00:10.05 | psokolovsky | cr2, yep, we don't |
01:05.22 | *** join/#htc-linux Hawk||- (n=Hawk@p57A52DC9.dip0.t-ipconnect.de) |
05:52.13 | *** join/#htc-linux tudenbart (n=willi@xdsl-81-173-170-39.netcologne.de) |
08:09.10 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
08:54.13 | *** join/#htc-linux iggy (n=iggy@64.190.183.194) |
08:55.48 | *** join/#htc-linux RoEn_PC (n=roen@p54A653B4.dip.t-dialin.net) |
09:46.42 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
09:54.28 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
10:33.34 | *** join/#htc-linux pH5 (n=ph5@e178228142.adsl.alicedsl.de) |
10:34.55 | goxboxlive | good morning everyone |
10:38.22 | pH5 | good morning |
10:51.48 | *** join/#htc-linux LunohoD_ (n=alex@e180127073.adsl.alicedsl.de) |
10:59.54 | *** join/#htc-linux RoEn_PC (n=roen@p54A653B4.dip.t-dialin.net) |
11:28.49 | goxboxlive | cr2: Are you around? |
11:28.57 | goxboxlive | I need some info about A780 |
11:29.21 | goxboxlive | Anyone else with A780 around? |
11:33.44 | *** join/#htc-linux Fossi (n=Fossi@e176105121.adsl.alicedsl.de) |
13:42.56 | *** join/#htc-linux psokolovsky (n=psokolov@236.usernat.ip.net.ua) |
13:44.41 | goxboxlive | psokolovsky: Do you own a Motorola A780? |
13:44.58 | psokolovsky | Hi, nope |
13:45.10 | goxboxlive | DO you know anything about it? |
13:49.33 | psokolovsky | nope ;-) |
14:16.30 | *** join/#htc-linux psokolovsky (n=psokolov@236.usernat.ip.net.ua) |
14:38.11 | *** join/#htc-linux renatostore (n=renato4@72-181-119-85.dyn4.digitaleitalia.it) |
14:39.00 | renatostore | hi all |
14:40.35 | goxboxlive | hi renatostore |
14:41.45 | renatostore | i'd like to go ahead with alpine development |
14:42.02 | renatostore | someone can help? |
14:43.20 | goxboxlive | There are someone who already has worked on it, here is some info : http://wiki.xda-developers.com/index.php?pagename=AlpineResearch |
14:44.17 | goxboxlive | It also exist in the current hh.org kernel. I dont know if it compiles and run. |
14:45.33 | renatostore | starting from that point ... go ahead |
14:45.57 | renatostore | i tested kernel |
14:46.04 | renatostore | touchscreen works |
14:46.14 | renatostore | also usb interface |
14:46.14 | goxboxlive | good |
14:46.45 | renatostore | we are using an universal image |
14:46.52 | goxboxlive | good |
14:46.53 | renatostore | with ad-hoc kernel |
14:47.02 | goxboxlive | i made some new few days ago |
14:47.20 | renatostore | but i see a lot of error when i open a console |
14:47.27 | goxboxlive | i c |
14:47.42 | renatostore | ? |
14:48.40 | goxboxlive | i c = I see :-) |
14:50.32 | renatostore | maybe is for we use an universasl image |
14:53.04 | goxboxlive | Do you remove the htcuniversal modules before booting? |
14:53.29 | goxboxlive | What is the error messages? |
14:55.51 | renatostore | i compiled alpine kernel and modules and imstalled them on sd card |
14:57.01 | goxboxlive | ok, and before install you detelte the htcuniversal stuff right? And also delete the contens of htcuniversal in /etc/modutils/ , right? |
14:57.15 | goxboxlive | s/detelte/delete/ |
14:57.16 | renatostore | no |
14:57.36 | goxboxlive | I think you shuold also try that, just to get rid of all htcuniversal stuff. |
14:58.04 | goxboxlive | Because there are some htcuniversal conf files in /etc/modutils |
14:58.24 | goxboxlive | liek snd-htcuniversal-audio......... htcuniversal_acx |
14:59.36 | renatostore | so what a have to do before installing modules |
15:00.39 | goxboxlive | As i did whgen i used the blueangel image: |
15:01.05 | goxboxlive | 'rm -rf /media/usbdisk-2/lib/modules/*' |
15:01.55 | goxboxlive | And |
15:02.57 | goxboxlive | delete htcuniversal_acx and snd-htcuniversal_audio in /etc/modutils |
15:03.27 | goxboxlive | and maybe also the rest of the sound modules if they causes any problems |
15:03.53 | goxboxlive | What is the error messages saying during boot? |
15:06.32 | goxboxlive | Thats as far as i know the only htcuniversal specific that exist. Maybe there are something with bluetooth and IRDA, but i mean that they are set as machine specific. |
15:07.29 | goxboxlive | And of course you should delete the included keymap-2.6 in the /etc folder |
15:08.21 | renatostore | i have problem with usb: when i plug it i lost x11 and appear a console with some messages |
15:09.11 | renatostore | but i can see the device by desktop |
15:11.27 | goxboxlive | You should first try out the bootstrap-image. Just to see if everyting works. (no gui)http://gnulinux.biz/files//Universal/sd-mmc/snapshot/Bootstrap-image/Angstrom-bootstrap-image-test-20070125-htcuniversal.rootfs.tar.bz2 |
15:12.06 | goxboxlive | You can try out ts with running ts_calibrate or ts_test or something. It should be included. |
15:12.34 | goxboxlive | I have to go now, be back later. |
15:13.40 | renatostore | now i have this error: |
15:13.55 | renatostore | ext2 get inode |
15:14.15 | renatostore | unable to read inode block |
15:14.44 | renatostore | inode=98314, block=155654 |
15:14.53 | renatostore | numer changing |
15:15.16 | renatostore | this occurs without usb now |
15:25.05 | *** join/#htc-linux Tjikkun (n=tjikkun@ip5451133d.direct-adsl.nl) |
15:56.53 | cr2 | renatostore: can you run 'fsck' on this filesystem ? |
15:59.06 | *** join/#htc-linux crashmatrix (n=quakepro@s5590785f.adsl.wanadoo.nl) |
16:14.08 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
16:17.06 | goxboxlive | cr2: are you there? |
16:17.40 | goxboxlive | ~change 1000 NOK to euro |
16:17.55 | goxboxlive | ~change 1000 DKK to euro |
16:20.05 | crashmatrix | Is there a not pseudo-random function in D? |
16:20.31 | crashmatrix | sorry, rephrase : is there a real random function in D |
16:21.30 | cr2 | goxboxlive: change ? |
16:21.37 | goxboxlive | cr2: Good |
16:21.41 | cr2 | crashmatrix: what is D ? |
16:21.51 | crashmatrix | oh crap |
16:21.52 | goxboxlive | cr2: You own a A780 |
16:21.55 | crashmatrix | sorry, wrong channel |
16:21.58 | cr2 | ~change 1000 EUR to SEK |
16:22.24 | cr2 | hm. instresting. |
16:22.26 | goxboxlive | I have one i can buy now for 1000DKK |
16:22.34 | cr2 | with gps ? |
16:22.38 | goxboxlive | I dont know |
16:22.49 | cr2 | ask. |
16:23.21 | cr2 | not all a780 have gps. only the european model. |
16:23.23 | goxboxlive | But, isnet there a GPS in everyone of them? At least if it is bought here in europe right |
16:23.24 | *** join/#htc-linux RoEn_PC (n=roen@p54A653B4.dip.t-dialin.net) |
16:23.28 | goxboxlive | right |
16:23.51 | cr2 | ask if the copilot is installed. |
16:24.04 | goxboxlive | But if it is the Asian one, there arent any gps chip in it? |
16:24.09 | cr2 | the gps module is missing. |
16:24.25 | cr2 | that's why they are cheaper. |
16:24.34 | goxboxlive | B But if the module is missing, could i install it? |
16:24.36 | cr2 | no module, no software, no maps. |
16:24.45 | cr2 | where will you get it ? |
16:25.09 | goxboxlive | On a ebay site, scandinavian actuion |
16:25.09 | cr2 | then you will need a warez'd copilot. |
16:25.14 | goxboxlive | right |
16:25.19 | cr2 | i don't know how to do it :) |
16:25.38 | goxboxlive | But if i install OpenMoko on it will i get the GPS funtionality on then? |
16:25.49 | cr2 | unlikely. |
16:25.54 | goxboxlive | ok |
16:26.08 | cr2 | the radio rom uploads the gps firmware, afaik. |
16:26.17 | cr2 | openmoko ? |
16:26.21 | goxboxlive | But the chip is there, only think that misses is the module |
16:26.32 | cr2 | you mean this weird software from GL ? |
16:26.38 | goxboxlive | Yes, Stefan Smith will make one |
16:26.39 | cr2 | it is useless for you. |
16:26.47 | goxboxlive | hoe |
16:26.49 | goxboxlive | how |
16:26.50 | cr2 | which chip ? |
16:26.53 | goxboxlive | gps |
16:27.01 | cr2 | hmm. |
16:27.12 | cr2 | i'll look for a link for you. |
16:27.38 | goxboxlive | thx |
16:30.25 | cr2 | http://wiki.openezx.org/GPS |
16:31.58 | *** join/#htc-linux Zule (n=Zule@24-155-71-190.dyn.grandenetworks.net) |
16:33.38 | goxboxlive | yes i have read it. It says the same as you. Only the EU version is suplied with gps chip. |
16:35.02 | cr2 | yes, so you need the MG4100-based module, _and_ the copilot. |
16:35.43 | cr2 | there is some crypto involved, probably it is tied to the DoC G3 id. |
16:36.17 | cr2 | i still need to learn how to dump the complete flash contents. |
16:36.45 | cr2 | btw, keep in mind that you don't need the full EU map license. |
16:36.53 | cr2 | 1 region is ok. |
16:37.47 | goxboxlive | ok |
16:37.49 | cr2 | i have 2 cards for germany and scandinavia, and use then separately. |
16:37.57 | goxboxlive | right |
16:38.21 | cr2 | it's only a minor annoyance to switch the cards on the DK border, and remove 1 file. |
16:38.26 | *** join/#htc-linux Niacin (n=Niacin@cpe-76-187-29-166.tx.res.rr.com) |
16:38.27 | goxboxlive | I have asked the seller now about to let him check if copilot is installed. |
16:38.34 | cr2 | ok. |
16:39.44 | goxboxlive | I just want a reliable phone. Me and thoose who tries to call me is sick and tired of the Universal. When they call me they allmost everytime get the answermachine. The phone is totaly crap. I need a new one. |
16:40.12 | cr2 | :) |
16:40.24 | goxboxlive | I could also just bought a SonyEricsson or a Nokia but what's fun about that. |
16:40.27 | cr2 | btw, a780 has 3X SAR value. |
16:40.35 | goxboxlive | ANd what is that |
16:41.14 | cr2 | emitted microwave power directed at your brain :) |
16:41.28 | cr2 | 3 times more than the universal. |
16:43.49 | goxboxlive | IS that healthy? |
16:44.26 | renatostore | cr2: do i need to run fsck on root fs (it's mounted)? |
16:45.06 | goxboxlive | They have done some reasearch abuot phones now and they are saying that it isent as dangeroues as they thought it where. |
16:45.31 | cr2 | goxboxlive: it's under the limit. at least in germany. |
16:45.39 | goxboxlive | ok |
16:45.43 | *** join/#htc-linux shido6 (n=shido6@d221-68-200.commercial.cgocable.net) |
16:45.55 | cr2 | renatostore: fsck in the usb cardreader. |
16:48.05 | cr2 | goxboxlive: directed powerful EMI is unnatural, so there is actually always a concern. like with the UV and higher energy radiation. |
16:49.22 | goxboxlive | right |
16:49.34 | parmaster | one of the best SAR's is the blueangel |
16:50.06 | parmaster | goxboxlive: thats what they always say.. its a lobby |
16:50.31 | parmaster | the huge corps who sell them are always defending themselves that way |
16:50.41 | goxboxlive | right |
16:51.17 | parmaster | its skin penetrating EMI regardless |
16:54.19 | cr2 | i have found a guy with a himalaya and broken backlight. |
16:56.19 | cr2 | it seems that the gps will be hacked on a780 soon. |
16:56.23 | cr2 | http://lists.openezx.org/pipermail/openezx-devel/2007-January/001003.html |
17:00.59 | goxboxlive | cr2: So the conclusion is cr2: If it is the Asian version it will have the GPS chip included, but not the the software/driver for it. And when the gps is hacked the Aisan version will also have gps support. I'm i right? |
17:03.15 | cr2 | no. |
17:03.32 | cr2 | asian version does not have the hardware and the software. |
17:03.55 | cr2 | european version has hardware and software. |
17:04.29 | cr2 | if you'll find the hardware (by some black magic) you will need the software. |
17:04.58 | cr2 | the software (complete FW) is available on sf.net (!) |
17:05.26 | cr2 | but i don't know if it will work. |
17:08.41 | cr2 | don't be fooled by the project name :) |
17:08.44 | cr2 | http://sourceforge.net/project/showfiles.php?group_id=149849 |
17:10.51 | goxboxlive | ok, i'l wait then until he has answered me about it has copilot installed or not. Thaks for the info. |
17:14.46 | Kevin2 | Hi |
17:16.00 | Kevin2 | It's been quiet here lately. |
17:20.48 | psokolovsky | Kevin2, Hi! |
17:21.06 | psokolovsky | Kevin2, so, how do we decide regarding matching ambiguity? |
17:23.41 | Kevin2 | What's the issue? Why don't we just make the strings more specific? |
17:23.58 | psokolovsky | Kevin2, see email |
17:24.15 | psokolovsky | "X5" vs "X50" |
17:25.23 | psokolovsky | Kevin2, is apache battery handling code in CVS? |
17:25.24 | Kevin2 | Can we just put "X50" first? Another option would be "X5\0" - but that is ugly. |
17:25.50 | psokolovsky | Kevin2, how have we control such stuff? second option won't work. |
17:29.08 | Kevin2 | psokolovsky: No, it doesn't look like my last apache patch ever went in. |
17:29.28 | psokolovsky | Kevin2, did you post it to k-d? |
17:30.33 | Kevin2 | Yes - dated 27th of December. |
17:31.35 | psokolovsky | Kevin2, yeah, I see. sorry about that. will commit soon. |
17:34.32 | Kevin2 | I originally intended detect to be for "architectures" instead of specific machines. I'm not against passing more info to detect(), but I would think simply moving the x50 to before x5 should be sufficient. I don't think this type of thing will come up very often. |
17:35.33 | psokolovsky | Kevin2, moving before where? do intend to depend on linking order or something? that will bite our ass sooner or later |
17:36.15 | psokolovsky | Kevin2, if we want to depend on order, we'd rather add explicit priority and sort on that before traversing machines |
17:37.12 | Kevin2 | See the Makefile comments. The link order is well defined and it has a specified impact on the exe. |
17:37.47 | Kevin2 | The order is already used -- for example, we need to check for pxa27x before checking for pxa2xx. |
17:39.42 | psokolovsky | Kevin2, hm, ok |
17:40.37 | *** join/#htc-linux cyrill62 (n=cyrill62@65.247.146.195.dynamic.adsl.abo.nordnet.fr) |
17:42.03 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
17:43.05 | psokolovsky | Kevin2, what about putting ads7877 to drivers/misc/adc/ ? |
17:44.02 | Kevin2 | I'm okay with that - I'd even prefer it - but I didn't want to hold up the apache support if others had objections or comments. |
17:45.13 | psokolovsky | Kevin2, ok, I'll put it there |
17:45.56 | Kevin2 | It would also mean moving adc.h. It would also suggest moving touchscreen-adc to some common directory also. |
17:46.21 | psokolovsky | Kevin2, ok, will look into that |
18:28.39 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
18:31.19 | psokolovsky | Kevin2, how good is your touchscreen-adc.c anyway? |
18:31.36 | *** join/#htc-linux Niacin_ (n=Niacin@cpe-76-187-29-166.tx.res.rr.com) |
18:32.01 | Kevin2 | psokolovsky: Works well for me. |
18:32.22 | psokolovsky | Kevin2, what the heck is x_plate_ohms? ;-) |
18:32.42 | Kevin2 | psokolovsky: BTW, I thought of another way to fix the aximx50 stuff - I sent an email. |
18:32.43 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
18:33.31 | Kevin2 | psokolovsky: It's from the ADC chip documentation. In order to know how much pressure the pen has, one should know the resistance of the screen.. |
18:34.12 | psokolovsky | Kevin2, what is recommended "don't care" value? ;-) |
18:35.22 | Kevin2 | .x_plate_ohms = 400, // XXX - don't know real value. |
18:35.33 | Kevin2 | I don't know what it is under apache, so I just made up a number. :-) |
18:35.34 | psokolovsky | I see ;-) |
18:36.37 | psokolovsky | Kevin2, I just wonder, if it really needs so many params. for exmple, irq and .ispendown() somehow complement each other... |
18:37.25 | psokolovsky | Kevin2, and btw, I don't see you doing any debouncing? |
18:37.35 | Kevin2 | Irq could be passed through the resources area. |
18:38.06 | Kevin2 | I guess we could limit it to pxa gpios, in which case one could remove ispendown. |
18:38.28 | Kevin2 | We could default x_plate_ohms and min_pressure. |
18:38.49 | psokolovsky | Kevin2, well, doesn't ad7877 just reports 0's in case pen is up? ads7846 behaves like that |
18:39.20 | psokolovsky | Kevin2, but most problem is debouncing. don't you need it? or do you do it in adc driver? |
18:39.36 | Kevin2 | psokolovsky: No, it is an ADC sense, so there could be some residual current sensed. |
18:40.16 | Kevin2 | psokolovsky: I don't need it on ad7877 - I have a kind of debounce, in that we check for min_pressure. |
18:41.02 | psokolovsky | Kevin2, hm, I wonder if that's really enough... |
18:41.28 | Kevin2 | The ad7877 does multiple senses internally and then reports the averaged value. Maybe that is why I don't need it. |
18:41.40 | Kevin2 | Can you point to a driver that does debounce so I can see what you mean? |
18:42.41 | psokolovsky | Kevin2, drivers/input/touchscreen/ads7846ts_ssp.c - just yesterday generalized h1910/h4000's ads7846 driver |
18:47.58 | *** join/#htc-linux tudenbart (n=willi@xdsl-87-78-46-229.netcologne.de) |
18:51.04 | Kevin2 | psokolovsky: Hrmm. You don't use the z2 position.. You're using a different algorithm. |
18:51.54 | Kevin2 | The touchscreen is stable on apache - I've played with it enough and have never seen a mis-sense. I don't know what it would be like on other pdas. |
18:52.21 | psokolovsky | Kevin2, yep. it's actually not mine algorithm, it was written by Pawel Kolodziejski, h1910 mainatiner. I doubt it bothers to sense pressure at all. |
18:52.55 | psokolovsky | Kevin2, but that driver also works very well for ads7846 |
18:53.19 | psokolovsky | Kevin2, well, we need to keep generalizing ADC/TS framework... |
18:54.34 | Kevin2 | psokolovsky: See http://www.analog.com/UploadedFiles/Data_Sheets/59289336AD7877_a.pdf page 16 |
18:55.08 | Kevin2 | It details both algorithms. |
19:04.34 | psokolovsky | Kevin2, well, ads7846 is of course the same stuff underlying, but uses another terms and I don't know if they talk about the same stuff there |
19:04.45 | psokolovsky | anyway: "First Conversion Delay (Control Register 2 Bits <5:4> )" - that's where magic appears to be |
19:04.50 | Kevin2 | psokolovsky: I find ads7846ts_ssp.c to be confusing - it has too many loops and sets a lot of "flags". I'm curious if touchscreen-adc would also work on those devices. |
19:05.03 | psokolovsky | Kevin2, btw, do you know what reference voltage is used on apache? |
19:05.43 | Kevin2 | I'm also doing a sort of "debounce" by checking for the pendown line after the sense to see if it is still asserted. |
19:05.50 | psokolovsky | Kevin2, I bet, it won't. but yes, ads7846ts_ssp.c has a bit confusing algo. I even wanted to rewrite it, but why, if it works perfectly ;-) |
19:06.11 | psokolovsky | Kevin2, that line itself will bounce ;-) |
19:08.33 | Kevin2 | I'm not sure what you mean by reference voltage -- I don't think it is needed in the ts measurement. |
19:09.28 | psokolovsky | Kevin2, yep, not for TS, but for battery |
19:10.26 | Kevin2 | Oh, I have a bunch of logs from wince - the numbers in htcapache-power make them match what wince reports. |
19:11.26 | Kevin2 | psokolovsky: You should be able to simulate the FCD by adding a small delay from when the first pen irq fires to when the first sense is taken. |
19:12.13 | psokolovsky | Kevin2, I see. But did you try to calc what ref voltage would be from them? do you think it makes sense at all? would ADC sense actual voltages, so some divided voltage anyway? |
19:13.16 | Kevin2 | BattVoltage = si->pins[0].value * 1106 / 1024; |
19:13.27 | psokolovsky | Kevin2, yep, I read an appnote which suggested following algo: throw away N1 samples, then average N2 next. As you see, ads7846ts* instead does bounding, and that works well too ;-) |
19:15.25 | psokolovsky | damn, where to put all those headers for thsi stuff? ;-) |
19:17.14 | Kevin2 | psokolovsky: The ad7877 has an internal / external voltage reference option. Wince (and hence my code) set it to internal. |
19:17.34 | psokolovsky | Kevin2, internal?! 2.5v? |
19:17.42 | psokolovsky | h4000 exactly has external |
19:18.11 | psokolovsky | and with internal, battery voltage value saturates |
19:18.39 | psokolovsky | and with external, 4095 is almost exactly full charge |
19:18.41 | Kevin2 | Yes. Maybe they have a resistor on the battery sense line. |
19:19.38 | psokolovsky | Kevin2, I just wonder where they take ref. voltage from then? ;-) |
19:19.48 | psokolovsky | (in my case, obviously ;-) ) |
19:25.33 | *** join/#htc-linux shido6 (n=shido6@d221-68-200.commercial.cgocable.net) |
19:29.29 | psokolovsky | Kevin2, htcapache-ad7877.c should be killed, right? |
19:32.30 | psokolovsky | Kevin2, and btw, how many pins apache's batt has? |
19:33.58 | *** join/#htc-linux rob_w (n=bob@p85.212.139.152.tisdip.tiscali.de) |
19:35.38 | Kevin2 | psokolovsky: Yes, once ad7877 is committed, the htcapache-ad7877 is no longer needed. |
19:35.45 | Kevin2 | Physical pins on the battery? |
19:39.14 | psokolovsky | Kevin2, yes. h4000 has 5. |
19:42.55 | Kevin2 | psokolovsky: Six |
19:43.57 | Kevin2 | Hrmm. Could be seven - I see what could be a ground separate from the other pins - though it might just be a physical battery detect. |
19:44.08 | psokolovsky | Kevin2, ok. so, do you think there's a chance that that tempretaure is really battery's, i.e sensed from diod in battery? |
19:47.16 | Kevin2 | All I know on temperature is that wince reports a value "BatteryTemperature" that comes directly from AD7877_SEQ_BAT1. |
19:47.40 | psokolovsky | ok |
19:48.12 | Kevin2 | The value changes as the phone charges, so it certainly could be a temperature. |
20:11.59 | *** join/#htc-linux renatostore (n=renato4@72-181-119-85.dyn4.digitaleitalia.it) |
20:13.07 | renatostore | cr2: hi, now universal image is working in alpine, i'm very impressioned |
20:13.35 | renatostore | i think that 2 thinks last |
20:13.59 | renatostore | phone and battery |
20:14.29 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
20:16.08 | renatostore | ph5: hi, can we close our testing to make phone works? |
20:22.10 | pH5 | hi renatostore, what phone? the magician phone works with gomunicator, but I want to wait for the openmoko stack before investing any time in application software. |
20:23.05 | renatostore | Alpine |
20:24.11 | pH5 | what parts are missing? do you already have the functions to reset / turn off the gsm chip like in magician_phone.c? |
20:25.07 | renatostore | last time we test by haret a found gpios for phone |
20:26.19 | renatostore | 10 36 108 |
20:27.04 | pH5 | renatostore: so try to copy magician_phone.c and replace the gpios with the ones you found out. the alpine has the Ti GPRS chipset, too, right? |
20:27.40 | pH5 | 10 and 108 are inputs, but you need to find the gsm reset and gsm power gpios |
20:28.34 | renatostore | how |
20:28.36 | renatostore | ? |
20:28.57 | pH5 | gsm power is easy to find with haret, it is turned off in flight mode |
20:29.55 | pH5 | about the gsm reset one, you could find it by disassembling rilgsm.dll or with a haret memory trace on the gpio registers when turning on the phone |
20:30.50 | renatostore | what is 36? |
20:31.28 | pH5 | I need to clean up the magician phone driver, we can use it for both blueangel and alpine too. |
20:31.36 | renatostore | could it be power? |
20:31.37 | pH5 | gpio 36 is the ffuart data carrier detect pin |
20:33.29 | pH5 | 35-45 are ffuart/btuart gpios. on magician the gsm power and reset gpios are 26 and 86 |
20:37.13 | renatostore | dump gpio could reveal power and reset pins? |
20:37.44 | *** join/#htc-linux cyrill62 (n=cyrill62@ble59-5-82-233-205-36.fbx.proxad.net) |
20:38.18 | pH5 | renatostore: the power pin yes, the reset pin no (it is only set for one clock cycle and then cleared again) |
21:33.01 | renatostore | ph5: dump gpio chages: 10, 31, 36, 42-45, 58-60, 69, 108, 72, 76-77 |
21:41.25 | renatostore | ph5: are you there? |
21:43.10 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
21:44.54 | pH5 | renatostore: 58-77 are lcd related, so maybe 31 is gsm power? but only if it is configured as output, no alternate function. |
21:47.05 | renatostore | so how can we be sured? |
21:48.25 | pH5 | is gpio 31 configured as output, is it set to an alternate function? dump gpio tells. |
21:48.33 | pH5 | also it would be certainly helpful to disassemble rilgsm.dll. |
21:49.18 | renatostore | i don't know where to read if it is output? |
21:49.59 | renatostore | 31 changes form o010 to o110 |
21:51.00 | renatostore | i think that it is output |
21:53.12 | renatostore | how to disassembre rilgsm.dll |
21:53.15 | renatostore | ? |
21:53.20 | pH5 | the second number (A column) is the alternate mode, so gpio 31 is used as SYNC_I2S pin |
21:53.43 | pH5 | we are looking for output pins with alternate mode set to 0. |
21:55.24 | pH5 | to disassemble rilgsm.dll you need objdump with arm support |
21:55.33 | pH5 | there is one in http://www.videolan.org/pub/testing/wince/wince-arm-gcc4.0.0-cvs-linux-crosscompiler-20050517.tar.bz2 for example |
21:56.09 | pH5 | do you have a dump of the alpine rom? |
21:57.45 | renatostore | too difficult for me |
21:58.32 | renatostore | i have a cross compiler by which i compiled kernel |
21:58.39 | renatostore | for alpine |
21:59.12 | pH5 | if you can point me to a alpine rom dump I could try to extract rilgsm.dll and disassemble it. |
22:01.02 | renatostore | you can find one here: ftp://ftp.xda-developers.com/Alpine/ |
22:01.36 | renatostore | maybe i can mail you the dll |
22:01.48 | renatostore | form my alpine |
22:04.22 | pH5 | that'd be cool, but I don't think you can just copy files out of the rom with wince. |
22:08.37 | renatostore | it's true: can't access file |
22:09.56 | renatostore | so 31 is not power |
22:09.59 | renatostore | ? |
22:10.00 | pH5 | hm, I just got ftp.xda-developers.com/Alpine/Extracted%20Rom/ms_.nbf, now to remember how to extract files from that.. |
22:10.20 | pH5 | right, 31 cannot be power, I guess it is used for sound. what sound chip does the alpine have? |
22:16.26 | renatostore | i don't know |
22:17.03 | renatostore | i had a picture of main boad but i can't find it now |
22:18.50 | renatostore | but ... if you look for pins with O in D column and 0 in A column that change passing in flight mode ... dump gpio can't help you |
22:19.18 | renatostore | none of the above list mach the case |
22:19.41 | *** join/#htc-linux cyrill62 (n=cyrill62@ble59-5-82-233-205-36.fbx.proxad.net) |
22:21.19 | pH5 | renatostore: hm, that is strange indeed |
22:22.12 | renatostore | i can post output of dump gpio if you want |
22:23.25 | cr2 | pH5: i think it was uda1380 |
22:23.38 | pH5 | hm, except for 10 (is this D=O,A=0?) all the other gpios you mentioned are known to be used otherwise. but it can't hurt, right. |
22:23.54 | pH5 | hey cr2, thanks. |
22:23.58 | cr2 | hi |
22:24.49 | renatostore | 10 is D=I A=0 |
22:25.33 | renatostore | but change in D=O A=0 |
22:26.24 | pH5 | renatostore: it is D=0 when the phone is turned off, right? it is set to output and cleared to save energy. |
22:27.12 | pH5 | gpio 10 is used by the GSM chip to wake up the PXA. |
22:27.54 | pH5 | ah ok, extracted rilgsm.dll |
22:32.10 | renatostore | are you going to change kernek source in cvs? |
22:32.58 | pH5 | renatostore: what do you mean? |
22:33.51 | pH5 | ok, gsm reset is gpio 86, just like on the magician |
22:34.40 | cr2 | experience :) |
22:34.46 | renatostore | with all pins maybe you can change magician_phone.c |
22:35.37 | renatostore | or better create alpine_phone.c |
22:36.05 | renatostore | but this is a module, isn't it? |
22:38.38 | pH5 | yes, it is a module. we should rename/move it and have a common tisomething_phone driver. |
22:40.03 | renatostore | what else do we need |
22:40.04 | renatostore | ? |
22:41.05 | cr2 | pH5: ba/hima too ? |
22:41.19 | cr2 | i hope to get hima soon. |
22:41.57 | pH5 | cr2: cool. ba/hima most probably. I should look at their gpio tables again. |
22:42.08 | pH5 | I wonder if the neo1973 has the same gsm chipset |
22:43.03 | cr2 | ti something afaik. |
22:43.09 | pH5 | renatostore: now I need to find the alpine's gsm power gpio. I would have expected that it is 26, but that one didn't change in the dumps. |
22:43.48 | renatostore | no |
22:43.48 | pH5 | cr2: right, ti something, like the magician. I wish I had some documentation about those chips. |
22:44.00 | cr2 | ;) |
22:44.07 | pH5 | yeah, who wouldn't :) |
22:44.10 | cr2 | they will not give it out. |
22:44.37 | pH5 | Or I need someone with enough spare time to rev-engineer the radio rom dumps :) |
22:44.37 | cr2 | afaik it is not possible to update the radio rom on the fic phone. |
22:45.03 | cr2 | i have tried in on universal. |
22:45.19 | cr2 | but it's a real bloatware written in c++. |
22:45.38 | cr2 | hmm. different chipset. |
22:45.48 | pH5 | universal has qualcomm, right? |
22:46.11 | cr2 | yes. |
22:48.16 | renatostore | ph5: i put a list of changes. It should be there. how to find it? |
22:55.47 | pH5 | I guess GSM power isn't a cpu gpio at all, but is controlled by the cpld |
22:57.38 | pH5 | renatostore: can you have a look at the cpld registers at phys 0x0c000000 (virt 0xa9500000) with the phone turned on and off? |
22:58.33 | renatostore | First this is what change in dump gpio: |
22:58.36 | renatostore | <PROTECTED> |
22:58.36 | renatostore | <PROTECTED> |
22:58.36 | renatostore | <PROTECTED> |
22:58.36 | renatostore | <PROTECTED> |
22:58.36 | renatostore | <PROTECTED> |
22:58.37 | renatostore | <PROTECTED> |
22:58.39 | renatostore | <PROTECTED> |
22:58.41 | renatostore | <PROTECTED> |
22:58.43 | renatostore | <PROTECTED> |
22:58.45 | renatostore | <PROTECTED> |
22:58.47 | renatostore | <PROTECTED> |
22:58.49 | renatostore | 108 I 0 0 0O 0 0 0 |
22:58.51 | renatostore | <PROTECTED> |
22:58.53 | renatostore | <PROTECTED> |
22:58.55 | renatostore | <PROTECTED> |
22:59.14 | pH5 | yeah, you see all the outputs are set to some alternate function. that's not it. |
23:01.18 | *** join/#htc-linux rob__w (n=bob@p85.212.136.221.tisdip.tiscali.de) |
23:04.20 | renatostore | what does it mean? |
23:05.53 | pH5 | it means that the gsm power is not controlled by the pxa itself but (hopefully) by the xilinx cpld. |
23:05.55 | *** join/#htc-linux cyrill62 (n=cyrill62@ble59-5-82-233-205-36.fbx.proxad.net) |
23:08.06 | *** join/#htc-linux meredydd (n=meredydd@caiusnat1.acad.cai.cam.ac.uk) |
23:15.09 | renatostore | it's a task for tomorrow ... bye |
23:16.51 | pH5 | hm ok. just for the record, I think the cpld egpio nr. 19 might be gsm power (bit 0x10 at 0xa9500002) |
23:16.51 | pH5 | good night |
23:16.57 | *** part/#htc-linux pH5 (n=ph5@e178228142.adsl.alicedsl.de) |
23:20.15 | meredydd | Okay, the website seems a little unclear - does the camera work at all |
23:20.16 | meredydd | ? |
23:21.16 | cr2 | on which phone ? |
23:22.05 | meredydd | Oh, sorry. BlueAngel. |
23:22.21 | meredydd | (from what I gather, the only one with a working radio) |
23:24.43 | cr2 | radio works on many phones. |
23:25.15 | cr2 | the camera on ba does not work, because the ati capture interface is not known. |
23:25.32 | cr2 | i.e. needs some disassembling. |
23:26.08 | cr2 | atividcap.dll calls into the main dll. |
23:29.14 | meredydd | Ah, interesting. |
23:29.20 | meredydd | Sorry - didn't quite get that part. |
23:29.50 | meredydd | I'm looking to throw together a "concept phone" for a company I work for |
23:30.11 | meredydd | and just thought that the best place to start would probably be something like this I can get Linux working on. |
23:31.45 | meredydd | So, is there any easy way to tell which phones have working radio-and-sound support? |
23:31.53 | meredydd | (cause frankly, the rest is just window-dressing) |
23:33.04 | cr2 | magician |
23:33.08 | cr2 | universal |
23:33.35 | cr2 | these are 2 sound chipsets: uda1380 and ak4641 |
23:37.58 | meredydd | Does the Universal radio work these days? Gah - I hate out-of-date webpages |
23:38.45 | cr2 | it works. more or less :) |
23:39.15 | meredydd | "More or less"? |
23:39.17 | cr2 | just no sensible gui |
23:39.43 | meredydd | Oh, bugger, no GPRS. Grr - deal-killer. |
23:39.45 | cr2 | and no sane mixer interface. |
23:40.11 | cr2 | i can't test gprs/umts. not data contract. |
23:40.21 | cr2 | s/not/no/ |
23:41.41 | cr2 | universal. |
23:43.28 | meredydd | Ah, sweet. |