| 04:46.58 | *** join/#htc-linux eremias (n=fdsfs@ACB0ADA8.ipt.aol.com) |
| 06:10.13 | *** join/#htc-linux RoEn_PC (n=roen@p54a65e8b.dip.t-dialin.net) |
| 06:26.11 | *** join/#htc-linux ellisway (n=ellis@host-87-74-241-174.bulldogdsl.com) |
| 06:36.18 | *** join/#htc-linux tudenbart (n=willi@xdsl-81-173-174-115.netcologne.de) |
| 07:02.54 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
| 07:03.37 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
| 07:09.16 | *** join/#htc-linux Ro_En_PC (n=roen@p54a6728e.dip.t-dialin.net) |
| 07:38.13 | *** join/#htc-linux rob_w (n=bob@p85.212.10.79.tisdip.tiscali.de) |
| 08:02.19 | *** join/#htc-linux psokolovsky (n=psokolov@82.193.98.7) |
| 08:39.17 | *** part/#htc-linux lilliput (n=Touriste@82-47-146-8.cable.ubr11.brad.blueyonder.co.uk) |
| 08:40.15 | *** join/#htc-linux eremias (n=fdsfs@ACB0ADA8.ipt.aol.com) |
| 08:40.15 | psokolovsky | Hi! |
| 08:40.31 | psokolovsky | cr2: going to rename himalaya to htchimalaya, as we agreed |
| 08:43.45 | *** join/#htc-linux Ranpha1 (n=Ranpha@hlm-n-73ed.adsl.wanadoo.nl) |
| 08:45.29 | Ranpha1 | Stupid question. I have a sprae sd-card and i really like to have linux on my qtek 9000. But do i need to replace the WM5 rom or can i run it when i insert the card pull it out and have WM5 back? |
| 08:54.17 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
| 09:46.13 | cr2 | Ranpha1: linux run from wince5 works as a wince program. no known side effects afaik. |
| 09:53.38 | *** join/#htc-linux twostupiddogs (n=twostupi@84.18.144.39) |
| 09:55.11 | cr2 | twostupiddogs: there are no problems to build opie on 64bit cpus since 1.5 years |
| 09:55.26 | twostupiddogs | hi all. |
| 09:55.35 | cr2 | i don't have any other cpus :) 32bit only on the arm phones . |
| 09:56.04 | twostupiddogs | I'm trying to build opie-image for htcuniversal with an x86_64, but have some troubles building qt |
| 09:56.16 | cr2 | sorry, the old vax is 32bit too ;) |
| 09:56.20 | cr2 | works for me. |
| 09:56.44 | cr2 | are these problems linked to 32bit ? |
| 09:56.47 | twostupiddogs | don't think so. |
| 09:57.01 | twostupiddogs | it failes during configuration of the qt |
| 09:57.12 | cr2 | i'm updating with monotone |
| 09:57.27 | twostupiddogs | me too |
| 09:57.30 | cr2 | did the last opie compilation 2 days ago. |
| 09:58.09 | twostupiddogs | I set it up just few days ago, but never finished the build |
| 09:58.49 | cr2 | 33 revs |
| 09:59.03 | twostupiddogs | I had to remove the locale for qemu problems with x86_64 on glibc, and compilation wen on |
| 09:59.16 | cr2 | sometimes things break, but it's usually working soon. |
| 09:59.24 | cr2 | yes, it's an old problem. |
| 09:59.26 | twostupiddogs | doing an update now too be sure |
| 09:59.40 | cr2 | i don't understand why it is still not solved. |
| 10:00.04 | twostupiddogs | well, at this time locales ain't a particular problem to me. |
| 10:00.38 | twostupiddogs | I just started the OE adventure so am not that expert. how do you get the version? |
| 10:01.06 | cr2 | which version ? |
| 10:01.32 | twostupiddogs | sorry revision |
| 10:02.05 | cr2 | with monotone. |
| 10:02.11 | twostupiddogs | yes. |
| 10:02.14 | cr2 | actually i compile just for fun. |
| 10:02.37 | cr2 | it's pH5 and goxboxlive that do serious work there. |
| 10:02.40 | twostupiddogs | right now it's like this for me too, but looking forward to help out |
| 10:02.53 | cr2 | i dont' try to do everything :) |
| 10:03.00 | twostupiddogs | well you look pretty active anyway :-) |
| 10:03.27 | cr2 | i know that it compiles, but never use the output myself. |
| 10:03.48 | twostupiddogs | I did an mnt update, but compilation still stops in the qt package. would it help if I paste the output? |
| 10:04.19 | cr2 | yes. |
| 10:05.00 | twostupiddogs | FATAL: Configuring qt failed. EXTRA_OECONF was -system-jpeg -system-libpng -system-zlib -no-qvfb -no-xft -no-vnc -gif -xplatform linux-gnueabi-arm-g++ -qconfig qpe -depths 8,16,32 -static -thread |
| 10:05.47 | twostupiddogs | checking the log |
| 10:06.10 | cr2 | compiling myself. |
| 10:06.32 | twostupiddogs | actually that's all... |
| 10:07.40 | twostupiddogs | I have a bunch of output telling me all the available compilers |
| 10:09.38 | cr2 | oh. it's recompiling gtk and some weird gnome stuff. |
| 10:09.51 | twostupiddogs | with opie? |
| 10:10.26 | cr2 | somebody who writes gint instead of int, has included gdk dependency for bluez-something ;) |
| 10:10.28 | twostupiddogs | maybe this might be the problem... " |
| 10:10.28 | twostupiddogs | No static library support for platform/compiler linux-gnueabi-arm-g++-stati |
| 10:10.28 | twostupiddogs | c |
| 10:10.28 | cr2 | yes. |
| 10:11.03 | cr2 | i've seen that problem half a year ago. |
| 10:11.24 | twostupiddogs | you remember the solution? |
| 10:11.28 | twostupiddogs | should I google? |
| 10:11.47 | cr2 | it was solved somehow. |
| 10:11.57 | cr2 | btw, which local.conf are you using ? |
| 10:12.14 | cr2 | pango, cairo. wtf ? |
| 10:12.30 | twostupiddogs | well the original in OE with the machine modifications.. |
| 10:12.53 | twostupiddogs | I followed the howto on angstrom... |
| 10:13.30 | cr2 | can you pastebin it ? |
| 10:13.56 | twostupiddogs | ok, but have to remove all the comments that are in, wait a sec. |
| 10:15.16 | twostupiddogs | DL_DIR = "/OE/downloads" |
| 10:15.20 | twostupiddogs | BBFILES := "/OE/org.openembedded.dev/packages/*/*.bb" |
| 10:15.24 | twostupiddogs | BBMASK = "" |
| 10:15.28 | twostupiddogs | PREFERRED_PROVIDERS = "virtual/qte:qte virtual/libqpe:libqpe-opie" |
| 10:15.32 | twostupiddogs | PREFERRED_PROVIDERS += " virtual/libsdl:libsdl-x11" |
| 10:15.36 | twostupiddogs | PREFERRED_PROVIDERS += " virtual/${TARGET_PREFIX}gcc-initial:gcc-cross-initial" |
| 10:15.40 | twostupiddogs | PREFERRED_PROVIDERS += " virtual/${TARGET_PREFIX}gcc:gcc-cross" |
| 10:15.52 | twostupiddogs | PREFERRED_PROVIDERS += " virtual/${TARGET_PREFIX}g++:gcc-cross" |
| 10:15.56 | twostupiddogs | MACHINE = "htcuniversal" |
| 10:15.56 | twostupiddogs | DISTRO = "angstrom-2007.1" |
| 10:15.56 | twostupiddogs | ENABLE_BINARY_LOCALE_GENERATION = "0" |
| 10:15.56 | twostupiddogs | IMAGE_FSTYPES = "jffs2 tar" |
| 10:15.56 | twostupiddogs | BBINCLUDELOGS = "yes" |
| 10:15.57 | twostupiddogs | maybe I should change something in the PROVIDERS... |
| 10:16.25 | *** join/#htc-linux BabelO (n=kvirc@lun34-2-82-238-28-28.fbx.proxad.net) |
| 10:16.52 | cr2 | i think my local.conf is in the xanadux CVS. |
| 10:17.11 | cr2 | you must only adapt the path |
| 10:17.18 | twostupiddogs | ok I'll try checking that out |
| 10:18.02 | twostupiddogs | tnks |
| 10:20.58 | cr2 | no. it seems to be an old version. |
| 10:21.13 | cr2 | http://xanadux.cvs.sourceforge.net/xanadux/oe |
| 10:23.25 | twostupiddogs | cr2: just to know, you use htcuniversal.conf or local.conf or both? |
| 10:26.26 | cr2 | local.conf |
| 10:26.37 | cr2 | i'll put my current version there. |
| 10:27.10 | cr2 | gtk+ |
| 10:27.19 | twostupiddogs | yep, 'cause I think that distro familiar-unstable isn't the right one... |
| 10:27.19 | cr2 | that's just insane. |
| 10:32.59 | twostupiddogs | the local.conf there looks pretty old... |
| 10:33.28 | twostupiddogs | cr2: you there? |
| 10:34.24 | twostupiddogs | bbiaf |
| 10:35.23 | cr2 | done. |
| 10:35.38 | cr2 | gnome-common-2.4.0-r0 |
| 10:35.47 | cr2 | now i really say WTF ? |
| 10:36.00 | *** join/#htc-linux ballessay (n=balles@client420.amh.kn.studentenwohnheim-bw.de) |
| 10:37.01 | twostupiddogs | agreed |
| 10:37.51 | cr2 | ok, we will clean this crap away eventually. |
| 10:38.24 | twostupiddogs | ok. what should I do to get the bulid done? |
| 10:39.07 | cr2 | did you create a working directory ? |
| 10:39.17 | cr2 | i have it in ~/stuff |
| 10:39.32 | cr2 | not /stuff like in the oe docs. |
| 10:40.03 | cr2 | libglade |
| 10:40.08 | twostupiddogs | yes I created /OE and then a build one |
| 10:40.16 | cr2 | my eyes are hurt. |
| 10:40.33 | cr2 | so your working dir is /OE ? |
| 10:40.39 | twostupiddogs | yes |
| 10:41.06 | twostupiddogs | like the angstrom sayd, I'll move it once it is ok. |
| 10:41.20 | twostupiddogs | wiki I mean |
| 10:41.23 | cr2 | so you have /OE/org.openembedded.dev |
| 10:41.29 | twostupiddogs | yep |
| 10:41.46 | twostupiddogs | and a /OE/build also |
| 10:41.46 | cr2 | then create /OE/build/conf |
| 10:41.53 | twostupiddogs | done |
| 10:42.01 | cr2 | and put local.conf in this directory. |
| 10:42.04 | cr2 | ok. |
| 10:42.10 | cr2 | edit local.conf |
| 10:42.18 | twostupiddogs | which one? the one in CVS? |
| 10:42.32 | cr2 | get the one from CVS and edit it. |
| 10:42.34 | cr2 | BBFILES := "${HOME}/stuff/org.openembedded.dev/packages/*/*bb" |
| 10:42.42 | cr2 | should be for you |
| 10:42.49 | cr2 | BBFILES := "/OE/org.openembedded.dev/packages/*/*bb" |
| 10:43.16 | twostupiddogs | I've already done it, but it complains about the familiar-unstable... |
| 10:43.21 | cr2 | you may change this one too |
| 10:43.24 | cr2 | DL_DIR = "${HOME}/sources" |
| 10:43.43 | cr2 | huh ? |
| 10:43.45 | cr2 | DISTRO = "angstrom-2007.1" |
| 10:43.57 | cr2 | what local.conf do you have ? |
| 10:44.02 | twostupiddogs | ok i missed the distro. |
| 10:44.15 | cr2 | http://xanadux.cvs.sourceforge.net/xanadux/oe |
| 10:44.24 | twostupiddogs | the one from xanadux cvs now |
| 10:44.29 | twostupiddogs | wait tring a build |
| 10:44.30 | cr2 | README needs an update too. |
| 10:45.44 | twostupiddogs | by the way, to use psyco should I get a 32bit version of perl? does it speed up the process a lot? |
| 10:46.38 | cr2 | don't use psyco |
| 10:46.52 | cr2 | i think it does not work on 64bit. |
| 10:47.16 | cr2 | lkcl said it is not doing anything useful anyway. |
| 10:47.31 | cr2 | ok. opie-image built. |
| 10:48.17 | twostupiddogs | configuring now... |
| 10:48.32 | twostupiddogs | but not the qt though. |
| 10:49.05 | twostupiddogs | ho, missed that, qt did configure ... |
| 10:49.07 | twostupiddogs | tnx |
| 10:50.23 | twostupiddogs | qte_mt compile ... |
| 10:51.29 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.250) |
| 10:54.20 | twostupiddogs | ok as this is gonna take long, going to have a pizza for lunch. |
| 10:54.29 | twostupiddogs | bbl. cr2: thanks a lot |
| 10:54.39 | cr2 | :) |
| 10:58.40 | BabelO | hi all |
| 10:59.30 | BabelO | cr2: is there any thing i can to for lcd resume ? another dump for spi maybe ? |
| 11:00.59 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
| 11:06.22 | cr2 | BabelO: no, if you have a boardid5 |
| 11:08.32 | BabelO | cr2: ok :( sniff i would like to help fo this :( i m adding gpio to sound |
| 11:08.41 | cr2 | BabelO: i'm slowly thinking of opie2. that will require a qt4 adaptation of roadmap. when i'll be finished, i'll switch. |
| 11:08.53 | cr2 | ok. |
| 11:10.44 | BabelO | cr2: i finish to add gpio and swithc to my OPIEII version adding gsmd instead of gnokii maybe, or i need to add a new HTC model to gnokii |
| 11:11.42 | cr2 | i don't think it's wise to drop gnokii completely. |
| 11:12.20 | cr2 | do you know how the greenphone gui looks like ? |
| 11:12.59 | cr2 | lkcl has disappeared, and i'll better buy athena instead of the greenphone :) |
| 11:13.39 | cr2 | imho, the greenphone gui look-and-feel should be cloned, so the users are not confused. |
| 11:14.57 | BabelO | :) athena wow, greenphone gui is really nice... they use xml approach to define gui and it is fully included in OPIEII too |
| 11:15.37 | cr2 | ok, where can i see the screenshot ? |
| 11:16.44 | BabelO | there is some screenshot in Qtopia documentation of themeview class |
| 11:17.10 | cr2 | BabelO: ok. i'll remove bleabgel_audio.c, because it's obsolete. |
| 11:17.15 | BabelO | i ve tried greephone vmware demo too |
| 11:17.20 | cr2 | hehe. |
| 11:17.31 | BabelO | cr2: ok and blueangel_battery is obsolete too |
| 11:17.35 | cr2 | ok. |
| 11:18.05 | cr2 | blueangel_leds is not nice, but we don't have a replacement yet. |
| 11:18.45 | BabelO | yes and we don't have same charging led as in wince |
| 11:20.00 | BabelO | cr2: http://doc.trolltech.com/qtopia2.2/html/qtopiatheming.html |
| 11:20.17 | BabelO | there is some screen and themed xml |
| 11:21.25 | cr2 | yes :) |
| 11:21.46 | cr2 | blueangel_tsc2200.c |
| 11:21.51 | cr2 | is it still needed ? |
| 11:23.01 | BabelO | cr2: yes it is setup of tsc2200 soc |
| 11:23.41 | cr2 | ok. |
| 11:23.43 | cr2 | menuconfig MACH_BLUEANGEL_AUDIO |
| 11:23.50 | cr2 | this is obsolete now ? |
| 11:24.00 | BabelO | yes |
| 11:24.14 | cr2 | ok. |
| 11:24.25 | cr2 | menuconfig MACH_BLUEANGEL_BATTERY ? |
| 11:25.04 | BabelO | yes, i use it for blueangel_power ;) |
| 11:25.26 | cr2 | maybe it should be renamed BLUEANGEL_POWER ? |
| 11:25.44 | cr2 | MACH_BLUEANGEL_KEYPAD |
| 11:25.50 | BabelO | yes, i don't difference between power and battery :) |
| 11:25.55 | cr2 | <PROTECTED> |
| 11:25.55 | cr2 | <PROTECTED> |
| 11:26.01 | cr2 | is it still true ?? |
| 11:26.39 | BabelO | let me try again keyboard, last time i try it, it works, but each time i press a key, got a lot of repeat |
| 11:28.04 | cr2 | POWER "Does not work yet" |
| 11:29.39 | BabelO | humm power works.. |
| 11:30.01 | BabelO | it hang after some times because of ds1wm issue |
| 11:30.20 | cr2 | ok. |
| 11:30.25 | cr2 | DPRAM is broken ? |
| 11:31.21 | BabelO | yes i think, no dpram wwhen i try gprs |
| 11:31.42 | cr2 | afaik nobody tried writing to PXA flash :) it may not work, and some gpio enable is needed. |
| 11:31.52 | cr2 | no dpram or no ppp ? |
| 11:32.06 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
| 11:33.24 | BabelO | no ppp dpram is found and tts/8 is added |
| 11:33.38 | BabelO | s/ppp/ppp,/ |
| 11:33.40 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
| 11:34.02 | cr2 | ok, so it's buggy. |
| 11:34.21 | cr2 | suspend "Does nothing useful right now" |
| 11:35.30 | cr2 | what should we do with PCMCIA ? |
| 11:35.42 | BabelO | cr2: yes suspend is not working because of LCD resume, if we comment out lcd_hw_off() function it seems to work |
| 11:35.48 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
| 11:36.56 | BabelO | pcmcia, i don't know, you have added missing gpio in htcblueangel_acx.c... on my platform, i ve used the both missing gpio to drive ACX power over pcmcia insert and eject |
| 11:37.14 | BabelO | and we can see acx card into pcmcia module of opie or gpe |
| 11:37.23 | cr2 | ok. |
| 11:38.22 | BabelO | how it is working on universal ? you can set up and set down easyly ? |
| 11:39.06 | goxboxlive | test |
| 11:39.23 | BabelO | test ok ;) |
| 11:39.56 | goxboxlive | I have ust installed the new kubuntu distro so i have to setup xchat again :-) |
| 11:40.04 | goxboxlive | hi btw |
| 11:40.07 | cr2 | BabelO: yes. there are some programming bugs, but they are irrelevant in t his context. |
| 11:40.34 | cr2 | hi goxboxlive. do you have a new notebook that does not crash because of overheating ? |
| 11:40.45 | goxboxlive | :-) no |
| 11:41.00 | cr2 | goxboxlive: i'm thinking about athena :) |
| 11:41.14 | goxboxlive | my good old ine, It's not that old either, it's just 2 1/2 year |
| 11:41.17 | cr2 | BabelO: do we really need 159 modules ??? |
| 11:41.40 | BabelO | cr2: 139 modules ;) but no not needed |
| 11:41.43 | goxboxlive | cr2: Are you going to buy it? |
| 11:41.49 | BabelO | need to cleanup maybe |
| 11:45.13 | *** join/#htc-linux twostupiddogs (n=towstupi@84.18.144.39) |
| 11:45.53 | cr2 | goxboxlive: i'm thinking. |
| 11:45.56 | BabelO | cr2: asylumed say to me that it use BA with rndis connected to windowws host with a samba share |
| 11:47.50 | BabelO | cr2: athena demo video is really nice |
| 11:50.48 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
| 11:50.53 | *** join/#htc-linux gox (n=goxboxli@176.84-48-210.nextgentel.com) |
| 11:53.44 | cr2 | BabelO: but can one telnet over rndis now ? |
| 11:55.58 | BabelO | yes i think he can see a share on is BA in Windows |
| 11:56.00 | BabelO | so telnet is a formality |
| 11:57.29 | cr2 | ok. |
| 11:57.43 | cr2 | compiling the cleaned tree now. |
| 11:57.50 | BabelO | ok ;) |
| 11:58.10 | BabelO | maybe we have to remove drivers/input/touchscreen/tsc2200_ts.c too |
| 11:58.24 | cr2 | can we revise the old ramdisk ? |
| 11:59.15 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
| 11:59.33 | BabelO | ramdisk ? i don't use ramdisk |
| 12:00.28 | cr2 | yes, but where do you install the modules ? |
| 12:00.48 | BabelO | on sd card |
| 12:00.50 | cr2 | i'd like to have a fallback option. |
| 12:00.59 | cr2 | imho it's a bad idea. |
| 12:01.31 | BabelO | cr2: why a bad idea to put modules on sd card ? |
| 12:01.38 | cr2 | if we will get the LCD resume working, i'll think about linux flashing. |
| 12:01.52 | cr2 | because the kernels change too often. |
| 12:02.14 | cr2 | and the same image can be reused on multiple devices. |
| 12:02.29 | cr2 | while the modules (even with the same name) are incompatible. |
| 12:03.17 | BabelO | we can compile without setting kernel version module into it |
| 12:03.24 | cr2 | i'd like to test on 4 devices without taking care about the modules. |
| 12:03.49 | cr2 | you can't use pxafb from pxa27x on pxa26x, for example. |
| 12:03.57 | cr2 | or will not be able soon ;) |
| 12:04.29 | cr2 | ok, compiled. |
| 12:08.04 | cr2 | you can update from CVS now. |
| 12:08.50 | BabelO | ok i do that in another directory and if it works i use this new base |
| 12:09.13 | cr2 | ok. |
| 12:12.10 | BabelO | you update defconfig too ? |
| 12:14.15 | cr2 | done. |
| 12:17.46 | BabelO | how this ramdisk works ? modules are embedded into zImage, or it is a standard initrd ? |
| 12:19.00 | cr2 | standard initrd |
| 12:19.06 | BabelO | cr2: do we have adcm i2c modules code somewhere ? |
| 12:19.29 | cr2 | get one from hh.org, gzip -d, and mount as ext2 over loop |
| 12:19.37 | cr2 | i think yes, i'll have a look. |
| 12:19.58 | cr2 | i think wiki had a link. |
| 12:22.34 | BabelO | link was removed |
| 12:23.46 | BabelO | he he my hardrive is full, but lot of interesting thing ;) i have the patch locally ;) |
| 12:24.03 | BabelO | adcm2650 i2c module |
| 12:24.09 | cr2 | look here |
| 12:24.15 | cr2 | http://wiki.openezx.org/CodeAnalysis |
| 12:24.28 | cr2 | it's also in the motorola a780 kernel |
| 12:24.51 | cr2 | hm 2700 |
| 12:25.00 | BabelO | he he there is a 22c module and a video modules for this chips |
| 12:25.02 | cr2 | and not very clean code. |
| 12:25.22 | cr2 | yes, the i2c side is easy. |
| 12:25.39 | cr2 | ati capture port is painful. |
| 12:26.37 | BabelO | but it is not in a780 module, there is adcm2700 and 3800 but no 2650 |
| 12:32.00 | BabelO | cr2: i send to you 2.6.9 kernel patch for adcm and pxa_camera |
| 12:32.15 | BabelO | it is old patch from link in the wiki |
| 12:33.00 | cr2 | ok, maybe it's also available somewhere else ? |
| 12:34.18 | BabelO | it is to keep it somewhere, maybe i had adcm2650 i2c chip detect if it is not hard |
| 12:35.03 | cr2 | ok, let's find the latest version, and use it locally. |
| 12:35.22 | cr2 | CONFIG_BLUEANGEL_CAMERA or something like that. |
| 12:35.43 | BabelO | yes |
| 12:38.19 | *** join/#htc-linux twostupiddogs (n=towstupi@84.18.144.39) |
| 12:38.21 | BabelO | last time i try to apply this patch, got lot of error |
| 12:38.37 | BabelO | but if i remember, i can load adcm module |
| 12:41.41 | cr2 | check the a780 patches |
| 12:41.43 | cr2 | i2c_adcm2650_inc_use |
| 12:41.58 | cr2 | i think these are not needed anymore |
| 12:42.01 | *** join/#htc-linux twostupiddogs (n=towstupi@84.18.144.39) |
| 12:42.30 | *** join/#htc-linux twostupiddogs_ (n=towstupi@84.18.144.39) |
| 12:42.34 | *** part/#htc-linux twostupiddogs (n=towstupi@84.18.144.39) |
| 12:43.21 | *** join/#htc-linux twostupiddogs (n=towstupi@84.18.144.39) |
| 12:43.23 | *** part/#htc-linux twostupiddogs_ (n=towstupi@84.18.144.39) |
| 12:44.34 | *** join/#htc-linux rob__w (n=bob@p213.54.35.44.tisdip.tiscali.de) |
| 12:44.59 | cr2 | this driver uses pxa27x-cif |
| 12:45.09 | cr2 | and needs a big rewrite. |
| 12:45.52 | BabelO | on ba we need imageon instead of this ? |
| 12:46.55 | cr2 | yes. |
| 12:46.59 | cr2 | +// Sets the bit to 60 Hz; sample duration is 25 |
| 12:46.59 | cr2 | +WS( 0x0e, 0x19); |
| 12:47.16 | cr2 | you should do extensive i2c traces, |
| 12:47.27 | cr2 | otherwise we can blow this sensor ;) |
| 12:48.23 | cr2 | it's probably easier to document the chip using this driver first. |
| 12:48.52 | BabelO | yes, but first step is to have a loadable module for this and can detect chip when powered on |
| 12:49.14 | cr2 | nice code: |
| 12:49.18 | cr2 | +// Note!! Document is wrong. 0x0076 shall be 0x0074 instead. |
| 12:49.18 | cr2 | +RL( STATUS_FLAGS, 0x0008, 0x0008 ); |
| 12:49.19 | psokolovsky | cr2: good cleanup on himalaya! I assume you didn't kill too much ;-) |
| 12:50.02 | cr2 | psokolovsky: no, the drivers are the same as on blueangel, and the gpio config is documented. |
| 12:50.09 | psokolovsky | cool! |
| 12:50.20 | psokolovsky | I hope uni is next ;-) |
| 12:50.57 | cr2 | :) |
| 12:51.04 | cr2 | asic3 ? |
| 12:52.14 | cr2 | this &foobar_asic3.dev stuff is sick. can we reuse the same trick as in asic3_mmc with pdev->dev.parent ? |
| 12:52.16 | psokolovsky | cr2: I mean deprecated drivers in htcuni dir |
| 12:52.20 | cr2 | ok. |
| 12:52.29 | psokolovsky | cr2: reuse where? |
| 12:52.43 | cr2 | in other asic3 slave drivers. |
| 12:53.09 | psokolovsky | cr2: which exactly? |
| 12:53.28 | cr2 | i will really need asic3_spi and asic3_pwm |
| 12:53.35 | cr2 | ok, it's a bit another issue. |
| 12:54.00 | psokolovsky | cr2: well sure, use parent-> in them. they just don't exist yet ;-) |
| 12:54.31 | cr2 | i'm asking because you are doing a major asic3_base cleanup |
| 12:54.47 | cr2 | maybe we should keep in mind the asic3 API too. |
| 12:56.49 | cr2 | BabelO: we can add the adcm2650 headers anyway. |
| 12:57.09 | cr2 | register definitions & such. |
| 12:59.02 | psokolovsky | cr2: I do syntactic cleanup mostly, preparing it for mainline submit. |
| 12:59.34 | psokolovsky | cr2: I want to renaming GPIO accessors and add timebase for LED setter though. |
| 13:00.29 | bd2 | psokolovsky, I guess it's pointless to think about mainline stuff, before we resolve dependency issues, no? |
| 13:00.49 | psokolovsky | bd2: Hi! which deps? |
| 13:01.12 | cr2 | bd2: i agree that it's too early to include it into mainline. |
| 13:01.19 | bd2 | psokolovsky, look for hx4700_power2.c you can't merge it with hx4700.c because of asic3 stuff |
| 13:01.56 | psokolovsky | bd2: so what? h4000_udc is also kinda not mergeable because of that. so what? |
| 13:02.06 | cr2 | hx4700_acx too. |
| 13:02.51 | cr2 | i only wish there was a better asic3 -> such_driver communication mechanism. |
| 13:02.53 | bd2 | psokolovsky, so we should implement something in soc core, that will register asic3 dependant *devices* only when asic3_base actually loaded |
| 13:03.16 | cr2 | bd2: they can do request_module(asic3_base) |
| 13:03.38 | cr2 | and fail to load if it fails. |
| 13:04.17 | bd2 | cr2, asic3_base module registers base driver. the driver needs asic3 device, which should be loaded at proper order too |
| 13:04.36 | cr2 | ok. |
| 13:04.49 | psokolovsky | bd2: I see no connection first of all, secondly, it's crappy-linux's not our problem. stupid linux doesn't provide any dependency management in kernel, offloading it to userspace. so, let's have it in userspace. |
| 13:05.00 | bd2 | so, it's pointless to request asic3_base, because it just registers driver (which may, or may not bound to the device) |
| 13:05.58 | bd2 | psokolovsky, we can pass to platform data devices which should be registered after asic3_probe() |
| 13:06.14 | cr2 | well, how do we solve it ? request_module() is a usespace way of doing things. |
| 13:06.27 | psokolovsky | bd2: what you mean "we can"? we *do*! |
| 13:07.50 | cr2 | psokolovsky: ? can you boot with asic3 not loaded ? |
| 13:08.06 | bd2 | psokolovsky, hmmmm. wait wait. so, we can put hx4700_power to hx4700_asic3_devices, and it will load it in proper time?? |
| 13:08.12 | psokolovsky | cr2: I have no idea, never tried. |
| 13:08.33 | psokolovsky | bd2: kinda can, if there're no other pre-dependencies. |
| 13:09.06 | bd2 | hmmm.. I should really try that, thus we *could* merge power to hx4700 |
| 13:09.09 | psokolovsky | bd2: issues is exactly few pre-dependencies, and the fact that we need to make "proxy" device and driver, instead of mere module. |
| 13:09.09 | bd2 | .c |
| 13:09.55 | psokolovsky | bd2: see for example h4000_bl.c vs hx4700_bl.c. mine is *driver*. yours is *module* |
| 13:10.48 | cr2 | psokolovsky: all the slave drivers should oops, i guess. |
| 13:10.51 | bd2 | s/*driver*/*device*/ ;-) but yeah |
| 13:11.17 | bd2 | psokolovsky, so, you can put it in h4000.c... that's great |
| 13:11.35 | bd2 | I'll do it for hx4700, including power stuff |
| 13:11.52 | cr2 | because they attempt to manipulate the asic3 registers directly, which is not mapped. |
| 13:12.22 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.250) |
| 13:12.38 | psokolovsky | bd2: ah damn, yeah, it's no longer a driver ;-). so well, you know, static loading issues of asic3 vs friends are solved now. and module loading is compeltely another matter. |
| 13:13.31 | psokolovsky | cr2: what exactly wants to manipulate? and I wrote about that some time ago to k-d. there's way to solve it too. |
| 13:13.37 | bd2 | yup.. so, everything is great, I guess. sorry for the panic |
| 13:14.07 | psokolovsky | cr2: solve, for static case. because there's nothing to solve for module case. just load modules in write order, voila. |
| 13:14.14 | psokolovsky | ok |
| 13:14.49 | cr2 | psokolovsky: if a random driver calls asic3_*** , and asic3_base is not registered, then it's an oops() |
| 13:14.58 | cr2 | ok, but we are not doing it right now. |
| 13:15.29 | cr2 | htcuniversal_acx removal crashes the machine for a simmilar reason. |
| 13:15.33 | bd2 | psokolovsky, yup. but for driver to call anything asic3_**, it should be bound to device. And devices which depends on asic3, should be passed as platform_data |
| 13:15.41 | cr2 | i guess hx4700_acx will crash too. |
| 13:15.46 | bd2 | s/psokolovsky/cr2/ |
| 13:16.02 | psokolovsky | cr2: so we should (load modules in the right order; note that modprobe doesn't necessarily should be able to do that) |
| 13:16.34 | cr2 | yes, that's the problem. |
| 13:16.48 | psokolovsky | cr2: can you give more details why uni acx crashes? |
| 13:17.02 | cr2 | yes, i'll look at the code now. |
| 13:17.45 | bd2 | because hx4700_acx.c : return platform_device_register( &acx_device ); |
| 13:18.19 | bd2 | thus it's registers device itself. instead it should be passed as child device to asic3 |
| 13:18.50 | cr2 | mem.c:2224 (void) hwdata->start_hw(); |
| 13:18.57 | bd2 | s/hx4700/ba/ |
| 13:19.13 | cr2 | this pointer is dead if foo_acx is removed. |
| 13:19.50 | psokolovsky | bd2, cr2: good catch! let's fix it. |
| 13:22.10 | psokolovsky | bd2: and I don't even want to tell how long I want to cleanup BT crap in hx4700.c/*_core.c ;-) |
| 13:22.30 | cr2 | mem.c:5164 (void) hwdata->stop_hw(); |
| 13:22.37 | cr2 | it's a point of failure too. |
| 13:24.34 | psokolovsky | cr2: ok, so as bd2 points out, there should no be hx4700_acx.c/htcuni_acx.c/etc. in the first place. |
| 13:27.00 | bd2 | psokolovsky, yup. but actually hx4700_acx.c seem to not use asic3. but htcuni/ba do use asic3, thus they must be asic3's child devices |
| 13:27.28 | cr2 | bd2: good idea. |
| 13:28.24 | bd2 | though, hx4700_acx should be placed in hx4700.c too. just for consistency/good sense |
| 13:28.39 | psokolovsky | bd2: yes. but we must adopt consistent pattern here. so, apparently all machines must have device registration as machine def. code - either directly, or via SoC child devices. But never as separate linux modules (but of course, can be in separate file). |
| 13:28.46 | psokolovsky | yep |
| 13:28.54 | cr2 | there should be a more general solution. |
| 13:29.20 | psokolovsky | general solution for what? |
| 13:29.34 | cr2 | psokolovsky: a module is not allowed to oops even if a user removes7inserts it a a random time. |
| 13:30.26 | cr2 | general ? i have a cpld in hermes which replaces asic3. should i do the slave driver model for it too ? |
| 13:30.28 | psokolovsky | cr2: once again: hx4700_acx or htcuniversal_acx or anything like that MAY NOT be module. |
| 13:31.03 | cr2 | ok, that's a point. |
| 13:31.10 | psokolovsky | cr2: yes, you have. Alternatively, we can add to linux generic device dependency model. But *you* will push it to mainline ;-F |
| 13:31.43 | cr2 | :) it's not going to be accepted. |
| 13:31.49 | bd2 | :-D |
| 13:32.48 | cr2 | because it's a business of a module A to know if it needs module B, and that can be solved in userspace. |
| 13:33.04 | psokolovsky | yep ;-) |
| 13:36.10 | cr2 | mem.c (acx) obviously needs foo_acx, so it's not allowed to be a module by your logic :) |
| 13:37.12 | cr2 | another question. can you confirm that hx4700 enables CDEX_SDBUS clock ? |
| 13:37.52 | cr2 | because i don't see it on other 4 asic3-SD using devices |
| 13:38.10 | psokolovsky | cr2: not by my logic, but by mem.c's logic ;-) I just accept that logic ;-) |
| 13:38.16 | cr2 | btw, what about h4000 ? |
| 13:38.20 | cr2 | LOL. |
| 13:38.24 | bd2 | mem.c is *driver*, thus it can be a module |
| 13:39.09 | psokolovsky | cr2: h4000 & hx4700 work pretty well with asic3_mmc as is. I dunno much about internal settings of asic3-sd ;-) |
| 13:39.36 | psokolovsky | cr2: see asic3_base.c:901: |
| 13:39.39 | cr2 | bd2: ok, but it blows on null pointers. and it's an obvious bug. |
| 13:39.40 | psokolovsky | /* XXX: should get correct SD clock values from pdata struct */ |
| 13:40.10 | cr2 | psokolovsky: run "dump asic3" if you have wince installed :) |
| 13:40.16 | cr2 | that's all what i need. |
| 13:40.27 | bd2 | cr2, sure. the question is why pointers becoming null |
| 13:40.34 | psokolovsky | cr2: it just doesn't expect a device to be gone behind its back |
| 13:40.50 | psokolovsky | cr2: will do a bit later ;-) |
| 13:40.57 | cr2 | yeah, but goxboxlive did it all the time . |
| 13:41.18 | bd2 | psokolovsky, indeed. but not device it's bound to. but device/driver which used by foo_acx |
| 13:41.35 | cr2 | and that should be ok. |
| 13:42.19 | psokolovsky | bd2: yep, so foo_acx is unloaded, pointer starts to point to "somewhere", then in crashes/hangs |
| 13:43.38 | bd2 | cr2, psokolovsky, the only thing mem.c is wrong, is thet is calls stop_hw (not start) at module_exit function, instead of acxmem_e_remove (i.e. in driver's remove() handler) |
| 13:44.02 | psokolovsky | bd2: hm, ok, maybe that too ;-) |
| 13:44.53 | bd2 | yeah, but commed at acxmem_e_cleanup_module() (which calls stop_hw) tells about that it's not all that easy |
| 13:45.16 | psokolovsky | ;-) |
| 13:45.29 | psokolovsky | let's just disallow foo_acx to be module ;-) |
| 13:46.18 | psokolovsky | note btw that it can't done right away either, due to botched Kconfig ;-D |
| 13:46.22 | bd2 | sure. but mem.c needs some fixing anyway (not a priority at all, because whole acx stuff is not in mainline yet) |
| 13:46.32 | psokolovsky | so, instead indeed should be moved towards machine.c ;-) |
| 13:47.05 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
| 13:48.58 | bd2 | psokolovsky, ready to proceed x/50 patchset? ;-) |
| 13:49.32 | psokolovsky | bd2: what?! 8-o |
| 13:50.13 | bd2 | hehe.. yeah, that's diff to my battery tree ;-) |
| 13:50.35 | psokolovsky | oh my. but you sent yesterday */10 one? |
| 13:50.57 | bd2 | yea, that's another patchset ;-) |
| 13:52.11 | bd2 | too bad we aren't on git rails yet.. merging would be much easier ;-) I'll send it soon, please test it in spare time, and feel free to commit whole stuff as "syncing with battery work tree" |
| 13:52.11 | psokolovsky | oh my ;-) |
| 13:52.46 | psokolovsky | bd2: well ok, if iit's really required. or we can do it a bit later (committing I mean) ;-) |
| 13:53.34 | bd2 | psokolovsky, sure. but then it will grow to x/100 patchset ;-)) |
| 13:53.46 | psokolovsky | ok ;-) |
| 13:56.11 | psokolovsky | bd2: I hope you also monitor commits/diff with CVS, because there were few commits to battery-related files |
| 13:57.16 | bd2 | psokolovsky, yup. I've done rebaseing few minutes ago |
| 13:57.22 | psokolovsky | ok |
| 14:15.22 | bd2 | ok, sent. I hope I've not exceed size limit |
| 14:21.32 | bd2 | crap, it seems I've really exceed. base64 encoding increased mail size |
| 14:25.06 | psokolovsky | bd2: nothing arrived so far at least |
| 14:26.33 | psokolovsky | cr2: so, I'm goint to kill ipaq-asic-led.h, and move stuff directly to ipaq-asic2.h/ipaq-asic3.h. sounds good? |
| 14:52.50 | cr2 | yes. |
| 14:52.59 | psokolovsky | bd2: ok, both versions arrived ;-) |
| 14:53.05 | psokolovsky | cr2: ok, committed |
| 14:53.10 | cr2 | great. |
| 14:55.04 | bd2 | psokolovsky, %-O |
| 14:55.43 | bd2 | probably some mailing list admin got notified about size exceed email |
| 14:56.50 | psokolovsky | yep ;-) |
| 15:02.29 | bd2 | kick me now if I promised something to do soon, and forgot. going to dig in adc (ssp) stuff, thus might become unresponsive for some time :-)) |
| 15:10.10 | psokolovsky | bd2: nope, all's ok ;-) |
| 15:11.14 | psokolovsky | and time to make new tag... |
| 15:16.24 | Kevin2 | Hi |
| 15:29.47 | cr2 | Kevin2: hi. i have a patch and a question for you. |
| 15:32.01 | Kevin2 | cr2: Ok. |
| 15:33.20 | cr2 | why do you use 16MHz clock on hermes ? |
| 15:34.10 | cr2 | the haret patch implements 'playsound N' function. |
| 15:34.18 | Kevin2 | cr2: Not sure what you mean - where do I do that? |
| 15:35.05 | Kevin2 | cr2: Send the patch to haret@handhelds.org -- sounds like a useful feature (no pun intended). |
| 15:35.52 | cr2 | ok :) |
| 15:36.43 | cr2 | s3c24xx_init_clocks(12000000); |
| 15:36.46 | goxboxlive | Kevin2: I have made a linboot.exe but it isent not executeable when i try to launch it in wince? Could this be that i skipped the ramdisk in it? |
| 15:37.03 | cr2 | what value do you have here ? |
| 15:38.01 | Kevin2 | cr2: I use s3c24xx_init_clocks(16934400); |
| 15:38.21 | Kevin2 | because that was your recommendation - I haven't really looked at the clocks yet. |
| 15:39.10 | Kevin2 | goxboxlive: No, you don't need an initrd. Can you create the file earlyharetlog.txt and see what it says? |
| 15:40.25 | goxboxlive | Kevin2: Ok, yes i can try that. |
| 15:40.26 | Kevin2 | goxboxlive: Can you do "cvs up" also. I just added a patch which will cause commands being executed to get logged. |
| 15:41.16 | Kevin2 | What was your make line? |
| 15:43.19 | cr2 | Kevin2: sent. |
| 15:44.51 | cr2 | Kevin2: i know a wince function in a dll, how can i dump r0,r1,r2 when it is called and catch it's return value ? |
| 15:45.43 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
| 15:45.52 | cr2 | imho you should try to set the clock to 12MHz -> s3c24xx_init_clocks(12000000); |
| 15:46.17 | psokolovsky | cr2: now point me to binary which does "dump asic3" ;-) |
| 15:46.19 | Kevin2 | cr2: On Hermes or PXA? |
| 15:46.22 | psokolovsky | Kevin2: Hi! |
| 15:46.32 | cr2 | and if it will be wrong, we need to fix the CVS. |
| 15:46.34 | Kevin2 | Hi psokolovsky |
| 15:46.54 | cr2 | psokolovsky: http://jornada820.sf.net/files/haret |
| 15:47.05 | psokolovsky | ok, thanks |
| 15:47.18 | cr2 | the asic3 code is so terrible that i don't want to port it to the mainline. |
| 15:47.42 | goxboxlive | Kevin2: after make (to build haret.exe) i ran: make linload KERNEL=zImage SCRIPT=startup.txt with mt zImage and startup.txt in the root of haret |
| 15:47.46 | cr2 | Kevin2: on hermes. |
| 15:49.04 | Kevin2 | goxboxlive: What is in your startup.txt? Is the last line "ramboot"? |
| 15:49.17 | goxboxlive | Kevin2: No, it's like a regular one with boot2 |
| 15:49.46 | cr2 | Kevin2: sorry. 12MHz on hermes, and to trace a function on PXA. |
| 15:49.53 | Kevin2 | cr2: I haven't implemented arch generic break points. Only pxa can do it. |
| 15:50.34 | psokolovsky | cr2: doesn't run. again hardcoded phone dependencies? |
| 15:50.57 | cr2 | psokolovsky: hm. yes. i'll remove it. |
| 15:51.02 | Kevin2 | cr2: set insn 0xaddr ; wi 99 |
| 15:52.12 | Kevin2 | you can change insnreg1 and insnreg2 to display any two registers. It doesn't currently support showing a third. You could modify the code to do so, or you could set a second breakpoint using insn2, insn2reg1, insn2reg2. |
| 15:53.07 | cr2 | Kevin2: thanks. |
| 15:53.25 | cr2 | psokolovsky: removed the RIL library now. |
| 15:54.33 | psokolovsky | cr2: ok, this starts |
| 15:55.10 | Kevin2 | goxboxlive: Haret is lacking good documentation. You need to use "ramboot" to boot the kernel stored in the exe. |
| 15:55.24 | goxboxlive | Kevin2: ok thx, i will try that |
| 15:55.41 | Kevin2 | goxboxlive: see docs/linboot.txt |
| 15:57.22 | goxboxlive | Kevin2: I thought that was just needed when using a initrd. |
| 15:58.17 | Kevin2 | goxboxlive: You should really only need to run: make linload KERNEL=my-zimage |
| 15:58.30 | *** join/#htc-linux eremias (n=fdsfs@ACB0ADA8.ipt.aol.com) |
| 15:58.31 | psokolovsky | cr2: http://pastebin.ca/451489 - h4000 |
| 15:58.32 | Kevin2 | The initrd defaults to empty and the script defaults to txt/linload.txt |
| 15:58.54 | psokolovsky | cr2: on hx4700, gnuharet segfaults on dump asic3 %) |
| 15:58.54 | goxboxlive | o |
| 15:58.56 | goxboxlive | k |
| 15:59.58 | cr2 | psokolovsky: hm. really. SDBUS is on. i guess it's the same on hx4700. |
| 16:00.19 | psokolovsky | cr2: tried again, segfaults. |
| 16:01.06 | cr2 | check the latest version, i've made an update. |
| 16:03.15 | cr2 | so we have 3 combinations in wince. |
| 16:05.23 | cr2 | btw, your wince h4000 setup also does not match the current asic3_mmc setup. |
| 16:09.00 | goxboxlive | I have asked france at hh.org if we could host files/images at hh.org and he said it is ok. I was thinking that we could have a common one for the htc-pxa devices. |
| 16:09.48 | cr2 | goxboxlive: i think we are close. |
| 16:09.49 | goxboxlive | But to make it i have ti send france my public ssh key. |
| 16:10.03 | goxboxlive | How do i make such a key? |
| 16:10.21 | cr2 | generate one. |
| 16:10.46 | cr2 | man ssh-keygen |
| 16:11.18 | goxboxlive | cr2: I didnt think about a common opie-image and s, i was thinking about a common place to have all our files. I think it is bether to have it at hh.org than at gnulinux.biz since the *biz ftp server is private |
| 16:11.23 | goxboxlive | ok thx |
| 16:11.36 | cr2 | i think my key is 8192 bits. just for fun ;) |
| 16:12.04 | cr2 | 16384 will probably work too. |
| 16:12.52 | cr2 | ok. |
| 16:14.58 | cr2 | psokolovsky: i'm inclined to create a ${MACHINE}_asic3.c and do all the asic3-related stuff there. |
| 16:15.10 | goxboxlive | cr2: so i should send france this key? id_dsa.pub |
| 16:15.19 | cr2 | yes. |
| 16:15.29 | psokolovsky | cr2: create for what? |
| 16:15.55 | cr2 | to separated the machine init from asic3 init. |
| 16:16.23 | goxboxlive | cr2: But what if i have two machines i would like to connect to hh.org? SHould i make two then? |
| 16:16.25 | cr2 | i'm already removing all device driver init to the appropriately named files. |
| 16:16.33 | psokolovsky | cr2: I don't understand. it is already separated. |
| 16:16.49 | cr2 | goxboxlive: you can connect from any machine that has your private key. |
| 16:17.14 | psokolovsky | cr2: that's strange, because we just recently spoke with bd2 that it's time to merge all stuff into single machine file ;-) |
| 16:17.42 | goxboxlive | ok, so if i just copy my id_dsa.pub over to my other machine then that will be enough. good |
| 16:18.38 | cr2 | psokolovsky: it's for testing purposes. not each device is at the same stage as hx4xxx |
| 16:19.36 | cr2 | eventually we may put all htc devices in a htc subdir, like the 'eseries'. |
| 16:19.48 | psokolovsky | cr2: I think it's superfluous. I alreday suggested extending platform_data structures, if there're additional info to be passed. |
| 16:20.41 | cr2 | it is not linked to each other. |
| 16:21.58 | *** join/#htc-linux eremias (n=fdsfs@ACB0ADA8.ipt.aol.com) |
| 16:22.34 | bd2 | merging stuff is exactly matter of stable enough machines. |
| 16:25.06 | psokolovsky | yep, I just try to hope all active ports in cvs are on ~same stage, and we can move in one front, and elaborate how we do stuff, not fall back to older state ;-) |
| 16:25.52 | psokolovsky | like goxboxlive said, otherwise it feels like one step ahead, 2 steps back. |
| 16:25.56 | psokolovsky | but nevermind |
| 16:26.53 | bd2 | yup. And actually, sparating *devices* to modules makes a sense only if platform_data stuff is about to debug. because usually driver is being debugged, not device (which is usually just two structs: device + its platform data) |
| 16:27.57 | psokolovsky | cr2: http://jornada820.sourceforge.net/files/haret/gnuharet-200704211800.exe still segfaults |
| 16:29.17 | cr2 | bd2 psokolovsky: i can't boot on himalaya from the SD. it's some timing issue, so i'm trying to debug asic3_mmc which works perfectly on many other machines. and there are some differences. or course all the research will come back to the common driver. |
| 16:29.38 | cr2 | psokolovsky: ok, then it's some different problem ;) |
| 16:29.54 | psokolovsky | cr2: so it's not like different asic3 reg values after all? |
| 16:29.55 | cr2 | psokolovsky: can the dump the CDEX register with 'pd' ? |
| 16:30.11 | psokolovsky | cr2: especially if you tell me address ;-) |
| 16:30.28 | cr2 | hmm. bus_shift is 1... |
| 16:30.44 | psokolovsky | cr2: ok, will look up myself ;-) |
| 16:30.57 | cr2 | http://wiki.xda-developers.com/index.php?pagename=ASIC3 |
| 16:31.09 | cr2 | i need to move this page to hh.org |
| 16:31.59 | psokolovsky | cr2: |
| 16:32.00 | psokolovsky | HaRET(1)# pd 0x0c000500 0x20 |
| 16:32.00 | psokolovsky | 0c000500 | 000563cb 00000000 00000000 00000000 | .c.............. |
| 16:32.00 | psokolovsky | 0c000510 | 00000000 00000000 00000000 00000000 | ................ |
| 16:32.07 | cr2 | 0x0c000000 + 0x500 |
| 16:32.09 | cr2 | ok. |
| 16:32.26 | cr2 | cdex is 63cb, sel is 0005 |
| 16:33.01 | cr2 | 0110.0011.1100.1011 |
| 16:37.20 | cr2 | spi off, pwm off, leds on, SDhost on, sdbus off. |
| 16:37.24 | cr2 | hmm. |
| 16:37.40 | psokolovsky | cr2: and sd clock different? |
| 16:38.27 | cr2 | so h4000 and rx3000 have sdbus on, everything else sdbus off. |
| 16:39.07 | cr2 | and sel can differ. |
| 16:39.41 | cr2 | 0x5 vs 0x7 if sdbus is on. |
| 16:40.25 | cr2 | so if sdbus is set on by wince, the speed is set to 24MHz. |
| 16:40.27 | psokolovsky | cr2: ah, ipaq-asic3 says "sel" for sdio, so they kinda match... |
| 16:41.16 | cr2 | the asic3_mmc has sdbus on, 24MHz off (=12MHz). |
| 16:41.42 | cr2 | that's the only visible difference. |
| 16:42.08 | cr2 | well, and this code too. |
| 16:42.11 | cr2 | <PROTECTED> |
| 16:42.11 | cr2 | <PROTECTED> |
| 16:44.09 | cr2 | is asic3 capable only of 3.3v ? |
| 16:44.25 | cr2 | ASIC3_MMC_REG(host, SD_CONFIG, SDHC_Power1) = SD_CONFIG_POWER1_PC_33V; |
| 16:45.09 | cr2 | probably i need to check it for different cards. also for sdio. |
| 16:48.13 | Kevin2 | cr2: Would it be possible to add the asic3 register definitions to haretconsole (eg, in memalias.py)? This would make it simple to monitor those things under mainline haret. |
| 16:52.42 | cr2 | Kevin2: good idea. is it possible to name the bits ? |
| 16:52.52 | cr2 | i've zero python experience. |
| 16:53.09 | cr2 | all names are in include/asm-arm/hardware/ipaq-asic3.h |
| 16:53.10 | Kevin2 | cr2: Yes. See Regs_s3c2442. |
| 16:53.16 | cr2 | ok. |
| 16:53.46 | Kevin2 | You shouldn't need any python experience - you just need to populate the variable with register addresses and bit names. |
| 16:54.07 | cr2 | yes, i've got it. |
| 16:55.01 | cr2 | Kevin2: there are some power-management related structures and defines missing from cegcc. where should i put them ? |
| 16:57.08 | Kevin2 | We need to push them to the cegcc folks. But in the meantime, you can put them into a header in include. Use the same header the MSN docs say they should be in, and put #include_next <header> at the top of the file. |
| 16:57.24 | cr2 | ok. |
| 16:57.55 | cr2 | Kevin2: memalias.py obviously needs to be per-machine. |
| 16:58.23 | cr2 | because the asic3 register locations depend wildly on a machine. |
| 16:59.00 | Kevin2 | It is - RegsList['Hermes'] = Regs_Hermes - will add the registers defined in Regs_Hermes for the machine that identifies itself as "Hermes". |
| 16:59.40 | cr2 | ok. |
| 17:21.22 | *** join/#htc-linux twostupiddogs (n=towstupi@62.123.180.130) |
| 17:23.59 | Kevin2 | cr2 / psokolovsky : See http://www.handhelds.org/hypermail/haret/current/0861.html |
| 17:35.31 | Kevin2 | I'll be back later. |
| 18:13.27 | *** join/#htc-linux MrVanx (n=chatzill@82-69-60-250.dsl.in-addr.zen.co.uk) |
| 18:41.20 | *** join/#htc-linux tudenbart (n=willi@xdsl-213-196-240-174.netcologne.de) |
| 19:11.34 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
| 20:31.55 | *** join/#htc-linux Ranpha1 (n=Ranpha@hlm-n-73ed.adsl.wanadoo.nl) |
| 20:58.54 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
| 21:05.01 | cr2 | goxboxlive: does your new haret boot ? |
| 21:06.33 | goxboxlive | cr2: i havent testet it yet. Since i setup my laptop to day i havent had time. |
| 21:07.52 | cr2 | ok. |
| 21:08.15 | goxboxlive | and also cdgcc is failing when i try to compile it on my laptop |
| 21:08.52 | goxboxlive | it compiled with debian, but with kubuntu it failes. Probably missing some libs |
| 21:23.39 | goxboxlive | cr2: tested linload now, and it works |
| 21:24.17 | goxboxlive | Kevin2: Yes with ramboot in startup.txt linload boots. Great, thx. |
| 21:26.05 | goxboxlive | cr2: Seems like it can be a problem to host files at hh.org. Rene says that we also should host the source to our images because of GPL. |
| 21:30.18 | cr2 | sounds a bit overzealous, but ok. |
| 21:30.30 | cr2 | post them at sf.net ? |
| 21:30.46 | cr2 | i don't know hot to do it, btw. |
| 21:34.23 | goxboxlive | i guess we can host files at linuxtogo, but since our kernel source is at hh.org i thought it was great to have it at same place. |
| 21:34.43 | *** join/#htc-linux alsuren (n=alsuren@host86-148-18-74.range86-148.btcentralplus.com) |
| 21:34.43 | *** join/#htc-linux Whacko (n=Whacko@s5593e8a2.adsl.wanadoo.nl) [NETSPLIT VICTIM] |
| 21:34.44 | *** join/#htc-linux Kevin2 (n=Kevin@207-237-65-88.c3-0.avec-ubr12.nyr-avec.ny.cable.rcn.com) [NETSPLIT VICTIM] |
| 21:34.44 | *** join/#htc-linux ljp (n=lpotter@203.94.178.46) |
| 21:34.58 | cr2 | btw, i have compiled opie today. |
| 21:34.59 | goxboxlive | good |
| 21:35.01 | cr2 | in includes tons of completely irrelevant stuff |
| 21:35.03 | goxboxlive | without any hacks? |
| 21:35.09 | goxboxlive | does it, i dint know |
| 21:35.16 | cr2 | beginning with gtk/atk/pango etc, and even such things like gnome-whatever ;) |
| 21:35.17 | cr2 | without any hacks. |
| 21:35.18 | goxboxlive | Yes i have seen it, i think it is because of dbus |
| 21:35.20 | cr2 | imho the bluez-utils should be removed from the distro |
| 21:35.20 | goxboxlive | why? |
| 21:35.20 | cr2 | unless somebody fixes the source. |
| 21:35.20 | goxboxlive | place a bug at OE |
| 21:35.20 | cr2 | because bluez-utils have nothing to do with gtk. |
| 21:35.24 | cr2 | if comebody can't type 'int' instead of 'gint' that does not mean we should compile a lot of irrelevant software. |
| 21:35.26 | goxboxlive | ok |
| 21:35.39 | cr2 | i don't want to engage in opie oe bugfixing. |
| 21:35.58 | cr2 | i prefer to fix opie2 instead and write the phone gui for it. |
| 21:36.27 | cr2 | i'm also working on the qt4 compatibility layer for roadmap. |
| 21:36.36 | cr2 | so it can be included in opie2. |
| 21:36.51 | goxboxlive | ok |
| 21:37.24 | cr2 | we talked with BabelO today, and will probably use gsmd for opie2. |
| 21:38.45 | goxboxlive | ok, that sounds good. Has he worked on Opie2 lately? |
| 21:39.01 | goxboxlive | has he done any progress? |
| 21:39.11 | cr2 | don't know but he plans to work. |
| 21:39.39 | goxboxlive | Seems like no one is intrested in Opie2 |
| 21:39.53 | goxboxlive | It isent in OE yet, and nobody is talking about it |
| 21:40.37 | cr2 | i need to copy some stuff from the virtual machine to have a normal setup. |
| 21:40.54 | cr2 | openmoko does not compile from oe too. |
| 21:41.15 | cr2 | git pull fails for me since 1 month and nobody cares. |
| 21:43.18 | goxboxlive | well, it might be that no one is aware of it. |
| 21:43.35 | goxboxlive | You have to place a bug so they are informed about it. |
| 21:43.44 | goxboxlive | Have you tried this? |
| 21:44.13 | goxboxlive | http://wiki.openmoko.org/wiki/MokoMakefile |
| 21:47.00 | cr2 | hm. it will setup its own tree ?? |
| 21:47.33 | goxboxlive | Yes, fully automatic, but you have to abort change machine in local.conf before it start to build the image. |
| 21:48.06 | goxboxlive | And it also creates openmoko distro insted of angstrom (eabi vs non-ebai ) |
| 21:48.53 | cr2 | ok, probably it's too much for me :) |
| 21:50.11 | goxboxlive | :-) |
| 21:50.23 | cr2 | goxboxlive: is it possible to read the "other" rtc ? |
| 21:50.39 | goxboxlive | what do you mean? |
| 21:50.44 | cr2 | do we have this chip on the motherboard at all ? |
| 21:50.59 | cr2 | sable aka paq69xx has it. |
| 21:51.21 | cr2 | the wince kernel on universal has software support for it. |
| 21:51.35 | goxboxlive | i dont know, we are using two versions as far as i know |
| 21:51.47 | cr2 | sa1100-rtc |
| 21:51.54 | goxboxlive | but i dont know wich one is the right to use |
| 21:51.58 | cr2 | i'm talking about the i2c chip. |
| 21:52.45 | cr2 | you can just load the driver and check. |
| 21:52.57 | cr2 | or even use i2cdump. |
| 21:53.16 | cr2 | ok. maybe i'll try it myself eventually. |
| 21:54.10 | goxboxlive | yes i think that is better, i don't know what to look for anyway |
| 21:54.50 | cr2 | :) |
| 21:55.10 | cr2 | what do you think about creating a new eabi ramdisk ? |
| 21:55.33 | *** join/#htc-linux MrVanx (n=chatzill@82-69-60-250.dsl.in-addr.zen.co.uk) |
| 21:56.01 | cr2 | we have an old .12 ramdisk for blueangel and the apache ramdisk by Kevin2. |
| 21:57.07 | goxboxlive | well, do we rely need one? |
| 21:57.49 | cr2 | for the installation, and for the modules. |
| 21:58.48 | cr2 | goxboxlive: does openmoko-dialer support PIN ? |
| 21:58.56 | goxboxlive | But how will that help us for the installation. A user will still be neded to create a linux partition on his SD card |
| 21:59.32 | cr2 | you can do it from the ramdisk ? |
| 21:59.45 | goxboxlive | cr2: I have never tried the pin stuff. I am not using it wheni have tried it out. I can try now |
| 22:00.02 | cr2 | the blueangel ramdisk even has an Xserver :) |
| 22:00.42 | cr2 | somebody at xda-dev asks where to set a pin and enter the provider. |
| 22:01.26 | goxboxlive | ok, i am trying it now |
| 22:04.54 | goxboxlive | No does not seems like pin os working in opie/openmoko-dialer |
| 22:05.30 | cr2 | ok, does it work on neo1973 ? |
| 22:05.38 | goxboxlive | no idea |
| 22:05.59 | *** join/#htc-linux RoEn_PC (n=roen@p54A6728E.dip.t-dialin.net) |
| 22:06.35 | goxboxlive | if it did it would have been working on universal too. It might be that it works with openmoko but i haven't tried it |
| 22:07.26 | cr2 | it's a pure AT gsm issue, so i don't see a reason why it fails. |
| 22:07.44 | goxboxlive | If we make a ramdisk, and if there are some updates how do we update the ramdisk, do we have to make a new one? in example bluez |
| 22:08.10 | goxboxlive | Are we able to build ramdisk in OE? |
| 22:14.48 | cr2 | don't know, but it's probably not an issue. |
| 22:15.26 | cr2 | i'll try to unpack the busybox.ipk to the ramdisk and write some minimal scripts. |
| 22:16.02 | cr2 | i'm curious if we can do a 16MB ramdisk, and what are the actual limits. |
| 22:16.12 | goxboxlive | ok |
| 22:16.33 | cr2 | there was a 4M limit on the compressed initrd, but i think Kevin has solved the problem. |
| 22:16.46 | goxboxlive | ok |
| 22:17.03 | cr2 | going to bed now. good night. |
| 22:17.08 | goxboxlive | good noght |