00:01.04 | awelux | funny |
00:48.30 | *** join/#htc-blueangel awelux (n=awelux__@dslb-084-058-148-208.pools.arcor-ip.net) |
00:48.41 | awelux | last findings: |
00:49.37 | awelux | Somehow I'm loading the kernel or zImage to the wrong address. I'm loading to 0x10000 instead of 0x8000 |
00:50.09 | awelux | Probably my misstake. |
00:53.33 | awelux | But it looks as if the problem is back in the c copde building the kernel bundle (I touched that as well). |
00:53.40 | awelux | More tomorrow |
00:53.41 | awelux | gn8 |
07:08.24 | *** join/#htc-blueangel rob_w|mis (n=rob_w@p549B8BC7.dip0.t-ipconnect.de) |
08:19.34 | *** join/#htc-blueangel awelux_ (n=awelux__@dslb-084-058-175-245.pools.arcor-ip.net) |
11:00.42 | *** join/#htc-blueangel pH5 (n=ph5@p54866371.dip.t-dialin.net) |
15:09.54 | *** join/#htc-blueangel philverb (n=philverb@c-24-5-65-9.hsd1.ca.comcast.net) |
15:10.14 | philverb | Hi. Anybody in here with an Apache? |
15:16.01 | *** join/#htc-blueangel ekkobelli (n=michi@p54A3DAD2.dip.t-dialin.net) |
15:31.24 | cr2 | hi philverb |
15:31.46 | cr2 | no. but there is a lot of common hardware between the htc phones. |
16:23.34 | *** join/#htc-blueangel pH5 (n=ph5@e178213031.adsl.alicedsl.de) |
18:36.12 | *** join/#htc-blueangel rob_w (n=bob@p213.54.202.126.tisdip.tiscali.de) |
18:36.12 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
19:15.09 | *** part/#htc-blueangel pH5 (n=ph5@e178213031.adsl.alicedsl.de) |
19:46.27 | Funklord | damnit, I accidentally overclocked my BA.. and it hard reset |
19:46.33 | Funklord | using wince |
19:47.16 | Funklord | another reason why the wince ramdisk idea is flawed |
20:30.51 | awelux_ | hi |
20:31.41 | awelux_ | Funklord, I understand that it's just a matter of finding out how to initialize the devices. |
20:31.47 | awelux_ | "just" |
20:33.03 | Funklord | hmm? |
20:34.00 | awelux_ | if you want to flash the pda to boot linux directly there is more work to do than writing the bootloader |
20:34.08 | Funklord | I'm using winCE atm, it knows how to init my devices, just that it's such a pile of shit |
20:34.25 | awelux_ | ^^ |
20:34.33 | Funklord | ;P |
20:34.40 | awelux_ | that's why we do this work ;) |
20:37.53 | Funklord | yes, I should do some too.. but haven't had the time.. :( |
20:38.16 | Funklord | awelux_: are you working on the oe stuff? |
20:38.46 | awelux_ | no. |
20:38.53 | awelux_ | I'm working on the magician basics |
20:43.07 | cr2 | hi awelux |
20:43.45 | cr2 | i have finally removed the binary toolhelp.lib from haret, and found a way to generate it in the makefile. |
20:43.56 | cr2 | i think i will commit the source into CVS. |
20:45.54 | cr2 | i have also converted the linuxbios logo into an *.ico, so it can be used for the minibootloader. |
20:46.53 | cr2 | hmm. |
20:48.57 | cr2 | DLL Name: (null) |
20:53.52 | cr2 | through studying the dirty windows internals i will become a decent windows programmer :lol: |
20:53.58 | cr2 | it works now. |
21:13.26 | awelux_ | nice |
21:14.30 | awelux_ | I'm searching for a bug why I load the kernel to 0xa0210000 instead of 0xa0208000 |
21:14.49 | awelux_ | It seams that I make a mistake in the C part of haret already |
21:21.51 | *** join/#htc-blueangel lamikr (n=chatzill@aragorn.kortex.jyu.fi) |
21:24.15 | cr2 | ok. it it committed and even compiles and boots after the checkout. |
21:26.14 | awelux_ | nice. |
21:26.28 | awelux_ | Any progress on the 2.6.15 kernel ? |
21:26.42 | awelux_ | I got my new pda pen's today |
21:28.26 | cr2 | no. i'm looking after the gpios again. |
21:32.40 | *** join/#htc-blueangel husam (n=kvirc@217.206.111.105) |
21:32.50 | husam | hi cr2 |
21:33.11 | lamikr | hi all |
21:33.21 | awelux_ | hi |
21:34.04 | husam | awelux_ : ... any lock with your boot :) |
21:34.06 | lamikr | Did I understood 2 days ago correctly that you can boot to Linux without modifying the bootloader? |
21:34.33 | awelux_ | husam, not yet. Maybe a little progress. At least something to investigate. |
21:34.53 | husam | Thats good |
21:36.06 | cr2 | hi husam |
21:36.21 | cr2 | there are some regs missing in the acx100 driver. |
21:36.21 | lamikr | Something like putting special filesystem to mmc/sd card could allow the boot to Linux from the mmc card instead of booting to wince... |
21:36.26 | husam | hi ... |
21:36.37 | cr2 | 0x900 |
21:36.49 | cr2 | vendor/product. |
21:37.20 | husam | Let ne have look ... I have header files I got from tiacx group |
21:37.20 | cr2 | lamikr: is h6300 a htc phone ? |
21:37.42 | husam | No It's i[aq |
21:37.45 | husam | ipaq |
21:37.47 | lamikr | cr2: No, it's HP's omap based ipaq. |
21:38.01 | cr2 | most ipaqs are from htc. |
21:38.04 | cr2 | and loox |
21:38.07 | cr2 | and some others. |
21:38.10 | lamikr | But it has same MDOC with blueangel. |
21:38.15 | husam | good to know |
21:38.43 | cr2 | most, if not all. that's why i am asking about the bootloader. |
21:39.12 | cr2 | it's called "outsourcing" ;) |
21:39.28 | husam | Yah .. alots of compnay do that |
21:40.14 | lamikr | There are actually two new ipaq models 6800 and 6900... One of the them looked exactly like blueangel. |
21:40.20 | cr2 | lamikr: doc+ ? it is used only by himalaya. the rest uses g3. |
21:40.42 | husam | cr2: yes h6300 uses mdoc+ |
21:48.42 | cr2 | i have looked up the register table here: http://195.66.192.167/linux/acx_patches/current/ |
21:48.43 | *** join/#htc-blueangel immolo (i=tyactctn@cpc3-cosh1-4-0-cust151.cos2.cable.ntl.com) |
21:50.07 | husam | rc2: the header file I have is part of windows driver ... and it list all registers |
21:50.20 | lamikr | cr2: I had thought that the "DoC G3 64MiB" mentioned in http://www.handhelds.org/moin/moin.cgi/BlueAngel is same mdoc. |
21:50.57 | husam | Lamikr: we have G2 not G3 |
21:52.16 | cr2 | is it secret ? :) |
21:52.40 | husam | Nooo ... givme your email and I will send it to you :) |
21:53.50 | cr2 | i have enough things to do :) is the 0x900 register there ? |
21:54.05 | husam | Yes |
21:54.09 | cr2 | ok. |
21:54.49 | cr2 | i have seen in your driver that you set the buffer memory pointer to 0. |
21:55.22 | cr2 | is it possible to find this address from the control area ? |
21:55.56 | cr2 | i searched on google for the pci logs, and the positioning of this area is quite random. |
21:56.20 | husam | According to the code you should be able to get the offset from the control register |
21:57.02 | cr2 | which register should i read ? is the latest hx4700_wlan.c of any use ? |
21:57.29 | husam | the buffer is set to zero becuase I'm using the controller address to refernce |
21:58.13 | cr2 | i don't want to work with the driver myself, but this address seems to be the key to the whole story. |
21:58.23 | cr2 | so what is the "actual" address used by h6300 ? |
21:58.43 | cr2 | i like gathering such statistics in wiki :) |
21:59.04 | husam | No the register returns the offset and then you can use control register to address the memory location |
21:59.55 | cr2 | http://wiki.xda-developers.com/index.php?pagename=HTC_Partnumbers |
21:59.57 | cr2 | http://wiki.xda-developers.com/index.php?pagename=HTC_Serial |
22:00.03 | husam | IO_ACX_INFO_MAILBOX_OFFS |
22:00.25 | cr2 | http://wiki.xda-developers.com/index.php?pagename=HTC_DiskOnChip |
22:00.28 | cr2 | ok. |
22:00.42 | husam | :) wow |
22:01.48 | husam | and I'm using IO_ACX_SLV_MEM_ADDR, IO_ACX_SLV_MEM_DATA to read |
22:02.07 | husam | I never manage to directoly map the memory area |
22:02.53 | cr2 | it is the right way for the "control" registers, but is it the same for the buffer ? |
22:03.07 | husam | cr2: do you know if Himalaya has the new drivers for mdoc+ 32M!!! |
22:03.37 | cr2 | no. drivers. |
22:03.52 | husam | I think so .. let me see I think I have anther driver uses that for the buffer too |
22:04.14 | cr2 | there is a binary driver for the moto a780. it is for g3, but i think g2 is also supported. not sure though. |
22:04.38 | cr2 | but only for 2.4 |
22:05.11 | husam | ohh ... I manage to get the new driver to read partition table ... but not sure if i have the right info |
22:06.18 | cr2 | cp15 has written a wrapper for the motorola driver, and read some info, but i don't know about his further progress. |
22:06.53 | husam | ok ... how did you guys manage to boot from mmc/sd !!! |
22:06.55 | lamikr | Has cp15 worked also with the himalaya? |
22:07.46 | cr2 | no. lkcl has written the himalaya stuff. |
22:07.50 | husam | cr2: the code I have suggest thet you can use slave mode to with buffer |
22:08.15 | cr2 | husam: the asic3 sd/mmc driver is used on ipaq hx4700. blueangel and universal use the same chip. |
22:08.47 | husam | so I need to find out if h6300 uses teh same one |
22:09.17 | cr2 | don't think so. omap has a builtin sd ? |
22:09.34 | cr2 | magician uses the builtin pxa sd/mmc . |
22:09.37 | husam | yes thats true ... |
22:09.44 | cr2 | and all the later zaurus. |
22:10.42 | husam | I had look at the bootloader code for h6300 ... so far I can't see MMC/SD card code |
22:11.12 | cr2 | but if the buffer can be mmaped (like the pci) then it is much more comfortable and probably faster. |
22:11.41 | husam | Agree ... but I don't have the right memory layout |
22:12.21 | lamikr | cr2: So does the htc bootloader contain method for reading this special MMC/SD chipset during the boot? |
22:12.34 | cr2 | of course. |
22:13.19 | cr2 | lamikr: is the bootloader using virtual memory mapping ? |
22:13.45 | husam | cr2: yes bootloader using virtual memory mapping |
22:14.05 | cr2 | do you know the map ? |
22:14.08 | cr2 | http://wiki.xda-developers.com/index.php?pagename=UniversalMemoryMap |
22:14.13 | cr2 | somethig like that. |
22:14.13 | husam | yes |
22:15.09 | cr2 | well, but the there must be a sd/mmc control area there. |
22:16.06 | husam | omap uses control register to interface mmc/sd card |
22:16.18 | lamikr | cr2: So is the htc bootloader always trying to boot first from the SD/MMC card and if it does not find something that it is looking, then continue the boot by loading wince from the flash? |
22:16.58 | cr2 | in principle, yes. it is just a bit more complicated. |
22:17.33 | cr2 | because there are many options. and buttons that can change the boot process. |
22:18.37 | cr2 | we boot from wince right now, but once the suspend/resume works, wince can be dropped. |
22:18.52 | cr2 | like on hx4700. |
22:18.56 | husam | lucky you :) |
22:18.56 | lamikr | oh? Just by using the couple of buttons that are available in the htc? |
22:19.24 | lamikr | cr2: So is hx4700 booting directly from the MMC/SD? |
22:19.39 | cr2 | you can press some buttons and get a bootloader command line. |
22:19.59 | cr2 | hx4700 flashes its own bootloader. SD/mmc is an option. |
22:20.16 | cr2 | blueangel can do the same. |
22:20.18 | husam | I did that .. I manag to get h6300 into stage where it listen to usb connection ... |
22:20.30 | husam | but when I connected I got random data |
22:20.35 | cr2 | universal is more compilated, because everything is on the m-sys g3 |
22:21.10 | cr2 | is it a builtin usb slave ? |
22:22.24 | husam | I think so ... omap has built in USB |
22:23.03 | lamikr | I have actually newer tested this myself, but everett has put following from the h6300 boot: |
22:23.14 | lamikr | <PROTECTED> |
22:23.16 | lamikr | You must reboot the ipaq, and hold down the 'email' and 'contacts' buttons before the tmobile screen appears. Then insert the ipaq into the cradle, you should then see "BL" in bottom left hand corner. When the ipaq is in the cradle you should see "USB" at the bottom. If your ipaq is not plugged in, you should see "No AC"... now to configure minicom to work.... |
22:23.45 | cr2 | looks like htc :) |
22:23.48 | cr2 | http://wiki.xda-developers.com/index.php?pagename=Universal_Technical |
22:24.05 | husam | cr2: I'm looking at tiacx100 slavedma.c ... it looks like you can use controller register to write to buffer the same you you do with mailbox |
22:26.21 | cr2 | the virtual memory map shows 1MB for acx100. so there are only 15 64K options :) |
22:27.48 | husam | I guess the first area for contol registers and the other for shared memory area |
22:28.10 | cr2 | yes. the first is control. |
22:28.43 | cr2 | the question is which of the rest is the buffer. |
22:29.01 | husam | yes :) ... |
22:29.54 | husam | this why I was planinig to use control register ... lazy man approch :) |
22:30.01 | cr2 | i will look with haret :) |
22:31.14 | husam | Thats good |
22:35.49 | cr2 | don't see anything... |
22:36.08 | cr2 | it's the same pattern all over this megabyte. |
22:36.44 | husam | you the driver does build link list of buffer though |
22:37.39 | cr2 | maybe the driver is mapping some other area. |
22:37.56 | husam | maybe |
22:39.28 | cr2 | the driver is a monster. disassembling it is not fun. |
22:40.00 | husam | I did that with one on my device ... someone did that for me using ida ... |
22:40.10 | husam | but still not easy to follow |
22:41.13 | husam | Damm ... I been pressing all kind of key combination ... I can only get the one we know :( |
22:42.06 | husam | cr2: try an monitor Data control register .. when you send and receive packets .. maybe this whats been used |
22:45.40 | cr2 | but it is not that easy. i need to write its offset into +4 and then read +8 ? |
22:46.09 | husam | Ohh I see |
23:10.31 | *** join/#htc-blueangel pandreas21 (n=a@static-np1-203.cytanet.com.cy) |
23:11.21 | pandreas21 | hello:) |
23:12.10 | cr2 | hi |
23:12.14 | pandreas21 | trying to boot linux as per the instructions in http://www.handhelds.org/moin/moin.cgi/BlueAngel |
23:12.27 | cr2 | do you want to destroy your phone too ? :) |
23:12.32 | pandreas21 | hehehehe |
23:12.42 | cr2 | so. |
23:12.57 | pandreas21 | yesterday i "destroyed" my ipaq 2210 as well |
23:13.17 | cr2 | poor bill is crying on the way to the bank. |
23:13.30 | pandreas21 | have an mmc card 1gb |
23:13.36 | cr2 | should be ok. |
23:13.46 | pandreas21 | formatted as per the instructions 30Mb fat16 and rest ext3 |
23:13.56 | cr2 | ok. |
23:14.09 | pandreas21 | mounted fat partition to /mnt/sd0 |
23:14.15 | pandreas21 | and ext3 to /mnt/sd1 |
23:14.20 | cr2 | ok. |
23:14.50 | pandreas21 | created /mnt/sd0/linux and /mnt/sd0/2577 |
23:15.32 | pandreas21 | copied all files except the rootfs to linux and autorun.exe to 2577 |
23:15.57 | pandreas21 | so far everything according to the guide |
23:16.01 | pandreas21 | ? |
23:16.26 | cr2 | ok. |
23:17.08 | pandreas21 | got opie rootfs instead of gpe. is it ok? |
23:17.32 | cr2 | yes. |
23:17.44 | pandreas21 | untarred into /mnt/sd1 |
23:17.54 | pandreas21 | and synced |
23:18.03 | cr2 | will need some tweaks though. |
23:18.05 | cr2 | ok. |
23:18.25 | pandreas21 | unmounted and inserted the card in pda2k |
23:18.37 | pandreas21 | nice tux logo with thermometer |
23:18.40 | cr2 | ok. |
23:18.58 | pandreas21 | error reading initrd-2.6.12-hh2.gz |
23:19.09 | pandreas21 | when thermo almost full |
23:19.29 | cr2 | hmm. i have already heard something like that. |
23:19.47 | pandreas21 | yeah i saw it on an irc log in google |
23:20.32 | cr2 | try to format the fat16 within wince. |
23:21.00 | pandreas21 | ok, i am not sure about something.... |
23:21.13 | cr2 | another option is to compile the kernel with the sd/mmc driver and mount the linux partition directly. |
23:21.23 | pandreas21 | formatted first partition with fat16 <32M, maybe this could be the cause? |
23:21.45 | cr2 | i have a blueangle, but hacking on the universal right now. |
23:21.51 | cr2 | i think it does not matter. |
23:22.00 | cr2 | i have 10MB partition. |
23:22.37 | cr2 | you can even copy this stuff to the wince ramdisk. inconvenient after the reset, but should work. |
23:22.52 | pandreas21 | aha! |
23:23.36 | pandreas21 | and the universal is? |
23:24.16 | cr2 | i use the same opie, and mount it directly. |
23:24.26 | cr2 | without the ramdisk. |
23:24.56 | pandreas21 | any docs on that or i should just google it? |
23:25.25 | *** join/#htc-blueangel husam (n=kvirc@217.206.111.105) |
23:26.06 | pandreas21 | oh jarjar |
23:26.08 | cr2 | on what ? how to compile the kernel ? |
23:27.01 | cr2 | here. but it is not yet so functional as the blueangel. |
23:27.04 | cr2 | http://forum.xda-developers.com/viewtopic.php?t=35972 |
23:27.26 | *** join/#htc-blueangel immolo (i=lwhxpf@cpc3-cosh1-4-0-cust151.cos2.cable.ntl.com) |
23:27.47 | pandreas21 | is the phone working now for blueangel or? |
23:30.20 | cr2 | blueangel - yes, universal - no. |
23:31.41 | pandreas21 | another question and i am gone. how on earth do i format the card under ce? |
23:32.08 | cr2 | format it as ext2, then wince will suggest you to "format" it. |
23:32.36 | cr2 | but with the fat16 partition flag, so it will see it. |
23:32.43 | pandreas21 | hahahaahahaha |
23:32.54 | pandreas21 | thnx cr2 |
23:33.40 | cr2 | universal is fun, after i will finish it, i will do a user-friendly install for the blueangel. |
23:34.14 | pandreas21 | great |
23:34.18 | pandreas21 | bye now |
23:40.59 | cr2 | husam: |
23:41.02 | cr2 | WLAN: io addr 0x24=0x01030505 |
23:41.21 | husam | Nice |
23:41.32 | cr2 | WLAN: mem addr 0x0=0x00000000 |
23:41.42 | cr2 | WLAN: mem addr 0x4=0x00010004 |
23:41.59 | cr2 | WLAN: mem addr 0x8=0x00020008 |
23:42.02 | cr2 | and so on. |
23:42.17 | cr2 | i have adapted the hx4700_wlan debug code. |
23:42.38 | cr2 | have a look, you know it better :) |
23:42.53 | husam | Was that the oofset from the beginig of the memory area |
23:43.04 | cr2 | ACX_SLV_MEM_ADDR_LO |
23:43.13 | cr2 | ACX_SLV_MEM_ADDR_HI] |
23:43.36 | cr2 | <PROTECTED> |
23:43.39 | cr2 | <PROTECTED> |
23:43.40 | cr2 | <PROTECTED> |
23:43.40 | cr2 | <PROTECTED> |
23:43.40 | cr2 | <PROTECTED> |
23:43.41 | cr2 | <PROTECTED> |
23:43.41 | cr2 | <PROTECTED> |
23:44.11 | cr2 | the driver is from the hh.org CVS. |
23:44.27 | husam | is so they did but it there |
23:44.54 | husam | I think I had look around last time it wasn't there |
23:45.13 | cr2 | looking for a link. |
23:46.32 | cr2 | http://handhelds.org/cgi-bin/cvsweb.cgi/linux/kernel26/arch/arm/mach-pxa/hx4700/hx4700_wlan.c?rev=1.8&content-type=text/x-cvsweb-markup |
23:47.09 | cr2 | hmm. this is after the soft reset. |
23:47.24 | cr2 | not on the running wince config. |
23:48.14 | husam | I though these register could have different values depend where you reading or writing |
23:49.09 | cr2 | well, i did what i could, the power init/down/reset code is ok, the chip is detected. |
23:49.10 | husam | I think this the device which have very limited number of control registers to access |
23:51.36 | cr2 | i think eeprom & friends could be dumped too from the control area. |
23:52.03 | cr2 | i will move to the phone/modem port now :) |
23:52.03 | husam | yes and the acx driver has the interface |
23:52.12 | husam | :) |
23:52.28 | husam | I'm trying to get y-boot to work with h6300 |
23:52.45 | cr2 | blueangel can make calls, and i can't. |
23:53.12 | husam | hmm!!! .. do you have builtin phone |
23:54.58 | cr2 | i have umts. on the second cpu. |
23:55.01 | cr2 | http://wiki.xda-developers.com/index.php?pagename=HTC_Universal |
23:56.16 | husam | OK |