**** BEGIN LOGGING AT Sun Aug 26 02:59:57 2007 Aug 26 07:55:58 * * OE Bug 2875 has been created by bluelightning(AT)bluelightning.org Aug 26 07:56:00 * * Opie first-run wizard broken due to path changes Aug 26 07:56:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2875 Aug 26 10:14:25 03koen 07org.oe.dev * r0db5a631... 10/ (7 files in 2 dirs): gimp: add 2.3.19, unify old recipes Aug 26 11:02:41 koen, ping Aug 26 11:05:10 Crofton|home: pong Aug 26 11:05:35 hey, I have a bb file that needs at least libtool-1.5.24 Aug 26 11:05:50 is there a way to bb this, or do I just use a comment Aug 26 11:06:12 * Cyberdeck wonders how Crofton|home and koen have setup their routing tables, because he receives the ICMP stuff too ;) Aug 26 11:06:21 http://www.maniacchallenge.org/index.html Aug 26 11:06:25 :) Aug 26 11:06:40 use a comment Aug 26 11:06:43 ok Aug 26 11:06:59 DEPENDS = "libtool (>= 1.5.24)" is valid, but does nothing (yet) Aug 26 11:07:04 I am trying to setup up something small to run the MANIAC Challenge SW Aug 26 11:07:18 I'll stick with a comment for now Aug 26 11:07:44 I also have a hacked up olsr I need to add Aug 26 11:07:48 hiya Aug 26 11:07:57 is olsr_maniac an OK name? Aug 26 11:08:19 I will set def_pref = "-1" to avoid clashing with the 0.3 we already have Aug 26 11:10:23 whats news folks? Aug 26 11:10:41 catching up on recreational OE hacking Aug 26 11:11:52 me too! Aug 26 11:12:03 just about to slip into a little hot cairo/rsvg love in fact .. Aug 26 11:20:20 Crofton: no underscores in PV Aug 26 11:20:23 ehm Aug 26 11:20:25 PN Aug 26 11:21:45 I was thinking of using maniac is PV :) Aug 26 11:22:05 after coming to my senses, I will use olsr-maniac_0.1 or something like that Aug 26 11:22:25 ~lart people who released hacked up things Aug 26 11:22:25 * ibot raises middle finger to people who released hacked up things Aug 26 11:50:47 03crofton 07org.oe.dev * rae2a216b... 10/ (3 files in 3 dirs): libnetfilter_queue : Add libnetfilter_queue to OE. Needs libtool 1.5.24 to build. Aug 26 11:51:01 03koen 07org.oe.dev * r71490693... 10/ (1 packages/sane-backends/sane-backends_1.0.17.bb): sane-backends: fix packaging and staging Aug 26 11:51:12 03koen 07org.oe.dev * rd0274d4f... 10/ (4 files in 3 dirs): sane-frontends: add 1.0.13 Aug 26 11:52:16 does libtool .24 have our rpath patch? Aug 26 11:52:42 probably Aug 26 11:52:58 it builds for everything I try, except Efika Aug 26 11:53:11 I would have to double check patches Aug 26 11:53:19 tried angstrom-2008 for efika? Aug 26 11:53:24 yeah Aug 26 11:53:27 no joy :( Aug 26 11:54:03 this hacked olrsd is annoying Aug 26 11:54:21 what is olsrd anyway? Aug 26 11:54:30 www.olsr.org Aug 26 11:54:40 ad-hoc network routing Aug 26 11:55:11 cool stuff Aug 26 12:05:24 hmmm Aug 26 12:05:35 as with all cool stuff, it has a broken buildsystem Aug 26 12:06:50 right Aug 26 12:07:06 there are multiple implemenatations Aug 26 12:07:32 the one from olsr.org looks sane Aug 26 12:07:37 the one I have suxor Aug 26 12:07:59 I will complain to the guy who manages the guys responsible for it next weekend :) Aug 26 12:09:39 03koen 07org.oe.dev * r046d6d94... 10/ (4 files in 3 dirs): olsrd: update to 0.5.3 Aug 26 12:09:52 there we go Aug 26 12:23:44 koen, we also have a olsr directory with an older version Aug 26 12:24:13 I noticed Aug 26 12:24:24 and ingored it :) Aug 26 12:24:28 I am going to put the maniac specific version in olsrd Aug 26 12:24:32 yeah me too Aug 26 12:24:47 it will really annoy you wrt to lack of build system :) Aug 26 12:42:40 hey zecke Aug 26 13:12:09 koen, Crofton|home: Please do start adding versioned DEPENDS. Whilst they don't yet do anything the syntax is now valid and should be supported Aug 26 13:12:36 (supported as in the syntax shouldn't berak anything, if it does we need to find out and fix it) Aug 26 13:32:33 what is the syntac again? Aug 26 13:33:01 Crofton|home: likely the debian/ipk one Aug 26 13:33:07 DEPENDS = "libtool (>= 1.5.24)" ? Aug 26 13:33:15 I assume that Aug 26 13:35:35 the space is important :) Aug 26 13:44:32 technically the depend is on libtool-cross also Aug 26 14:01:45 03crofton 07org.oe.dev * r641c0ed5... 10/ (1 packages/libnetfilter-queue/libnetfilter-queue_0.0.13.bb): libnetfilter-queue : Add versioned dependency on libtool-cross Aug 26 14:32:26 hello oe , can change kernel code in the work directory and rebuild or are patches a must Aug 26 14:33:19 basically, patches are the "easiest" way Aug 26 14:33:54 Unless there is some guru command to recompile withotu unpacking new source Aug 26 14:33:56 so then just a a patch in the work directory or is a bb file needed ? Aug 26 14:34:10 bitbake needs to read a bb file Aug 26 14:34:28 now, you can set your path so that the staging cross compilers are in your path Aug 26 14:34:40 that should work for kernel hacking Aug 26 14:35:28 i just need a small change to redboot.c Aug 26 14:35:56 so 1st step is to make a bb file ? for that patch ? Aug 26 14:36:33 When I have a complete built of an image for one device and want to build an image for a different device, changing the machine type. Can I then reuse the old finished packages from the first device again when I rechange machine type? Aug 26 14:36:58 dcordes, depends on how similar the machines are Aug 26 14:37:03 is kernel hacking for oe setout by example anywhere ? Aug 26 14:37:20 I know I can resue a lot between OSK, davinci and gusmtic Aug 26 14:37:24 interfaith, not really Aug 26 14:37:26 Crofton|home: I have things setup for angstrom akita and now want to build an image for angstrom htc-wizard Aug 26 14:37:34 htc-wizard is omap850 Aug 26 14:37:53 look in the machine files, that should give you some ideas Aug 26 14:38:11 yes thx..i will try and replicate Aug 26 14:38:49 let me rephrase that- are the packages from the akita build overwritten when I start building for htc? Aug 26 14:39:08 perhaps some WAIT command could be made ..for a hacker to change code after kernel is unpacked in work dir Aug 26 14:39:30 ha ha Aug 26 14:41:26 or hack the .tar.bz2 file .. there must be some notes on how to change the kernel code Aug 26 14:43:33 i think, that kernel is just a package so you can do bitbake -c compile kernel again whenever you want Aug 26 14:44:41 so if i change kernel code in work dir then , bitbake -c compile kernel, i will get the change ? Aug 26 14:57:31 * koen finishes debuging the winding mechanism in his camera Aug 26 14:59:22 koen: will bitbake -c compile kernel/virtual ->build changes made to kernel code i work dir ? Aug 26 14:59:43 bitbake -b path/to/kernel/recipe.bb -c compile -f Aug 26 14:59:54 nice trick ! thx Aug 26 15:00:08 someone needs to do a web page on guru mode :) Aug 26 15:00:13 hey i installed midori but i don't see no menu item for it .. is there one in the pkg? Aug 26 15:00:48 torpor: midori doesn't have a .desktop file Aug 26 15:01:03 ah i see.. terminal only eh? Aug 26 15:01:57 I suspect the developer has it registered as 'default browser', so gnome launches it when he clicks on a link Aug 26 15:05:52 koen: there are no .bb files in ixp4xx-kernel , just defconf file Aug 26 15:07:19 hm, why 'bitbake -c compile virtual/kernel' won't work ? Aug 26 15:07:47 i will try it now Aug 26 15:08:10 kernel/virtual ? Aug 26 15:09:41 yes it does compile.. so i will make a change and try Aug 26 15:10:47 there's PROVIDES += "virtual/kernel" in kernel.bbclass Aug 26 15:11:24 AFAIK -f only works with -b Aug 26 15:11:39 http://rafb.net/p/v6LMNM42.html Aug 26 15:11:47 from my stab at autobuilding Aug 26 15:12:19 script - fb.net/p/UVo0II73.html Aug 26 15:13:42 and local.conf http://rafb.net/p/OUEqTa97.html Aug 26 15:15:11 Crofton: TINDER_HOST = "localhost" -> TINDER_HOST = "ewi546.ewi.utwente.nl" Aug 26 15:15:58 heh Aug 26 15:16:09 i made the change , then did the recompile , no tasks rerun it says Aug 26 15:16:09 I guess we know where you do autobuilds .... Aug 26 15:17:06 there must be a way to recompile a change in kernel code ? Aug 26 15:17:23 yes, delete the compile stamp Aug 26 15:17:30 oh.. ok Aug 26 15:17:33 no, stay away from stamps Aug 26 15:17:40 or use -b -f as koen suggested Aug 26 15:17:42 bitbake -b path/to/kernel/recipe.bb -c compile -f Aug 26 15:17:52 there is no .bb file in ixp-kernel dir Aug 26 15:18:05 people messing with stamps void their right to support Aug 26 15:18:05 just the default config files Aug 26 15:18:07 interfaith, there must have been a bb fiel somewhere that ran Aug 26 15:18:09 find it Aug 26 15:18:42 look in machine dir in tmp/work for ideas what bb file ran Aug 26 15:18:48 interfaith, yes Aug 26 15:18:51 y Aug 26 15:18:54 the bb file knows to look there Aug 26 15:19:13 what directory is above the ixp-kernel dir? Aug 26 15:19:36 03crofton 07org.oe.dev * r73ca8061... 10/ (1 MAINTAINERS): MAINTAINERS : Update entry for Philip Balister Aug 26 15:19:38 linux Aug 26 15:19:41 03crofton 07org.oe.dev * r5987b92f... 10/ (1 MAINTAINERS): MAINTAINERS : Another update to Philip Balister. Aug 26 15:19:43 so Aug 26 15:19:46 then ixp4xx-kernel Aug 26 15:19:54 there are bb files there Aug 26 15:19:58 in linux Aug 26 15:20:04 yes many Aug 26 15:20:08 yes Aug 26 15:20:32 i see the ixp4xx-kernel.inc file Aug 26 15:20:36 one of those is run, and it gets the defconfig from the dir that matches the machine name Aug 26 15:20:45 several ixp4xx bb files Aug 26 15:20:57 ok so just run that bb file.. will do Aug 26 15:21:00 look in tmp/work Aug 26 15:21:17 this might geive you an idea of the version Aug 26 15:21:23 you will need to poke around Aug 26 15:21:27 there is the ixp4xx-kernel_2.6.21.6.bb Aug 26 15:21:34 yes i know what version is used Aug 26 15:21:45 so just re-run that bb file Aug 26 15:21:57 yes, with the options mentioned Aug 26 15:22:03 will do Aug 26 15:22:13 have you found the source bitbale unpacked? Aug 26 15:23:00 koen, this looks better, I think I need to go back and turn off parallel make though Aug 26 15:23:27 yes it seems to be working ! Aug 26 15:24:44 that's GREAT is sure beats patch work ! Aug 26 15:26:14 yes one file redboot.c in the work dir was changed to use the P30 flash chip Aug 26 15:27:40 will -D or -v show the files that are re-compiled ? Aug 26 15:27:54 -D Aug 26 15:28:30 PREFERRED_PROVIDERS += "virtual/${TARGET_PREFIX}depmod:module-init-tools-cross" Aug 26 15:28:35 this line is no longer needed? Aug 26 15:29:37 I have a ton of DSP boards on my desk and no easy way to use them Aug 26 15:29:47 * koen curses TI, atmel and ADI Aug 26 15:30:50 these people are like drug dealers Aug 26 15:30:58 giving away free samples Aug 26 15:31:01 well Aug 26 15:31:04 maybe not Aug 26 15:31:20 only the atmel board was a free smaple Aug 26 15:31:51 hmm Aug 26 15:31:51 yeah Aug 26 15:31:58 I'm out of snapple as well Aug 26 15:37:01 koen, is there a way to tag any bugs reports I generate from the autobuild with build inf? Aug 26 15:37:16 like me using libtool-cross-1.5.24? Aug 26 15:38:52 yes Aug 26 15:39:21 sepukku.bbclass, line 325 Aug 26 15:40:51 I suppose we should create a way to configure that from local.conf .... Aug 26 15:41:04 my current build has no overrides :) Aug 26 15:41:05 yes Aug 26 15:41:12 put python makes my eyes bleed Aug 26 15:41:33 I am glad there are productive python people here Aug 26 15:41:50 the ones at school make me worried python make syou lazy Aug 26 15:42:06 koen: such as? Aug 26 15:42:14 <-- used to do DSP-related work for a synth company Aug 26 15:42:24 am currently playing with some nice toys from freescale .. Aug 26 15:42:45 so many toys, so little time Aug 26 15:46:51 morning Aug 26 15:50:30 thx all that did buld in the changes to redboot, one step up :) Aug 26 15:52:29 has anyone gotten bluetooth fully worked out yet so that it can be used instead of USB (i.e. networking etc)? Aug 26 15:53:40 torpor: man pand Aug 26 15:55:10 03koen 07org.oe.dev * r031c7752... 10/ (1 packages/bluez/bluez-utils.inc): bluez-utils: depend on gstreamer for headset stuff Aug 26 16:14:17 koen: i wanna gui for it. Aug 26 16:14:39 poke bmidgley or holtmann Aug 26 16:14:56 there's a good chance marcel will add it to bluez-gnome Aug 26 16:15:03 ok Aug 26 16:15:05 good to know Aug 26 16:15:13 when initramfs-image is made, then that is the cpio image to use in .config ? Aug 26 16:15:24 cbrake: any chance you are logged in? Aug 26 16:15:33 would be so sweet to have a working gui i could use to bluetooth myself onto the internet .. Aug 26 16:15:51 seeking a simple boot kernel that has all it needs Aug 26 16:16:09 torpor: but PAN is pretty easy to set up from the cmdline Aug 26 16:16:19 (it was 2 years ago) Aug 26 16:17:00 oh i see Aug 26 16:17:04 i'll check it out Aug 26 16:24:56 All - is there a way to build images without relying on modules? I build my kernel statically and when I boot up the angstrom distro, it is looking for a bunch of modules. Aug 26 16:25:07 I see the modules gz...but it is 78M when decompressed Aug 26 16:30:27 03mickeyl 07org.oe.dev * r5780868a... 10/ (1 packages/x11vnc packages/x11vnc/x11vnc_0.9.3.bb): add x11vnc -- exports your X11 session on-the-fly via VNC Aug 26 16:30:43 hi gods Aug 26 16:33:44 * Crofton|home looks around Aug 26 16:44:15 Crofton|home, well its all relative so sometimes that is my view ! ;-) Aug 26 16:49:22 ixp4xx hacker ? Aug 26 16:49:36 mickey|offline: how did your talk go? Aug 26 16:54:18 grr, how do I see the build status in tinderbox again? Aug 26 16:55:11 ok, click on distro Aug 26 16:59:29 hackers: can changes be made to rootfs , then build ? Aug 26 17:01:17 can anyone tell me how to make the rootfs NOT include the kernel/uImage ??? Aug 26 17:02:13 i have seen that in the conf files Aug 26 17:03:14 interfaith: are you saying you saw the uImage in conf files for rootfs or are you talking about something else? Aug 26 17:03:48 i am looking for it now, in the oe examples there is a case where the kernel is removed Aug 26 17:04:34 interfaith: ok. cool. I was trying to see where that was. I didn't find it but guess I wasn't looking at the right spot. Aug 26 17:04:43 see slugos-image.bb Aug 26 17:05:07 interfaith: I was hoping not to have it depend on the modules either. when it loads, I see tons of messages about modules.dep not found etc Aug 26 17:05:08 at the end of the bb file the kernel is removed Aug 26 17:06:10 several other places do this as well, perhaps grep IMAGE_ROOTFS Aug 26 17:07:38 interfaith: cool. so, it basically just deletes the file before making the rootfs. Aug 26 17:45:29 steliosk: are u responsible for the ML403 and ML410 oe stuff? Aug 26 18:09:49 morning Aug 26 18:10:08 koen: when will bluez req half of gnome/kde just to be build? Aug 26 18:17:58 * * OE Bug 1391 has been REOPENED by cwixon(AT)usa.net Aug 26 18:18:00 * *  minimo package bit rotted badly Aug 26 18:18:02 * * http://bugs.openembedded.org/show_bug.cgi?id=1391 Aug 26 18:21:51 hrw: never Aug 26 18:22:45 check http://www.angstrom-distribution.org/repo/?action=details&pnm=bluez-utils Aug 26 18:23:57 26 17:55 < CIA-35> koen org.oe.dev * r031c7752... / (1 packages/bluez/bluez-utils.inc): bluez-utils: depend on gstreamer for headset stuff Aug 26 18:24:19 to build bluez we need to build cups and gstreamer Aug 26 18:24:27 it is getting insane Aug 26 18:40:51 anyone got a moment for a few machine.conf questions Aug 26 18:41:03 go Aug 26 18:41:13 like how are TARGET_ARCH and PACKAGE_EXTRA_ARCHS sued? Aug 26 18:41:17 er used Aug 26 18:41:31 tool chain selection? Aug 26 18:41:46 TARGET_ARCH is architecture (arm/i386/i686/mipsel etc) Aug 26 18:41:47 binary compatiblity info? Aug 26 18:42:01 and that drives tool chain selection? Aug 26 18:42:06 PACKAGE_EXTRA_ARCHS is list of binary compatible archs (i386/i486/i586 like) Aug 26 18:42:37 for OSK we have target = arm and extra = armv5te Aug 26 18:43:20 also what does MACHINE_EXTRA_RRECOMMENDS do? Aug 26 18:44:32 its list of stuff recommended for machine - usually kernel modules Aug 26 18:44:41 yeah Aug 26 18:44:50 Crofton|home: look at conf/documentation.conf Aug 26 18:44:52 so they are very likely to get installed in the image? Aug 26 18:45:04 I did, it is not super clear :) Aug 26 18:45:23 M_E_R is not there :) Aug 26 18:45:43 task-base depends on task-machine-base which recommends MACHINE_EXTRA_RRECOMMENDS and depends on MACHINE_EXTRA_RDEPENDS Aug 26 18:46:42 what is the process for installing from a recomends list on to the image Aug 26 18:46:56 the RDEPENDS go for sure, the recommends confuse me Aug 26 18:47:10 recommends are installed if available and user can remove it Aug 26 18:47:14 on device Aug 26 18:47:21 user cant remove depends stuff Aug 26 18:47:38 ok, so image creation succeeds even if recomends are not available! Aug 26 18:47:43 yes Aug 26 18:47:45 thanks Aug 26 18:49:11 np Aug 26 18:50:07 bye all Aug 26 18:51:18 Crofton|home: see h2200.conf on how PACKAGE_EXTRA_ARCHS should like like on armv5 Aug 26 18:53:00 so an arm5 machine can run arm4 Aug 26 18:53:29 so h2200 runs armvte Aug 26 18:53:47 these are the ARM instruction sets right? Aug 26 18:53:57 armvvN can run armvN-1, N-2, N-3, etc Aug 26 18:54:00 yes Aug 26 18:54:22 (t=thumb, e=dsp, j=java s=no idea) Aug 26 18:54:48 these are the Instruction Set Architectures Aug 26 18:54:50 and in OE arm == armv4 Aug 26 18:54:56 ah Aug 26 18:55:07 eabi needs armv4t Aug 26 18:55:18 since BX is a thumb instruction Aug 26 18:58:54 so I need to update extra archs for OSK :) Aug 26 19:00:20 yes Aug 26 19:03:44 guys, could you suggest me please, which distro has minimal size in order to buld it quikly? Aug 26 19:04:27 yakov: your best bet would be angstrom and angstrom-minimal-image Aug 26 19:04:27 angstrom-minimal-image Aug 26 19:04:39 thanks! Aug 26 19:04:53 angstrom-minimal-image can be used in DISTRO? Aug 26 19:05:36 DISTRO="angstrom-2008.1" Aug 26 19:05:45 IMAGE_TARGET = "angstrom-minimal-image" Aug 26 19:06:14 or DISTRO="angstrom-2007.1" Aug 26 19:08:24 depends on which glibc version you want Aug 26 19:08:36 2007 is 2.5, 2008 is 2.6.1 Aug 26 19:08:45 both use the same uclibc version Aug 26 19:08:54 (although it seems uclibc doesn't like gcc 4.2) Aug 26 19:13:42 koen, BTW which GCC version suits best for 2008.1? Aug 26 19:13:52 it's latest release AFAIU.. Aug 26 19:14:12 koen: so is 2007 still the release candidate or are you switching to 2008? Aug 26 19:31:12 ok i made a simple patch for mtd, does the patch file go next to the default config file ? Aug 26 19:34:44 koen, you saud this line needs to go from machine files? Aug 26 19:34:49 PREFERRED_PROVIDERS += "virtual/${TARGET_PREFIX}depmod:module-init-tools-cross" Aug 26 19:52:04 For your viewing pleasure http://ossie.wireless.vt.edu/~balister/ankle/ Aug 26 19:56:43 :) Aug 26 19:56:48 what Aug 26 19:56:49 the Aug 26 19:56:50 fuck Aug 26 19:57:00 the video? Aug 26 19:57:01 man i thought it was a demo of ossie! sheee-it! Aug 26 19:57:15 it is my ankle :) Aug 26 19:57:30 hahahaha Aug 26 19:57:58 * mwester takes away Crofton|home's remaining beers. Aug 26 19:58:27 my friends and collect our medical incidents on film :) Aug 26 19:59:45 my med student friends watched my wifes knee surgery on the big scren TV Aug 26 20:00:54 * Crofton|home is drinking coffee Aug 26 20:01:01 and trying to write Aug 26 20:04:35 mwester: did you ever get a chance to publish those kexec patches? Aug 26 20:06:21 hvontres1home: No - last week was out of town, still trying to catch up with everything (including rebuilding my main file server that failed in the storm last week). Aug 26 20:07:45 mwester, you are in midwest? Aug 26 20:07:52 mwester: np. btw, did you see the message from Uli Luckas on the OE list? Aug 26 20:09:13 Crofton|home: yes - Chicago suburbs. I was flying back from Denver on Thursday PM, whilst my kids were trying to keep the basement from flooding. The server in the basement got damaged somehow, although it did stay dry. Aug 26 20:09:46 well I'm glad you missed the worst of it Aug 26 20:10:20 hvontres|home: Yes, I've seen that message, and a bunch of other on the arm-kernel list. I need to get the core patch cleaned up and submitted to l-a-k, then we can update the patchset in OE, and from there work on the new UI patch for kexec. Aug 26 20:13:38 Crofton|home: i've got a few pics from my surgery around as well .. i know how it is to be proud of that stuff. Aug 26 20:13:45 but for the most part, everyone is grossed out. ;) Aug 26 20:14:15 as opposed to my friend who remarked he had not seen one of those tools b4 Aug 26 20:14:25 I thinnk he has had both ankles done .... Aug 26 20:16:48 mwester: which suburb? Aug 26 20:43:42 Crofton|home: I found the spot in the connex boot where the tmpfs mount fails Aug 26 20:43:51 Very strange! Aug 26 20:44:01 how so? Aug 26 20:44:19 the kernel does have support for tmpfs? Aug 26 20:44:23 In /etc/init.d/udev this line fails: Aug 26 20:44:37 if ! mount -n -o size=$tmpfs_size,mode=0755 -t tmpfs tmpfs /dev; then Aug 26 20:44:42 hello Aug 26 20:44:46 gm Aug 26 20:44:52 size is not a legal mount option! Aug 26 20:45:01 oe.org warped back in time :-) Aug 26 20:45:10 heh Aug 26 20:45:13 the web site? Aug 26 20:45:21 hmmm Aug 26 20:45:24 hence the error message: Aug 26 20:45:28 ramfs: Bad mount option size Aug 26 20:45:30 I think busybox provides moutn Aug 26 20:46:04 lrwxrwxrwx 1 root root 16 Jan 1 00:00 /bin/mount -> mount.util-linux Aug 26 20:46:27 steve, is your ethernet working? Aug 26 20:46:31 can you look at your moutn? Aug 26 20:46:35 I wish :-) Aug 26 20:46:54 root@gumstix-connex:/dev$ ls -l /bin/mount Aug 26 20:46:54 lrwxrwxrwx 1 root root 7 Jan 1 00:00 /bin/mount -> busybox Aug 26 20:46:59 likewise, are you one of the busybox sucks and why guys? Aug 26 20:47:00 ah Aug 26 20:47:13 I bet the busybox mount does not support size Aug 26 20:47:31 so was does the OSK get the other mount Aug 26 20:48:51 Hmm . . . did a man mount on ubuntu and size isn't an option there either Aug 26 20:50:22 Crofton|home: huh, what kind of guys? Aug 26 20:50:54 on gumstix they have busybox mount Aug 26 20:51:00 03woglinde2 07org.oe.dev * r1ea115ab... 10/ (1 MAINTAINERS): MAINTAINERS: fix spelling of cacao-cldc Aug 26 20:51:05 on omap5912osk, I have mount.util Aug 26 20:51:57 Crofton|home: especially the mount differs a bit, on my targets I use busybox + linux utils mount Aug 26 20:52:04 I speculate, this is because I hae HOTPLUG = "udev" in my machine file .... Aug 26 20:52:14 yes Aug 26 20:52:27 I think that is what happens on the OSK Aug 26 20:52:42 I don't know the exact difference really, I think busybox is catching up on their mount support. I am not sure why I didn't select the busybox mount, but I remember it had a reason. Aug 26 20:52:49 I just do not understand the voodoo that made this happen on the OSK Aug 26 20:53:10 likely because it doesn't support size when mounting tmpfs .... Aug 26 20:53:14 Crofton|home: probably is in the distro Aug 26 20:53:27 we both use angstrom :) Aug 26 20:53:40 target maybe? Aug 26 20:53:52 angstrom_minimal_image? Aug 26 20:53:53 try adding HOTPLUG = "udev" Aug 26 20:53:59 to machine Aug 26 20:54:07 birbake -c clean task-base Aug 26 20:54:09 Ok, will do Aug 26 20:54:22 add the include/gumstix.inc Aug 26 20:57:42 sakoman, here is the result of "ipkg list" Aug 26 20:57:44 http://pastebin.ca/671647 Aug 26 20:57:48 on the osk. Aug 26 20:58:07 this does not come from angstrom-minimal-image, but should be a useful ref Aug 26 21:00:32 Crofton: thanks! Aug 26 21:01:16 root Aug 26 21:01:22 heh Aug 26 21:01:32 don't type the password Aug 26 21:01:48 :-) attempting to log in the wrong window :-) Aug 26 21:02:26 everyone should google their passwrods ocasionally Aug 26 21:02:39 :-) Aug 26 21:02:48 You change it every month, right? Aug 26 21:06:26 of course! Aug 26 21:06:57 :) Aug 26 21:07:12 Just googled my current password, no hits! Aug 26 21:07:28 I'm 99.9% sure that HOTPLUG="udev" is bogus Aug 26 21:07:55 Don't tell me that after I started the build with it :-) Aug 26 21:08:11 hmmm Aug 26 21:08:38 well, how do we get the util-linux mount on the gumstix? Aug 26 21:09:12 gusmtix has busybox mount, which does not support size= option Aug 26 21:09:32 Powerbook-2:~/Projects/OpenEmbedded/org.openembedded.dev/conf/distro koen$ grep -rn mount * Aug 26 21:09:32 angstrom-2007.1.conf:406: util-linux-mount util-linux-umount \ Aug 26 21:10:05 lrwxrwxrwx 1 root root 7 Jan 1 00:00 /bin/mount -> busybox Aug 26 21:10:41 so any image that honours DISTRO_EXTRA_RDEPENDS will have it Aug 26 21:11:06 he is buils angstrom-minimal-image Aug 26 21:12:24 I am pretty sure he does have udev, from a boot log I saw .... Aug 26 21:12:45 Crofton: yes udev is there Aug 26 21:13:11 * Crofton|home es a note to drop teh udev line from omap5912osk Aug 26 21:13:11 the udev init script is where mount fails Aug 26 21:13:38 can you try ipkg list | grep mount Aug 26 21:14:59 Crofton|home: http://rafb.net/p/ewLLOH23.html Aug 26 21:15:00 Hmm . . . no ipkg in the build Aug 26 21:15:06 mine has ipkg Aug 26 21:15:16 it's showing util-linux-mount Aug 26 21:15:25 2.12r-r12 Aug 26 21:15:27 doh Aug 26 21:15:40 asmola and I are running ossie-image ..? Aug 26 21:15:54 hmmm Aug 26 21:16:10 but i still get the file system errors when i boot, even with this image Aug 26 21:16:16 this begs the question how the correct mount gets into ossie image Aug 26 21:17:57 so task-base contains the util-linux-mount? Aug 26 21:18:03 but not task-boot Aug 26 21:18:24 asmola, can you post ls -l /bin/mount? Aug 26 21:18:36 thats what i was just doing before you asked Aug 26 21:18:37 you have all manner of issues with that image ... Aug 26 21:18:41 asmola: does your boot message log: ramfs: Bad mount option size Aug 26 21:18:52 basically, busybox, or util-linux Aug 26 21:19:01 util-linux Aug 26 21:19:09 but i am still getting the ramfs: bad mount option size Aug 26 21:19:39 * Crofton reboots the osk Aug 26 21:19:51 I'm confused, cause when I look at the docs for standard linux mount size is not a legal option! Aug 26 21:20:18 INIT: /dev/initctl is not a fifo Aug 26 21:20:24 is the only interesting thing I see Aug 26 21:21:05 it is on mine Aug 26 21:21:14 you need to look under options for tmpfs Aug 26 21:21:22 ah, ok Aug 26 21:21:28 my bad Aug 26 21:21:53 np, I like it when it looks like I know what I am doing .... Aug 26 21:22:01 :-) Aug 26 21:22:06 crap, build failed on ossie-image Aug 26 21:22:14 how? Aug 26 21:22:29 in python-native Aug 26 21:22:35 ill pastebin in a sec Aug 26 21:22:37 urg Aug 26 21:23:14 i'm signing off - too tired Aug 26 21:23:17 cu all tomorrow Aug 26 21:23:17 gn Aug 26 21:23:34 gn for later Crofton|home Aug 26 21:23:56 http://pastebin.com/m54b59bd2 Aug 26 21:24:57 hmm, maybe i am just doing something stupid here Aug 26 21:25:16 prior to this i built angstrom and then did the osie image build Aug 26 21:25:18 that is so familair Aug 26 21:25:22 why? Aug 26 21:25:31 well just casue it was already built Aug 26 21:26:02 Should I move to ossie-image too, so we are all on the same image? Aug 26 21:26:09 sakoman, no Aug 26 21:26:13 OK! Aug 26 21:26:26 that is my double secret (well not really) SDR image Aug 26 21:26:44 Promise, I won't peak Aug 26 21:26:45 uh oh, is my life endanger for exposing the secret? Aug 26 21:26:48 I'll have an angstrom-minimal-image fopr OSK later Aug 26 21:26:51 no Aug 26 21:26:59 it is on a public svn Aug 26 21:27:02 :) Aug 26 21:27:37 ok, so it look slike I (or koen) need to fix angstrom-minmal-image Aug 26 21:27:58 koen, should I fix and push per your patch? Aug 26 21:29:43 well, I am going to push it, since koen likely fell asleep in his controls book ... Aug 26 21:29:52 :) Aug 26 21:32:15 sakoman, mtn pull, mtn update Aug 26 21:32:24 bitbake -c clean task-boot Aug 26 21:32:35 bitbake angstrom-minimal-iamge Aug 26 21:32:41 Build with koen's patch already in progress Aug 26 21:41:22 Build is done, sanity check shows proper mount in build/tmp/rootfs/bin :-) Aug 26 21:41:56 hopefully this improves things ... Aug 26 21:42:11 reflashing . . . Aug 26 21:46:19 re the ossie build dying, i just did an mtn pull \ update since i hadn't since thurs night or fri morning, lookslike it might be ok now Aug 26 21:46:56 that may be why the poblem seemed familiar Aug 26 21:47:07 mickey had done some python work Aug 26 21:48:01 I'm working on getting a script together to do auto builds on machines I care about Aug 26 21:48:18 it also reports problems straight tp bugzilla Aug 26 21:48:22 oh slick Aug 26 21:57:24 still ixp4xx hack how to change default config settings ? Aug 26 21:58:25 hey ppl Aug 26 22:03:09 Crofton, asmola: Sigh, not there yet -- http://www.sakoman.net/oe/boot.txt Aug 26 22:05:08 03crofton 07org.oe.dev * r6b0ced9a... 10/ (1 packages/images/angstrom-minimal-image.bb): angstrom-minimal-image : Use mount from util-linux. Needed to moutn tmpfs. Aug 26 22:07:33 mount still doesn't like size, but the util version spits out a more verbose complaint :-) Aug 26 22:09:47 ah Aug 26 22:09:51 interesting Aug 26 22:10:51 I'll build an OSK version of minimal image tongiht and flash in the AM Aug 26 22:11:25 Will be interesting to see what you get Aug 26 22:11:37 yes Aug 26 22:11:52 you can rty editing the scripts to see if you can make it work Aug 26 22:12:09 then we can try to fix the problem at the source Aug 26 22:12:12 That's my next step! Aug 26 22:12:21 great minds thik a like Aug 26 22:12:33 just can't spell Aug 26 22:12:34 Wonder what the default is if you don't specify size Aug 26 22:12:42 half of ram ..... Aug 26 22:12:50 from the man page I saw Aug 26 22:12:59 Worth a try for starters . . . Aug 26 22:13:03 yeah Aug 26 22:17:04 Crofton: With default size: http://www.sakoman.net/oe/boot2.txt Aug 26 22:17:39 what does mount look like? Aug 26 22:17:40 Only one complaint: about /sysfs Aug 26 22:17:47 yeah Aug 26 22:18:10 same on osk Aug 26 22:19:30 pasted mount output onto end of previous link, hit refresh Aug 26 22:20:17 mine looks about the same Aug 26 22:20:38 so does your software use autotools? Aug 26 22:20:48 So we just need to figure out why either mount or the script is broken :-) Aug 26 22:20:53 yeah Aug 26 22:21:03 Which software? Aug 26 22:21:04 hopefully, I can reprodiuce here in the mornign Aug 26 22:21:17 that you want to run on the gumstix? Aug 26 22:21:23 No Aug 26 22:21:59 hopefully it is not super complex build method then :) Aug 26 22:22:28 We'll see! Aug 26 22:22:46 what is the module for usb mass storage? Aug 26 22:22:47 sane autotools stuff is easy to build in OE Aug 26 22:23:00 since we can abuse autotools to make it fins all teh libs Aug 26 22:23:07 Seems like I'm just about at the point I can begin to port my stuff over from buildroot Aug 26 22:23:10 man my typing has become poor Aug 26 22:23:14 yeah Aug 26 22:23:51 I checked an most of the stuff I need package-wise is already in OE Aug 26 22:23:58 cool Aug 26 22:24:04 And hopefully builds :-) Aug 26 22:24:04 are you missing anything? Aug 26 22:24:16 microperl perhaps Aug 26 22:24:19 let me know if something does not, I can take a look at it Aug 26 22:24:41 need gd, gnuplot Aug 26 22:25:25 what versions? Aug 26 22:25:36 I see gd-2.0.33 Aug 26 22:25:43 not critical (I don't think :-) Aug 26 22:25:56 gnuplot 4.2 Aug 26 22:26:13 the problem with ahving around 1000+ packages is bitrot :) Aug 26 22:26:34 currently using gd 2.0.33 and gnuplot 4.0.0 Aug 26 22:26:54 there is that version of gnuplt also :) Aug 26 22:26:58 So buildroot has more bitrot :-) Aug 26 22:27:40 as this internet thing grows, bitrot gets worse Aug 26 22:27:58 I ahve been working with OE for a couple of years Aug 26 22:28:21 we are definitely gettign leary of people that appear and want to add their favorite package Aug 26 22:28:41 we need people who will stick around and pay attention :_) Aug 26 22:29:11 looking at other stuff I use: freetype 2.1.9, mDNSResponder 107.5, msmtp-1.4.11, ntp, boa Aug 26 22:29:44 Crofton|home: can somebody comit my latest rt73 bb? :D Aug 26 22:30:00 where is it? Aug 26 22:30:04 and what is it? Aug 26 22:30:11 rt73 is wifi chipset Aug 26 22:30:19 I do not like to commit things I do not understand Aug 26 22:30:34 inside /packages/rt2x00/rt73* Aug 26 22:30:39 is version 1.0 Aug 26 22:30:44 I would like to comit version 2.0 Aug 26 22:31:05 can you put a mtn diff in bugs.openembedded.org? Aug 26 22:31:24 I am not familiar with wifi chipsets Aug 26 22:31:26 I know what a diff is but what is an mtn diff? Aug 26 22:31:39 basically only the SRC_URI changed Aug 26 22:31:47 you have added it locally right? Aug 26 22:32:09 did you modify a file, or add a new one? Aug 26 22:32:11 I did add it nowhere- just made it a nd gave it to Zero_Chaos to test it Aug 26 22:32:18 I modified 1.0 Aug 26 22:32:26 as I said- only SRC_URI changed Aug 26 22:32:29 ok Aug 26 22:32:49 remind me tomorrow, I have been working on thesis all day and am pretty brian dead Aug 26 22:33:04 and I am not familiar with wifi stuff Aug 26 22:33:21 okay thanks :) what is thesis? Aug 26 22:34:18 basically embeded SDR stuff Aug 26 22:34:33 what's SDR :X Aug 26 22:34:38 basically, I am lowering my standards to get teh paper :) Aug 26 22:34:44 software defined radio Aug 26 22:34:48 oh ok Aug 26 22:35:03 is it any related to gnu radio? Aug 26 22:35:54 yes and no Aug 26 22:36:05 gnu radio is sdr, but I am not working with gnu radio Aug 26 22:36:18 although, I did help write teh bb file for gnuradio in OE Aug 26 22:36:48 The GNU radio guys rock! Aug 26 22:36:59 I would like to experiment with this radio stuff but then I saw devices supporting it cost lots of $s Aug 26 22:37:38 the USRP is afairly "cheap" Aug 26 22:37:43 for the field Aug 26 22:37:50 ....found my connex... Aug 26 22:38:37 I am going to pretend to mow my lawn Aug 26 22:38:46 I need to escape from the computer Aug 26 22:38:58 no! resist! Aug 26 22:39:30 I need to move around Aug 26 22:39:32 I think I am going to *actually* mow my lawn Aug 26 22:39:41 But before I do Aug 26 22:40:01 yes? Aug 26 22:40:24 Crofton, can you suggest an image that contains a bit more than minimal-boot? Aug 26 22:40:35 angstrom-console-image :) Aug 26 22:40:48 will you have a display? Aug 26 22:40:50 OK, I'll give it a try after the lawn Aug 26 22:40:52 in the end? Aug 26 22:40:59 No, no display Aug 26 22:41:04 good Aug 26 22:41:05 Just a web UI Aug 26 22:41:09 ah Aug 26 22:41:11 cool Aug 26 22:41:24 l8r Aug 26 22:41:29 l8r! Aug 27 00:56:43 all: anyone clued in on how to add a kernel patch for ixp4xx Aug 27 01:00:22 interfaith: at the moment, it is pulled from the nslu2-linux svn repo Aug 27 01:00:38 so we can do one of three things: Aug 27 01:00:47 yes thx i need to patch redboot.c Aug 27 01:00:56 where can it go ? Aug 27 01:01:04 what's the patch? Aug 27 01:01:26 a p30 fix simple few lines of code to find the partions Aug 27 01:01:39 bbiab Aug 27 01:02:00 shall i cat the patch here ? Aug 27 01:02:04 pastebin it for me Aug 27 01:02:06 ~pastebin Aug 27 01:02:08 somebody said pastebin was a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Aug 27 01:02:26 ok it is just a few lines Aug 27 01:04:36 --- mtd/redboot.c 2007-07-06 23:47:55.000000000 -0500 Aug 27 01:04:36 +++ mtdx/redboot.c 2007-08-26 09:49:58.000000000 -0500 Aug 27 01:04:36 @@ -64,17 +64,24 @@ Aug 27 01:04:36 if (!buf) Aug 27 01:04:36 return -ENOMEM; Aug 27 01:04:37 Aug 27 01:04:39 - if ( directory < 0 ) Aug 27 01:04:41 - offset = master->size + directory*master->erasesize; Aug 27 01:04:43 - else Aug 27 01:04:45 - offset = directory*master->erasesize; Aug 27 01:04:47 + //if ( directory < 0 ) Aug 27 01:04:49 + // offset = master->size + directory*master->erasesize; Aug 27 01:04:51 + //else Aug 27 01:04:53 + //offset = directory*master->erasesize; Aug 27 01:04:55 + Aug 27 01:04:57 + offset =0; Aug 27 01:05:01 Aug 27 01:05:03 printk(KERN_NOTICE "Searching for RedBoot partition table in %s at offset 0x%lx\n", Aug 27 01:05:05 master->name, offset); Aug 27 01:05:07 Aug 27 01:05:09 - ret = master->read(master, offset, Aug 27 01:05:11 + //ret = master->read(master, offset, Aug 27 01:05:13 + // master->erasesize, &retlen, (void *)buf); Aug 27 01:05:15 + Aug 27 01:05:17 + ret = master->read(master, master->size - (2*master->erasesize), Aug 27 01:05:19 master->erasesize, &retlen, (void *)buf); Aug 27 01:05:21 Aug 27 01:05:23 + Aug 27 01:05:25 + Aug 27 01:05:27 if (ret) Aug 27 01:05:31 goto out; Aug 27 01:05:33 Aug 27 01:05:35 offset is 0 Aug 27 01:05:37 and ret is set to the end of the chip Aug 27 01:05:39 this is the way i setup the p30 Aug 27 01:05:41 anyways that works fine Aug 27 01:05:55 my problem is how to add patches and change cmdline etc Aug 27 01:06:49 rwhitby: the kernel boots OK and loads redboot ok, but cannot load a ramdisk or load a root disk yet Aug 27 01:07:55 rwhitby: the redboot works fine for 2.4 kernel.. mv..now i want to escape from mv and go with oe and 2.6 for the ixdpg425 model of hw Aug 27 01:08:11 interfaith: in a meeting now, back in 30 mins Aug 27 01:08:17 fine Aug 27 01:21:47 interfaith: changing redboot.c is usually not the correct way to do something. Aug 27 01:30:44 ok well the mtd driver has been working fine for us for a year or so Aug 27 01:31:33 rwhitby: at this point i could move out of redboot and go to uboot Aug 27 01:32:02 rhwitby: my goal is to get an OE solution with or without redboot Aug 27 01:33:20 rhwitby: so where can patches be placed ? or must one hack and force a compile Aug 27 01:53:06 interfaith: I'm still in that meeting Aug 27 01:54:19 thx Aug 27 02:03:34 rwhitby:as i have all the 2.4 setup code, perhaps that can be used for 2.6 Aug 27 02:12:30 interfaith: back now Aug 27 02:13:28 interfaith: so the deal with ixp4xx-kernel patches is that we only put stuff in that we think has a chance of being accepted upstream, and we have a common ixp4xx kernel which supports all the various ixp4xx machines that OE supports. Aug 27 02:13:59 can you explain what your redboot patch does, and whether it would be accepted in the mainstream linux kernel or not? Aug 27 02:14:14 ok Aug 27 02:14:26 the p30 is the larger flash chip 32meg Aug 27 02:15:05 the change in redboot locates that needed partition info Aug 27 02:15:42 i have no problem with another solution for the p30 Aug 27 02:16:16 so where is the redboot partition table located in this larger chip? Aug 27 02:16:22 i think intel suggested or stated some details for the p30 that required that redboot code Aug 27 02:16:36 bbiab (another call) Aug 27 02:16:41 ok Aug 27 02:17:58 yes.the partition table is located at the top end of the chip..via the intel datasheet i recall Aug 27 02:18:19 anyway.. redboot is working fine it boots and finds all the partitions ok Aug 27 02:18:45 the problem is how to patch the code Aug 27 02:18:55 let me show you Aug 27 02:21:11 i was thinking the new mtd driver had new code for the p30, as they do include the 2nd chip select Aug 27 02:21:41 however the mtd code as is does not work with the p30 here **** ENDING LOGGING AT Mon Aug 27 02:59:57 2007