00:09.09 | *** join/#htc-linux SmallR2002 (n=SmallR20@88-107-35-245.dynamic.dsl.as9105.com) |
00:17.16 | *** join/#htc-linux unicat_ (i=foobar@dslb-088-073-016-075.pools.arcor-ip.net) |
00:19.52 | *** join/#htc-linux Ranpha1 (n=Ranpha@hlm-n-73ed.adsl.wanadoo.nl) |
00:55.40 | *** join/#htc-linux ellisway (n=ellis@83.167.181.177) |
01:01.01 | *** join/#htc-linux psokolovsky_ (n=psokolov@82.193.98.14) |
03:08.01 | *** join/#htc-linux ellisway (n=ellis@83.167.181.177) |
03:26.00 | *** join/#htc-linux GSV (i=GSV@c-71-202-155-188.hsd1.ca.comcast.net) |
06:22.12 | *** join/#htc-linux RoEn (n=roen@p54A67546.dip.t-dialin.net) |
06:26.40 | *** join/#htc-linux ellisway (n=ellis@83.167.181.177) |
07:02.57 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
07:38.51 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
07:39.33 | *** join/#htc-linux dion (n=dion@inhex.net) |
08:05.20 | *** join/#htc-linux psokolovsky_ (n=psokolov@82.193.98.14) |
08:37.18 | *** join/#htc-linux pH5 (n=ph5@p5485D3EA.dip.t-dialin.net) |
08:41.53 | *** join/#htc-linux ellisway (n=ellis@83.167.181.177) |
09:00.47 | *** join/#htc-linux psokolovsky_ (n=psokolov@82.193.98.14) |
09:02.36 | goxboxlive | Good morning pH5 & psokolovsky_ |
09:18.54 | *** join/#htc-linux psokolovsky__ (n=psokolov@empty-crossing.volia.net) |
09:27.13 | goxboxlive | cr2, i have bluetooth mouse working with Universal :-) It's created as /dev/input/event4. |
09:27.38 | goxboxlive | Havent manage to get it working in OpieII yet, but i will :-) |
09:28.31 | goxboxlive | <7>PM: Adding info for bluetooth:acl0050F27F2B2E |
09:28.31 | goxboxlive | <6>input: Microsoft Five Button Mouse as /class/input/input4 |
09:28.31 | goxboxlive | root@htcuniversal:~$ |
09:52.14 | pH5 | hi goxboxlive, isn't that mouse bigger than the universal? I have used a logitech mx900 with the ipaq hx4700 once, but it felt a bit weird. |
09:54.25 | *** join/#htc-linux hugo_nz (n=hugo@132.181.52.99) |
09:56.38 | goxboxlive | pH5, just 4 testing :-) |
09:56.58 | goxboxlive | it also work wit OpieII now :-) |
10:51.00 | goxboxlive | psokolovsky__, is there any profit for htcuniversal, magician or hx4700 to compile with iwmmxt instead of Xscale ? |
10:51.27 | psokolovsky__ | hi |
10:51.36 | goxboxlive | hi btw |
10:51.56 | psokolovsky__ | at the current stage, IMHO it would cause more trouble than benefit |
10:52.20 | goxboxlive | i c |
10:52.23 | pH5 | goxboxlive: only for programs that explicitly make use of it, I'd guess. |
10:52.44 | pH5 | mplayer is already compiled for iwmmxt, I think (at least hope) it has some optimized routines. |
10:52.48 | psokolovsky__ | it's better to hand-optimize assembly for selected algorithms, like video, and make them be selected dynamically based on running CPU ;-) |
10:52.52 | pH5 | liboil/gstreamer/ffmpeg would be other candidates. |
10:53.22 | goxboxlive | ok |
10:53.42 | pH5 | I had built complete iwmmxt images for some time though, everything except glibc didn't cause a problem with -mtune=iwmmxt |
10:55.24 | psokolovsky__ | pH5: did you ever try to look at objdump to see how well it "optimizes"? |
10:57.05 | *** join/#htc-linux ellisway (n=ellis@83.167.181.177) |
10:59.48 | pH5 | psokolovsky__: heh, no. actually, I haven't looked at any iwmmxt assembler code at all, yet. |
11:01.33 | pH5 | iwmmxt optimized blitting/scaling/alpha blending code would be cool for the pxafb machines. |
11:02.20 | psokolovsky__ | I just doubt it could get much speedup on generic code, and it would be nice to know how smart it with vector/array code. big mmx told to be bale to do at least sth there. |
11:02.43 | psokolovsky__ | yeah, sure. still think that hand-coding would give much more |
11:03.13 | psokolovsky__ | and sigh, that we still don't have atomatical optmization infra (for example, based on genetical algos) |
11:04.59 | *** join/#htc-linux Tjikkun (n=tjikkun@82-204-54-115.dsl.bbeyond.nl) |
11:24.21 | *** join/#htc-linux psokolovsky__ (n=psokolov@empty-crossing.volia.net) |
11:26.09 | *** join/#htc-linux psokolovsky (n=psokolov@empty-crossing.volia.net) |
12:45.14 | *** join/#htc-linux psokolovsky_ (n=psokolov@82.193.98.14) |
13:17.13 | *** join/#htc-linux psokolovsky__ (n=psokolov@empty-crossing.volia.net) |
13:21.18 | *** join/#htc-linux psokolovsky (n=psokolov@82.193.98.14) |
13:46.04 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
13:46.31 | goxboxlive | . |
14:12.13 | cr2 | goxboxlive: hi. |
14:12.33 | cr2 | opie-networksettings-1.2.2-r0 |
14:12.37 | cr2 | ERROR: function do_compile failed |
14:13.45 | psokolovsky | 1.2.2 is too old |
14:14.11 | cr2 | hi psokolovsky |
14:14.18 | psokolovsky | hi |
14:14.29 | cr2 | that's what bitbake opie-image picks. |
14:14.38 | psokolovsky | cr2: ever bothered with "industrial" pdas? ;-) |
14:14.53 | cr2 | no. |
14:15.13 | psokolovsky | here's possibility to buy pda with barcode scanner, always was my idiot's dream ;-) |
14:15.33 | cr2 | sounds cool :) |
14:16.37 | psokolovsky | yep ;-) |
14:16.52 | psokolovsky | it's some bare wince 4.1 thing though |
14:17.37 | cr2 | if the barcode scanner is ps2 or serial it should be easy. |
14:17.40 | *** join/#htc-linux ellisway (i=twer@83.167.180.99) |
14:17.44 | cr2 | which cpu is it using ? |
14:17.58 | *** join/#htc-linux dion (n=dion@inhex.net) |
14:18.25 | goxboxlive | hi cr2 i have problems building too. It's the dam* libstdc++. Now when i changed the extra install too libstdc++ instead of libstdc++6 it cant find it's package. |
14:19.49 | goxboxlive | Here is the pastebin. http://pastebin.no/2182 |
14:21.23 | cr2 | is it just MACHINE="magician" ? |
14:22.22 | cr2 | i'm compiling openmoko just for fun. |
14:22.31 | cr2 | a780 fails for me too. it was ok a week ago. |
14:22.48 | psokolovsky | cr2: it is pxa255, pretty bare. only cf+sd besides scanner |
14:22.59 | cr2 | cf is good. |
14:23.05 | psokolovsky | I really hope that scanner is on serial, or it all just waste of time and money ;-) |
14:23.11 | cr2 | no pain. |
14:23.31 | psokolovsky | goxboxlive: you really should raise that question to RP |
14:23.41 | cr2 | it can be ps2, but unlikely. |
14:24.03 | cr2 | psokolovsky: it can be sdio ;) |
14:24.03 | psokolovsky | goxboxlive: in the meantime you could try 2-levekl sceme, task-magician-console + magician-console-image |
14:24.16 | psokolovsky | cr2: yes, that's announced in specs |
14:24.32 | goxboxlive | psokolovsky, right, i think ljp is about to remove the OpieII listdc++ depends . He is trying it out now. |
14:24.33 | psokolovsky | cr2: but that's only worse, as mmc is not on pxa then ;-I |
14:25.22 | psokolovsky | goxboxlive: and as you know, the *right* way to solve that is to add opie2 to OE in the first place. you're working on it for quite some time, you could do that long ago... |
14:25.41 | cr2 | psokolovsky: the sdio on pxa is supported. and the interface can be like a serial port like gps (which is supported too). |
14:26.18 | goxboxlive | psokolovsky, it exist a bb file for OpieII, but it is just a signle file wich builds everything, and zecke told that it would not be accepted. |
14:26.29 | psokolovsky | cr2: pxa supports sdio? didn't know that. and yeah, I hope for exactly internal serial connection ;-) |
14:26.37 | goxboxlive | to split it up is kind of complex i heard. |
14:27.36 | psokolovsky | cr2: I just wondered if you have an idea what protocol such scanners use... |
14:27.47 | cr2 | psokolovsky: http://source.mvista.com/~dsingleton/sdio |
14:28.18 | cr2 | i think there must be some linux barcode scanner driver. |
14:28.44 | cr2 | the protocol must be very primitive. |
14:28.53 | cr2 | cuecat ? |
14:29.10 | cr2 | the famous dotcom fiasco ? |
14:29.54 | cr2 | http://users.skynet.be/ppc/cuecat |
14:31.33 | psokolovsky | cr2: thanks for link! |
14:31.55 | *** join/#htc-linux eway (n=ellis@83.167.181.177) |
14:35.43 | cr2 | psokolovsky: why are the sa1100fb mach_info structs are commented out ? is there any device that really uses the new feature ? |
14:37.40 | *** join/#htc-linux scorpio16v (n=scorpio1@ACB4B9F8.ipt.aol.com) |
14:39.05 | psokolovsky | cr2: I added it just ~1week ago. so no, but all should, and that was intended, as you see by those commented out structs ;-) |
14:48.29 | cr2 | yes, the current driver looks very ugly with the #ifdefs |
14:52.09 | *** part/#htc-linux scorpio16v (n=scorpio1@ACB4B9F8.ipt.aol.com) |
14:57.39 | cr2 | git://git2.kernel.org/pub/scm/linux/kernel/git/drzeus/mmc.git |
15:29.18 | *** join/#htc-linux unicat (i=foobar@dslb-088-073-045-102.pools.arcor-ip.net) |
15:54.14 | *** join/#htc-linux rob_w (n=bob@M87a5.m.pppool.de) |
16:09.57 | *** join/#htc-linux Pi}|{amka (n=Katya@55.gprs-nat.mtsnet.ru) |
16:10.30 | Pi}|{amka | Привет |
16:14.40 | *** part/#htc-linux Pi}|{amka (n=Katya@55.gprs-nat.mtsnet.ru) |
16:17.40 | *** join/#htc-linux ellisway (n=ellis@83.167.181.177) |
16:32.54 | Kevin2 | pH5: Are you around? |
16:33.39 | pH5 | hi Kevin2, yes I am |
16:34.05 | Kevin2 | Hey - you have a kernel that can boot directly from the bootloader, right? |
16:34.45 | Kevin2 | Would you be able to run a test with it on haret? I'm trying to see if "resumeintoboot" works, but I'm pretty sure my kernel can't cope with uninitialized hardware. |
16:35.45 | Kevin2 | http://handhelds.org/~koconnor/haret/haret-20070624.exe and http://handhelds.org/~koconnor/haret/haret-20070624b.exe |
16:36.48 | Kevin2 | pH5: If you could try haret with those binaries - just setup like you would normally boot the kernel, but use "resumeintoboot" instead of "boot". |
16:37.55 | Kevin2 | BTW, I just found that I need to manually init the MSC2 registers in order for me to talk to the egpio chip. |
16:38.20 | pH5 | Kevin2: I can try that. But I'm not sure if maybe the bootloader already sets up some hardware state which it wouldn't when booting directly from flash. |
16:39.05 | Kevin2 | pH5: It looks like the resume code path is pretty sparse on the bootloader. Doesn't seem to do much of anything besides jump to wince. I could be wrong though. |
16:40.03 | pH5 | Kevin2: no, I have the same impression. if the SMR bit is set, it just jumps to wince here. |
16:42.02 | Kevin2 | Cool, so jumping to linux kernel init on resume should be a pretty good test for linux's hardware setup. |
16:45.32 | pH5 | except for the memory setup etc. that is done in the wince resume path before doing this PSPR crc thing and jumping back into action. |
16:46.46 | pH5 | I still couldn't figure out why resume via wince's resume path doesn't work here when I boot from the bootloader instead of via haret. |
16:47.25 | Kevin2 | Hrmm. But wouldn't the normal bootloader config memory also? |
16:48.36 | pH5 | Ah right, linux needs configured memory. |
16:50.29 | Kevin2 | The latest versions of haret have a "resumetraces" list. You can add registers and memory addresses to query during a wince resume. Run "wi 30", suspend the device, resume the device, and you can query the status of the machine just before wince resumes. |
16:51.06 | Kevin2 | Using this method, I just found that the bootloader only restores part of the memory config. For example, the egpio accesses are not setup. |
16:53.47 | pH5 | Kevin2: very nice. |
16:54.19 | *** join/#htc-linux FossiFoo_ (n=Fossi@e176097144.adsl.alicedsl.de) |
16:56.41 | *** join/#htc-linux ellisway (n=ellis@83.167.181.177) |
17:01.32 | pH5 | Kevin2: both stop after loading the kernel with an "unexpected resume vector (ea000000 e3e06000)" |
17:17.14 | Kevin2 | pH5: What is the wince resume address on your machine? |
17:19.47 | *** join/#htc-linux sinmsinm (n=alex@1.Red-83-43-229.dynamicIP.rima-tde.net) |
17:20.09 | sinmsinm | hi |
17:20.38 | sinmsinm | anybody is trying to linuxize s200? |
17:20.51 | sinmsinm | i would like to help |
17:21.34 | *** part/#htc-linux sinmsinm (n=alex@1.Red-83-43-229.dynamicIP.rima-tde.net) |
17:22.12 | *** join/#htc-linux sinmsinm (n=alex@1.Red-83-43-229.dynamicIP.rima-tde.net) |
17:22.54 | pH5 | Kevin2: I don't understand what you mean. the bootloader jumps to 0x40000 in flash where it does the mem setup and crc check. the resume vector (resume address, arm control register, translation table base, domain access control) is stored in memory at 0xa0000000. |
17:23.04 | pH5 | Kevin2: do you want me to read what is stored at 0xa0000000? |
17:23.12 | pH5 | sinmsinm: what is s200? |
17:23.23 | sinmsinm | qtek s200 |
17:23.40 | sinmsinm | it has prophet processor |
17:23.41 | pH5 | htc prophet? |
17:23.44 | sinmsinm | isn't it? |
17:23.47 | pH5 | ah, ok. |
17:23.47 | Kevin2 | pH5: At some point the bootloader hands control back to wince. Is that at 0xa0000000? If so, then yes, can you dump the first few addresses there? |
17:24.38 | sinmsinm | now i'm trying to get all information i can |
17:25.22 | pH5 | Kevin2: http://en.pastebin.ca/587524 <-- this is the last part of the resume path. let me check what's at 0xa0000000. |
17:27.16 | pH5 | sinmsinm: prophet is one of the omap850 phones iirc? - you should ask the people that are doing the wizard or tornado ports. |
17:28.09 | sinmsinm | ok |
17:28.13 | sinmsinm | thanks |
17:28.44 | pH5 | http://vivien.chappelier.free.fr/typhoon/index.html |
17:28.44 | pH5 | http://linwizard.sourceforge.net/ |
17:29.16 | pH5 | unfortunately they don't seem to hang around here. thus I know next to nothing about omap850. |
17:31.03 | sinmsinm | it's time to boost this project arround prophet |
17:31.08 | sinmsinm | :-) |
17:31.19 | Kevin2 | pH5: As a guess, your wince resume vector is stored at 0xa0000000. |
17:37.03 | pH5 | Kevin2: a0000000 | 0000040a 0010040a 0020040a 0030040a | .......... ...0. |
17:37.09 | pH5 | that looks like it's not filled in yet |
17:37.34 | pH5 | how do I check the value at 0xa0000000 during resume? |
17:38.06 | Kevin2 | Well, you need to know the resume vector to test things during resume. :-} |
17:39.01 | Kevin2 | You could see if "set mmutrace p2v(0xa0000000)" shows you what it is. |
17:39.02 | pH5 | cat I set the resume vector address from within haret? |
17:40.07 | Kevin2 | pH5: Yes. But then I'm not sure wince wouldn't overwrite it when it suspended the device. |
17:40.12 | pH5 | HaRET(1)# set mmutrace p2v(0xa0000000) |
17:40.13 | pH5 | line 1: MMUTRACE(1 args) expected |
17:40.47 | Kevin2 | Oops. Meant "set trace p2v(0xa0000000)". |
17:41.07 | Kevin2 | Then run "wi 30", suspend / resume, and see if you see a trace for what wince throws into that address. |
17:43.13 | pH5 | argh |
17:43.22 | pH5 | the connection dies when I suspend |
17:43.39 | Kevin2 | It doesn't come back when you resume? |
17:44.17 | pH5 | no. synce-serial over ttyUSB0 dies completely. |
17:44.38 | pH5 | let's try that again and log to a file. |
17:44.45 | Kevin2 | Oh. If I set everything back up with the same IPs it usually resumes okay for me. |
17:45.01 | Kevin2 | Yeah, just add "log foo.txt" first and you'll have it on disk. |
17:45.29 | pH5 | Kevin2: I know, I wrote the log patch for gnuharet because of this issue :) |
17:46.08 | Kevin2 | Heh |
17:51.14 | Kevin2 | You could also try searching the first few megs of ram. On my apache the signature is 0xea0003fe (b +0x1000) followed by a bunch of 0x0 values. (The real wince code starts at 0xa0041000, but I hook the address at 0xa0040000.) |
17:52.58 | pH5 | Kevin2: is that wm2005? |
17:53.03 | Kevin2 | Yes. |
17:53.23 | pH5 | ah, wm2003se here. I have the b+0x1000 signature at 0x40000, so there it is still in flash. |
17:54.47 | pH5 | the log doesn't seem to contain anything useful |
17:55.00 | pH5 | but after resume I have this at 0xa0000000: |
17:55.00 | pH5 | a0000000 | 80080bb4 00003a7f a0660000 00000001 | .....:....f..... |
17:55.56 | pH5 | bbl, food |
18:00.25 | psokolovsky | cr2: so, I bought that crap. turned out to be SA ;-( |
18:01.29 | *** join/#htc-linux FossiFoo (n=Fossi@e182095206.adsl.alicedsl.de) |
18:01.53 | psokolovsky | or maybe ;-), as now I'd have more chances to hack on SA ;-) just regret few extra bucks I paid for it ;-) |
18:02.38 | *** join/#htc-linux GSV (i=GSV@c-71-202-155-188.hsd1.ca.comcast.net) |
18:03.18 | psokolovsky | it has touchscreen on some hacky microcontroller - they just copy h3600 design ;-\ |
18:06.42 | GSV | see the new qualcomm cpu!! http://www.mymobile911.com/dnn/Home/tabid/36/EntryID/56/Default.aspx#Comments |
18:07.34 | GSV | see the difference between hermes http://www.mymobile911.com/dnn/Home/tabid/36/EntryID/58/Default.aspx |
18:08.14 | *** join/#htc-linux psokolovsky_ (n=psokolov@empty-crossing.volia.net) |
18:11.55 | Kevin2 | I found the wince resume address on Hermes - it is at 0x30100000. Haret resumetraces work too! |
18:15.01 | GSV | kevin2: did you see the links i posted few minutes ago? |
18:19.59 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
18:21.16 | Kevin2 | GSV: Hrmm. Interesting. But qualcomm doesn't release any specs, so I doubt we'll be able to do anything with it. |
18:21.35 | *** part/#htc-linux sinmsinm (n=alex@1.Red-83-43-229.dynamicIP.rima-tde.net) |
18:24.54 | GSV | kevin2:i just wanna know that is this qualcomm better or s3c2442? |
18:27.44 | GSV | http://www.cdmatech.com/products/msm7200_chipset_solution.jsp |
18:28.46 | GSV | "Support for Linux® and other third-party operating systems" |
18:31.34 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
18:32.09 | Kevin2 | GSV: Looks like marketing fluff to me. We need to be able to download detailed data sheets. To date, I haven't seen anything like that from Qualcomm. (And, yes, I've looked.) |
18:43.55 | goxboxlive | hi BabelO , are you awake? |
18:43.57 | BabelO | hi |
18:44.09 | BabelO | goxboxlok, same idea :) |
18:44.35 | goxboxlive | :-) Did you build Opie2 ? |
18:45.22 | BabelO | yes, but can't start it :( |
18:45.41 | BabelO | problem with dbmigrate |
18:46.45 | *** join/#htc-linux dion (n=dion@inhex.net) |
18:48.56 | goxboxlive | BabelO, realy, hmm i dont have that. But this version is more unstable than the 4.2.1 i think |
18:49.15 | goxboxlive | BabelO, did you install it or copy it to your sd/NFS? |
18:54.43 | BabelO | i just use make install and then share image folder over nfs |
18:54.44 | BabelO | but still have black screen |
18:54.53 | goxboxlive | hmm |
18:55.17 | goxboxlive | Did Helix/alsa build ? |
18:56.22 | BabelO | yes |
18:56.56 | BabelO | i ll fix error |
18:57.23 | pH5 | Kevin2: what is the purpose of ResumeSJ->returnCode? |
18:58.04 | pH5 | for wm2003 we can't replace the vector at 0xa0000000 because it gets overwritten during resume. |
18:58.42 | pH5 | maybe overwrite the instructions at 80080bb4? (what is up with that - haret gets killed when I try to pd 80080bb4) |
18:58.54 | pH5 | *overwritten during suspend |
18:59.54 | Kevin2 | 80080bb4 doesn't look like a valid memory address |
19:00.55 | Kevin2 | returnCode is misleading - it is the address of code to jump to when done with the handler. |
19:03.04 | cr2 | hi |
19:03.09 | Kevin2 | Hi cr2 |
19:03.27 | goxboxlive | hi cr2 |
19:04.29 | Kevin2 | pH5: Maybe try dumping top few megabytes of ram and search for ea0003fe -- maybe there is another match somewhere in ram? |
19:04.47 | pH5 | hi cr2 |
19:05.00 | cr2 | i've done a lot of work on roadmap, and have reached most results i've expected. |
19:05.04 | pH5 | Kevin2: wm2003 works differently. the ea0003fe is at 0x40000 still in flash |
19:05.55 | pH5 | maybe we have a chance to intercept wince's suspend code so we can let it write the address of the trampoline to 0xa0000000? |
19:06.21 | cr2 | Kevin2: wince3 resumes into flash. |
19:07.32 | Kevin2 | pH5: Yeah, a trace on "a0000000" should help find writes to that address. Maybe they are using a different virtual address to write to it? |
19:08.00 | BabelO | goxboxlive: for power button it is already handheld, maybe try to remove comment in htcphonehardwqare.cpp |
19:08.49 | goxboxlive | BabelO, what do you mean? What is the powerbutton suposed to do? Suspend? |
19:08.52 | cr2 | pH5: is the resume code different from ba on magician ? |
19:11.24 | Kevin2 | Is wince 2003 == ce4? |
19:16.33 | cr2 | Kevin2: ~ |
19:17.38 | pH5 | cr2: don't know, I don't have decoded ba nk.exe at hand. |
19:18.01 | Kevin2 | cr2: Huh? |
19:18.30 | pH5 | I strongly expect it to be similar, but I don't think I have seen this gpio storage at 0xa0000c00 that ba uses anywhere in the magician disassembly. |
19:19.37 | cr2 | Kevin2: $\approx$ |
19:19.52 | cr2 | BabelO: http://mapcenter.cgpsmapper.com/maplist.php?author=6796 |
19:21.17 | Kevin2 | cr2: Did you see message from GSV earlier? Looks like there is another arm processor variant coming - qualcomm. |
19:21.56 | Kevin2 | That means we'll have samsung, ti, intel, and qualcomm processors to worry about.. |
19:23.55 | cr2 | Kevin2: afaik it runs l4 linux microkernel |
19:24.06 | cr2 | so you'l have linux on linux. |
19:24.46 | *** join/#htc-linux ellisway (n=ellis@83.167.181.177) |
19:27.23 | pH5 | :) wow, wince kernel code is pretty when you have spent some time looking at the mess that c++ produces |
19:30.04 | cr2 | pH5: the oem part :) |
19:30.26 | pH5 | aw right :) |
19:30.48 | Kevin2 | cr2: It would seem that HTC Kaiser runs the wince6 OS on the qualcomm "apps arm core". Do you really think they got wince to run under l4? |
19:32.01 | cr2 | Kevin2: and how is the phone controlled ? by wince6 ? |
19:33.51 | cr2 | http://www.embeddedstar.com/press/content/2005/11/embedded19219.html |
19:34.19 | pH5 | Kevin2: 80080bb4 is still in flash (80080bb4 might be uncached for 80bb4 I assume?) - there the wm2003 resume code loads the stack pointer and restores the gpios |
19:34.57 | pH5 | cr2: this is going to be complicated. I hope there will be docs for those chips. |
19:36.46 | cr2 | http://ertos.nicta.com.au/software/kenge/iguana-project/latest |
19:36.58 | cr2 | pH5: i doubt there will be docs. |
19:37.30 | cr2 | do you have docs for ti calypso ? |
19:39.17 | pH5 | cr2: just hoping for docs for the application processor part. |
19:41.03 | cr2 | omap850 ? |
19:41.53 | cr2 | i think it's better documented that this new one ;) |
19:42.23 | cr2 | they went microkernel for "trusted" computing. |
19:43.51 | cr2 | pH5: the stack pointer in flash ? |
19:46.14 | BabelO | goxboxlive: line 178 htcphonehardware.cpp |
19:47.29 | BabelO | cr2: good map :) but my town is not on this one |
19:49.45 | pH5 | cr2: no, it is loaded from 0x90000010. this is the code: http://rafb.net/p/E76Ynz15.html |
19:50.56 | goxboxlive | BabelO, yes, but what does it do? Will it shutdown linux with power button or? |
19:51.02 | pH5 | (at 881fc is WakeupProcess() that checks wakeup interrupts, etc.) |
19:51.38 | pH5 | all of this is still in flash :( |
19:51.49 | BabelO | goxboxlive: i think yes, power button is set by htcbuthandler |
19:54.08 | goxboxlive | BabelO, ok, but the strange part about it is that powerbutton in universal suspends the device, while with magician it doesnt do anything. hmmm maybe it has something to do with that i am not using htckbdhandler with magician. |
19:54.34 | BabelO | goxboxlive: i ll look |
19:55.26 | goxboxlive | i'll test to add htckbdhandlerwith magician also. |
19:55.27 | GSV | have you guys heard of HTC omni?!! |
19:56.42 | cr2 | pH5: but it's +0x10 from the start of RAM?? do you know the value ? |
19:57.59 | cr2 | GSV: it's vaporware. |
19:58.28 | GSV | cr2:what does that mean? |
19:59.20 | GSV | oh got it |
19:59.33 | GSV | but if it's true |
20:01.44 | cr2 | it will be a strange device. |
20:02.18 | cr2 | if you want a phone you go hermes, if you want a mini-notebook you go athena. |
20:02.36 | pH5 | cr2: currently the value of 0xa0000010 is ffffc764, but I can see the in the suspend code where it writes the current stack pointer to 0x90000010. |
20:02.56 | BabelO | goxboxlive: it is ok, it is in htcbuthandler.cpp #define PWR_BUTTON 116 |
20:04.15 | goxboxlive | right |
20:06.00 | cr2 | pH5: ffffc764 is a virtual address. |
20:06.21 | cr2 | pH5: looks like a kernel stack. |
20:08.57 | Kevin2 | bbl |
20:15.13 | *** join/#htc-linux rob_w (n=bob@M87a5.m.pppool.de) |
20:17.48 | *** join/#htc-linux Hawk|- (n=Hawk@pD9507AE1.dip0.t-ipconnect.de) |
20:26.50 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
20:51.00 | goxboxlive | hmm whats this: QAudioInput: snd_pcm_open: error -2 |
20:52.24 | BabelO | goxboxlive: maybe something missing here /dev/snd/* |
20:52.44 | goxboxlive | BabelO, have you manage to start OpieII yet? |
20:53.50 | BabelO | no, infinite loop due to DBMigrate i think |
20:53.57 | BabelO | so black screen |
20:54.17 | goxboxlive | BabelO, have you tried to start single apps with qws? |
20:54.17 | BabelO | what is your opieII revision number ? |
20:54.40 | BabelO | goxboxlive: yes, not working |
20:56.09 | goxboxlive | BabelO, i think 87-88. The qt and qte versions is 70-80 i think. I do not reconfigure qt and qte when i configure. It takes less time to compile |
20:56.19 | BabelO | i ll try mixer -qws, but always black screen |
20:56.31 | goxboxlive | BabelO, mixer is buggy |
20:57.12 | BabelO | all is buggy then, no app starting |
20:57.17 | goxboxlive | :-) |
20:57.45 | BabelO | what is your QWS_DISPLAY line ? |
20:57.59 | goxboxlive | i did an update to source now and i saw that ljp had added a lot from 88-89 |
20:58.06 | goxboxlive | hold i'll check |
20:59.14 | goxboxlive | BabelO, this is my startup script: http://pastebin.no/2185 (the QWS keyboard is just a test now. |
21:01.12 | BabelO | wouah :) i try your settings, i see thing that are written on my screen ( console mode) but no application |
21:03.32 | BabelO | it is a real VNC for command line mode ;) |
21:04.04 | goxboxlive | :-) |
21:04.56 | BabelO | goxboxlive: this line is new for me AudioVolumeManager::registerHandler "Media" "QPE/MediaServer/MediaVolumeControl" |
21:05.59 | goxboxlive | BabelO, yes i got it myself. The volume button manage to lower the volume, but not set the volume up. |
21:06.05 | goxboxlive | I can see it in mixer. |
21:06.22 | BabelO | ok that a progress ;) |
21:06.43 | goxboxlive | BabelO, i want to remove htckbd handler for magician. Is it just to dele the folders in htcphone? |
21:07.37 | BabelO | goxboxlive: you can yes, it is just a plugin in/opt/Qtopia/qt_plugins |
21:08.23 | goxboxlive | so i can remove kernelkeymap folder and htckbd folder .No need to edit any files |
21:08.39 | BabelO | yes, for magician |
21:09.04 | goxboxlive | good |
21:11.56 | cr2 | goxboxlive: but you have keys on magician ? |
21:12.00 | goxboxlive | BabelO, did you had any issues with mediaserver when you builded with helix and alsa? |
21:12.32 | goxboxlive | cr2, yes, but we have htcbutton handler. |
21:12.37 | BabelO | yes, it does not take include path from qtopia |
21:13.00 | goxboxlive | BabelO, anything about ATOMIC_OPERATORS? |
21:13.22 | BabelO | goxboxlive: yes need to add #define somewhere |
21:13.31 | BabelO | but i already do that before |
21:13.43 | goxboxlive | BabelO, ahh and you forgot of course |
21:14.07 | BabelO | yes i forgot i think i do it in middle of the night |
21:14.34 | goxboxlive | When i compile with helix it aplies patches for i586 and not for xscale. |
21:14.45 | BabelO | but it is not same issue i got, i look into my old source and did not find such define |
21:15.05 | goxboxlive | have you done a diff ? |
21:15.23 | goxboxlive | did you edit the helix sources or the OpieII sources? |
21:16.56 | goxboxlive | cr2, still here |
21:18.27 | goxboxlive | cr2, do you know why magician and universal show the year 1994 and not 2007? Is there a why it can be modified without syncing against a online timeserver? Could it be syncing against the gsm modem? |
21:20.45 | BabelO | goxboxlive: same on BA |
21:21.02 | BabelO | it is from internal RTC after reset |
21:22.13 | goxboxlive | BabelO, but i have noticed that with Opie(1) it shows right time and year. |
21:22.48 | cr2 | goxboxlive: it's how wince uses the RTC |
21:23.04 | goxboxlive | cr2, ok, anyway to override it? |
21:23.24 | cr2 | what does 'date' say ? |
21:23.30 | cr2 | on the tty0 |
21:36.59 | BabelO | cr2: this is froom log |
21:37.03 | BabelO | sa1100-rtc sa1100-rtc: setting the system clock to 1994-09-02 12:01:28 (778507288) |
21:39.09 | cr2 | BabelO: look at the sa1100-rtc source |
21:40.55 | goxboxlive | root@magician:~$ cat /sys/class/rtc/rtc0/date |
21:40.55 | goxboxlive | 1994-10-02 |
21:41.54 | goxboxlive | Thme minutes are allmost correct (differ + 1 hour |
21:46.31 | cr2 | 1 hour is ok. |
21:51.15 | goxboxlive | cr2, yes absolute, but 13 years is not that ok :-) |
21:57.09 | goxboxlive | BabelO, still up, listen to this: |
21:57.21 | BabelO | music ? |
21:57.23 | BabelO | :) |
21:58.35 | goxboxlive | BabelO, so you are saying that you did the #ifdef for a long time ago. And you builded the new svn release now without changing anything with helix/mediaserver. So the i guess you must have done something with the helix sources. |
21:58.52 | goxboxlive | Could you do for me, a diff between your helix sources and a new fresh one? |
22:03.30 | BabelO | goxboxlive: i think i do same change than you, adding a #define in helix.h, wait i ll look |
22:04.09 | goxboxlive | BabelO, i added in hxcom.h |
22:05.13 | cr2 | BabelO: isn't this message from /etc/init.d/* ? |
22:05.16 | cr2 | setting the system clock |
22:05.26 | goxboxlive | But i added #define NO_ATOMIC_OPERATORS |
22:06.16 | cr2 | no. |
22:06.17 | goxboxlive | cr2, i noticed for a while ago, it always used the time and date the image where builded |
22:06.32 | goxboxlive | but thats a while ago |
22:07.21 | goxboxlive | there is noi file named helix.h |
22:07.41 | BabelO | cr2: it is in dmesg with other line like this |
22:07.51 | BabelO | 7>[ 0.710000] sa1100-rtc sa1100-rtc: rtc intf: sysfs |
22:07.52 | BabelO | <7>[ 0.710000] sa1100-rtc sa1100-rtc: rtc intf: proc |
22:07.52 | BabelO | <7>[ 0.710000] sa1100-rtc sa1100-rtc: rtc intf: dev (254:0) |
22:07.52 | BabelO | <6>[ 0.710000] sa1100-rtc sa1100-rtc: rtc core: registered sa1100-rtc as rtc0 |
22:07.52 | BabelO | <6>[ 0.930000] sa1100-rtc sa1100-rtc: setting the system clock to 1994-09-02 12:01:28 (778507288) |
22:08.21 | BabelO | and yes it is comming from etc/init.d |
22:16.41 | goxboxlive | cr2, did you see my post earlier about bluetooth mouse ? I had it working in OpieII with Universal :-) |
22:18.09 | cr2 | goxboxlive: yes .) |
22:29.27 | BabelO | goxboxlive: i do same thing than you, i add it to hxcom.h |
22:29.28 | goxboxlive | BabelO, ok, then i dont understand why it works for you and not for me. I have opie-4.2.2 image at ~htcpxa. When you have time, could you test it to see if helix layer is working for on BA. |
22:29.29 | goxboxlive | It's not an rootfs image, just the /opt directory |
22:29.29 | goxboxlive | BabelO, http://www.linuxtogo.org/~htcpxa/htcmagician/images/OpieII/no_rootfs/opieII-4.2.2.tar.bz2 |
22:29.30 | BabelO | goxboxlive: i ll test it now |
22:29.31 | goxboxlive | BabelO, cool :-) |
22:29.31 | goxboxlive | Then i stay up a little longer to hear if it works or not |
22:31.23 | BabelO | unpacking |
22:34.23 | BabelO | starting qpe |
22:35.58 | BabelO | oups, battery down, i ll change it |
22:37.33 | BabelO | goxboxlive: your script is in bz2 ? |
22:37.55 | goxboxlive | BabelO, yes in the root of it |
22:42.10 | BabelO | goxboxlive: starting slowly |
22:42.31 | goxboxlive | BabelO, yes it's slow. It's buggy |
22:42.47 | BabelO | ah mouse cursor appear |
22:43.09 | goxboxlive | :-) |
22:43.21 | goxboxlive | BabelO, i get a lot of this: |
22:43.23 | goxboxlive | Segmentation fault |
22:43.24 | goxboxlive | QAbstractSocket::waitForReadyRead() is not allowed in UnconnectedState |
22:43.24 | goxboxlive | ApplicationLayer: Application layer server unexpectedly terminated. |
22:43.24 | goxboxlive | ApplicationLayer: Application layer server unexpectedly terminate |
22:43.50 | BabelO | yes segfault |
22:43.57 | BabelO | for me too |
22:44.35 | goxboxlive | try again, it will start at last |
22:45.22 | BabelO | yes started |
22:45.55 | BabelO | segfault again :( |
22:46.11 | BabelO | same error as you |
22:47.07 | goxboxlive | :-) |
22:47.15 | BabelO | ok starting again with qtopia theme |
22:47.22 | BabelO | i notice it is more stable |
22:47.37 | goxboxlive | BabelO, just so you are aware of it,it's the first buildable image with 4.2.2 |
22:48.05 | goxboxlive | yes it's unstable, but ljp told me that he had a crappy update script so all pathces wherent added. |
22:48.27 | BabelO | ok, so i ll try mediaplayer ? |
22:49.52 | goxboxlive | yes please |
22:50.29 | BabelO | lot of segfault :) |
22:51.05 | BabelO | ok i start mediaplayer by hand |
22:52.24 | BabelO | no music :( |
22:53.26 | goxboxlive | BabelO, i c |
22:53.34 | BabelO | no music in database |
22:53.42 | BabelO | need to change storage.conf |
22:53.53 | goxboxlive | well, then i have no idea what it is. It must have something to do with the Atomic operators |
22:54.10 | goxboxlive | well, it's late. g'night |
22:54.18 | BabelO | good night |
22:57.07 | BabelO | yes sstrange not working for me too |
23:13.51 | *** join/#htc-linux ellisway (n=ellis@83.167.181.177) |