00:00.07 | mrevil | but it's lost after the next hard-reset, isn't it? |
00:00.21 | mrevil | how can haret "autostart"? |
00:00.38 | awelux | yes. that's why I use a pc side script to copy all the files to the magician including haret. |
00:00.56 | awelux | Then I use the synce.prun to start haret remotely |
00:01.37 | mrevil | ok, i'm seeing light. |
00:02.22 | mrevil | sorry, concentration is leaving me, |
00:03.26 | awelux | never mind. It's a lot of new stuff. I'm on it since january. And others are working on this for years. |
00:05.33 | mrevil | it's really a lot of stuff. without your help it would take me a long long time just to start the system - even without writing a single line of code |
00:07.04 | mrevil | hmm, even with initrd: kernel-panic |
00:07.16 | mrevil | cannot mount rootfs |
00:07.21 | awelux | yes. I know what you mean. It took about two month untill the system first bootet. And after that I didn't change a single line of kernel code. Just found out which modules to load and to use Image instead of zImage |
00:08.41 | awelux | when the system bootet. Did you notice if the red bar was thicker |
00:08.45 | awelux | ? |
00:09.10 | mrevil | yes, it seemed to be thicker |
00:09.59 | awelux | ok.then the initrd was loaded into memory. |
00:10.13 | mrevil | yes, it was |
00:10.43 | mrevil | but init was not executed. gave just the usage-int |
00:10.47 | mrevil | ..hint |
00:11.46 | awelux | ok. |
00:11.57 | awelux | I will try to recreate that |
00:15.52 | awelux | can you mount the rd file under your normal linux ? |
00:15.57 | awelux | maybe it got corrupted ? |
00:16.19 | awelux | probably no |
00:16.21 | awelux | *not |
00:17.06 | mrevil | oh. might be i took the incomplete one.. |
00:17.14 | awelux | ah another idea |
00:17.30 | awelux | how big is the file ? |
00:17.53 | awelux | Because it should be an compressed ext2 image |
00:18.09 | awelux | so if you gunzip it you can mount it |
00:18.37 | awelux | and |
00:18.42 | awelux | no |
00:19.27 | mrevil | the size is 5503300 |
00:19.32 | mrevil | about 5M |
00:20.13 | mrevil | ah - that one works |
00:20.25 | mrevil | next try |
00:24.15 | awelux | btw. do you know where I can find a bigstorage rom for magician ? |
00:25.12 | mrevil | i am using one :) |
00:25.34 | mrevil | got it from xda-developers.com |
00:26.07 | mrevil | but give me a few seconds to remember the procedure |
00:27.26 | mrevil | there was a nice text on the side... hm, was it inside the forum or the wiki?... |
00:28.00 | awelux | http://forum.xda-developers.com/viewtopic.php?t=46628&highlight=bigstorage+magician |
00:28.05 | awelux | this maybe? |
00:28.53 | mrevil | ftp://ftp.xda-developers.com/Magician |
00:29.29 | mrevil | yes, that is the page |
00:32.10 | mrevil | i step-by-step used that page (linked from the one above) :http://wiki.xda-developers.com/index.php?pagename=MagicianRomUpgrade |
00:33.20 | mrevil | the second part is the interesting one |
00:34.37 | awelux | Did you also use the netherland rom or a german one? |
00:35.18 | mrevil | oh, sorry |
00:35.29 | mrevil | i oversaw that |
00:35.51 | mrevil | i'm O2 customer but took the Tmobile one |
00:36.42 | mrevil | tmobile germany |
00:38.11 | awelux | ok.thx. |
00:38.14 | mrevil | at first i had problems with the update-program |
00:38.34 | mrevil | it told me about "wrong provider" or something like that |
00:38.37 | awelux | yes. it seams to be fow windows only ;) |
00:39.17 | mrevil | at that time i was not able to connect it to linux |
00:39.35 | mrevil | but there is a way with help of an sd-card.. |
00:40.25 | awelux | I'm about to activate a windows notebook. that should do the trick |
00:40.28 | mrevil | that should work without windows |
00:42.05 | mrevil | if you press reset while holding the camera and the "power"-switch |
00:42.14 | awelux | I've used that windows update procedure before for a normal rom upgrade |
00:42.57 | mrevil | ok, that shold be not really different from a normal update |
00:43.48 | mrevil | if you like your providers pre-installed tools you shold remember to backup the 'hidden' partition |
00:44.15 | mrevil | do you know how to access this partition? |
00:46.23 | awelux | yes. the trick with pressing reset in the first 3 seconds |
00:46.28 | mrevil | yes |
00:49.12 | mrevil | if you don't want to use windows, the sc-card-procedere should be your thing |
00:49.34 | i386 | awelux: :) |
00:50.00 | i386 | awelux: how goes the magician port? |
00:50.11 | mrevil | awelux: if you press the buttons, mentioned above you get into the "bootloader" |
00:50.51 | awelux | i386, very well. We are just installing the current state on mrevil's magician |
00:50.54 | mrevil | awelux: backlight is of but if you connect to usb and use a lamp you see "usb" on the screen |
00:51.39 | i386 | awww |
00:51.47 | awelux | I've seen that. But the notebook is already booted up |
00:51.55 | i386 | do you have something I can run from my flash card? |
00:52.17 | mrevil | hehe, ok |
00:52.18 | i386 | or demo? |
00:52.19 | i386 | :P |
00:52.23 | awelux | yes. |
00:52.37 | awelux | Do you have haret i386 ? |
00:53.40 | awelux | We are not yet in easy to use demo mode yet. |
00:54.29 | awelux | But with haret it should be fairly easy to boot a kernel. |
00:54.43 | awelux | The next step would be to mount some root partition. |
00:55.05 | awelux | mrevil, how does the other (complete) initrd fare? Better? |
00:55.05 | i386 | I have haret |
00:55.44 | awelux | do you have a sourceforge account ? |
00:56.01 | i386 | me? |
00:56.10 | i386 | or mrevil? |
00:56.19 | mrevil | just at that moment it failed again :( |
00:56.25 | awelux | i386, you. |
00:56.30 | i386 | yes I do |
00:56.34 | i386 | "jameswdumay" |
00:56.55 | i386 | btw |
00:57.06 | i386 | say if I helped out with testing/debugging |
00:57.17 | awelux | Because I would like to see if you can check out the right cvs tree if you use a sf account instead of anonymous cvs |
00:57.46 | i386 | is there a real chance of bricking my htc phone? |
00:57.51 | i386 | awelux: sure |
00:58.06 | i386 | just give me the CVSROOT string |
00:58.17 | awelux | cvs -z3 -d:ext:jameswduma@xanadux.cvs.sourceforge.net:/cvsroot/xanadux co -P -r K2-6-12-hh2-xda0-un0 linux-2.6-xd |
00:58.25 | awelux | that should do the trick. I hope |
00:58.40 | awelux | i386, there is a slight chance yes. |
00:59.04 | awelux | But right now things are fairly harmless |
01:00.16 | awelux | We don't use much functionality of the magician and I don't see where it could break the device. |
01:00.44 | i386 | hold on |
01:00.49 | mrevil | awelux: how far is your bigrom-expedition? |
01:00.49 | awelux | Right now I'm just carefull with the USB host options in the kernel. I'm having them disabled for now. |
01:00.55 | i386 | I forgot to install cvs on this new box |
01:01.07 | awelux | mrevil, not far. Eating my soup now |
01:01.22 | mrevil | ah, ok, bon apetit! |
01:01.47 | awelux | but it's strange that it stops at the same place. |
01:02.27 | i386 | checkling out |
01:05.47 | i386 | awelux: ok |
01:05.54 | i386 | what do you want me to look at? |
01:05.54 | awelux | i386, can you see something in the folder arch/arm/mach-pxa/magician/ ? |
01:06.05 | awelux | are there some c files in there ? |
01:06.44 | i386 | james@cabbage:~/hacking/linux-2.6-xda/arch/arm/mach-pxa/magician$ ls |
01:06.45 | i386 | CVS htcmagician.c htcmagician_core.h htcmagician_lcd.c htcmagician_lcd.h htcmagician_ts.c Kconfig Makefile |
01:06.52 | awelux | good |
01:07.26 | awelux | Thanks. It seams it's not possible to get these files with anonymous access. |
01:07.47 | awelux | And I was really puzzled for a while because I was sure I checked them in. |
01:08.15 | awelux | Then I feared that only I could see these files with my account. |
01:08.29 | awelux | But obviously any sf account can see them. |
01:08.46 | i386 | hmm |
01:08.51 | i386 | sounds odd |
01:09.02 | awelux | yes. But no action needed. |
01:09.11 | awelux | Do you have an arm linux crosscompiler ? |
01:09.25 | awelux | arm-linux-gcc ? |
01:10.22 | i386 | No I dont |
01:10.52 | i386 | oh is a 256mb SD card enough? |
01:11.10 | awelux | yes. right now the distros are very small |
01:11.16 | awelux | 64mb would be sufficient |
01:12.00 | i386 | awelux: I dont have the time to build the code right now |
01:12.14 | i386 | someone is about to pick me up |
01:12.51 | mrevil | awelux: when i try to co anonymously it tells me "cannot find module `linux-2.6-xd' - ignored" |
01:12.56 | awelux | Ah. It will probably take an hour anyway to get the system booted |
01:13.23 | i386 | ahh ok |
01:13.27 | i386 | awelux: I have to go |
01:13.29 | awelux | mrevil, interesting. |
01:13.34 | awelux | i386, cu later |
01:13.51 | i386 | keep up the good work man |
01:13.53 | mrevil | cu, i386 |
01:13.54 | i386 | :) |
01:14.07 | awelux | mrevil, and now what do we do with your problem. |
01:14.07 | i386 | cya mrevil |
01:14.08 | i386 | goodluck :) |
01:14.15 | mrevil | thx :) |
01:14.27 | mrevil | hm, don't know. |
01:14.49 | mrevil | perhaps i should take a look into the initrd start-up-scripts |
01:14.50 | awelux | did you try mto unzip the rd file and moun it on your normal pc ? |
01:15.07 | awelux | there are no initrd startup scripts. |
01:15.14 | mrevil | but i will give up for tonight... |
01:15.42 | awelux | take a look at init/main.c of the kernel |
01:15.49 | awelux | Right at the end |
01:16.58 | awelux | that's where the kernel is loading the init process. A little before it mounted the root fs |
01:17.07 | mrevil | therer is alinuxrc inside the initrd, is it ignored? |
01:17.17 | awelux | should have mounted |
01:18.01 | awelux | no. that's why you pass the line init=/linuxrc to the kernel |
01:18.20 | awelux | instead of the default init it will start linuxrc |
01:18.32 | awelux | so linuxrc is the first thing that's executed. |
01:18.37 | mrevil | ok |
01:18.43 | awelux | But the initrd has to be mounted by the kernel before that |
01:19.46 | awelux | If you have 15 more minutes we could try to use my kernel and my haret version |
01:20.03 | awelux | or I will upload them and you can try it tomorrow |
01:21.49 | mrevil | my concentration is down, i should better put it away now.. |
01:22.04 | awelux | ok. it'S really late. |
01:22.11 | mrevil | ah - the haret-version |
01:22.50 | mrevil | wait, i made a screenshot, perhaps i was not precise enough.. |
01:23.53 | awelux | It's unlikely haret has anything to do with it when you reach the kernel. It'S also possible that I missed something in the magician_defconfig. |
01:25.06 | awelux | Anyway. good night mrevil |
01:25.21 | mrevil | goodnight awelux! |
01:28.07 | mrevil | good n8 |
05:11.10 | parmaster | see the movie "Slipstream" |
07:16.03 | *** join/#htc-linux awelux (n=awelux__@dslb-084-058-168-148.pools.arcor-ip.net) |
10:00.47 | *** join/#htc-linux hlbot (n=adm@iclem.net) |
10:52.52 | *** join/#htc-linux goxboxlive (n=goxboxli@ti500710a080-11983.bb.online.no) |
11:27.43 | m1nd | how goes the wifi hacking? |
11:39.46 | goxboxlive | awelux: I tried with your settings, and i got allmost the same result. But why is your *-ts.c so diffrents from ours? I tried to just replace Magician with Universal in the file, but it refused to compile. |
11:40.10 | awelux | hi. |
11:40.32 | goxboxlive | hi , so youre here allready :-) |
11:40.41 | awelux | I started with the hx4700_ts.c file instead of the universal one |
11:41.01 | goxboxlive | ok i c. |
11:41.42 | goxboxlive | Is there many files i have to change to try with your ts.c file? |
11:43.18 | awelux | No, but I don't think that will make any difference. |
11:43.33 | goxboxlive | ok, i c. |
11:43.41 | awelux | The TI2046 TS chip has very few settings |
11:44.02 | awelux | You send a 7 bit and you get a response |
11:44.27 | awelux | the upper 3 bit are 100% clear |
11:44.42 | awelux | the lower 4 bit depend on the TS hardware. |
11:45.32 | awelux | you could try and google for tsc2046.pdf . That's the reference of the chip |
11:46.18 | goxboxlive | ok, i am not so in to this, this is way over my knowlegde so i guess i'll only make it worse :-) |
11:46.23 | awelux | bit 3 is 8bit/12bit mode so it's almost 100% certain that we will use 12 bit mode there. |
11:46.46 | awelux | 8bit mode is just faster with aquireing samples |
11:47.17 | awelux | So it's really just the bits 0-2 which can be changed to something else. |
11:49.00 | awelux | Bit2 is most likely 0. That's recommended for good precission and it's the case anywhere else. |
11:49.28 | awelux | So for starters I would play around with the last 2 bits. |
11:49.32 | goxboxlive | But why does the magician use SSDR_P2 and while the universal use SSDR_P1? |
11:50.31 | awelux | ah. Yes I remember now. It didn't work for me until cr2 reminded me that the tsc2046 is connected to different gpios on the magician |
11:50.46 | awelux | There are three ssp interfaces on the pxa272 |
11:51.16 | goxboxlive | ok, but with the last two bits you mean the one after "d" and the one after "9" right? |
11:51.27 | awelux | It's certain that SSDR_P1 is the right one for you since you got some data already |
11:51.42 | awelux | yes. exactly. the one after the d or 9. |
11:51.59 | awelux | after that we just send two filler bytes. |
11:52.11 | awelux | and in this time we expect an answer |
11:52.22 | awelux | which will be written into an internal fifo and read out later |
11:53.19 | goxboxlive | ok, i'll try some out :-) |
11:54.40 | awelux | When you try them out I recommend to start ts_test, draw a few horizontal and vertical lines, start ts_calibrate, then ts_test again. |
11:55.23 | awelux | That's probably the safest way to recognice changes in the TS behavior. |
11:55.33 | goxboxlive | Ok, i have tried ts_calibrate first, and then ts_test. I have problems with closing the ts_test |
11:55.54 | awelux | I closed it with a second ssh login and kill |
11:56.15 | goxboxlive | ok. |
11:57.01 | awelux | From the way things are wroking x and y calibration should be independent. So you will probably get straight lines in one direction first. |
11:57.02 | awelux | And |
11:57.35 | awelux | when you are in ts_test, can you check if the measured TS position is sensitive to preassure? |
11:58.21 | awelux | For me the position changed slightly when I applied more preassure until I had the right settings |
12:01.39 | awelux | I would like to know if the universal TS shows the same behavior sometimes. |
12:03.18 | goxboxlive | hmm i tried now with 0xd9* and 0x98* but i got the same result : http://forum.xda-developers.com/files/jasjar_247.jpg |
12:04.08 | awelux | ok. why 0xd90000 ? |
12:05.42 | awelux | I thought you would try out 0xd10000, 0xd20000, 0xd30000, 0x910000, 0x920000, 0x930000 ? |
12:05.47 | goxboxlive | beacause regular is 0xd0 and i thougt that if i changed it as much as i can, the imybe would se an diffrents with ts_test |
12:05.57 | awelux | :) |
12:05.59 | awelux | no |
12:06.20 | awelux | it's not a nob. it's more like 7 buttons |
12:07.07 | awelux | the last two button/bits are the interesting ones |
12:08.04 | goxboxlive | ahh, i c. i'll try out :-) |
12:09.47 | awelux | actually, |
12:10.37 | awelux | looking at this picture I would bet on 0xd20000 adn 0x930000 to work |
12:10.57 | awelux | because I had a similar picture once but turned 90 degree |
12:12.30 | goxboxlive | ok |
12:12.37 | awelux | if you have one of the values right one axis should be linear already |
12:12.51 | awelux | <- crossing my fingers |
12:17.39 | goxboxlive | hmm i tried out 0xd1 and 0x91 now, the ts_test cganged a bit. |
12:18.05 | awelux | good |
12:18.29 | awelux | I consider this progress :) |
12:19.07 | goxboxlive | i'll try out 0xd20000 adn 0x93000 now |
12:19.18 | awelux | <- crossing my fingers again |
12:19.26 | goxboxlive | :-) |
12:22.14 | goxboxlive | hmm with SSDR_P1 = 0xd20000 and SSDR_P1 = 0x930000 the touchscreen is dead |
12:24.02 | goxboxlive | hmm this is not good. It's dead even when WM5 start's. I have to take the battery of and reset it again |
12:24.06 | awelux | ok |
12:24.44 | awelux | I guess it's because the ts IRQ was powered down |
12:25.15 | goxboxlive | ok, well let's hope that :-) |
12:25.31 | awelux | We might have to swap the sampling of the X and Y coordinates |
12:26.18 | goxboxlive | It helped to tka eout the battery. Maybe i should try d1 and 93 |
12:26.49 | awelux | so when we get a penirq we can switch off the penirq to measure the y part and then switch it on while meassuring the x part. |
12:26.53 | awelux | no |
12:26.58 | awelux | or wait |
12:27.18 | awelux | cay you try 0xd20000 and 0x920000 ? |
12:27.25 | goxboxlive | yes |
12:27.28 | goxboxlive | hold on |
12:27.44 | awelux | this will leave the irq enabled and we should get linear meassurements on one axis |
12:28.15 | awelux | if it works we can swap the x and y sampling in the code and retry 0xd2 and 0x93 |
12:29.48 | awelux | that's probably why noone succeded before |
12:36.39 | goxboxlive | d2 and 91 the same as d1 and 91 |
12:36.52 | goxboxlive | i mean d2 and 92 |
12:38.13 | awelux | draw a cross over the middle of the screen |
12:39.38 | goxboxlive | the cross is perfect, except it dosent start in the corner. It's allmost same as http://forum.xda-developers.com/files/jasjar_247.jpg in the corners |
12:39.41 | awelux | I wonder if there is a way that the nonlinear calibration in one direction will cause the other to appear nonlinear too |
12:40.35 | awelux | I meant a cross with horizontal lines going through the middle |
12:41.17 | awelux | I drew a grid on the screen. |
12:41.53 | awelux | But it's probably a good bet to do the swap of the x and y sampling right away |
12:41.57 | goxboxlive | its an x² function |
12:42.37 | awelux | ok I will think about this for a while |
12:42.48 | goxboxlive | how do i swap the x and y sampling |
12:43.09 | awelux | a ) swap the for loops with the 0xd20000 and 0x93000 . |
12:43.58 | goxboxlive | but the but the derivated of x² ( dont know the exact name of the function) is very low |
12:44.33 | awelux | b) swap the sections in ts_time_call back which read out the x and y samples |
12:44.42 | awelux | they are similar in structure: |
12:45.01 | awelux | <PROTECTED> |
12:45.01 | awelux | <PROTECTED> |
12:45.24 | awelux | another for loop |
12:45.32 | awelux | and then either x= or y= |
12:45.42 | awelux | actually swapping the x and y there is enough |
12:45.49 | goxboxlive | ok |
12:47.47 | awelux | hmm. the raw values of a good ts setting will look like a parallelogram |
12:48.20 | awelux | the calibration will map this to a box |
12:48.57 | awelux | using the Y (in your case probably X) value to calculate the X (Y for you). |
12:49.03 | goxboxlive | can you swap the x and y for me awelux: Here is the ts.c at pastebin. Jsut modify it and oastebin it again,: http://pastebin.com/715166 |
12:49.51 | awelux | someone changed that driver :) |
12:51.17 | awelux | swap the while sections at lines 200-209 |
12:52.50 | awelux | and swap lines 298 and 299 |
12:53.05 | awelux | and use 0x930000 |
12:53.09 | awelux | that should be all |
12:54.12 | goxboxlive | ywiyh this you mean: move the while * to line 209? |
12:56.20 | goxboxlive | ywiyh = with |
12:57.05 | awelux | line 204 and it's comment in 202 should be swapped with 209 and the comment 207 |
12:57.35 | awelux | while(!(SSSR_P1 & (1 << 2))) |
12:57.35 | awelux | ; |
12:57.47 | awelux | <PROTECTED> |
12:58.05 | awelux | SSDR_P1 = 0xd20000 /* | (lower_bits << 0) */; /* May want bits 0 and 1 set for keeping power up */ |
12:59.01 | awelux | this structure is twice at the lines 200 to 209. The first tells the tsc2046 to get an x sample the second to get an Y sample |
12:59.21 | awelux | We want to swap the to demand the Y sample first. |
13:02.56 | goxboxlive | so it look like this? http://pastebin.com/715184 |
13:03.45 | awelux | exactly change the 0x92 to 0x93 and swap lines 298/299 |
13:03.53 | awelux | http://pastebin.com/715185 |
13:04.31 | awelux | uh. I forgot to change the x920000 to 0x930000 too |
13:07.21 | goxboxlive | like this? http://pastebin.com/715192 |
13:08.17 | goxboxlive | and of course line 298/299 like this: |
13:08.20 | goxboxlive | <PROTECTED> |
13:08.20 | goxboxlive | <PROTECTED> |
13:08.30 | awelux | now we swapped lines 200- 209 twice :) |
13:08.55 | goxboxlive | hmmm i not so in to this u know. But am i ready to compile now? |
13:09.32 | awelux | no wait |
13:09.37 | goxboxlive | :-) |
13:09.57 | awelux | http://pastebin.com/715195 |
13:10.22 | awelux | compare the lines 200 -209 again. |
13:10.42 | awelux | and the X and Y are hopefully uppercase in 298/299 right? |
13:12.02 | goxboxlive | no they where in lowercase before i edited it |
13:12.22 | goxboxlive | i just swapped the x to line 299 and y to line 298 |
13:13.56 | awelux | ok. strange. they look like uppercase in pastebin |
13:16.28 | goxboxlive | you right :-) |
13:17.08 | goxboxlive | i changed it and try to compile it |
13:17.15 | awelux | good. |
13:24.54 | *** join/#htc-linux jeanseb (n=jeanseb@gazypan.dyndns.org) |
13:25.20 | goxboxlive | exactely the same result |
13:27.18 | awelux | :( |
13:28.27 | awelux | but at least the ts is mworking with 0x930000 now? |
13:29.00 | awelux | because before this didn't work. |
13:29.38 | goxboxlive | yes it does |
13:30.06 | awelux | can you start ts_print |
13:30.31 | goxboxlive | hold on, i have to boot again. |
13:30.37 | awelux | oh and what about the preassure ? |
13:30.56 | awelux | does it make a difference how much pressure you apply to the ts? |
13:33.45 | goxboxlive | the pressure seems more stable now. Doesnt have to press har. I'll try ts_print |
13:34.55 | awelux | do you get a movement in the X or Y direction when you apply more preassure ? |
13:35.43 | awelux | try that somewhere in the middle where you almost have horizontal and vertical lines |
13:36.30 | goxboxlive | ts_print doesnt do a thing. Just hangs |
13:36.41 | awelux | touch the ts ? |
13:36.57 | goxboxlive | yes a lot data appera |
13:37.17 | awelux | is there a ts_printraw or something like that? |
13:37.55 | goxboxlive | yes |
13:38.13 | awelux | otherwise you can also do od -x -v /dev/input/tsraw0 |
13:38.42 | goxboxlive | a list of cordinates? appear when i touch the screen. |
13:38.47 | awelux | hmm. It's tricky to interpret this data. |
13:39.11 | awelux | probably mthe most interesting part is if the raw data changes only mon one coordinate if you apply more preassure |
13:40.16 | awelux | and note down which coordinate changes when you use different vaules of 0x[d9][0123]0000 |
13:41.27 | awelux | hopefully you will see that different coordinates change for different values and sometimes an axis remains stable regardless of preassure. |
13:41.30 | goxboxlive | ok, i'll have to try later. Have to take a trip out now. Thanks alot for your help |
13:41.39 | goxboxlive | ok i c |
13:41.55 | awelux | yes. cu around. |
13:43.34 | awelux | there is a slight probability that we have to use both 0x93 and 0xd3. the preassure tests could indicate that. |
13:44.06 | awelux | if that#s the case we will have to reactivate the penirq after we aquire the samples. |
15:35.38 | *** join/#htc-linux BabelOued (n=BabelOue@lun34-2-82-238-28-28.fbx.proxad.net) |
15:58.16 | awelux | cr2, There are huge changes in the universal TS driver. Why? It looks like there is code to do nonlinear calibration. |
15:59.59 | awelux | In case the TS is not the problem but the setup of the tsc2046 is, would this build in "calibration" transform a first order linear behavior into a higher order one that can't be corrected by ts_calibrrate? |
16:00.56 | *** join/#htc-linux rmoravcik (n=rmoravci@icm1-orange.orange.sk) |
18:30.51 | *** join/#htc-linux AlphaPi[jr] (i=aaa@host30-182.pool872.interbusiness.it) |
18:39.01 | *** join/#htc-linux amgad_ (n=amgad@196.205.176.235) |
19:50.36 | goxboxlive | cr2: Are you there? |
19:59.50 | awelux | hi goxboxlive |
20:00.13 | awelux | Did you use the time to buy something for mothers day? |
20:02.44 | goxboxlive | Mothers day? Wasen't that in february? |
20:05.02 | awelux | hmm. |
20:05.22 | goxboxlive | Actually, my son has birthday today ( he was 2 years :-) ) |
20:06.11 | awelux | ic. It seams mothers day is celebrated on different days in different countries. |
20:15.15 | goxboxlive | probably |
20:25.12 | goxboxlive | hmm kind of strange. I rolled back a few versions of the ts driver (the latest one without all equations) nad edited the 0xd and 0x9 . Now the edges is way bether, but in the midles it draws about 1cm over my point ( using ts_test) |
20:26.54 | goxboxlive | acyually the upper 1/4 is quite good, but the longer down i come the worse it gets.. |
20:35.09 | goxboxlive | awelux: Do you want here something? |
20:35.31 | goxboxlive | I have a perfect screen calibration on my Universal |
20:35.51 | awelux | :) |
20:35.59 | awelux | yea!! |
20:36.02 | goxboxlive | I used d1 and 91 and now the ts_test is perfect :-) |
20:36.03 | awelux | what did you use? |
20:36.09 | goxboxlive | Thanks a lot for your help. |
20:36.35 | goxboxlive | And i rolled back some versions of the ts.c driver. |
20:37.15 | awelux | yes. I wanted to ask cr2 about the recent changes of the driver. |
20:37.40 | goxboxlive | man, this is great. Have to try Opie. :-) |
20:38.07 | awelux | You mentioned right at the beginning that the universal driver is very different from the magician one |
20:38.38 | goxboxlive | yes, i saw it where a lpt of mathematics eqution in it, and your didnt have that. |
20:39.18 | awelux | And I thought it was just that I restarted based on the hx4700 driver. |
20:39.40 | goxboxlive | :) |
20:40.01 | awelux | The old universal driver is somewhat similar with a few modifications. The new one needs work or a revoke to the earlier version. |
20:40.14 | awelux | How is opie when you can hit the buttons ? ;) |
20:41.01 | goxboxlive | hold on, it's booting, had to reset it because of the ts_test. |
20:41.41 | *** part/#htc-linux rmoravcik (n=rmoravci@icm1-orange.orange.sk) |
20:42.31 | goxboxlive | it works great. cool, have to post on xda forum so people get intrest again. |
20:42.59 | goxboxlive | even the exit cross in the corner is a peace of cake to handle now :-) |
20:49.30 | awelux | The really nice thing is that it's now possible to work with applications. |
20:49.55 | awelux | So the people who wanted the universal for applications can start using it. |
20:50.03 | awelux | well program for it. |
20:50.37 | awelux | Now we need power management. To be able to reload the batteries |
20:51.15 | awelux | Then people can start running linux 24/7 :) |
20:59.00 | awelux | gee. I just bought an application for the ppc but this crazy double activation register process is slower and more work than searching for a crack :/ |
21:04.58 | goxboxlive | look at the screenshot :-) http://forum.xda-developers.com/files/ts_test_uni.jpg |
21:06.33 | *** join/#htc-linux goxboxlive (n=goxboxli@ti500710a080-11983.bb.online.no) |
21:12.10 | awelux | wow beautifull :) |
21:17.33 | goxboxlive | yes cool this is cool. |
21:17.51 | awelux | I've finally a bigstore rom installed, everything is setup again, some applications moved to the new bigstorage and I've a complete backup. |
21:19.16 | awelux | ah. I should again try to checking the htcmagician_ts.c |
21:21.59 | goxboxlive | yes i should that to, bit i'll wait to i have talked to cr2. |
21:23.33 | awelux | question: cvs commit -t -m "linear ts" htcmagician_ts.c |
21:23.40 | awelux | shouldn't that work ? |
21:24.04 | awelux | But I get: |
21:24.04 | awelux | <PROTECTED> |
21:24.04 | awelux | ssh: connect to host cvs.sourceforge.net port 22: Connection refused |
21:24.46 | goxboxlive | have no idea. |
21:27.22 | awelux | Strange. The total file size of the files in the /windows/ foilder without it's subfolders is already 71.5MB |
21:37.30 | goxboxlive | hmm yes it's big |
21:40.34 | cr2 | hi guys ! i'm excited by your success ! |
21:48.43 | goxboxlive | cr2: Yes this was cool. Have to go to bed now, have oversleep visitors, and they sleep where my computer is. night :-) |
22:03.24 | cr2 | awelux, it does not work with 2.6.16 driver. not good ;) |
22:03.39 | awelux | ah hi cr2 |
22:03.57 | awelux | I guess you are refering to the ts? |
22:04.08 | cr2 | yes. |
22:04.43 | awelux | I'm having trouble with the cvs. Do you havbe any idea why I can't check in the magician_ts ? |
22:05.11 | awelux | and, shouldn't I be able to connect to sf server via ssh with my account ? |
22:05.28 | cr2 | i can't too. |
22:05.46 | awelux | ah. ok. So it's a sf problem and not my problem? |
22:05.49 | cr2 | ssh: connect to host cvs.sourceforge.net port 22: Connection refused |
22:05.52 | cr2 | i think so. |
22:06.49 | awelux | ok. being the cvs noob that I am I thought I forgot something. |
22:07.34 | awelux | Anyway. what to do now? Does the 2.6.16 kernel boot for you? What's the tag you use to check it out? |
22:08.45 | awelux | What about power management? Any idea what's needed to get the battery charging working? |
22:09.56 | awelux | And, we should think of the propper was to share the TS driver between the magician and universal. |
22:11.14 | awelux | Beside the port SSP1/2 and the tsc sample request byte the systems are similar. |
22:13.11 | cr2 | irq |
22:13.24 | cr2 | 2.6.16 does not work for me. |
22:14.05 | cr2 | power management is big: power monitor, charging, and suspend/resume. 3 different problems. |
22:14.28 | awelux | charging would be most usefull I think. |
22:15.43 | cr2 | you should not use it without the monitor to avoid overcharging. |
22:16.29 | cr2 | i have found a charge enable gpio and the monitor code partly works, but there are some things to resolve. |
22:26.33 | awelux | how does the monitor work in principle? |
22:26.42 | awelux | does it monitor the voltage somehow? |
22:26.56 | cr2 | yes. |
22:27.01 | awelux | does it try to keepm track of the amount of elictricity used? |
22:27.17 | cr2 | ba and uni have a custom chip. |
22:27.42 | cr2 | through measuring current (voltage on a known resistor) |
22:28.18 | awelux | ok. So the chip keeps track of the battery status. |
22:28.35 | cr2 | yes. but you don't have such chip in magician. |
22:28.41 | awelux | Do you know if we have to give the chip feedback or calibration points? |
22:28.41 | cr2 | maybe tsc2046 ? |
22:28.57 | cr2 | ba uses the ts chip, btw. |
22:29.01 | awelux | the tsc2046 should be able to meassure the voltage |
22:29.38 | cr2 | no, but you need to know how to decode its output. |
22:30.21 | cr2 | universal has a chip on the w1 bus of asic3. |
22:33.56 | awelux | does the no mean that the tsc doesn't meassure the voltage ? |
22:36.07 | awelux | It does look like it has a function to meassure the voltage. I'm not sure if it's used for the power management so. |
22:36.23 | awelux | It's not trivial to map voltage to charge status. |
22:39.33 | cr2 | the uni is a expensive device and it has a special chip that measures a lot of things, has eeprom, etc. it is like on older ipaqs. DS2762. |
22:40.23 | cr2 | i don't see it on magician, and you need something that understands w1. on uni it is asic3, which you also don't have. |
22:41.46 | awelux | still, somehow wince manages to track the charge status. |
22:42.31 | cr2 | look how ba does it. |
22:42.59 | cr2 | you can measure the charge current and integrate it over time. |
22:43.36 | awelux | for charging this might work. But how about power usage? |
22:43.57 | awelux | It seams that the power usage is highly dynamically. |
22:44.11 | awelux | Just measuring from time to time might not be enough |
22:59.30 | cr2 | i don't know how wince calculates the % from voltage and current. but obviously both are used. |