22:50.37 | lkcl | and i can try testing gpe, properly. |
| 22:51.02 | rob_w | lkcl, ur talking himalaya now , right ? |
| 22:51.49 | cr2 | lkcl: is it finally working ? |
| 22:54.48 | lkcl | YEP! |
| 22:55.17 | cr2 | GREAT! |
| 22:55.38 | cr2 | can we get rid of CONFIG_MACH_BLUEANGEL in include/asm/hardware/ipaq-asic3.h ? |
| 22:55.44 | lkcl | awwwwww |
| 22:55.52 | rob_w | cr2, why ? |
| 22:56.03 | lkcl | cr2 - i got other fish to fry. |
| 22:56.06 | lkcl | i wanna do sound, next. |
| 22:56.36 | cr2 | portability. i though bus_shift was supposed to solve this problem. |
| 22:56.49 | lkcl | once it works, i really don't care, unless i'm doing long-term maintenance, it's not pretty, and/or ... yeh, i know. |
| 22:57.06 | cr2 | ok. sound and DoC :) |
| 22:57.10 | lkcl | but none of asic3_mmc.c is designed to use the asic3_xxx functions that are explicitly and specifically for the purpose. |
| 22:57.21 | lkcl | gimme a chance to do a cvs commit first, damnit! |
| 22:57.29 | cr2 | :))))))) |
| 22:59.04 | cr2 | rob_w: include/asm/hardware/ipaq-asic3.h is not our header, we should not put our trash into it. |
| 22:59.55 | rob_w | afterwards i had simliar thoughts ,either the file name was wrong or "we" should not be there |
| 23:00.28 | lkcl | i'm gonna hit it wiv an aaaaaaaammah |
| 23:00.36 | lkcl | i'm gonna it it wiv an aaaammah |
| 23:00.40 | cr2 | the name is not the best, but it _is_ not our business :) |
| 23:00.46 | lkcl | i'm gonna hit it wiv an aaaaammer |
| 23:00.53 | lkcl | and it ain't gonna work no mooorrrrr |
| 23:01.10 | rob_w | lkcl, all ok |
| 23:01.21 | lkcl | oh. |
| 23:01.22 | lkcl | rob. |
| 23:01.23 | lkcl | btw. |
| 23:01.25 | lkcl | the blueangel? |
| 23:01.29 | lkcl | i have board id 0x5. |
| 23:01.35 | rob_w | yeah . what ` |
| 23:01.35 | lkcl | and suspend works _absolutely_ fine. |
| 23:01.44 | lkcl | or i should say: resume works fine. |
| 23:01.47 | rob_w | oh yes |
| 23:01.51 | lkcl | *shrug*. |
| 23:01.55 | lkcl | was _i_ surprised. |
| 23:02.03 | lkcl | and irritated :) |
| 23:02.13 | rob_w | even with your split thingi ? |
| 23:02.40 | lkcl | yes, even with the split thingy. because there's nothing in the post section - i don't think. |
| 23:02.50 | lkcl | that's the way i left it when i was doing the 0x6 code. |
| 23:03.22 | rob_w | sorry , i dont get it also i am too drunken .. to understand anything but clear things |
| 23:03.44 | lkcl | you don't have cvs commit access do you, rob? |
| 23:04.03 | rob_w | but anyway it is interessting to know that the latest cvs kernel does resume to you .. |
| 23:04.09 | rob_w | xanadux cvs ? |
| 23:04.13 | lkcl | yes, xanadux. |
| 23:04.24 | rob_w | ? u did have |
| 23:04.41 | lkcl | yes, i do. i was just wondering if someone had given you cvs commit access to xanadux. |
| 23:04.54 | rob_w | yes i have , |
| 23:05.13 | rob_w | until now i did not use it taht much .. |
| 23:05.19 | lkcl | oh dear. staying away from the keyboard tonight, then? keyboard, drink, keyboard, drink... :) |
| 23:05.28 | rob_w | nice that commit now was at 0:00 |
| 23:05.52 | rob_w | what is it , let me see ... |
| 23:06.54 | rob_w | ~larts himalaya for not having a keyboard |
| 23:07.07 | rob_w | crap |
| 23:07.11 | rob_w | ~lart himalaya for not having a keyboard |
| 23:07.31 | rob_w | hehe |
| 23:08.21 | rob_w | lkcl, why where you asking about my cvs ? |
| 23:12.03 | lkcl | oh, nothing, dude *whistle*. have another drink. |
| 23:12.36 | rob_w | yeah yeah , i know |
| 23:12.48 | rob_w | fill me up with beer |
| 23:12.56 | rob_w | but it wont work , sunny |
| 23:12.58 | rob_w | :-) |
| 23:13.07 | iclem | lkcl: SD working on Himalaya ? |
| 23:13.19 | lkcl | yep, iclem. |
| 23:13.29 | rob_w | ~chant lkcl |
| 23:13.30 | apt | lkcl ! lkcl !! lkcl !! lkcl!!!!!!! |
| 23:13.35 | lkcl | cool, huh? |
| 23:13.40 | iclem | yess really great news !! |
| 23:14.34 | lkcl | doubled the #defines in the macros referring to the ASIC3 stuff, set the hardware address to 0x0c800000 and guessed that the IRQ was 11. |
| 23:14.47 | lkcl | not too difficult - the code's already there (and working on other devices). |
| 23:15.01 | rob_w | lkcl, if you have a working kernel on ba , you my wanna paste it somewhere so i can double check if did introduce that problem somewhere else |
| 23:15.02 | rob_w | bbl |
| 23:15.05 | lkcl | just a friggin pain in the neck for two days trying to work it out. |
| 23:15.17 | lkcl | k. |
| 23:15.26 | lkcl | http://hands.com/~lkcl/blueangel |
| 23:15.33 | lkcl | in a minute. |
| 23:15.49 | rob_w | bookmarked |
| 23:17.19 | lkcl | hmmmmm.... sound... sound.... |
| 23:17.42 | iclem | although I would not be able to make progress the development, I could contribute by cleaning the code a bit and make it portable, if you give me the main line ideas |
| 23:18.00 | rob_w | lkcl, hmmm so you say that is not the kernel ... hmm so some module , which i didn not unload ... ( rob_w loads all stuff on boot |
| 23:19.49 | rob_w | we need to move up the tsc2200 |
| 23:20.20 | lkcl | iclem: talk to cr2 about it, but there's a stack of #defines in asic3_mmc.c which could really do with using asic3_somefunctionorother |
| 23:20.31 | lkcl | asic3_read_blah_blah |
| 23:20.32 | iclem | cr2, how can the code be modified to share a same asic_mmc.c without ifdefs and others quick hacks ? |
| 23:21.00 | iclem | ok |
| 23:22.24 | cr2 | lkcl: why is the irq=11 ? |
| 23:22.25 | lkcl | take a look at asic3_base.c |
| 23:22.31 | lkcl | i have no idea, cr2! |
| 23:22.41 | lkcl | i don't care, either :) |
| 23:22.52 | cr2 | iclem: i think yes. but i have other priorities now. |
| 23:23.27 | cr2 | damn irda!!!!!!!!!!!!!!!!!!! |
| 23:24.28 | cr2 | lkcl: what happend if you set it to 10 ? can you make a dump gpio in wince with inserted sd card ? |
| 23:24.51 | cr2 | iclem: can you also make a dump gpio in wince with the inserted sd card ? |
| 23:32.40 | iclem | only gpio 74 differ with SD card inserted |
| 23:34.45 | cr2 | is gpio11 configured as interrupt ? |
| 23:35.43 | iclem | <PROTECTED> |
| 23:36.26 | iclem | FE means it is an irq ? |
| 23:36.31 | cr2 | hmm. |
| 23:36.34 | cr2 | yes. |
| 23:36.57 | cr2 | why is the wiki data different ? |
| 23:37.01 | iclem | I will update wiki |
| 23:37.36 | cr2 | ok. i have created all other pages, but not this one :) |
| 23:37.42 | iclem | yes in wiki it's not flagged as interrupt |
| 23:41.05 | cr2 | yes. it was probably my mistake. |
| 23:41.45 | iclem | lkcl has already updated it |
| 23:42.06 | cr2 | ok. |
| 23:42.18 | lkcl | yep. |
| 23:42.20 | iclem | time to test this sd access ;) |
| 23:42.59 | cr2 | :) |
| 23:53.26 | lkcl | hmmmm.... uda1380 sound ain't thingy... very up-to-date... hmmm.. |
| 23:53.57 | cr2 | why ? |
| 23:54.10 | lkcl | lots of functions don't exist |
| 23:54.34 | cr2 | the driver or the hardware ? |
| 23:54.36 | lkcl | a higher-order-function table |
| 23:54.38 | lkcl | the driver. |
| 23:55.39 | cr2 | iclem: i'd like to know the IR transceiver model on himalaya. |
| 23:55.55 | lkcl | uda1380_device_controls_t doesn't exist... hmmm. |
| 23:56.27 | iclem | cr2: hm how could I help ? |
| 23:56.40 | cr2 | handhelds.org too ? |
| 23:57.12 | cr2 | iclem: do you know which is it ? |
| 23:57.25 | iclem | no |
| 23:57.34 | cr2 | for ba there is a BOM list... |
| 23:57.49 | iclem | BOM ? |
| 23:57.54 | lkcl | bill of materials |
| 23:58.31 | cr2 | well, maybe these photos can help. http://forum.xda-developers.com/viewtopic.php?t=5002 |
| 23:59.11 | cr2 | wow. i must has ir on my linuxppc mac. |