**** BEGIN LOGGING AT Wed Dec 27 03:00:00 2006 Dec 27 07:38:53 03arjan 07org.oe.dev * r189d5833... 10/ (1 conf/machine/h1940.conf): h1940: clean up machine config Dec 27 07:55:46 koen: ping Dec 27 08:17:55 mmp: pong Dec 27 08:18:57 koen: you botmail'ed me about USB support in h5xxx; did you mean UDC, or USB host? Dec 27 08:19:34 UDC works with 2.6.16 for me (should also with HEAD) Dec 27 08:20:27 I got the Xfbdev also working (and not crashing) few days ago, but this seems to be more complicated issue. Dec 27 08:20:33 it's just ugly hack Dec 27 08:21:06 udc Dec 27 08:21:17 I wanted to install angstrom-bootstrap-image + distcc Dec 27 08:21:26 so I only need udc + serial Dec 27 08:22:31 udc should work with lastest 2.6.16 and lastest HEAD Dec 27 08:22:46 Erik Hovland commited that patch to both Dec 27 08:25:06 * mmp browses through CVS Dec 27 08:25:12 'latest 2.6.16' is in OE? Dec 27 08:26:38 koen: I'm just looking if 2.6.16 has the patch commited... Dec 27 08:27:21 hmm, so just HEAD it seems contains it. Dec 27 08:27:46 but the patch is rather simple, if you wait few minutes, I'll put it somewhere... Dec 27 08:28:33 or, even better, export Erik's :) Dec 27 08:31:17 http://www.handhelds.org/hypermail/kernel-discuss/22/2217.html Dec 27 08:31:20 this should be it Dec 27 08:43:10 mmp: should h5000 in OE use HEAD or 2.6.16+patch? Dec 27 08:44:09 koen: Well, I don't know, it seems that Erik did not commit the patch to 2.6.16 (or I forgot to send him mail:( ). I'm doing my development on 2.6.16. actually. Don't know what state is "there upwards" :) (in newer versions) Dec 27 08:45:08 NOTE: Applying patch 'h5400-udc-mod-from-milan.patch' Dec 27 08:45:08 NOTE: package linux-handhelds-2.6-2.6.16-hh8-r1: task do_patch: completed Dec 27 08:45:17 :) Dec 27 08:46:11 do you have any more patches for that? Dec 27 08:46:19 s/for that/against 2.6.16/ ? Dec 27 08:53:17 hey zecke|23c3 Dec 27 08:53:33 ho Dec 27 08:55:30 03koen 07org.oe.dev * r5d3b23f0... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.16-hh8.bb): linux-hh 2.6.16-hh8: add patch from Milan Plzik to fix UDC on h5000 Dec 27 08:55:38 mmp: there you go Dec 27 08:55:48 :-) Dec 27 08:55:54 thanks:) Dec 27 09:00:29 hi someone here Dec 27 09:01:22 http://pastebin.com/844679 Dec 27 09:03:25 ariari_frenchy1: hi, where is the question? Dec 27 09:03:36 2.4.20_mvlcee31-omap16xx_gsm_gprs Dec 27 09:03:44 I don't think we support MontaVista here at all Dec 27 09:03:54 i have pb with driver Dec 27 09:04:04 not in that pastebin another Dec 27 09:04:24 I think you are in the wrong channel Dec 27 09:04:35 send your requests to your MVista contact person Dec 27 09:07:46 http://pastebin.com/845637 Dec 27 09:09:30 funny guy Dec 27 09:10:38 u see something wrong ? Dec 27 09:11:05 yes you use MontaVista stuff, proprietary NDIS and GDI stuff. You are wrong in this channel, please leave Dec 27 09:11:38 wut is funny i just kill my skypephone netgear one and trying to revive it that all Dec 27 09:12:45 sure, but leave now Dec 27 09:13:13 morning all Dec 27 09:13:23 ariari_frenchy1: e.g. get your manual. check for contact page, call the call-center Dec 27 10:05:35 hi zecke Dec 27 10:05:55 hi niv_one_three Dec 27 10:06:12 happy holidas . any news about the mq200 (simpad screen drv) Dec 27 10:06:15 ? Dec 27 10:08:07 dunno . Dec 27 10:20:00 koen: hi Dec 27 10:40:18 koen: (ad h5000) if you will see any weird behavior of that kernel, pls. botmail me or send me an e-mail (milan point plzik /at/ gmail) - it might be possible that there is some caching weirdness... . Dec 27 10:52:13 anyone @ 23c3? Dec 27 10:55:16 harrydg: <11:04> * zecke|23c3 has quit ("leaving") Dec 27 10:55:20 no one me too waiting help Dec 27 10:56:31 morning guys Dec 27 10:58:02 g morning Dec 27 10:58:29 hey Marcin Dec 27 11:00:23 hi Dirk Dec 27 11:01:09 RP: MACHINE=mtx-2 DISTRO=angstrom-2007.1 - unpack of util-linux will fail because bitbake DOES NOT check in proper paces for files Dec 27 11:10:01 my angstrom build for all machines fails in so many places.. Dec 27 11:10:55 koen: ping? Dec 27 11:14:24 koen: http://www.openembedded.org/filebrowser/org.openembedded.dev/packages/gtk%2B Dec 27 11:14:26 ji Holger Dec 27 11:14:28 koen: filebrowser fails on gtk+ Dec 27 11:15:07 php escaping bug? Dec 27 11:15:16 zecke|23c3: probably Dec 27 11:15:32 :) Dec 27 11:15:52 koen: do we build gtk+ with atk support? I currently try to hookup mickeyl with dogtail Dec 27 11:16:56 zecke|23c3: http://www.openembedded.org/repo/org.openembedded.dev/packages/gtk+/ Dec 27 11:17:34 hi koen ; Dec 27 11:17:47 zecke|23c3: yes, we build with atk support Dec 27 11:17:51 hrw: pong Dec 27 11:18:50 koen can i instal that distrib on that phone ,netgear ?sph101 Dec 27 11:19:21 koen: angstrom build for all machines progressing quite nice but also show new bugs in mipsel support and in bitbake trunk Dec 27 11:22:58 ariari_frenchy1: if your kernel has eabi support, yes Dec 27 11:23:38 eabi ? Dec 27 11:23:44 wut is that ? Dec 27 11:24:57 you know how to instal koen Dec 27 11:25:04 koen or cohen ? Dec 27 11:28:34 ariari_frenchy1: do us a favor, leave Dec 27 11:29:57 i thoungt you can port openembedded on my device Dec 27 11:30:21 ariari_frenchy1: without device? Dec 27 11:30:39 no i can give access to the device if you want Dec 27 11:31:19 no pb ssh is open on that machine and the device plugged in serial usb Dec 27 11:31:40 hrm if you are interessted . Dec 27 11:31:50 i ll give you access Dec 27 11:32:00 ariari_frenchy1: 1. does this device has working 2.6 kernel with sources? Dec 27 11:32:23 ariari_frenchy1: 2. does this device original rootfs is available without NDA like papers? Dec 27 11:32:47 no i only see 2.4 on it Dec 27 11:32:56 ariari_frenchy1: 3. are you brave enough to use untested rootfs which can brick device? Dec 27 11:33:02 2.4??? argh.. Dec 27 11:33:16 have a look Dec 27 11:33:39 Linux version 2.4.20_mvlcee31-omap16xx_gsm_gprs Dec 27 11:34:00 http://pastebin.com/845637 Dec 27 11:34:45 ariari_frenchy1: no one here is crazy enough to work on 2.4 vendor kernels Dec 27 11:35:21 is it possible to install 2.6 kernel on it ? Dec 27 11:35:22 03hrw 07org.oe.oz354x * rfd16a9e2... 10/ (3 files in 3 dirs): prismstumbler: fix CROSS COMPILE BADNESS - close #1679 Dec 27 11:35:54 or other os on it ? Dec 27 11:36:30 if some one ported 2.6 for it... Dec 27 11:36:49 where can i check ? Dec 27 11:37:11 no idea from me Dec 27 11:37:27 u can port it to 2.6 ? Dec 27 11:37:50 no Dec 27 11:38:28 koen|awa1: can you add list of supported archs into angstrom.conf? so it will be skipped for arm4 etc Dec 27 11:39:41 debian has it i think Dec 27 11:43:13 qemu-native-0.8.2+cvs20060723-r1 breaks with /usr/include/linux/usbdevice_fs.h:49: error: variable or field `__user' declared void on my new glibc-2.5 system .. any hints ? Dec 27 11:43:56 update SRCDATE to something newer? Dec 27 11:44:24 oh Dec 27 11:44:48 03dirk 07org.oe.dev * r05d1f975... 10/ (1 packages/openvpn/openvpn_2.0.9.bb): add 2.0.9 version Dec 27 11:44:56 03dirk 07org.oe.dev * re15e7138... 10/ (6 files in 3 dirs): add 2.4.7 Dec 27 11:44:56 i didnt watch this date .. Dec 27 11:45:01 03dirk 07org.oe.dev * r2040fbc4... 10/ (4 files in 3 dirs): wlan-ng-modules: new snapshot 20061109 Dec 27 11:46:26 hmm i cant find where this date is introduced ,, Dec 27 11:47:00 do13: 'add 2.4.7'? Dec 27 11:50:13 the srcdate was introduced by an commented SRCDATE .. how can this be !! Dec 27 11:51:09 in 2007 comments still will start with Dec 27 11:51:12 # Dec 27 11:53:03 will Dec 27 11:53:28 hrw: sorry forgot openswan :) Dec 27 11:53:54 do13: check bugtracker then - we have open bug for openswan Dec 27 11:54:19 hrw: ok. I overlooked this Dec 27 11:55:19 * rob_w thought his perception of the known world will breake down Dec 27 11:55:24 hrw: Hmm need to resubscribe to the bugtracker ml Dec 27 11:57:38 hrw: I'd like to avoid setting COMPATIBLE_ARCH in angstrom, but an INCOMPATIBLE_ARCH would be acceptable Dec 27 11:57:50 * koen heads downtown Dec 27 12:01:21 * niv thinks about his simpad booting anstrong * Dec 27 12:04:08 hail zecke|23c3 Dec 27 12:04:27 master! Dec 27 12:05:50 * pb_ back from holiday Dec 27 12:07:21 zecke|23c3: do I take it from your name that you are at the eleet ccc event? Dec 27 12:10:54 cu Dec 27 12:24:19 morning Dec 27 12:55:26 morning XorA|gone Dec 27 13:03:50 re Dec 27 14:19:20 pb_: correct Dec 27 14:22:19 ich: very good Dec 27 14:23:58 in my eleetnes I'm searching a power plug Dec 27 14:33:08 someone here will go to FOSDEM 2007 ? Dec 27 14:34:07 hehe, ALL will be there Dec 27 14:38:13 hi zecke|23c3 i hee u are there surelly :) ... Dec 27 14:38:28 i'm thinking about to come :) Dec 27 14:39:47 gremlin[it]: without zecke there can be no fosdem Dec 27 14:40:03 uhh ?! Dec 27 14:40:24 hehe :) Dec 27 14:40:37 pb_: thanks master Dec 27 14:47:43 zecke|23c3 mainly i would know about u all :) Dec 27 14:48:02 i am working on a recipe for swsqueeze, a cli audio app. it depends on libmad and flac. both of those build fine and i see their headers setup in the staging dir. but swsqueeze fails to build because it can't find those headers like . does the staging dir like tmp/staging/i486-linux/include automatically get pulled by ccache i486-linux-gcc or should i do something specific to fix swsqueeze's makefile? Dec 27 14:49:17 jayakumar2: OE sets CFLAGS, CPPFLAGS and CXXFLAGS Dec 27 14:49:32 jayakumar2: maybe your buildsystem is not using them? or change them? Dec 27 14:51:45 zecke|23c3, you are right. swsqueeze seems to be redefining CFLAGS Dec 27 14:52:22 i guess i should sed before oe_runmake and set their CFLAGS to be += instead of = Dec 27 14:55:58 niv_one_three: I saw your bluetooth mod based on alberts one. Do you want to add a description about it on opensimpad.org? Dec 27 14:59:55 niv_one_three: or I am allowed to take your pictures? Dec 27 15:04:27 mr_mice : I am waiting to have more info Dec 27 15:04:45 like better shoots and actually see it work Dec 27 15:04:56 all albert left is the source code Dec 27 15:05:22 to use it - I must manually add it to an OE env and build it Dec 27 15:05:40 I am still unable to do that Dec 27 15:05:53 seems 2.4 is long gone Dec 27 15:06:26 and the new 2.6 can't run on the simpad - for the screen driver wasn't writen Dec 27 15:07:03 I even tried bribing zecke - but it didnt help much Dec 27 15:07:06 oh well Dec 27 15:07:35 may soon I will have more time to get in to this Dec 27 15:07:35 hey likewise Dec 27 15:07:52 but I still seek to have a way to build a OE image at home Dec 27 15:07:52 hey dudes and dudettes Dec 27 15:07:53 Jin^eLD: hi Dec 27 15:08:06 btw, the screen driver - can't be that hard to code Dec 27 15:08:34 I mean - I m no hacker - but it should be kind of copy paste when I see other drivers - and the open spec Dec 27 15:08:53 niv_one_three: you could do a short description about what you have done until now Dec 27 15:09:10 YES feel free to use any info i put on the web Dec 27 15:09:19 ok Dec 27 15:09:32 niv_one_three: You know better what you have done ;) Dec 27 15:09:38 I will do that when I get home in 2.5 hours Dec 27 15:09:39 likewise: today at 4 in the morning, a historic moment :> we checked in our current development upnp server code to public svn at sourceforge :> Dec 27 15:09:47 niv_one_three: :) Dec 27 15:09:48 :-) Dec 27 15:09:53 1.5 years since last release hehe Dec 27 15:10:02 niv_one_three: many thx Dec 27 15:10:03 you worked through the nite? Dec 27 15:10:08 no problem Dec 27 15:10:09 we often do that, yes Dec 27 15:10:17 well, we started around 1700 and worked till 04 in the morning Dec 27 15:10:29 niv_one_three: what is your problem with the 2.6 mq200 driver? Dec 27 15:10:32 Jin^eLD: is this for your work? Dec 27 15:10:34 koen: hi Dec 27 15:10:55 likewise: no, this is for fun, our personal opensource project Dec 27 15:11:00 Jin^eLD: is there a .bb yet? :-) Dec 27 15:11:12 the mq200 driver is not compatible with current kernel - this is what I understod Dec 27 15:11:26 likewise: I have to adapt it for svn, but there is.. and then I have to submit it, koen told me to append it to bugtracker Dec 27 15:11:38 Can someone explain the purpose of LEAD_SONAME? Especially if a package builds multiple libs. Dec 27 15:11:53 niv_one_three: wich version of 2.6 do you want to use? Dec 27 15:12:01 mr_mice: got my email? Dec 27 15:12:16 I seek to use the current stable kernel Dec 27 15:12:23 Jin^eLD: I would love to play with it sometime. Do you know if iPod or iTunes can use UPnP (a colleague wants to know)? Dec 27 15:12:34 I think the latest 2.6.19 is good Dec 27 15:13:11 niv_one_three: when did you wrote that mail? Dec 27 15:13:19 likewise: as far as I know - no Dec 27 15:13:31 ipod has no ethernet anyway Dec 27 15:13:35 mr_mice: I was reffering to the address Dec 27 15:13:47 do u have my email address? Dec 27 15:14:15 likewise: one possibility would be to mount ipod as mass storage and access the mp3 files directly Dec 27 15:14:25 and then make them available via upnp to other devices on the network Dec 27 15:14:41 Jin^eLD: no I meant the other way around. Dec 27 15:14:48 niv_one_three: yes Dec 27 15:14:51 ok Dec 27 15:15:01 so feedback me tommorow Dec 27 15:15:04 you mean as a mediarenderer? no.. not that I would know Dec 27 15:15:07 niv_one_three: ok Dec 27 15:15:14 bye for now Dec 27 15:15:26 niv_one_three: bye Dec 27 15:15:53 likewise: you need some udp/tcp stuff for upnp Dec 27 15:16:11 likewise: you could do that if you had a bluetooth enabled player with a tcp stack Dec 27 15:16:17 or something that has wifi Dec 27 15:16:19 Jin^eLD: understood Dec 27 15:16:28 which actually would be pretty cool :> Dec 27 15:16:55 Jin^eLD: I am thinking of creating a WiFi MP3 player connected to my car stereo using the EFIKA Dec 27 15:17:18 Jin^eLD: Using this: http://gnunilink.sourceforge.net/ Dec 27 15:18:53 likewise: wow, cool Dec 27 15:19:58 likewise: seems kind of complicated though? Dec 27 15:20:29 Jin^eLD: unilink looks alot like an SPI bus, so the EFIKA would need to mimick the unibus. Dec 27 15:20:50 aah, ok I think I start to get the idea Dec 27 15:20:57 It simply tells the car stereo there is a new device on the bus. Audio goes over simple RCA jack plugs. Dec 27 15:21:26 The EFIKA would synchronize with my home collection of music when my car is close to home. Dec 27 15:22:00 nice Dec 27 15:22:59 actually it is quite cool that so much stuff is available, that allows to build devices etc. yourself Dec 27 15:39:10 likewise: ping? Dec 27 15:45:16 likewise: can the bootloader in the bios image you gave me load a kernel from network or usb? I'm suddenly getting a kernel panic while the kernel is booting. Appears to be PCI-related, but I'm not entirely sure since I can't see the whole dump. Dec 27 15:55:59 NAiL: pong, yes I think I have tested net booting. Dec 27 15:57:53 NAiL: I'm about to leave here. CU Dec 27 16:35:11 ~botmail for likewise: LEAD_SONAME tells OE which library to use for debian-style shlib renaming Dec 27 16:35:55 koen: is anything special needed for largefile support? I have a configure script that does an AC_SYS_LARGEFILE, turned out it did not work correctly; at least the compiled program did not work as expected; it did work when disabled largefile via the configure option Dec 27 16:37:12 I am preparing a .bb file for you.. we moved the sources of our upnp server to public SF svn.. Dec 27 16:37:22 Jin^eLD: no idea on that, sorry Dec 27 16:37:33 hmm Dec 27 16:37:39 I'll disable largefile for now then... Dec 27 16:39:58 so I create a bugzilla account and then just file a bugreport vs. OE/metadata with my .bb file attached? Dec 27 16:41:29 yes Dec 27 16:41:53 although the 'attachment' option won't appear when you create the bug Dec 27 16:42:05 so you have to create the bug first to be able to attach the .bb Dec 27 16:42:14 ok Dec 27 16:44:00 (which is my major annoyance with bugzilla) Dec 27 16:44:05 :) Dec 27 16:44:42 my annoyance seems to be with my mailserver, not sure if the bugzilla mail got blocked - since I am not getting anything back... and the admin is not around Dec 27 17:03:05 Jin^eLD: pastebin the .bb and I'll add it Dec 27 17:04:17 roger that Dec 27 17:04:17 ah, it's in svn :) Dec 27 17:05:18 http://pastebin.ca/293485 Dec 27 17:05:28 yep :> Dec 27 17:05:51 allthough I think largefile is still not disabled in the svn version Dec 27 17:07:08 so.. I can stop typing my bugzilla report? :) Dec 27 17:08:04 yes :) Dec 27 17:08:04 ok, thanks :) Dec 27 17:08:30 I ran it through oe-stylize.py as well Dec 27 17:09:56 I'll update our file with your changes when you are done.. Dec 27 17:10:39 brb 10 min Dec 27 17:11:58 tmbinc: hey?! Dec 27 17:12:29 mickeyl: hey Dec 27 17:12:37 heya Dec 27 17:12:52 mickeyl: do you plan to come back here? meet el Dec 27 17:12:52 zecke|23c3: ah cool Dec 27 17:13:01 zecke|23c3: where are you? Dec 27 17:13:23 zecke: I don't think I can stay awake for more than a dinner now Dec 27 17:13:28 tmbinc: I don't know how you look like but I'm in the hackcenter Dec 27 17:13:32 tmbinc: micekyl is in berlin too Dec 27 17:13:50 tmbinc: I'm near to the 'Angels' Dec 27 17:14:48 zecke|23c: that is - where? near the entrance? Dec 27 17:16:12 tmbinc: Hackcenter, at the end of the hack center Dec 27 17:16:45 i'm there, too Dec 27 17:16:55 near that non-working DVB-T screen Dec 27 17:17:14 tmbinc: I think I'm a bit more left Dec 27 17:17:40 tmbinc: I can get to the elevator and dance? Dec 27 17:17:52 small moment please Dec 27 17:18:02 I'm waiting for my compiler and mtn anyway :) Dec 27 17:18:32 hey mickeyl Dec 27 17:18:35 yo koen Dec 27 17:18:42 * mickeyl inspects tomorrow's program Dec 27 17:19:14 zecke|23c3:ok now :) Dec 27 17:19:26 we just chinese-copied a 2 layer PCB using a digicam ;) Dec 27 17:19:27 mickeyl: are you interested in a faster libpoppler? Dec 27 17:19:36 koen: absolutely! Dec 27 17:20:10 tmbinc: okay :) Dec 27 17:20:20 although i have yet to see how PDF reading @ 280 dpi looks like... Dec 27 17:20:22 :D Dec 27 17:20:23 mickeyl: we overlooked --enable-fixedpoint for over a year Dec 27 17:20:32 koen: d'oh!? Dec 27 17:20:51 coolness! tried on a pda? Dec 27 17:20:56 how does it 'feel'? Dec 27 17:20:59 mickeyl: the next step is to see which backend is faster, 'cairo' or 'splash' Dec 27 17:21:05 I have yet to try it Dec 27 17:21:05 rock Dec 27 17:21:09 can't wait Dec 27 17:21:29 I'm poking the irex dudes to benchmark it, since pdf is their core business :) Dec 27 17:21:35 righto, good idea Dec 27 17:21:44 Jin^eLD: NOTE: package mediatomb-0.8+0.9pre1+svn20061227-sqlite: completed Dec 27 17:22:08 that irix thing is on sale yet? Dec 27 17:22:19 yes, €650 Dec 27 17:22:25 k, thank Dec 27 17:22:26 s Dec 27 17:22:29 * mickeyl ponders Dec 27 17:22:35 bbl, going to shower, then dinner Dec 27 17:23:27 03koen 07org.oe.dev * r42910a78... 10/ (1 packages/libexif/libexif_0.6.13.bb): libexif: fix do_install and staging Dec 27 17:27:45 ~lart SPI buses Dec 27 17:27:45 * ibot does a little 'dpkg -P SPI buses' action Dec 27 17:30:36 03koen 07org.oe.dev * r43c1d1ca... 10/ (1 packages/mediatomb packages/mediatomb/mediatomb_svn.bb): Dec 27 17:30:36 mediatomb: add svn version Dec 27 17:30:36 * this is a cleaned up version of http://mediatomb.svn.sourceforge.net/viewvc/mediatomb/trunk/mediatomb/scripts/mediatomb_svn.bb?view=markup Dec 27 17:30:53 re Dec 27 17:31:07 koen: cool, thanks :> Dec 27 17:31:27 Jin^eLD: could you sync that back to svn? Dec 27 17:31:40 yep, doing that now Dec 27 17:43:48 03koen 07org.oe.dev * r55c211ee... 10/ (4 files in 3 dirs): gnome-vfs: add 2.6.13 (partially from poky). This version uses dbus instead of corba, hence the missing gnome-vfs-dbus.bb. One version to rule them all :) Dec 27 17:44:26 mickey_berlin_bb: note that commit :) Dec 27 17:48:48 koen: ok, updated with your changes; thx again Dec 27 18:00:46 Jin^eLD: thank you for writing the .bb :) Dec 27 18:00:56 koen: thanks for adding it :> Dec 27 18:54:26 Hey -- What kind of boot times can one expect to get down to with an ARM9 (Actually, an OMAP) in a pretty minimal configuration? It is the deciding factor between linux and say nucleus or whatever Green Hills has to offer in a project I am working on. Dec 27 19:06:38 * koen suggests abusing suspend to disk to get 'boot' times that are ~2 seconds Dec 27 19:37:10 RP: brave warrior :} Dec 27 19:37:26 zecke|23c3: ? Dec 27 19:38:13 RP: cache bug hunting Dec 27 19:38:26 zecke|23c3: ah :} Dec 27 19:38:53 zecke|23c3: I running out of ideas with it :-/ Dec 27 19:40:00 hey RP Dec 27 19:41:14 any ideas on boot times? Dec 27 19:41:56 [20:34] • koen suggests abusing suspend to disk to get 'boot' times that are ~2 seconds Dec 27 19:42:14 Well... Dec 27 19:42:22 We need a real hard-reset Dec 27 19:42:37 hi XorA Dec 27 19:43:04 We're doing a digital satelite radio, it's not radiation hard and if bits get flipped and we go unresponsive or whacky, they need to hit a reset line and have us reset Dec 27 19:44:01 * XorA would suggest in that situation compressing RAM after boot into ROM and having the bootloader just uncomress that and run again Dec 27 19:44:19 done that for military applications before Dec 27 19:44:23 RP: it is not that DMA thing? Dec 27 19:44:29 smaller CPU's though Dec 27 19:44:34 zecke|23c3: Which DMA thing? Dec 27 19:44:39 XorA: That would work way well. Dec 27 19:44:50 RP: russel send a mail which ended "Oh wait" Dec 27 19:44:54 sent... Dec 27 19:44:56 and also did that for a DNA PCR machine in labs where power fail is regular Dec 27 19:45:09 XorA: is that doable with OE? Dec 27 19:45:10 zecke|23c3: Yes. I have no clue which DMA thing he means though... Dec 27 19:45:24 Hi! Dec 27 19:45:36 RP: maybe DMA mem-to-mem xfer? Dec 27 19:45:36 kholz: well youd have to build your image etc, boot it on the board, then use something like a jtag debugger to freeze the system Dec 27 19:45:39 psokolovsky: hi:) Dec 27 19:45:53 We could do that. Dec 27 19:46:07 kholz: its a common technique when you need to "reboot" fast Dec 27 19:46:17 psokolovsky: DMA is mem-mem xfer, yes. Which kernel function would that be though? Dec 27 19:46:36 What about how to reload the saved ram from flash back into ram? Anything you'd reccomend looking into? Dec 27 19:46:50 RP: sorry, I have no clues beyond that ;-) Dec 27 19:46:53 kholz: Ive only done it using my own OS Dec 27 19:46:59 ahh. Dec 27 19:46:59 k Dec 27 19:47:14 kholz: but koens suggestion of looking at suspend to disk is sound Dec 27 19:47:15 I'll look into it. Dec 27 19:47:23 Yeah, though we have rad hard issues Dec 27 19:47:28 kholz: suspend2 attempts to do something similar on x86_64 I believe... Dec 27 19:47:36 or plain x86:) Dec 27 19:47:42 kholz: you would suspend to disk once, then burn that disk image to rom Dec 27 19:47:54 kay Dec 27 19:48:06 Thats a good way to go about it. Dec 27 19:48:16 Thanks for the ideas. Dec 27 19:48:24 * XorA remebers losing flash contents when the UV lasers discharged so I assume same problems happen in satelites Dec 27 19:48:45 yeah Dec 27 19:48:59 Lemme ask if we could get a rom to burn onto the board. Dec 27 19:49:23 mmp, btw, if you'd like to have 2.6.16-hh9 with h5000 fixes, or sth, feel free to ask (that's re: Koen adding your udc patch for -hh8) Dec 27 19:50:19 psokolovsky: Yes, I will, (Erik Hovland even said that he will commit the udc patch, to older series of kernels, but I probably forgot to ask him to do so...) Dec 27 19:50:32 mmp: he did Dec 27 19:50:37 kholz: the ts people use something like that Dec 27 19:50:53 kholz: you can steal some ideas from http://www.embeddedarm.com/linux/fastboot-7400.htm Dec 27 19:51:12 kholz: but most of that page is useless marketing blurb ("Optimized!!!") Dec 27 19:51:17 psokolovsky: also in older 2.6.16? I've seen the patch only in mainline (but maybe I have not looked the right way, I'm not familiar with viewcvs) Dec 27 19:51:34 older 2.6.16 = 2.6.16 HEAD :) Dec 27 19:53:32 koen: thanks for the link. Lemme read up on what uboot can do, see if I could get it to copy flash to ram and begin execution. Dec 27 19:53:57 * RP ponders his utterly useless complicated finite state machine :) Dec 27 19:54:16 psokolovsky: if the patch is in 2.6.16-HEAD, it would be great if it would be also included in -hh9 Dec 27 19:54:51 RP: Don't mention Finite State Machine/Automata!!! (/me is just trying to solve assignments from Formal languages and automatas:) Dec 27 19:55:18 (or how is the proper declination of Automat)on?) Dec 27 19:55:24 Aho - Hopcroff - Ullman Dec 27 19:55:34 hehe Dec 27 19:55:51 gremlin[it]: direct hit:) Dec 27 19:55:52 mmp, http://handhelds.org/cgi-bin/cvsweb.cgi/linux/kernel26/arch/arm/mach-pxa/h5400/h5400.c?only_with_tag=K2-6-16-DEV Dec 27 19:56:24 i study on the book of that trio :) Dec 27 19:56:53 studied ... past ... far far far past :) Dec 27 19:57:04 mmp: so yes, it will go onto -hh9 if/when that will be tagged Dec 27 19:57:17 psokolovsky: ahaa, I have not noticed that:) Thanks for pointing Dec 27 19:57:49 gremlin[it]: :)) Dec 27 19:57:54 gremlin[it], mmp: yep, that's what real computer science is, not pale kernel hacking ;-I Dec 27 19:58:47 psokolovsky: :-)) Yeah, a bit too abstract sometimes, but pretty interesting, though :) Dec 27 19:59:08 ~lart sharp remotes Dec 27 19:59:08 * ibot drops a truckload of VAXen on sharp remotes Dec 27 19:59:30 ouch Dec 27 19:59:35 that's a lot of vaxen Dec 27 19:59:46 VAX on a chip Dec 27 20:01:21 mmp ... i use more times what i learn on that exam than on the C++ course :) Dec 27 20:03:06 gremlin[it]: :) I believe so; for me it's rather interesting course but... sometimes a way difficult too:) Dec 27 20:03:23 mmp: I saw your problem and the mail to linux-arm-kernel btw - I'm not sure if your problem is related or not atm Dec 27 20:04:04 RP: I'm not sure, too, but those two issues might be related together. I should write more information about it, but I was a bit in hurry... . Dec 27 20:05:11 I don't know enough about how caching works on ARMs to state that those issues are connected or not; but it definitely seems that disabling caching in fb area helped... . Dec 27 20:05:23 so at least problem descriptions use some common substrings:) Dec 27 20:06:09 mmp yes was difficult for me too ... Dec 27 20:06:54 gremlin[it]: not mentioning that assignment is quite trivial (with possibly trivial solution), I can't get rid of the feeling that I'm walking around it for few hours already:) Dec 27 20:07:11 (the solution) Dec 27 20:07:43 mmp: say a turing machine can solve this issue and the solution is so trivial that it is not nore worthy?! Dec 27 20:08:47 zecke|23c3: :-))) My problem is that I'm restricted to FSA and regular grammars... Dec 27 20:09:35 But my assignment already took too much lines here; sorry for starting an OT:) Dec 27 20:10:32 took _too many_ lines... . Sorry also for the errors I've made (and I'm making:) Dec 27 20:12:10 if u get rid of my USB Device Controller i'll do your assignements :) *LOL* :) Dec 27 20:12:49 gremlin[it]: :-)) Dec 27 20:20:30 RP: to the caching issue - I have few other drivers behaving quite interesting way, although I think they should work; I'll let you know if I find something directly related to caching... . Dec 27 20:21:38 on the other hand h5550 has actually a bit outdated drivers, so I can't say it can be only caching issue. Dec 27 20:27:53 hi XorA Dec 27 20:30:13 hi ssvb Dec 27 20:30:47 did you check benchmark results at http://www.oesf.org/forums/index.php?showtopic=22280 ? Dec 27 20:31:40 i wonder if it is really memory bus limiting performance on zaurus Dec 27 20:33:36 do you know what optimization options are used to build mplayer for zaurus? maybe just the set of gcc options is not optimal? Dec 27 20:36:37 ssvb: possibly, I did get a 2% improvement by doing an iwmmxt build Dec 27 20:39:04 XorA, do you know what is the effect of -Os, -O2, -O3 options on Zaurus? My tests showed that standard -O4 is the fastest on Nokia 770 for mpeg4, but it results in somewhat worse performance for h264 than -O2 Dec 27 20:39:48 but mpeg4 playback seems to be more important so I decided to keep -O4 Dec 27 20:40:22 ssvb: I havent done things to that detail Dec 27 20:44:25 XorA: I did not do thorough tests either, but as every little bit helps, I at least tested some combination of options and ended up with "-O4 -mcpu=arm926ej-s -fomit-frame-pointer -ffast-math" Dec 27 20:45:04 koen , could you add blueangel into packages/linux/linux-handhelds-2.6.inc at COMPATIBLE_MACHINE ?= Dec 27 20:45:42 XorA: Is it possible to do some quick tests with -mcpu=arm926ej-a replaced with 'xscale' or 'iwmmxt'? Dec 27 20:47:00 rob__w: sure Dec 27 20:47:23 great Dec 27 20:48:01 XorA: by the way, what is the current status of AAC, AC3 and WMA support in mplayer on ARM? Do you know about any useful patches? Dec 27 20:48:40 03koen 07org.oe.dev * r97e37e0d... 10/ (1 packages/linux/linux-handhelds-2.6.inc): linux-handhelds: add blueangel to COMPATIBLE_MACHINE Dec 27 20:49:45 ssvb: my benchmark were with "-Os" Dec 27 20:49:53 bam Dec 27 20:50:51 ssvb: I did enable fixed point AAC decoding once, but it didnt make any different so speed Dec 27 20:52:39 XorA: fixed point is always used on ARM architecture for AAC, no matter you enable it or disable :) Dec 27 20:52:50 there is some ifdef in the sources Dec 27 20:53:14 ssvb: maybe this is new, was a few versions back, it used to be hardcoded to never use fixed point Dec 27 20:54:40 XorA: hmm, that's strange, anyway with fixed point AAC, main profile is not supported at all (and that's the default for encoding AAC audion in mencoder) :( Dec 27 20:55:08 ssvb: I tried playing some ipod video once, it was insanely slow Dec 27 20:57:11 koen: Is it possible to do one more run of the benchmarks with -O4 (or -O3 as I believe these options are the same according to gcc docs)? just to see if it makes any significant difference Dec 27 20:58:10 sure, if I find the time Dec 27 20:58:27 * koen needs to translate a webapp first Dec 27 20:58:32 * koen stabs php Dec 27 21:01:46 ~prod crazy remote code Dec 27 21:01:47 * ibot zaps crazy remote code with a high voltage cattle prod Dec 27 21:18:04 morning all Dec 27 21:31:11 03pfalcon 07org.oe.dev * r3c8de256... 10/ (4 files in 3 dirs): Dec 27 21:31:11 blueangel: Rename machine to htcblueangel. Dec 27 21:31:11 * Because otherwise soon we'll have dozen machines, with most people have no Dec 27 21:31:11 idea what they are, and a bit later - cases of duplicate machine names Dec 27 21:31:11 (HTC nicknames are pretty commonwords). Dec 27 21:31:12 * Discussed with Blueangel people. Dec 27 21:31:14 * 'magician' remains on pH5's conscience ;-). Dec 27 21:37:45 03rwhitby 07org.oe.dev * r634b0214... 10/ (1 packages/images/slugos-image.bb): slugos-image: Fixed missing change to SLUGOS_FLASH_IMAGE Dec 27 22:10:58 RP: this remote is crazy assed :-) Dec 27 22:27:40 XorA: yes. Any luck with it? Dec 27 22:28:23 RP: Well button held down max111 returns 0, button up max111 returns 255 Dec 27 22:28:28 well 255 ish Dec 27 22:28:36 accounting for ADC jitter Dec 27 22:28:54 XorA: Sounds like where JustinP got too :-/ Dec 27 22:29:31 RP: I went back to basics and just loop in interrupt 10 times for values, when I get sensible values I try and make a decent state machine Dec 27 22:30:00 reminds me, where the hell is me multimeter Dec 27 22:30:36 RP: I did go through the pxa270 and max1111 data sheets and I think we are doing the right thing to read the chip Dec 27 22:30:57 XorA: Yes, also note that we use the max1111 for other things and those readings work Dec 27 22:31:37 RP: yeah, I stuck some debug in there and saw sensible stuff Dec 27 22:31:50 and channel 0 does appear to be the remote Dec 27 22:33:52 thats new, white screen of death Dec 27 22:34:42 XorA: Remeber there is an AK Enable GPIO Dec 27 22:35:08 RP: yes, I was playing with that when I got white screen of death :-) Dec 27 22:36:43 ~praise kexec Dec 27 22:36:48 All hail kexec! Dec 27 22:38:44 XorA: Its definetly CH0 and that GPIO appears to pull the AKIN line high, Dec 27 22:40:06 CH0? Dec 27 22:40:34 XorA: Channel 0 Dec 27 22:40:42 yes Dec 27 22:41:52 Putting a scope on it might be interesting... Dec 27 22:42:30 * XorA might be persuaded to crack open a c7x0 but not his terrier Dec 27 22:43:41 I was thinking of a substitue headphone plug. I guess a 4 position one might be tricky to find... Dec 27 22:45:26 RP: well the remote is deconstuctable Dec 27 22:46:07 That would show you what the voltage levels were doing on the line... Dec 27 22:46:49 looks like a late night in the office one night :-) Dec 27 22:47:03 XorA: :) Dec 27 22:49:50 RP: does a schematic exist for that bit of circuit? Dec 27 22:50:41 03awilcox 07org.oe.dev * rba38f1ad... 10/ (1 packages/imagemagick/imagemagick_6.2.9.bb): imagemagick: fix typo, and remove xml and perl support in openprotium Dec 27 22:55:07 RP: is set_scoop_gpio(&spitzscoop_device.dev, SPITZ_SCP_AKIN_PULLUP); Dec 27 22:55:22 RP: working on the right scoop device? that looks like CF power line to me Dec 27 22:55:33 and would explain the wierd CF power errors I am seeing Dec 27 23:07:09 RP: a result Dec 27 23:08:02 RP: I can get readings from the remote :-) Dec 27 23:08:53 is there a way of building a filesystem image manually from ipks which have been generated with openembedded? i've tried to build bootstrap-image and i keep getting errors when it comes to building the image complaining about unsatisfied dependencies (though the ipks have been build for all the needed programs) Dec 27 23:08:54 ~dance Dec 27 23:09:00 * ibot <("<) <('_')> (>")> Dance Kirby Dance! Dec 27 23:09:51 RP: I am getting steady readings from each button that are distinquishable Dec 27 23:11:53 XorA: Yay. So it was using the wrong CF device? Dec 27 23:11:53 here are the errors i'm getting : http://rafb.net/p/N0643Q49.html Dec 27 23:12:17 RP: yeah, wrong GPIO on the wrong scoop Dec 27 23:12:48 RP: for spitz anyway, I see it differs on the other models Dec 27 23:13:20 XorA: Time to audit that I guess... Dec 27 23:13:58 Orkie: Have a look at the full log and see if it helps you find which package is missing Dec 27 23:14:25 RP: my numbers dont correspond with those in the driver, but at least they are consistant Dec 27 23:14:25 where will i find the log? Dec 27 23:14:51 Orkie: Task failed: /home/adan/Projects/open2x/OE/tmp/work/tune-arm920t-linux/bootstrap-image-1.0-r0/temp/log.do_rootfs.12392 ? Dec 27 23:15:05 oh yes, sorry :D Dec 27 23:15:08 RP: they do however correspond with 2.4 code Dec 27 23:16:34 XorA: sounds good :) Dec 27 23:16:52 RP: I shall leave it at that for tonight and see about fixing up a driver tomorrow morning Dec 27 23:16:52 XorA: Which GPIO are you using for Terrier? Dec 27 23:17:31 RP: SCOOP_GPCR_PA12 which corresponds with your comments in spitz.h Dec 27 23:17:34 here is the full log, i can't see anything obviously missing: http://rafb.net/p/UqOPK983.html Dec 27 23:17:55 XorA: and what were you using before that? GPIO13? Dec 27 23:18:50 SCOOP_GPCR_PA16 was in the code you pointed me at Dec 27 23:19:14 oe.reversefold.com/sharpsl-rc.patch Dec 27 23:19:26 and for terrier, scoop2 Dec 27 23:19:47 PA16 is for c7x0 Dec 27 23:20:01 Ah well, we know what the problem was now... Dec 27 23:20:21 RP: well at least I can test on c7x0 as well Dec 27 23:20:36 XorA: I thought the c7x0 needed a dfferent remote? Dec 27 23:20:50 RP: Ill bet its just the resister values that changed Dec 27 23:21:01 XorA: It'd be nice to know Dec 27 23:21:07 RP: giving the wrong numbers for their crappy driver Dec 27 23:21:41 RP: but Ill get something going on spitz then try c7x0 Dec 27 23:22:02 XorA: Sounds good. JustinP will also be pleased :) Dec 27 23:22:12 RP: though is doing it as key events the right way? having zaurusd do the values matching would make more sense to me Dec 27 23:22:49 XorA: How would you feed the values to zaurusd? Dec 27 23:23:30 RP: my knowledge of 2.6 interfaces lacks on that, can you just have an event device return a number 0-255? Dec 27 23:23:55 XorA: Not without breaking the rules. The correct approach is to send keys Dec 27 23:24:02 RP: ok Dec 27 23:24:17 RP: I was just seeing the fun of homebrew remotes Dec 27 23:24:51 XorA: That would be nice to support Dec 27 23:25:15 RP: well we can get what we have working, then someone else can extend it :-) Dec 27 23:25:52 XorA: Bonus marks for a driver that will detect and support both sharp remote controls ;-) Dec 27 23:26:04 * XorA hides Dec 27 23:26:46 I'm actually quite pleased, it means that driver code I threw together wasn't that wrong :) Dec 27 23:27:22 RP: no, I think for spitz it would have workedish with those two simple changes Dec 27 23:28:07 was JustinP testing on spitz as well Dec 27 23:28:25 I think so... Dec 27 23:29:23 RP: well Im going to call it a night, back tomorrow Dec 27 23:30:38 'night XorA Dec 27 23:30:47 gnight folks Dec 27 23:35:35 RP: horror, is preempt guilty? Dec 27 23:48:17 zecke|23c3: its possible. I should disable it and check... Dec 27 23:49:08 PREEMPT is still marked Experimental? Dec 27 23:51:18 zecke|23c3: it is Dec 28 00:06:39 wow I'm watching a good disturbing movie Dec 28 00:07:24 hi. i've done a patch adding a recipe for softsqueeze. is it the right thing for me to post a bug on oe bugzilla with the mtn diff attached? Dec 28 00:07:48 jayakumar2: yes, and bitch us if we forget to comment :) Dec 28 00:07:58 ok, cool :-) Dec 28 00:08:19 seriously, grab a stick and poke us Dec 28 00:32:10 cool. added under bug 1717 . hope it doesn't look too bad, it is my first oe patch Dec 28 00:41:22 03pfalcon 07org.oe.dev * rcc9abaf9... 10/ (3 files in 2 dirs): (log message trimmed) Dec 28 00:41:22 linux-handhelds-2.6: Add 2.6.19-hh7. Dec 28 00:41:22 * Replaces 2.6.19-hh6. Dec 28 00:41:22 * Changes since -hh6 include: Dec 28 00:41:22 1. Further cleanup of codebase for 2.6.19. Dec 28 00:41:23 2. Migrate away from legacy function/symbol names in Generic IRQ API. Dec 28 00:41:27 3. h3900 SD/MMC and LCD pwoer management support. Dec 28 01:24:27 Hey guys Dec 28 01:24:43 me and a buddy of mine wrote this script to send files from your computer to your zaurus Dec 28 01:24:57 it finds the IP of your zaurus and then uses scp Dec 28 01:24:59 http://crimelords.org/~hoax/zcp-FINAL.sh.txt Dec 28 01:25:02 that's the code Dec 28 01:25:06 it works great Dec 28 01:25:11 nice and simple syntax Dec 28 01:26:13 2 Dec 28 01:28:20 it's for people with unix machines <__< **** ENDING LOGGING AT Thu Dec 28 02:59:57 2006