00:28.42 | *** join/#htc-linux don102 (n=don@p54BAE325.dip.t-dialin.net) |
00:28.56 | don102 | hi |
00:28.59 | don102 | i have a big problem |
00:29.10 | don102 | with my htc ... |
00:29.46 | don102 | i upgraded a ROM version.. and it boots up with a white screen... |
00:31.56 | don102 | but now i have a big problem |
00:32.03 | don102 | i can't come into the bootloader |
00:32.07 | don102 | what can I do... |
00:32.32 | don102 | this are the facts. i have no possibility to boot into windows... |
00:33.54 | don102 | and i have no chance to get into the bootloader with "Camera + On + softreset" |
00:33.56 | don102 | what can i do? |
00:34.18 | don102 | I really need help :-/ |
00:56.43 | *** join/#htc-linux thinko (i=jdoe6alp@smaug.rackdragon.com) |
02:20.25 | *** join/#htc-linux eway (n=ellis@homealone.good1.com) |
04:00.31 | *** join/#htc-linux Zoolooc (n=fredsiba@84.149.41.29) |
05:22.32 | *** part/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
05:32.10 | *** join/#htc-linux goxboxlive (n=jrs@195.159.97.196) |
05:32.26 | goxboxlive | good morning |
06:33.28 | *** join/#htc-linux _law_ (n=law@mail.stiftadmont.at) |
07:22.38 | *** join/#htc-linux kyppcgeek (n=PocketIR@156.sub-75-205-66.myvzw.com) |
07:28.46 | kyppcgeek | ? |
07:29.51 | *** part/#htc-linux kyppcgeek (n=PocketIR@156.sub-75-205-66.myvzw.com) |
07:42.35 | *** join/#htc-linux kyppcgeek (n=PocketIR@156.sub-75-205-66.myvzw.com) |
07:42.55 | kyppcgeek | anyone here? |
07:53.12 | *** join/#htc-linux dcordes (n=user0815@f048228117.adsl.alicedsl.de) |
07:54.00 | dcordes | moin |
07:56.54 | *** join/#htc-linux sadeness_ (n=LamersIn@nat1.rlan.ru) |
07:58.53 | dcordes | can somebody give me a hand extracting the bootloader out of htc devices? I have several romdumps of htc kaiser but have no idea how to get a bootloader out |
08:07.57 | dcordes | they are in *.raw and consist of 3 files eache |
08:43.14 | *** join/#htc-linux Pryan2 (i=ircap8@84.125.95.179.dyn.user.ono.com) |
08:55.56 | dcordes | hi Pryan2 6 |
08:58.14 | *** join/#htc-linux rmoravcik (n=moravcik@gtsgw.ttc.cz) |
08:59.18 | *** join/#htc-linux rmoravcik (n=moravcik@gtsgw.ttc.cz) |
09:20.10 | *** join/#htc-linux BabelO (n=fcr@81.80.236.75) |
09:25.44 | BabelO | hi |
09:29.31 | BabelO | goxboxlive: around ? :) |
09:52.20 | dcordes | hi BabelO |
09:52.56 | dcordes | can you help me with kaiser bootloader? |
09:53.34 | BabelO | dcordes: for what ? |
09:55.24 | dcordes | BabelO: i need to dump it and dont know how |
09:55.36 | dcordes | i have 3 *.raw files |
09:58.39 | dcordes | BabelO: i sarched the xda wiki/forums for solutions but only found 'rom coking things' |
10:00.57 | dcordes | BabelO: any idea? |
10:06.49 | BabelO | dcordes: i never remember, boot is 1Mb or 2Mb after something, don't know if the first Mb after screen buffer is boot or Wince mapped kernel file |
10:07.24 | dcordes | BabelO: ok but how do i get that stuff out of the raw file? |
10:12.35 | BabelO | dcordes: with objdump i think, you can disassemble by hand |
10:12.57 | BabelO | you need to found the start |
10:14.34 | dcordes | BabelO: i just received email from fellow kaiseruser. he has the location of the bootloader so i can dump it easily |
10:22.19 | BabelO | ok |
10:27.00 | dcordes | BabelO: it is at 0x00001000 can that be? |
10:29.10 | BabelO | dcordes: don't know |
10:29.23 | dcordes | ok will just try when home |
10:33.37 | goxboxlive | BabelO: hi, yes i'm around. I have just been at the Liquor Store for bying liquor to a christmas party at work tomorrow. Up here in Norway we are not allowed to buy wine or other alcohol drinks in a regular stor. Only beer can be buoght in a regular store. |
10:36.01 | BabelO | goxboxlive: ok :) http://linuxbox2.free.fr/imlife/home.png and http://linuxbox2.free.fr/imlife/weather.png |
10:36.52 | BabelO | goxboxlive: i do that quickly :) i ve to finish coding all, after i ask someone for graphics help ... lol |
10:40.14 | goxboxlive | BabelO: i think it looks cool. BTW, what is this ORange F ? Is it on your Greenphone? |
10:40.56 | sadeness | but the sun is a bit oval :D |
10:41.39 | BabelO | sadeness: yes, i know, i do that quicly, i want to have an homescreen like it is on wince6 with htc plugin ;) |
10:42.01 | BabelO | goxboxlive: yes it is on gp, my provider |
10:44.21 | goxboxlive | BabelO: I havent got the provider working on universal. |
10:44.38 | goxboxlive | BabelO: Have you looked further on the BA modem and Qtopia? |
10:45.14 | BabelO | goxboxlive: not at the moment |
10:45.38 | goxboxlive | ok |
10:47.23 | BabelO | goxboxlive: i like the idea to have a nice homescreen on qtopia ;) |
10:47.33 | BabelO | need to do better graphics |
10:49.00 | goxboxlive | BabelO: i agree, but i think first of all we need to have the modem working. Why do we want a nice homescreen when it's not usable if you see what i mean. Has ljp responded you about the modem yet? |
10:49.55 | goxboxlive | I have compiled all the mux plugins now in the source and also tried editing them, but no luck. It stil segfaukts |
10:50.28 | goxboxlive | I just wish there where a dirty trick we could do to just have it working like universal does. |
10:56.20 | BabelO | goxboxlive: yes agree with you, need to look more on it |
10:58.19 | *** join/#htc-linux rob_w|laptop (n=rob_w@gw.suellberg-hamburg.de) |
11:04.58 | *** join/#htc-linux lieter_ (n=lieter@a62-251-100-248.adsl.xs4all.nl) |
11:24.35 | goxboxlive | good :-) |
11:24.49 | *** join/#htc-linux wireddd (n=wired@66-188-77-4.dhcp.athn.ga.charter.com) |
12:02.57 | *** join/#htc-linux moby_wan_kenoby (n=kvirc@80.168.34.147) |
12:56.48 | *** join/#htc-linux goxboxlive (n=goxboxli@81.80-202-132.nextgentel.com) |
14:55.56 | dcordes | BabelO: hey to get back to kaiser spl: It is at 0x00001000 - which haret command downloads it? |
14:57.23 | BabelO | dcordes: it is dump command like pd or pf to dump to file |
14:57.35 | BabelO | but this is physic adress |
14:57.50 | dcordes | BabelO: oh- so I need logic? |
14:58.40 | BabelO | so something like pd p2v(0x00001000) 0xfff for example |
15:04.19 | dcordes | BabelO: dude that made my device reboot |
15:04.34 | dcordes | is this a bad thing? |
15:05.03 | dcordes | BabelO: I put "pd p2v(0x00001000) 0xfff" and it reset |
15:37.23 | BabelO | dcordes: yes wrong idea p2v is for physic -> virtual and pd use physic adress |
15:57.10 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
15:58.32 | *** join/#htc-linux Stamrogh_ (n=Stamrogh@c83-254-131-136.bredband.comhem.se) |
15:59.26 | *** join/#htc-linux GPFerror (n=gpferror@76.187.41.132) |
16:41.31 | *** join/#htc-linux Mineque (n=Mineque@195.60.65.50) |
17:37.59 | *** join/#htc-linux pH5 (n=ph5@e178212102.adsl.alicedsl.de) |
18:11.07 | *** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
18:24.27 | *** join/#htc-linux Zoolooc (n=fredsiba@p54953FA2.dip0.t-ipconnect.de) |
18:32.09 | Pryan | hi |
18:35.07 | dcordes | Zoolooc: hi. Did you try getting the kais bootloader? |
18:41.50 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
18:42.02 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
18:45.55 | Pryan | reboot in linux |
18:51.16 | Zoolooc | hello |
18:51.27 | Zoolooc | dcordes: no, unfortunately no time at all |
18:51.40 | *** join/#htc-linux dimoniet (n=mal@81.203.51.20.dyn.user.ono.com) |
18:54.01 | *** join/#htc-linux Pryan (n=pryan@84.125.95.72.dyn.user.ono.com) |
18:58.28 | dcordes | Zoolooc: ambition gone? |
18:58.30 | *** join/#htc-linux goxboxlive (n=goxboxli@81.80-202-132.nextgentel.com) |
18:58.44 | dcordes | I don't see us getting anywhere if nobody contributes :( |
19:01.19 | dcordes | pH5: http://git.android.com/?p=linux-msm.git could you take a look at the last but one entry? Might this be of revealance for kaiser so I can try a rebuild? |
19:02.18 | *** join/#htc-linux TimRiker (n=timr@216.49.181.128) |
19:04.29 | pH5 | dcordes: you can rebuild all you want. but it won't help you a bit until haret is fixed :) |
19:05.44 | dcordes | I didn't understand it was a haret problem |
19:06.28 | dcordes | pH5: will it be possible at all for Kevin2 to fix w/o machine? |
19:06.59 | pH5 | dcordes: as long as it hangs in haret, it doesn't even matter if you try to boot a proper zImage or a mp3 file. the problem is that (to my understanding) haret hangs where it tries to disable the interrupts. |
19:11.04 | dcordes | pH5: ah of course... I'm already getting confused on the whole thing. Did you notice Kevin2's attempts working around the problem? http://irclog.iclem.net/?chan=htc-linux&day=1 see from 17:58 if you have a minute spare time |
19:20.27 | pH5 | ah, now I understand why the msm_fb driver needs this strange updater thread |
19:21.22 | *** join/#htc-linux dcordes (n=dcordes_@unaffiliated/dcordes) |
19:22.51 | dcordes | pH5, sorry term died. I didn't know there was an msm_fb. I thought the fb would be device(display) specific. |
19:23.18 | pH5 | dcordes: if you look at the sources, there is a mddi_client_toshiba.c file |
19:23.36 | dcordes | let me see |
19:23.42 | pH5 | all the gpio/voltage parts are of course display specific |
19:24.06 | pH5 | but the core (msm_fb, mdp, etc.) responsible for frame buffer and dma transfer is universal. |
19:25.33 | dcordes | pH5, in how far can we make use of this? |
19:27.10 | pH5 | dcordes: no idea. i don't even know now if haret hangs or actually jumps into the kernel |
19:27.43 | pH5 | because the haret boot messages obviously can't work if the frame buffer -> mddi dma is software-driven and we just have turned off the interrupt controller. |
19:28.03 | dcordes | that's a thing I'm also wondering. I thought about possible led or backlight drivers from other devices to see if it jumps into the kernel or not |
19:28.18 | pH5 | dcordes: do you have some easy feedback mechanism like a vibrator that is controlled by a simple gpio pin? |
19:28.29 | pH5 | dcordes: or that :) |
19:29.02 | dcordes | Don't ask me how these are controlled..... |
19:29.29 | pH5 | dcordes: does watch gpios work on msm already? |
19:29.45 | dcordes | it has front keypad backlight, display backlight, keyboard backlight, vibrator |
19:29.50 | pH5 | you have to go one step back and map the hardware events. |
19:30.05 | pH5 | turn on/off vibrator, lights, etc. and see which gpios (if any) change. |
19:30.41 | dcordes | ok I think vibrator is best place to start because you can make it have an event without big hassle on the device |
19:31.23 | dcordes | pH5, I don't know what's watch gpios form top of the head, only know cr2_ and me researched some touchscreen location by me touching the screen and tracing stuff |
19:31.34 | dcordes | that will be the thing? |
19:32.27 | pH5 | dcordes: yes |
19:32.36 | dcordes | pH5, connection open |
19:36.04 | dcordes | pH5, how did I initiate the recording again? |
19:36.16 | pH5 | watch gpios? |
19:36.27 | dcordes | yes so I can see where vibra is |
19:36.47 | pH5 | watch gpios <timeout in seconds> |
19:37.52 | dcordes | 5 |
19:38.17 | dcordes | damnation |
19:38.22 | dcordes | it spams the whole screen in one scond |
19:38.50 | dcordes | tricky |
19:40.21 | pH5 | dcordes: use ibits to mask out those spamming gpios |
19:40.29 | pH5 | ibit gpios <number> |
19:40.33 | dcordes | cool |
19:40.51 | dcordes | number==077.400 ? |
19:41.32 | pH5 | no, the gpio number. |
19:41.41 | dcordes | 92b00840: |
19:41.45 | pH5 | (or numbers, you probably have multiple gpios switching all the time) |
19:42.02 | dcordes | yes like 5 lines per second |
19:43.38 | dcordes | pH5, don't be angry with me but what part of the line is the gpio number now ^^ ? |
19:44.32 | dcordes | pH5, in this example, what's th gpio#? 077.864 GPIOS 92a00c20: ?(203)=800 |
19:45.25 | pH5 | first number is the time, the second is the register address, I think. 203 should be the gpio number |
19:45.42 | pH5 | try "ibit gpios 203" and look if this line still appears |
19:45.43 | dcordes | great |
19:45.53 | dcordes | i issued 203 and 290 because those are the spammers |
19:46.01 | dcordes | aha! perfect! |
19:46.13 | pH5 | if you have many gpios to disable you can even to "ibit gpios 203 290" etc. |
19:46.27 | dcordes | pH5, fsck, seems like vibrator is one of the spammers |
19:46.48 | pH5 | dcordes: not necessarily. maybe the vibrator is not on a cpu gpio but connected to something else :( |
19:46.50 | dcordes | I don't see anything when I make it vibrrate |
19:47.10 | pH5 | the cpld perhaps, if there is any (and as this is a htc device, there probably is) |
19:47.13 | dcordes | pH5, a cpu gpio means it is directly controlled by the cpu? |
19:47.26 | pH5 | dcordes: yes, directly connected to one of the cpu pins |
19:47.29 | dcordes | pH5, hwat is cpld? |
19:47.30 | dcordes | ok |
19:47.52 | pH5 | ~wiki cpld |
19:47.57 | dcordes | dude how can they not connect the vibrator to the cpu. They knew I would use it as indicator for msm kernel booting or not ^^ |
19:48.20 | pH5 | it is a programmable chip. htc uses them quite often to add more gpios and irqs to the systems. |
19:48.31 | dcordes | to access mor hardware |
19:48.32 | pH5 | dcordes: lol :) |
19:49.07 | dcordes | pH5, how do I want to find out cpld existence/address? |
19:50.18 | pH5 | dcordes: KaiserMemoryMap says there is a CPLD at phys 0xa9800000 |
19:50.35 | pH5 | you can try to watch that region with the mmutrace mechanism |
19:51.19 | dcordes | do I have to restart to get rid of ibit gpios filters or is that irrelevant? |
19:51.26 | pH5 | irrelevant |
19:51.42 | dcordes | mmutrace add?? |
19:52.35 | pH5 | addlist mmutrace p2v(0xa9800000) 0x1000 |
19:52.42 | pH5 | wi 1 |
19:52.59 | pH5 | maybe you have to ignore some spamming irqs with "ibit irqs <number>", similar to gpios. |
19:53.28 | dcordes | http://pastebin.ca/raw/807529 |
19:53.35 | dcordes | no looks sleek |
19:54.46 | pH5 | that was with the vibrator switching? anything if you add irq 16 and 19 to ibits? |
19:55.45 | dcordes | pH5, <dcordes> http://pastebin.ca/raw/807529 not vibrator, it is the thing you just told me to do, cpld |
19:55.58 | dcordes | did I get your question twisted |
19:56.21 | pH5 | dcordes: what I'm looking for is the gpio that the vibrator is connected to. |
19:56.32 | dcordes | pH5, now your question reached my userspace. |
19:57.02 | pH5 | if you don't have something strange like a cpld connected over i2c, there should be a register *somewhere* in your memory space where a single bit switches with the vibrator being turned on or off. |
19:57.23 | pH5 | we know this register is not one of the cpu gpio registers (or otherwise we would have seen something with watch gpios) |
19:57.25 | dcordes | aye how do I find out? |
19:57.39 | dcordes | ok |
19:57.49 | pH5 | so the next idea was to watch the cpld's memory space (that I got from the MemoryMap, no idea if it is correct) while triggering the vibrator |
19:58.23 | pH5 | for that you should mask out the irqs that appear in that log (16 and 19) and do a "wi 10" again, and trigger the vibrator somehow. |
19:58.38 | dcordes | aaah ok ^^ |
19:58.51 | dcordes | HaRET(7)# bit irqs 16 19 |
19:58.54 | pH5 | if there is no signal either, that'd mean that either the cpld is not mapped at that location (a98..) or that the vibrator isn't connected to the cpld :) |
19:59.04 | dcordes | let's see first |
19:59.27 | dcordes | 001.734 IRQS 94e00000: ?(6)=40 |
19:59.27 | dcordes | 001.979 IRQS 94e00000: ?(6)=40 |
19:59.29 | dcordes | cha ching |
19:59.38 | dcordes | it appeared definetly on vibrator brrrr |
20:00.15 | pH5 | dcordes: nah, not good. we want to see something change at a9400000(and following) |
20:00.22 | dcordes | :-( |
20:00.32 | pH5 | the IRQs are only input signals. vibrator control is definitively an output |
20:00.51 | dcordes | pH5, http://pastebin.ca/807535 |
20:01.36 | dcordes | pH5, those two lines definetly appear when I hit the vibrator switch and it brrrs, not hwen I do normal touch |
20:01.39 | dcordes | anywhere |
20:02.21 | pH5 | interesting |
20:02.57 | dcordes | pH5, so does this get us anywhere? |
20:03.39 | pH5 | dcordes: no. can you do *anything* during wi 10 to make some memory accesses to a94... appear? (turn on/off lights, change backlight brightness, switch bluetooth, whatever) |
20:04.12 | dcordes | pH5, so we can point out different potential CPLD candidates? |
20:04.43 | pH5 | dcordes: to see that mmutrace actually works on the msm and/or that there is actually something mapped at a94... |
20:04.52 | dcordes | pH5, do you know a fucntion in windows mobile (or even haret) to turn offl ight? |
20:04.56 | dcordes | display backligt |
20:05.35 | pH5 | dcordes: no idea. on magician I could switch the backlight by holding the power button for a long time, but I guess that doesn't make sense on the kaiser, which doesn't have a transflective lcd. |
20:05.48 | dcordes | pH5, same here |
20:05.53 | dcordes | what's a transflective lcd? |
20:05.59 | pH5 | oh, or so it does? |
20:06.07 | pH5 | a lcd that you can still read when the backlight is turned off |
20:06.17 | dcordes | I hold down the "power" button on the side and display is off, keys locked |
20:06.26 | dcordes | lol didn't observe this on kaiser |
20:06.32 | dcordes | but wait let me try |
20:06.49 | dcordes | sorry I don't hold it down, I just press it short |
20:06.57 | dcordes | holding it down makes kaiser want to go shutdown |
20:07.17 | pH5 | yeh, I don't think you can read the display without the backlight being turned on on the kaiser. |
20:08.20 | cr2_ | hi |
20:08.31 | pH5 | hi cr2_ |
20:08.46 | cr2_ | does haret watch for the gpio changes on msm7x00 ? |
20:08.49 | dcordes | pH5, so I simply wi something and see what happens on baclight |
20:09.21 | pH5 | cr2_: it looks like it, Kevin2 added some addresses in the gpio register space to the GPIOS list for msm |
20:09.56 | dcordes | pH5, within my current haret session, I cannot see any IRQ reactions on display turn on/off |
20:10.01 | dcordes | on wi 20 |
20:10.10 | dcordes | turned it like 5x on/off |
20:10.13 | dcordes | nothing |
20:10.24 | dcordes | it only prints something afterwards |
20:11.11 | cr2_ | ah, there are no backlight control commands in haret. |
20:11.20 | dcordes | pH5, cr2_ sounds like I should upgrade my haret version? |
20:11.48 | dcordes | I have the last msm revision (#4) Kevin2 made on 1.12. during vsfb attempts |
20:18.53 | dcordes | pH5, still around? |
20:23.14 | *** join/#htc-linux Stamrogh (n=Stamrogh@c83-254-132-174.bredband.comhem.se) |
20:26.30 | dcordes | :-( |
20:26.39 | pH5 | dcordes: occasionally. looks like this doesn't get us any further, either. |
20:26.59 | dcordes | how come? |
20:27.29 | dcordes | you mean the whole CPLD thing we just did or are you refering to my haret upgrade question? |
20:27.31 | pH5 | dcordes: me being around occasionally or this not getting us any further? |
20:27.39 | pH5 | ah :) |
20:28.09 | pH5 | dcordes: well, until now we didn't see any access to the alleged CPLD registers yet. |
20:28.12 | *** join/#htc-linux Stamrogh__ (n=Stamrogh@c83-254-132-174.bredband.comhem.se) |
20:28.47 | dcordes | you take this form inactivity on wi and display backlight activity? |
20:28.55 | pH5 | maybe Kevin has some more insights into memory tracing, or you could try to give the other unknown memory regions from the MemoryMap the same treatment. |
20:29.40 | pH5 | dcordes: the recipe is always the same. watch some memory region with mmutrace, fiddle with some hardware, and as soon as you see something change in wi you know that this hardware is related to that memory region. |
20:32.28 | dcordes | pH5, what's the "aim of the exercise"? find hradware that can be abused as boot indicator? |
20:32.36 | pH5 | yes. |
20:33.00 | dcordes | ok |
20:33.13 | pH5 | leds or vibrator or backlight, anything that blinks or buzzes. |
20:33.24 | dcordes | I thought vibrator didn't work for the latter? |
20:33.50 | pH5 | dcordes: wrong. you just didn't find out how to control it yet. |
20:34.02 | dcordes | oh ok. So how would I go on about the vibrator? |
20:34.09 | dcordes | and what wer the two lines? |
20:34.18 | dcordes | which appeared after it buzzed? |
20:34.43 | pH5 | irq 6 |
20:35.03 | pH5 | you could try to find out if irq 6 has some significance from in the linux-msm sources. |
20:35.18 | dcordes | lol |
20:35.40 | dcordes | you know I'm a DAU in such things |
20:35.58 | pH5 | dcordes: you have to start somewhere :) |
20:36.03 | dcordes | true |
20:36.30 | dcordes | so how would I start finding places that refer to irq 6 in linux-msm source? |
20:36.41 | dcordes | hihi |
20:37.12 | pH5 | true, all this weird stuff takes a lot of time to understand. I'd start to look in irq.c or the headers that it references. |
20:37.23 | pH5 | maybe this is no use, but who knows.. |
20:37.59 | *** join/#htc-linux Stamrogh___ (n=Stamrogh@c83-254-132-174.bredband.comhem.se) |
20:38.33 | dcordes | I definetly need somebody of your engagement in kaiser lobby |
20:41.02 | dcordes | pH5, I will try my luck with backlight now |
20:42.33 | dcordes | pH5, kaiser has a light sensor |
20:42.54 | dcordes | it knows when it is in a dark enviornment and turns on keyboard backlight on keybaord strok in such situation |
20:45.43 | dcordes | pH5, what do you think, keyboard backlight hooked up to CPLD/GPIOS? |
20:49.35 | pH5 | dcordes: probably. unless the msm has pwm registers that are used for the backlight. |
20:49.40 | dcordes | keyboardstrokes show resonance on watch gpios |
20:50.00 | pH5 | dcordes: that is good always the same gpio number? |
20:50.03 | dcordes | touch screen doesn't which is good so I can easily switch the backlight on |
20:50.07 | dcordes | lemme see |
20:51.24 | dcordes | aaah wtf no |
20:51.34 | dcordes | I entered which made haret boot my lockupkernel |
20:51.50 | pH5 | whoops :) |
20:52.06 | dcordes | but I made some observations |
20:52.15 | dcordes | _most_ of the keys show same resonance in haret |
20:52.29 | dcordes | they give few lines (<10) with differing gpio numbers |
20:52.54 | dcordes | what's intersting is that all 4 arrow keyys show no resonance |
20:54.24 | dcordes | I will try to get a clean result by pressing all the keys one by one (in a text editor, lol) |
20:55.44 | pH5 | dcordes: I think those gpios are used for the internal matrix keyboard controller in the msm |
20:55.51 | pH5 | the arrow keys are connected to gpios on the cpld |
20:56.17 | pH5 | at least that is the way they did it on the herald (omap850) |
20:56.58 | pH5 | now you only need to find the memory region (with mmutrace) that changes when you press the arrow keys ;) |
20:57.49 | pH5 | but if you are unlucky, as on the herald, you have the cpld connected via i2c (which means it is not easy to control from early linux boot or during the last haret stages) |
20:58.11 | cr2_ | pH5: cpld is mmaped |
20:58.19 | pH5 | cr2_: for sure? |
20:58.23 | cr2_ | on kaiser |
20:58.30 | cr2_ | at 0x98000000 |
20:58.33 | pH5 | that's good. but then why don't we see any response in mmutrace? |
20:58.45 | pH5 | ah, wait. then I got the wrong address. |
20:58.48 | cr2_ | i think so, but it needs to be checked |
20:59.16 | cr2_ | kaiser and titan have different cpld mappings |
20:59.27 | pH5 | cr2_: KaiserMemoryMap says 0xa9800000 |
20:59.58 | dcordes | I'm here to try. please... |
21:00.27 | pH5 | dcordes: repeat the wi experiment with p2v(0x98000000) as address |
21:00.34 | dcordes | alright |
21:00.39 | pH5 | and change the buzzer, press the arrow keys etc. during wi |
21:00.43 | dcordes | brrr one? |
21:01.01 | pH5 | dcordes: everything you got ;) |
21:01.23 | dcordes | I didn't addlist mmutrace stuff before I made keybaord experiment |
21:01.52 | dcordes | only ibitted spamming gpios |
21:02.10 | pH5 | dcordes: that's ok, most of the keyboard keys are connected to gpios as you found. |
21:02.12 | dcordes | anyway I'm in a new session now |
21:02.14 | pH5 | but the arrow keys are not |
21:02.21 | pH5 | so for them we have to look at the cpld again |
21:02.37 | pH5 | (and not at a9800000, but 98000000, as cr2_ says) |
21:02.58 | dcordes | pH5, just to make sure we don't spend valuable time for nothing, where can I put keyboard stroke wi output from last experiment? |
21:03.55 | dcordes | I will just have it local.. |
21:04.56 | pH5 | dcordes: You should make a new KaiserGPIO wiki page, not immediately, but at least before your computer crashes the next time and you lose the locally stored information :) |
21:05.45 | cr2_ | pH5: hmm. maybe 0xa9800000 is correct. |
21:06.17 | dcordes | pH5, where I will list it e.g. like "keytroke T => gpio 23,456,345 etc? |
21:06.41 | pH5 | I'd say try both then. Unless somebody knows for sure that 0x98000000 is not a CS region. |
21:07.04 | dcordes | is there a way to easily figure out the wrong one? |
21:07.23 | dcordes | ah i will retry vibrr experiment as you said, will that clearify situation? |
21:08.08 | pH5 | dcordes: I wouldn't worry about the matrix assignments right now. just a list of gpio by number (like MagicianGPIO) and mark all of the keyboard related gpios. this list can be filled in as you learn more. |
21:08.38 | pH5 | dcordes: it can, but only if you get a result for one of the addresses.. |
21:09.25 | pH5 | cr2_: do you remember where you got that kaiser cpld information from? |
21:10.28 | dcordes | I need a small stand up bird that hits the listen button on kaiser. like the one homer uses to maintain the power plant remotely |
21:12.20 | dcordes | ok I have a dump file now where I pressed all buttons |
21:14.33 | cr2_ | pH5: SPL |
21:15.04 | cr2_ | pH5: i'll recheck it. |
21:21.41 | cr2_ | pH5: the main source of confusion is the different mapping used by SPL compared to wince |
21:33.28 | dcordes | pH5, cr2_ "addlist mmutrace p2v(0x98000000) 0x1000; ibit irqs 16 19; wi 10" ===> "002.587 IRQS 94e00000: ?(6)=40; 002.793 IRQS 94e00000: ?(6)=40" |
21:35.11 | dcordes | pH5, that's pretty much the same output I got for the same with "a" prefixed |
21:36.54 | cr2_ | dcordes: what does 'show gpios' say ? |
21:37.17 | cr2_ | wine is completely confused on my display. |
21:39.17 | dcordes | I need a new favourite pastebin |
21:40.00 | dcordes | cr2_, http://paste.ubuntu-nl.org/47157/ |
21:40.17 | dcordes | cr2_, what do you try to run on it? |
21:45.03 | cr2_ | dcordes: why are you tracing '0x9' and not '0xb' ? |
21:45.40 | dcordes | cr2_, <pH5> dcordes: repeat the wi experiment with p2v(0x98000000) as address |
21:45.53 | dcordes | <cr2_> at 0x98000000 |
21:46.20 | cr2_ | dcordes: can you change the 0x9* to 0xb* on gpio trace ? |
21:46.25 | dcordes | sure |
21:46.31 | dcordes | I will reopen |
21:46.37 | cr2_ | 0x98* was my mistake i guess |
21:46.44 | dcordes | nvm |
21:46.51 | dcordes | good for my training |
21:47.34 | dcordes | cr2_, 0xb98000000? I'm not sure on zero count |
21:48.08 | pH5 | dcordes: this is 32bit arch, so 0xb9800000, 8 hex digits. |
21:48.27 | pH5 | cr2_: 0xb9800000 is phys? |
21:48.29 | dcordes | aha valuable bit of info :) |
21:48.35 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
21:48.59 | cr2_ | yes, haret has "addlist gpios p2v(0xa9200800)\n" hardcoded |
21:49.50 | cr2_ | which translates into 0x92xxxxx |
21:50.05 | cr2_ | which is not the address used by wince. |
21:50.15 | cr2_ | wince uses 0xb2xxxxx |
21:51.08 | goxboxlive | cr2_: did u look at this? http://www.compulab.co.il/x270em/html/x270-em-datasheet.htm |
21:51.13 | cr2_ | so either haret/you includes all virtual mappings, or p2v() in addlist is just evil. |
21:51.51 | cr2_ | goxboxlive: yes. how much does it really cost ? i don't buy into "starts at $120" ;) |
21:52.14 | cr2_ | the bare motherboard - maybe. |
21:52.38 | goxboxlive | cr2_: thats only the base. The complete one without case $649 |
21:53.03 | goxboxlive | it's expensive |
21:53.09 | cr2_ | +customs tax |
21:53.17 | pH5 | goxboxlive: devboards tend to be |
21:53.26 | goxboxlive | ok |
21:53.35 | cr2_ | so i don't see any advantage. htc is commodity hardware. |
21:53.38 | goxboxlive | cr2_: i think that is included for german |
21:54.20 | goxboxlive | cr2_: but here the drivers are finished, and with a case you can use it as a main phone. We cant do that toda. |
21:54.23 | goxboxlive | toda |
21:55.00 | cr2_ | goxboxlive: from what i have read at the customs homepage, you are in hot water, if it's not person->person post. |
21:55.31 | cr2_ | i don't know how do they actually sell moko1 phones. it's a bit of mystery for me. |
21:55.56 | cr2_ | goxboxlive: is there a list of hardware drivers ? |
21:57.54 | *** join/#htc-linux dcordes (n=dcordes_@unaffiliated/dcordes) |
21:58.11 | dcordes | ~lart alicedsl |
21:58.11 | apt | eats alicedsl and falls over dead |
21:59.17 | dcordes | pH5, cr2_ http://paste.ubuntu-nl.org/47162/ |
21:59.30 | dcordes | kaiser locked up seconds after I issued wi |
22:00.53 | cr2_ | dcordes: you should replace the 0x9 to 0xb for each of the addresses here: http://paste.ubuntu-nl.org/47157 |
22:01.09 | cr2_ | as in |
22:01.20 | cr2_ | addlist gpios 0xb2b00800 |
22:01.31 | dcordes | in a new session? |
22:01.52 | dcordes | or with addlist mmutrace before |
22:02.11 | cr2_ | yes, add all these commands before running 'wi' |
22:05.40 | cr2_ | pH5: hehe, this x270 has an i2c gpio extender. |
22:06.50 | dcordes | exhausting |
22:07.18 | dcordes | cr2_, now I wi and hit vibrator? |
22:07.34 | dcordes | s/vibrator/vibration_alert |
22:08.01 | cr2_ | dcordes: yes. |
22:08.31 | cr2_ | you can put these commands into a script file, and run it at the beginning. |
22:08.58 | dcordes | I didn't use sed n'either a script |
22:09.12 | dcordes | raw keyboard work |
22:11.39 | cr2_ | ok. do you see something coming out of wi ? |
22:11.52 | dcordes | I'm playing around with it |
22:12.02 | cr2_ | imho, you should. otherwise it's one more haret bug. |
22:12.09 | dcordes | trying to figure what is the actual brrr and what is the event that happens when I switch the "mode" |
22:12.19 | dcordes | I'm using the SOUND menu of htc today screen |
22:12.36 | dcordes | when bbrrr mode is active, and I hit it again, there is output |
22:12.37 | cr2_ | ok, let's check if you see the gpio changes at all. |
22:12.57 | cr2_ | and you create "brrrr" with the 'nled' command |
22:13.13 | dcordes | what's nled? |
22:13.21 | cr2_ | run help |
22:13.32 | cr2_ | NotificationLED |
22:13.39 | dcordes | how can I run it when I have the wi active? |
22:14.07 | cr2_ | listen 9998; telnet kaiser 9998 |
22:14.35 | cr2_ | listen 9998 in haret |
22:15.03 | cr2_ | and telnet to that port. so you'll have a "command" link and a "trace" link for haretconsole. |
22:15.51 | dcordes | http://paste.ubuntu-nl.org/47170/ |
22:16.32 | cr2_ | ok, this is strange. |
22:16.53 | dcordes | some activity is only initiated by the raw button hit |
22:18.21 | dcordes | cr2_, check this http://paste.ubuntu-nl.org/47171/ |
22:22.11 | cr2_ | dcordes: there are irq changes only. |
22:22.21 | dcordes | which means? |
22:22.40 | cr2_ | something is wrong. |
22:23.34 | dcordes | that's bad |
22:24.07 | dcordes | cr2_, with the cpld address? |
22:24.28 | cr2_ | no, the msm7x00 gpio |
22:27.20 | dcordes | I just tried to check led/gpio |
22:27.58 | dcordes | disabling spammer gpios with ibit and turning on one led with nledset 0 0 |
22:29.11 | dcordes | cr2_, no output |
22:29.24 | dcordes | bbiab |
22:36.15 | cr2_ | TD035STEE1 in this em-270 |
22:37.19 | cr2_ | Kevin2: and they init it over pxa-ssp |
22:42.26 | dcordes | cr2_, If you want me to try something just poke |
22:46.28 | *** join/#htc-linux adamg2 (n=adamg@80-192-177-87.cable.ubr03.blac.blueyonder.co.uk) |
23:28.33 | dcordes | pH5, cr2_ thanks a lot for your help today |
23:28.35 | dcordes | bye |
23:35.27 | paulproteus | http://wiki.openmoko.org/wiki/One_transistor_too_many |
23:35.36 | paulproteus | HTC-Linux doesn't have *that* problem, at least. |
23:41.42 | Foxdie | I can't seem to get HTC linux working properly on my universal |
23:41.54 | Foxdie | I can get wifi enabled but not bluetooth or gsm / gprs |
23:43.15 | cr2_ | paulproteus: you should talk to pH5 about calypso suspend |
23:43.57 | cr2_ | Foxdie: bt/gsm should work if you've loaded the respective modules |
23:44.20 | cr2_ | Foxdie: i must admit that the current gui controls are bad. |
23:45.33 | Foxdie | I don't even know how to get that working |
23:45.47 | Foxdie | I know how to get a shell but dialler is broken |
23:46.06 | Foxdie | Seems you guys need a QC guy ;) |
23:46.09 | Foxdie | (Quality Control) |
23:46.21 | Foxdie | if you want I'll help with that |
23:48.30 | paulproteus | cr2_, Ouch re: calypso |
23:48.41 | paulproteus | Foxdie, Are you using goxboxlive's images? |
23:49.32 | cr2_ | Foxdie: do you run the latest wince5se ? |
23:50.33 | paulproteus | If BTW I buy a Universal that has WinCE 6, not 5, would that be a problem? |
23:50.56 | Foxdie | I'm using his images yes, currently Angstrom X11 |
23:51.08 | cr2_ | there are no wince6 phones afaik. |
23:51.10 | Foxdie | and I'm running WM6, but its 6 months old now |
23:51.46 | Foxdie | Universal will never be released with WM6 because its not financially viable for cell networks to release updates |
23:51.50 | paulproteus | Foxdie, The OpenMoko images? |
23:52.07 | Foxdie | I think so, yes, they were under the OpenMoko section |
23:52.08 | cr2_ | what is called wm6 is just a minor wince5 modification |
23:52.16 | paulproteus | Foxdie, Try the Qtopia image. |
23:52.20 | paulproteus | cr2_, Oh, I didn't realize. |
23:52.23 | Foxdie | cr2_: its only really the shell thats different |
23:52.32 | paulproteus | Foxdie @ http://linuxtogo.org/~htcpxa/htcuniversal/index.html |
23:52.34 | Foxdie | Qtopia... *shudders* |
23:52.45 | Foxdie | I remember having a Zaurus C860 with Qtopia |
23:52.49 | Foxdie | couldn't really do anything with it |
23:52.57 | paulproteus | Foxdie, Please, just try it for me. (-: |
23:53.12 | paulproteus | Check out the modification time. |
23:53.15 | paulproteus | It's way more recent. |
23:53.15 | cr2_ | zaurus had qtopia2. nonphone |
23:53.19 | Foxdie | ooh |
23:53.32 | cr2_ | this is qtopia4 for the phones |
23:53.32 | Foxdie | okay, I will try it tomorrow, will email myself the link at work and put it on there |
23:54.05 | Foxdie | done |
23:54.10 | Foxdie | and now I must go to bed, take care all |
23:54.18 | cr2_ | good night |
23:54.29 | Foxdie | CR2, I will catch you tomorrow and offer to test / write down bugs in rom releases if you'd like |
23:56.39 | paulproteus | Night, Foxdie. |
23:58.04 | cr2_ | Foxdie: it's goxboxlive who is creating the roms. |
23:58.16 | cr2_ | i can't recreate them anyway. |
23:58.36 | cr2_ | i need to create my own then, because it's not documented anywhere. |
23:58.46 | cr2_ | how to create a rom. |