00:00.28 | cr2 | it's for the LCD. |
00:01.05 | cr2 | no, it's for gpio. |
00:01.05 | cr2 | 0x5e gpio is the SD card present. a nice start :) |
00:04.10 | Zoolooc | "pd 0xA9100834 4" gives exception as well |
00:04.20 | Zoolooc | or I did not understand what offset to ad |
00:04.22 | Zoolooc | add |
00:05.43 | cr2 | a92 not a91 |
00:06.08 | cr2 | 0x33 - 0x38 gpios are sd related. |
00:06.38 | cr2 | 0x3e - 0x43 |
00:07.03 | Zoolooc | thought also the one where in that table, GPIO_CONF2 appears, is needed too |
00:07.28 | cr2 | yes. |
00:08.06 | cr2 | but the offsets are 410 , 414 and 504 for GPIO_CONF2 |
00:09.47 | Zoolooc | alright, just now I understood how to interpret that table |
00:11.07 | Zoolooc | so, you would need for each of those GPIO-related addresses to run that? ok, a script will be the right thing I guess |
00:12.12 | Zoolooc | MSM_GPIO1 has so many, but you refered only to 834. how do I know which ones are neede, or are they needed alll?= |
00:14.23 | cr2 | all would be best. and put the dump into some wiki page. like KaiserGPIO or somehting. |
00:14.47 | cr2 | i've identified some pins already, but it's only the beginning. |
00:16.58 | Zoolooc | ok, then I will put those gpio related addresses in some script |
00:17.47 | Zoolooc | what are you sseing actually, what does the response "a9200834 | 00000001 | ...." to "pd 0xA9200834 4" actually mean? |
00:20.26 | cr2 | the value at the physical address 0xa9200834 is 1 |
00:20.34 | cr2 | it's the standard hexdump format |
00:22.58 | Zoolooc | yes, but how to relate to what a 1 at that address means, and what a 0 means? you said something about SD card earlir |
00:23.18 | Zoolooc | does the SD status show that way at some location? |
00:23.36 | cr2 | these are the gpio bits. |
00:23.41 | Zoolooc | if yes, how to know which location is responsible for that in the first place? |
00:23.54 | Zoolooc | and which one for something else |
00:24.09 | cr2 | 0x53 bit detects the SD |
00:24.18 | cr2 | you can calculate in which bank it is. |
00:25.24 | Zoolooc | why is ther no 53 in the TitanMemoryMap? |
00:25.41 | Kevin2 | Hi |
00:25.46 | Zoolooc | Hi |
00:25.59 | cr2 | the zero bank has bits 0..15 |
00:26.06 | cr2 | hi Kevin2 |
00:26.38 | Kevin2 | Hey. I'm going to read through the irc logs. |
00:27.26 | cr2 | Kevin2: i've identified the USB setup on athena. bad luck, pxa27x_udc is not used. now i need the spi keyboard ;) |
00:27.46 | cr2 | to login and debug ATI. |
00:28.40 | Kevin2 | cr2: Heh, that'll be a lot of typing. What is USB connected to? |
00:29.22 | cr2 | isp1582 |
00:29.46 | cr2 | i've found the linux driver for 1362 in the blackfin branch. |
00:30.19 | cr2 | i've also found the 1582 code in some attachment to an obscure ML, but can't find the official download. |
00:31.22 | cr2 | but it's GPL, so there is no problem. don't know if this source is going to compile, so my idea is to print and compare the 1362 and 1582 datasheets. |
00:32.15 | cr2 | the USB host is pxa27x-ohci, so i can put rootfs on the USB stick. at least in theory. |
00:32.49 | cr2 | the usb stick works with wince. i actually run haret off it :) |
00:34.03 | Kevin2 | Is it usb 2.0 or something? Odd that they wouldn't reuse the pxa one. |
00:34.28 | cr2 | isp1582 is usb2.0 |
00:34.40 | cr2 | pxa27x ohci is 1.1 |
00:34.54 | cr2 | i don't know why they didn't use 2.0 host in 1582. |
00:35.13 | cr2 | isp1582 even can otg. |
00:35.52 | *** join/#htc-linux Keizer (n=keizer@c-69-139-42-47.hsd1.fl.comcast.net) |
00:36.46 | Kevin2 | Supposedly the pxa can do OTG also. I don't know if the linux driver supports it though. |
00:37.20 | cr2 | i doubt it. |
00:43.19 | cr2 | 0x58, 0x59, 0x5d for sd1. |
00:51.17 | cr2 | Zoolooc: do you have FM radio on kaiser ? |
00:51.38 | Zoolooc | cr2: unfortunately not |
00:52.32 | Zoolooc | that, VGA, IR and OTG is missing from that device |
00:53.35 | cr2 | but it's possible ? |
00:55.19 | Hymie | cr2: let's bankrupt amazon and have the kindle ;) |
00:55.20 | Zoolooc | OTG shows in some MSM7k "data sheet" |
00:55.23 | Hymie | hack |
00:55.42 | Zoolooc | VGA is only possible with 7200A I guess |
00:56.01 | cr2 | vga out ? or vga LCD ? |
00:56.14 | Zoolooc | don't know about FM radio |
00:57.18 | Zoolooc | LCD I think |
00:57.36 | Zoolooc | cr2: waht about this script: http://rafb.net/p/nhSv2E54.html |
01:03.25 | Zoolooc | cr2: gave me this: http://rafb.net/p/BemNma94.html |
01:05.52 | Zoolooc | will go to bed now. cr2: the haretlog.txt file is on my uSD card in the Kaiser, so I will have it with me tomorrow @work, if you see me in the channel and want me to put it in a wiki page, please remind me |
01:06.13 | Zoolooc | good night folks |
01:07.20 | cr2 | ok, good night |
01:07.21 | Zoolooc | did you have a brief look, does it look as expected? |
01:08.11 | Kevin2 | Good night. |
01:09.29 | cr2 | Zoolooc: it's to early . |
01:10.56 | Zoolooc | ok :) really gone now, good night |
01:11.05 | dcordes | it's a bout time for me too |
01:11.07 | dcordes | gnite all |
01:13.33 | cr2 | good night. |
02:00.24 | *** join/#htc-linux marmotta (n=skodde@unaffiliated/skodde) |
02:21.54 | Kevin2 | cr2: Still around? |
02:38.18 | Kevin2 | cr2: Well, if you get this message later - I added register name decoding to mmutrace events in haretconsole. You'll need to "cvs up" and recompile haret. I've also added the framework (but not all the defs) for the ATI chips - see haretconsole/regs_ati.py. |
03:55.41 | *** join/#htc-linux Zoolooc_ (n=fredsiba@p549526BF.dip0.t-ipconnect.de) |
04:58.57 | *** join/#htc-linux marmotta (n=skodde@unaffiliated/skodde) |
05:06.59 | *** join/#htc-linux skodde2 (n=skodde@unaffiliated/skodde) |
05:23.12 | *** join/#htc-linux goxboxlive (n=jrs@195.159.97.196) |
05:23.18 | goxboxlive | Good morning |
05:57.34 | *** join/#htc-linux |dion| (n=dion@inhex.net) |
06:30.33 | *** join/#htc-linux _law_ (n=law@mail.stiftadmont.at) |
06:39.17 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
07:28.53 | *** join/#htc-linux kiozen (n=oeichler@p5492A373.dip0.t-ipconnect.de) |
07:32.59 | *** join/#htc-linux rmoravcik (n=moravcik@gtsgw.ttc.cz) |
08:06.59 | *** join/#htc-linux rob_w|laptop (n=rob_w@p549B9A85.dip0.t-ipconnect.de) |
08:28.00 | *** join/#htc-linux sadeness_ (n=LamersIn@nat1.rlan.ru) |
09:05.40 | *** join/#htc-linux AC_Mast (i=s@gateway/tor/x-0deaf0deaac549b4) |
11:37.04 | *** join/#htc-linux goxboxlive (n=jrs@195.159.97.196) |
11:46.10 | *** join/#htc-linux ltxda (n=ltxda@c-98-196-125-205.hsd1.tx.comcast.net) |
11:46.42 | *** join/#htc-linux zoolooc (n=lucian@p549526BF.dip0.t-ipconnect.de) |
11:48.42 | zoolooc | hello |
11:48.52 | zoolooc | cr2: I uploaded the haret log with preliminary GPIO dump linked at http://wiki.xda-developers.com/index.php?pagename=KaiserMemoryMap |
12:19.18 | *** join/#htc-linux TeringTu1y (n=maarten@195-241-125-243.ip.telfort.nl) |
12:41.54 | *** join/#htc-linux Jax184 (i=Jax184@75.153.81.166) |
12:41.56 | Jax184 | rah |
12:45.59 | Jax184 | anyone alive? |
12:49.58 | *** join/#htc-linux TeringTuby (n=maarten@195-241-125-243.ip.telfort.nl) |
12:55.14 | *** join/#htc-linux ltxda0 (n=ltxda@c-98-196-125-205.hsd1.tx.comcast.net) |
13:15.21 | *** join/#htc-linux ltxda (n=ltxda@c-98-196-125-205.hsd1.tx.comcast.net) |
13:25.41 | *** join/#htc-linux AC_Mast (i=s@gateway/tor/x-b1c8170bb81ff18b) |
13:25.57 | *** join/#htc-linux the_sys0p (i=the_sys0@gateway/tor/x-df2bbfab9da68c03) |
13:27.11 | *** join/#htc-linux Cyorxamp (n=Cyorxamp@212.57.232.254) |
13:33.03 | *** join/#htc-linux goxboxlive (n=jrs@195.159.97.196) |
13:33.06 | goxboxlive | .no |
13:33.56 | Jax184 | yes? |
13:58.23 | goxboxlive | what |
14:46.23 | *** join/#htc-linux dcordes (n=user0815@f048226146.adsl.alicedsl.de) |
14:46.58 | dcordes | moin |
14:55.46 | *** join/#htc-linux Kevin2 (n=Kevin@207-237-65-88.c3-0.avec-ubr12.nyr-avec.ny.cable.rcn.com) |
15:15.02 | dcordes | how would I go about reading the logs from last night? are there any live logs? |
15:18.07 | dcordes | I'm aware of http://apt.rikers.org/%23htc-linux/ but how about the active day? |
15:25.16 | Kevin2 | dcordes: I use the logs at http://irclog.iclem.net/?chan=htc-linux |
15:26.35 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
15:31.15 | dcordes | Kevin2: cool thanks. What are mmuctrace events? |
15:41.13 | Kevin2 | dcordes: Haret can alter the ARM cpu Memory Management Tables to force certain memory accesses to fault. Haret uses this (via addlist mmutrace) to "watch" all the ways wince reads/writes to certain chips. |
15:41.34 | Kevin2 | s/Tables/Unit/ |
15:41.55 | Kevin2 | bbl |
15:41.57 | *** part/#htc-linux Kevin2 (n=Kevin@207-237-65-88.c3-0.avec-ubr12.nyr-avec.ny.cable.rcn.com) |
15:53.44 | *** join/#htc-linux goxboxlive (n=goxboxli@81.80-202-132.nextgentel.com) |
16:00.05 | *** join/#htc-linux tonyk (n=tli@deng.tekla.com) |
16:23.48 | kiozen | silent today, cr2 any n560 news ? |
16:32.44 | *** join/#htc-linux law|home (n=_law_@213.173.86.202) |
17:00.14 | *** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
17:00.28 | *** join/#htc-linux Kevin2 (n=Kevin@207-237-65-88.c3-0.avec-ubr12.nyr-avec.ny.cable.rcn.com) |
17:01.50 | *** join/#htc-linux pH5 (n=ph5@e178242116.adsl.alicedsl.de) |
17:05.49 | cr2 | hi kiozen |
17:06.02 | cr2 | looking at the n560 LCD init now. |
17:09.43 | pH5 | good evening |
17:11.03 | cr2 | hi pH5 |
17:11.12 | Kevin2 | Hi |
17:14.05 | cr2 | hi Kevin2 |
17:16.26 | pH5 | hey Kevin2. I'm looking for a way to remove holes in my EGPIO space. Could apache work with 8bit accesses to its egpio cpld? |
17:17.30 | Kevin2 | pH5: I'm pretty sure 16 bit accesses are the "correct" way. The chip itself is configured in the pxa memory controller as a 16 bit chip. |
17:17.48 | kiozen | hi cr2 |
17:17.54 | Kevin2 | So, if you make an 8bit access, it will just discard half the data, and if you make a 32 bit access it will make two 16 bit reads. |
17:18.19 | pH5 | Kevin2: discarding half the data for single gpio access is just fine, the question is will strb work? |
17:19.52 | pH5 | I want to adapt htc-egpio to grow gpio_chip support, and for magician I could fit everything nicely into two gpio_chips (one in, one out), if I didn't have 8 dead GPIOs per register because of the 16bit access. |
17:19.53 | cr2 | pH5: it's always strh/ldrh by htc |
17:20.11 | cr2 | Kevin2: i have some problems with htc-egpio too. |
17:20.31 | cr2 | first, athena has 9 regs. |
17:20.58 | cr2 | second, the bus_shift value provided by platform_driver is never checked. |
17:21.33 | cr2 | and i get some stray (number>7) irqs from the cpld. |
17:22.07 | pH5 | cr2: which chip has 9 regs? cpld1? |
17:22.27 | cr2 | yes. and kaiser too. don't remember about others. |
17:22.43 | cr2 | pH5: the 0x10 reg is 9 ? |
17:22.55 | pH5 | hm. why is there no KaiserCPLD wiki page yet? where are all the kaiser owners? :) |
17:23.27 | cr2 | pH5: but there is Kaiser_MDP for them :) |
17:23.35 | Kevin2 | On a strb, the pxa will raise some kind of a "1-bit access" line. I think it's risky to use strb. |
17:23.36 | cr2 | so they'll be busy today. |
17:24.15 | pH5 | cr2: he, nice. |
17:24.57 | Kevin2 | s/1-bit/1-byte/ |
17:25.53 | Kevin2 | cr2: Did you see my message last night about haretconsole register name decoding for mmutrace? |
17:26.41 | *** part/#htc-linux zoolooc (n=lucian@p549526BF.dip0.t-ipconnect.de) |
17:26.52 | cr2 | Kevin2: yes. i don't get one thing, though. how are you tracing ati which has multiple virtual mappings ? |
17:27.07 | cr2 | is it enough to declare only the physical address ? |
17:30.15 | Kevin2 | Yes. I find the phys addr for all the virt mappings, then do the register lookups by the phys addr. |
17:30.30 | cr2 | ok. |
17:30.32 | Kevin2 | I'll be back later. |
17:32.15 | pH5 | Kevin2: I'm running fine with strh, although HTC use strb on magician. I guess I have to add some magic around readw to differentiate between 8bit and 16bit register width then. |
17:33.06 | Kevin2 | pH5: Can you check if magician registers the chip as 8 or 16 bit with memory controller? |
17:35.02 | cr2 | Kevin2: there is no memctl decoder in haret. |
17:35.30 | cr2 | btw, what should be the right way to decode the registers for a given cpu? |
17:35.50 | cr2 | i needed the pxa ohci settings, so i hacked on gnu-haret again. |
17:37.10 | cr2 | i'm also using the old perl scripts on occasion. |
17:38.20 | pH5 | Kevin2: would that be MSC1[RBW3] for CPLD on CS3? |
17:41.41 | *** join/#htc-linux kiozen (n=oeichler@rgnb-4db03d73.pool.einsundeins.de) |
17:41.59 | cr2 | hehe. wince crashed while running haret. |
17:42.27 | pH5 | MSC0 = 0x7ff001a0, MSC1 = 0x7ff01880, MSC2 = 0x16607ff0 |
17:42.39 | pH5 | so rom bus width = 32bit all the way |
17:45.43 | cr2 | insufficient available bus power. |
17:46.16 | cr2 | but works in wince ;-) |
17:48.32 | cr2 | LOL |
17:48.44 | cr2 | connected the usb keyboard to athena :) |
17:49.00 | cr2 | and it works. |
17:49.24 | cr2 | i guess the mouse will work too. |
17:50.54 | cr2 | pH5: i've used .power_budget = 100, is there any way to find out the correct value ? |
17:53.51 | pH5 | cr2: no idea. disassemble wince or connect to it with a fake gadget with increasing power requirements? |
17:54.27 | cr2 | pH5: where is it in wince ? |
17:55.01 | pH5 | I don't know |
17:55.52 | cr2 | ok. |
17:56.59 | cr2 | zaurus sets 150mA, some omap board 250mA |
17:57.16 | cr2 | i guess i can just set 0 for unlimited. |
17:57.29 | pH5 | it really depends on what kind of switches are on the board |
17:57.35 | cr2 | unlike on the zaurus, there is hw overcurrent protection. |
18:00.30 | pH5 | cr2: if it doesn't supply 500mA, unlimited is the wrong value. try connecting power-hungry usb hard drives. |
18:02.52 | cr2 | my SDcard adapter does not work in wince, but the sticks are. |
18:03.04 | cr2 | so i doubt it's providing 500mA |
18:03.15 | cr2 | and i always run with external power supply. |
18:04.49 | cr2 | maybe i should try a powered hub :) |
18:07.55 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
18:11.03 | cr2 | got a small problem though. |
18:11.43 | cr2 | which will prevent booting with rootfs on usb disk. |
18:13.49 | pH5 | yes? |
18:15.23 | cr2 | the hub needs a reconnect |
18:15.52 | cr2 | so i can have the disk,keyboard and mouse connected. |
18:16.19 | cr2 | otherwise i'll boot off the stick, but there will be no way to login. |
18:17.15 | cr2 | i don't know if a reconnect while rootdelay can help. |
18:18.38 | cr2 | heh. it seems that the most driver problems are with applying power ;) |
18:19.33 | cr2 | pH5: do we have a standard way to apply power to a pxa uart device ? |
18:19.50 | cr2 | then i'll login over bluetooth :) |
18:21.49 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
18:22.58 | cr2 | hi BabelO |
18:23.22 | BabelO | hi cr2 |
18:23.38 | pH5 | cr2: we have the platform_pxa_serial_funcs in hh.org |
18:24.11 | cr2 | pH5: is somebody using them ? |
18:24.55 | cr2 | BabelO: you are allowed to produce a 250GB qtopia image now :) using many keyboard, mice and joysticks :) |
18:25.51 | cr2 | BabelO: i've connected a 250GB disk, mouse and keyboard to athena. and they all work in linux. |
18:25.58 | *** join/#htc-linux |dion| (n=dion@inhex.net) |
18:27.08 | BabelO | bow, you need video out :) |
18:27.21 | cr2 | pH5: right now everybody follows the hx4700 for applying uart power. i guess if it'll be done right, everybody will benefit and copy it. |
18:27.29 | welp | BabelO: you made a newer gpe image yet? :> |
18:28.02 | cr2 | BabelO: it's an i2c chip with a datasheet, if you mean the TVout. |
18:28.27 | cr2 | BabelO: but it may include some ATI magic, i don't yet know. |
18:29.16 | BabelO | welp: no |
18:29.40 | BabelO | cr2: yes tv out, then you have a complete micro computer ;) |
18:29.56 | welp | boo, the one i made a couple of weeks back doesn't seem to work 100% - the battery monitor's not working |
18:31.22 | cr2 | BabelO: there is TVout and vga out. |
18:32.20 | BabelO | cr2: but signal go out from cinch connector for both ? |
18:32.32 | cr2 | welp: the kernel battery monitor on ba is not quite right. |
18:32.59 | cr2 | BabelO: there is FBAS,S-video and VGA connector. |
18:33.09 | welp | cr2: aah, fair enough |
18:33.47 | cr2 | welp: do you know your pasic2 register values ? |
18:34.23 | welp | i've never done any assembly or anything of the kind.. |
18:34.27 | welp | (unfortunately) |
18:34.41 | welp | but, i'm only 17! i have a long way to go! ^_- |
18:35.07 | cr2 | welp: it's a haret command in old gnu-haret |
18:36.47 | welp | i've only been messing about with this since about 2 or 3 weeks ago |
18:40.49 | dcordes | hi alö |
18:40.53 | dcordes | hi all |
18:43.04 | dcordes | cr2: any haret tasks for me? |
18:44.10 | cr2 | dcordes: i've created the Kaiser_MDP wiki page with the data from androids. |
18:44.33 | cr2 | dcordes: you can dump the actual kaiser values, and create a new column |
18:45.46 | cr2 | you can dump the MDP_FULL_BYPASS_WORD33 and MDP_FULL_BYPASS_WORD37 |
18:46.07 | cr2 | so we can check if your LCD parameters are consistent with the android driver. |
18:46.31 | cr2 | the page is here http://wiki.xda-developers.com/index.php?pagename=Kaiser_MDP |
18:47.00 | dcordes | already found ok |
18:47.49 | cr2 | you can also try to trace this area, but it may result in a big flood. |
18:48.09 | cr2 | so maybe make a static dump first. |
18:49.12 | dcordes | what was the command for static dump? |
18:49.25 | pH5 | pd |
18:50.15 | pH5 | cr2: the thing about pxa_serial_funcs is nobody had the heart to submit those patches upstream |
18:51.29 | cr2 | pH5: I didn't look at it in detail. can the current hx4700 implementation be copied to other devices ? |
18:52.22 | cr2 | i meand the hx4700_bt |
18:54.46 | pH5 | cr2: it's a hack, it should probably abstracted out into a brf6150 driver. |
18:55.04 | pH5 | is it possible to somehow modify this code so that it'll also work on s3c? (I see it accesses BTMSR directly) |
18:55.12 | pH5 | or omap |
18:56.07 | *** join/#htc-linux Zoolooc (n=fredsiba@p549526BF.dip0.t-ipconnect.de) |
18:56.27 | dcordes | hi Zoolooc |
18:56.42 | cr2 | pH5: i see a big "generic" value in such driver. |
18:56.52 | Zoolooc | hi folks |
18:56.59 | cr2 | pH5: i need something like that for gps on ffuart, for example. |
18:57.54 | dcordes | cr2: what do I set as <size> for 0x10184? |
18:58.06 | cr2 | pH5: and there is also an old problem with IR/rs232 on the same uart depending on the gpio setting. hmm. the AT radio/ bootloader is in the same category. |
18:58.34 | pH5 | right, I still didn't manage to talk to the radio bootloader from linux :/ |
18:58.35 | cr2 | dcordes: these are all 32bit (4byte) registers. |
18:59.41 | dcordes | ok for # "pd 0x10184 4" I got "00010184 | 756e204f | O nu" |
18:59.43 | cr2 | pH5: i don't mind to copy hx4700_bt for all these, but then we will have 20 such drivers, so it's nice when this will be done right. |
19:00.39 | pH5 | cr2: I agree. I already more or less copied this for magician_phone. |
19:01.16 | cr2 | and 20 is probably an underestimation. |
19:01.35 | dcordes | cr2: what part of the output am I supposed to put in which column of the wiki? |
19:02.04 | cr2 | dcordes: create a new column |
19:02.47 | cr2 | dcordes: hmm. sorry. |
19:02.53 | *** join/#htc-linux rob_w (n=bob@X0419.x.pppool.de) |
19:03.00 | cr2 | you forgot to add the MDP_BASE |
19:03.08 | cr2 | aka 0xaa200000 |
19:03.32 | cr2 | you are just dumping the SRAM now. that's why the numbers look weird. |
19:04.17 | cr2 | pd 0xaa210184 4 |
19:06.20 | dcordes | so I use pd 0xaa2 as prefix vor all values in the first column? |
19:06.32 | cr2 | yes. |
19:06.34 | dcordes | 0xaa2 I mean |
19:08.16 | dcordes | aa210184 |EXCEPTION while reading from address 92110184; 00000000 | ... |
19:10.25 | dcordes | shall I use thte mmutrace thing now? |
19:10.27 | cr2 | dcordes: how does your 'dump mmu' look like ? |
19:10.30 | cr2 | no. |
19:10.37 | dcordes | what is my dump mmu? |
19:10.52 | cr2 | dump mmu |
19:11.06 | cr2 | haret command that dumps the MMU tables. |
19:11.19 | dcordes | lot of output, 4 columns looks nice |
19:12.51 | dcordes | on the phone status bar is shown up while the stuff ist printed on my console |
19:13.04 | dcordes | End of virtual address space at the end |
19:13.08 | cr2 | what are the entries for the 0xaa200000 physical address |
19:13.29 | Zoolooc | how's going folks with kaiser spying? |
19:14.34 | dcordes | HaRET(3)# pd 0xaa210184 4 |
19:14.36 | dcordes | HaRET(4)# pd 0xaa210184 4 |
19:15.28 | dcordes | hm oh I'm not so fresh today ^^ greped the log. that's all it outputs when I grep for 0xaa2 |
19:15.44 | cr2 | dcordes: if it's mapped, you can't get an exception. |
19:16.07 | Zoolooc | aren't the responses in the second line? |
19:18.00 | cr2 | Zoolooc: you can try it too. |
19:18.43 | Zoolooc | will try, right now I'm a bit busy few minutes... |
19:19.46 | dcordes | I cleared the logs, reopened connection, and "mmu dump"ed again. the new log file does not contain any line with 0xaa2 in it |
19:20.00 | Zoolooc | can one connect to haret via the ppp interface established via usb with synce-serial? I couldn't, I can only connect via WLAN, because I run dnsmasq on my FritzBox |
19:21.07 | kiozen | BabelO: while cross compiling gdal the final linker command wants to link against my system's /usr/lib64/librt.so |
19:21.24 | dcordes | cr2: can I take from this that the physical address space 0xaa2* is not mapped on my device right now? |
19:21.27 | kiozen | you know where to hit configure to chose the right one from the arm tree |
19:21.45 | cr2 | kiozen: why don't you use oe for gdal ? |
19:21.52 | pH5 | Zoolooc: yes, that should be possible. you need to know the fixed IP address for that case. |
19:21.59 | kiozen | hm, good question |
19:22.12 | kiozen | I try.. |
19:22.39 | Zoolooc | pH5: oh, then I'll specify it in my udev rule for synce-serial-setup |
19:23.28 | Zoolooc | pH5: now I realize I tried with the desktop's local IP on ppp0, instead the remote (the PDA) |
19:26.36 | dcordes | cr2: how to go on? |
19:27.00 | Zoolooc | actually, I should setup a bridge to give Kaiser internet access via USB like on windows |
19:27.24 | cr2 | dcordes: can i download your "dump mmu" ? |
19:27.47 | dcordes | does it contain critical data? |
19:28.19 | cr2 | no. |
19:28.19 | dcordes | I don't want to paste it in that case. |
19:28.42 | cr2 | it's a mapping table, nothing useful. |
19:28.48 | dcordes | oke |
19:29.02 | dcordes | so I can pastebin w/o problems? |
19:29.12 | cr2 | <=150K |
19:29.24 | dcordes | dcc? |
19:29.32 | cr2 | will not work. |
19:29.34 | BabelO | kiozen: i ve some broken path too, i always fix it by hand in makefile |
19:29.38 | *** join/#htc-linux dcordes_ (n=dcordes_@f048226146.adsl.alicedsl.de) |
19:30.09 | cr2 | kiozen: there are 6 lcds, and each has its own init ;-) |
19:30.50 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
19:30.59 | dcordes_ | cr2, sending dcc |
19:32.50 | kiozen | 6 different one for the n560? |
19:33.36 | kiozen | Thanks BabelO. I wonder what is less PITA, fixing the makefiles or compiling gdal via oe |
19:34.18 | cr2 | kiozen: 5xx |
19:34.31 | dcordes | cr2: dcc timeout |
19:34.46 | cr2 | kiozen: only 1 vga lcd though. |
19:34.53 | BabelO | kiozen: for ùe it is fixing makefile, i ve strange problem with oe these days |
19:35.29 | kiozen | cr2: that explains why each loox has it's own problems. |
19:35.49 | kiozen | the c550 seems to be the worse one |
19:36.53 | dcordes | cr2: can I load it up somewhere? |
19:37.37 | dcordes | have to go in a minute. examination tomorrow |
19:39.13 | cr2 | ok. |
19:39.20 | cr2 | Zoolooc: can you run haret ? |
19:42.18 | Zoolooc | cr2: yes, can it wait few minutes, or do you also want to to leave, or do you want to get something finished about kaiser now? |
19:42.44 | Zoolooc | cr2: I'm flexible, if you want me to do it now |
19:42.57 | dcordes | cr2: received querry? |
19:44.24 | cr2 | 92100000 | aa200000 | 1MB section | D=0 AP=1 ?=2000 |
19:45.04 | cr2 | dcordes: there should not be an exception. |
19:45.24 | dcordes | let me retry |
19:45.48 | cr2 | Zoolooc: can you try 'pd 0xaa210184 4' ? |
19:45.57 | cr2 | dcordes: do a soft reset before. |
19:46.28 | dcordes | cr2: same thing! |
19:46.47 | dcordes | soft reset=reboot? |
19:46.50 | pH5 | eh.. |
19:47.05 | pH5 | pd takes physical addresses |
19:47.17 | pH5 | for virt there's vd, isn't there? |
19:47.37 | cr2 | pH5: this is a physical address. |
19:47.41 | dcordes | booting... |
19:47.50 | pH5 | 921 is the virt one? |
19:47.51 | cr2 | virtual will be 0xb21 |
19:47.55 | cr2 | yes. |
19:48.02 | pH5 | oy, sorry. |
19:48.05 | cr2 | that's the dump mmu format. |
19:48.13 | cr2 | yeah, msm is strange. |
19:48.39 | cr2 | pH5: you got used to pxa nad omap too much :) |
19:49.24 | cr2 | they are much more straightforward. |
19:49.26 | Zoolooc | cr2: that gives an exception here |
19:49.41 | cr2 | Zoolooc: ok. interesting. |
19:50.00 | dcordes | cr2: rebooted, opened connection as usual and get "pd 0xaa210184 4" > aa210184 |EXCEPTION while reading from address 92110184; 00000000 | .... |
19:50.15 | cr2 | what about 'vd 0xb110184 4' ? |
19:50.21 | cr2 | hm. |
19:50.23 | dcordes | same |
19:50.29 | dcordes | 0b110184 |EXCEPTION while reading from address 0B110184 |
19:50.30 | cr2 | vd 0xb210184 |
19:50.39 | cr2 | vd 0xb210184 4 |
19:50.53 | dcordes | same0b210184 |EXCEPTION while reading from address 0B210184 |
19:50.54 | Zoolooc | exception |
19:51.20 | pH5 | somebody needs to give a msm device to Kevin2 and have him update haret :) |
19:51.26 | cr2 | Kevin2: are you here ? |
19:51.49 | dcordes | I can arrange a nice root shell on console client tomorrow |
19:51.54 | cr2 | yeah, looks like a haret bug. |
19:52.09 | dcordes | if that would help |
19:52.31 | dcordes | he only has to poke me for clicking listen button on kaiser |
19:53.01 | dcordes | Kevin2: was mentioning something about mmu patches or so earlier this day |
19:53.32 | Zoolooc | didn't he commit those yesterday? I compiled hared last night |
19:54.09 | dcordes | Zoolooc: I run the binary you dcced me |
19:56.10 | Zoolooc | dcordes: yes, that's what I assumed |
19:56.19 | dcordes | 03:38 Kevin2 cr2: Well, if you get this message later - I added register name decoding to mmutrace events in haretconsole. You'll need to "cvs up" and recompile haret. I've also added the framework (but not all the defs) for the ATI chips - see haretconsole/regs_ati.py. |
19:56.55 | cr2 | yes, i know. |
19:57.28 | Zoolooc | dcordes: I remember this, I updated haret from CVS in the meantime, I might have dcced you my previous build, but it seems that's another issue |
19:57.40 | dcordes | ok |
19:58.08 | dcordes | sorry but I have work to do. bbl |
19:59.17 | cr2 | kiozen: i've hit some strange non-gpio code in the lcd init. |
19:59.53 | kiozen | That's the Steve Balmer special, developer,developer,developers |
20:00.12 | kiozen | what do you think it is? |
20:01.09 | cr2 | we can ignore it for now :) |
20:01.33 | cr2 | you've done tracing the lcd powerup once. |
20:01.51 | cr2 | i see 3 pxa gpios, and 2 cpld gpios. |
20:02.11 | kiozen | ok |
20:02.12 | cr2 | it goes like: |
20:03.25 | cr2 | 96=1, sleep, 95=1, sleep, 91=0, sleep, 91=1, sleep, cpld(1,5,4), sleep, cpld(1,5,8), sleep |
20:03.49 | cr2 | you also have the lccr4 top bit set, like the universal. |
20:04.00 | cr2 | heh. it's probably the lccr* stuff in the middle. |
20:05.36 | cr2 | pH5: is it possible to apply pxafb power in some obvious way without creating a n560_lcd driver ? |
20:05.59 | cr2 | i think it was possible once, but then was deprecated, or something like that. |
20:07.06 | pH5 | cr2: .pxafb_lcd_power in pxafb_mach_info |
20:07.15 | cr2 | linux is not a mobile os ;) just a big monolitic x86 kernel ;) |
20:07.48 | cr2 | pH5: hmm. ok. but i think it's used only on the old ipaqs... |
20:07.57 | J4k3 | linux is a pretty good embedded os |
20:08.04 | kiozen | what's the reason for such a complex sequence? I would have expected one signal for powering the lcd and some bit pattern to set the backlight |
20:08.06 | J4k3 | with uclibc |
20:08.07 | pH5 | it is deprecated, but I'm still using it on magician, because it flickers less than the new method :) |
20:09.35 | cr2 | pH5: yes, wince sets 3 gpios, configures the LCD regs, and applies 2 more gpios. it's the infamous _pre, _post stuff. |
20:10.38 | cr2 | J4k3: uclibc has nothing to do with the kernel. and linux 'modules' architecture is so braindead, that i'd better shut up now ;) |
20:12.21 | pH5 | cr2: yes, we need support for that in pxafb to get it right. |
20:12.46 | pH5 | currently it calls the lcd_power function after enabling the lcd controller, or before disabling it again. |
20:17.35 | cr2 | pH5: a similar problem exists on the LCDs with the spi config. |
20:23.12 | pH5 | cr2: yes, there is this nice lcd_device abstraction for lcd power in drivers/backlight/lcd.c, but it also pays no attention that the gpio fiddling has to be synchronized with the lcd controller setup. |
20:24.04 | pH5 | spi scares me, but for the pre/post things it would be enough to add some more options to pxafb_lcd_power, which is ugly, because it's pxafb specific. |
20:28.53 | cr2 | kiozen: yes, it's the lccr* fiddling in the middle. |
20:30.22 | cr2 | so it's like 'set 2 gpios, setup lccr, reset 1 gpio, set 2 cpld gpios' |
20:36.34 | cr2 | now i only need to find out the lcd id process. the board id has 2 bits, so you can't have 6 options ;) |
20:37.53 | pH5 | cr2: perhaps those are complex bits. |
20:39.31 | cr2 | or i just copy the stuff over to the n560 directory, and have 1 config :) |
20:44.17 | J4k3 | cr2: well, by comparison to any other offering, its the best thing available for that 'size' range. |
20:48.44 | goxboxlive | hmm Openmoko works kind of great now on universal, but still miss sms support. |
20:49.12 | cr2 | goxboxlive: is it fast ? |
20:49.17 | goxboxlive | cr2: nope |
20:50.15 | goxboxlive | cr2: i have made a easy install package of it. It's uploaded allready if you want to try it. The kinetic scrolling is amazing. |
20:50.43 | pH5 | goxboxlive: try the moko-gtk-engine |
20:50.49 | goxboxlive | And i dont get how Openmoko manage to adjust the correct time, year and date. |
20:51.39 | cr2 | goxboxlive: does it support the accelerometer ? i have one in athena :) |
20:52.28 | goxboxlive | cr2: no idea |
20:52.50 | goxboxlive | pH5: isent that included in the openmoko distro? I am not using angstrom with openmoko. |
20:53.18 | goxboxlive | pH5: i guess it's not |
20:56.04 | pH5 | goxboxlive: you have to bitbake it yourself. angstrom, openmoko, doesn't matter. |
20:56.12 | *** join/#htc-linux Zoolooc_ (n=fredsiba@p549526BF.dip0.t-ipconnect.de) |
20:57.14 | *** join/#htc-linux kiozen_ (n=oeichler@rgnb-4db03d73.pool.einsundeins.de) |
20:59.53 | goxboxlive | pH5: i have done it. Is it just to install and run openmoko or do i have to edit something? |
21:01.52 | pH5 | goxboxlive: I had to edit /etc/gtk-2.0/gtkrc, not sure if that is still the case. |
21:02.43 | goxboxlive | ok, i'll look |
21:03.41 | cr2 | pH5: do we have AT%EM on calypso ? |
21:03.59 | goxboxlive | pH5: gtkrc doesnt exist, and moko-engine isent running either |
21:04.12 | pH5 | cr2: it's not enabled in magician's firmware. |
21:04.15 | kiozen_ | BabelO: what's the difference between a host and a build system |
21:05.48 | cr2 | BabelO: do you have AT%EM on ba ? |
21:06.07 | BabelO | cr2: don't think |
21:06.25 | pH5 | cr2: we need ti's sources :) |
21:06.51 | pH5 | or somebody with magical abilities to can transplant thumb interwork code from the gta01 firmware into ours. |
21:06.55 | BabelO | kiozen_: don't know, the same ? or the host is a clean base for the build system, like OE do ? |
21:07.02 | pH5 | s/to/who/ |
21:07.33 | kiozen_ | hm.. |
21:08.32 | cr2 | pH5: do you know how deal with the radio.nbf ? |
21:08.41 | BabelO | kiozen_: do you use QSettings ? |
21:08.53 | kiozen_ | in QLandkarte, yes |
21:09.29 | pH5 | cr2: sure, itsme is the answer to everything |
21:09.34 | BabelO | kiozen_: i use arrayIndex do save profile settings , and i want to delete them, but i can't, don't know why |
21:09.43 | pH5 | there's a decoderadio.pl floating around somewhere |
21:10.00 | pH5 | but reverse engineering mixed thumb/arm code is no fun |
21:10.22 | cr2 | ok. |
21:10.29 | pH5 | (even more if you have no idea what the processor registers are) |
21:10.47 | kiozen_ | afaik you are not supposed to delete anything |
21:10.57 | kiozen_ | just rewrite the config and save |
21:11.36 | goxboxlive | http://wiki.xda-developers.com/index.php?pagename=RomTools/decoderadio.pl |
21:12.12 | kiozen_ | grr... still havent found out how to tell the linker not to use /usr/lib64/librt.so for arm targets |
21:12.17 | BabelO | kiozen_: yes, but sometimes, you want to delete profile ;) |
21:12.44 | kiozen_ | it's like the registry in windows, it never forgets |
21:13.30 | goxboxlive | pH5: could you pastebin your gtkrc file. It seems like i dont have it |
21:13.30 | BabelO | kiozen_: yes , but it lose data all 6 month when you format your pc again ;) |
21:13.51 | kiozen_ | what about clear() and rewriting the config? |
21:14.20 | BabelO | too long for the phone, i want to do less access as possible |
21:15.26 | kiozen_ | it has to dump the complete text file to disk anyway |
21:16.48 | BabelO | kiozen_: yeah i found my mistake, it is ok now ;) |
21:16.56 | kiozen_ | tell us :) |
21:17.17 | *** join/#htc-linux TimRiker (n=timr@209-33-233-142.dsl.infowest.net) |
21:18.27 | BabelO | kiozen_: i just do like everyone, just one line of code to delete 6 line in config... not a good idea ;) |
21:18.35 | cr2 | kiozen_: gave up for now. too much .asm today :) |
21:19.07 | kiozen_ | tell us :) |
21:19.24 | kiozen_ | ok cr2 :) |
21:20.40 | BabelO | cr2: kiozen_ a good question, what is the best icon to tell a user that its function is to set a default profile ? |
21:20.55 | BabelO | hummm not sure my qustion is smart |
21:21.22 | kiozen_ | head's profile with a star? |
21:21.42 | kiozen_ | like new folder |
21:22.00 | BabelO | kiozen_: ahh good idea :) |
21:22.28 | BabelO | kiozen_: it is for that http://linuxbox2.free.fr/imlife/ |
21:23.00 | BabelO | capture46.png |
21:24.01 | kiozen_ | add a new viking :) |
21:24.13 | BabelO | lol ;) |
21:25.41 | BabelO | all done with vi :) |
21:26.14 | kiozen_ | so use vi as input front end |
21:26.28 | kiozen_ | you will end up with a lot of friends :) |
21:27.07 | cr2 | goxboxlive: i've read a rather dumbass comment in the openmoko archive :) saying that universal is a rather slow device. compared to gta01 probably. |
21:27.10 | BabelO | kiozen_: now i use monkey studio, but there is lot of bug in it |
21:27.38 | kiozen_ | yes I removed imediately it after 2 min of use |
21:27.45 | cr2 | goxboxlive: i winder what they'll say about athena. |
21:27.51 | cr2 | s/win/won/ |
21:28.30 | BabelO | kiozen_: ok, but i want kdevelop on windows, to continue my work between 12:00am and 1:00pm |
21:28.42 | goxboxlive | cr2: Do they mean when running Openmoko? |
21:29.03 | kiozen_ | don't they have a windows port already? |
21:29.17 | BabelO | kiozen_: really ? not found then |
21:29.44 | kiozen_ | thought I have seen something on the kdevelop mailing list |
21:29.59 | BabelO | kiozen_: yes, but with cygwin installed :( |
21:30.17 | kiozen_ | and the whole kde stuff |
21:30.35 | cr2 | goxboxlive: no, just a general comment. |
21:31.08 | cr2 | BabelO: is it possible to make qtopia4 keyboard and mouse friendly ? |
21:32.14 | cr2 | BabelO: the current config is very greenphone/gta01 oriented. i.e. designed for the phones with limited capabilities. |
21:32.16 | BabelO | cr2: for mouse i see how to do it, there is an ability to show/hide mouse when you want |
21:32.23 | cr2 | sorry for the pun :) |
21:32.24 | pH5 | goxboxlive: sorry, that filesystem is long gone. I think I needed that card for the diagnostic disk experiments. |
21:32.56 | pH5 | but it was only a single line including the gtkrc from /usr/share/themes/moko or similar |
21:33.08 | BabelO | cr2: for keyboard i use it over vnc without problem |
21:33.40 | cr2 | BabelO: btw, is it possible to add the pixmap for 'Back' button ? |
21:33.47 | goxboxlive | pH5: :-) ok |
21:34.10 | cr2 | i find it very counterintuitive how it#s done now. |
21:34.36 | goxboxlive | cr2: Universal is slow running wince. But Qtopia and Opie are fast. X11/GPE is also slow. But i guess it is a lot faster than the GTA01 |
21:34.49 | BabelO | cr2: it is because i set it with label only ;) i ve to try both config, and with label, i can set the best size for all component |
21:35.12 | goxboxlive | pH5: what is the purpose of running moko-engine. Is it faster or will the messager work? |
21:35.18 | cr2 | goxboxlive: that's why i talk about a dumbass comment ;) |
21:35.47 | goxboxlive | :-) |
21:35.57 | cr2 | BabelO: with the text label it looks like the touchscreen-less phone. |
21:36.42 | BabelO | cr2: yes agree |
21:44.31 | cr2 | pH5: i still don't get it. why htc never had a problem with updating radio, but fic has ? |
21:45.53 | pH5 | cr2: this is gsm business, I think nobody can understand this. |
21:46.01 | cr2 | they blame ti, but somehow i don't believe it. |
21:46.05 | cr2 | ok. |
21:54.15 | goxboxlive | pH5: what is the purpose of using the moko-engine? |
21:54.51 | goxboxlive | I notced that openmoko on universal is faster in portrait mode than landscpe. Does that make any sense? |
21:55.37 | cr2 | weird. |
21:56.00 | cr2 | because portrait is the "native" LCD mode. |
21:56.18 | cr2 | since there is no hw rotation like on ati. |
22:03.59 | cr2 | hehe. booting with root=/dev/sda1 sounds funny :) |
22:06.45 | pH5 | goxboxlive: moko-gtk-engine is thomas wood's xlib only gtk engine, which is very fast compared to the pixmap scaling fest that is the standard theme. |
22:06.54 | goxboxlive | pastebin.no/2809 |
22:07.08 | pH5 | I guess that is also the reason why landscape < portrait, maybe some pictures have to be scaled more. |
22:07.09 | goxboxlive | Why is this happening pH5, i am trying to build a magian image |
22:07.30 | goxboxlive | ok |
22:07.53 | pH5 | goxboxlive: add imagemagick-native to ASSUME_PROVIDED if you have it installed locally. |
22:08.16 | goxboxlive | pH5: ok |
22:08.36 | pH5 | the openmoko-icon-theme-standard2-qvga hack needs imagemagick and I didn't want to bother creating a native OE package for that. |
22:23.10 | *** join/#htc-linux rob_w (n=bob@X0419.x.pppool.de) |
22:30.19 | Jax184 | rah again |
22:34.58 | *** join/#htc-linux TimRiker (n=timr@pdpc/supporter/bronze/TimRiker) |
22:36.00 | *** join/#htc-linux the_sys0p (i=the_sys0@gateway/tor/x-117f31f0f9d56ef6) |
22:48.04 | cr2 | pH5: i'm still doing something wrong. the kernel crashes on any attempt to do gpio_get_value from CPLD. |
22:50.02 | Jax184 | I picked up a Blue Angel last night for $10 |
22:51.02 | cr2 | Jax184: which one ? |
22:51.28 | Jax184 | it's branded by audiovox, and was sold through Telus |
22:51.29 | pH5 | hm. for $10 I would take one too :) |
22:51.41 | Jax184 | so it's CDMA |
22:51.50 | pH5 | hey, an alpine? |
22:51.54 | cr2 | no. |
22:52.06 | pH5 | there are real cdma blueangels? |
22:52.15 | cr2 | Jax184: what is the baorid ? 4 ? |
22:52.23 | Jax184 | hmm? |
22:52.24 | pH5 | cr2: did you try replacing the memory accesses in htc-egpio with debug printfs? |
22:52.31 | cr2 | pH5: i think it's called harrier, not blueangel. |
22:52.38 | Jax184 | that's what I'm thinking |
22:52.41 | cr2 | pH5: i'll try. |
22:52.46 | Jax184 | what I've read so far is a touch confusing |
22:52.51 | pH5 | ah, harrier. I never heard of that, know the name only from the bootloader. |
22:52.54 | *** join/#htc-linux TimRiker (n=timr@pdpc/supporter/bronze/TimRiker) |
22:53.18 | cr2 | Jax184: the boardis is probably 4. i've heard that it actually works. |
22:53.46 | cr2 | the lcd part on 4 is the same as on 5. i.e. no weird spi stuff. |
22:54.00 | Jax184 | I took it apart last night |
22:54.24 | cr2 | Jax184: do you have lcd ? :) |
22:54.56 | Jax184 | I didn't open up the screen casing, since I don't have a T4 torx driver |
22:55.29 | cr2 | t5 |
22:55.45 | Jax184 | well, I don't have a T5 any more than I do a T4 |
22:55.48 | cr2 | i know because i've replaced the lcd on mine. |
22:55.51 | Jax184 | the smallest I have is T6 |
22:56.03 | Jax184 | anyway, I think you're right about the board |
22:56.16 | cr2 | can you boot linux on it ? |
22:56.30 | Jax184 | no idea |
22:56.39 | Jax184 | haven't tried |
22:56.40 | cr2 | well, try. |
22:56.43 | cr2 | :) |
22:56.59 | *** join/#htc-linux dcordes (n=user0815@f048226146.adsl.alicedsl.de) |
22:57.49 | Jax184 | I know the upgrades to WM6 and WM5 state that they absolutly do not work on the harrier |
22:59.23 | cr2 | because of cdma i guess. |
22:59.31 | Jax184 | yeah |
23:00.52 | cr2 | pH5: gpio 357, is it realistic ? |
23:02.45 | cr2 | for cpld1 |
23:02.55 | cr2 | will check cpld2 now |
23:04.46 | pH5 | cr2: no. what is the gpio number you use in the gpio_set_value call, and what number is .gpio_base in the struct htc_epgio_platform_data? |
23:05.59 | pH5 | you have to decide the gpio numbers, and with the current hh.org implementation the gpio chip's gpio numbers should probably start at 0x100 and 0x200, respectively |
23:06.09 | cr2 | 525 for cpld2. |
23:06.30 | cr2 | yes, that's the numbers i use. |
23:06.55 | cr2 | i can check 357-256, 525-512 |
23:08.56 | pH5 | hm, ok. you have that huge number of registers on cpld1, and with the fake 16bit access leaving half of the gpio numbers empty you get up to 357. |
23:09.17 | cr2 | 256+16*6+5 |
23:09.19 | pH5 | so what register access does htc-egpio translate this to? |
23:09.35 | pH5 | cr2: yup, that's why I want to have an option for the 8bit chips. |
23:09.40 | cr2 | and 512+13 |
23:09.53 | cr2 | 525 |
23:10.00 | pH5 | this is ok |
23:10.01 | cr2 | cpld2 is ok. |
23:11.04 | cr2 | yes. and 357 for cpld1 |
23:11.18 | cr2 | but they obviously point to some wrong location. |
23:12.02 | cr2 | how i can debug that ? |
23:12.12 | cr2 | i'll look at /proc/ioports |
23:12.46 | cr2 | usb keyboard is quite handy. |
23:13.55 | dcordes | usb keyboard on htc? |
23:13.58 | pH5 | cr2: well, you have configured htc-egpio with the cpld base addresses, so htc-egpio calculates (for 256+16*6+5) the register address as u16 base[6<<bus_shift] |
23:14.16 | dcordes | I thought htc usb hosts were unpowered |
23:14.49 | pH5 | dcordes: not on athena it seems :) |
23:15.16 | cr2 | but with 250mA limit it still rejects my stick. |
23:15.26 | cr2 | will put 500 ;) |
23:16.58 | dcordes | usb mass storage one? |
23:18.45 | cr2 | yes. |
23:18.52 | cr2 | the keyboard/mouse is ok. |
23:19.05 | Jax184 | I still want to do something cool with this harrier |
23:19.06 | cr2 | w228x.0 is ok |
23:19.20 | cr2 | atiw_mmc is ok. |
23:19.32 | cr2 | htc-egpio.0 is ok |
23:19.33 | dcordes | amazing that usb stick pulls more than 250mA |
23:19.49 | cr2 | htc-egpio.1 is ok |
23:20.14 | cr2 | dcordes: it may be some kernel bug. but if i connect the stick to a powered hub, it works. |
23:20.45 | dcordes | w or w/o power supply on the hub? |
23:21.14 | dcordes | I had similar effect with my zaurus. Usb hub tells the host that there is enough currency although it has no power at all |
23:21.37 | cr2 | dcordes: with power |
23:21.58 | dcordes | there is a way to remove the currency check completely. XorA from oe once made me a kernel patch for this |
23:22.09 | cr2 | zaurus is the only pxa device that sets the current limit |
23:22.22 | dcordes | but I think it's dangerous. That's why I got a hub for external power needs |
23:22.43 | cr2 | pH5: no cpld interrupts in /proc/interrupts ? |
23:22.59 | dcordes | I believe that. from user power point of view I can tell I'm very satisfied with the zaurus |
23:23.07 | dcordes | -power |
23:23.11 | pH5 | cr2: /proc/interrupts only shows requested irqs |
23:23.17 | cr2 | pH5: ok, the htc-egpio does not register itself. ok. |
23:24.39 | cr2 | pH5: the cpld memory areas are those which i requested. why can an access fail ? |
23:26.03 | cr2 | going to blow athena with a 500mA limit :) |
23:28.51 | dcordes | what do you run on it? |
23:28.53 | cr2 | heh. with 500mA it works at boot. |
23:29.01 | cr2 | dcordes: ? |
23:29.07 | dcordes | operating system |
23:29.25 | cr2 | wince5 |
23:29.38 | cr2 | i'm not reflashing anything. |
23:29.55 | cr2 | i always use factory installed software :) |
23:30.13 | cr2 | unlike goxboxlive :) |
23:31.22 | dcordes | is there a way to exactly dump the interal flash of htcs? |
23:31.47 | dcordes | so in case of rma you can get them back to factory installed software? |
23:32.33 | cr2 | it's doc g4 here. |
23:32.41 | cr2 | unsigned bit = gpio & GPIO_BASE_MASK; |
23:33.36 | cr2 | hmm. where is GPIO_BASE_MASK |
23:33.59 | pH5 | that's gpiodev stuff |
23:34.38 | cr2 | it's 0xff |
23:34.41 | cr2 | ok. |
23:35.26 | cr2 | so i must get 16*6+5 |
23:35.35 | cr2 | <PROTECTED> |
23:35.52 | cr2 | bus_shift=0 |
23:35.54 | cr2 | i hope. |
23:36.10 | cr2 | set it to 2 before ;) |
23:36.54 | cr2 | don't see where is can fail... |
23:37.31 | cr2 | .bus_shift = 0, |
23:37.41 | cr2 | .gpio_base = 1*GPIO_BASE_INCREMENT, |
23:45.50 | cr2 | no, it's not the htc-egpio fault. |
23:45.55 | dcordes | between did you find out anything about the exception thing on kaiser? |
23:46.20 | cr2 | happens somewhere deep in the not very transparent gpiodev stuff. |
23:46.43 | cr2 | dcordes: need Kevin2 for that. it looks like a bug in haret. |
23:49.38 | cr2 | pH5: gpio_get_value does not come to egpio_get |
23:50.13 | cr2 | the printk in egpio_get does not show up. |
23:50.36 | cr2 | i'd be surprised if it was different. |
23:53.31 | cr2 | i think things like that prevent the screen output |
23:53.34 | cr2 | <PROTECTED> |
23:53.35 | cr2 | <PROTECTED> |
23:53.52 | cr2 | because the fb is not yet initialized. |
23:54.35 | dcordes | cr2: Ok I hope he will be able to help with it |
23:55.04 | BabelO | cr2: try with EFB_DEBUG :) |
23:56.23 | cr2 | yes, commenting the BUG() helped :) |
23:57.47 | cr2 | hmm. strange |