| 00:00.56 | lkcl | http://wiki.xda-developers.com/index.php?pagename=Universal_Technical |
| 00:00.57 | cr2 | lkcl: what was the ctrl- gsm multiplexing sequence ? |
| 00:01.02 | lkcl | i got some stuff. |
| 00:01.16 | lkcl | erk. ctrl-v AT..... ctrl-b or something like that. |
| 00:01.18 | JuNuKN | Sorry, I have to leave for today ... good night to everybody. |
| 00:01.24 | lkcl | night JuNuKN |
| 00:01.34 | lkcl | look on the himalaya gsm pages. |
| 00:01.36 | lkcl | i think |
| 00:01.51 | cr2 | good night JuNuKN |
| 00:02.23 | cr2 | i have a better picture of universal. but can't identify more chips. |
| 00:03.33 | cr2 | upper left corner. they have left out the agps chip. |
| 00:04.22 | cr2 | there is one more chip in the "sound" box |
| 00:05.44 | cr2 | there is a TI bt chip to the left of the acx100 wlan. |
| 00:07.22 | lkcl | the qualcom rtr6250 is a UTMS thing - there's separate chips for receive and transmit ?!!!! |
| 00:07.48 | lkcl | oh - no. the rtr6250 is a transceiver. whew. |
| 00:08.29 | lkcl | WLAN 802.11b: Texas Instruments TNETW1100B/ACX100. - also on the ipaq H4100 |
| 00:08.35 | cr2 | there is a short description of the 6250 chipset. it is quite capable. camera, 3D accel for CIF, etc. even the gps interface. |
| 00:08.53 | cr2 | yes. and on blueangel... |
| 00:09.04 | cr2 | acx100 |
| 00:10.30 | cr2 | the most intersting will be the data communication with 6250. it is serial over the pxa usb host! |
| 00:11.26 | lkcl | damn. |
| 00:13.46 | lkcl | http://www.opencores.org/forums.cgi/cores/2004/10/001282 |
| 00:14.10 | lkcl | mentions a coolrunner II 2c128 - from xilinx - as having only 128 flip-flops. |
| 00:14.18 | lkcl | i.e. it's a really small fpga. |
| 00:14.26 | lkcl | that's gonna be fun. |
| 00:15.49 | cr2 | i think it is only for some gpio and bt/wlan mux support. |
| 00:16.03 | lkcl | well - we could just shove data down the SPI interfaces and see what comes back. |
| 00:16.27 | lkcl | or just look in the .dlls and see what they do. |
| 00:16.35 | cr2 | we can just look closer at the touchscreen driver. |
| 00:16.45 | lkcl | i have a license for IDA-Pro (i only ever buy compilers and disassemblers :) |
| 00:16.58 | cr2 | linux version ? |
| 00:17.03 | lkcl | oh. and three copies of dark reign 2 and two copies of dark reign 1. |
| 00:17.15 | cr2 | :)) |
| 00:17.22 | lkcl | friggin windows unfortunately - this _was_ two years ago, they didn't have a linux version. |
| 00:17.32 | lkcl | bastards. |
| 00:17.57 | cr2 | the arm support was not that great. but better than objdump :) |
| 00:19.18 | cr2 | i have actually learned how to use objdump in a productive way ;) |
| 00:19.57 | cr2 | can't you just get an upgrade at no cost ? |
| 00:21.07 | lkcl | i have no idea. i haven't had a need to use it since i stopped using it 18 months ago. i'll ask them |
| 00:24.56 | cr2 | worth a try. |
| 00:27.00 | lkcl | sent a message to pierre@datarescue.com... |
| 00:27.02 | lkcl | anyway. |
| 00:27.10 | lkcl | can't keep doing late nights. |
| 00:28.50 | cr2 | i 'll try the haret irda and will go to bed. great. found the soundchip on hima |
| 00:29.27 | lkcl | wheeee :) |
| 00:49.37 | cr2 | it does not work. need to think out something else. |
| 00:55.45 | cr2 | lkcl: i'm going to bed. good night. |
| 00:56.28 | *** join/#htc-blueangel Funklord (n=cow@c-cbd572d5.014-46-73746f28.cust.bredbandsbolaget.se) |
| 01:49.46 | *** join/#htc-blueangel MC-JetLi (n=jet@ppp-69-222-183-92.dsl.chcgil.ameritech.net) |
| 05:36.06 | *** join/#htc-blueangel MC-JetLi (n=jet@ppp-69-222-183-92.dsl.chcgil.ameritech.net) |
| 08:18.03 | *** join/#htc-blueangel rob_w-cgi (n=txuser@dd1010.kasserver.com) |
| 08:18.03 | *** mode/#htc-blueangel [+o rob_w-cgi] by ChanServ |
| 08:18.38 | rob_w-cgi | morning |
| 11:21.42 | lkcl | mornin |
| 11:31.18 | *** join/#htc-blueangel gherdo (n=gherdo@193.41.76.133) |
| 11:31.31 | gherdo | morning |
| 11:32.49 | rob_w-cgi | he all |
| 11:33.09 | rob_w-cgi | i am testing a bluetooth hack soon to toggle the blue led |
| 11:45.56 | lkcl | oh dear :) |
| 11:51.41 | rob_w-cgi | cool feels like working |
| 13:41.12 | *** join/#htc-blueangel iclem (n=clement@iclem.net) |
| 14:31.51 | *** join/#htc-blueangel pH5 (n=ph5@p5485FE3B.dip.t-dialin.net) |
| 14:32.37 | rob_w-cgi | hi pH5 |
| 14:32.56 | pH5 | hi rob_w-cgi |
| 14:33.05 | cr2 | hi guys |
| 14:33.19 | pH5 | hi cr2. did that haret build work? |
| 14:33.46 | cr2 | the binary is working, but no irda output :( |
| 14:33.51 | pH5 | :-/ |
| 14:34.09 | cr2 | i can't link the wince binaries for some reason. |
| 14:35.01 | cr2 | gnu zealotry, baby! i want haret to be compiled like that: http://trac.videolan.org/vlc/file/trunk/INSTALL.wince |
| 14:40.21 | cr2 | need to get rid of gapi, make some minor changes, add the uart_putc() in asm and write a normal Makefile |
| 14:41.44 | cr2 | cool. arm-wince-pe-gcc-4.1.0 |
| 15:05.26 | cr2 | rob_w: something FYI http://nsis.sf.net |
| 15:10.32 | rob_w-cgi | oh dear |
| 15:13.16 | rob_w-cgi | i am refusing to learn this ! |
| 15:14.18 | rob_w-cgi | ok ok .. so lets see ,. it should be easy to get a initrd which , decreases , formats and copy`s the gpe image on one boot. |
| 15:14.50 | rob_w-cgi | but not this week ! |
| 15:15.23 | rob_w-cgi | we need to hunt the resume problems .. those are real showstoppers currently |
| 15:15.48 | rob_w-cgi | if we have suspend working again then this is a almost usable thingi ( with some more love to gomunicator |
| 15:21.08 | cr2 | i am not saying you should learn this. but i think we need to get rid of "it may be under /dev/mmc on your machine" and "These two commands may differ for your system" |
| 15:22.11 | cr2 | it requires quite a good knowledge of linux to understand the meaning |
| 15:25.09 | rob_w-cgi | agreed .. actually the mmc path is on a host is not the same always |
| 15:25.36 | rob_w-cgi | still those guys having a linux host will manage i think .. others without any linux knowledge will suffer |
| 15:26.56 | rob_w-cgi | who is actually left here with a BA ? |
| 15:28.13 | cr2 | i'm trying to fdisk a file. fdisk asks me to provide a cylinder count in a funny translation: Sie müssen angeben Zylinder. i 'll switch to LANG=C |
| 15:28.37 | cr2 | what should i provide for the SD partition ? |
| 15:29.05 | rob_w-cgi | ? |
| 15:29.13 | rob_w-cgi | hmm |
| 15:29.27 | rob_w-cgi | what size of sd do you target ? |
| 15:29.47 | cr2 | both 128 and 256. then 512 |
| 15:30.46 | rob_w-cgi | so you are creating a generic image for all of them ? |
| 15:33.12 | cr2 | i want to do: fdisk ; mkdosfs; mke2fs ; mount -o loop; cp ; tar xfvzp ; umount. then the image can be written directly to the sd on windows with activesync. |
| 15:33.31 | cr2 | then we will target writing on wince , and with synce. |
| 15:35.11 | cr2 | is 128 enough ? i don 't has a 256MB card myself. only 512 |
| 15:37.40 | cr2 | fdisk says now: Disk SDIMAGE: 255 heads, 63 sectors, 0 cylinders |
| 15:37.42 | rob_w-cgi | 128 is fully enough for gpe |
| 15:38.03 | cr2 | i think SD has 512byte blocks ? |
| 15:38.27 | rob_w-cgi | but that procedure is to have a linux host , right ? |
| 15:38.28 | cr2 | for gpe and dos/haret ? |
| 15:38.35 | rob_w-cgi | also |
| 15:39.38 | cr2 | well. we can create the images for download. |
| 15:40.35 | cr2 | compressed, of course. |
| 15:41.44 | cr2 | hmm. not all 128MB sd cards have 128MB or the same amount. we need to find a minimum. |
| 15:42.43 | rob_w-cgi | cr2 so what device do you have, universal ? |
| 15:42.56 | cr2 | yes. |
| 15:43.03 | cr2 | and headless BA. |
| 15:43.33 | rob_w-cgi | ah right the headless one .. why dont you ship it to warranty or did they refuse ? |
| 15:43.48 | cr2 | there are other options. like openzaurus. install base 16MB and update over the net. |
| 15:44.13 | cr2 | they will refuse. it is an obvious mechanical damage |
| 15:44.43 | rob_w-cgi | ah ,, too bad |
| 15:44.51 | cr2 | i have already thrown LCD away. |
| 15:45.35 | cr2 | one company has sent me the wrong LCD. |
| 15:46.04 | cr2 | i've sent it back, because i have explained them at the very beginning: i need Sony. |
| 15:48.23 | rob_w-cgi | ah crap |
| 15:48.54 | cr2 | it is possible to order the LCD in hongkong, but i despise credit cards. And a guy from xda-dev also had problems with the wrong LCD. |
| 15:50.06 | rob_w-cgi | tsts , feels like real crap ... also it seems i am one lonely man left with a working BA here , OR WHAT ? |
| 15:50.21 | cr2 | cp15 :) |
| 15:52.46 | rob_w-cgi | yeah and him i feel guilty as did not contributed in a noticeable amount |
| 15:53.15 | cr2 | i can boot the headless gpe :) but then i will need IR to control it anyway. |
| 15:53.42 | rob_w-cgi | how about investing some test to suspend with it ?!? |
| 15:53.51 | rob_w-cgi | via ssh via docking station ß |
| 15:54.34 | cr2 | no. it is not so easy. wince will not boot, so no haret for me. i must go the firmware update route. need cp15 for that. |
| 15:54.47 | cr2 | ;) |
| 15:56.02 | cr2 | then i must either boot it with the HTC bootloader or flash a new bootloader which does not exist yet. like on hx4700 |
| 15:58.03 | cr2 | it is doable, but not alone. |
| 15:58.28 | rob_w-cgi | argh |
| 15:58.39 | rob_w-cgi | then never mind |
| 15:59.44 | cr2 | i hope to meet him tomorrow, so we will see... |
| 15:59.53 | rob_w-cgi | running home .. say hello to cp15 !! |
| 16:50.20 | *** join/#htc-blueangel iclem (n=clement@iclem.net) |
| 17:37.32 | iclem | ~seen JuNuKN |
| 17:37.47 | apt | junukn <n=JuNuKN@Md7a8.m.pppool.de> was last seen on IRC in channel #htc-blueangel, 17h 36m 29s ago, saying: 'Sorry, I have to leave for today ... good night to everybody.'. |
| 17:38.13 | iclem | hello all |
| 17:41.44 | *** join/#htc-blueangel rob_w (n=bob@p85.212.195.154.tisdip.tiscali.de) |
| 17:41.45 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 17:55.42 | rob_w | nice ,, got a 2nd battery for the ba for little money |
| 18:01.41 | *** join/#htc-blueangel lkcl (n=lkcl@bb-87-82-4-146.ukonline.co.uk) |
| 18:01.49 | lkcl | hi again. |
| 18:01.53 | lkcl | i got my xda-2 back. |
| 18:04.41 | rob_w | himalaya ? |
| 18:04.45 | lkcl | yep! |
| 18:04.50 | rob_w | ohh |
| 18:11.15 | cr2 | lkcl: what are your plans ? |
| 18:13.31 | lkcl | cr2: THAT'S IT! you need to put mem=63M on the kernel-boot line |
| 18:13.44 | lkcl | i got a boot on the universal. |
| 18:13.59 | cr2 | with un0 ? |
| 18:14.10 | lkcl | no with the generic one. |
| 18:14.21 | lkcl | will try the un0 one next. |
| 18:14.42 | lkcl | it's terminated at the serial initialisation stage. |
| 18:14.48 | cr2 | where do i get the compiled generic kernel ? |
| 18:14.52 | lkcl | thinks that ttyS0 is at 0x40100000 |
| 18:14.59 | lkcl | ttyS1 at 0x40200000 |
| 18:15.07 | lkcl | and ttyS2 at 0x40700000 |
| 18:15.25 | cr2 | it is right. |
| 18:15.54 | lkcl | but yeh: the key (thanks to cp15) was putting mem=63M in the startup.txt kernel params line. |
| 18:16.00 | lkcl | hoo friggin rah. |
| 18:16.24 | lkcl | damn the screen is small with a 4x6 font. |
| 18:16.41 | cr2 | hmm. how is the LCD wired then ? |
| 18:16.53 | lkcl | have a look at htcuniversalfb.c |
| 18:17.00 | iclem | hi all |
| 18:17.05 | lkcl | hiya iclem |
| 18:17.29 | iclem | lkcl: have you tried last 2.6.12 kernel on himalaya ? |
| 18:17.37 | lkcl | okay: i am spoiled for choice, now. |
| 18:17.51 | lkcl | no iclem, i haven't. i just got it back about... 2 hours ago. |
| 18:17.57 | iclem | ok ;) |
| 18:18.11 | lkcl | am spoiled for choice.... so i think i'll go make some tea instead :) |
| 18:18.20 | cr2 | you can test the sound and DoC. |
| 18:18.25 | lkcl | yep :) |
| 18:18.35 | lkcl | tea first. |
| 18:28.25 | cr2 | i get the blank white screen with un0 and mem=63M. not surprising. pxafb has no idea about this strange framebuffer layout. need a console... |
| 18:29.09 | cr2 | i think the same applies to pH5 with the magician. |
| 18:34.01 | pH5 | cr2: what is un0? |
| 18:35.34 | cr2 | the universal branch in CVS. |
| 18:36.54 | pH5 | ouch, right :) |
| 18:38.12 | cr2 | we have ba0, waiting for ma0 :) |
| 18:41.14 | lkcl | no, you need htcuniversalfb.c not pxafb.c |
| 18:41.48 | lkcl | i've put in some untested code that appends the two memory locations together that make up the framebuffer. |
| 18:43.40 | cr2 | i have compiled with htcunversalfb.c. don't see anything. can you dcc the compiled binary ? |
| 18:44.07 | lkcl | _what_??? :) |
| 18:44.20 | lkcl | ok. |
| 18:44.25 | lkcl | i may have got the code wrong. |
| 18:44.35 | lkcl | do a diff of vsfb.c and htcuniversalfb.c |
| 18:44.50 | lkcl | basically there are two memory areas at different physical memory locations. |
| 18:45.05 | lkcl | and they need to end up at _contigous_ virtual memory locations. |
| 18:45.12 | lkcl | if you know how to do that, then the problem is solved. |
| 18:45.28 | lkcl | the code that i've added is untested guess-work. |
| 18:45.41 | cr2 | i have seen your code. |
| 18:45.57 | lkcl | with one of the functions being a "non-public" function that you're not supposed to use :) |
| 18:47.55 | cr2 | you can reserve two adjacent areas at htcuniversal_map_io stage ? |
| 18:48.17 | cr2 | or is it too late ? |
| 18:48.17 | lkcl | haven't a clue. hmmm... interesting thought though... |
| 18:48.34 | cr2 | no. with mem=63M it is not too late |
| 18:52.47 | lkcl | splat... |
| 18:53.00 | lkcl | hmmm.... |
| 18:53.01 | rob_w | lkcl, if this is need more often we could get it as kernel boot command |
| 18:53.33 | lkcl | needed what? sorry, i no unnerstand |
| 18:53.47 | rob_w | i had my bluetooth led nicely toggled by the gpe-bluetooth applet :) |
| 18:53.56 | rob_w | brb |
| 18:53.58 | lkcl | he he |
| 18:54.48 | lkcl | no - i definitely got that wrong, cr2. |
| 18:55.29 | lkcl | bugger. |
| 18:56.20 | lkcl | oka. |
| 18:56.22 | lkcl | okay. |
| 18:56.46 | lkcl | need to remap a couple of pages, in pxa_map_io. |
| 18:57.16 | rob_w | lkcl, i thought that this contigious memory is something for a boot command line |
| 18:57.43 | lkcl | no, it's two areas of memory that comprise the screen buffer. 2*640*480 bytes. |
| 18:58.04 | rob_w | why like this ? |
| 18:58.30 | cr2 | rob_w: it is for pxa27x owners :) |
| 18:58.41 | lkcl | because there wasn't enough memory on-chip so they had to have an extra 0x40000 bytes off-chip!!! |
| 18:58.52 | lkcl | _i_ don't know :) |
| 18:59.00 | cr2 | who have on-cpu sram. |
| 18:59.16 | cr2 | but only 256K |
| 18:59.26 | lkcl | okay gimme a few mins to sort this out... |
| 19:05.05 | lkcl | okay i'm putting the virtual memory locations at 0xf000 0000 just because i can, which makes the address of the screen at 0xf0aa 0000. |
| 19:09.01 | lkcl | about the only thing i can think of is that they made the chip, decided the screen was to be 480x360 and then changed their minds, and made it 640x480 instead. |
| 19:10.47 | cr2 | memory costs money. and not every phone needs 640x480x16bit. look at msm6250 |
| 19:11.05 | lkcl | _no_ phone needs 640x480!!!!! |
| 19:11.46 | lkcl | i mean - _damn_ - you could watch telly on the universal, and get xine to run quite nicely on it, thank you |
| 19:13.14 | cr2 | :) |
| 19:13.38 | lkcl | i mean... it'd hurt your eyes as it's too small, but... |
| 19:14.17 | cr2 | i want to see a good map in qpegps. i don 't use gsm/umts that much :) |
| 19:14.55 | lkcl | request_mem_region - any ideas how it works? |
| 19:16.53 | lkcl | hmmm... let's suck it and see... |
| 19:18.50 | lkcl | run.... |
| 19:18.55 | lkcl | aaaanndd... |
| 19:19.00 | lkcl | splat. |
| 19:19.01 | lkcl | damn. |
| 19:20.43 | lkcl | tryinbg the whoops |
| 19:24.59 | lkcl | trying the "generic" one... |
| 19:25.21 | lkcl | nope.... i've messed that up, too, now, as well :) |
| 19:25.50 | cr2 | why do you need request_mem_region ? |
| 19:26.18 | lkcl | well you don't _need_ it but if you don't use it then the area of memory might get allocated for general purpose use |
| 19:26.23 | lkcl | i guess. |
| 19:27.53 | cr2 | look at arch/arm/mach-sa1100/jornada720.c |
| 19:28.47 | lkcl | yep. |
| 19:29.10 | lkcl | frame buffer of size 0x00100000. we have twice that. |
| 19:29.15 | lkcl | i'm working on it ... :) |
| 19:30.41 | lkcl | okay - e1356fb_init doesn't do request_mem_region - they just use it. |
| 19:30.45 | lkcl | okay, i can live with that :) |
| 19:32.05 | lkcl | run, go onnnnn.... |
| 19:32.17 | lkcl | darn. |
| 19:32.27 | lkcl | okay - brb, gotta go say hi to wife... |
| 19:35.06 | cr2 | static struct map_desc jornada720_io_desc[] |
| 19:48.55 | lkcl | silly bugs.... or not pressing the button YAAY! |
| 19:48.57 | lkcl | it worked. |
| 19:52.47 | cr2 | what ? |
| 20:03.50 | lkcl | i got a full black screen. |
| 20:04.08 | lkcl | uploading a patch to hands.com/~lkcl/universal, now. |
| 20:04.14 | lkcl | gotta go eat. back later. |
| 20:14.52 | *** join/#htc-blueangel eme (n=bosty@BSN-210-231-5.dial-up.dsl.siol.net) |
| 20:14.57 | eme | hi all |
| 20:15.03 | rob_w | hey eme |
| 20:15.14 | eme | see on wiki that you need some logs for suspend |
| 20:15.24 | rob_w | oh yes |
| 20:15.31 | eme | for power managment |
| 20:15.39 | eme | i can install right now if you want |
| 20:15.41 | rob_w | i worked some wiki things today |
| 20:15.54 | eme | joust tell me wich log do you need |
| 20:15.56 | rob_w | sure , lets see if you get logs , i didnt |
| 20:16.19 | eme | ok will download from wiki and put ond sd now |
| 20:16.38 | rob_w | make sure the actual files com from gnulinux.biz |
| 20:16.51 | eme | ok |
| 20:17.03 | eme | link from wiki is ok or not? |
| 20:17.12 | rob_w | yes |
| 20:17.20 | eme | tks |
| 20:18.53 | lkcl | YAAY. got a boot screen for CONFIG_HTC_UNIVERSAL. |
| 20:18.54 | lkcl | hurrah. |
| 20:19.14 | pH5 | lkcl: congrats |
| 20:19.22 | eme | great |
| 20:19.33 | lkcl | a full screen, not this thingy rubbish :) |
| 20:19.37 | lkcl | half-screen |
| 20:19.52 | lkcl | good idea on the io resources, cr2. much simpler. |
| 20:20.50 | cr2 | lcd memory is io resource. where is the patch ? :) |
| 20:21.06 | lkcl | burble? |
| 20:21.14 | lkcl | http://hands.com/~lkcl/universal/f |
| 20:21.58 | cr2 | ok. |
| 20:23.07 | rob_w | nice |
| 20:30.58 | lkcl | has the universal got an ASIC3? |
| 20:32.26 | lkcl | answers on a postcard... or at xdadev wiki: page UniversalASIC3 :) |
| 20:38.52 | rob_w | how is it going eme ? |
| 20:40.48 | lkcl | WHOA, cr2 - something not quite right in htcuniversal_core.c or maybe in the asic3 code :) |
| 20:41.29 | rob_w | lkcl , what was that discussion you had with cp15 about suspend the the order of lcd vs. framebuffer ... or so ? |
| 20:41.33 | lkcl | i just ended up with the screen being reset to "white" and also the camera bright-light being switched on :) |
| 20:41.37 | lkcl | oh. yes. |
| 20:41.47 | lkcl | the sequence is very specific. |
| 20:41.55 | lkcl | including timings. |
| 20:42.18 | lkcl | i will cvs commit what i have, so you can see. |
| 20:43.01 | rob_w | currently the lcd and backlight comes back but the unit freezes right after ,, no time to interact on resume , also it looks are there a no logs at all |
| 20:43.29 | lkcl | you have to do two of the GPIOs, followed by a specific delay, followed by one more, followed by a delay, followed by the w100fb initialisation sequence, followed by another delay, followed by another GPIO activation, followed by another delay, followed by yet another GPIO activation. |
| 20:43.32 | rob_w | so do you think that it the problem which locks up on resume ?? |
| 20:43.33 | lkcl | and then you're done. maybe. |
| 20:43.38 | lkcl | i haven't a clue. |
| 20:44.15 | lkcl | i had it working - mostly - in that i could press the power button to reactivate the device, do ifup usb0, press return and it'd be okay. |
| 20:44.24 | lkcl | but i DID NOT allow the device to auto-suspend |
| 20:44.26 | rob_w | to me there is not much left ( beside the asic`s ) as i ve unloaded and tryed with very little modules |
| 20:44.36 | lkcl | i was ssh'd in to the box and MANUALLY doing apm --suspend |
| 20:44.53 | lkcl | yes, it's going to be with timings. |
| 20:45.12 | lkcl | have you got to the point where the framebuffer itself is a module? |
| 20:45.26 | lkcl | because if you have, that could prove a very useful test. |
| 20:45.38 | lkcl | do modprobe w100fb |
| 20:45.42 | rob_w | no didnt get that far |
| 20:45.44 | lkcl | and rmmod w100fb |
| 20:45.48 | lkcl | and modprobe w100fb again. |
| 20:46.00 | rob_w | hang on a sec |
| 20:46.09 | lkcl | that's the ultimate test, to see if it's something to do with suspend or to do with the framebuffer code. |
| 20:46.58 | lkcl | if you can't do that, then the other test is to write a new module which interacts with the framebuffer one and calls functions in it - the initialisation code. |
| 20:47.13 | lkcl | to re-initialise the framebuffer. |
| 20:47.23 | cr2 | lkcl: i get an lcd reset and then the white screen. |
| 20:47.39 | lkcl | yeh, the white screen means that the initialisation has not succeeded. |
| 20:47.40 | cr2 | we don't need htcuniversal_core.c |
| 20:48.11 | lkcl | the w100fb code has not successfully woken up the ATI chip. |
| 20:48.39 | lkcl | last time i looked at it i thought it might be due to the GPDRs and GAFRs etc. not being initialised in the correct directions etc. |
| 20:48.54 | lkcl | which is why it's so important to make w100fb etc. all modules. |
| 20:49.02 | lkcl | not compiled-in. |
| 20:49.20 | lkcl | alternatively, some printks would do the trick, before suspend and after resume. |
| 20:49.46 | lkcl | but you _must_ run apm --suspend and _not_ let that stupid gpe code get hold of the suspend stuff. |
| 20:49.46 | cr2 | lkcl: but then the kernel runs up to the fb init on universal. not bad. |
| 20:50.07 | lkcl | oh - it works, it's useable! it's just that suspend on the LCD is buggered. |
| 20:50.31 | lkcl | cp15 had it going to "partial sleep" and got an 18 hour battery life out of it - basically by switching off the backlight. |
| 20:50.41 | lkcl | and most other devices. |
| 20:50.53 | rob_w | lkcl, but why then is that the only thing which seems to come back on resume ,, as i ve backlight and a pictures ?!? |
| 20:51.23 | lkcl | your sentence is grammatically incorrect, rob - i don't understand it. |
| 20:51.24 | lkcl | sorry. |
| 20:51.38 | rob_w | hehe |
| 20:51.43 | rob_w | ok , so |
| 20:52.38 | rob_w | i dont understand why the bug to resume is inside the video parts , as they actuallly look as if they would resume ok . All the rest is frozen |
| 20:52.43 | rob_w | .. better ? |
| 20:53.01 | lkcl | the bit i don't understand: "all the rest is frozen". |
| 20:53.13 | lkcl | what do you mean, "all the rest is frozen"? |
| 20:53.22 | lkcl | because when i was testing this, i had no "other stuff frozen". |
| 20:53.24 | lkcl | it worked. |
| 20:53.29 | lkcl | everything but the screen. |
| 20:53.33 | rob_w | ahm the BA freezes ( not ice or snow ) on resume |
| 20:53.37 | lkcl | i ssh'd in. |
| 20:53.41 | lkcl | i ran apm --suspend |
| 20:53.47 | lkcl | i then pressed the power button. |
| 20:53.54 | lkcl | i then, on my laptop, did ifup usb0 |
| 20:54.05 | lkcl | i then pressed return a couple of times on the ssh session and hey presto, it was back. |
| 20:54.10 | lkcl | only, with a white screen. |
| 20:54.30 | lkcl | and if you did echo ajdhsadasldj > /dev/fb0 |
| 20:54.30 | lkcl | and then hexdump -C /dev/fb0 |
| 20:54.39 | lkcl | you could see aashjdahskdjahkj as the first line of the hexdump. |
| 20:55.09 | lkcl | so it works - as long as you DO NOT let the gpe power management daemon do the suspend. |
| 20:55.14 | lkcl | does that make sense? |
| 20:55.23 | rob_w | so you say your device was able to sshed after resume ... |
| 20:55.27 | lkcl | what are _you_ doing that makes "all the rest freez". |
| 20:55.30 | rob_w | i was using apm -s ?!? |
| 20:55.56 | lkcl | no, i never re-ssh'd back in after a resume. |
| 20:56.04 | lkcl | i typed "ifup usb0" on my laptop. |
| 20:56.07 | rob_w | i am so confused now that i actually need to test again if i can ssh in |
| 20:56.15 | lkcl | i then pressed returnb a few times ON THE ORIGINAL SSH SESSION. |
| 20:56.20 | lkcl | i did NOT "ssh back in". |
| 20:56.27 | rob_w | yes |
| 20:56.35 | lkcl | when the device suspends, your laptop detects that the usb has gone to sleep. |
| 20:56.38 | rob_w | so the ssh was there on suspend and reactivated ;-) |
| 20:56.50 | lkcl | it then REMOVES the usbnet module (ifconfig usb0 down). |
| 20:57.04 | lkcl | you must MANUALLY reactivate the usb0 interface - ifup usb0. |
| 20:57.14 | lkcl | ifconfig usb0 up. whatever. |
| 20:57.19 | rob_w | sure |
| 20:57.26 | lkcl | then you can carry on as if nothing had happened. |
| 20:57.36 | rob_w | .. eh so that was during the ba was "suspended " ? |
| 20:58.03 | rob_w | hang on hang .. |
| 20:58.05 | lkcl | when the BA is suspended, your laptop detects that it has gone away, and automatically does an ifconfig usb0 down, yes. |
| 20:58.33 | rob_w | ok so now i am sshed in .. so i do the magic apm -s ? |
| 20:58.42 | lkcl | if you do lsmod | grep usbnet on your laptop whilst the BA is suspended, you will find "0" uses or quite possibly that usbnet is missing. |
| 20:58.43 | lkcl | yep. |
| 20:58.56 | lkcl | then press the power button to wake it up |
| 20:59.08 | lkcl | DON'T press any keys in the ssh console window. |
| 20:59.16 | rob_w | ok , ba is sleeping(tm) |
| 20:59.19 | lkcl | go to a _separate_ console on your laptop. |
| 20:59.27 | lkcl | press the power button to resume. |
| 20:59.34 | lkcl | log in as root on your laptop |
| 20:59.41 | lkcl | type "ifconfig usb0" |
| 20:59.46 | lkcl | and you will see that the interface is down. |
| 20:59.57 | rob_w | yep down |
| 21:00.01 | lkcl | bring the interface up - ifup usb0 on my debian/stable system. |
| 21:00.23 | rob_w | ba is still sleeping -> usb0 no such device |
| 21:00.24 | lkcl | if you're using some bullshit os like deadrat you'll need some different voodoo magic to bring the interface up. |
| 21:00.35 | lkcl | did you press the power button to wake it up? |
| 21:00.39 | rob_w | tahts fine this i can handle |
| 21:00.43 | rob_w | no |
| 21:00.46 | lkcl | you haven't pressed the power button have you :) |
| 21:00.51 | rob_w | no |
| 21:00.52 | rob_w | :-) |
| 21:00.53 | lkcl | then of course it's not there :) |
| 21:00.59 | rob_w | :-( |
| 21:01.06 | rob_w | pressing |
| 21:01.16 | lkcl | i did say - twice :) |
| 21:01.23 | lkcl | okay - white screen? |
| 21:01.35 | rob_w | i didnt believed it for both times |
| 21:01.45 | lkcl | have faith, dude :) |
| 21:01.49 | rob_w | no still black ?!?! hang on |
| 21:02.02 | lkcl | press it only once... |
| 21:02.21 | lkcl | twice puts it into suspend again... |
| 21:02.25 | rob_w | ah crap ... i didnt had the blueangel_suspend.ko loaded .. now it doesnt come back at all |
| 21:02.30 | lkcl | oops :) |
| 21:02.54 | rob_w | ok we try again in 2 min |
| 21:03.08 | rob_w | then you show me your /dev/fb0 voodoo |
| 21:03.16 | lkcl | okay, try it again - oh, and do a cvs update (or see the patch at... http://hands.com/~lkcl/blueangel/lcd_stuff.patch |
| 21:03.30 | lkcl | look in the irc logs, or page-up!!! |
| 21:03.39 | lkcl | what kind of irc client are you using, dude?? :) |
| 21:04.03 | lkcl | how many lines on your irc screen? |
| 21:04.14 | rob_w | about 12 ?? |
| 21:04.26 | rob_w | yeah ok ok |
| 21:04.28 | lkcl | twelve?!?!?! you need a better irc client :) |
| 21:04.38 | rob_w | silence |
| 21:04.48 | lkcl | try bitchx - it's _rubbish_ but there's a page-up thing mmhmhmmhmhmmhmmh |
| 21:04.57 | lkcl | mhmmmmhmmmmhnnnn! |
| 21:05.08 | lkcl | mmmm! |
| 21:05.29 | eme | rob_w: booting... |
| 21:05.34 | rob_w | ok eme |
| 21:05.54 | rob_w | ok lets see if that cvs is already available |
| 21:06.32 | lkcl | mmm mmmm mhhhmm! |
| 21:07.03 | rob_w | ooops i dumped the whole kernel ,, |
| 21:11.03 | cr2 | lkcl: the flash led is soooo bright :) |
| 21:11.31 | eme | amm, which modul on host for usbnet? |
| 21:12.05 | lkcl | tell me about it, cr2... |
| 21:12.08 | cr2 | the vibra and flash led gpios are not in universalasic3 page. |
| 21:12.55 | cr2 | i get a white screen and switch on flash led with the current code. we need to set the same values on asic3 as wince... |
| 21:13.06 | lkcl | ack. |
| 21:13.28 | lkcl | any good ideas on how to get them? |
| 21:13.29 | cr2 | i have set the direction and alt as on my universal. |
| 21:13.35 | lkcl | ah :) |
| 21:13.52 | cr2 | i have then, and other have sent me their. |
| 21:14.06 | cr2 | what is your board id ? 5 ? |
| 21:14.20 | lkcl | clueless. |
| 21:14.53 | rob_w | with your patches do i still need to run w100fb as module ? |
| 21:15.28 | cr2 | lkcl: aah. init=0 for all... |
| 21:15.50 | cr2 | what should i put there. |
| 21:16.20 | cr2 | can you boot wince and try with haret: pd 0x10000000 0x400 ? |
| 21:21.32 | eme | rob_w: which log should I check? |
| 21:24.22 | rob_w | i had to logread > mylogfile to get a backup |
| 21:24.48 | rob_w | but it seems as logread will not catch the suspend/resume circle`s |
| 21:24.54 | rob_w | or whatever |
| 21:27.04 | eme | so need to wait for susspend or can send it maunaly ? |
| 21:27.49 | rob_w | apm -s |
| 21:28.04 | rob_w | hang on |
| 21:28.19 | rob_w | do you have blueangel_suspend loaded ? |
| 21:28.38 | rob_w | u need that to resume |
| 21:28.56 | eme | not yet |
| 21:29.01 | eme | modprobe ? |
| 21:29.26 | rob_w | yep |
| 21:29.30 | eme | done |
| 21:29.59 | rob_w | so do have ssh currently ? |
| 21:30.04 | eme | jp |
| 21:30.07 | lkcl | cr2: done. |
| 21:30.10 | rob_w | ok then apm -s |
| 21:30.11 | eme | yes logged in |
| 21:30.14 | eme | ok |
| 21:30.34 | eme | root@blueangel:~# apm -s |
| 21:30.34 | eme | -sh: apm: not found |
| 21:30.41 | cr2 | lkcl: i have put the correct (my) init values for asic3. |
| 21:30.47 | rob_w | su - ? |
| 21:30.48 | lkcl | just upload... to... hands.com/~lkcl/universal |
| 21:30.57 | lkcl | it's probably renamed to apm.old, eme. |
| 21:31.07 | lkcl | or apm.disable |
| 21:31.16 | eme | yes it is |
| 21:31.25 | lkcl | so either run apm.disable -s |
| 21:31.28 | lkcl | or rename it. |
| 21:31.34 | eme | blank screen |
| 21:31.42 | rob_w | black =? |
| 21:31.47 | eme | black |
| 21:31.53 | eme | and ssh freez |
| 21:31.55 | lkcl | cr2: you got them? in universal.c, now? |
| 21:31.55 | rob_w | nice .. |
| 21:32.00 | eme | on host |
| 21:32.02 | rob_w | keep the ssh terminal |
| 21:32.06 | eme | ok |
| 21:32.13 | rob_w | wake it up |
| 21:32.17 | lkcl | don't touch any keys in the ssh terminal, eme. |
| 21:32.20 | lkcl | yet. |
| 21:32.27 | eme | roger |
| 21:32.43 | cr2 | damn. removed the wrong directory |
| 21:32.50 | eme | grey screen |
| 21:32.57 | lkcl | cr2: http://hands.com/~lkcl/universal/typescript.haret.pd.0x10000000 |
| 21:33.02 | lkcl | eme: GOOD. |
| 21:33.07 | rob_w | hmm ... ifconfig your host usb0 up |
| 21:33.08 | cr2 | lkcl: ok. |
| 21:33.15 | lkcl | okay. now do "ifup usb0" on a separate window. or whatever. |
| 21:33.31 | lkcl | check the ip address with some pings |
| 21:33.44 | lkcl | and then if that responds, press a couple of keys on the ssh window. |
| 21:33.51 | lkcl | don't leave it too long! |
| 21:33.51 | rob_w | eme report ? |
| 21:34.00 | eme | second plz |
| 21:34.05 | lkcl | because the gpe "suspend" might kick in soon. |
| 21:34.20 | lkcl | cr2: you got it? |
| 21:34.31 | lkcl | you know what you're looking at? |
| 21:34.31 | rob_w | killall ipaq-sleep will prevent auto suspend |
| 21:34.39 | lkcl | that's the one. |
| 21:34.44 | lkcl | forgot what it was called. |
| 21:35.07 | eme | cant get usb0 on host |
| 21:35.17 | eme | no such device |
| 21:35.27 | cr2 | lkcl: your gpioc is different. i need to write down the bits to decode it. |
| 21:35.30 | rob_w | eme how is the screen looking on the ba ? |
| 21:35.43 | eme | clean but grey |
| 21:35.56 | rob_w | ~orders one small beer please |
| 21:37.05 | rob_w | hmm to me i saw the gpe-session ?? did you had gpe running or only kernel console ? |
| 21:37.18 | eme | gpe too |
| 21:37.52 | rob_w | eme where you able to see and understand the logs ;-) |
| 21:38.27 | eme | vy hard in this position ;) |
| 21:39.38 | eme | will try agn. I make on single enter after apm maybe this is the reason?? |
| 21:39.51 | eme | single enter on ssh terminal |
| 21:40.25 | rob_w | if that dummy cvs checkout ever finishes, i will test lkcl `s patches against this ! |
| 21:41.17 | rob_w | hmm the grey is strange , i didnt had this , yet it DOES point into the video things |
| 21:41.48 | eme | booting again |
| 21:42.45 | lkcl | eme: if you run "script" before ssh'ing in... |
| 21:43.01 | eme | what script? |
| 21:43.07 | lkcl | then after you finish ssh, type "exit" and what you see on-screen will be written into a file "typescript". |
| 21:43.19 | lkcl | type "man script" for an explanation. |
| 21:43.53 | eme | checking |
| 21:43.54 | lkcl | then, after the resume, run dmesg in the ssh console. |
| 21:44.01 | eme | ok |
| 21:44.08 | rob_w | lkcl, will that record more things as already seen on the terminal ? |
| 21:44.13 | lkcl | once it's woken up of course. |
| 21:44.35 | lkcl | everything that you type and everything that appears on-screen will end up in a file called typescript |
| 21:44.39 | lkcl | it's very useful. |
| 21:44.59 | lkcl | esp. in situations like this, where it's a bit dodgy getting data across. |
| 21:45.27 | lkcl | cr2: whatwhatwhat? i have a different gpioc? burble? :) |
| 21:45.51 | lkcl | i have the wireless activated, and i don't have the GSM activated. |
| 21:47.10 | eme | if I get it right sript -f filename ?? and then ssh to ba |
| 21:47.54 | cr2 | lkcl: your boardid is 5. |
| 21:48.35 | cr2 | the change of .init didn't help. looking for others. |
| 21:49.35 | lkcl | eme: just script would do, and the default filename is going to be "typescript". |
| 21:49.43 | lkcl | eme: so - yep :) |
| 21:50.24 | lkcl | cr2: i bet you that the xilinx gets reset... |
| 21:51.04 | eme | if you are not logged in gpe jous waiting on login screen adn do on terminal amp -s after that wont come up to grey screen anymore |
| 21:51.07 | rob_w | lkcl, why is that changes in blueangel_lcd.c so heavy to board id 6 ? i ve 5 ;-( |
| 21:51.19 | cr2 | lkcl: how ? through pxa gpio ? |
| 21:52.05 | rob_w | lkcl, i think that patch will not work for me , or ? |
| 21:52.44 | cr2 | lkcl: i want a serial console. hmm. maybe to try bluetooth ? :) |
| 21:53.17 | lkcl | rob - i have no idea. |
| 21:53.32 | lkcl | you will need to do something similar, then, rob. |
| 21:54.40 | rob_w | hmm comparing your patch and the previous one against that file , it seems board id 5 is quite a looser , right ? |
| 21:54.53 | lkcl | don't know yet... |
| 21:55.07 | lkcl | it's worth taking a look at the assembler |
| 21:57.19 | lkcl | hmmm... my brain hurts, looking at this assembly code. |
| 21:59.33 | eme | joust the first time wakes up and show grey screen. now wont wake up at all |
| 21:59.52 | rob_w | eme, did you load blueangel_suspend ? |
| 21:59.59 | eme | damm |
| 22:00.24 | rob_w | got ya |
| 22:00.24 | eme | repeating exercise... |
| 22:00.45 | rob_w | put it in /etc/modules for now :-) |
| 22:01.16 | eme | i like cut/paste tutorial... ;) |
| 22:06.33 | rob_w | crap i didnt get the patched cvs sources ... tss that makes this evening a failure ... will go to bed |
| 22:11.32 | eme | again get after wakeup grey screen. du you ned log from script? |
| 22:25.48 | *** join/#htc-blueangel iclem (n=clement@iclem.net) |
| 22:27.59 | lkcl | okay folks am half-asleep. |
| 22:28.22 | eme | zombie lol |
| 22:58.31 | eme | good night all |
| 23:00.40 | lkcl | night eme |
| 23:01.10 | eme | like to help u guys |
| 23:01.22 | eme | despertly waiting for linux to come on ba |
| 23:01.41 | lkcl | it's there - just not ... _quite_ there :) |
| 23:01.59 | eme | close than ever |
| 23:02.14 | lkcl | try compiling your kernel with the w100fb as a module. |
| 23:02.19 | lkcl | and the asic3 as a module. |
| 23:02.33 | lkcl | and the blueangel_lcd as a module. |
| 23:02.44 | lkcl | and buy a serial cable :) |
| 23:03.09 | eme | already have on homemade ;) |
| 23:03.14 | lkcl | excellent |
| 23:03.38 | eme | why modul is beter than static compiled or joust mather of testing |
| 23:04.03 | eme | an i know how to compile kernel for debian and not sure how to do this for ba. |
| 23:06.19 | eme | im stuck in plc programing and need some 'push' to start in c ... |
| 23:07.43 | eme | like linux vy much and work with him since kernel 2.0.36 |
| 23:08.10 | eme | all pc-s at home are on linux exept wifes and ba :(( |
| 23:09.32 | lkcl | it's a bit complicated, but it boils down to this: |
| 23:09.51 | lkcl | follow the instructions for bitbake (on the openembedded site) |
| 23:10.13 | lkcl | download the patch at http://hands.com/~lkcl/blueangel |
| 23:10.23 | lkcl | copy the local.conf file there, too |
| 23:10.28 | lkcl | do bitbake gpe-image |
| 23:10.58 | lkcl | then, after 6 hours, you will have a full system - including an arm cross-compiler, and there will be a subdirectory linux-2.6-xda. |
| 23:11.13 | lkcl | you can use the compiler in the staging/ subdirectory. |
| 23:11.22 | eme | 6 hours?? like on old 486?? |
| 23:11.23 | lkcl | that's ... sort-of it :) |
| 23:11.34 | lkcl | oh if it's a 486, then make it about 5 days :) |
| 23:11.42 | eme | haha |
| 23:11.58 | lkcl | yes, it's a lot of software. 6 hours on an AMD64 3000+ with 1gb of RAM. |
| 23:12.09 | eme | thanks for instructions and ofcourse I will try it |
| 23:12.25 | lkcl | btw you need about 8gb of disk space. |
| 23:12.37 | eme | you can check my other hoby if u like on http://lea.hamradio.si/~s51kq |
| 23:12.43 | eme | I have it |
| 23:12.51 | lkcl | i've missed out a couple of steps but there's enough to keep you occupied for at least a day or two :) |
| 23:13.13 | lkcl | he he |
| 23:13.43 | lkcl | you're a ham radio person? cool. |
| 23:13.57 | eme | yp since 1984 |
| 23:14.25 | lkcl | do you know how to "look people up" if you know their callsign? |
| 23:14.52 | eme | do not understand this |
| 23:15.12 | lkcl | don't worry... |
| 23:15.27 | eme | my call sign is S52FT |
| 23:18.31 | eme | openembedet is on oe.handhelds.org right? |
| 23:19.37 | eme | must be this one ;) |
| 23:24.10 | iclem | ~botmail help |
| 23:30.12 | lkcl | lemme see..., eme... |
| 23:30.31 | lkcl | yep. |
| 23:34.33 | eme | what? |
| 23:36.47 | iclem | ~botmail for JuNuKN Try this kernel: http://iclem.net/himalaya/zImage-2.6.12-hh2 with any BA or himalaya initRD, or the last familiar/GPE snapshot via NFS. LCD, TS works, that's how I made screenshots on http://wiki.xda-developers.com/index.php?pagename=HimalayaDevelopment |
| 23:54.15 | lkcl | yes it's oe.handhelds.org |
| 23:54.44 | lkcl | oh - iclem: you got a himalaya kernel? excellent :) |
| 23:54.59 | iclem | i got many ;) |
| 23:55.04 | lkcl | can you let me know what config you're using? |
| 23:55.06 | lkcl | ah _ha_ :) |
| 23:55.22 | cr2 | guys, what about asic3_mmc for himalaya ? |
| 23:55.26 | lkcl | i'm trying to bitbake a himalaya config, and not having much luck |
| 23:55.36 | lkcl | burble? |
| 23:55.41 | lkcl | haven't got that far, cr2 |
| 23:55.50 | cr2 | why ? |
| 23:56.06 | iclem | cr2: i've not managed to make asic3_mmc work |
| 23:56.17 | iclem | module load but no block device appear |
| 23:56.24 | cr2 | do you know the asic3 irq ? |
| 23:56.30 | cr2 | sdio irq. |
| 23:56.55 | iclem | llange thinks it's 11, same as on BA |
| 23:57.09 | cr2 | iclem: of course it is not working. some things must be corrected. |
| 23:57.20 | iclem | lkcl: do you want my .config ? |
| 23:57.22 | lkcl | damnit am going to have to do a bitbake rebuild... |
| 23:57.27 | lkcl | yes please iclem. |
| 23:57.50 | cr2 | is there himalaya_asic3 struct ? |
| 23:58.02 | cr2 | i'll look myself. |
| 23:58.58 | lkcl | it uses the h3900_asic3 struct. |
| 23:59.14 | lkcl | okay - it's _named_ h3900_asic3 :) |
| 23:59.25 | iclem | lkcl: http://iclem.net/himalaya/config.txt |
| 23:59.44 | cr2 | static int __init mmc_init(void) |
| 23:59.49 | iclem | i'm not enough skilled to understand asic3_mmc stuff |
| 23:59.53 | cr2 | extern struct platform_device blueangel_asic3; |
| 23:59.57 | lkcl | ta. |