| 08:22.46 | *** join/#htc-blueangel awelux_ (n=awelux__@dslb-084-058-169-112.pools.arcor-ip.net) |
| 09:07.04 | *** join/#htc-blueangel JuNuKN (n=JuNuKN@Mc6d0.m.pppool.de) |
| 10:03.22 | JuNuKN | Awelux_ are you here ? |
| 10:14.33 | *** join/#htc-blueangel n00b (i=ant@ip70-174-180-62.dc.dc.cox.net) |
| 10:19.05 | n00b`d | is there a reason i can't login to the cvs anonymously? |
| 10:21.31 | *** join/#htc-blueangel n00bish (i=ant@ip70-174-180-62.dc.dc.cox.net) |
| 10:44.36 | JuNuKN | n00xyz: The server is often to bussy to give you a connect retry it later again ! |
| 10:46.10 | n00b`d | oh, okay, thanks |
| 10:46.21 | n00b`d | i don't use the cvs system much, so i didn't know it'd reject access due to it being busy |
| 10:46.54 | JuNuKN | You'r welcome |
| 11:11.52 | JuNuKN | awelux_ : online? |
| 12:24.10 | *** part/#htc-blueangel JuNuKN (n=JuNuKN@Mc6d0.m.pppool.de) |
| 12:44.51 | awelux_ | n00b`d, the sourceforce cvs is under heavy load. |
| 12:45.26 | awelux_ | I guess at any given time a few thousand users check out files there |
| 14:05.30 | *** join/#htc-blueangel JuNuKN (n=JuNuKN@Mc6d0.m.pppool.de) |
| 14:07.44 | JuNuKN | Hey awelx_ ... at the moment cp15 (which originally did the port for ba) is online on #navit , he is interrested in the current status, maybe he can give you some hints about your problem! just ask ...Bye |
| 17:35.34 | *** join/#htc-blueangel awelux (n=awelux__@dslb-084-058-111-183.pools.arcor-ip.net) |
| 18:00.01 | *** join/#htc-blueangel mndctrl (n=mind@81.167.1.2) |
| 18:28.19 | awelux | hi question: |
| 18:29.03 | awelux | on himalaya (or ba?) you run the kernel starting at 0xa0208000 right? |
| 18:29.17 | awelux | why 2MB and not 500k or 1MB? |
| 18:29.40 | awelux | Does someone know what is going on in these lower 2MB ? |
| 18:31.42 | awelux | now I'm running at 2MB |
| 18:31.49 | awelux | but it's still not working :/ |
| 18:35.50 | awelux | ooooooooohhh |
| 18:35.59 | awelux | maybe I should reconsider the watchdog thing |
| 20:10.09 | *** join/#htc-blueangel cr2 (n=konversa@crpl22.physik.uni-wuppertal.de) |
| 20:26.49 | awelux | hi cr2 |
| 20:27.20 | awelux | I'm now ignoring the first 2MB of RAM |
| 20:27.50 | cr2 | hi awelux |
| 20:27.54 | cr2 | so ? :) |
| 20:28.14 | awelux | And start at 0xa0208000. But I still have the problem that the execution aborts without apparent reason. |
| 20:31.58 | awelux | What I don't understand is this: |
| 20:32.33 | awelux | The execution always stops within a narrow band of code. |
| 20:32.49 | awelux | and the same band might work a few hundred times |
| 20:33.07 | awelux | and always fail on the 85 or 87th pass |
| 20:33.12 | awelux | something like that |
| 20:37.49 | *** join/#htc-blueangel JuNuKN (n=JuNuKN@Md3d0.m.pppool.de) |
| 20:38.29 | awelux | now when I change something minor and recompile it will fail a lot earlier or later. maybe around pass 24-25. |
| 20:38.36 | cr2 | hmm. did you change something in the kernel ? |
| 20:38.59 | awelux | I'm adding the debug pixels to the kernel. |
| 20:39.00 | cr2 | to tell it about the 0xa02 ? |
| 20:39.05 | awelux | no functional change |
| 20:39.31 | awelux | oh. yes. I changed haret, the bootloader and the kernel |
| 20:39.55 | cr2 | i don't know about 2.6, but in 2.4 there were some constants and ldscript. |
| 20:40.20 | awelux | I'm pretty certain that this is not the problem. |
| 20:40.42 | awelux | I can checked and it's compiled as it should |
| 20:41.00 | awelux | I visualized the memory area the kernel is copied to. |
| 20:41.14 | cr2 | kernel or zimage ? |
| 20:41.30 | awelux | both |
| 20:41.35 | awelux | both are fine |
| 20:41.54 | cr2 | zimage can be put anywhere in the memory, but the decompressor should use the right memory and the kernel should be put at a right place. |
| 20:42.59 | awelux | the physical kernel location is in htcuniversal |
| 20:43.50 | cr2 | hmm. decompressor ? |
| 20:43.57 | awelux | arch/arm/mach-pxa/Makefile.boot |
| 20:44.41 | cr2 | ok. |
| 20:45.33 | awelux | Now I wonder, could the watchdog timer be that precise to always interrupt in a certain area of code? |
| 20:45.54 | awelux | And why would that interrupt depend on the kernel image |
| 20:46.15 | awelux | And why is it independent of a time ? |
| 20:47.06 | awelux | If I use just the vmlinux kernel and not the zImage I get the same problem but the execution time is much lower |
| 20:47.22 | cr2 | btw, there is the rtc in pxa27x |
| 20:47.52 | awelux | yep. it has a watchdog function. |
| 20:48.05 | cr2 | you can reset it. |
| 20:49.00 | cr2 | do you zero the memory before copying/decompressing ? |
| 20:49.20 | awelux | no |
| 20:49.24 | awelux | should I try that? |
| 20:49.46 | awelux | JuNuKN, thought we could fill the memory with nop's |
| 20:49.54 | cr2 | you can try to do it. |
| 20:50.00 | cr2 | zero is nop ? |
| 20:50.03 | awelux | and check at the end if something jumped out of the code. |
| 20:50.25 | cr2 | yes, if you will reach the end. |
| 20:50.25 | awelux | I think something else is nop |
| 21:06.09 | awelux | there is no expicit nop |
| 21:06.17 | awelux | they suggest to use mov r0, r0 |
| 21:06.45 | *** join/#htc-blueangel JuNu_KN (n=JuNuKN@Md3d0.m.pppool.de) |
| 21:09.57 | JuNu_KN | Better take a one byte command instead of a three byte commad, otherwise, you don't know, whether it interprets the command correct! |
| 21:10.21 | awelux | ^^ |
| 21:10.29 | awelux | there is no one byte command |
| 21:10.43 | awelux | all instructions are one word in size |
| 21:11.09 | JuNu_KN | On other processors NOP is such ... No Operation ... uses only one processor step |
| 21:11.16 | awelux | and a word is 32 bit for arm |
| 21:11.48 | awelux | that's a risc processor |
| 21:11.58 | JuNu_KN | OK ... you see my last hardware developement and asm was a 8Bit Processor. |
| 21:12.11 | awelux | ^^. A while ago. |
| 21:12.20 | cr2 | but there is nop in gas. |
| 21:12.30 | awelux | ok |
| 21:13.00 | JuNu_KN | But it should be ok to use a word command, because then it will access only the "unpair" adresses as start for interpretation |
| 21:13.08 | cr2 | ok. |
| 21:13.10 | cr2 | e1a00000 nop (mov r0,r0) |
| 21:13.12 | awelux | the arm will throw an exception if an instruction doesn'T start at a word boundary. |
| 21:13.41 | cr2 | so it is not a zero. |
| 21:14.01 | awelux | so if there is a valid interrupt pointer somewhere is must be word aligned. |
| 21:14.31 | awelux | nSo I've to fill the memory after the kernel and the first 2mb with e1a00000 |
| 21:14.58 | JuNu_KN | I think that interrupt vector is unintialized ... therfore fill it with nop to fetch that escape and at the end detect, where it starts |
| 21:14.59 | cr2 | yes. |
| 21:15.52 | JuNu_KN | Only a Idee, what I would try in my good old 8Bit world ;o) |
| 21:16.02 | awelux | if it's uninitialized it will be hard to catch. I've only 64MB ram in 4GB address space |
| 21:16.37 | awelux | btw. I bought an irda dongle today |
| 21:16.43 | awelux | And a bluetooth adapter |
| 21:16.48 | awelux | both work |
| 21:16.50 | JuNu_KN | Oh correct, but it is a chance, and I expect the vector should be in the first page |
| 21:17.15 | awelux | So I could try and setup the IR port for debugging. |
| 21:18.00 | awelux | but that's only interesting when I overcome the current problem |
| 21:18.38 | cr2 | ok. |
| 21:19.02 | cr2 | how to find out the phone number ? i want to call my universal :) |
| 21:21.55 | JuNu_KN | Send a SMS to me, I will answer you with your nummer ;o) Or try all numbers, until one is always bussy.... thats yours ;o)) |
| 21:22.54 | cr2 | good idea. i will send an sms to my a780 :) |
| 21:23.13 | cr2 | while running haret, of course. |
| 21:24.15 | awelux | lol. the second suggestion is great junukn |
| 21:24.38 | cr2 | linux is so much better. it is in the toplevel phone menu: "Meine Nummer" :lol: |
| 21:25.09 | cr2 | he has chosen the second option. |
| 21:25.30 | JuNu_KN | Who? That was my doublette :o) |
| 21:26.54 | awelux | I just discovered that I can create a startup file, send it with haret and the kernel to the pda |
| 21:27.03 | awelux | use synce-prun to run haret |
| 21:27.37 | awelux | and it will run even though I still get the windows mobile startup screen |
| 21:27.55 | cr2 | yes. i have also tested linking the kernel to haret. |
| 21:27.55 | awelux | Always adjusting the touchscreen got old after a few days |
| 21:28.17 | awelux | I had to fix the startup script name though |
| 21:28.27 | awelux | haret was searching for startupN.txt |
| 21:28.58 | awelux | no wonder it didn'T work to unmask gpio's via a startup script |
| 21:29.22 | cr2 | aah. |
| 21:29.43 | cr2 | i was also wondering why it does not work. |
| 21:29.50 | awelux | lol |
| 21:30.20 | awelux | src/wince/haret.cpp |
| 21:30.32 | cr2 | the idea was to have a different file, since i use boot2 which is incompatible with original haret. |
| 21:30.47 | awelux | ic |
| 21:30.57 | awelux | these hacks are easy to forget |
| 21:31.06 | cr2 | yep. |
| 21:39.01 | awelux | hmm. It looks like there ae zero page exception vectors in the lowest 0x100 byte |
| 21:39.34 | awelux | And we place pur haret bootloader there |
| 21:39.38 | awelux | *our |
| 21:39.52 | cr2 | how ? |
| 21:43.27 | awelux | ah no. we don't I thought we copy it in the preloader. |
| 21:43.35 | awelux | ah no, we don't. I thought we copy it in the preloader. |
| 21:44.54 | awelux | but we just copy the taglist. |
| 22:02.16 | awelux | http://www.simtec.co.uk/products/SWLINUX/files/booting_article.html |
| 22:02.21 | awelux | nice article |
| 22:06.05 | awelux | JuNu_KN, that's also very interesting for you |
| 22:06.42 | JuNu_KN | I am just reading ... ;o) I follow everything you write down :o) |
| 22:06.43 | awelux | or wait |
| 22:07.09 | awelux | You didn't want to boot wince and linux on the same device right ? |
| 22:09.55 | JuNu_KN | ... as long, as linux does not work perfect, I would prefer to still have Ce arround and start linux from SD |
| 22:12.06 | awelux | with this taglist you should be able to pass the ram areas linux can use to the kernel |
| 22:20.36 | JuNu_KN | cr2. How about you? Does your phone work? Any progresses? |
| 22:21.38 | JuNu_KN | awelux I thinks you did only a few assembler commands and not the kernel at the moment? |
| 22:22.08 | awelux | ? |
| 22:22.33 | awelux | it's to early to use the taglist with the magician yet |
| 22:22.44 | awelux | but it could be usefull for other devices |
| 22:23.01 | awelux | I'm currently debugging the asm of the kernel. |
| 22:23.25 | JuNu_KN | Ahh ... I understand ... I understood, that you expected errors in your tag definition! |
| 22:23.32 | awelux | It seams the asm is ok but something triggeres an nexcfeption |
| 22:24.40 | awelux | no. But we could allocate noncontinous ram from wince and use the taglist to tell the kernel where it can find useabloe ram. |
| 22:24.57 | awelux | of course the kernel must still be in a continous area |
| 22:25.12 | awelux | but the rest of the ram can be fragmented |
| 22:25.50 | JuNu_KN | Yes, I understand so far ... pluss the space, direct behind for the uncompressed kernel :o) |
| 22:26.08 | JuNu_KN | You see I read |
| 22:26.31 | JuNu_KN | I am willing, but not able |
| 22:28.02 | awelux | np. take your time |
| 22:28.57 | JuNu_KN | Not possible beside fulltime job 45 Hour a week) and familie |
| 22:29.12 | awelux | familie is the difference for me ;) |
| 22:30.24 | JuNu_KN | And the biggest disadvantage ... till now all customers prefer windows ... and so most of the dayly surroundings are windows. |
| 22:30.55 | awelux | not for me. lol |
| 22:31.10 | cr2 | two cams and asic3 . scary. and egpio. |
| 22:31.29 | awelux | you are trying to get the cams working ? |
| 22:31.35 | JuNu_KN | Therefore I will carry my linux system arround (Smartphone) and collect experiences with it and it's possibilities |
| 22:31.49 | cr2 | no. but they flood me with gpios. |
| 22:32.22 | awelux | mask them in the startup.txt |
| 22:32.33 | JuNu_KN | startupN.txt :o) |
| 22:33.04 | awelux | in your case |
| 22:33.10 | awelux | I simply removed the N |
| 22:33.33 | awelux | Where should I zero the memory? |
| 22:33.49 | awelux | directly in the kernel code = |
| 22:33.50 | awelux | ? |
| 22:34.01 | awelux | that's the only part that doesn't get overwritten later |
| 22:34.36 | awelux | hmm. I think I will do that. I can start an uncompressed kernel and get toi this part immediately. |
| 22:37.03 | JuNu_KN | I have to leave now .... Good night and much success |
| 22:45.03 | cr2 | awelux, is it possible to run haret over bluetooth ? i can't run it over usb (wince2005) and want to test wifi. |
| 22:46.06 | awelux | you can't run it over usb? |
| 22:46.12 | awelux | that's strange |
| 22:46.16 | cr2 | is activesync over bt ip-based or serial ? |
| 22:46.27 | awelux | I've to check |
| 22:46.32 | cr2 | synce does not work in wince5. |
| 22:46.35 | awelux | Actually I've no idea where to start |
| 22:46.42 | awelux | the bluetooth donge is brand new |
| 22:46.53 | cr2 | usb bt ? |
| 22:47.11 | awelux | yes usb bt. I just confirmed that it can find the pocketpc |
| 22:47.32 | awelux | I can see my pc in the detected device list |
| 22:48.32 | cr2 | http://blog.ptch.de/index.php?/archives/60-Mit-dem-PDA-ins-Netz....html |
| 22:50.03 | cr2 | hehehe. |
| 22:50.52 | awelux | hmm? |
| 22:51.29 | awelux | oh. das ist ja alles auf deutsch dort. |
| 22:51.29 | cr2 | yes. it's ppp . should work. |
| 22:51.37 | cr2 | PAN ist von Hause aus mit WindowsMobile 2003 nicht möglich, weil der MS Bluetooth Stack dieses nicht unterstützt. |
| 22:52.31 | cr2 | we should think about those who read the logs :) |
| 22:53.28 | awelux | yep. I always do ;) |
| 22:53.40 | awelux | development is in english |
| 22:53.44 | awelux | Here and at work |
| 22:54.19 | awelux | that's very suse specific |
| 22:57.11 | awelux | But I'm happy with the usb connection. For bt I would have to do the ntouchscreen thingy again |
| 22:59.20 | cr2 | i need to sleep today ;) good night. |
| 22:59.58 | awelux | gn8 |
| 23:08.51 | *** join/#htc-blueangel awelux__ (n=awelux__@dslb-084-058-144-038.pools.arcor-ip.net) |
| 23:09.53 | awelux__ | don't try to strace bluepin |