| 08:24.16 | *** join/#htc-blueangel awelux_ (n=awelux__@dslb-084-058-173-255.pools.arcor-ip.net) |
| 08:34.54 | *** join/#htc-blueangel geers_ (n=DevzMan@c51008F31.inet.catch.no) |
| 08:38.35 | *** join/#htc-blueangel rob_w|mis (n=rob_w@p549BC6CA.dip0.t-ipconnect.de) |
| 17:28.59 | *** join/#htc-blueangel suma (i=thambi@203.212.228.23) |
| 17:29.06 | suma | hi |
| 17:29.18 | suma | anybody here ? |
| 17:31.51 | dullard | noone's been here for days |
| 17:32.01 | suma | i c |
| 17:32.10 | dullard | :-( |
| 17:32.12 | suma | i have blue angel handheld |
| 17:32.27 | suma | i want to load with linux kernel |
| 17:32.51 | dullard | I have a universal... I've yet to get any linux on it |
| 17:33.06 | suma | i c |
| 17:33.55 | dullard | I think there are some instructions on the xda-developers wiki for loading linux on to the blue angel |
| 17:34.56 | suma | checking it |
| 17:50.15 | suma | is it not possible to load linux without mmc card ?? |
| 17:51.49 | dullard | I think so... only off mem card. That gives you the chance to dual boot anyway |
| 17:52.25 | suma | http://handhelds.org/moin/moin.cgi/BlueAngel |
| 17:52.30 | suma | i'm checking this site |
| 17:52.42 | suma | this says i need to format the sd card |
| 17:52.47 | suma | and use fdisk ... |
| 17:53.04 | suma | I use windows xp, shall i do from that ? |
| 17:55.04 | dullard | I think one of the partitions needs to be a linux partition... I may be wrong though |
| 17:55.31 | suma | you mean on the sd card ? |
| 17:56.14 | dullard | yeah |
| 19:45.00 | *** join/#htc-blueangel JuNuKN (n=JuNuKN@Med35.m.pppool.de) |
| 19:49.37 | JuNuKN | Hello everybody. |
| 19:49.48 | JuNuKN | hello cr2 and awelux_ |
| 19:53.47 | cr2 | hi JuNuKN. my universal got a replacement motherboard and is back now! |
| 19:54.23 | JuNuKN | Wow ... great! Was ist without costs now ? |
| 19:54.33 | cr2 | yes. |
| 19:54.51 | cr2 | soon i will get a ba LCD. |
| 19:56.00 | JuNuKN | ... also great. Now you can start again to do your researches and play with it ;o) |
| 19:56.10 | JuNuKN | Lucky boy ;o) |
| 19:56.51 | suma | will the company won't doubt if the device often get replaced ? |
| 19:57.15 | cr2 | :) it is still working. |
| 19:59.14 | suma | can i partion with partition magic from windows to upload linux on my htc ? |
| 19:59.19 | JuNuKN | Did you risk again to do some linux tests with it )) |
| 19:59.30 | cr2 | can you download the latest gnuharet version and type "boot2" and tell me what you will see on the screen ? |
| 20:00.04 | JuNuKN | sumo: better take dammsmalllinux or knoppix and format it from there with a external sd-card reader |
| 20:00.22 | cr2 | suma: i think yes. |
| 20:00.38 | cr2 | at least you can try. |
| 20:00.46 | suma | thanks |
| 20:01.26 | JuNuKN | cr2 give me instructions, what it shall do and on which device |
| 20:02.14 | cr2 | partitioning is only a half of the task. you need to format it with ext2fs and untar the gpe archive. i don 't know how it can be done on windowz. |
| 20:02.50 | JuNuKN | ext3 ! |
| 20:03.33 | cr2 | JuNuKN: go to jornada820.sf.net/files/haret ,get the latest version and run "boot2" on the haret command line. |
| 20:03.43 | cr2 | first on magician. |
| 20:04.16 | JuNuKN | What will it do ? |
| 20:04.48 | cr2 | load the kernel an ramdisk into memory and plot where they are on the screen. |
| 20:05.03 | cr2 | you need to reset after this command. |
| 20:05.24 | cr2 | because the kernel and ramdisk will stay in the memory. |
| 20:05.34 | suma | is it not possible to |
| 20:05.48 | suma | http://uranus.it.swin.edu.au/~jn/linux/explore2fs.htm |
| 20:06.17 | suma | i think with this software i can untar to the mmc |
| 20:07.16 | *** join/#htc-blueangel coolhp (n=crap@mtl149-99-190-66.dedicated.sprintdsl.ca) |
| 20:07.25 | coolhp | Hi all ! |
| 20:07.43 | coolhp | I was wondering if I could bother you for some quick assistance. |
| 20:08.00 | JuNuKN | cr2 : I don't want to do it on Magician, because it may double again my Contacts afterwards when I sync with the second PC. |
| 20:08.05 | cr2 | suma: i have no experience of running windows :) |
| 20:08.17 | cr2 | coolhp: hi |
| 20:08.18 | JuNuKN | Is it from help to test with himalaya, which is currently not in use |
| 20:09.00 | cr2 | hmm. ok. try hima. |
| 20:09.05 | coolhp | I'm trying to boot linux on my blueangel... using Haret-0.3.6. I got to the tux+thermometer part. It seems to load ok but it then hangs there (when the thermometer reaches the top and tux has a wasted look) :-) |
| 20:09.09 | coolhp | Any ideas ? |
| 20:10.19 | cr2 | not really. the most interesting things happen just after that. for many people it just worked. debugging is not really easy. |
| 20:10.32 | coolhp | I guess so :-) |
| 20:10.48 | coolhp | I would love to have some more verbose information from haret.. but oh-well :-) |
| 20:11.10 | coolhp | Would anyone have a URL from which I could donwload the latest zImage and initrd ? |
| 20:11.24 | cr2 | we have a similar problem on universal and magician. |
| 20:11.31 | cr2 | handhelds.org wiki. |
| 20:11.40 | coolhp | Also, does the vfat partition size really matter ? (I had a 1G SD split into 800M vfat and 200M ext3) |
| 20:12.01 | cr2 | do you have a smaller card ? |
| 20:12.12 | coolhp | Unfortunetly not :-( |
| 20:12.49 | *** join/#htc-blueangel immolo (i=vlgjkhop@cpc1-cosh1-4-0-cust126.cos2.cable.ntl.com) |
| 20:13.25 | coolhp | Is there any way of getting haret to spit out a little more info ? |
| 20:13.31 | cr2 | actually, the vfat partition is not absolutely necessary. you can boot from wince directly. vfat is just a bit more convenient. |
| 20:13.54 | coolhp | Not that tux's picture isnt nice but some debug info is always better. |
| 20:14.38 | cr2 | no, because the interrupts are disabled. if you reach the kernel code, then you can use the serial console. |
| 20:15.07 | coolhp | I'll try not loading the initrd and see what happens :-) |
| 20:15.29 | cr2 | i 'm trying to debug the assembler code by direct writes into the framebuffer. |
| 20:15.46 | cr2 | but it is not completed now. |
| 20:16.05 | cr2 | yes. it should work. |
| 20:16.46 | coolhp | I think I may have bricked my blueangel |
| 20:16.54 | coolhp | Darn thing wont reboot. |
| 20:17.19 | cr2 | why ? you need to do a hard reset. |
| 20:20.15 | awelux_ | hi |
| 20:20.47 | JuNuKN | cr2 ... your man is there for trying to boot on magician ... |
| 20:20.48 | cr2 | hi awelux. i have a haret version that writes into the framebuffer. you can test it. |
| 20:20.56 | cr2 | :) |
| 20:21.00 | awelux_ | great. |
| 20:21.13 | awelux_ | I've a question about that mov pc,r3 |
| 20:21.33 | awelux_ | it moves the physical address of the preloader into the program counter. a jump |
| 20:21.41 | cr2 | but not yet in assembler. but it think we will do it soon. i have my universal back :) |
| 20:21.50 | cr2 | yes. |
| 20:22.12 | coolhp | How big should the initrd file be ? |
| 20:22.19 | awelux_ | why does it use the physical address there and not the virtual one? |
| 20:22.51 | cr2 | if you have 1:1 entry in the l1 table they are equal. |
| 20:23.01 | cr2 | in this 1 MB. |
| 20:23.23 | awelux_ | ok. I don't understand then how this can happen. |
| 20:23.27 | cr2 | coolhp: it is better when initrd <4MB. don 't ask me why. |
| 20:23.33 | awelux_ | brb. |
| 20:23.45 | awelux_ | have to take something out of the car first |
| 20:23.48 | coolhp | Hmmmm |
| 20:23.50 | coolhp | That might be my problem. |
| 20:24.01 | coolhp | All the initrd's I've found are around 14m |
| 20:24.20 | cr2 | compressed not uncompressed. |
| 20:24.26 | coolhp | .gz |
| 20:25.17 | cr2 | i think the last initrd on hh.org is about 5MB . i could not boot universal with it. the older one (3.4MB) worked. |
| 20:25.38 | cr2 | but it works for many people on blueangel. |
| 20:25.52 | cr2 | the initrd from hh.org |
| 20:26.50 | coolhp | LOL |
| 20:26.58 | coolhp | hh.org = Top-Quality Dating Services At Lower Prices!. |
| 20:27.28 | cr2 | :))) i meant handhelds.org of course. |
| 20:27.38 | coolhp | Oh god... LOL hh2.org is even worse.... ROFLMAO |
| 20:27.57 | coolhp | Would you happen to have a URL for a copy of the old initrd ? (the <3M one) ? |
| 20:28.24 | cr2 | i will look for it. |
| 20:28.37 | awelux_ | is the new gnuharet with the visual page moves on your server cr2 ? |
| 20:28.54 | cr2 | yes. |
| 20:29.01 | cr2 | the latest version. |
| 20:29.11 | awelux_ | k |
| 20:29.54 | cr2 | not the source, but i can upload it too. it is a bit tuned for universal 480*640, but not really. |
| 20:30.42 | awelux_ | I'm about to dig into the bootloader code now. So I might as well start from your version. Would be easiest for question. |
| 20:31.07 | cr2 | ok. wait a second. |
| 20:32.42 | JuNuKN | cr2 .... I am sorry, now wlan driver and everything is prepared on Himlaya, but now the battery is empty. I will charge it tonight and try (perhaps) tomorrow. |
| 20:34.31 | cr2 | JuNuKN: i think we will move forward the debugging code now, so it does not matter that much. |
| 20:35.23 | JuNuKN | Thanks .... I will go to bed soon and follow your progresses for a few minutes. |
| 20:36.04 | cr2 | coolhp: i have uploaded the small initrd to http://jornada820.sf.net/files/haret |
| 20:36.20 | coolhp | Thx cr2 ! |
| 20:36.23 | cr2 | awelux: the source is uploaded. |
| 20:36.30 | coolhp | In the mean time, I acually got to boot the 5Meg initrd. |
| 20:36.39 | coolhp | And the kernel did boot. |
| 20:36.49 | cr2 | ok. so what was the problem ? |
| 20:37.09 | coolhp | Looks like the initrd was the issue. |
| 20:37.18 | coolhp | I dont know why but my file was 14M |
| 20:37.25 | coolhp | Now I have another issue. |
| 20:37.44 | coolhp | It seems the modules folder on my rootfs do not match my kernel version. |
| 20:38.01 | coolhp | It seems to be looking for /lib/modules/2.6.12-hh2 |
| 20:38.19 | coolhp | but I have /lib/modules/2.6.12-hh2-xda0-ba0 |
| 20:38.47 | cr2 | then uncompress,mount the initrd and update the modules. |
| 20:39.23 | cr2 | hmm. i hate this mess... |
| 20:39.44 | coolhp | I am not too sure of what you mean by uncompress,mount the initrd and update the modules. |
| 20:39.45 | cr2 | have you looked at hands.org/~lkcl ? |
| 20:39.50 | *** join/#htc-blueangel iclem (n=clement@iclem.net) |
| 20:40.53 | iclem | hi all |
| 20:40.59 | cr2 | hi iclem. |
| 20:41.36 | cr2 | can you compile a module-less kernel for blueangel ? |
| 20:41.41 | JuNuKN | Hi iclem ..... ... I will stay for a few more minutes ;o) |
| 20:42.08 | iclem | cr2: yes |
| 20:42.21 | iclem | hu JuNuKN |
| 20:42.40 | iclem | sorry, no news for himalaya ;) |
| 20:43.32 | cr2 | iclem. i have looked on the motherboard pictures and came to the conclusion that hima,ba and magician have exactly the same sound/gsm config. |
| 20:43.51 | coolhp | cr2 : More progress... I have succeeded in booting all the way up to the login screen. |
| 20:44.13 | cr2 | coolhp: gpe login? |
| 20:44.19 | iclem | great, if one work the others will |
| 20:44.19 | coolhp | Nope |
| 20:44.25 | coolhp | Shell |
| 20:44.31 | coolhp | I've logged in a root. |
| 20:44.42 | coolhp | no gpe :-( |
| 20:44.43 | cr2 | on console? |
| 20:44.48 | coolhp | yeah |
| 20:45.26 | cr2 | look in the logs. it was discussed. my ba has no lcd right now, so i can't help you very much. |
| 20:46.26 | cr2 | iclem: the gpios may be (and are) different, but the general idea and chips are all the same. |
| 20:46.30 | iclem | cr2: what about putting your gnuharet source in xanadux cvs |
| 20:47.30 | cr2 | when it will boot linux. it has some implicit pxa27x things that should be sorted out. the asic3 stuff is broken right now. |
| 20:48.22 | cr2 | i have no idea how it will behave on hima and ba. |
| 20:49.17 | cr2 | but i will test it when the lcd arrives. |
| 20:49.22 | iclem | we are the alpha testers |
| 20:50.16 | cr2 | i was pulling the wrong gpio and then waited 1.5 months for the repair ;( |
| 20:50.58 | cr2 | i have put the source on sf.net, so alpha-alpha testers can hack on it. |
| 20:52.33 | coolhp | Do you have to hard-reset every time you want to leave linux ? |
| 20:52.53 | cr2 | on wince 2003 - yes. |
| 20:53.15 | cr2 | 2005 is a bit better ;) |
| 20:53.34 | cr2 | soft reset is enough. |
| 20:53.57 | coolhp | I'm not going back to 2005 on my BA... not in a million years.... |
| 20:54.03 | coolhp | Are you using 2005 on your BA ? |
| 20:55.02 | cr2 | i have 2005 on my universal. it is a real dog. ba is headless and lies under the table. |
| 20:55.33 | coolhp | What is the main difference between the BA and Universal ? |
| 20:55.39 | coolhp | 3G ? |
| 20:56.29 | cr2 | 480*640 lcd with 2x dpi. really great. all the rest is worse. yes, and the umts. |
| 20:59.50 | coolhp | LOL |
| 21:00.35 | dullard | nice keyboard tho |
| 21:00.57 | cr2 | that's true. |
| 21:01.48 | cr2 | but universal is heavy like a brick. i use my moto a780 as a phone. universal is a small notebook ;) |
| 21:04.23 | coolhp | Is anyone here using 2005 on their BA ? |
| 21:04.33 | coolhp | I found it extremely slow. |
| 21:05.10 | cr2 | hehe. even though ba has 128MB ram vs. 64 MB on universal. sigh. |
| 21:09.24 | coolhp | Oh... by the way.. is anyone using a BA with the 1.15 radio rom ? I've been hesitating to upgrade... is it any help to the GPRS connection issues ? |
| 21:13.38 | awelux_ | the graphic does appear |
| 21:13.54 | awelux_ | a blue line ater a large black block |
| 21:15.22 | cr2 | after or in ? |
| 21:15.31 | awelux_ | in the black block |
| 21:15.57 | awelux_ | but over the half of the black block. |
| 21:16.15 | awelux_ | about the virtual/physical mapping of the preloader: |
| 21:16.22 | awelux_ | HaRET(3)# pd 0xa3031000 16 |
| 21:16.23 | awelux_ | a3031000 | e28f30f8 e2822c01 e1a0a002 e3a00000 | .0...,.......... |
| 21:16.23 | awelux_ | HaRET(4)# vd 0xa3031000 16 |
| 21:16.23 | awelux_ | a3031000 | 00740074 003a0070 002f002f 00770077 | t.t.p.:././.w.w. |
| 21:16.54 | cr2 | the large black block is the pages in 64MB. the blue pixels are the pages where the kernel and initrd are loaded. |
| 21:17.13 | awelux_ | I guessed as much. nice |
| 21:17.20 | cr2 | do you have the 1:1 entry for this memory block ? |
| 21:17.36 | awelux_ | what is the 1vs1 entry? |
| 21:17.47 | awelux_ | I'm after goodbye cruel world now |
| 21:17.54 | awelux_ | and it's not right it seams |
| 21:18.36 | cr2 | bootlinux sets this enrty for the 1MB ram (aligned) where the preloader lives |
| 21:18.37 | awelux_ | actually I don't understand how a virtual address and a physical address should match at all |
| 21:19.07 | awelux_ | ok. but why doesn't it show up with vd then ? |
| 21:20.07 | cr2 | if you have switched the mmu on, then the translation is given by your page tables. if it maps everything 1:1 then you have phys=virt. |
| 21:21.09 | awelux_ | ok. so why do I jump to a physical address when the mmu is still turned on? Doesn't this correspont to a random virtual address? |
| 21:21.11 | cr2 | you can't do vd at such stage where virt=phys. it is also only for 1MB. |
| 21:21.23 | awelux_ | ok. |
| 21:21.37 | cr2 | no. for this 1MB they are equal. |
| 21:22.06 | awelux_ | the problem is still: the jump to the preloader fails. |
| 21:22.09 | cr2 | so it is virtual, but because of the respective page table entry it is also physical. |
| 21:22.39 | awelux_ | uh. I hoped this would make sense after a while |
| 21:22.46 | cr2 | i will look at it. |
| 21:23.10 | awelux_ | 0xa3031000 is really a physical address? it seams awfull large. |
| 21:23.28 | cr2 | no. |
| 21:23.37 | cr2 | 0xa4* is 64MB. |
| 21:23.50 | awelux_ | so the a4 don't matter |
| 21:24.06 | awelux_ | wait |
| 21:24.09 | awelux_ | calculating. |
| 21:24.28 | cr2 | 16*4=64. |
| 21:26.43 | awelux_ | ic. 0x4000000 is 64MB. So 0x00031000 and 0xa3031000 is the same for the cpu right ? |
| 21:27.38 | cr2 | 0xa0000000 is where the physical memory starts. |
| 21:28.37 | cr2 | so the physical addresses are from 0xa0000000 to 0xa3ffffff |
| 21:28.52 | awelux_ | oh. ok. And jumping to physical memory with mov pc,0xa3031000 is ok while the mmu is turned on? |
| 21:29.44 | cr2 | yes. when the l1 page enrty for the 0xa3000000 points to 0xa3000000 |
| 21:30.21 | awelux_ | ok. So I've to make sure that it does now |
| 21:30.29 | cr2 | you are always jumping to the virtual address if the mmu is on. |
| 21:31.10 | cr2 | that's what the mmu is for. |
| 21:37.21 | awelux_ | taking a closer look at memory.cpp now. |
| 21:38.46 | awelux_ | I tried to find a description of the pxa MMU but no success so far. |
| 21:39.02 | awelux_ | do you have a reference guide from intel cr2? |
| 21:40.11 | cr2 | only the standard manual. i think memory.cpp is already irrelevant when you are in the asm code. |
| 21:40.46 | awelux_ | not really. there seams to be a problem with exactly this 1:1 mapping. |
| 21:41.15 | cr2 | but it is made without any memory.cpp code. |
| 21:41.44 | awelux_ | any idea how I could test it? All the memphysAddr functions are from memory.cpp |
| 21:42.20 | cr2 | where do you need the memphysAddr ? |
| 21:43.05 | awelux_ | around where the memory for the preloader get's allocated and the preloader gets copied to it's final location. |
| 21:44.59 | awelux_ | <PROTECTED> |
| 21:44.59 | awelux_ | <PROTECTED> |
| 21:45.12 | cr2 | the preloader is allocated very early. |
| 21:45.22 | cr2 | is the preloaderPA reasonable ? |
| 21:45.32 | awelux_ | the first line alignes the preloader somewhere in the just allocated virtual ram. |
| 21:45.55 | awelux_ | the second line get's the corresponding physical ram address with black magic |
| 21:46.22 | awelux_ | yes. the prelaoderPA points into physical ram |
| 21:48.33 | awelux_ | then the preloader code get's copied into the virtual ram with memcpy. It shows up at the expected address with pd. |
| 21:48.56 | cr2 | well. we can replace the preloader allocation code with something else, but it will be tricky to push it to the top of the memory. i'm not excited by the memory allocation with malloc. It calls for trouble. |
| 21:49.38 | cr2 | how big is the preloader ? 1 page ? |
| 21:49.40 | awelux_ | now somehow the virtual address 0xa3031000 would be set to point to the physical address 0xa3031000 |
| 21:49.48 | awelux_ | 1 page shouold be enough. |
| 21:49.56 | cr2 | ok. |
| 21:50.05 | awelux_ | size is around 150 byte I think |
| 21:50.44 | cr2 | preloaderPA>>20 is the l1 entry for this 1MB. |
| 21:51.46 | cr2 | after that line you have the 1:1 mapping for this 1MB. |
| 21:52.15 | awelux_ | the MMU works with 1MB blocks ? |
| 21:52.24 | awelux_ | I thought it's finer |
| 21:52.44 | cr2 | the level1 - yes. |
| 21:53.11 | awelux_ | ah. k. I guess reading up on MMU's won'T hurt ;) |
| 21:53.20 | cr2 | then there are 64K and 4K |
| 21:54.34 | cr2 | hmm. the bootMachineType and memPhysAddr can be used as globals. probably no need to transfer them in registers. |
| 21:55.09 | awelux_ | I'm pretty sure the ext "C" calls work well |
| 21:55.21 | awelux_ | I checked them with test functions. |
| 21:55.25 | cr2 | the cpuFlushCache() - ist it pxa27x -specific ? |
| 21:55.54 | cr2 | there should be a wince function for that. |
| 21:56.06 | cr2 | without any wild hacks. |
| 21:57.55 | cr2 | the linux on palm lifedrive (pxa27x) has some inline asm in the bootloader. we can look at it. |
| 21:58.28 | awelux_ | Good idea. |
| 21:58.59 | awelux_ | The cpuFlush is using assembler. |
| 22:03.28 | cr2 | http://garux.sf.net |
| 22:04.55 | cr2 | asmloader.c |
| 22:08.03 | cr2 | it contains asm only when really necessary. otherwise it is in C. |
| 22:08.56 | awelux_ | I see the part with "Set Page table entries for 0xA0000000 and 0xA1000000 to identity. |
| 22:08.57 | awelux_ | <PROTECTED> |
| 22:09.23 | cr2 | it is palm-specific. |
| 22:09.58 | cr2 | the preloaderPA>>20 is much more portable. |
| 22:10.00 | awelux_ | that's the thing about the dissaaaembly they wrote in the descriptio? |
| 22:10.39 | cr2 | we don 't care . wince is simplier and better documented :) |
| 22:12.02 | cr2 | my guess is that something pxa27x specific is missing making the haret very fragile. after all i have booted the kernel on the universal with this code. |
| 22:12.58 | cr2 | with the standard haret-0.3.6, actually. |
| 22:14.06 | awelux_ | did it work every time for you? |
| 22:14.34 | cr2 | no. not always. there was a problem with the >4MB initrd. |
| 22:15.50 | awelux_ | ok. the initerd I'm using is larger than 5mb :/ |
| 22:15.58 | awelux_ | might be part of the problem |
| 22:16.07 | cr2 | can you try without it ? |
| 22:16.20 | awelux_ | yes |
| 22:16.52 | awelux_ | set INITRD "" ? |
| 22:18.07 | *** join/#htc-blueangel rob_w|mis (n=rob_w@p213.54.203.215.tisdip.tiscali.de) |
| 22:18.23 | cr2 | yes. |
| 22:19.38 | cr2 | now i have a new idea. the last 1MB of ram. when we switch the mmu off, the dma descriptor will be invalid... we need to deal with it ... |
| 22:19.55 | cr2 | ah. it is only on the universal. |
| 22:20.26 | cr2 | hmm. but the sram is also mapped somewhere in the virtual memory. |
| 22:20.47 | cr2 | what is your fb phys/wincevirt address ? |
| 22:20.59 | rob_w|mis | and guys .. how is it going ? |
| 22:21.37 | cr2 | hi rob. debugging is not easy ;) i have my universal back now. |
| 22:21.51 | awelux_ | ah. great news. |
| 22:21.57 | awelux_ | take care of it ;)( |
| 22:21.59 | awelux_ | ;) |
| 22:22.17 | rob_w|mis | cr2, i can feel with ya |
| 22:22.52 | rob_w|mis | today i had a very unsuccessfull day .. |
| 22:22.59 | cr2 | rob, i will get the ba lcd tomorrow or on wednesday. |
| 22:23.03 | cr2 | why ? |
| 22:23.22 | rob_w|mis | i cracked up to make a good rs485 connection |
| 22:23.38 | rob_w|mis | cool .. so will you get the correct lcd this time ? |
| 22:23.51 | cr2 | i hope ;) |
| 22:24.02 | rob_w|mis | wish you the best |
| 22:24.26 | cr2 | thanks. so what was with the rs485 ? |
| 22:24.46 | rob_w|mis | it is NOT a rs232 so to speak |
| 22:25.01 | cr2 | differential ? |
| 22:25.20 | rob_w|mis | yep , i guess that you called it |
| 22:25.47 | cr2 | good for speed and noise, i guess. |
| 22:25.58 | rob_w|mis | so tx+ tx - rx+ rx- where 5+ and 5- |
| 22:26.09 | rob_w|mis | they say |
| 22:26.14 | rob_w|mis | well |
| 22:26.18 | cr2 | like the twisted pair. |
| 22:26.29 | rob_w|mis | speed i dont know but noise it is |
| 22:26.47 | cr2 | ok. |
| 22:26.51 | rob_w|mis | but i dont i even think i wired it up correctly |
| 22:27.23 | rob_w|mis | then also you need to toggle sending /recieving with RTS , but i didnt manage it |
| 22:27.27 | awelux_ | interesting. I tired boot2 twice and the second time it crashed the device with the blue line being smaller. |
| 22:27.40 | cr2 | tx+ to rx+ , rx- to tx- |
| 22:27.40 | rob_w|mis | <PROTECTED> |
| 22:28.11 | cr2 | awelux: the same on the universal |
| 22:29.07 | cr2 | rob_w: there must have been some linux-lab project |
| 22:30.07 | cr2 | i have also seen a terribly fragmented blue line. |
| 22:30.10 | rob_w|mis | oh any examples are greatly appriciated |
| 22:31.07 | cr2 | search for linux lab rs485 |
| 22:32.04 | awelux_ | can we think of a test for the virtual/physical 1:1 mapping? If I do a getmem right after the mmu [preloaderPA >> 20] I should find the preloader code with a normal val = *preloaderPA right ? |
| 22:32.54 | cr2 | yes. but how you will give it out ? |
| 22:33.19 | awelux_ | bit by bit would work. What about the FB ? |
| 22:33.40 | cr2 | ok. but how ? |
| 22:34.01 | awelux_ | I could switch off/on the key led and vibra again. |
| 22:34.15 | cr2 | you can memcpy it into the fb, but does it help ? |
| 22:34.16 | awelux_ | maybe even morse the bit's when I press a key. |
| 22:34.32 | cr2 | if you understand morse :) |
| 22:34.56 | awelux_ | would it be displayed right away if it's memcopied into the FB? |
| 22:35.26 | awelux_ | ok not real morse code. Just something that works ;) |
| 22:35.41 | cr2 | if you will think about rgb565 encoding, then yes. |
| 22:35.44 | rob_w|mis | cr2, search querys are broken at that lab server .. and i cant find something digging the ftp |
| 22:36.37 | cr2 | well. look for the other links then. |
| 22:37.45 | rob_w|mis | will do |
| 22:42.50 | *** join/#htc-blueangel rob_w|mis (n=rob_w@p213.54.203.215.tisdip.tiscali.de) |
| 22:47.32 | awelux_ | Right now I'm jumping from haret back to wince and wince freezes |
| 22:49.33 | awelux_ | no. not freezes. but nquite broken. I will check where haret stops. |
| 22:49.41 | awelux_ | It's different from the old code |
| 22:49.55 | awelux_ | I guess it has something to do with the disable dma or peripherals |
| 22:50.26 | cr2 | or course. this is very destructive. |
| 22:52.38 | awelux_ | (cpu->shutdown_peripherals)(); isn't liked |
| 22:53.00 | cr2 | comment it out. |
| 22:53.24 | awelux_ | I wonder why it worked before. |
| 23:17.09 | *** join/#htc-blueangel JuNuKN (n=JuNuKN@Med35.m.pppool.de) |
| 23:31.57 | awelux_ | ok. after I commented out the (cpu->shutdown_peripherals)() I get to the mysterious jump to preloader again. |
| 23:32.23 | awelux_ | I can send pixels to the FB up to the end. |
| 23:33.16 | awelux_ | So an approach is now to give feedback over memory content at the preloader address after the 1:1 mapping to the FB. |
| 23:33.42 | awelux_ | I will do this tomorrow though. |
| 23:34.38 | cr2 | i think the best thing is to blacken the blue pixels again and set the red pixels at the bottom of the memory as the pages are copied. then make a buzz at linux boot :) |
| 23:37.15 | awelux_ | hehe. yes that would be nice. |
| 23:37.50 | awelux_ | But I also suggerst to place dot's somewhere in a bar for almost every line of the boot progress. |
| 23:38.03 | awelux_ | Somehow line lilo's dot's , |
| 23:38.17 | awelux_ | the number of printed dots indicate the problem |
| 23:38.45 | cr2 | there is plenty of space at least on the universal. |
| 23:38.59 | awelux_ | on the magician as well. |
| 23:39.16 | awelux_ | Just a tiny part of the middle of the screen get's blackened. |
| 23:39.43 | cr2 | yes. it is only 16384 pixels. |
| 23:39.45 | awelux_ | I've to look into the alignment. the block starts somewhat random at the middle |
| 23:40.05 | cr2 | you never know how it is mapped. |
| 23:40.41 | cr2 | but it is possible to call allocphysmem() to make it contiguous, but it is really not worth it. |
| 23:41.38 | awelux_ | true. the graphic stuff should be left for later. |
| 23:41.52 | awelux_ | indication about the boot process is nice though. |
| 23:42.27 | cr2 | yes. even this strip tells a lot about the boot process. |
| 23:42.44 | cr2 | a lot of useful debug info. |
| 23:43.36 | awelux_ | I placed a lot of small dot's on the screen like that: |
| 23:43.42 | awelux_ | <PROTECTED> |
| 23:43.42 | awelux_ | <PROTECTED> |
| 23:43.42 | awelux_ | <PROTECTED> |
| 23:43.42 | awelux_ | <PROTECTED> |
| 23:43.42 | awelux_ | <PROTECTED> |
| 23:44.06 | *** join/#htc-blueangel eXi (n=exile@linux-geeks.de) |
| 23:44.27 | awelux_ | Just copied your fb code |
| 23:44.30 | cr2 | yes. it is just so easy :) |
| 23:45.07 | awelux_ | about the mmu mapping. |
| 23:45.23 | awelux_ | is it possible mthat the 2nd and 3rd level mapping is off ? |
| 23:46.19 | awelux_ | And do you have a suggested reading about the mmu of the pxa? |
| 23:46.20 | cr2 | not really, but you can look at the "dump mmu" code. |
| 23:46.26 | awelux_ | k |
| 23:47.00 | cr2 | but this code is not very reliable and crashes most computers. |
| 23:47.35 | cr2 | probably because of the page tables changing during the dump and no locking. |
| 23:50.58 | awelux_ | dump mmu displays a lot |