05:26.04 | *** join/#htc-linux rob_w (n=bob@p213.54.224.218.tisdip.tiscali.de) |
05:55.05 | *** join/#htc-linux goxboxlive (n=goxboxli@ti500710a080-14919.bb.online.no) |
06:33.45 | *** join/#htc-linux goxboxlive (n=goxboxli@ti500710a080-14919.bb.online.no) |
07:17.11 | *** join/#htc-linux awelux__ (n=awelux__@dslb-084-058-165-238.pools.arcor-ip.net) |
07:18.17 | *** join/#htc-linux BabelOued (n=BabelOue@lun34-2-82-238-28-28.fbx.proxad.net) |
13:13.19 | *** join/#htc-linux goxboxlive (n=goxboxli@ti500710a080-14919.bb.online.no) |
13:30.18 | *** join/#htc-linux rob_w|mis (n=rob_w@p549BBACF.dip0.t-ipconnect.de) |
18:18.49 | *** join/#htc-linux apt (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
18:18.49 | *** topic/#htc-linux is HTC LINUX ! THIS IS THE FORWARD CHANNEL from #htc-blueangel ; check also -> http://www.handhelds.org/moin/moin.cgi/BlueAngel | http://wiki.xda-developers.com/index.php?pagename=BlueangelResearch | http://gnulinux.biz/files/blueangel/sd/linux/ LOGS: at http://ibot.rikers.org/%23htc-linux/ |
18:19.06 | Funklord | awelux__: how do you make a backup on the SD? |
18:19.34 | Funklord | can it be restored without access to a computer? |
18:21.37 | awelux__ | I can recommend Backup Manager from Sunnysoft. You can create self extracting backups. |
18:29.04 | *** join/#htc-linux mikearthur (i=mike@82-41-205-190.cable.ubr11.edin.blueyonder.co.uk) |
18:29.28 | mikearthur | Do any HTCs actually work with Linux and allow you to make outgoing calls? |
18:30.07 | mikearthur | (yet :D ) |
18:35.59 | awelux__ | hmm. I thought yes. |
18:36.31 | mikearthur | what models? |
18:38.41 | awelux__ | I think the blueangel was working and there was some progress with the universal. Check the progress of the blueangel |
18:39.31 | Funklord | awelux__: thanks for the tip |
18:39.53 | Funklord | awelux__: this really caused me a big headache |
18:40.07 | Funklord | I was stuck in london with no place to go |
18:40.21 | awelux__ | I was here when you were at the icafe |
18:40.54 | Funklord | hehe ok |
19:10.13 | *** join/#htc-linux rob_w (n=bob@213.54.222.47) |
19:25.13 | cr2 | hi awelux |
19:25.28 | cr2 | i think you are right, but there are two things: |
19:25.50 | cr2 | 1. we will drop this driver when 2.6.16 will work |
19:27.53 | cr2 | 2. the weighted linear combination done by the driver can't correct the nonlinear effects. |
19:32.02 | goxboxlive | cr2: I know the 2.6.16 doesnt boot, but is there a way i can get it? I wanne try a whole lot of combinations to try to boot it. |
19:44.50 | cr2 | good question. i will try to create a patch for the hh.org CVS. |
19:45.44 | cr2 | i have tried to switch off everything, but it still does not want to boot. |
19:49.44 | goxboxlive | cr2: Great. |
19:54.04 | cr2 | hmm. asic3_mmc and acx/mem changes. |
20:12.11 | awelux__ | cr2, did you switch off serial ttys ? |
20:12.40 | awelux__ | What is the status of the power management currently? |
20:12.46 | cr2 | where ? |
20:12.50 | cr2 | nonexistent. |
20:13.07 | awelux__ | when you try to boot 2.6.16 |
20:13.14 | cr2 | well. the lcd power management on universal works. |
20:13.24 | awelux__ | And, are you at the linuxtag on saturday per chance ? |
20:13.30 | cr2 | i can't switch off FFUART. |
20:13.33 | cr2 | no. |
20:13.50 | awelux__ | why can't you switch off ffuart ?` |
20:13.57 | cr2 | then i cant boot. |
20:14.06 | cr2 | even the 2.6.12 kernel |
20:14.30 | awelux__ | interesting. Currently I can't boot if I compile the pxa serial ports module at all. |
20:14.43 | awelux__ | probably wrong gpios |
20:14.44 | cr2 | i have looked at the suspend/resume code for h2200 |
20:15.02 | cr2 | compared that with universal |
20:15.14 | cr2 | and with the memory dump |
20:15.43 | cr2 | h2200 loads the htc bootloader as firmware and patches some addresses, so linux looks like wince :) |
20:16.11 | cr2 | the wince kernel is at 0xa0040000 |
20:16.29 | cr2 | probably on all htc devices. |
20:16.52 | cr2 | below i see the USB host buffers + evil ohci clock. |
20:17.14 | cr2 | and the remains of the bootloader, which the usb host have not overwritten. |
20:17.43 | cr2 | in the zero page there is some stuff needed for wakeup, it seems. |
20:17.48 | awelux__ | interesting idea |
20:18.26 | awelux__ | probably wince works very much like linux from the integrators point of view |
20:19.27 | cr2 | i looked in the logs, your ohci sturuct is higher than mine ? |
20:19.29 | awelux__ | Since they shouldn't have access to the source they will probably just be able to define which modules to include and have to provide the initial setup in the bootloader |
20:20.06 | awelux__ | hmm? what do you mean? |
20:20.13 | cr2 | normally the htc bootloader takes 256K |
20:20.21 | awelux__ | are you refering to an haret dump ? |
20:20.24 | cr2 | in pxa flash in your case |
20:20.29 | cr2 | or on hima/ba |
20:20.39 | cr2 | of in ram |
20:20.45 | cr2 | on universal |
20:20.49 | awelux__ | It looked the same on magician |
20:20.57 | cr2 | that's why wince is at +40000 |
20:21.12 | awelux__ | About 150k were used |
20:21.24 | cr2 | funny is that they put ohci buffers over the bootloader. |
20:22.35 | awelux__ | I'm not sure I understand. Are you refering to the stuff at around 0x32000 which seams to be usb related? |
20:22.55 | cr2 | i have it at 0x14000 |
20:22.59 | cr2 | <PROTECTED> |
20:23.01 | cr2 | ok. |
20:23.04 | cr2 | yes. |
20:23.35 | awelux__ | that is indeed a little strange. |
20:23.46 | cr2 | this is ohci struct, and you have the evil 16bit reg at +0x80 offset. |
20:24.09 | awelux__ | 16 bit reg? |
20:24.20 | cr2 | th ohci counter. |
20:24.25 | awelux__ | ah ic |
20:24.34 | awelux__ | which causes mem corruption |
20:24.44 | cr2 | the USB hosts writes there when it is on. |
20:24.46 | awelux__ | or dma access violations probably |
20:24.54 | cr2 | yes. even if the DMA is off. |
20:25.11 | awelux__ | because I've seen the process stop when writing to that address. |
20:25.12 | cr2 | it probably writes directly over the system bus. |
20:25.26 | cr2 | the LCD dma is also somewhat special. |
20:26.13 | cr2 | i think it is a good idea to check if this counter runs before booting. |
20:26.44 | cr2 | otherwise the kernel will be put over this counter and ... |
20:26.55 | awelux__ | no |
20:27.08 | cr2 | this problem is known since the jornada820 age ;) |
20:27.14 | awelux__ | If you use the zImage it will unpack and halt because of a crc msismatch |
20:27.27 | cr2 | yes. |
20:28.15 | awelux__ | The reason for the different memory locations of the usb struct could be that it's controlled by wince and wince doesn'T know a thing about it's bootloader |
20:29.58 | cr2 | but wince itself is always at 0x40000 (try to dump the page at this address) and it knows where to put the ohci mem. |
20:30.44 | awelux__ | why doest it know where to put the ohci mem? And why shouldn't it know to put it at 0x32000 or 0x14000 |
20:32.08 | awelux__ | cr2, about the TS. Because the current user mode driver can't correct the nonlinearities we will need a new one. |
20:34.08 | awelux__ | The new user mode driver should be able to read files with a detailed information about how to map the TS coordinates. Your universal will need a different file than my magician. |
20:35.28 | awelux__ | A very nice architecture should have an api to query a arbitrary number of calibration points. |
20:36.30 | awelux__ | The calibration result which will always work relative to the TS raw mapping file will be saved very similar to now. |
20:36.57 | cr2 | yes. i think we will reuse the ads7846 driver. |
20:37.17 | cr2 | this is a task for the tslib plugin. |
20:37.48 | awelux__ | I've to look at that one. But it's probably a good idea sicne the TI2046 seams to be similar to some ads chips. |
20:38.30 | awelux__ | Another advantage is that we can implement a very clean TTS kernel driver. |
20:38.49 | cr2 | the kernel: my idea is to forget about the lower 256K and decompress the zImage to 0xa0040000 |
20:39.11 | awelux__ | And a without much work a very ugly and very individual tslib so that at least the developers can use the TS. |
20:39.31 | cr2 | ok. |
20:39.57 | awelux__ | Writing a better user mode TSlib is easier than writing new kernel code so I hope this will happen sooner or later if the code is left alone ;) |
20:40.31 | cr2 | :) |
20:41.36 | awelux__ | I will study the ads7846 driver now. |
20:42.29 | cr2 | it relies on the generic ssp bus driver. |
20:42.44 | cr2 | s/ ssp / spi / |
20:43.07 | awelux__ | ok. a bit like the himalaya ts driver ? |
20:44.46 | cr2 | hmm. maybe. but in a more clean way. spi is now like i2c, it has api, and you don't need to do bitshifting stuff ;) |
20:45.55 | awelux__ | good. |
20:55.00 | BabelOued | hi |
20:59.09 | BabelOued | cr2: could you have a look at http://linuxbox2.free.fr/blueangel/htcblueangel_acx.c |
21:05.09 | cr2 | what is wlan_irq and wifi_irq ? |
21:06.29 | BabelOued | these two are input it's GPIO17 and 75 |
21:07.04 | BabelOued | it's a test when i start |
21:07.20 | BabelOued | acx100 is detected |
21:07.43 | BabelOued | but after give me a message eCPU is already running |
21:08.19 | BabelOued | i've tried with and wothout msleep, but no difference |
21:08.25 | cr2 | ok. then you didn't reset it. |
21:09.17 | BabelOued | ok maybe i have wrong gpio reset |
21:09.21 | cr2 | do you boot with wifi running ? |
21:09.23 | cr2 | maybe. |
21:09.53 | BabelOued | yes, i try with wifi running and without |
21:09.54 | BabelOued | but |
21:11.00 | BabelOued | if i don't set GPIO_RESET before power other, card isn't detected |
21:11.27 | BabelOued | and i'm not sure with pxa_gpio_mode |
21:36.12 | BabelOued | acxmem_l_reset_mac: enable soft reset... |
21:36.12 | BabelOued | acxmem_l_reset_mac: disable soft reset and go to init mode... |
21:36.12 | BabelOued | 00014541 <== acxmem_l_reset_mac |
21:36.12 | BabelOued | acx: eCPU is already running. reset_dev() FAILED |
21:36.13 | BabelOued | 00015042 <== acxmem_s_reset_dev: 00000001 |
21:36.13 | BabelOued | 00015042 <== acxmem_e_probe: FFFFFFFB |
21:46.10 | *** join/#htc-linux _cronos_ (n=cronos@ron34-1-82-224-149-102.fbx.proxad.net) |