**** BEGIN LOGGING AT Wed Sep 26 02:59:56 2007 Sep 26 06:51:38 03pfalcon 07org.oe.dev * rad349285... 10/ (1 packages/linux/linux-handhelds-2.6.inc): linux-handhelds-2.6: Generate sample HaRET boot script just like reflash.ctl. Sep 26 06:51:42 03svolpe 07org.oe.dev * r6658e557... 10/ (1 packages/pango/pango-directfb_1.18.1.bb): pango-directfb: Removed since it broke pango builds for other distros. Sep 26 06:51:47 03svolpe 07org.oe.dev * rd6922094... 10/ (1 packages/cairo/cairo-directfb_1.4.8.bb): cairo-directfb: R1 Added RPROVIDES to stop interference with cairo Sep 26 06:51:53 03svolpe 07org.oe.dev * rddd0bbe8... 10/ (1 packages/gtk+/gtk+-directfb_2.10.14.bb): gtk+-directfb: added RPROVIDES Sep 26 06:51:59 03svolpe 07org.oe.dev * rbe38a621... 10/ (1 packages/pango/pango-directfb_1.18.1.bb): pango-directfb: Added it back with RPROVIDES to stop conflicts Sep 26 06:52:04 03pfalcon 07org.oe.dev * r2bd3aa80... 10/ (1 packages/images/liveramdisk-image.bb): Sep 26 06:52:04 liveramdisk-image: No longer stuff payload during, doesn't scale. Sep 26 06:52:04 * Instead, this is done during actual LiveRamdisk building. Sep 26 07:21:23 ...you may be a code geek with lots of experience in your particular niche who believes that some combination of cross-platform development tools and binary translation is good enough to make ISA irrelevant to any hardware design decision, but you are wrong, wrong, wrong. http://arstechnica.com/articles/paedia/cpu/Intels-x86-ISA-grows-down-today-laptops-tomorrow-the-iPhoneBlackberry.ars Sep 26 07:38:16 morning. how do i get the python-psyco bitbake file compiled? Sep 26 07:38:26 bitbake python-psyco doesnt work Sep 26 07:38:35 bitbake nameofbbfile.bb neither Sep 26 07:47:05 did someone ever play with saa711x ? Sep 26 08:21:14 morning Sep 26 08:56:07 where are the checksums stored ? I got "The MD5Sums did not match." for tree Sep 26 08:58:04 i wannt to install python-psyco, the .bb file exists Sep 26 08:58:40 but i cant install it with bitbake python-psyco Sep 26 08:59:08 for bitbake itself or for your target arch ? Sep 26 09:10:57 the issue with tree seems to be that the package bumped silently its version number to 1.5.1.1, with a symlink to the old 1.5.1 tarball Sep 26 09:14:36 nud: for bitbake (host_tools) Sep 26 09:15:16 nik0n: then you should install the one from your distro Sep 26 09:16:58 no i want a complete selfcompiled toolchain and hosttools Sep 26 09:17:01 in staging dir Sep 26 09:39:01 hi Sep 26 10:17:56 mwester, what log are you refering to? what I don't understand is on my system bitbake does choose pango over pango-directfb if I just do bitbake pango Sep 26 10:20:44 hi all Sep 26 10:25:23 re Sep 26 10:30:19 03svolpe 07org.oe.dev * r067d06cf... 10/ (1 packages/pango/pango-directfb_1.18.1.bb): pango-directfb: Remove it again do to conflicts with pango Sep 26 10:44:31 s/do/due Sep 26 10:44:32 ;] Sep 26 11:04:39 hm.. at what point are files in tmp/deploy being removed or overwritten ? should I copy/move everything right away or do those files remain untouched unless I change/rebuild the same target? Sep 26 11:07:42 For me, building open-moko, they're all still there Sep 26 11:08:20 The builds named with date and time, and an automatic symlink to the latest build Sep 26 11:08:25 well i backup up the files just to be on the safe side nevertheless hehe Sep 26 11:08:46 Well, only takes 10h or so to build from scrat... :P Sep 26 11:09:01 haha Sep 26 11:09:02 ;] Sep 26 11:09:16 i had that "experience" yesterday ;] Sep 26 11:09:32 not open-moko though, and not 10h Sep 26 11:09:37 but still like 3h or so Sep 26 11:12:26 morning all Sep 26 11:12:43 hi RP Sep 26 11:13:49 Philippe, was about to say something in finish, but realized that I don't know a single decent word in Finnish hehe.. ;] Sep 26 11:20:44 morning RP Sep 26 12:14:25 zecke, ping Sep 26 12:14:30 pong Sep 26 12:14:42 still looking for hotels ..... Sep 26 12:14:48 http://www.maerkischer-hof-berlin.de/ Sep 26 12:14:53 everything booked our or too expensive? Sep 26 12:14:53 these guys are fax only Sep 26 12:14:58 yeah Sep 26 12:15:06 booked or not emailing me Sep 26 12:15:23 checking on partials, one palce has Sat-Tues open Sep 26 12:15:57 anyway you can check for roms at the link above? Sep 26 12:16:00 Crofton|home: try the crappy^beautiful hostel :) Sep 26 12:16:08 soon :) Sep 26 12:16:14 how is the hostel? Sep 26 12:16:23 zecke: why is Berlin so busy that week? Sep 26 12:16:31 Crofton|home: it is nearby and it looks small Sep 26 12:16:41 Crofton|home: well, I would call it, but I'm not even in Germany atm Sep 26 12:16:46 arg Sep 26 12:16:53 forgot about that Sep 26 12:17:08 XorA: no idea, OEDEM? the 3rd of october is a national holiday though Sep 26 12:17:08 wait, they have an email address Sep 26 12:17:44 zecke, I iwill; try email Sep 26 12:18:28 what is the holiday? Sep 26 12:18:55 Crofton: only a minor one, unification of Germany ;) Sep 26 12:19:18 ah Sep 26 12:19:32 what year was that? Sep 26 12:19:35 * Crofton|home should know Sep 26 12:19:37 http://en.wikipedia.org/wiki/German_Reunification Sep 26 12:22:53 * Crofton|home still thinks the iron curtain fell "recently" Sep 26 12:25:19 Crofton|home: it did, compared to the history of the universe Sep 26 12:25:29 or my age :) Sep 26 12:39:00 Crofton|home: http://www.hotel-albrechtshof.de/ checked that one? Sep 26 12:39:33 wow, that is way out... :) Sep 26 12:39:57 urg Sep 26 12:40:00 yeah :) Sep 26 12:40:07 Crofton|home: no, we have two albrecht streets in berlin :) Sep 26 12:40:16 Crofton|home: it is not too bad, ~10 minutes of walk to the office Sep 26 12:40:22 I have to multiply everything by 1.4 also :) Sep 26 12:40:38 lol Sep 26 12:41:17 Crofton|home: there are some bars in that area, you are not requested to go to the office :) Sep 26 12:45:54 Crofton|home: http://maps.google.com/maps?f=q&hl=en&geocode=&q=Albrechtstra%C3%9Fe+8,+10117+Berlin&sll=52.456794,13.324356&sspn=0.012108,0.030427&ie=UTF8&ll=52.525517,13.386068&spn=0.012089,0.030427&z=15&om=1 Sep 26 12:46:27 this is? Sep 26 12:46:54 Crofton|home: hotel-albrechtshof.de Sep 26 12:47:00 the address :) Sep 26 12:47:07 ah Sep 26 12:47:27 jackpot Sep 26 12:47:37 http://www.gaestehaus-berlin-mitte.de/preise/index.php Sep 26 12:48:28 Is this close? Sep 26 12:49:05 Crofton|home: let me find the address Sep 26 12:49:45 Crofton|home: not bad Sep 26 12:49:54 Crofton|home: http://maps.google.de/maps?f=d&hl=de&daddr=Habersaathstra%C3%9Fe+40a,+10115,+Berlin&ie=UTF8&om=1&ll=52.529956,13.385081&spn=0.012897,0.030427&z=15 Sep 26 12:50:24 .5 K? Sep 26 12:50:33 ok, I am going to book Sep 26 12:50:59 Crofton|home: more than 500 meter Sep 26 12:51:17 but not much right? Sep 26 12:51:30 at this point, it looks like my best bet Sep 26 12:51:35 Crofton|home: route planning says 1km Sep 26 12:52:15 I won't get mugged, right? Sep 26 12:52:40 why? while walking? Sep 26 12:52:47 * Crofton|home can walk a K no problem Sep 26 12:52:48 yeah Sep 26 12:52:55 Crofton|home: depends if I see you with cool hardware :-|D Sep 26 12:53:00 heh Sep 26 12:53:18 alright, I am going to make a reservation Sep 26 12:53:37 I need the exercise :) Sep 26 12:56:03 Crofton|home: damn, I promised to not make any long walks :( Sep 26 12:59:10 let me try a couple otehrs Sep 26 12:59:42 do someone have experience with i2c devices ? Sep 26 13:01:07 nud: what do you mean experience Sep 26 13:01:53 XorA: like: is there usually something special to do to get a i2c device working, or just modprobing the device module should be enough? Sep 26 13:02:05 nud: just modprobing normally enough Sep 26 13:03:26 I was wondering because there is a i2c_pxa module but it's not loaded Sep 26 13:03:49 well you will also need to load the correct i2c driver for your device Sep 26 13:04:12 and the device I want to communicate with is a saa7113 chip (saa7115 module) Sep 26 13:04:47 nud: depending what you want to do, I2C is pretty easy to use from user space as well: http://pastebin.ca/715731 Sep 26 13:05:22 cbrake: well I'm still on the video input issue for the cm-x270 Sep 26 13:06:09 I have i2c_pxa, and saa7115 which is the chip that does frame-grabbing on the sb-x270 board Sep 26 13:06:22 now I'm wondering if it should just work or if there is something subtle in there Sep 26 13:06:44 nud: at the kernel level, things get a little more complex with the I2C driver model Sep 26 13:06:45 given that the saa7115 module uses i2c as well Sep 26 13:06:58 I need a V4L interface Sep 26 13:07:14 which should be provided by the driver Sep 26 13:08:35 cbrake: did you use the cm-x270 i2c bus ? Sep 26 13:08:49 nud: yes, that is what the sample code I pastebin'd is for Sep 26 13:08:59 nud: it talks to a small A/D I2C chip Sep 26 13:09:17 ok so you just modprobe i2c_pxa ? Sep 26 13:09:41 nud: booting my system so I can check ... Sep 26 13:11:13 I don't even have a /dev/i2c* node Sep 26 13:11:47 i2c_pxa 6240 0 Sep 26 13:11:47 i2c_dev 6204 0 Sep 26 13:11:47 i2c_core 20264 2 i2c_pxa,i2c_dev Sep 26 13:11:52 OK, I have a reservation at the Hotel Berlin Sep 26 13:11:58 the /dev/i2c is provided by i2c_dev Sep 26 13:12:00 49 euro per night Sep 26 13:12:11 nud: if you are only using I2C in kernel space, you probably don't need that Sep 26 13:12:30 nud: but, it is a easy way to test things Sep 26 13:13:05 ah ok, i2c_dev makes the /dev Sep 26 13:14:04 nud: neat, I've never even looked at the video capture stuff on the sb-x270 ... Sep 26 13:14:28 I still don't have the /dev/video0 I need Sep 26 13:15:08 despite saa7115 is used by v4l2_common Sep 26 13:25:14 I've been told that there are doxygen docs for libipkg somewhere... Anyone know where they are? Sep 26 13:25:39 is there a ramdisk image target? Sep 26 13:26:04 i have now cpio cramfs ext2/3 and jffs2 images and a glomation evaluation board with a redboot bootloader Sep 26 13:26:14 i flashed every image and tried to access it Sep 26 13:26:34 but i always get a UDF-fs: No VRS found Sep 26 13:26:34 mount: Mounting /dev/mtdblock2 on /mnt failed: Invalid argument Sep 26 13:26:44 error. anyone got a suggestion? Sep 26 13:29:20 Jin^eLD, ping Sep 26 13:33:45 hrm... "checking if image is initramfs...it isn't (bad gzip magic numbers); looks like an initrd" /me scratches his head -- anyone using compulab hardware around? :) Sep 26 13:35:56 gilligan_: yes Sep 26 13:36:08 gilligan_: I maintain the cm-x270 support in OE Sep 26 13:36:41 aaah.. you're charles from bec systems? Sep 26 13:36:54 gilligan_: close -- Cliff :-) Sep 26 13:37:00 cliff cliff.. okay :) Sep 26 13:37:13 cbrake, i mailed you some time ago.. "Tobias Pflug" .. from atm computer .. Sep 26 13:37:14 ;] Sep 26 13:37:26 I'm also trying to get stuff working on cm-x270 :p Sep 26 13:37:36 gilligan_: ahh, very good -- nice to connect the IRC :-) Sep 26 13:37:45 cbrake, right :) Sep 26 13:38:14 cbrake, i've just followed your description in the compulab-pxa270.conf file.. Sep 26 13:38:24 so has anyone been able to build gcc-cross-initial 4.2.1 successfully? Sep 26 13:38:40 cbrake, angstrom based uclibc etc.. Sep 26 13:39:27 cbrake, however the kernel doesn't seem to recognize the gzipped cpio archive as intitramfs for some reason Sep 26 13:39:55 gilligan_: very strange. You set "ramdisk on" in the BL I assume Sep 26 13:40:24 BL? Sep 26 13:40:24 gilligan_: and the cpio.gz image was about 1MB in size? Sep 26 13:40:28 gilligan_: bootloader Sep 26 13:40:42 ah.. oh.. let me double check.. Sep 26 13:41:24 oh and the next question right away.. there is one .bin zImage and one .cmx270 which is slightly smaller -- what is the difference ? Sep 26 13:41:49 there is one that works and one that doesn't ;-) Sep 26 13:41:54 iirc you have to use the .bin one Sep 26 13:42:04 Chase: yes, I build a complete 4.2.1 toolchain for PXA270 a few days ago. kicking off another build now. Sep 26 13:42:30 gilligan_: depends if your are downloading using USB or TFTP. USB use .cmx270, TFTP use .bin Sep 26 13:42:33 cbrake, ramdisk option was disabled in the bootloader.. that certainly "didn't help" hehe Sep 26 13:42:34 everything builds fine with angstrom 2007.1, but I'd like to use 2008.1 for maverick crunch support Sep 26 13:43:03 cbrake, oh alright.. i'm using tftp w/ .bin so everything is fine Sep 26 13:43:11 so lets try this again.. Sep 26 13:44:02 Chase: I suspect that crunch pulls in some things that I'm not seeing in my Xscale build then Sep 26 13:44:33 cbrake, that worked.. Sep 26 13:44:34 Chase: Koen may know as I think he worked on the maverick crunch stuff at one point Sep 26 13:44:53 it fails compiling regex.c in libiberty cause it can't find sys/types.h Sep 26 13:45:01 I don't think it has anything to do with crunch support Sep 26 13:45:11 I've grepped all the patches in 4.2.1 and 4.1.2 Sep 26 13:45:18 and there's nothing that affects libiberty Sep 26 13:45:33 any1 knows how to bring a jffs2 image to a redboot device? Sep 26 13:45:42 Chase: you are building uclibc if I recall? Sep 26 13:45:46 yes Sep 26 13:45:51 uclibcgnueabi Sep 26 13:46:09 I was able to build angstrom-minimal-image-with-mtd-utils just fine Sep 26 13:46:29 after a few hacks to conf/machine/include/tune-ep9312.inc Sep 26 13:46:33 Chase: we had to remove crunch from OE because the patches totally b0rked gcc even when doing non crunch builds Sep 26 13:46:46 Chase: they fucked gcc's internals in a bad way Sep 26 13:46:56 Chase: ok, I'll try that here for Xscale Sep 26 13:46:57 it's in gcc 4.2.1 on oe Sep 26 13:47:02 cbrake, so.. almost there.. what do I have to specify as root ? /dev/ram ? /dev/nor ? /dev/mtd ? bit clueless on that Sep 26 13:47:19 rp: is ross on irc ? Sep 26 13:47:25 gilligan_: with initramfs, you don't have to specify root=, it will always use the initramfs Sep 26 13:47:34 gilligan_: you will however need to specify init=/sbin/init Sep 26 13:47:47 cbrake, ah right.. that's why Sep 26 13:47:59 gilligan_: as initramfs only looks for init in /init -- never figure out why Sep 26 13:48:31 cbrake, huh.. ok.. Sep 26 13:49:13 Chase: perhaps file a bug with a build log if you have not already Sep 26 13:49:34 cbrake, i fiddled around with some special purpose live distribs at work where i used some self-written init.. but i always thought default would be /sbin/init ;] Sep 26 13:49:40 I'm going to try to debug it today Sep 26 13:49:48 wondering if other see the problem Sep 26 13:49:56 others Sep 26 13:52:43 Chase: are you building on a 64-bit host? Sep 26 13:52:52 x86 Sep 26 13:53:20 as in, x86, not x86_64 Sep 26 13:54:04 Chase: ok, there are some libiberty issues with building binutils on x86_64 Sep 26 13:54:27 cbrake, ATAG_INITRD is deprecated; please update your bootloader. <-- is that critical or can I ignore that for now? Sep 26 13:55:33 gilligan_: you can ignore that Sep 26 13:56:00 gilligan_: the error messages mean the initramfs did not take .. Sep 26 13:56:35 well it does not seem to start the initramfs init Sep 26 13:59:29 it does however report "checking if image is initramfs... it is" Sep 26 14:00:26 Chase: ucblic/2008.1/xscale build started ... Sep 26 14:00:36 ok Sep 26 14:01:02 my current thoughts are that gcc 4.1.2 and 4.2.1 are having different configure switches possibly Sep 26 14:01:26 I'm going to compare logs between an angstrom 2007.1 and 2008.1 build Sep 26 14:01:32 and see what's different Sep 26 14:01:37 Chase: sounds good Sep 26 14:01:47 Chase: you from OH@USA? Sep 26 14:01:54 yep Sep 26 14:02:05 where'd you get that from? Sep 26 14:02:10 Chase: neat -- I'm from Clev/Akron area Sep 26 14:02:20 oh, ok Sep 26 14:02:31 my wife goes to osu med school, so I Sep 26 14:02:37 'm stuck hanging around here Sep 26 14:02:43 Chase: heh Sep 26 14:03:12 Chase: the RR ip address gives you away :-) Sep 26 14:03:22 yeah, that's what I figured Sep 26 14:03:38 cbrake, wow, there are quite a few PA/Ohio/WV OE members lately :-) Sep 26 14:04:03 Gerrath: yeah, maybe we need to host a OEDEM/midwest :-) Sep 26 14:04:17 cbrake, you beat me to the punch. Sep 26 14:05:05 cbrake: Good idea. Sep 26 14:05:39 cbrake, any idea what I am doing wrong here ? ;) Sep 26 14:06:09 gilligan_: you may want to look in the tar version of the image and make sure init is in /sbin/init Sep 26 14:06:30 gilligan_: I'll try it later today after/if my uclibc build finishes Sep 26 14:06:33 cbrake, i did.. it is a link to /sbin/init.sysvinit Sep 26 14:07:13 mwester, you mentioned last night a log file I should view to see the pang-directfb issue.. I'm not sure I know which log your refering to. FYI: when I do bitbake pango it builds pango and not pango-directfb Sep 26 14:08:29 cbrake, i also tried to specify /sbin/sysvinit directly, just to be sure.. but doesn't help Sep 26 14:08:39 Right - the bitbake command is looking for a specific recipe name (filename.bb), rather than a dependency to satisify. So the problem will appear only if you bitbake something that has a stated dependency on pango. Sep 26 14:09:12 mwester, easy enough, I can just make a very simple dumby file :-) Sep 26 14:09:13 gilligan_: so you get the "checking for initramfs -- it is" type message? Sep 26 14:09:23 cbrake, yes i do Sep 26 14:09:37 bitbake -g -n recipe-that-requires-pango.bb Sep 26 14:10:05 That command will not do the build, but create "depends.dot", which will have the entire dependency tree outlined. Sep 26 14:10:08 gilligan_: I don't know then -- this is a very difficult area to debug -- handoff of kernel to init :-( Sep 26 14:10:25 I just grep that for "directfb"... Sep 26 14:10:46 mwester, great, thanks. Sep 26 14:12:13 cbrake, odd.. well the kernel should only try to mount root if the execution of /sbin/init fails i assume? Sep 26 14:12:32 gilligan_: yes, seems that way Sep 26 14:12:53 hm Sep 26 14:19:05 hmm, I create a OeLocalGroups page on the OE wiki, and and there is no edit button? Sep 26 14:19:14 anyone know how to use this wiki? Sep 26 14:19:17 Does anyone have experience using OE with a chip that doesn't have an MMU? Specifically, I am looking at building a minimal system for an ARM7TDMI microcontroller. I have been spoiled by OE, and have all my bitbake recipes working, and I am looking at the uClinux build system and thinking "how could this possibly work" :-) Sep 26 14:21:17 koen|away, has done some black fin work, which I think is uclinux Sep 26 14:24:01 Crofton: that's encouraging. I grepped all the kernel configs in the tree, and didn't see anything without an MMU. After my experience getting OE running with a 2.4 kernel, I know that there are a number of assumptions in the distros about running 2.6, and I am hoping to avoid a repeat. Sep 26 14:25:03 Guess I will have to give it a try and see how far I get. Sep 26 14:26:03 yeah, too bad he is working at the moment Sep 26 14:26:41 Crofton: Working on what? Sep 26 14:27:44 working for his day job :) Sep 26 14:28:23 Crofton: This is my day job (and nights, lately) :-) Sep 26 14:28:27 heh Sep 26 14:36:00 cbrake, any info on your build? Sep 26 14:36:12 I see the some of the source of my problems Sep 26 14:36:51 in gcc 4.1.2 in the base Makefile all-target-libiberty does not do anything Sep 26 14:36:59 but in 4.2.1, it does stuff Sep 26 14:37:08 and it's in that stuff that gcc is breaking Sep 26 14:42:31 Chase: uclibc-0.9.29-r7: task do_package: started Sep 26 14:43:02 has it completed gcc-cross-initial? Sep 26 14:43:25 it doesn't take me long before I hit gcc-cross-initial, so I would have thought you would see the error by now Sep 26 14:44:42 Chase: yes, gcc-cross-initial is needed to build uclibc Sep 26 14:44:48 Chase: so it seems to be working Sep 26 14:44:56 hmmm... Sep 26 14:45:23 can you check your log.do_compile for gcc-cross-initial to see if it mentions all-target-libiberty? Sep 26 14:47:32 I guess you probably won't see anything about that Sep 26 14:47:39 make won't say anything Sep 26 14:48:08 could you look for "Checking multilib configuration for libiberty..." Sep 26 14:51:12 Chase: http://pastebin.ca/715844 Sep 26 14:51:33 thanks Sep 26 14:55:55 ummm, cbrake, that's for 4.1.2 Sep 26 14:56:19 has anyone used setup.py to build a C++ project? Sep 26 14:59:14 Chase: it looks like OE uses gcc 4.1.2 for gcc-cross-initial, and then 4.2.1 for gcc-cross Sep 26 14:59:44 cbrake@happy:/build/oe/build_targets/cm-x270-uclibc/tmp/work/armv5te-angstrom-linux-uclibcgnueabi$ ls |grep gcc Sep 26 14:59:47 gcc-cross-4.2.1-r6 Sep 26 14:59:50 gcc-cross-initial-4.1.2-r8 Sep 26 15:00:42 oh geez, I forgot that I had added gcc-cross-initial = 4.2.1 to my preferred versions Sep 26 15:01:11 see, if I left it at 4.1.2 for initial, then -mfp=maverick would fail on compilations Sep 26 15:01:26 so I'm trying to figure out how to build 4.2.1 as initial Sep 26 15:01:55 Chase: ahh, you are breaking new ground then :-) Sep 26 15:01:55 this may be the reason why it's left at 4.1.2 as initial Sep 26 15:02:00 Chase: no doubt Sep 26 15:02:04 who would have left it there Sep 26 15:02:11 as in, who might know more about this? Sep 26 15:02:36 Chase: Koen, but he is not here right now Sep 26 15:02:47 are there certain times when he's on? Sep 26 15:03:35 Chase: 9-5 GMT+1, but it looks like he might be doing something else today Sep 26 15:03:41 Chase: unknown -- I'd post a message to the mail list or wait for koen|away to change to koen Sep 26 15:04:05 alright Sep 26 15:04:11 he said he would not be around most of this week due to work commitments Sep 26 15:04:35 and just for the record, I work with a bunch of guys who do not understand Linux/Unix Sep 26 15:37:19 * cbrake mumbles something about getting loosing his mind Sep 26 15:37:33 gilligan_: init is not the right cmdline param -- looking for the right one Sep 26 15:39:07 cbrake, huh.. i've used that before and it worked fine.. even if on totally different projects Sep 26 15:39:17 gilligan_: rdinit=/sbin/init Sep 26 15:40:04 aah.. i Sep 26 15:40:06 alright Sep 26 15:40:24 cbrake, well i always used initrd and not ramfs ... Sep 26 15:40:42 gilligan_: that works :-) Sep 26 15:41:23 gilligan_: it looks like the console is not set up for the serial port, but I do get a console on the display Sep 26 15:42:01 ah good Sep 26 15:42:08 * gilligan_ gets his old tft from the attic Sep 26 15:42:46 gilligan_: make that "login prompt", not console Sep 26 15:49:09 Gerrath: mwester: I started the following page: http://elinux.org/LocalUserGroups -- maybe we could do a little planning there Sep 26 15:51:47 cbrake, s/console/login\ prompt/ .. yes.. got it ;] Sep 26 15:52:04 gilligan_: great! Sep 26 15:53:49 cbrake, think I found the culprit Sep 26 15:54:10 so libiberty is only built if there are other gcc lib* being built Sep 26 15:54:23 I have no clue what libiberty actually does, but that's how it's set up Sep 26 15:54:46 well, between 4.1.2 and 4.2.1, they added libgomp, a gnu openmp lib Sep 26 15:54:53 and it builds by default Sep 26 15:55:12 thus now, libiberty is set to build in 4.2.1 because libgomp is set to build Sep 26 15:55:16 I think it exists so that someone can type "-Liberty" on a command line. Sep 26 15:55:32 I tried --disable-libiberty, but gcc failed to compile Sep 26 15:55:50 so I tried --disable-libgomp, which wiped out both libiberty and libgomp Sep 26 15:55:53 and it compiled Sep 26 15:56:12 so I'm going to clean up my tmp dir and start from scratch just to make sure it still works Sep 26 16:02:18 ade|desk: Probably in gnome places Sep 26 16:02:58 cbrake, actually... darn.. i only get a blank red screen on the vga port.. Sep 26 16:11:24 cbrake, what's the last message that you get on the serial ? Sep 26 16:15:59 cbrake, when I get "Freeing init memory: 104K" the monitor goes red, and that's it Sep 26 16:19:00 cbrake, thats great! Sep 26 16:21:47 gilligan_: you should get lots of messages -- starting init tasks, etc Sep 26 16:22:18 cbrake, alright.. so init just ain't starting here then.. at first i thought maybe for some odd reason it would switch to monitor at this stage suddenly Sep 26 16:22:47 gilligan_: the only thing that does not come out on the serial port right now is the login prompt Sep 26 16:22:56 i see Sep 26 16:23:35 well it just idles after freeing init memory here now Sep 26 16:23:50 s/idles/stops Sep 26 16:25:10 cbrake, Angstrom-minimalist-image-mtdutils-glibc-ipk-2007.9-test-20070926-compulab-pxa270.rootfs.cpio.gz <-- you used this as well right ? Sep 26 16:25:45 cbrake, then downloaded to SRAM and flashed to NROM Sep 26 16:26:26 gilligan_: yes, that is the image I used -- downloaded to flash using: download ramdisk tftp Sep 26 16:26:31 gilligan_: and then "flash ramdisk" Sep 26 16:26:51 x_X Sep 26 16:28:25 reflashing kernel/ramdisk to be sure that nothing is corrupt.. Sep 26 16:37:47 sigh.. flashed again and now the ramdisk actually /is/ corrupt Sep 26 16:40:37 btw... are the two red LEDs on the board between PS/2 and serial/VGA always on? just noticed that and was wondering.. Sep 26 16:46:36 cbrake, reflashed everything.. but it still stops at freeing init memory Sep 26 16:46:53 cbrake, how big are your rootfs/zimage files? Sep 26 16:50:13 gilligan_: zImage 1.28M, cpio.gz 1.17M Sep 26 16:50:42 almost every time I change oe files, it does "Handling BitBake files" for a while Sep 26 16:50:52 is there any way around this? Sep 26 16:50:59 seems like a big waste of time Sep 26 16:51:25 I've tried setting CACHE="" in my local.conf, but bitbake won't run like that Sep 26 16:51:46 cbrake, zImage 1.3M, cpio.gz 2.3M .. i suppose that is because forgot to disable locales ? Sep 26 16:52:23 Chase: if you change a conf file, this changes the global variable space. And the global variable space can affect any recipe, so they must all be reparsed Sep 26 16:53:18 Chase: psycho may help some. I can't use it though as I run amd64 Sep 26 16:53:35 can't we just read the relevant files for each package as it is being built? Sep 26 16:54:04 cbrake, so maybe ramdisk&kernel is too big? but I don't quite think so, because in that case the kernel would have told me right away during loading Sep 26 16:54:21 Chase: I don't think so because dependencies can only be determined after everything has been parsed Sep 26 16:54:33 Chase: you may be able to set BBMASK during dev to speed things up a little Sep 26 16:55:09 I use gentoo and it doesn't take this long to go through files when I change some configuration Sep 26 16:55:18 it just seems like there's something awry Sep 26 16:55:53 what bb file did you change? Sep 26 16:56:04 cbrake, you are also using angstrom-2007.1 right ? Sep 26 16:57:22 gilligan_: I'm using 2008.1 right now, but 2007.1 should produce about the same results Sep 26 16:57:37 gilligan_: 2.3M sounds a little big Sep 26 16:58:19 cbrake, well I guess I should set ENABLE_BINARY_LOCALE_GENERATION=1 and GLIBC_GENERATE_LOCALES = "en_US.UTF-8" Sep 26 16:58:26 gilligan_: the size of the NOR flash partition is only 2359296 bytes Sep 26 16:58:42 eh.rather.. ENABLE_BINARY_LOCALE_GENERATION=0 Sep 26 16:59:09 practically changing any file will do it Sep 26 16:59:28 Chase: changing a bb file will only result in reparsing just that file Sep 26 16:59:29 for instance conf/machine/include/tune-ep9312.inc Sep 26 16:59:44 Chase: right, that affects the global var space Sep 26 16:59:50 ok Sep 26 17:00:25 cbrake, i'll rebuilt with ENABLE_BINARY_LOCALE_GENERATION = "0 and see how big the ramdisk gets.. Sep 26 17:03:31 cbrake, and by the way.. isn't NOR flash 4MB and not just 2359296 ? Sep 26 17:04:02 gilligan_: kernel takes up some, BL takes some, param block, and you are left with 2.3M for your rootfs Sep 26 17:05:00 cbrake, i see.. so i have to rebuild angstrom without the locales Sep 26 17:05:46 gilligan_: let me check, there may be a shortcut ... Sep 26 17:06:13 cbrake, well i could perhaps just remove some files from the cpio archive.. Sep 26 17:07:15 gilligan_: you could try export IMAGE_LINGUAS = "" in the image file Sep 26 17:07:39 gilligan_: already there Sep 26 17:08:06 gilligan_: hmm, well seems like something could use some fixing there unless the uclibc libs themselves are bloated Sep 26 17:09:12 92K lib/udev Sep 26 17:09:12 48K lib/modules Sep 26 17:09:12 2.5M lib/ Sep 26 17:09:53 gilligan_: pastebin a recursive listing of lib/ Sep 26 17:10:16 sec.. Sep 26 17:10:52 http://pastebin.com/m47c9af04 Sep 26 17:14:38 gilligan_: I think you are building glibc Sep 26 17:15:06 gilligan_: instead of uclibc Sep 26 17:16:05 hum.. well I did specify linux-uclibc in local.conf Sep 26 17:16:37 but i suppose i should have done that in the angstrom config file as well Sep 26 17:16:39 ? Sep 26 17:18:25 hm.no..doesn't seem so Sep 26 17:18:43 gilligan_: ANGSTROM_MODE = "uclibc" in local.conf Sep 26 17:18:56 d'oh Sep 26 17:19:08 couldn't find that info anywhere ;/ Sep 26 17:20:24 gilligan_: OeFaq Sep 26 17:20:41 gilligan_: perfectly obvious :-( Sep 26 17:20:48