**** BEGIN LOGGING AT Thu Jun 16 02:59:57 2011 Jun 16 08:48:32 zumbi, to answer yesterdays question, imx51 was just removed in ubuntu in favor of mx5, omap4 is one of our supported arches so needs to stay Jun 16 08:50:22 IMAGES !!!! Jun 16 08:50:24 http://cdimage.ubuntu.com/daily-preinstalled/20110616/ Jun 16 08:50:31 yipiee Jun 16 08:51:22 hmm, nearly 1G Jun 16 09:06:16 ogra_: what's the difference between omap and omap4? Jun 16 09:06:32 in the cdimage build Jun 16 09:06:35 omap4 has a special kernel package Jun 16 09:06:46 and special boot options Jun 16 09:07:13 for omap4 we also provide additional packages that make it easy to enable the TI Jun 16 09:07:15 PPA Jun 16 09:07:30 which holds the sgx drivers and HD video codecs etc Jun 16 09:08:15 I see.. thanks Jun 16 09:08:48 the prob is, while you can have a jiont kernel packqage for all omap arches, you still need a separate binary bootloaders and cmdline options atm Jun 16 09:08:56 -s Jun 16 09:10:12 ogra_: part of that can probably be solved with boot scripts Jun 16 09:10:20 not really Jun 16 09:11:03 TI initializes half the HW in their bootloaders instead of the kernel ... Jun 16 09:11:40 if you would have x-loader and u-boot binaries that support all omap arches and only switch through boot scripts, you would end up with huge binaries Jun 16 09:12:08 and i even doubt a multiarch x-loader would be possible at all Jun 16 09:14:39 i fear until UEFI gets default for ARM we will have to live with SoC specific bootloader binaries here Jun 16 09:14:56 yes, x-loader upstream maintainer was a bit sick of all that Jun 16 09:16:14 Why doesn't TI just boot directly into kernel and this one is responsible to do everything Jun 16 09:16:22 no idea Jun 16 09:16:44 don't really need u-boot, unless you rely on it to initialize everything Jun 16 09:16:49 for the build cluster we tried to use a raw kernel boot ... and initially ended up with no MMC and no USB Jun 16 09:17:17 getting that initialized needed a good bunch of first stage bootloader hacking Jun 16 09:17:53 sadly you need x-loader, there is not enough memory to carry u-boot Jun 16 09:18:08 so x-loader needs to init the ram first Jun 16 09:18:29 now, everything starts to make sense Jun 16 09:18:33 sadly it grew more heads over the years it seems Jun 16 09:18:57 as well as u-boot did Jun 16 09:19:34 The processor only has 64K memory on boot, that's why x-loader is the way it is. It initializes ram (and a bunch of other crap), then u-boot can load and do more. Jun 16 09:19:48 so even if you wanted to do something like booting a raw minimal kernel and then kexec into something full, you would still need x-loader and u-boot Jun 16 09:20:06 Well, x-loader at least. Jun 16 09:20:15 which makes kexec moot since it just adds boot time for no benefit Jun 16 09:20:20 You can boot a kernel from xloader. Jun 16 09:21:03 well, you should be able to dd a binary header in front of uImage that replaces x-loader Jun 16 09:21:16 ogra, would an xloader/uboot that supports OMAP3 and 4 be that huge? Even if it were double the current size - which it probably would not be - it's not an issue Jun 16 09:21:20 at least with newer omaps Jun 16 09:21:26 Can't. not enough memory to load both. Jun 16 09:21:47 janimo, for x-loader it is ... as GrueMaster said above, 64k is your limit Jun 16 09:22:26 not having both binaries but one that handles both. Is the extra code/data so much different it would increase the size beyond what it fits? Jun 16 09:22:28 for u-boot it might be possible to be a big multisoc binary Jun 16 09:22:38 but surely for the cost of bootspeed Jun 16 09:24:26 is it true that TI had problems with memory controllers as those did not match, so omap was unable to use memory (as LPDDR) Jun 16 09:24:39 on XM, yes Jun 16 09:24:54 (beagle XM omap3) Jun 16 09:24:56 yes, I heard on 3530 early revs Jun 16 09:25:22 all this is just insane.. :) Jun 16 09:25:30 XM used 36/37 iirc from start - or not? Jun 16 09:25:40 34/35 was on normal beagleboard Jun 16 09:25:42 the wonderful world of ARM :) Jun 16 09:26:21 I hope for cheap ARM board with 2MB of flash on board Jun 16 09:26:43 so there will be space for bootloader so boot from usb/sd/sata/network will be easy Jun 16 09:26:44 i think you will see flash on board less and less Jun 16 09:26:57 eMMC costs only half of i t Jun 16 09:27:16 ogra_: flash/emmc/esomething/whatever Jun 16 09:27:39 a way to just connect power and get board booted Jun 16 09:27:45 hrw: sounds about right, xM using DM3730 Jun 16 09:28:25 ogra_: if you want to use sdio on panda you have to solder second sdmmc or use other computer to send bootloader over usbdevice Jun 16 09:28:42 there is "Automotive eMMC" much more reliable/tested than current eMMC Jun 16 09:28:52 I miss real developer boards Jun 16 09:30:36 I need to play with TI Wilink module, I hope it performs as expected Jun 16 10:18:40 oh, there goes tomboy Jun 16 10:19:07 Good morning Jun 16 10:24:03 A little question, if I may. I cross compiled kernel for Beagle (that is, for arm). In parent directory (relative to kernel sources) I have now 5 .dev biles and a lot of .udeb files. What are .udeb files? Is it sufficient if I install linux-image-2.6.38-10-omap_2.6.38-10.44_armel.deb and linux-libc-dev_2.6.38-10.44_armel.deb? What are -versatile packages? (ok. Maybe not a little question, but I am tired as hell...) Jun 16 10:24:36 udeb files are packages used by debian-installer Jun 16 10:24:45 udebs (micro-debs, but created with an assuption of weak unicode support) are only for the installer. Jun 16 10:24:51 you dont need them if you dont build an installer image based on d-i Jun 16 10:25:12 Ok, good. Jun 16 10:25:22 So only *.deb are of my interest then.. Jun 16 10:25:27 you should really make your build command more explicit Jun 16 10:25:42 -versatile is another arm subarch Jun 16 10:26:06 I just followed manual pointed by rsalveti ;-) Jun 16 10:26:19 It was not very... um, explanatory. Jun 16 10:26:23 That doesn't have instructions for only building some flavours of the kernel? Jun 16 10:26:38 fakeroot debian/rules binary-debs should just build the linux-image and headers packages Jun 16 10:26:42 http://www.omappedia.org/wiki/Ubuntu_kernel_build_alternatives Jun 16 10:26:52 sigh Jun 16 10:26:55 ogra, That still builds all the flavours though Jun 16 10:27:05 why didnt you use any ubuntu documentation instead ? Jun 16 10:27:11 in manual there is binary-arch Jun 16 10:27:15 not binary-debs Jun 16 10:27:20 persia, oh, right, its the main ubuntu soucretree Jun 16 10:27:26 * ogra_ forgot Jun 16 10:27:31 ogra_: I used what rsalveti told me... Jun 16 10:27:47 garagoth, You did the right thing. We're unhappy with the docs, not with your execution. Jun 16 10:27:55 well, there are about three very detailed wikipages on the ubuntu wiki Jun 16 10:27:56 garagoth, You did more than you needed, because the docs suck. Jun 16 10:28:07 Just use the "omap" kernel .deb on your beagle. Jun 16 10:28:19 Ok. Many thanks. Jun 16 10:28:29 ogra, I thought the omappedia page was based on one of them Jun 16 10:28:30 * persia is looking Jun 16 10:28:57 ogra_: If you would be so king and provide a link to better tutorial... Jun 16 10:29:14 persia, well, as the omappedia pages are all based on something ... the prob is there was a snapshot taken but the updates dont move over Jun 16 10:29:32 ogra, https://wiki.ubuntu.com/Kernel/Action/BuildKernelPackage doesn't help much :( The wiki hates me. Jun 16 10:29:57 I agree that it's better to use the w.u.c source, but we have to know a good one, and I fear the messiness of w.u.c is biting us here. Jun 16 10:30:05 persia: heh, this is even shorter then omappedia... Jun 16 10:30:31 garagoth, Sure, but it skips a lot, doesn't explain how to cross compile, and still results in what you got. Jun 16 10:30:43 :-) Jun 16 10:31:26 https://wiki.ubuntu.com/Kernel/BuildYourOwnKernel is one Jun 16 10:31:36 https://wiki.ubuntu.com/Kernel/BuildYourOwnKernel is a bit nicer, but I thought there was a cross-compile one. Jun 16 10:32:02 there is amitk's blog Jun 16 10:32:15 but thats not up to date, i think hrw had a more up to date one Jun 16 10:32:29 ogra_: yes, go to hrw's blog post Jun 16 10:33:09 No, there's a wiki page about it. Jun 16 10:33:18 If that's not up-to-date, then we fail at docs. Jun 16 10:33:26 http://marcin.juszkiewicz.com.pl/2010/10/19/how-to-cross-compile-arm-kernel-under-ubuntu-10-10/ one? Jun 16 10:33:37 * ogra_ likes https://wiki.ubuntu.com/SergeHallyn_ppakernels Jun 16 10:34:00 short and simple Jun 16 10:34:11 Doesn't cover the use case though. Jun 16 10:34:16 no Jun 16 10:34:26 covers mine though :) Jun 16 10:34:38 https://wiki.linaro.org/Resources/HowTo/PackageYourOwnKernel is more complicated Jun 16 10:34:42 Only because you're refusing to participate in Ubuntu these days. Jun 16 10:34:46 who uses cross compilers anyway :P Jun 16 10:34:52 ... me? Jun 16 10:34:56 persia, huh ? Jun 16 10:34:58 me Jun 16 10:35:00 :) Jun 16 10:35:11 ogra, Pushing kernels to PPAs when they ought be in the archive. Jun 16 10:35:21 pfft, *REAL MEN* compile natively :) Jun 16 10:35:25 no way Jun 16 10:35:33 two hours to get a kernel Jun 16 10:35:44 i would prefer to pull myself a tooth Jun 16 10:35:44 persia, it cant be in the archive yet ... Jun 16 10:35:51 I cross-compiled this kernel for few hours. I took looots of disk space. Jun 16 10:35:59 No way I would do this natively... Jun 16 10:36:03 ogra, Why not? Jun 16 10:36:04 garagoth: define looots Jun 16 10:36:11 ppisati, i just always have a tree around where i only remobuild changes Jun 16 10:36:18 *rebuild Jun 16 10:36:28 only if i know they work i roll a package Jun 16 10:36:44 persia, because i cant upload to natty :P Jun 16 10:37:03 (and even if i could i wouldnt gain any benefit from that) Jun 16 10:37:10 The rest of us are working on oneiric these days :p Jun 16 10:37:11 hrw: looots == 7.9 GB Jun 16 10:37:14 ogra_: add kernel to oneiric and provide it also in ppa? Jun 16 10:37:15 * persia stops teasing ogra Jun 16 10:37:20 garagoth: 8GB for kernel tree? Jun 16 10:37:25 * ppisati tries to figure out what's the plan for oneiric/ti-omap4 Jun 16 10:37:33 hrw, if you donate me 24h every day :P Jun 16 10:37:38 this is what I have in kernel dir when cross-compile finished Jun 16 10:37:52 garagoth: huge Jun 16 10:37:56 inclugind generated packages Jun 16 10:37:59 garagoth, you did it wrongly Jun 16 10:38:07 garagoth: good that my panda has still 153GB free on /home Jun 16 10:38:16 ops. 135GB Jun 16 10:38:16 you only wanted an omap3 beagle deb, but built *everything* Jun 16 10:38:20 hrw: USB HDD ? Jun 16 10:38:27 ogra, Yeah, but that's our fault: the documentation sucks. Jun 16 10:38:28 garagoth: yes - 320GB sata Jun 16 10:38:40 ogra_: Now I know. I blame documentation :D Jun 16 10:38:43 actually to cross compile i do: Jun 16 10:38:50 export $(dpkg-architecture -aarmel); export CROSS_COMPILE=arm-linux-gnueabi- Jun 16 10:38:52 hrw: How you connected SATA to Panda? Jun 16 10:38:57 fakeroot debian/rules clean; fakeroot debian/rules binary-$flavour Jun 16 10:38:58 garagoth, ogra@isis:~/kernel/linux-ac100-2.6.37$ du -hcs Jun 16 10:38:58 501M . Jun 16 10:39:38 oh, wait, thats unbuilt Jun 16 10:39:47 :D Jun 16 10:39:48 [flag@newluxor ubuntu-natty]$ du -sk . Jun 16 10:39:48 3937700 . Jun 16 10:39:49 :P Jun 16 10:40:08 3.8G Jun 16 10:40:28 still to much if you only build the linux-image and -headers debs Jun 16 10:40:30 binary-omap -> omap3 Jun 16 10:40:39 fakeroot debian/rules binary-omap Jun 16 10:40:42 natty/master Jun 16 10:40:52 if you have any recepie to cut it, i'm all ears :) Jun 16 10:41:11 doesnt that roll udebs and -docs and the like (and ddebs) Jun 16 10:41:16 when i'm doing bisect/bug finding i use linus vanilla Jun 16 10:41:25 copy .config from some ubuntu kernel Jun 16 10:41:32 and then make uimage Jun 16 10:41:44 it works and it takes much less time and space Jun 16 10:41:49 nope Jun 16 10:41:50 yeah Jun 16 10:41:55 onlye kernel and header Jun 16 10:41:58 thats how i do it for the ac100 too Jun 16 10:42:03 Oh my. The page on the wiki got removed as part of Kernel Wiki Gardening, and replaced with a link to amitk's old blog post. Jun 16 10:42:11 only for a release i actually roll a test package Jun 16 10:42:26 btw, one day i want to find out what are the minimun required configs to make a vanilla kernel boot an ubuntu userland Jun 16 10:42:38 garagoth: usb->sata enclosure? Jun 16 10:42:40 garagoth, Sorry. For now, the omappedia page is the best one, but if you look at the Ubuntu pages, you can find ways to build fewer kernels to achieve your goals. Jun 16 10:42:47 persia: Linking to someones blog from wiki for official instructions... well, looks unprofessional :-) Jun 16 10:43:15 Especially when the author of the blog post recommends a different resource (see backscroll). Jun 16 10:43:24 hrw: Mm. And what transfer rates you have? I wanted to replace my home server with a panda/beagle cluster Jun 16 10:43:25 In general, we don't do much with cross-compiling (except hrw). Jun 16 10:43:54 garagoth: poor - 20MB/s max Jun 16 10:43:57 hrw, You should take over some area of the wiki, and build the comprehensive cross-compilation resource, and make all the other wikis (e.g. omappedia) point there. Jun 16 10:44:30 persia: sounds like a plan Jun 16 10:44:45 \o/ Real Documentation! Jun 16 10:48:46 ppisati, didnt you have a spec for that ? Jun 16 10:48:59 * ogra_ thought he saw one Jun 16 10:49:45 ogra_: about what? Jun 16 10:50:11 btw, one day i want to find out what are the minimun required configs to make a vanilla kernel boot an ubuntu userland Jun 16 10:50:29 ogra_: nope, unfortuntaly no Jun 16 10:50:37 oh ? Jun 16 10:50:43 did it get dropped ? Jun 16 10:51:02 never heard of a spec about it Jun 16 10:54:28 YESSS!!!!! i2c bus 2 working !!! Jun 16 10:54:47 But for some reason there is no entry about expansion board in dmesg... hum... Jun 16 10:55:32 ppisati, yeah, i only saw the workitem on https://launchpad.net/ubuntu/+spec/other-kernel-o-config-review ... seems i mixed that up Jun 16 10:58:49 ogra, The minimums are tricky (but there is work to document them). There's a number of cases where things work for *some* systems and break horribly for others, because of assumptions in userspace about kernel config that only are exposed in non-default environments. Jun 16 10:59:22 sure Jun 16 10:59:38 * ogra_ didnt say it was an easy task Jun 16 10:59:43 :) Jun 16 11:03:20 Heh, yeah. But seriously, I believe the kernel team has been collecting this data since ~karmic because of some regressions discovered at karmic release because the release config didn't match userspace expectations. Jun 16 11:03:51 I doubt they have *all* the quirks documented yet, but there ought be a fair collection of information. I forget where, but remember apw telling me that it was in git somewhere. Jun 16 11:04:42 Hrm. The magic summon-the-man-who-remembers-stuff-about-the-kernel incantation didn't work :( Jun 16 11:05:32 persia, collecting which data ? Jun 16 11:05:38 LOL Jun 16 11:05:44 it worked :) Jun 16 11:06:12 apw, Which kernel options are required to match userspace expectations for Ubuntu. Jun 16 11:06:45 ahh i see, we have been codifying the values of ones we have found to be required, in the enforcer file, but that doesn't cover everything which is required sadly Jun 16 11:07:15 Where is the enforcer file, so ogra can have a peak at the current knowledge? Jun 16 11:08:22 persia, i know where it is i think Jun 16 11:08:54 ogra@isis:~/kernel/linux-ac100-2.6.37$ ls debian.ac100/config/enforce Jun 16 11:08:55 debian.ac100/config/enforce Jun 16 11:08:58 here we go Jun 16 11:09:13 apw, Thanks! Jun 16 11:09:34 ogra, Good luck. If you find more required configs, please make sure they reach that file for all the kernels. Jun 16 11:09:43 i dont get why its in the flavour subdir though Jun 16 11:10:09 since it checks lots of general values ... that should just be in debian/ Jun 16 11:10:11 Because debian.foo/ generates debian/ so anything in debian/ is subject to wiping. Jun 16 11:10:27 only parts of debian/ Jun 16 11:10:45 Yeah, well. debian.foo/ is where all the kernel team work happens. Jun 16 11:10:59 They like to leave debian/ alone if possible, considering it strange black magic (which it is) Jun 16 11:11:35 s/black magic/insane perl/ Jun 16 11:11:37 :) Jun 16 11:12:39 * persia agrees with Arthur C. Clarke, and continues to call it "magic" in compliance with common duck-typing algorithms. Jun 16 11:12:48 heh Jun 16 11:13:08 magic always has something positive in it for me ... Jun 16 11:13:24 sick parl code somewhat misses that positivity :) Jun 16 11:13:33 *perl Jun 16 11:13:36 * persia uses "black" and "white" to indicate relative positions on that continuum Jun 16 11:14:01 And "strange" relates to it's unrelation to anything else anywhere else Jun 16 11:14:55 well, even 'black' magic might make me sit in awe ... while some perl code just lets me sit in tears ;) Jun 16 11:15:04 ;-) Jun 16 11:15:16 Right. That's the "strange" part. Jun 16 12:14:05 morning Jun 16 12:14:09 * rsalveti reading backlog Jun 16 12:15:21 Good morning. Jun 16 12:32:53 garagoth: awesome, good you got it working Jun 16 12:33:02 yeah, next step is to build just the omap flavor :-) Jun 16 12:33:08 then it'll be even faster Jun 16 12:33:26 persia: problem is that there is no official cross-build documentation anywhere at wiki.ubuntu Jun 16 12:33:33 that's why I pointed the omap one Jun 16 12:33:41 as it describes well how to do it Jun 16 12:33:54 rsalveti: hrw volunteered to create some, so that's a solved problem. Jun 16 12:34:22 well, it's solved once it's done ;-) Jun 16 12:34:24 No criticism was implied to giving garagoth helpful advice: the critism was at *all of us* for not having the pages on the wiki. Jun 16 12:34:29 lool, i'm looking at debian bug 550584, i dont really understand the purpose of your highest-abi file ... cant we determine the highest abi from just listing /boot ? Jun 16 12:34:31 Debian bug 550584 in flash-kernel "flash-kernel not run when going to new upstream kernel version" [Important,Open] http://bugs.debian.org/550584 Jun 16 12:35:28 rsalveti: The problem of "What are we goind to do about the lack of docs" is solved. The problem "There are no good cross-compilation docs on the wiki" is pending :) Jun 16 12:36:14 garagoth: I believe if you run with binary-omap it should just build for omap flavor, but need to check Jun 16 12:36:27 yes Jun 16 12:36:56 Indeed. Jun 16 12:37:40 rsalveti: :-) Jun 16 12:37:50 Took me to compile this 3-4 hours Jun 16 12:38:51 garagoth: that's because you compiled for both versatile and omap Jun 16 12:38:56 so just omap should be half Jun 16 12:39:08 and takes a while because there are tons of modules Jun 16 12:39:12 well, thats still quite long for cross building Jun 16 12:39:23 rsalveti: I need to verify one more thing, as in dmesg there was no info that expansion board was detected. But i2c works fine! Jun 16 12:39:27 the ac100 package natively takes 2h here Jun 16 12:39:28 well, that depends on the host ;-) Jun 16 12:39:34 I have dual core 2Ghz Jun 16 12:39:45 ogra_: Right. Two flavours ~ 3-4 hours. One flavour ~ 2 hours. Jun 16 12:39:51 4255 bogomips per core Jun 16 12:39:59 persia, natively ... Jun 16 12:40:13 persia, vs 4 for two flavours cross Jun 16 12:40:14 Heh, I know. See prior point about ARM hardware being good these days :) Jun 16 12:40:22 heh Jun 16 12:40:29 garagoth: can you paste me the patch you used? Jun 16 12:40:33 (and cross-compilation not actually being an interesting problem anymore) Jun 16 12:40:39 yeah Jun 16 12:40:48 as i said, real men build native Jun 16 12:40:56 uh. I used one that you pointed me to and tweaked it a little. Jun 16 12:40:59 * ogra_ wants a T-Shirt with that *g* Jun 16 12:41:11 * persia suspects this also applies to real women and real small green furry things from Alpha Centuri Jun 16 12:41:27 one chunk was failing, so I added it manually... and irq_set_irq_type I corrected to set_irq_type Jun 16 12:41:38 not sure about the furry things ... but you could read men as "humans" Jun 16 12:41:56 garagoth: ok, that should be fine Jun 16 12:42:28 i doubt green furry things from alpha centauri use C ... they will likely use green furry C ... Jun 16 12:42:29 I gen generate a patch for you Jun 16 12:42:36 I still have original file Jun 16 12:46:12 http://nevander.eu.org/patch_expansionboards_2.6.38.10.patch Jun 16 12:48:43 garagoth: your u-boot needs to set the buddy kernel argument Jun 16 12:49:01 I believe the extension support is enable at u-boot already Jun 16 12:49:16 but then you probably need something like buddy=${buddy} at your kernel cmd line Jun 16 12:51:32 let me check u-boot sources Jun 16 12:51:49 aaah. Jun 16 12:51:54 true!!! Jun 16 12:52:14 Let me check :-) Jun 16 12:52:38 garagoth: see if you're u-boot is already saying that you have an expansion board Jun 16 12:52:40 ogra_: one way is to scan /boot, the other is to note down the versions when installing kernels; but note that the bug concludes we should copy the grub logic Jun 16 12:52:49 Yes, it is detecting it correctly Jun 16 12:53:29 ogra_: debbugs are a bit hard to read because they are threaded; I usually download the mbox and either read by thread or by date Jun 16 12:53:52 garagoth: so try adding buddy=${buddy} at your kernel cmd line arguments Jun 16 12:54:06 garagoth: edit /boot/boot.script and then run flash-kernel Jun 16 12:54:50 lool, yeah, i read that ... do you think the patch will make it into debian soon ? i would like to have it before A2 8we have massive upgrade probs due to that bug) Jun 16 12:55:38 if it doesnt make debian in time, i'll just addi it to ubuntu until i can sync it, but if avoidable i wouldnt like to waste time on having to carry it in ubuntu Jun 16 12:57:08 ogra_: the patch needs to be reworked to use the grub bits and heavily tested; I'm getting asked about f-k stuff in Debian frequently, but I have a hard time making progress on it Jun 16 12:57:23 should improve in the next weeks hopefully Jun 16 12:57:48 well, oneiric can be your guinea pig Jun 16 13:23:41 rsalveti: It works !!! Jun 16 13:29:06 garagoth: awesome Jun 16 13:29:22 Definitely. Fully working beagle now... Jun 16 14:01:10 hello, is ubuntu-omap4-extras still instable on natty? Jun 16 14:01:38 instable ? Jun 16 14:01:47 uragano2: should be, why? Jun 16 14:01:54 tested yesterday with a fresh image and it worked fine Jun 16 14:03:07 * ogra_ still wonders if instable is a typoed installable or unstable :) Jun 16 14:04:06 because i installed it already one time and my system become unstable, so searching on google i found this https://bugs.launchpad.net/ubuntu/+source/jasper-initramfs/+bug/759817 Jun 16 14:04:08 Ubuntu bug 759817 in jasper-initramfs "ubuntu-omap4-extras from the TI PPA is not installable in natty" [Medium,Fix released] Jun 16 14:04:29 thats long fixed Jun 16 14:04:59 your unstability surely doesnt have to do anything with ubuntu-omap4-extras itself Jun 16 14:05:21 probably with something this pulls in, but these are plenty of packages Jun 16 14:05:27 ok...i try to install it again hoping that this time it'll work fine :D Jun 16 14:05:41 we didnt chnage anything since release Jun 16 14:05:45 *change Jun 16 14:17:42 chromium-browser build passed 310 minutes. just 2.5h to possible breakage Jun 16 14:22:15 hrw, how long does it usually take ? Jun 16 14:22:29 last time it took 440 minutes to fail Jun 16 14:22:39 (and why is everyone building chromium today ?) Jun 16 14:22:42 on panda Jun 16 14:22:49 ogra_: it ftbfs on armel Jun 16 14:22:56 ah Jun 16 14:23:08 well, someone in #ac100 compiles it too today Jun 16 14:23:22 I have a patch for 440 minutes breakage applied Jun 16 14:23:32 * ogra_ crosses fingers Jun 16 14:23:43 bug 791283 btw Jun 16 14:23:44 Launchpad bug 791283 in chromium-browser "chromium-browser version 11.0.696.71~r86024-0ubuntu1 failed to build on armel" [Undecided,Confirmed] https://launchpad.net/bugs/791283 Jun 16 14:24:02 12.something.x.y also fails Jun 16 14:24:18 yeah, i saw it on the fstbfs list since a while Jun 16 14:29:41 this time the installation finished successfully :D Jun 16 14:45:25 audio throws hdmi doesn't work...can u help me? Jun 16 14:45:57 *through Jun 16 15:47:10 Hi. The oneiric daily headless OMAP 4 doesn't want to boot the kernel Jun 16 15:47:34 heh, that can well be Jun 16 15:47:37 Is there a place to file bugs or does one just report here? Jun 16 15:47:47 see the topic Jun 16 15:48:02 Ah, cool Jun 16 15:48:13 we just changed the build system completely, so its no surprise there are issues :) Jun 16 15:48:19 Okay. Jun 16 15:48:38 No problem. Is there a more stable version that one might try? Jun 16 15:48:44 Like a natty build? Jun 16 15:49:01 Or what do people recommend for getting a debian based OS onto the Panda? Jun 16 15:49:09 ah Jun 16 15:49:54 https://wiki.ubuntu.com/ARM/OMAP Jun 16 15:49:57 jeremiah, There are working natty images. Jun 16 15:49:58 see that page Jun 16 15:50:14 persia: Oh really? Okay, I'm happy to try on of those. :) Jun 16 15:50:25 I guess I'll use my google-fu :) Jun 16 15:50:34 No, check the page ogra gave you Jun 16 15:50:49 Ah, okay, will do. Jun 16 15:50:51 Has direct links and everything. Without sponsored links or ads or anything :( Jun 16 15:50:55 s/(/)/ Jun 16 15:50:58 heh Jun 16 15:53:43 Well, I'll try the 11.04-preinstalled-headless-armel+omap4 and see what happens. Jun 16 15:54:16 Beware that you have selected a *very* minimal install: you likely have to install bundles of things to make it do what you want. Jun 16 15:54:22 jeremiah: what rev board do you have? Jun 16 15:54:44 persia: I plan to do that, I'm going to try to spin up the Automotive respin Jun 16 15:54:52 GrueMaster: Let me check . . . Jun 16 15:55:14 ah, for the automotive respin that image might actually still be a bit big :) Jun 16 15:55:14 Should be a label on the bottom of the board. Jun 16 15:55:25 GrueMaster: Looks like the lable says REV A# Jun 16 15:55:30 Sorry, Rev A3 Jun 16 15:55:32 ogra, Isn't it just ubuntu-minimal? Jun 16 15:55:51 jeremiah, you might want to talk to Dr_Who i know he knows some tricks to make them even smaller Jun 16 15:55:57 persia, -standard iirc Jun 16 15:56:02 ogra_: Really? Okay. :) Jun 16 15:56:06 A3??? Didn't know they were available. Jun 16 15:56:10 jeremiah, Do you particularly want to do it as a remix (as it is called in our nomenclature)? Jun 16 15:56:17 GrueMaster: I just got onw Jun 16 15:56:28 cool. let me know how it goes. Jun 16 15:56:32 persia, yes, it was defined at UDS Jun 16 15:56:36 persia: Well, I've been working with some Canonical folks to get this up and running. Jun 16 15:56:38 an IVI remix Jun 16 15:56:44 I work the GENIVI consortium Jun 16 15:56:50 ogra, It's not tricks: it's a smaller seed collection (linaro nano) Jun 16 15:57:02 Hi jeremiah Jun 16 15:57:04 And Canonical is planning on releasing the official GENIVI "respin" Jun 16 15:57:06 persia, its also omitting the installation of any docs etc Jun 16 15:57:07 Dr_Who: Hi! Jun 16 15:57:29 nano is surely so shrunk down for a proper ubuntu Jun 16 15:57:33 Dr_Who: I was hoping to do some testing of the Ubuntu / GENIVI respin for automotive Jun 16 15:57:42 yeah and in the next incantation, replacing some of the core elements with their busybox counterparts Jun 16 15:57:44 but something between minimal and nano should be cleanly possible Jun 16 15:57:50 Ah. I'd be in favour of having an automotive flavour from which Canonical derived, but that may require more coordination. Jun 16 15:57:54 I have a list of packages and I'll add some of the GENIVI components Jun 16 15:57:55 * GrueMaster wanders off ins search of something called the "loo". Jun 16 15:58:01 jeremiah: sounds like fun! Jun 16 15:58:11 Dr_Who: It is! :) Jun 16 15:58:13 GrueMaster, two times right Jun 16 15:58:20 Any time I get to fool around with Ubuntu its fun Jun 16 15:58:25 down the corridor Jun 16 15:58:49 jeremiah: you might be interested in https://wiki.linaro.org/LiveHelper/Hacking Jun 16 15:59:01 basically how to roll your own images Jun 16 15:59:01 persia: Well, there has been a bit of co-ordination with Chris Kenyon, Ken Edwards, et. al. from Canonical. Jun 16 15:59:09 Dr_Who: Okay, I'll read that. Jun 16 15:59:11 Thanks Jun 16 15:59:26 I can point you at the things we did for nano which may or may not be of interest to you Jun 16 16:00:02 I think if I can get a kernel to boot on the OMAP I'd be happy. I'm willing to leave the userland pretty plain for a bit. Jun 16 16:00:09 jeremiah: o actually its in that wiki page Jun 16 16:00:17 We don't have an official OMAP release until October. Jun 16 16:00:28 ah .. plenty of time Jun 16 16:00:43 ah, well in that case the headless image is surely the least painful startin point Jun 16 16:00:44 (Even though the official Canonical press release claiming compliance is out!) =) Jun 16 16:00:53 ogra_: Cool, I'll start there. Jun 16 16:01:02 I already have the headless running on the OMAP 3 Jun 16 16:01:10 Really nice. Jun 16 16:01:12 and then apply nano imporvements where possible Jun 16 16:02:46 So the steps would be to boot Natty on the panda then use Live Helper? Jun 16 16:03:20 image is the same on omap4. Jun 16 16:03:32 Booting Natty on the panda gives you a development environment. Jun 16 16:03:44 Live Helper lets you build images based on some archive (you might use natty). Jun 16 16:03:56 jeremiah, live helper is to actually produce rootfses Jun 16 16:03:56 persia: Ah, okay. Thanks. Jun 16 16:04:05 Running live-helper (or rather live-build) on the panda lets you build images for use on the panda. Jun 16 16:04:20 right, it was renamed Jun 16 16:04:30 ogra_: okay, now I get it. So it would allow me to create a custom root file system Jun 16 16:04:34 That is frickin' great. Jun 16 16:04:41 Just what I'm looking for. Jun 16 16:04:41 If you're testing experience, you can run live-build on an arbitrary platform where you know you have support, and build images for that platform. Jun 16 16:05:08 If you're testing platform support, you want to run it natively (at least on the same architecture) Jun 16 16:05:08 persia: Hmm, interesting. I need to do some testing but don't really have a testing environment yet. Jun 16 16:05:32 Do you need to test images, or just test software? Jun 16 16:05:38 The goal is to port some automotive software to the OMAP Jun 16 16:05:50 In that case, I'd recommend completely ignoring live-build for now. Jun 16 16:05:52 So mostly just testing apps Jun 16 16:06:00 Ah, okay. Fair enough. Jun 16 16:06:01 yeah, what persia said Jun 16 16:06:02 Just install Natty on the panda. Jun 16 16:06:12 I plan to do some building and then package the apps so others can test Jun 16 16:06:16 Then install the application you want to test (package it for easy uninstall/upgrade/etc.) Jun 16 16:06:21 Later we'll do the integration. Jun 16 16:06:40 Cool, sounds like a plan. Jun 16 16:06:57 If you're doing packaging, I'd strongly recommend either doing packaging on *some other* machine, and just build on the panda *OR* install the Netbook image on the panda and use that as a development environment. Jun 16 16:07:03 (depends how many Ubuntu boxes you have around) Jun 16 16:07:18 I have a Tegra2 box Jun 16 16:07:23 Starting from the headless image as a development environment will get annoying very quickly. Jun 16 16:07:25 And a Intel Core 2 duo Jun 16 16:07:27 Ah, that works :) Jun 16 16:08:02 Either is good for packaging. For building, you want to build the package on the tegra or the panda. If both are running Ubuntu, the results ought be identical. Jun 16 16:08:24 I won't need an magic VFP flags for the tegra? Jun 16 16:08:41 Once you know the set of packages you want to deploy, then it's worth looking at seeds, live-build, etc. and attempting to make an image to distribute as the remix. Jun 16 16:08:52 No. Jun 16 16:08:55 I thought that I might face some NEON issues. But the box I have, called the TrimSlice seems to be working pretty okay Jun 16 16:09:14 by default ubuntu doesnt use NEON at build time anywhere Jun 16 16:09:19 thats a requirement Jun 16 16:09:22 Ah okay, good to know Jun 16 16:09:23 No. Jun 16 16:09:42 Rather, Ubuntu doesn't default to NEON instructions anywhere, and only uses NEON where runtime detection is available. Jun 16 16:09:44 your Sw either needs runtime detection two binary packages Jun 16 16:10:04 Ubuntu *definitely* uses NEON at build time, or we wouldn't be able to support NEON runtime detection :p Jun 16 16:10:23 ubuntu doesnt support packages using neon at build time Jun 16 16:11:05 The buildds can't execute NEON? Jun 16 16:11:14 sure they can Jun 16 16:11:40 So, how isn't it supported? Jun 16 16:11:53 By policy, packages aren't allowed to depend on NEON being available at runtime. Jun 16 16:11:58 But I don't see how this affects build-time. Jun 16 16:12:02 it isnt supported to default to NEON and to hardcode that at build time Jun 16 16:12:18 And, further, I suspect that test code run at build time *would* run NEON, as the runtime detection should notice it is available. Jun 16 16:12:29 Ah, yes, that's not supported :) Jun 16 16:12:34 phew Jun 16 16:12:52 finally i got the formulation of the sentence in a way that persia agrees Jun 16 16:13:01 Amazing. Jun 16 16:13:08 P.S. It's easier to talk about not supporting NEON-by-default *at runtime* than fussing about build-time :p Jun 16 16:15:08 jeremiah, Anyway, my understanding is that the base VFP stuff is implemented the same for a range of SoCs, and we're mostly using that. This does mean that we don't take advantage of some of the capabilities of TEGRA (but that would need runtime detection, etc. for all the same reasons it's required for NEON) Jun 16 16:36:25 Okay, cool! Jun 16 16:39:56 somebody know why during the boot i see IP-Config: eth0 hardware address MAC-A mtu 1500 DHCP [ 2.775718] eth0: link up IP-Config: no response after 60 secs - giving up Jun 16 16:42:20 uragano2, Maybe there is software attempting to bring up a network link that you aren't using? Jun 16 16:43:30 i see this since the first time that i installed natty Jun 16 16:46:01 persia: no, sorry. the error was an other i confused! anyway i have installed only ubuntu-omap-extras and dropbear, Jun 16 17:08:27 Anyone know where I can get the stuff to enable the sound on my pandaboard on Oneiric? Jun 16 17:08:50 stuff ? Jun 16 17:08:58 ogra_: not sure what I need to make it work Jun 16 17:09:09 ogra_: So I punted and was non-specific. :) Jun 16 17:09:35 ogra_: If I plug in a USB sound "card" that works fine. Jun 16 17:09:51 bug 746023 Jun 16 17:09:51 Launchpad bug 746023 in alsa-utils "No sound on omap4" [High,In progress] https://launchpad.net/bugs/746023 Jun 16 17:10:04 see the pre-last comment from tobin Jun 16 17:10:19 it is also in the release notes and on various wikipages Jun 16 17:10:20 Ah, thanks, will look. Jun 16 17:41:15 i am experiencing many issues with natty!!! did u do too??? Jun 16 17:42:58 uragano2: natty is past... Jun 16 17:43:31 what do u mean? Jun 16 17:43:48 I am experiencing issues with oneiric Jun 16 17:48:49 by any chance does a pandaboard compiling ubuntu kernel page exist? Its been years since i built a kernel, and that was on familiar hardware. I have no idea what to enable/disable for a pandaboard. Jun 16 17:52:07 MrCurious, Grab the kernel from the archive, and look at the config as a base. Jun 16 17:52:41 you mean apt-get install kernel? Jun 16 17:54:31 MrCurious: the config files for an ubuntu kernel are installed to /boot/config-*-omap4 Jun 16 17:54:42 oh sweet Jun 16 17:54:55 now the end of persia's sentence parses (in my head) Jun 16 17:54:59 apt-get source, but that was the idea. Jun 16 17:55:14 just to locate the kernel. reading through a apt-cache search now Jun 16 17:56:51 It's "linux-to-omap4" for the panda, isn't it? Jun 16 17:56:59 Err, "linux-ti-omap4" Jun 16 17:59:00 from looking at the uname compared to the avail kernel sources it seems there were some patches put against it? Jun 16 17:59:24 the -1208-omap4 Jun 16 18:04:48 That's the name of the binary package. GO ahead and grab that, and apt will do the right thing. I'm not sure it does the right thing with the source name for kernel packages (I know it does the wrong thing for the main Ubuntu kernel, and filed a bug about it a few years ago) Jun 16 18:05:37 so grab the source on teh -1208-omap4 binary kernel? Jun 16 18:06:37 Right. That gets you the source, all the patches, and the configuration used. Jun 16 18:07:00 You can fiddle with that with relative confidence that you can undo what you did and get back to what was installed by default. Jun 16 18:07:02 oh sweet, the source tag is going to become my new best friend of the minute Jun 16 18:07:25 apt-get source is a handy tool: it's part of what makes using an open source operating system fun :) Jun 16 19:08:32 what means if i run arp -a form my pc and it returns "pandy.homenet.telecomitalia.it (192.168.1.217) at on wlan0"..why is it incomplete? Jun 16 19:42:20 to build the image for pandaboard, on pandaboard, its make uImage then make uinstall (will that work right when the msdos partition is on SD, and the root partition is on a USB drive? Jun 16 19:49:56 here is to hoping that all the configuration options are auto set Jun 16 19:54:55 well, make uImage fails 3 files in with no rule to make target Jun 16 19:55:06 it has been a long time. make dep is now un-needed Jun 16 19:58:39 guessing i have to copy the config from /boot to defconfig in the kernel dir? Jun 16 19:59:09 MrCurious, Build a kernel package, and then install it. Let flash-kernel take care of the details. Jun 16 19:59:49 having trouble getting the kernel build to take off Jun 16 20:00:09 trying to put in teh patch to usb that prpplague mentioned in teh pandaboard post from a few hours ago Jun 16 20:04:02 trying to work out how to get /boot/config-2.6.38-1208-omap4 to apply to the kernal-image-2.6.38-1208-omap4 sources i downloaded Jun 16 20:04:06 and get that image built Jun 16 20:05:22 ubuntu/aufs/magic.mk no such file or directory Jun 16 20:05:29 thinking i am on teh wrong path Jun 16 20:05:48 make clean doesnt work Jun 16 20:06:07 https://wiki.ubuntu.com/Kernel/BuildYourOwnKernel might help. It'S written from the i386/amd64 perspective, but most things are basically the same. Jun 16 20:06:32 ty Jun 16 21:31:58 persia: when i get to the fakeroot debian/rules step i discover i lack a debian/rules Jun 16 21:32:03 any hings? Jun 16 21:32:16 hints rather Jun 16 21:32:17 How did you get the source? Jun 16 21:32:25 apt-get source linux-image-$(uname -r) Jun 16 21:32:36 which gives me /usr/src/linux-headers-2.6.38-1208-omap4 Jun 16 21:32:43 i wonder if make clean nuked that Jun 16 21:32:51 It shouldn't. Jun 16 21:33:33 It should give you a .dsc file, a .gz file, and an unpacked directory "linux-omap" in the directory in which you ran apt-get source. Jun 16 21:34:23 it did make a src dir in the dir i ran that Jun 16 21:34:36 is that the kernel dir i should be messing with instead of the one in /usr/src? Jun 16 21:34:41 Yes. Jun 16 21:34:50 I'm not sure how you got the one in /usr/src Jun 16 21:36:08 i grabbed the kernel in a variety of ways, i guess Jun 16 21:36:49 building the kernel outside of /usr/src feels ODD Jun 16 21:37:27 Heh. We do offer several choices, so that people who want to do different things can. Sometimes that seems like a good idea, and sometimes it makes it awkward (as many fok randomly search the internet for solutions, and find all sorts of recommendations out there) Jun 16 21:37:49 yup Jun 16 21:38:02 i am documenting my travels, so i can only ask once Jun 16 21:38:19 Ask as much as you like. Jun 16 21:38:30 ty Jun 16 21:38:42 We don't always agree on things, but if you ask, we're more likely to point you to reasonably heavily trod paths :) Jun 16 21:38:55 If you don't ask, you never know what you might find. Jun 16 21:39:09 thanks! oh, a bud pointed me to a rather simple way of getting ubuntu to bring up wireless on boot Jun 16 21:39:54 Which? I have a suspicion I won't like it, but still ... :) Jun 16 21:40:18 i pm you it Jun 16 21:40:53 it may not be perfect, as i noticed the box lost networking after being on for a day or two Jun 16 21:41:17 make: *** No rule to make target `binary-generic'. Stop. Jun 16 21:41:24 that sounds ominous Jun 16 21:41:30 wonder if i need binary-generic Jun 16 21:41:55 It's a rule in debian/rules. Not something external you need. Jun 16 21:42:32 You shouldn't have to uninstall Network Manager to do that: Network Manager is supposed to check /etc/network/interfaces and ignore anything defined there. Jun 16 21:42:46 hmmm Jun 16 21:43:22 hmm debian/rules lacks a binary-generic Jun 16 21:44:02 root@unwin-desktop:~/linux-ti-omap4-2.6.38# grep binary debian/rules Jun 16 21:44:03 binary: binary-indep binary-arch Jun 16 21:44:03 include $(DROOT)/rules.d/2-binary-arch.mk Jun 16 21:44:03 include $(DROOT)/rules.d/3-binary-indep.mk Jun 16 21:44:38 perhaps the build mechanism has mutated since the howto page was last touched Jun 16 21:45:33 Oughtn't have been. Jun 16 21:45:41 Maybe the ti-omap kernel is different somehow. Jun 16 21:46:04 Anything in rules.d/* ? Jun 16 21:46:11 well, i think i will just push on, and hope that dependency management catches any missed steps Jun 16 21:47:21 3 different rules.d dirs (debian.ti-omap4, debian, and debian.master Jun 16 21:47:34 ti-omap4 dir only has armel Jun 16 21:47:39 .mk Jun 16 21:49:36 debian/rules.d/* Jun 16 21:49:59 0-common-vars.mk 1-maintainer.mk 2-binary-arch.mk 3-binary-indep.mk 4-checks.mk 5-udebs.mk Jun 16 21:50:08 debian.ti-omap4 *should* have been used to populate debian/ when the source package was uploaded, so you oughtn't need to worry for making local packages. Jun 16 21:50:09 binary-arch is the one we want Jun 16 21:50:23 ? Jun 16 21:50:25 I don't remember what debian.master does. Jun 16 21:50:39 binary-arch.mk probably includes the definition for binary-arch: Jun 16 21:50:50 make sure debian/rules includes debian/rules.d/binary-arch.mk Jun 16 21:51:11 nope Jun 16 21:51:56 well it does i think Jun 16 21:52:08 include $(DROOT)/rules.d/2-binary-arch.mk Jun 16 21:54:35 binary-arch appears to be building stuff Jun 16 21:54:45 back in a bit. going to let it cook and see what it makes Jun 16 21:55:50 Oh, right. Jun 16 21:56:05 "binary-generic" is the expansion of "binary-${FLAVOUR}". Jun 16 21:56:26 You probably wanted "binary-ti-omap4" Jun 16 21:56:46 Oughtn't matter in practice: I believe this kernel source only has the one flavour. Jun 16 21:59:36 make: *** No rule to make target `binary-ti-omap4'. Stop. Jun 16 21:59:37 hmmm Jun 16 22:00:50 well binary-arch was making something think i will let that cook Jun 16 22:02:51 We'll see what it makes :) Jun 17 02:41:24 persia: it finished cooking Jun 17 02:41:56 and it made a bunch of _armel.udeb files and some armel.deb **** ENDING LOGGING AT Fri Jun 17 02:59:56 2011