| 02:01.55 | *** join/#htc-blueangel keith (n=keith@pool-71-112-104-182.sttlwa.dsl-w.verizon.net) |
| 02:57.26 | *** join/#htc-blueangel keith (n=keith@pool-71-112-104-182.sttlwa.dsl-w.verizon.net) |
| 09:15.29 | *** join/#htc-blueangel rob_w (n=bob@p213.54.207.172.tisdip.tiscali.de) |
| 09:15.29 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 12:10.10 | rob_w | lkcl, you there ? |
| 12:46.29 | lkcl | no, dude, am invisible |
| 12:46.35 | lkcl | oops he's gone :) |
| 15:53.19 | *** join/#htc-blueangel Al_3sjad (n=al_3sjad@85.154.147.215) |
| 15:54.05 | Al_3sjad | Good evening MR.lkcl |
| 16:01.03 | Al_3sjad | I think I come very early today. :) any way I will come later. |
| 16:25.25 | *** join/#htc-blueangel jurop88 (n=kvirc@81-208-36-84.ip.fastwebnet.it) |
| 16:57.16 | *** join/#htc-blueangel rob_w (n=bob@p213.54.207.172.tisdip.tiscali.de) |
| 16:57.16 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 17:06.56 | lkcl | heya folks. oops he's gone too |
| 17:07.28 | cr2 | hi lkcl |
| 17:07.44 | cr2 | what do you want to compile today ? |
| 17:09.43 | lkcl | muhahaha |
| 17:09.53 | cr2 | i am thinking about adding the magician to the un0 tree. pH5 has disappeared... |
| 17:10.10 | lkcl | i am off to europe for a month, btw. |
| 17:10.16 | lkcl | to the un0 tree? yeh, why not. |
| 17:10.26 | lkcl | un0... un0.... |
| 17:10.32 | lkcl | why the un0 tree? |
| 17:10.41 | cr2 | off to europe ? |
| 17:11.08 | cr2 | because it is the same pxa270, and no asic3. should be no big problem. |
| 17:12.43 | cr2 | looking for somebody with alpine and hw651x :) |
| 17:14.27 | cr2 | can gpe/opie be compiled "as-is" from the monotone tree, or do i need some patches ? oh. so many questions at once. |
| 17:30.45 | *** join/#htc-blueangel rob_w (n=bob@p213.54.207.172.tisdip.tiscali.de) |
| 17:30.45 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 17:30.51 | lkcl | ack. |
| 17:31.05 | lkcl | i got an orange c500 :) |
| 17:31.09 | lkcl | spv c500 |
| 17:31.25 | cr2 | what is that ? |
| 17:31.31 | lkcl | it's an htc tornado. |
| 17:31.34 | lkcl | or something. |
| 17:31.46 | cr2 | ooh. a hackable one. |
| 17:31.49 | lkcl | yehh. |
| 17:31.52 | lkcl | anyway. |
| 17:32.08 | lkcl | i _must_ go - if where i'm staying tonight has internet access i'll be able to thingy. |
| 17:32.11 | lkcl | wotsit. |
| 17:32.24 | lkcl | i don't know the conditions for where i will be staying in europe. |
| 17:32.41 | cr2 | ok. |
| 17:32.42 | lkcl | certainly there won't be internet access unless i can wrangle some directional wireless |
| 17:32.52 | lkcl | and at work i don't know either! |
| 17:33.05 | lkcl | bit of a pisser. |
| 17:33.14 | lkcl | okay gotta pack. |
| 17:33.25 | lkcl | ttfn! |
| 17:33.28 | *** part/#htc-blueangel lkcl (n=lkcl@bb-87-82-4-146.ukonline.co.uk) |
| 18:42.33 | *** join/#htc-blueangel iclem (n=clement@iclem.net) |
| 18:46.45 | iclem | hi all |
| 18:47.19 | iclem | cr2 ? |
| 19:29.09 | cr2 | hi iclem |
| 20:25.32 | iclem | hi cr2 |
| 20:26.16 | iclem | Could you help me with IDA, I have some questions |
| 20:27.00 | cr2 | i am not an expert in ida. i prefer objdump ;) |
| 20:27.40 | cr2 | it should work on normal coff executables. |
| 20:28.21 | iclem | I would like to know how to find a GPIO action in a wince DLL |
| 20:28.37 | cr2 | the bootloader is more complicated. i don't know how to use two "segments" simultaneously. |
| 20:28.53 | cr2 | look in the data section. |
| 20:28.56 | cr2 | hima? |
| 20:29.13 | iclem | i've loaded asic3 IDC but I can find any ref to the ASIC area |
| 20:29.21 | iclem | yes with hima dlls |
| 20:29.46 | iclem | s/can/cannot |
| 20:30.06 | cr2 | in the data segment you should see the full table of addresses. |
| 20:30.46 | cr2 | they don't declare them const, so these things end in data. |
| 20:30.53 | cr2 | .data |
| 20:32.55 | cr2 | gpioA is a9300000, b a9300100, and so on. look at the memory map and asic3 wiki pages. |
| 20:33.43 | cr2 | the gpioX is normally referenced from .text, and then an offset is added in the .text |
| 20:34.23 | cr2 | rename a9300000 as asic3gpioa, etc. then you can search for the name. |
| 20:35.16 | cr2 | very painful and tedious process... try to use watch program first. many things are more obvious this way. |
| 20:35.33 | iclem | This is done by the asic3-wince.idc in IDA, but I do not find any of these address |
| 20:35.45 | cr2 | get my latest haret version, and tell me if wadd16 works. |
| 20:36.06 | iclem | it watches change at a particular address ? |
| 20:36.55 | cr2 | you don't need *.idc it works also without. maybe the *idc is wrong. i have removed a lot of incorrect info and unfounded things from wiki. |
| 20:41.52 | cr2 | the sd part of the script is obviously wrong. the rest is incomplete. |
| 20:44.11 | cr2 | the asic5 serial port is at +0x700 |
| 20:44.29 | cr2 | himalayamemorymap wiki page is now correct. |
| 20:46.08 | iclem | how do you proceed with objdump ? |
| 20:46.23 | cr2 | forget it. it is a joke. |
| 20:47.31 | cr2 | you can just dump the data section with objdump and grep for the addresses. |
| 20:47.42 | cr2 | which dll is that? |
| 20:50.38 | iclem | battery.dll or HSDCard.dll |
| 20:52.33 | cr2 | sd controller ? |
| 20:53.37 | iclem | yes, SDIO |
| 20:53.48 | cr2 | section data: 01f* - these are dll references. |
| 20:53.51 | cr2 | then something. |
| 20:54.12 | cr2 | then 0xaee00000 - pxa gpio base |
| 20:54.35 | cr2 | then a93* - all asic3 base registers. |
| 20:54.54 | cr2 | the sd controller has only 1 interrupt. is it not known ? |
| 20:56.17 | cr2 | do you have sdio cards? |
| 20:56.43 | iclem | no, it was just to test to find well known addresses |
| 20:57.22 | *** join/#htc-blueangel JuNuKN (n=JuNuKN@Me58f.m.pppool.de) |
| 20:57.39 | JuNuKN | Hello iclem and cr2 |
| 20:57.45 | iclem | in IDA, in .data, i see groups of DCB, could thez be addresses ? |
| 20:57.54 | iclem | hi JuNuKN |
| 20:58.06 | cr2 | hi JuNuKN. |
| 20:58.28 | cr2 | iclem: if they are referenced, then they are dcd |
| 20:58.37 | JuNuKN | Hello together .. I read in actual log, that you are discussing about hima and sdio ;o) |
| 20:58.38 | iclem | ie : DCB 0 DCB 0x10 DCB 0x30 DCB 0xA9 |
| 20:58.46 | cr2 | not all are referenced. |
| 20:58.55 | cr2 | yes. modify it in dcd. |
| 20:59.10 | iclem | how can this be done ? |
| 20:59.43 | iclem | found, changed as double word |
| 20:59.46 | cr2 | it was probably just #include "asic3.h" with a structure. |
| 21:00.19 | cr2 | iff there is no reference, they are just dummies. |
| 21:01.07 | iclem | yes but i can check reference in code then |
| 21:01.21 | iclem | ok some changed to thei names, ie GPIO_A_MASK |
| 21:01.36 | cr2 | press "n", and give it a name. |
| 21:01.37 | cr2 | ok. |
| 21:02.33 | iclem | cool, thank you, now I think I can dig a bit more |
| 21:02.40 | cr2 | JuNuKN: sdio is very complicated. not technically, but other things. |
| 21:03.31 | JuNuKN | Shure I only thought maybe you need some more logs from SDIO Access with Hima ... |
| 21:04.14 | cr2 | you can join forces with iclem, and fill the missing gpios in wiki. |
| 21:04.15 | JuNuKN | No further plans ;o) My only plan is to somewhen have linux on may hima with Phone enabled and sound, |
| 21:06.27 | *** join/#htc-blueangel Al_3sjad (n=al_3sjad@85.154.147.215) |
| 21:06.31 | JuNuKN | Like it is today ... but stored and accessible on SD card, to customize it and try some userland changes |
| 21:07.00 | cr2 | JuNuKN: can you run "kmalloc" on hima ? |
| 21:07.32 | JuNuKN | I can do .... If you like ... |
| 21:08.23 | cr2 | yes. if you can. |
| 21:09.19 | cr2 | we need to write a wince program that periodically plays some short sound, and trace it with haret. |
| 21:09.39 | cr2 | open,play,close,open,play,close,... |
| 21:09.58 | JuNuKN | By the way ... I tested an other Magician with my SD card and it worked fine with it :o( Then I asked for testing an other SD card, to work with mine ,,,, :o)) It worked with that. So I bought it ! 1 Gbyte 64Euro :o( Maybe I will sell mine now on Ebpay |
| 21:10.46 | cr2 | great. |
| 21:10.56 | JuNuKN | Why not playing a mp3 song for 3 minutes or so, while tracing ? |
| 21:11.30 | cr2 | it is open/close i"m interested in, not play . |
| 21:11.42 | JuNuKN | Or maybe a book as mp3 for half an hour and trace ? |
| 21:12.06 | JuNuKN | Ahh ok ... Would you prepare it ? |
| 21:12.18 | cr2 | all you will see is spi traffic on pxa gpio 23,24,25. |
| 21:12.35 | JuNuKN | Otherwise I can do it by hand Open Play Open Play some ding-sounds or so |
| 21:12.41 | cr2 | we need the asic3 gpios for enabling sound. then there is also a headset. |
| 21:13.23 | cr2 | you will use touchscreen and do many other nasty things that can't be separated from each other. |
| 21:13.25 | JuNuKN | So lets start with the jobs ... |
| 21:13.51 | cr2 | run kmalloc, it is the easiest. |
| 21:14.04 | JuNuKN | First do a kmalloc with actual gnuharet version ? |
| 21:14.39 | cr2 | yes. i don't have a newer version now. |
| 21:15.24 | cr2 | i will try to do "wmalloc" today. for some reason it fails on my jornada820, but was working on bluenagel. |
| 21:15.49 | JuNuKN | So lets check if I have the newest on my Hima 200601202311 |
| 21:15.57 | cr2 | ok. |
| 21:16.34 | JuNuKN | Is it the newest ? I don't remember the download lionk |
| 21:16.55 | cr2 | 23:11 sounds good :) |
| 21:19.38 | JuNuKN | OK it took a while but now we are connected ;o) |
| 21:20.13 | cr2 | i have uploaded a new version, which has "wmalloc" |
| 21:20.26 | cr2 | but kmalloc is the same. |
| 21:21.24 | JuNuKN | 600Blocks still with old Haret |
| 21:21.34 | JuNuKN | .. till now 900 |
| 21:23.14 | JuNuKN | 1600 till now |
| 21:23.57 | JuNuKN | You know, that we have around 100MB free ? |
| 21:24.11 | cr2 | 100 ? |
| 21:24.35 | cr2 | but necessarily contiguous. |
| 21:25.06 | JuNuKN | Shure 100MB ! |
| 21:25.19 | cr2 | we will see. |
| 21:26.05 | cr2 | after i will calculate the pixclock for the magician, we can try to create a magician kernel :) |
| 21:26.07 | JuNuKN | must be arround 7000 Blocks ? |
| 21:26.20 | cr2 | divide by 4 |
| 21:26.34 | cr2 | 25000 |
| 21:27.16 | JuNuKN | On the mag we have 1700 with 35 MB and here we have 100MB free !? |
| 21:27.25 | cr2 | but i don't believe. there must be a 25bit (32MB) limit. if not, it is interesting. |
| 21:28.17 | cr2 | one stupid program allocates a page in the middle, and... kmalloc breaks. |
| 21:28.19 | JuNuKN | Isn't the space of magician not to small to start a kernel ? |
| 21:28.43 | cr2 | 6MB ? |
| 21:28.59 | cr2 | haret can't boot the kernel >4 MB by design. |
| 21:29.44 | cr2 | it will overwrite the ramdisk. |
| 21:29.59 | JuNuKN | Ahh, OK and after the initial boot it takes care of the mmu !? It pushes away the Wince out of memory !? |
| 21:31.35 | cr2 | haret switches off the mmu, after it has loaded the kernel and ramdisk. wince is dead after the interrupts are swtched off :) then zImage is unpacked and linux takes over the memory you told it on the command line. |
| 21:32.42 | JuNuKN | Isn't it using the mmu ? |
| 21:33.54 | cr2 | yes, it is. but it does not care if the memory is empty of filled with some wince stuff. |
| 21:33.55 | JuNuKN | till now 3900 blocks |
| 21:33.56 | cr2 | what is the virtual address ? |
| 21:33.58 | JuNuKN | It is fastcinating |
| 21:34.17 | JuNuKN | always 2b0000 |
| 21:34.26 | cr2 | ok. |
| 21:34.47 | JuNuKN | paddr = a4833000 |
| 21:35.27 | cr2 | 16MB, not bad. you can, in principle, tell linux to use only this memory. then wince will not even know that linux was running :) |
| 21:35.32 | JuNuKN | it switches from time to time ... you will see in the logs |
| 21:35.50 | cr2 | the physical base address ? |
| 21:37.09 | JuNuKN | yes paddr a4eesometing a53a0someting and so oin |
| 21:37.25 | JuNuKN | uuups ready 4421 pages |
| 21:38.13 | cr2 | a4* is the upper 64MB, the lower is reserved for ramdisk. we can try to push it lower, but wince kernel will resist ;) |
| 21:39.12 | JuNuKN | there are also some a57ba000 |
| 21:39.30 | JuNuKN | Ready to send log? |
| 21:39.44 | cr2 | there is a setmemorydivision function. |
| 21:39.53 | cr2 | has it stopped ? |
| 21:40.09 | JuNuKN | Yes ... I reported ,,, 4421 pages |
| 21:40.17 | cr2 | ok. |
| 21:41.21 | cr2 | now you can get the latest version and try "wmalloc". it is one more way to deprive wince of ram :) |
| 21:41.27 | JuNuKN | Should I send log or doing other stuff first ? |
| 21:41.59 | cr2 | don't send the log. it is more or less clear now what is going on. |
| 21:42.56 | JuNuKN | Please give me again the link |
| 21:43.42 | cr2 | http://jornada820.sf.net/files/haret/ |
| 21:46.31 | JuNuKN | only wmalloc? without dump? Do I have to hardreset it afterwards ? |
| 21:47.02 | cr2 | only wmalloc. i think you should reset anyway. |
| 21:47.22 | JuNuKN | ut only softreset !? |
| 21:47.32 | cr2 | yes. |
| 21:47.52 | JuNuKN | uuups thats fast! |
| 21:48.04 | cr2 | we just allocate the memory, do destructing anything. |
| 21:48.07 | cr2 | it works ? |
| 21:48.38 | JuNuKN | ... of 423 |
| 21:50.51 | JuNuKN | .. of 800 |
| 21:51.57 | JuNuKN | iclem how about you ? What are you working on ? |
| 21:52.45 | iclem | I'm learning how to disassemble a wince dll ;) |
| 21:53.31 | JuNuKN | .. of 1100 |
| 21:53.33 | iclem | found reference to charger_enable, ac_in and board_id3 in battery_dll |
| 21:54.04 | cr2 | with "strings" ? :) |
| 21:54.12 | JuNuKN | Give me a link to tool and instructions ;o) I will read it too ... |
| 21:54.27 | cr2 | strings wince.dll |
| 21:54.33 | iclem | lol |
| 21:55.09 | iclem | cr2 i need a confirmation |
| 21:55.10 | cr2 | strings gives the most useful info, iclem :) |
| 21:55.42 | iclem | let me find this strange code |
| 21:55.47 | cr2 | and 'strings -el' for unicode. |
| 21:56.10 | iclem | MOV R1, R0,LSL#16 |
| 21:56.19 | iclem | MOV R2, R1,LSR#16 |
| 21:56.37 | cr2 | it is just &0xffff |
| 21:56.47 | iclem | is the same as moving helf word from R0 to R2 ? |
| 21:56.54 | iclem | s/helf/half |
| 21:57.05 | cr2 | cut 16bit from a register. |
| 21:57.12 | cr2 | yes. |
| 21:57.34 | cr2 | asic3 regs are16bit. |
| 21:58.04 | iclem | thanks |
| 21:59.26 | JuNuKN | iclem again a question : Is your hima eally still working with sd card and GPE image? |
| 21:59.36 | iclem | SO if i've got GPIO_C_MASK+0x30 tested with 0x40, it means GPIOC05 is tested ? |
| 21:59.56 | iclem | JuNuKN: yes |
| 22:00.31 | JuNuKN | Great! Do you belive in deifferences of SD card access by board ID 's ? |
| 22:00.42 | iclem | JuNuKN: given the howto i give you |
| 22:00.56 | iclem | I'm on battery.dll right now |
| 22:02.17 | iclem | JuNuKN: lkcl and you have the same boardid, AS asic3_mmc work for him, it should work for you too |
| 22:02.39 | JuNuKN | Ohhh I didn't know that! |
| 22:03.47 | cr2 | iclem: can you compile the asic3_mmc into the kernel and boot with root=/dev/mmc* like all other handhelds projects ? |
| 22:04.40 | iclem | of course |
| 22:05.08 | cr2 | i think we should do it to avois all these troubles. |
| 22:05.42 | iclem | okay, let's go |
| 22:06.04 | JuNuKN | Is someone of you familar with OE? Maybe we can later setup the instructions for OE to download the stuff, compile it , download a prepared GPE image and intrd and doing the things automatically !? |
| 22:06.51 | cr2 | i am not familiar, but eventually we must do it. |
| 22:07.12 | cr2 | for all 5-6 platforms we have :) |
| 22:08.16 | JuNuKN | In principle, I think it should only be a script, so cvs, wget export path make and so on ... It should save a lot of time and errors |
| 22:08.23 | cr2 | JuNuKN: is your hima still alive ? |
| 22:08.43 | JuNuKN | Yes it is ... of 2100 |
| 22:09.37 | JuNuKN | It shows me the face of billy with tieres in his eyes ;o) |
| 22:09.54 | cr2 | :))) |
| 22:10.16 | JuNuKN | ment tears |
| 22:10.17 | cr2 | make such a logo for haret. |
| 22:11.07 | JuNuKN | I am no designer and not familar with photo stuff |
| 22:11.27 | cr2 | penguin takes the dollar packs from crying billy :) |
| 22:11.32 | JuNuKN | ... of 2240 till now |
| 22:11.59 | cr2 | ok. it must me more than kmalloc |
| 22:12.01 | JuNuKN | Yes would be great joke ;o) |
| 22:12.56 | JuNuKN | By the way, you where on ccc ? How was the try of a hack to the xbox360 ? |
| 22:13.14 | cr2 | animated. |
| 22:13.30 | JuNuKN | What does it mean`? |
| 22:13.51 | cr2 | i was talking with cp15 and some other nice guys at that time. |
| 22:14.09 | JuNuKN | Is it hacked or not? I don't found infos on the congress page till now |
| 22:14.23 | cr2 | a screen animation while kmalloc gets the memory . |
| 22:14.34 | cr2 | don't know. |
| 22:16.28 | JuNuKN | Nice animation ... Good Idea for a linux logo for the HTC devices ... somehow showing the dollars flowing from Microsoft to HTC and to telecom and smaller parst from telecom to customer |
| 22:16.45 | cr2 | i should allocate the memory by bisection. that will be faster. |
| 22:16.47 | cr2 | :))) |
| 22:17.27 | JuNuKN | ... of 2500 till now |
| 22:18.52 | JuNuKN | and the big trouser pockets from telecom. |
| 22:20.22 | JuNuKN | ... enought for the moment ... look whether you can help iclem or answer some open questions to him ... I will report, if it has stopped |
| 22:25.44 | cr2 | t-mob will sue us for using the magenta colour ;) |
| 22:31.44 | JuNuKN | ...of 3040 till now ... |
| 22:32.56 | JuNuKN | Lets take the bubbles from O2 and the T.Mobile Logo, which lets them out ... and a penguin in a boot, fishing for it :o) |
| 22:46.48 | JuNuKN | Now it reports from time to time something with bytes at the end |
| 23:15.31 | JuNuKN | till now we are at 4300 |
| 23:19.46 | cr2 | it is going slow. |
| 23:20.30 | JuNuKN | now same speed as at the beginning 4416 now |
| 23:20.52 | JuNuKN | nearly same speed |
| 23:21.53 | JuNuKN | Yes I am tired |
| 23:22.17 | cr2 | ok. the most important thing is that it works! :) |
| 23:22.45 | cr2 | i will also go to bed. |
| 23:23.02 | cr2 | good night. |
| 23:25.55 | JuNuKN | good night löast status 4550 |