00:04.43 | cr2 | yeah. call for uninitialized... |
00:04.45 | cr2 | wtf ? |
00:05.32 | cr2 | if (!get_value) |
00:06.26 | cr2 | i call platform_add_devices( devices, ARRAY_SIZE(devices) ); |
00:06.36 | cr2 | and after that want to get the gpio. |
00:06.48 | cr2 | devices are |
00:06.52 | cr2 | static struct platform_device *devices[] __initdata = { |
00:06.52 | cr2 | <PROTECTED> |
00:06.52 | cr2 | <PROTECTED> |
00:08.10 | *** join/#htc-linux the_sys0p` (i=the_sys0@gateway/tor/x-ee45bfac8df573cd) |
00:08.12 | cr2 | why does it work for magician,hx4700, apache, hermes, but not for me ? |
00:11.42 | cr2 | well, but i need to call it in machine_init. there is no other way... |
00:12.03 | cr2 | fscking generic gpio api ;) |
00:12.26 | BabelO | cr2: so cpld still a problem ? |
00:13.01 | dcordes | Good fight, good night ppl |
00:13.03 | cr2 | it seems that i'll be forced to remap the gpios and pick them directly and then unmap. |
00:13.27 | cr2 | BabelO: it seems that the driver is initialized too late. |
00:14.01 | BabelO | cr2: buit in ? |
00:14.08 | cr2 | yes. |
00:14.23 | cr2 | why do i need the silly api, if i can't use it anyway ? |
00:15.10 | BabelO | hmmm i need to set some default button shortcut in qtopia |
00:15.16 | BabelO | doh ! |
00:15.58 | cr2 | BabelO: if i'll find how to disable the ati powersave timer, i can check qtopia on athena. that will be a cool screenshot. |
00:16.26 | BabelO | cr2: yes ;) with roadmap, so good |
00:16.41 | cr2 | yeah. 5" vga LCD. |
00:17.38 | cr2 | BabelO: i've turned my local kernel tree into a real mess :) |
00:18.52 | BabelO | cr2: yes me too with artemis... can't do any patch now, it take me two days to redo all in separate tree |
00:19.22 | cr2 | i've patched a lot and have created many new drivers. |
00:19.38 | BabelO | save your data :) |
00:19.46 | BabelO | i use gmail to do that ... lol |
00:19.56 | BabelO | i hope they have a good backup |
00:20.42 | cr2 | #else /* start early for dependencies */ |
00:20.42 | cr2 | subsys_initcall(egpio_init); |
00:20.42 | cr2 | #endif |
00:20.55 | cr2 | seems to be not early enough ;) |
00:24.33 | cr2 | arch_initcall ? |
00:28.53 | cr2 | no, does not help. |
00:29.05 | cr2 | Kevin2: your advice is needed. |
00:29.20 | cr2 | ok. good night. |
00:29.34 | BabelO | good night |
00:43.58 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
00:53.52 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
01:12.50 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
02:01.24 | *** join/#htc-linux the_sys0p` (i=the_sys0@gateway/tor/x-21612c7a2cd8b8a8) |
03:18.57 | *** join/#htc-linux ltxda (n=ltxda@c-98-196-125-205.hsd1.tx.comcast.net) |
03:24.36 | *** join/#htc-linux dcordes|1zZzzZz (n=user0815@f054212115.adsl.alicedsl.de) |
04:06.07 | Kevin2 | Hi |
04:09.29 | Jax184 | Rah |
04:09.37 | Jax184 | someone here has to know |
04:09.50 | Jax184 | how does the camera in the Blue Angel interface? |
04:10.03 | Jax184 | it wouldn't happen to be something nice and standard like USB, would it? |
04:12.24 | *** join/#htc-linux Zoolooc_ (n=fredsiba@p549525EF.dip0.t-ipconnect.de) |
04:14.05 | Kevin2 | Bah! That would take all the fun out of it. |
04:14.57 | Jax184 | I'm just trying to think of ways to make my Harrier useful |
04:15.23 | Jax184 | if I could replace the utterly useless camera with a USB wifi device, it would be a big improvement |
04:15.38 | Jax184 | then I wouldn't have to choose between a net connection and storage space |
04:16.07 | Kevin2 | Most camera chips don't use usb. You can run haret and watch the interrupts to find out what harrier is using. (Or search the wiki, someone has already probably documented it.) |
04:17.03 | Jax184 | also, I haven't installed linux on this thing since everything I've seen so far says no one gives a damn about supporting the CDMA Harrier |
04:18.29 | Kevin2 | Haret is a wince app. |
04:19.07 | Jax184 | oh? |
04:21.07 | Kevin2 | http://www.handhelds.org/~koconnor/haret/haret-20071122.exe |
04:24.28 | Jax184 | would there be a set of instructions to go with this? |
04:25.45 | Kevin2 | It's kind of light on documentation. |
04:26.13 | Jax184 | well then it's not of much use to me as it is |
04:27.40 | Kevin2 | You can read through the stuff at http://www.handhelds.org/moin/moin.cgi/HaRET |
04:28.50 | Jax184 | well that doesn't tell me a thing |
04:41.26 | Jax184 | I've had enough of this thing for tonight |
04:42.25 | Jax184 | want to buy a Harrier? |
04:42.25 | Jax184 | hehe |
05:06.17 | *** join/#htc-linux GPFerror (n=GPF@76.187.41.132) |
05:30.52 | *** join/#htc-linux goxboxlive (n=jrs@195.159.97.196) |
05:31.26 | goxboxlive | good morning |
05:35.50 | ellisway | mornin goxboxlive |
05:36.14 | goxboxlive | hi ellisway, up early too |
05:36.19 | ellisway | yeh |
05:36.33 | ellisway | just trying your newest qtopia image |
05:37.53 | ellisway | goxboxlive what would i have to do to add helix ? |
05:38.24 | ellisway | would i have to compile my own image or can i just compile helix ? |
05:38.53 | goxboxlive | ellisway: You have to build Qtopia youself. Its lame i think, but Helix is only for developer issue, and are buidlt with single linece AFAIK |
05:39.10 | goxboxlive | ellisway: it's enough just tp compile Qtopia |
05:39.11 | ellisway | that sucks |
05:39.53 | ellisway | still waiting on your current image to finish booting its just starting qtopia |
05:39.57 | goxboxlive | Yes totaly. There is an workaround, and that is to se gstreamer instead of helix, but i havent manage to build it. |
05:40.34 | goxboxlive | ellisway: i uploaded a new Openmoko image yesterday, it's quite impressive now. But still a little slow compared to Qtopia. |
05:41.29 | goxboxlive | ellisway: was it you that offer me to build Qtopi and other images at a powerfull machine? |
05:42.03 | ellisway | i have a few dual xeon boxes that could be used for building on |
05:42.29 | ellisway | but there all in a private non routed vlan atm |
05:42.40 | goxboxlive | ellisway: ok |
05:43.39 | ellisway | not sure which distros are on some of the servers either atm |
05:43.57 | ellisway | i think will probably be a slackware based distro |
05:44.15 | goxboxlive | ok |
05:44.42 | ellisway | ok well your current qtopia image still gives me the non network thing with it saying please wait on the dialer |
05:45.07 | goxboxlive | ellisway: a friend of mine had the same problem |
05:45.18 | ellisway | oh |
05:45.46 | ellisway | i`m fairly sure that its not down to the radio rom aswell now |
05:46.03 | ellisway | as i`ve had the same thing with 1.18 1.15 and 1.10.03 |
05:46.10 | goxboxlive | ellisway: no i dont think so either. Wich radio rom are you using? |
05:46.31 | ellisway | on the device i`ve just tried it has 1.10.03 |
05:46.41 | ellisway | tmobiles current uk image |
05:47.35 | ellisway | i`m gonna try the openmoko image |
05:47.47 | ellisway | just copying it to sd card |
05:48.08 | goxboxlive | i c, i have no idea what it might be. You could try to dump the atdb.log and send it to me and hen i can compare it with mine to see if your device is sending som eother AT chats than mine is. |
05:48.49 | ellisway | you`d have to guide me through |
05:49.00 | ellisway | but i`ll try the moko image quickly 1st |
05:49.10 | goxboxlive | i'll check, first you need to get haret |
05:49.50 | ellisway | Kevin2`s 1 ? |
05:50.12 | ellisway | [04:20] <Kevin2> http://www.handhelds.org/~koconnor/haret/haret-20071122.exe |
05:50.55 | goxboxlive | yes thats good enough. |
05:51.28 | ellisway | downloaded |
05:51.29 | goxboxlive | I dont have the howto here at work, you have to wait to i get home, or until cr2 gets on. He know how to do it. |
05:51.41 | ellisway | np |
05:51.53 | ellisway | i`ll be about this evening |
05:52.10 | ellisway | will even try and get 1 of the xeons onto a public vlan in a bit |
05:52.21 | ellisway | if you want somwhere to build |
05:53.00 | ellisway | cant remember what the boot drive is on the box i`m thinking about i no the storage is 3 x 500gb sata hdd`s in raid 5 |
05:58.07 | ellisway | nice system startup sou nd from moko :) |
05:58.45 | goxboxlive | ellisway: Do not waste your time with a lot of work setting up a build machine. I thought is was ready. beacuse i was just curioues how fast Qtiopia where able to build. My own machine builds a Qtopia image in about three hours. But Trolltechs buildmachines (15) builds a image in 15mins. :-) |
05:59.22 | ellisway | damn |
05:59.48 | ellisway | theres a few dual xeons running but nothing thats setup with a build enviroment as such |
06:00.06 | ellisway | well not cross compile anyway |
06:01.15 | ellisway | i need to get 1 setup for building on just find it a tad confusing when reading through some of the quides on getting oe setup etc |
06:02.30 | goxboxlive | ellisway: A crosscompiler is builded by Openembedded it self. But, there are a lot of applications and libs that needs to be installed to have it running |
06:03.06 | goxboxlive | ellisway: http://www.openembedded.org/wiki/RequiredSoftware |
06:03.26 | goxboxlive | and http://www.openembedded.org/wiki/OEandYourDistro |
06:06.01 | Zoolooc_ | if you use gentoo, you can compile yourself almost any cross-toolchain by just entering a commnad line and wait for it to compile |
06:07.56 | ellisway | i`ll have to wait till i get 1 of the boxes online before i can do a build anyway |
06:08.32 | ellisway | but i think most of the required stuff is in slackware by default and probably in most other distro`s by default aswell looking at it |
06:11.59 | ellisway | well the moko image connects to the network without issue |
06:12.46 | ellisway | and rings when i ring the phone |
06:16.00 | *** join/#htc-linux sadeness_ (n=LamersIn@nat1.rlan.ru) |
06:24.16 | ellisway | hmm thats another thing i notice with the qtopia images the rtc always gets set to 1997 but the moko image botted with the correct date |
06:35.19 | *** join/#htc-linux _law_ (n=law@mail.stiftadmont.at) |
07:08.31 | *** join/#htc-linux |dion| (n=dion@inhex.net) |
07:28.22 | *** join/#htc-linux rob_w (n=bob@Mb95f.m.pppool.de) |
07:34.23 | goxboxlive | ellisway: Yes, the date is screwed up in all images beside oipenmoko. I think i figured out how thhey are doing it yesterday so with my next Qtopia image i will have it working i think. But i think it sets the date/time/year when the image where build, not when it's booted at frist boot. |
07:35.04 | goxboxlive | And next time it sets the time when it was booted last time. But at least it will give us tge right year :-). |
07:37.05 | *** join/#htc-linux rmoravcik (n=moravcik@gtsgw.ttc.cz) |
07:47.46 | ellisway | k |
08:00.35 | *** join/#htc-linux kiozen (n=oeichler@84.146.163.103) |
08:05.01 | *** join/#htc-linux zoolooc (n=lucian@84.149.37.239) |
09:09.55 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
09:29.52 | *** join/#htc-linux Foxdie (i=Foxdie@77-98-48-193.cable.ubr05.harb.blueyonder.co.uk) |
09:33.16 | *** join/#htc-linux rob_w|laptop (n=rob_w@p549B8C8A.dip0.t-ipconnect.de) |
10:49.24 | cr2 | goxboxlive: can you explain that to me ? http://wiki.xda-developers.com/index.php?pagename=UniversalBatteryIssue |
10:56.43 | goxboxlive | hold on |
11:08.45 | cr2 | goxboxlive: i'll be doing something with the last magenta line here http://www.handhelds.org/moin/moin.cgi/UniversalStatus |
11:09.17 | cr2 | sdio on asic3, because i have some discrepancies with the ati sdio. |
11:10.05 | cr2 | and the test mdoc g3 driver is available for tornado, so if you are brave enough, you can test it on the universal :) |
11:11.06 | cr2 | but it 's a character, readonly driver though. |
11:17.53 | goxboxlive | cr2: I have tried all that, i had to buy a new battery two times |
11:18.11 | goxboxlive | yues, change the wiki if there are something new or upodated |
11:19.15 | goxboxlive | cr2: Can i break it? |
11:19.40 | cr2 | unlikely. |
11:19.49 | goxboxlive | do you dare to try it? |
11:20.04 | cr2 | looks very similar to what wince does in the bootloader. |
11:20.15 | cr2 | but not 100% exactly. |
11:21.08 | cr2 | but it'll be a good answer on the question if you can really backup the mdoc. |
11:21.14 | goxboxlive | So, what do i have to do to test it? Bu do i have to put a linux image on the devoice, or may i still run from sd card? |
11:21.47 | cr2 | run from sd, compile the driver as a module, insmod it, use dd or cat. |
11:21.57 | cr2 | bbl |
11:22.51 | goxboxlive | so do not replace anything then. I will be able to boot wince again? |
11:59.32 | *** join/#htc-linux AC_Mast (i=s@gateway/tor/x-ee614e086f6cc100) |
12:09.45 | cr2 | goxboxlive: the only "irreversible" thing is to destroy the ipl and spl. |
12:10.11 | cr2 | and even that can be resolved with some soldering. |
12:12.48 | goxboxlive | cr2: ok, does the driver exist in the hh.org kernel source? |
12:13.21 | goxboxlive | bbl, going home from work. |
12:13.23 | *** part/#htc-linux goxboxlive (n=jrs@195.159.97.196) |
12:14.19 | dcordes|1zZzzZz | cr2: moin- seen Kevin2 yet? |
12:17.16 | cr2 | dcordes: no. |
12:17.45 | dcordes | is there anything I can do now w/o the possible haret patch? |
12:18.44 | cr2 | can you try to read from the other mapped areas ? |
12:18.54 | cr2 | reading gpio was possible. |
12:19.11 | cr2 | this will help to narrow the search. |
12:19.24 | dcordes | from Titan memory map? |
12:19.42 | cr2 | yes. the non-RAM mappings. |
12:29.14 | dcordes | cr2: are zeros at the end of a memory address relevant? |
12:34.20 | dcordes | set permissivemmutrace 1; addlist mmutrace 0xa9a01000 1; wi 30; is active when I touch and move around stylus on the screen |
12:37.35 | dcordes | but also when I do other things like writing something with word mobile. maybe it is really wifi related because it seams to react on all kind of activity and haret is communicating over wifi |
12:39.37 | dcordes | at least the comment says it is uart wifi something |
12:46.36 | cr2 | "uart" comes from android kernel. wifi is one of the sdio controllers, don't know what does it have to do with this area. |
12:47.12 | dcordes | pastebin the output? |
12:51.39 | cr2 | ok. |
12:52.36 | *** join/#htc-linux goxboxlive (n=goxboxli@81.80-202-132.nextgentel.com) |
12:52.45 | cr2 | lol. |
12:52.47 | *** join/#htc-linux Marex-notebook (n=marex@vasut.kolej.mff.cuni.cz) |
12:52.49 | cr2 | Applying patch 'disable_java.patch' |
12:53.43 | dcordes | cr2: http://pastebin.ca/794791 |
13:03.19 | cr2 | 1. addlist mmutrace 0xa9a01000 1 |
13:03.35 | cr2 | i doubt you can trace 1 byte, but ok. |
13:03.46 | cr2 | all you can see there are the interrupts. |
13:05.05 | cr2 | probably one of them is from the SDIO slot, where you have your wifi. |
13:05.14 | dcordes | hm |
13:09.01 | cr2 | don't do 'wi' |
13:09.33 | cr2 | try just to do 'pd 0xphys 4' for each starting physaddress |
13:09.40 | dcordes | I just tried to do what we did when we looked for touchscreen |
13:10.01 | dcordes | ok |
13:10.04 | cr2 | wifi is not a big problem to trace. |
13:10.30 | dcordes | cr2: but it's always active because I run the telnet through it |
13:10.40 | cr2 | it's easier to make it run on pxa first. |
13:10.44 | cr2 | ok. |
13:11.09 | cr2 | you can login over usb. if it will work ;) |
13:11.31 | dcordes | what do I need on the host? |
13:11.33 | cr2 | if bluetooth. |
13:11.52 | dcordes | bluetooth? |
13:11.52 | cr2 | rndis needs a patch afair. |
13:12.28 | cr2 | since activesync can work over BT, you'll be able to telnet over bt too. |
13:13.08 | dcordes | I have some bt1 usbstick in some drawer |
13:13.20 | dcordes | but what about usb? It needs patching on the host? any documentation? |
13:13.32 | dcordes | Bus 004 Device 007: ID 0bb4:0b0b High Tech Computer Corp. |
13:13.49 | dcordes | [56203.102020] eth1: register 'rndis_host' at usb-0000:00:1a.7-3.1, RNDIS device, 80:00:60:0f:e8:00 |
13:21.32 | cr2 | yes, but afaik it does not work winth wince5 as expected. |
13:21.54 | cr2 | does not work for me, and even crashes wince on hermes. |
13:22.39 | cr2 | http://www.synce.org/index.php/Connecting_your_Windows_Mobile_2005_device_via_USB_(usb-rndis-lite) |
13:23.55 | dcordes | don't I have wince6 on kaiseR? |
13:26.05 | dcordes | oh CE OS 5.2.1620 Build 18125.0.42 |
13:26.42 | dcordes | you pay that much for crapsoftware and even get obsolete version |
13:33.46 | cr2 | there are no wince6 devices afaik. |
13:34.22 | dcordes | ok ^^ |
13:34.39 | kiozen | wince5 or better -> linux |
13:35.46 | cr2 | m$ want you to buy a 128MB device to run "real" wince6. it's an endless upgrade money grubbing cycle. software plays only secondary role. |
13:36.40 | cr2 | kiozen: i've hit a big problem with egpio yesterday. but it hopefully does not affect n560. |
13:37.12 | kiozen | what is it? |
13:37.28 | cr2 | it's initialized too late in the kernel boot process. |
13:38.40 | kiozen | hm, any official way to change the order? |
13:39.57 | cr2 | dcordes: there are 63 >1Mb mappings. how much of them are readable ? |
13:40.36 | cr2 | kiozen: yes, but the side effects need to be considered. |
13:41.29 | kiozen | thus how good are the chances I have got a 440.-€ wince only device? |
13:46.16 | cr2 | wince-only is impossible. |
13:46.52 | cr2 | i'll do the screen, and will switch to making the SD boot reliably. |
13:47.29 | *** join/#htc-linux the_sys0p (i=the_sys0@gateway/tor/x-f6ae236ec2e0cf9f) |
13:48.13 | dcordes | cr2: what do you mean by > 1mb mappings? |
13:49.57 | cr2 | dcordes: the non-RAM entries in the list. |
13:51.23 | dcordes | no clue |
13:51.41 | dcordes | how to identify them? |
13:52.39 | dcordes | please have mercy with me and be a bit more specific with the tasks. I'm completely new to all this |
13:55.52 | goxboxlive | cr2 there? |
13:56.03 | cr2 | dcordes: the second column contains the base physical addresses of "mappings" |
13:56.07 | cr2 | goxboxlive: yes. |
13:56.21 | goxboxlive | cr2: is the mdoc driver in the kernel source? |
13:56.47 | cr2 | dcordes: so pick them out, remove the RAM areas (0x0, 0x10000000, 0x12000000) and write a script. |
13:57.17 | cr2 | goxboxlive: no, but i'll add it today. to the universal directory. |
13:57.24 | goxboxlive | ok |
13:57.39 | cr2 | dcordes: i'm at work too, and just using a free minite to write something here :) |
13:58.24 | dcordes | cr2: I was assuming that and really appreaciate you still help |
13:59.17 | dcordes | I copied the whole table to oo calc. The first column is completely irelevant? Which of the second column are to be "picked out" ? |
14:00.13 | cr2 | the 'phys' column, the entries without + at the beginning. |
14:00.55 | Kevin2 | Hi |
14:01.49 | cr2 | hi Kevin2 |
14:02.23 | cr2 | Kevin2: we have a problem on msm7200. the read from a mapped area gives an exception. |
14:02.36 | dcordes | cr2: ok so I a make a list of all fields from the second column w/o the + eg 0xa8600000 but not +0x3c , +0x40 , +0x54 etc? |
14:02.38 | cr2 | for some areas. |
14:02.47 | cr2 | dcordes: yes. |
14:03.14 | cr2 | dcordes: then just add 'pd ' before and ' 4' after. that'll be your script. |
14:03.25 | dcordes | great :) |
14:03.45 | Kevin2 | This one: (02:44:23 PM) cr2: 92100000 | aa200000 | 1MB section | D=0 AP=1 ?=2000 |
14:04.18 | cr2 | yes. |
14:05.09 | Kevin2 | Well, right away, we see that MSM is using an undocumented bit (the ?=2000) in the page tables. |
14:05.41 | Kevin2 | Can someone point me to the full "dump mmu" output along with the haretlog.txt file? |
14:05.45 | cr2 | what can it be ? |
14:06.04 | cr2 | dcordes: can you do it ? |
14:06.24 | cr2 | Kevin2: noexec ? |
14:09.01 | dcordes | cr2: sorry was busy with the "script" I came to 88 lines. You think that fits? |
14:09.23 | dcordes | cr2: didn I pastebin my dump mmu yesterday? |
14:09.28 | Kevin2 | cr2: I don't know. I think PXA had some bits for controlling access to the "mini cache". It could also just be sloppy wince programming and not properlly initializing the fields to zero... |
14:10.32 | dcordes | Kevin2: http://pastebin.ca/raw/793959 |
14:10.51 | dcordes | I think that's my dump mmu output |
14:12.58 | cr2 | 51c00000 | aa200000 | Tiny (1K) | AP=3 ?=80 |
14:13.03 | cr2 | looks strange too. |
14:13.09 | Kevin2 | dcordes: You have Titan or Kaiser? |
14:13.24 | dcordes | Kaiser. |
14:13.25 | cr2 | Kaiser. it's msm7200 |
14:14.21 | dcordes | cr2: http://pastebin.ca/794873 that's the script. Does it look right? How to execute it with console script? |
14:15.58 | *** join/#htc-linux BabelO (n=fcr@81.80.236.75) |
14:16.08 | cr2 | dcordes: replace xxxx with zeros. |
14:16.19 | dcordes | ok |
14:17.42 | dcordes | done |
14:18.00 | dcordes | now but it on kaiser and run it with haret's script launcher? |
14:18.37 | cr2 | is there a "run" haret command ? |
14:19.13 | dcordes | I mean the thing with Default.txt |
14:19.25 | dcordes | the one also used to run bootscripts |
14:20.00 | Kevin2 | Is "pd 0xaa210184 4" the command that fails? |
14:20.08 | cr2 | yes. |
14:20.24 | Kevin2 | dcordes: Can you run "show p2v(0xaa210184" ? |
14:20.31 | Kevin2 | show p2v(0xaa210184) |
14:23.44 | Kevin2 | dcordes: Also, can you create a file "earlyharetlog.txt", start haret, and then post the "haretlog.txt" file? (Or point me to one if you've already done this.) |
14:26.32 | dcordes | ok a second |
14:27.16 | dcordes | Kevin2: touch earlyharetlog.txt inside haretconsole dir? |
14:27.28 | cr2 | dcordes: yes. |
14:27.33 | dcordes | ok let's go |
14:27.34 | Kevin2 | No, in the same directory as haret on the phone. |
14:27.44 | dcordes | oh ok |
14:28.17 | dcordes | any way to do that in wince? |
14:28.27 | cr2 | use pword |
14:28.30 | dcordes | (creating empty file with .txt) |
14:29.11 | dcordes | cr2: where can I find that if any? |
14:29.21 | cr2 | pocket word |
14:29.28 | cr2 | text editor. |
14:29.47 | cr2 | type a space and save it to earlyharetlog |
14:29.51 | Kevin2 | dcordes: You can run "redir earlyharetlog.txt help" from haret. |
14:30.10 | dcordes | ok |
14:31.26 | cr2 | Kevin2: we already have 'ps' , it's time for an 'ls' implementation :) |
14:31.32 | dcordes | great it's there |
14:31.44 | dcordes | ps didn't work for me |
14:31.51 | dcordes | Unable to create tool help snapshot |
14:31.59 | cr2 | interesting. |
14:32.01 | dcordes | HaRET(4)# show p2v(0xaa210184) |
14:32.01 | dcordes | 0x92110184 |
14:32.25 | cr2 | yes, that's from the 1MB entry. |
14:33.21 | dcordes | is there smb on windows mobile? |
14:33.34 | dcordes | so I can use samba to access kaiser's filesystem? |
14:33.40 | dcordes | It's a pain to always use sd |
14:33.44 | cr2 | dcordes: can you check 'vd 0xb2110184 4' ? |
14:33.56 | dcordes | HaRET(5)# vd 0xb2110184 4 |
14:33.56 | dcordes | b2110184 |EXCEPTION while reading from address B2110184 00000000 | .... |
14:34.03 | cr2 | ok. |
14:34.43 | cr2 | i think you can mount it as the usb storage. maybe ;-) |
14:35.00 | dcordes | ok I will try |
14:35.33 | cr2 | if your rndis will be replaced with appropriate software on the wince side. |
14:36.08 | dcordes | http://pastebin.ca/794890 |
14:36.29 | kiozen | BabelO: ever compiled libssl for arm? |
14:36.29 | dcordes | Kevin2: that's earlyharetlog |
14:37.09 | Kevin2 | dcordes: Actually, I need the output from "haretlog.txt". |
14:37.28 | Kevin2 | When haret first runs, if it sees "earlyharetlog.txt", then it will create a log in "haretlog.txt". |
14:37.39 | dcordes | on the remote host? |
14:38.01 | cr2 | kiozen: isn't it all in oe ? |
14:38.01 | Kevin2 | No - in the same directory as haret.exe on the phone. |
14:38.38 | kiozen | oe -> crash boom bang, I am sick watching it compile and fail |
14:39.15 | dcordes | ah I see didn't rerun haret after I created the file... |
14:40.16 | cr2 | kiozen: NOTE: package bluez-libs-3.22: completed |
14:40.38 | cr2 | kiozen: i've started oe compilation in the background. |
14:41.30 | kiozen | I tried to compile gdal with it. First it stateded to compile the tools again and finally failed compiling gdal |
14:42.10 | dcordes | is there a way to umount in wince? Or do you always remove the sd uncleanly? |
14:42.22 | cr2 | i'll check after it'll build opie(tm) |
14:42.30 | dcordes | I don't want to reformat the fs all few days |
14:42.47 | cr2 | dcordes: wince does sync writes. like the dos floppies ;) |
14:43.40 | cr2 | dosfsck -y |
14:43.43 | dcordes | Kevin2: hope that's the right one now: http://www.pastebin.ca/794899 |
14:43.44 | cr2 | may help too. |
14:44.18 | dcordes | cr2: what does that mean - sync writes? |
14:44.40 | cr2 | dcordes: nothing is cached in ram. |
14:44.43 | dcordes | everything you copy is directly registred in fs? |
14:44.46 | dcordes | i see |
14:44.58 | cr2 | if you say write to disk, it's written. |
14:45.13 | dcordes | ok |
14:45.22 | cr2 | if you take out the card while writing, it's "your fault"(tm) |
14:45.34 | dcordes | hrhr |
14:46.32 | cr2 | kiozen: lol. compiling libopie(tmp) failed :) |
14:46.46 | kiozen | <evilgrin> |
14:46.51 | cr2 | wireless.h:646: error: '__s32' does not name a type |
14:47.17 | cr2 | tryinf gdal. |
14:47.44 | cr2 | i'm not interested in opie(tm). it was just a test. |
14:48.33 | cr2 | bitbake gdal |
14:48.37 | cr2 | omg. 1.3.2 |
14:48.50 | cr2 | i think the latest is 1.4.2 |
14:49.12 | dcordes | you need that version? |
14:49.30 | Kevin2 | dcordes: Can you try the following: Run "set PHYSMAPMETHOD 0" and then "show p2v(0xaa210184)" ? |
14:49.34 | kiozen | you see what I mean? |
14:49.53 | cr2 | kiozen: failed. |
14:50.33 | dcordes | Kevin2: unfortunately I have no idea what that means. but of course I try |
14:51.11 | Kevin2 | You can run "help vars" to see a description of all variables. The PHYSMAPMETHOD alters the way haret looks to resolve physical addresses. |
14:51.53 | dcordes | HaRET(1)# set PHYSMAPMETHOD 0 |
14:51.53 | dcordes | HaRET(2)# show p2v(0xaa210184) |
14:51.53 | dcordes | 0x002a0184 |
14:52.15 | cr2 | kiozen: configure: line 22175: syntax error near unexpected token `else' |
14:52.29 | kiozen | yes, me too |
14:52.45 | cr2 | kiozen: maybe change to .bb to compile 1.4.2 ? |
14:52.54 | dcordes | is WI=WIRQ? |
14:53.20 | kiozen | still eager to find out what it needs to compile the official packet |
14:53.56 | kiozen | CONFIG_LIBS is not set to the cross compiler lib dir but to /usr/lib64 |
14:54.11 | kiozen | try to learn and understand... |
14:55.34 | Kevin2 | dcordes: Yes. WI=WIRQ |
14:55.37 | dcordes | ok |
14:55.50 | dcordes | the HELP and help vars functions are useful |
14:55.56 | Kevin2 | Okay, can you run set PHYSMAPMETHOD 0 and then pd 0xaa210184 4 |
14:56.15 | Kevin2 | There is "help dump" too. |
14:56.48 | dcordes | Kevin2: I don't need to reset the session or anything to get rid of the previously set vars? |
14:56.54 | cr2 | kiozen: stuff/org.openembedded.dev/packages/gdal/gdal_1.4.2.bb, do_fetch |
14:57.06 | dcordes | cr2: NOTE: package gdal-1.4.2-r0: task do_compile: started |
14:57.16 | cr2 | NOTE: package gdal-1.4.2-r0: task do_compile: started |
14:57.32 | Kevin2 | Vars are set until you exit haret. |
14:58.17 | dcordes | Kevin2: that's what I mean- do I have to exit before I run the set PHYSMAPMETHOD 0 and then pd 0xaa210184 4 task? |
14:58.35 | Kevin2 | No. Run them both in the same session. |
14:58.38 | dcordes | HaRET(10)# pd 0xaa210184 4 |
14:58.38 | dcordes | aa210184 |EXCEPTION while reading from address 002A0184 00000000 | .... |
14:59.20 | dcordes | cr2: very good recipe since you only need to cp bb file to new {PV} |
14:59.30 | cr2 | dcordes: agreed. |
15:00.01 | dcordes | cr2: did you check for change in dependencies? |
15:00.02 | Kevin2 | Okay. This doesn't look like a haret issue. I'd guess the chip itself is raising the exception. |
15:00.11 | dcordes | hm is that a bad thing? |
15:01.00 | cr2 | dcordes: it's rather standalone. i think proj4 may be the only dependency, and it's very stable. |
15:01.17 | Kevin2 | I'll be back in a few. |
15:02.17 | dcordes | cr2: NOTE: package gdal-1.4.2: completed |
15:02.32 | dcordes | for armv5te |
15:03.56 | cr2 | dcordes: good. what package does angstrom have ? |
15:05.14 | cr2 | [ ] gdal-dev_1.3.2-r0_armv5te.ipk |
15:05.43 | kiozen | cr2: how do I get gdal_1.4.2.bb? mtn update said it's uptodate |
15:05.46 | cr2 | dcordes: how many cpus do you have :) |
15:06.25 | cr2 | kiozen: mv *1.3.2.bb -> 1.4.2.bb |
15:06.32 | kiozen | oh :) |
15:07.08 | dcordes | cr2: I have evil machine. model name : Intel(R) Core(TM)2 Duo CPU E6750 @ 2.66GHz with 4mb cache and 2gig ram |
15:07.28 | dcordes | cr2: there are packagin QA issues. We need to set install parameters in .bb |
15:07.39 | dcordes | I didn't get main ipk only dev and dbg |
15:07.57 | cr2 | yeah. stupid bitbake uses only 1 cpu on my dual opteron. |
15:07.59 | cr2 | ok. |
15:08.34 | dcordes | cr2: did you get the same errors? |
15:08.53 | dcordes | if so I can take care of it |
15:08.55 | cr2 | i'm waiting the compile to finish. |
15:09.06 | dcordes | I'm pretty sure you'll have the same issues |
15:09.14 | kiozen | cr2: did you try PARALLEL_MAKE = "-j2" in local.conf? |
15:09.32 | cr2 | i also run raid1. too much 250GB disks were dead before ;) |
15:09.39 | cr2 | i'll have a look. |
15:10.34 | cr2 | ERROR: QA issue: libsupc++.la failed sanity test (reference to workdir or installed) |
15:10.46 | dcordes | kiozen: is that required in order to make bitbake have use of both cpus? |
15:10.52 | dcordes | cr2: exactly |
15:11.45 | dcordes | cr2: do you know where those missing la files should go? Then it's no problem. |
15:11.53 | kiozen | well thw -j2 option will make the compiler use 2 CPUs the rest will still use one |
15:12.18 | dcordes | good to know |
15:12.37 | dcordes | so my bb always used one cpu for gcc lol |
15:12.47 | cr2 | also got dev and dbg |
15:13.21 | dcordes | so do you know where the la files should go on target machine? |
15:14.27 | cr2 | <PROTECTED> |
15:14.58 | cr2 | libgfortran was one of them. |
15:16.10 | kiozen | crash boom bang - no libcurl ... |
15:16.40 | cr2 | Kevin2: it's interesting that the CPU does not let to read the LCD controller settings ;) |
15:16.55 | cr2 | i'll check the other "forbidden" areas. |
15:17.21 | dcordes | as QA issues I get ERROR: QA issue: libstdc++.la, libmudflapth.la, libsupc++.la, libmudflap.la failed sanity test (reference to workdir or installed) |
15:18.59 | dcordes | FILES_${PN} += "${datadir}" <- that adds all files supposed to go to ${datadir} to the main package. But I don't know the variable for the dir where those files should go |
15:19.35 | cr2 | dcordes: it's a normal general library. |
15:19.38 | cr2 | like zlib |
15:19.47 | cr2 | of libpng |
15:19.50 | cr2 | or libtiff |
15:20.31 | dcordes | I think we both speak of a different ERROR |
15:24.04 | kiozen | funny for gdal it only uses 1 cpu, for curl & co it used 2 |
15:24.35 | dcordes | cr2: to get back to kaiser: How do I go on about the script? |
15:24.40 | dcordes | how to execute it |
15:24.46 | cr2 | help run ? |
15:27.21 | cr2 | "RUNSCRIPT <filename> [<ignoreNotFound>]\n" |
15:36.06 | cr2 | Kevin2: what does 'Tiny' page mean ? |
15:37.34 | cr2 | dcordes: btw, if you have a windows machine, you can try itsutils on kaiser. |
15:37.34 | Kevin2 | cr2: A 1K page instead of a normal 4K page. |
15:37.46 | cr2 | Kevin2: is it possible on arm ?? |
15:37.57 | Kevin2 | Yes. |
15:38.05 | cr2 | hm. interesting. |
15:39.07 | Kevin2 | Arm has large pages (64K) small pages (4K) and tiny page (1K). It also has the 1Meg section map. |
15:39.53 | cr2 | i'm surprised that wince may really use 1K pages. |
15:40.22 | dcordes | cr2: I have one windows machine |
15:40.39 | Kevin2 | It looks like they want to set some kind of extra bit in the page tables. The Tiny page descriptors have 4 extra bits. The small page doesn't have any extra room. |
15:40.57 | dcordes | cr2: runscript script.txt makes Kaise die and reboot. |
15:41.02 | cr2 | it may be some mechanism to prevent access to these areas. |
15:41.31 | cr2 | dcordes: remove the ram parts. |
15:41.45 | cr2 | i guess it did not like 'pd 0x0 4' |
15:42.09 | dcordes | http://pastebin.ca/794956 |
15:42.14 | dcordes | that's the script |
15:42.35 | dcordes | remove all 0x0* lines? |
15:42.39 | cr2 | remove everything before pd 0x80000000 4 |
15:43.07 | cr2 | because it's ram or something looking like ram ;) |
15:43.19 | dcordes | line 1-25? |
15:43.28 | cr2 | we know that it's different from titan. |
15:43.45 | cr2 | yes |
15:43.58 | dcordes | I can do that with word pocket hihi |
15:44.30 | Kevin2 | Maybe swetland would know what the extra page table bits do. |
15:46.17 | cr2 | good idea. |
15:47.02 | Kevin2 | One could also write to the page table to clear the bits and then try reading the data. |
15:51.18 | dcordes | cr2: http://pastebin.ca/794964 line 82=softreset |
15:51.28 | cr2 | yes, disable irq, store bit, clear bit, copy page, restore bit, enable irq. |
15:52.23 | cr2 | dcordes: remove all entries that cause exception. |
15:53.40 | cr2 | Kevin2: maybe some other bits are set to instruct the cpu to check these bits. |
15:54.23 | Kevin2 | cr2: Sure, but you can do it with haret commands. |
15:54.23 | cr2 | now it may be interesting to check titan. didn't pitaboy put the mmudump somewhere ? |
15:54.36 | Kevin2 | http://www.handhelds.org/moin/moin.cgi/HTCTitan |
15:54.56 | cr2 | Kevin2: i think you can' disable irq with haret commands. |
15:55.10 | cr2 | how will you see anything on your screen then. |
15:55.19 | Kevin2 | You don't need to disable irqs - just clear the bit - flush the tlb - and then dump the memory. |
15:55.27 | cr2 | ok. |
15:55.51 | Kevin2 | Of course, it may lockup the phone (or worse).. |
15:56.48 | *** join/#htc-linux law|home (n=_law_@213.173.86.202) |
15:57.12 | cr2 | btw, i had another problem yesterday. the htc-egpio is probed too late, so i can't read the gpio bits early enough. |
15:57.35 | Kevin2 | What needs to use the bits so early? |
15:58.02 | cr2 | i need to map the second cpld depending on the bit of the first cpld. |
15:58.22 | cr2 | and the second cpld keeps the board number. |
15:58.37 | Kevin2 | You can't map unconditionally? |
15:59.27 | cr2 | i can work around it, but just was unhappy that it happens. |
16:01.51 | Kevin2 | I understand your frustration - take a look at the htc-bbkeys driver. It needs to be archinit because the lcd init needs to use the gpios, but it can't init the input functionality because that subsystem isn't ready by that time. |
16:06.47 | cr2 | ok. |
16:18.41 | dcordes | cr2: http://pastebin.ca/794991 it kills kaiser as before with the same exception output- I think I did go wrong about removing the correct entries? |
16:18.51 | dcordes | could you tell me the lines to be removed? |
16:22.28 | cr2 | ok, the problem is more complex than i thought. |
16:22.53 | cr2 | btw, have you seen in your log that it's possible to read 0xaa200000 ? |
16:23.24 | cr2 | it also may be the difference between 7200 and 7200A ;-) |
16:24.03 | cr2 | we can't just assume the 1:1 equivalence. |
16:24.33 | cr2 | there are differences between titan and kaiser btw. for example in the CPLD position. |
16:26.41 | cr2 | since swetland has the datasheets, we may ask him to name the files appropriately |
16:32.25 | dcordes | I don't really understand :) |
16:36.58 | dcordes | I only know 7200 is my chipset, 7200A is swetland's. what do you mean by 1:1? similarity between 7200 and 7200A? what's CPLD? (position as in where it is in the memory map?) |
16:38.32 | cr2 | 7200 is different from 7200A. CPLD is the gpio extender used by htc. |
16:41.35 | dcordes | so the next step is to poke swetland for altering some filenames to make htc's CPLD usable with the android kernel? |
16:42.15 | kiozen | is there a way to tell bitbake to move all those files it stored in seperate directories into one single directory structure? |
16:44.31 | dcordes | kiozen: do you talk about do_package? |
16:44.53 | cr2 | kiozen: i think they are already. |
16:45.33 | kiozen | no the are all in their oen lengthy path with bin, lib, include and share at the end |
16:45.55 | cr2 | ~/stuff/tmp/staging/arm-angstrom-linux-gnueabi/include |
16:46.06 | cr2 | ~/stuff/tmp/staging/arm-angstrom-linux-gnueabi/lib |
16:46.18 | kiozen | one rootfs with bin, lib, include .. would be enought |
16:46.45 | cr2 | ~/stuff/tmp/staging/arm-angstrom-linux-gnueabi/bin |
16:47.02 | cr2 | if you need bin. |
16:47.18 | cr2 | i think for compiling the incllude and lib are enough. |
16:47.23 | kiozen | ah great, the docs pointed me to tmp/work.... |
16:48.21 | cr2 | tmp/work is the working directory |
16:48.58 | cr2 | and the toolchain is in ~/stuff/tmp/cross |
16:49.17 | cr2 | i've even managed to compile qtopia with this toolchain. |
16:49.28 | cr2 | some time ago. |
16:51.52 | kiozen | hm, always thought gdal needs proj4 |
16:53.52 | dcordes | cr2: anything I can do right now? |
16:55.49 | cr2 | dcordes: don't think so. |
16:56.38 | cr2 | go to suspend mode :) |
16:57.11 | dcordes | ^^ |
16:57.27 | cr2 | you have a new device without any docs. even htc needs one more year to make linux run on it :) and they have all docs. |
16:58.05 | dcordes | that does not sound any good for me |
16:58.14 | cr2 | dcordes: if you enjoy c programming and want to do something useful, also for kaiser, write the .bts compiler. |
16:58.43 | dcordes | I can make helloworld.c :D |
16:58.49 | cr2 | it's actually not so bad as it sounds. |
16:58.58 | dcordes | one year sounds really bad |
16:59.27 | cr2 | dcordes: they did not promise that linux will run on kaiser, btw. |
16:59.40 | dcordes | who? |
16:59.46 | cr2 | htc |
17:00.11 | dcordes | did they ever tell linux will run on any of their devices? |
17:00.33 | cr2 | ok, i've added the poweron to ati... |
17:00.42 | cr2 | dcordes: yes, they are in oha. |
17:00.50 | dcordes | oha? |
17:02.05 | cr2 | http://www.openhandsetalliance.com |
17:02.50 | kiozen | bbl |
17:03.46 | dcordes | cr2: which devices do they mention? |
17:08.10 | dcordes | cr2: do you mean Our participation in the Open Handset Alliance and integration of the Android platform in the second half of 2008 enables us toblablabla? |
17:08.12 | *** part/#htc-linux zoolooc (n=lucian@84.149.37.239) |
17:17.13 | *** join/#htc-linux rob_w (n=bob@Mb95f.m.pppool.de) |
17:19.18 | cr2 | dcordes: i've read something less buzzword-compliant which was more promising. |
17:19.52 | dcordes | lol buzzword compilant... |
17:20.15 | dcordes | you shouldn't habe said the one year thing. that were buzzword which ruin my day |
17:24.12 | cr2 | heh. htc is not like fic |
17:24.24 | cr2 | they really produce working hardware. |
17:25.35 | *** join/#htc-linux kiozen (n=oeichler@rgnb-4db03060.pool.einsundeins.de) |
17:28.12 | cr2 | Kevin2: i've created a compilable and working ati SD config now. with applying power and so on. it does not work, as you can guess :) but it's a good starting point. need to compile it for hermes though. |
17:28.40 | cr2 | kiozen: now returning to your LCD :) |
17:30.31 | kiozen | :) faster than one year |
17:30.31 | dcordes | cr2: so we cannot do anything at this time? |
17:31.31 | dcordes | hmm |
17:32.31 | *** join/#htc-linux ellisway (n=ellis@homealone.good1.com) |
17:32.53 | dcordes | would it be any useful to try booting a 7200A kernel? |
17:38.04 | cr2 | dcordes: if we will remove the LCD driver, replace it with vsfb, change the gpio stuff and remove other android drivers that may access nonexisting hardware, then we may boot. but what's next ? |
17:40.01 | cr2 | kiozen: pxa270 is a very well known CPU. |
17:40.03 | dcordes | cr2: don't know. just boot :D |
17:40.51 | cr2 | kiozen: i'll compile an new kernel now. |
17:42.07 | dcordes | cr2: maybe having a bootable kernel would pull other interested kaiser users? |
17:42.56 | cr2 | dcordes: htc wallaby has a bootable kernel too. |
17:43.55 | dcordes | since 2005? |
17:44.08 | cr2 | 2004 |
17:44.21 | cr2 | if not earlier. |
17:44.27 | dcordes | what does that tell me? |
17:44.35 | dcordes | return kaiser and get hermes? |
17:45.22 | cr2 | if you want to run linux, and don't care about the qualcomm gps, then yes. |
17:46.26 | dcordes | I was almost sure about hermes at the beinning. 128mb made me decide for kaiser |
17:46.59 | *** join/#htc-linux Zoolooc (n=fredsiba@p549525EF.dip0.t-ipconnect.de) |
17:47.17 | dcordes | wanting to run linux is of high importance for me |
17:47.30 | cr2 | gps and 128MB are on the + side. |
17:47.36 | cr2 | everything else is - |
17:47.44 | Kevin2 | I fear we still have a ways to go on hermes too. :-( |
17:48.37 | cr2 | Kevin2: i've already documented the w228x init. |
17:49.01 | *** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
17:49.02 | cr2 | and i'll do SD. |
17:49.45 | cr2 | don't promise 2D accel, but we don't really have it on w32x0 too in any stable form. |
17:50.12 | cr2 | Kevin2: what else do we have on hermes ? |
17:51.02 | *** join/#htc-linux rob_w (n=bob@Mb95f.m.pppool.de) |
17:52.07 | Kevin2 | My biggest concern is the ATI chip. There is so much functionality using it (graphics, SD, sound, etc) and we have no docs. |
17:53.10 | Kevin2 | I appreciate (and am highly impressed) by your efforts to RE it. It's just not a process that is going to happen in a short time period. |
17:53.37 | cr2 | yes, the I2S DMA may be very difficult. |
17:53.54 | cr2 | and we don't really target the camera. |
17:53.57 | goxboxlive | cr2: can you add the driver. |
17:54.17 | cr2 | goxboxlive: yes. i'll add all the athena changes too. |
17:54.39 | goxboxlive | good |
17:55.31 | cr2 | if i'll find how to disable the ATI power watchdog timer, then it will be usable as a pda. |
17:55.49 | goxboxlive | ok, good. |
17:55.54 | cr2 | and for watchdog only one register block comes to mind. |
17:56.11 | goxboxlive | Do you ever use your AThena, or is only collecting dust? |
17:56.37 | cr2 | i use it for gps |
17:56.55 | cr2 | still can't find the pin for the sim card. |
17:57.30 | cr2 | sable. that one i didn't touch for a long time. |
17:57.40 | goxboxlive | :-) |
17:57.56 | Kevin2 | cr2: In the logs, I read you were looking for the best way to decode registers (eg, pxa mem controller) with mainline haret. What I've been doing is adding the register definitions to haretconsole, and then do a "watch" on them. |
17:57.57 | cr2 | because it has weird ericsson modem. |
17:58.49 | cr2 | Kevin2: so how should i parse the bits ? |
17:59.16 | Kevin2 | cr2: I just describe each bit - see regs_pxa.py |
17:59.36 | cr2 | Kevin2: i need to grab some regs, and unstuff them. |
17:59.53 | Kevin2 | What does "unstuff" them mean? |
18:00.00 | cr2 | ok, these are the bit, but sometimes i need more tricky operations. |
18:00.30 | cr2 | ok, let's take the pxafb pixclock as an example. |
18:01.21 | cr2 | i need to call some 'mcr p14', get the lccr3, and do some math. |
18:01.34 | cr2 | should i do it in python ? |
18:01.39 | cr2 | which i don't really know, btw. |
18:01.40 | Kevin2 | I understand what you're saying - sometimes funky math is needed to get to the "real" pix clock. |
18:02.13 | Kevin2 | Accessing the coprocessor registers is no problem. But doing math on the results isn't currently supported. |
18:02.45 | Kevin2 | However, these "funky math" cases are pretty rare - shouldn't we document all the bits first, and then look to do the math for the user? |
18:03.19 | cr2 | Kevin2: i have no problem to port the c code. it's just a question if you'd like to have so much simple code in haret. |
18:04.13 | cr2 | it's actually the fastest solution. |
18:04.22 | cr2 | but it's obviously pxa-only. |
18:05.06 | Kevin2 | I'm okay with checking in pxa only code. If you want to port it over - go ahead. |
18:05.22 | cr2 | i've also one other weird candidate. htc AIC2/AIC3 aka pasic2/pasic3, but i'll port it today. |
18:05.31 | cr2 | ok. |
18:05.42 | cr2 | what about 'dump gpio' ? |
18:06.01 | Kevin2 | I do wonder, if it's worth making custom code for things we can do generically. |
18:06.25 | cr2 | it runs outdated pxa25x decoder now on all archs. |
18:06.34 | Kevin2 | Same thing with "dump gpio" - if you're willing to port the code over, I'm okay with checking it in. |
18:07.02 | Kevin2 | "dump gpio" should only be available on pxa. But yes, it only supports the first 86 gpios. |
18:07.03 | cr2 | yes, but the final result is that we discuss it for months, and when i need something done fast, i recompile gnu-haret |
18:07.11 | goxboxlive | Kevin2: IS thee a way we could add multiple image support for the easy installer'? Like, it looks for openmoko.rootfs.tar.bz2 and if it finds it it will make a openmoko-linux.img file. And if we have a Qtopia.rootfs.tar.bz2 it wil make a qtopia-linux.img file. With this we could have multiple linux systems on a partition. |
18:07.27 | cr2 | bbl |
18:07.36 | cr2 | need to buy some food. |
18:07.45 | goxboxlive | bon apetit |
18:09.02 | Kevin2 | goxboxlive: I think we should do that. I was thinking of passing in a parameter from haret on the command line - something like "distrofile=abc", and then have the installer only unpack abc.rootfs.tar.bz2, and have it unpack it to abc.img. |
18:11.23 | goxboxlive | Kevin2: Ok, so we choose that when we make the installer. |
18:11.58 | Kevin2 | goxboxlive: I think we should. The installer needs to have a kernel that matches what is in the rootfs - so we may as well explicitly bind them. |
18:12.29 | Kevin2 | I need to get the scripts into a source code repository somewhere.. |
18:12.50 | Kevin2 | You recommend the linuxtogo site? Do you know how I can get a repo? |
18:13.42 | goxboxlive | Kevin2: Isent there a way we can use the kernel in the boot folder of the rootfs, and instead have a common zImage for the installer? |
18:14.25 | goxboxlive | Kevin2: repo? You mean access to the linuxtogo/~htc-linux? |
18:15.08 | goxboxlive | I have allready given you that. I sent a common mail to you, ph5, cr2, babel0 and lkcl |
18:16.14 | Kevin2 | goxboxlive: Okay - I see your email. |
18:16.20 | goxboxlive | good |
18:17.45 | Kevin2 | Do you know if the linuxtogo people are okay with setting up source code control stuff? I'd like to setup a svn (or cvs or git or whatever) repository to hold the scripts that make the install initramfs. |
18:18.33 | goxboxlive | Kevin2: i have no idea, but florian who runs the site has a htc device hinself so think it's ok |
18:18.45 | *** part/#htc-linux Mineque (n=Mineque@195.60.65.50) |
18:19.13 | *** join/#htc-linux Mineque (n=Mineque@195.60.65.50) |
18:21.58 | Kevin2 | Okay. I'm going to create an account and see what I can do. Also, it looks like florian (and others) are in #linuxtogo - but I'll check on my own first. |
18:23.07 | goxboxlive | Kevin2: you could register a project. Everyone can di that |
18:23.14 | goxboxlive | s/di/do |
18:23.29 | goxboxlive | you just have to be registrated |
18:24.22 | Kevin2 | At some point, we're going to have to find a new home for the kernel cvs on handhelds.org. Since the mailing list went away everything has been in flux. It's been 4 months without a mailing list. |
18:24.42 | *** join/#htc-linux Marex-notebook (n=marex@gwfm10-3-250.802.cz) |
18:25.30 | goxboxlive | Kevin2: i have noticed. Seems like hh.org is fading out. |
18:26.26 | bd2 | GIT, please! |
18:28.37 | Kevin2 | bd2: Yes - I think git would be the natural choice. Maybe we can get the lkml people to give us one. Otherwise, maybe the linuxtogo people would be willing to setup a server. |
18:30.10 | bd2 | Kevin2, well... honestly, I think we need bunch of git trees. "MFD/SOC", "ADC", "HP iPAQs", "HTC", "all-in-one alike current -hh". ;-) |
18:30.39 | bd2 | though, htc/hps intersects.. |
18:31.13 | bd2 | ^^^ bunch of git trees if somebody wants to push things mainline. |
18:32.02 | bd2 | unfortunately, nobody from the mainline will ever agree to pull all-in-one tree. it's just very huge ;-) |
18:43.43 | dcordes | somebody knows a working ssh client for windows mobile? |
18:46.33 | Kevin2 | dcordes: If you find one, would you let me know? :-) |
18:48.33 | the_sys0p | There's a PuTTY for Windows Mobile, I believe. |
18:48.49 | dcordes | the_sys0p: pocketputty's keyboard didn't work for me |
18:49.02 | dcordes | I enter the host and it says can't connect to host §(&§($&)"§$"$= |
18:49.36 | the_sys0p | Hmm...It worked fine for me on my Wizard... I think I tried one other, but I don't remember what it was. :( |
18:49.36 | the_sys0p | Sorry. |
18:49.51 | dcordes | thanks anyways |
18:50.02 | dcordes | also tried different j2me ones |
18:50.05 | dcordes | no success |
19:04.33 | *** join/#htc-linux |dion| (n=dion@inhex.net) |
19:10.11 | *** join/#htc-linux pH5 (n=ph5@e178194088.adsl.alicedsl.de) |
19:11.15 | pH5 | good evening |
19:12.01 | dcordes | hey pH5 |
19:12.22 | pH5 | hej dcordes. how is the kaiser? |
19:13.07 | dcordes | my Kaiser is not feeling good. It wants 2.6.24. |
19:13.33 | pH5 | :) no news yet about the haret issue? |
19:14.15 | cr2 | hi pH5 |
19:14.44 | dcordes | pH5: there is. Kevin2 says it is not haret but msm7200 specific issue |
19:15.14 | cr2 | btw, what are our chances to update the hh.org kernel to a more decent version ? |
19:20.19 | cr2 | goxboxlive: can you create an installer for me if i'll upload the athena zimage ? |
19:26.44 | cr2 | goxboxlive: or i'll just commit my changes to CVS |
19:27.09 | dcordes | cr2: how do you make usb input devices work with your windows mobile machines? |
19:30.20 | dcordes | keeboard didn't work with powered hub either |
19:30.25 | dcordes | keyboard |
19:33.49 | Kevin2 | I got pocketputty to work. It's pretty slow though. |
19:34.04 | goxboxlive | cr2: yes of course, just send me the zImage |
19:34.12 | goxboxlive | hi pH5 |
19:34.36 | cr2 | goxboxlive: recompiling universal |
19:35.03 | goxboxlive | OPenmoko on magian is working real good now. Even neod is working. How did you solve the keys, i cant make it work in universal |
19:35.12 | goxboxlive | that one was for pH5 |
19:45.44 | cr2 | goxboxlive: added the docg3 module |
19:46.15 | goxboxlive | ok i'll test |
19:47.02 | goxboxlive | cr2: in next mail, attach the zImage :-9 |
19:47.10 | cr2 | already done. |
19:49.01 | Zoolooc | Kevin2: I too noticed pocketputty to be so slow, did you install the provided binary, or compiled it yourself? |
19:51.41 | cr2 | pH5: can i use gpio_set_value directly on the pxa gpios ? |
19:52.04 | goxboxlive | cr2: i had to compress it, google refused me to send a exe file |
19:52.11 | goxboxlive | just bunzip2 and utar |
19:53.38 | cr2 | goxboxlive: thanks. |
19:53.49 | goxboxlive | np |
19:54.56 | pH5 | goxboxlive: use EXTRA_OECONF_htcuniversal = "--with-platform-htc" |
19:54.57 | goxboxlive | cr2: what can i expect with the mdoc3 driver. What will happend when i load the modules |
19:55.14 | pH5 | or look at htc.patch in OE if that doesn't work and do something similar for universal. |
19:55.47 | goxboxlive | pH5: Where do i add this? EXTRA_OECONF_htcuniversal = "--with-platform-htc" In my local.conf? |
19:56.07 | pH5 | cr2: yes, and if the gpio nr. is const it is even optimized into a inlined ldr + and |
19:56.21 | cr2 | pH5: ok. |
19:56.23 | pH5 | goxboxlive: no, in packages/openmoko2/neod_svn.bb |
19:56.42 | goxboxlive | ok, i found the htc.pathc. I'll take a look at it. thanks |
19:56.50 | cr2 | goxboxlive: check modinfo first. you need to supply the base MDOC address. it's 0x0 on universal. |
19:57.21 | goxboxlive | cr2: how, show me exactly how i shall load the modules |
19:57.31 | cr2 | and it defaults to 16bit access now, i think. so it will be unhappy on the ba. |
19:57.46 | cr2 | kiozen: still here ? |
19:57.50 | kiozen | yup |
19:57.59 | kiozen | still compiling |
19:58.08 | cr2 | kiozen: i'll create a new kernel now. |
19:58.17 | cr2 | with LCD powerup. |
19:58.29 | kiozen | ok, heat up the SD card |
20:00.28 | goxboxlive | cr2: hwo do i supply the base MDOC adress? modprobe docg3 ????? |
20:00.58 | cr2 | goxboxlive: what does 'modinfo docg3' say ? |
20:01.22 | Kevin2 | Zoolooc: I just downloaded the binaries. |
20:01.48 | goxboxlive | cr2: i havent got that far yet, i had to compile on my laptop. It get hot :-) and slow |
20:03.26 | dcordes | also got Pocketputty working. But slowass |
20:04.59 | goxboxlive | pH5: I think magian and universal has diffrents keycodes. Thats why i use diffrents keyboard driver in Qtopia. So i guess i have to use another patch. But i'll check first |
20:05.55 | pH5 | goxboxlive: the current patch is just a hack to emulate gta01 buttons anyway. |
20:06.15 | goxboxlive | pH5: We shoudl expand it so we can have all our buttons working |
20:06.19 | pH5 | we should make use of the many buttons we have on our devices |
20:06.33 | goxboxlive | :-) i was faster than you:-) |
20:06.53 | pH5 | :) but I didn't want to spend too much time on this, not knowing whether or not neod would be scrapped next month and maybe replaced by something else |
20:08.15 | cr2 | goxboxlive: where should i put the .exe ? |
20:08.31 | goxboxlive | root of the sd card together with the rootfs image |
20:08.43 | goxboxlive | pH5: ok, i c |
20:09.09 | goxboxlive | cr2: Lets break my universal :-) |
20:09.11 | goxboxlive | root@htcuniversal:~$ modinfo docg3 |
20:09.11 | goxboxlive | filename: /lib/modules/2.6.21-hh18/kernel/arch/arm/mach-pxa/htcuniversal/docg3.ko |
20:09.11 | goxboxlive | description: Test modules for Diskonchip G3 device description |
20:09.11 | goxboxlive | author: Thomas Cougnard <tgnard@free.fr> |
20:09.12 | goxboxlive | license: GPL |
20:09.14 | goxboxlive | depends: |
20:09.16 | goxboxlive | vermagic: 2.6.21-hh18 preempt mod_unload ARMv5 |
20:09.18 | goxboxlive | parm: doc_config_location:Physical memory address at which to probe for DiskOnChipG3 (ulong) |
20:09.20 | goxboxlive | root@htcuniversal:~$ |
20:09.45 | cr2 | modprobe docg3 doc_config_location=0x0 |
20:10.06 | goxboxlive | [ 275.830000] Doc_G3: chip id=200 |
20:10.06 | goxboxlive | [ 275.840000] Starting flash stuff |
20:10.06 | goxboxlive | [ 275.850000] end |
20:10.13 | cr2 | good. |
20:10.30 | cr2 | hmm. what was the major/minor :) |
20:10.30 | goxboxlive | now what? |
20:10.42 | cr2 | pm.c:(.text+0x1d68): undefined reference to `__bad_udelay' |
20:10.44 | cr2 | wtf. |
20:11.15 | cr2 | 264 0 |
20:11.23 | cr2 | that's not very good. |
20:11.28 | cr2 | 254 0 |
20:11.38 | goxboxlive | what are you talking about? AThena? |
20:11.42 | cr2 | it's the same major/minor as mmc |
20:11.49 | cr2 | no, n560 |
20:12.07 | goxboxlive | Ok but what about universal. What did i get? |
20:12.09 | pH5 | cr2: "This function intentionally does not exist; if you see references to it, it means that you're calling udelay() with an out of range value." |
20:12.18 | goxboxlive | can i mount it or something? |
20:12.32 | cr2 | pH5: copied from htcapache.c :) |
20:12.38 | pH5 | max delay for udelay on arm is currently 2000us afaik. |
20:12.52 | pH5 | you should rather use mdelay for bigger delays, or better msleep. |
20:13.03 | cr2 | pH5: it it allowed to have block 254,0 and char 254,0 ? |
20:13.06 | cr2 | ok. |
20:13.38 | cr2 | don't know why Kevin2 used udelay(2000) . |
20:14.14 | pH5 | cr2: yes, block and char major/minor are separate. |
20:15.06 | goxboxlive | ok, removing module |
20:15.11 | cr2 | goxboxlive: mknod -m 600 /dev/mydocg3 c 254 0 |
20:15.25 | cr2 | and then |
20:15.44 | cr2 | cat /dev/mydocg3 > mdocdump |
20:16.13 | cr2 | kiozen: kernel ready. |
20:16.17 | goxboxlive | hmm it hangs |
20:16.52 | cr2 | goxboxlive: who hangs ? |
20:17.00 | Kevin2 | cr2: I think I copied the udelay from somewhere. |
20:17.02 | goxboxlive | i just get docg3_open and it hangs |
20:17.03 | kiozen | ok, btw what *bb is used to get libasound? package/alsa has no *bb |
20:17.26 | cr2 | Kevin2: lol. i've upped 2000 to 3000 in one place, so it's my fault. |
20:17.39 | goxboxlive | cr2: |
20:17.41 | goxboxlive | [ 632.960000] docg3_open |
20:17.41 | goxboxlive | [ 725.980000] Failed for 4000000 2000 |
20:17.41 | goxboxlive | [ 726.000000] docg3_release |
20:17.58 | pH5 | kiozen: that's OT ;) I think alsa-lib. |
20:18.01 | cr2 | goxboxlive: ok, you have 128MB g3. this driver is not smart. |
20:18.12 | goxboxlive | cr2: :-) |
20:18.19 | cr2 | goxboxlive: it can only 64MB |
20:18.22 | goxboxlive | should i unload the module? |
20:18.24 | goxboxlive | ok |
20:18.28 | cr2 | yes. |
20:18.37 | cr2 | what did you get in the dump ? |
20:18.38 | goxboxlive | That shoudl be easy to rewrite, or? |
20:19.00 | goxboxlive | cr2: i didnt dump anything |
20:19.12 | goxboxlive | how do i dump? |
20:19.18 | cr2 | goxboxlive: the main rewrite should be to filter out the ECC/EDC and to have a block device, so you can mount it. |
20:19.55 | goxboxlive | cr2: ok, but it should be easy to make it support 128MB rught? |
20:20.02 | goxboxlive | s/rught/right/ |
20:21.00 | cr2 | goxboxlive: try dd if=/dev/mydocg3 bs=512 of=mdocg3.raw count=2048 skip=4 |
20:22.01 | goxboxlive | cr2: it spam my screen |
20:22.03 | goxboxlive | [ 990.210000] Alignment trap: dd (1554) PC=0x00016538 Instr=0xe1c321d0 Address=0x00086ca4 FSR 0x013 |
20:22.03 | goxboxlive | [ 990.240000] Alignment trap: dd (1554) PC=0x0001641c Instr=0xe1ce21d0 Address=0x00086ca4 FSR 0x013 |
20:22.03 | goxboxlive | [ 990.270000] Alignment trap: dd (1554) PC=0x00016428 Instr=0xe1cc21f0 Address=0x00086ca4 FSR 0x813 |
20:22.03 | goxboxlive | [ 990.300000] Alignment trap: dd (1554) PC=0x00016534 Instr=0xe1c301d8 Address=0x00086cac FSR 0x013 |
20:22.03 | goxboxlive | [ 990.330000] Alignment trap: dd (1554) PC=0x00016538 Instr=0xe1c321d0 Address=0x00086ca4 FSR 0x013 |
20:22.04 | goxboxlive | [ 990.350000] Alignment trap: dd (1554) PC=0x0001641c Instr=0xe1ce21d0 Address=0x00086ca4 FSR 0x013 |
20:22.06 | goxboxlive | [ 990.380000] Alignment trap: dd (1554) PC=0x00016428 Instr=0xe1cc21f0 Address=0x00086ca4 FSR 0x813 |
20:22.08 | goxboxlive | [ 990.410000] Alignment trap: dd (1554) PC=0x00016534 Instr=0xe1c301d8 Address=0x00086cac FSR 0x013 |
20:22.12 | goxboxlive | [ 990.440000] Alignment trap: dd (1554) PC=0x00016538 Instr=0xe1c321d0 Address=0x00086ca4 FSR 0x013 |
20:22.14 | goxboxlive | [ 990.460000] Alignment trap: dd (1554) PC=0x0001641c Instr=0xe1ce21d0 Address=0x00086ca4 FSR 0x013 |
20:22.16 | goxboxlive | [ 990.490000] Alignment trap: dd (1554) PC=0x00016428 Instr=0xe1cc21f0 Address=0x00086ca4 FSR 0x813 |
20:22.18 | goxboxlive | [ 990.520000] Alignment trap: dd (1554) PC=0x00016534 Instr=0xe1c301d8 Address=0x00086cac FSR 0x013 |
20:22.20 | goxboxlive | [ 990.550000] Alignment trap: dd (1554) PC=0x00016538 Instr=0xe1c321d0 Address=0x00086ca4 FSR 0x013 |
20:22.22 | goxboxlive | [ 990.580000] Alignment trap: dd (1554) PC=0x0001641c Instr=0xe1ce21d0 Address=0x00086ca4 FSR 0x013 |
20:22.24 | goxboxlive | [ 990.600000] Alignment trap: dd (1554) PC=0x00016428 Instr=0xe1cc21f0 Address=0x00086ca4 FSR 0x813 |
20:22.26 | goxboxlive | [ 990.630000] Alignment trap: dd (1554) PC=0x00016534 Instr=0xe1c301d8 Address=0x00086cac FSR 0x013 |
20:22.28 | goxboxlive | [ 990.650000] Alignment trap: dd (1554) PC=0x00016538 Instr=0xe1c321d0 Address=0x00086ca4 FSR 0x013 |
20:22.30 | goxboxlive | [ 990.680000] Alignment trap: dd (1554) PC=0x0001641c Instr=0xe1ce21d0 Address=0x00086ca4 FSR 0x013 |
20:22.32 | goxboxlive | [ 990.710000] Alignment trap: dd (1554) PC=0x00016428 Instr=0xe1cc21f0 Address=0x00086ca4 FSR 0x813 |
20:22.34 | goxboxlive | [ 990.730000] Alignment trap: dd (1554) PC=0x00016534 Instr=0xe1c301d8 Address=0x00086cac FSR 0x013 |
20:22.36 | goxboxlive | [ 990.760000] Alignment tra |
20:22.38 | goxboxlive | and so on |
20:22.39 | cr2 | goxboxlive: don't spam. |
20:22.42 | goxboxlive | cr2: should i break it? |
20:22.44 | goxboxlive | stop it? |
20:22.50 | cr2 | yes. |
20:23.15 | cr2 | ok, then i need to look myself. |
20:23.24 | goxboxlive | ok, |
20:23.27 | cr2 | kiozen: do you want to get the kernel ? |
20:23.38 | kiozen | yes |
20:23.51 | cr2 | gmx ? |
20:23.59 | kiozen | sure |
20:24.18 | cr2 | ok. |
20:24.29 | cr2 | sent. |
20:24.51 | kiozen | received, trying... |
20:24.55 | cr2 | prepare for blowing your nice vga lcd :) |
20:25.07 | cr2 | heh. there is already a bug. |
20:25.13 | kiozen | ok, I wear googles |
20:25.16 | cr2 | the screen will be flickering. |
20:25.32 | kiozen | and then explode nicely? |
20:25.32 | cr2 | but only after resume, which does not work :) |
20:25.47 | cr2 | we had the same problem on the universal. |
20:26.11 | cr2 | it's the silly pxafb which does not support pixclock doubling through .lccr4 |
20:26.19 | cr2 | on pxa27x |
20:26.22 | pH5 | cr2: add it! |
20:26.47 | cr2 | pH5: ifdef spaghetti again ;) |
20:29.09 | cr2 | goxboxlive: preparing to destroy athena now. |
20:29.19 | cr2 | goxboxlive: hmm. it will not work. |
20:29.42 | cr2 | goxboxlive: do i need rootdelay=10 root=/dev/sda1 ? |
20:30.03 | cr2 | i boot from the usb stick |
20:33.05 | cr2 | Mounting SD card |
20:33.15 | cr2 | FAIL FAIL FAIL FAIL |
20:34.51 | kiozen | cr2: booting, but same root device SNAFU |
20:35.18 | kiozen | screen flickers shortly after mmutrace |
20:35.27 | cr2 | kiozen: the SD my be wrong, i don't yet touch the wifi/SD enable bit. |
20:35.40 | cr2 | kiozen: mmutrace ?? |
20:36.14 | kiozen | sorry: turning of mmu |
20:36.48 | cr2 | but it flickers ? |
20:37.09 | kiozen | yes very short |
20:37.23 | cr2 | ok. |
20:37.38 | cr2 | i need to research the wifi power gpios. |
20:37.48 | cr2 | and switch over to SD. |
20:38.18 | cr2 | they share the same pxamci controller, so some arbitration is happening. |
20:38.20 | kiozen | the device will stall at 'turning off mmu' if it had been in wince/suspend befor |
20:39.20 | cr2 | kiozen: can you dump the CPLD registers ? |
20:39.34 | kiozen | how? |
20:39.42 | cr2 | while running wince with wifi on and off. |
20:39.52 | cr2 | i'll tell you the address to dump. |
20:42.31 | kiozen | what was the activesync ip address again? (missed to note that one) |
20:43.51 | cr2 | pd 0x10000000 0x20 |
20:45.00 | cr2 | 192.168.55.100 ? |
20:45.48 | cr2 | it's the gw |
20:45.51 | cr2 | 192.168.55.101 |
20:46.33 | kiozen | ok have to change os, bbl |
20:47.35 | cr2 | goxboxlive: how do i switch from SD to the usb ? |
20:49.59 | *** join/#htc-linux AC_Mast (i=s@gateway/tor/x-60c7240041c60c8b) |
20:59.39 | *** join/#htc-linux kiozen (n=oeichler@rgnb-4db03060.pool.einsundeins.de) |
21:00.25 | kiozen | cr couldn't see any change http://pastebin.ca/795128 |
21:00.33 | kiozen | cr2 |
21:00.46 | cr2 | kiozen: ok. |
21:06.02 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
21:34.56 | cr2 | Kevin2: i've captured and decoded the asic3 SD and SDIO processing, so i can compare the wince ops directly, without the hacked linux driver. |
21:38.19 | cr2 | goxboxlive: after athena universal looks really small :) |
22:00.17 | cr2 | hehe. wince rejects to detect a partitioned stick. |
22:02.06 | pH5 | cr2: even if the first partition is fat16/fat32 type and has a vfat system? |
22:02.32 | cr2 | wee :) |
22:02.38 | cr2 | pH5: way my fault. |
22:02.50 | cr2 | angstrom boots on athena :) |
22:03.02 | pH5 | cr2: congratulations :) |
22:03.04 | cr2 | generating key |
22:03.17 | *** join/#htc-linux lieter (n=lieter@a62-251-100-248.adsl.xs4all.nl) |
22:03.18 | cr2 | pH5: still, it a long way to go. |
22:03.32 | cr2 | running off the usb stick is not practical. |
22:04.12 | cr2 | mdoc g4 may be better, and the SD needs to be made working. |
22:04.29 | cr2 | i'm not talking about the 8GB microdrive... |
22:05.52 | pH5 | yup, I just notice that I've got my magician for two years now, and I'm still not running from flash :) |
22:05.52 | cr2 | well, now i can't login, because i need the usb slot. |
22:06.06 | cr2 | pH5: sd boot ;) |
22:06.24 | cr2 | starting qtopia... |
22:06.30 | cr2 | segfault |
22:06.43 | cr2 | tslib installation!) |
22:07.20 | cr2 | goxboxlive: you have some funny kernel module versioning. |
22:07.41 | cr2 | mod_unload ARMv5 |
22:08.13 | Kevin2 | cr2: You were able to boot from ATI SD? |
22:08.24 | cr2 | Kevin2: no, from USB stick. |
22:08.33 | cr2 | it's athena, not hermes. |
22:09.36 | cr2 | goxboxlive: hardcoding /dev/event/* is not nice ;) |
22:09.50 | cr2 | Kevin2: what can we do on hermes now ? |
22:09.54 | goxboxlive | pH5: lets change that and run from the flash |
22:10.07 | goxboxlive | cr2: i know, but it's only the startup script |
22:10.19 | goxboxlive | even greenphone does it that way AFAIR. |
22:10.22 | goxboxlive | also fic |
22:10.32 | cr2 | goxboxlive: what's the problem with tslib ? |
22:10.51 | cr2 | goxboxlive: if others do something wrong, it does not mean you should do it the same way :) |
22:11.08 | goxboxlive | cr2: wich image are you using. ? Yesteday it where missing so i hade to add it. |
22:11.24 | goxboxlive | missing in console image, it where suplied before. |
22:11.24 | cr2 | goxboxlive: the image you have given me. |
22:11.42 | goxboxlive | if it worked oon universal it should work in athena too |
22:11.42 | cr2 | do you have a new one ? |
22:11.56 | cr2 | goxboxlive: 17 Nov |
22:11.58 | goxboxlive | no, not now. Use the one at linuxtogo |
22:12.09 | goxboxlive | should include tslib |
22:12.10 | goxboxlive | bbl |
22:12.17 | cr2 | ok. |
22:12.49 | Kevin2 | cr2: I'm not sure what you mean. I'm still planning to write hermes keyboard and adc drivers. |
22:15.13 | cr2 | Kevin2: ok, i'll debug the ATI SD, and then ATI powerup. |
22:16.24 | cr2 | pH5: do we have a chance to move to 2.6.24 ? |
22:17.09 | cr2 | i've seen that zaurus 2.6.23 kernel includes the patch for universal. |
22:18.35 | cr2 | i don't want to debug the mmc on .21 knowing that it got a complete rewrite in .23 |
22:23.56 | pH5 | cr2: I'm still hoping psokolovsky will get up one day and decide to just do it ;) |
22:24.05 | Kevin2 | I agree. I wonder if we should ditch hh.org cvs and setup/find a new repository somewhere. |
22:24.22 | pH5 | Kevin2: git |
22:24.29 | Kevin2 | With the mailing list gone, everything is in flux. |
22:24.51 | pH5 | Kevin2: actually, I'd like to upstream as much as possible, as soon as possible. |
22:24.54 | Kevin2 | pH5: I agree. bd2 had the same comment earlier also. |
22:24.59 | cr2 | hh.org has a lot of weird ipaq stuff |
22:25.12 | cr2 | the asic3 driver need a serious cleanup. |
22:25.35 | pH5 | my current agenda is getting gpiolib and then htc-egpio into mainline. |
22:25.37 | cr2 | if i was a mainline mainteiner i'd have rejected it from the very beginning ;) |
22:25.40 | Kevin2 | I don't think mainline has the pxa27x usb gadget driver. Otherwise, I think htcapache could go mainline pretty easily. |
22:26.20 | pH5 | Kevin2: I don't know enough about usb to make anything of the discussion that comes up on lak once in a while and then dies again :( |
22:26.24 | Kevin2 | (And honestly, I'd reject the pxa27x driver from mainline - it's just an ugly cut and paste of the original pxa usb driver.) |
22:26.54 | Kevin2 | I don't know anything about usb either. :-( |
22:26.56 | cr2 | pH5: i'm thinking about a generic pwm driver. |
22:26.58 | pH5 | Kevin2: that's about what david brownell told rodolfo giometti when it last came up. but no fixes came after that. |
22:27.17 | cr2 | i'm not an usb expert too... |
22:27.50 | Kevin2 | I wasn't aware of the history. But I tried to diff the two, and it's a mess. Clearly mainline pxa driver has been maintained where the pxa27x has not. |
22:34.44 | cr2 | ok, i've booted with usb hub reinserting during rootdelay=10 |
22:36.32 | cr2 | Run Mode clock: 208.00MHz (*16) |
22:36.32 | cr2 | Turbo Mode clock: 624.00MHz (*3.0, active) |
22:36.32 | cr2 | Memory clock: 208.00MHz (/2) |
22:36.32 | cr2 | System bus clock: 208.00MHz |
22:36.52 | cr2 | EGPIO phys=08000000 virt=c8800000 |
22:36.52 | cr2 | EGPIO phys=09000000 virt=c8802000 |
22:36.52 | cr2 | W228X: remapping=4000000 |
22:36.52 | cr2 | W228X: num_resources=4 |
22:36.52 | cr2 | W228X: registering atiw_mmc @ 4200000 |
22:36.52 | cr2 | W228X: base_init+register ret=0 |
22:37.46 | cr2 | atiw_mmc: ATIW MMC/SD Driver, controller at 0x4200000 |
22:37.46 | cr2 | mmc0: clock 0Hz busmode 1 powermode 0 cs 0 Vdd 0 width 0 timing 0 |
22:37.46 | cr2 | atiw_mmc: requested clock 0Hz busmode 1 powermode 0 Vdd 0 |
22:38.16 | cr2 | atiw_mmc: requested clock 312500Hz busmode 1 powermode 2 Vdd 20 |
22:38.16 | cr2 | egpio set: reg 3 = 0x0040 |
22:38.44 | cr2 | hmm. i can apply the power earlier. |
22:39.45 | cr2 | mmc0: starting CMD0 arg 00000000 flags 00000040 |
22:39.55 | cr2 | mmc0: req done (CMD0): 0/0/0: 00000000 00000000 00000000 00000000 |
22:40.03 | cr2 | does not look good. |
22:40.13 | cr2 | atiw_mmc: requested clock 312500Hz busmode 1 powermode 2 Vdd 20 |
22:40.13 | cr2 | egpio set: reg 3 = 0x0040 |
22:40.13 | cr2 | mmc0: starting CMD8 arg 000001aa flags 00000075 |
22:40.14 | cr2 | atiw_mmc: Opcode: 8, base: c8808000 reg_base: c880a000 |
22:40.37 | cr2 | now it's dead. only 1 irq was generated. and obviously not acked. |
22:41.52 | cr2 | <PROTECTED> |
22:41.53 | Kevin2 | Not acked at the egpio layer or the ati irq layer? |
22:41.56 | cr2 | <PROTECTED> |
22:42.00 | cr2 | <PROTECTED> |
22:42.00 | cr2 | <PROTECTED> |
22:42.00 | cr2 | <PROTECTED> |
22:42.00 | cr2 | <PROTECTED> |
22:42.00 | cr2 | <PROTECTED> |
22:42.00 | cr2 | <PROTECTED> |
22:42.17 | cr2 | Kevin2: it should be acked at the ATI CTRL layer. |
22:42.23 | cr2 | at least wince does it. |
22:42.40 | Kevin2 | Can you download the "x" program and poll the registers? Perhaps even manually ack them? |
22:42.41 | cr2 | it's not a egpio irq, but a PXA irq. |
22:43.13 | cr2 | Kevin2: i'd like to compare the driver with the wince dump from asic3 and atiw2284 |
22:43.42 | cr2 | i'm missing some register commands on atiw compared to asic3_mmc |
22:47.12 | cr2 | iomem looks good. |
22:47.16 | cr2 | 04000000-04020000 : w228x.0 |
22:47.17 | cr2 | 04200000-04204000 : atiw_mmc |
22:47.17 | cr2 | <PROTECTED> |
22:47.17 | cr2 | 04800000-04895fff : vsfb |
22:47.17 | cr2 | 08000000-08000020 : htc-egpio.0 |
22:47.17 | cr2 | 09000000-09000004 : htc-egpio.1 |
22:49.27 | *** join/#htc-linux ltxda (n=ltxda@c-98-196-125-205.hsd1.tx.comcast.net) |
22:49.28 | cr2 | hmm. why there is no ad7877 touchscreen... |
22:50.18 | Kevin2 | cr2: Athena has ad7877 and the htc spi keyboard? Are they on the same spi bus? |
22:50.35 | cr2 | no. |
22:50.47 | cr2 | the spi keyboard is hotplug. |
22:51.21 | Kevin2 | hotplug? |
22:51.28 | cr2 | Kevin2: i have DAV and pendown. you've listed 3 pins, what is the third one ? |
22:51.39 | cr2 | yes, it can be disconnected. |
22:51.57 | cr2 | it has a magnet inside to attach. |
22:52.01 | dcordes|away | is there a way to make ctrl key in wm6 with htc keyboard/ |
22:52.46 | Kevin2 | cr2: The apache has only 2 pins for ad7877 also. I thought a third pin was also used, but it appears not. |
22:53.21 | cr2 | Kevin2: you get a gpio irq that the keyboard is attached/disconnected so you just apply/remove power and get the data from SPI. |
22:53.23 | cr2 | ok. |
22:53.37 | Kevin2 | But it's a different spi bus from ad7877? |
22:53.49 | cr2 | i'll look. |
22:54.23 | cr2 | heh. i've probably forgot the ssp driver. |
22:54.36 | cr2 | so much for missing the generic spi interface ;) |
22:56.28 | cr2 | Kevin2: where do you set the spi port for adc ? |
22:57.12 | Kevin2 | cr2: Heh, the ad7877 is just hardcoded for the first pxa ssp port. ;-) |
22:58.03 | cr2 | 23,24,25,26 is ssp1 ? |
22:58.41 | cr2 | yes, ssp2 is the spi keyboard. |
22:59.13 | cr2 | http://wiki.xda-developers.com/index.php?pagename=AthenaGPIO |
23:00.42 | cr2 | Kevin2: where is the ADC pxa_ssp driver ? |
23:01.18 | Kevin2 | arch/arm/mach-pxa/ssp.c |
23:01.46 | cr2 | in the menu |
23:02.13 | cr2 | aah. misc devices. |
23:03.19 | cr2 | hm. and i need select PXA_SSP |
23:03.35 | cr2 | oh. i want generic spi ;) |
23:08.16 | cr2 | it does not want to be selected . |
23:11.47 | cr2 | the menuconfig parser has some bugs. |
23:12.18 | cr2 | if there is no legend for the "tristate" line, it does not show the option. |
23:13.05 | cr2 | ok, now everything compiled. |
23:18.59 | *** join/#htc-linux Marex-notebook (n=marex@gwfm10-3-250.802.cz) |
23:34.52 | cr2 | booting the n560 kernel, and wondering why it does not boot ;) |
23:37.54 | pH5 | cr2: gettin late? |
23:41.41 | cr2 | yeah. rebooting. |
23:42.31 | cr2 | haret may probably check the zimage mtype |
23:48.09 | pH5 | cr2: a zImage doesn't necessarily have a single MTYPE |
23:48.18 | cr2 | assertion mixer failed. |
23:48.46 | cr2 | pH5: but it may be at least compatible. |
23:48.53 | cr2 | of you can force it. |
23:48.55 | pH5 | should be possible, but I suspect unpacking the kernel and finding the machine type table indepently of the kernel version is not trivial |
23:49.35 | cr2 | yes, linking to zlib in wince is not nice. |
23:51.14 | dcordes | cr2: is it complicated to remove the lcd driver from msm7200A kernel and replace it with vsfb? I would really like to try build and boot something. |
23:52.14 | dcordes | just a defconfig thing? |
23:53.32 | cr2 | dcordes: look how it's done on hermes in the hh.org CVS |
23:53.58 | cr2 | it's actually very simple. ( VerySimpleFrameBuffer) |
23:54.52 | dcordes | hh.org= global handhelds.org kernel? |
23:56.20 | cr2 | yes |
23:57.36 | dcordes | http://www.handhelds.org/sources.html I'm there and don't know where to go. |
23:59.45 | cr2 | http://www.handhelds.org/handhelds-faq/development.html#USING-CVS |