**** BEGIN LOGGING AT Wed Sep 22 02:59:57 2010 Sep 22 06:59:33 have you had a similar problem? http://paste.ubuntu.com/498254/ Sep 22 07:07:50 I haven't seen that reported previously. Sep 22 07:08:15 Please file a bug against qemu-kvm-extras-static (assuming you generated that on an Ubuntu system) Sep 22 07:16:16 yes I've generated that using rootstock Sep 22 07:23:26 Was that error produced as part of a rootstock build? Sep 22 07:32:06 morning Sep 22 08:22:53 dmart_, fyi http://forums.computers.toshiba-europe.com/forums/thread.jspa?threadID=56355&start=30&tstart=0 Sep 22 08:23:25 dmart_, so general bootloader access is there it seems, but nobody figured out the details Sep 22 08:23:45 hi ogra Sep 22 08:26:04 I'm always working with the boot process... I'm very frustrated with the new upstart mechanism (only few times all works fine, other times no: http://paste.ubuntu.com/498292/). Sep 22 08:27:05 I don't think upstart is your issue there. Sep 22 08:27:26 with this asyncronous method seems that some daemons wait something else but I don't understand when and what Sep 22 08:28:20 persia... I think so, because some times all the boot process goes ok, sometimes no Sep 22 08:28:50 and the problem is always similar to the one I reported in the pastebin Sep 22 08:32:51 when the problem occurrs, all devices in /dev have a wrong datetime (01-01-1970) and ownership (root:root) Sep 22 08:33:26 thats with ubuntu kernel ? Sep 22 08:33:47 and does your system have an rtc battery and/or a network connection ? Sep 22 08:34:55 the rtc is taken from another processor (a secure processor) Sep 22 08:35:00 so it's good Sep 22 08:36:29 the problem seems related to the devices generated by udev during the the initrd and the devices in the real filesystem Sep 22 08:36:43 but I'm using the same initrd of the ubuntu Sep 22 08:41:00 http://paste.ubuntu.com/498300/ Sep 22 08:41:22 it seems that udev fails to be killed in the initrd?? Sep 22 08:44:21 udev is supposed to be carried over from initrd, how did you generate the initrd assuming you use a different kernel Sep 22 08:44:44 also do your kernel config options match with the ubuntu ones Sep 22 08:45:32 if your devices are all tagged 1970 then your rtc part in the kernel that sets it on kernel bringup doesnt work (default in ubuntu) Sep 22 08:46:22 I copied the initrd from the rootstock using this kernel: http://rcn-ee.net/deb/lucid/v2.6.35.4-l4/linux-image-2.6.35.4-l4_1.0lucid_armel.deb Sep 22 08:46:36 ubuntu sets the rtc dircetly at unpackaing the kernel, if that doesnt work it falls back to ntpdate ... if *that* doesnt work you can still use the fixrtc cmdline option that sets the clock to last mount time of the disk Sep 22 08:47:05 you should recreate the initrd if you use a different kernel Sep 22 08:47:16 but I don't use any modules... Sep 22 08:47:31 so the initrd should be the same except that Sep 22 08:47:42 * ogra wonders why people always assume initrd is only for using modules Sep 22 08:47:55 i heard that three times this week Sep 22 08:48:11 I have read all scripts inside the initrd Sep 22 08:48:19 it wont break but will gain you unexpected results Sep 22 08:48:26 ogra, leftovers from embedded development models Sep 22 08:48:38 and they should work with my configuration Sep 22 08:48:39 like for example udev misbehaving with broken RTC :) Sep 22 08:49:12 what HW is that exactly ? Sep 22 08:49:15 furibondox, The key bit is not the scripts *in* the initrd, but rather the scripts used to genreate the scripts in the initrd which hardcode certain things depending on the kernel. Sep 22 08:49:22 a standard beagle ? Sep 22 08:49:49 no Sep 22 08:49:55 it's not a standard beagleboard Sep 22 08:50:04 it's a custom board Sep 22 08:50:08 with omap 3530 Sep 22 08:50:10 but you are trying to use a kernel that was compilde from a beagleboard defconfig ? Sep 22 08:50:16 no Sep 22 08:50:20 we use our kernel Sep 22 08:50:24 well, the above one surely is Sep 22 08:50:43 rcn-ee kernels are all beagle kernels afaik Sep 22 08:50:57 I've only used the kernel I said you above in order to generate the initrd with the rootstock Sep 22 08:51:03 ah Sep 22 08:51:12 only for that Sep 22 08:51:15 how much ram do you have on your board ? Sep 22 08:51:34 256MB Sep 22 08:51:35 the plymouth error is likely the ureadahead OOM Sep 22 08:51:41 yeah, that would fit Sep 22 08:52:08 currently uredahead needs more than 256M else it will OOM and tear down plymouth with it Sep 22 08:52:46 (you should actually see OOM messages for both in dmesg) Sep 22 08:53:09 yesterday i tried also to disable ureadahead and plymounth but the problems remained Sep 22 08:53:15 hi, while booting omap4 sdp i am getting this error .."init: ureadahead main process (487) terminated with status 5" ... any idea what status 5 suggests ?? Sep 22 08:53:23 you cant disable plamouth Sep 22 08:53:39 ogra: yes I saw ;-) Sep 22 08:53:40 avinashhm, that you are using an SD card which it cant sacn Sep 22 08:53:53 avinashhm, or any other flash device Sep 22 08:54:07 avinashhm, that message is fine, its only info, not an error Sep 22 08:54:50 ogra, its not related to network ... ?? it hangs after this .. i do nt get the prompt .... Sep 22 08:54:56 furibondox, try removing ureadahead Sep 22 08:55:05 ok Sep 22 08:55:13 I have removed all nw options from make menuconfig ... Sep 22 08:55:15 furibondox, which prompt on which console ? Sep 22 08:55:48 just disable in the /etc/init/ureadahead or aptitude purge ureadahead ? Sep 22 08:55:54 i.e. if you expect something on serial, did you make sure serial was enabled during rootstock with the --serial option ? Sep 22 08:56:03 apt-get purgs Sep 22 08:56:07 *purge Sep 22 08:56:12 ok Sep 22 08:56:24 (why do people always use aptiture is beyond me) Sep 22 08:56:43 yes I use --serial ttyS2 Sep 22 08:56:52 k Sep 22 08:57:08 and thats also the console you are looking at ? Sep 22 08:58:00 yes Sep 22 08:58:25 ogra: Hi! For a BB xM with possible memory corruptions issues, is "mem=256M" a reliable workaround? Sep 22 08:58:39 again... my kernel command line is this: rw console=ttyS2,115200e8 vram=5M omapfb.vram=2560K,2560K root=/dev/mtdblock7 rootfstype=yaffs2 lcdType=QVGA mpurate=600 omap_vout.vid1_static_vrfb_alloc=y mem=227M Sep 22 08:58:45 alf_, yes, according to some testers Sep 22 08:59:20 * ogra never had these RAM probs on his XM so i cant tell for sure Sep 22 08:59:37 ogra: Great, thanks. Hopefully this will make my BB xM experience more deterministic :) Sep 22 08:59:45 heh Sep 22 09:00:05 i think ricardo uses that option on his XM too Sep 22 09:00:19 ok now (w/o ureadahead) I've only this error in the boot.log Sep 22 09:00:20 init: procps main process (3090) terminated with status 139 Sep 22 09:01:27 thats an issue with your kernel config Sep 22 09:02:13 yes, I'm investigate on this... Sep 22 09:03:43 * ogra recommends pulling the ubuntu kernel deb and looking at the config file inside Sep 22 09:03:59 enabling devtmpfs is also helpful Sep 22 09:04:14 and there are a good bunch of other options i guess Sep 22 09:05:15 ogra, the next statement after the "init: ureadahead main process (487) terminated with status 5" is "Last login: " ... does this need any n/w support ?? Sep 22 09:05:43 "Last login: " ?? Sep 22 09:05:55 that looks like you have some kind of autologin going on Sep 22 09:06:03 its tells ... Last login time ... and other things right ?? Sep 22 09:06:08 that line comes from motd iirc Sep 22 09:06:19 yeah ... ubuntu @ ubuntu is logged by default ... Sep 22 09:06:51 n/w = network ? Sep 22 09:07:23 yeah ... network ... :-) .. sorry for confusion. Sep 22 09:08:24 network is only needed if you dont have an rtc battery or dont use the fixrtc cmdline option ... to avoid fsck on boot if your last mount time is worng Sep 22 09:08:46 but apparently you dont get an fsck so it all seems fine to me Sep 22 09:09:54 i am on AC supply ... so no rtc battery ... and i don't have fixrtc in my bootargs .. may be this is causing the erro r ... Sep 22 09:10:05 which error ? Sep 22 09:10:34 * ogra doesnt see an error in any of your questions above Sep 22 09:10:35 not error .. the hang .... it hangs after the print "init: ureadahead main process (487) terminated with status 5" ... Sep 22 09:10:49 i thought it gives you an autologin Sep 22 09:11:18 yeah ... it give autologon normally ...but when i disable network in menuconfig, it hangs Sep 22 09:11:50 do you disable the NIC or all network support ? Sep 22 09:12:06 indeed, as any linux system ubuntu needs a loopback device Sep 22 09:12:16 so you should have basic networking enabled Sep 22 09:13:20 Networking support ---> Networking options ---> TCP/IP networking .... whats NIC ?? and loopback ??? Sep 22 09:13:42 ugh, why do you play with these options if you dont know what you do ? Sep 22 09:13:50 NIC = network interface card Sep 22 09:13:57 (your network card driver) Sep 22 09:14:27 and loopback is the lo interface lostening for localhost connections on the virtual 127.0.0.1 address Sep 22 09:14:31 *listening Sep 22 09:14:50 there are some stability issues ... like system crashes @ rt_worker_func once in an hour or 2 ... so this funciton was tracked to some network file ... route.c ... so we wanted to disalbe that ... Sep 22 09:14:52 some apps expect loopback networking to be available to properly run Sep 22 09:16:05 ogra, so no way to boot ubuntu, with network disabled ??? Sep 22 09:16:15 well, i dont know Sep 22 09:16:26 but i know that some apps look for loopback Sep 22 09:16:40 X wont start without it iirc Sep 22 09:16:58 not sure about the low level, we never disable TCP/IP in ubuntu kernels Sep 22 09:17:10 so i cant really tell what would happen then :) Sep 22 09:17:55 thanks for hlep ogra ... :) ...let me give some more time on this ... and see what happens ..:-) Sep 22 09:18:05 heh, ok Sep 22 11:31:43 persia, did you work with dyfet on the hdf issues ? where is that at ? Sep 22 11:33:00 dyfet: ping? Sep 22 11:34:24 * dmart guesses it may still be too early Sep 22 11:35:15 or to late (in care of persia) Sep 22 11:35:21 *case Sep 22 11:35:39 dmart, btw, did you see my ping this morning ? Sep 22 11:36:12 ogra: hmm, guess not. Sep 22 11:36:17 dmart, http://forums.computers.toshiba-europe.com/forums/thread.jspa?messageID=218590𵗞 someone found access to the bootloader on the toshiba Sep 22 11:36:21 ogra: (very faintly) pong Sep 22 11:36:25 heh Sep 22 11:36:33 ooo Sep 22 11:36:41 so it seems to be just a matter of time now :) Sep 22 11:37:11 * ogra already bricked one, so i'm waiting for others to do the hard work until someone figures out how to change the boot options Sep 22 11:37:43 ogra, A little, but doko just -marm'd it, which makes the issue not visible. I believe dyfet was working on a test case for the underlying toolchain bug exposed in h5detect. Sep 22 11:37:57 dmart, ^^^ Sep 22 11:38:17 dmart, If you want to investigate, remove -marm, rebuild, and then look at the detection macros in h5detect. Sep 22 11:38:38 persia, there was discussion about bug #635199 in #linaro (why are you not in there ? :) ) Sep 22 11:38:39 Launchpad bug 635199 in hdf5 (Ubuntu) (and 1 other project) "[armel] the ./H5detect test segfaults when built for armv7 (affects: 1) (heat: 315)" [High,Confirmed] https://launchpad.net/bugs/635199 Sep 22 11:38:42 The code does some things that end up being very ugly for any architecture that can handle more than one instruction set :) Sep 22 11:39:07 * persia doesn't currently have attention to contribute to linaro Sep 22 11:39:19 we have a team wide agreement Sep 22 11:39:27 that everyone will be in both channels Sep 22 11:39:34 persia: well, it's ubuntu too :) Sep 22 11:40:19 Yao Qi's comment is correct, and matches what dyfet was telling me. Sep 22 11:41:20 I should have some spare attention soon, and will be there more :) Sep 22 11:46:40 Hmmm, the .diff.gz is almost twice as big as .orig.tar.gz for hdf5 :O Sep 22 11:46:44 that's a bit scary Sep 22 11:47:44 persia: yes, it doesn't look like a toolchain issue Sep 22 11:48:09 I'm also wondering whether the kernel does something wrong with respect to alignment fixups. Sep 22 11:48:17 (in addition to reporting them as SIGILL...) Sep 22 11:48:42 That's a deeper question than I can answer: best wait for dyfet. Sep 22 11:48:49 sure Sep 22 11:48:51 no problem Sep 22 11:49:02 h5detect is a good way to exercise alignment issues though :) Sep 22 11:49:10 :) Sep 22 11:50:39 wow software-center is really a myth Sep 22 11:51:10 i wonder if the design team is aware that they overengineerd a bit here :) Sep 22 11:51:26 how do you mean Sep 22 11:51:39 more than less /var/lib/apt/lists/*Packages ? Sep 22 11:51:41 well, i'm just testing a few things in it Sep 22 11:51:52 and i dont find it very intuitive Sep 22 11:52:10 i.e. trying to install all of OO.o i indeed click on the metapackage Sep 22 11:52:38 which only gets me an "update now" button (that in the backend seems to run apt-get update) Sep 22 11:53:12 you actually have to click on something like "openoffice writer" and then select the metapackage under "addon packages" Sep 22 11:53:24 hmmm, that sounds weird... surely installing the metapackage that ought to suck in everything? Sep 22 11:53:47 yes, it does, but only if you pick one of the sublevel packages and check a box Sep 22 11:54:39 hmmm Sep 22 11:54:56 I would have thought "install it all" would be a sensible default for inexperienced users Sep 22 11:55:02 heh, yeah Sep 22 11:55:13 i'm probably to experienced to understand it though Sep 22 11:55:17 who knows :) Sep 22 11:55:27 * dmart shrugs Sep 22 11:56:04 ah, seems to actually be a bug :) Sep 22 11:56:05 phew Sep 22 12:21:13 ogra: hi! we finally found a problem in the u-boot currently beeing used (we were searching into MLO, but the issue was u-boot). Basically, u-boot overrides the pin-mux settings done in MLO, and miss some recent pin-mux updates. Who shall I speak to for defining the best way to fix this issue and get it integrated? sakoman? Sep 22 12:24:32 sebjan, 1st a bug against u-boot-linaro 2nd yes, sakoman and 3rd we need to talk to slangasek and jcrigby to include the fixes in the package Sep 22 12:36:28 ok, thanks Sep 22 12:36:54 sakoman: we have this issue with the linaro u-boot used on omap4: bug 645167 Sep 22 12:36:56 Launchpad bug 645167 in u-boot-linaro "omap4: u-boot overrides pin-muxing settings from x-loader (affects: 1) (heat: 8)" [Undecided,New] https://launchpad.net/bugs/645167 Sep 22 12:40:42 eek Sep 22 12:40:51 you filed it in the upstream project :) Sep 22 12:42:42 * ogra triages properly Sep 22 12:45:16 that's ok Sep 22 12:45:25 as long as it indeed affects the git repo :-) Sep 22 12:47:14 lool, well, both, git and the package :) Sep 22 12:47:26 i do care more about the package in maverick though :) Sep 22 12:53:47 ogra: hi... any chance we could upgrade systemtap to 1.3 for 10.10 or this is way too late? it's in universe. Sep 22 12:53:58 lool: ^^^^ Sep 22 12:54:25 ndec, what does maverick have atm ? Sep 22 12:54:31 we can surely try Sep 22 12:54:32 ogra: 1.2 Sep 22 12:54:45 * ogra hopes 1.3 is in debian Sep 22 12:54:58 it gets tricky if it isnt, should be easy if it is Sep 22 12:55:13 ogra: the 1.3 release notes mention they add support for .35, 1.2 is supposed to work up to .34... i am surprised that was not catched since 10.10 is on .35 Sep 22 12:55:20 ndec: Given that systemtap is not seeded, it would seem possible Sep 22 12:55:50 ndec: We could take it from Debian experimental Sep 22 12:55:57 ndec: Did you try the 1.3 as packaged? Sep 22 12:56:26 ndec: if you could copy the 1.3 source into some non-virtual PPA and confirm the source built under maverick produces working binaries, that would help Sep 22 12:56:45 lool, oh, it is in experimental ? awesome Sep 22 12:56:50 lool: no... we just realized that 1.2 does not work on omap4 with .35. right now I am still unsure if it's because of systemtap of because of our scripts. Sep 22 12:56:52 that makes it easy Sep 22 12:57:31 ndec: Would be nice if you could track this in a bug and report results of your testing of 1.3 as packaged Sep 22 12:58:01 ++ Sep 22 12:58:19 lool: ok. will open a bug, and we will do testing today or tomorrow ... so it's still possible to get something in 10.10... wouah.. I thought it was too late. Sep 22 13:03:20 ndec: TBH new upstream releases are not the kind of things we're looking for at this point of time, but if you say systemtap is unusable with maverick kernels, it's a good reason to include it; <201009171327.17913.ubuntu@kitterman.com> has the detailed timeline for universe packages not on any media Sep 22 13:28:34 ndec, i would have set the bug to critical, but then i end up in discussions with lool again about bug status :) "high"+milestone means its very important to have it fixed befoire release, critical would mean it affects everyone using ubuntu (which is not true) Sep 22 13:28:54 mpoirier, yo Sep 22 13:29:13 ogra: wazup ! Sep 22 13:29:32 mpoirier, seems there is another bug with the IEGP2 that affects networking Sep 22 13:29:38 mpoirier: hi... did you get your new wild animal? Sep 22 13:29:53 and it seems to be fixed already for linaro's kernel Sep 22 13:29:57 ndec: yes, thaming it right now. Sep 22 13:29:59 just saw the bug Sep 22 13:30:02 mpoirier, and it seems linaro has a fix for it ... but i have no idea how it would affect other omap3 arches ... see bug 618734 Sep 22 13:30:03 Launchpad bug 618734 in linux-linaro (Ubuntu Maverick) (and 2 other projects) "[omap/igep] No eth0 with Linaro kernel (affects: 1) (heat: 106)" [High,Fix released] https://launchpad.net/bugs/618734 Sep 22 13:30:23 ogra: hold on. Sep 22 13:30:30 (i guess the main question here is why doesnt it work as a module though) Sep 22 13:36:00 ogra: yep, doesn't sound like a correct fix for me Sep 22 13:36:08 ++ Sep 22 13:36:30 but seems to be a workaround if it doesnt affect other oamp3 subarches Sep 22 13:36:37 yup Sep 22 13:36:41 and i'm not sure about the latter Sep 22 13:37:08 ogra: do we have one igepv2 around in our team? Sep 22 13:37:18 nope, only linaro has them Sep 22 13:37:19 seems lool has one Sep 22 13:37:36 got it Sep 22 13:37:36 right and some other folks in the linaro team Sep 22 13:39:23 jo-erlend: maybe you could also help us testing this issue Sep 22 13:39:33 jo-erlend: you said at the other bug that you don't have your eth0 working Sep 22 13:40:04 rsalveti, right. I have both wlan and wired on my board, but none of them is working. Sep 22 13:40:39 for wireless the cause is probably missing firmware Sep 22 13:41:00 for wired it seems that a workaround is to move the config to built-in Sep 22 13:41:13 still need to understand why this fixed the issue Sep 22 13:41:17 what does that mean? Sep 22 13:42:08 ogra: sorry for the delay - took me a while to find a wrothy tree. Sep 22 13:42:18 ogra: It may help with some arch but affect others. Sep 22 13:42:33 yeah, thats what i feared Sep 22 13:43:09 ogra: I'd like to have an explanation for what the real effect of turning off CONFIG_FIXED_PHY has. Sep 22 13:43:35 ogra: i.e, why it's fixing the problem on some arch and not others. Sep 22 13:43:41 ask on the bug, i just stumbled over the fix in linaro (well, lool pointed me to it) Sep 22 13:43:56 ogra: looks like a work around to me. Sep 22 13:44:00 after i had seen the request in the other bug you filed a merge request for Sep 22 13:44:04 rsalveti, "move the config to built-in"? Sep 22 13:44:08 yes, its surely just a workaround Sep 22 13:44:50 jo-erlend: currently the kernel is generating a module for your ethernet adapter Sep 22 13:44:57 the workaround is to move to built-in Sep 22 13:45:05 I have good news - armel cross compiler 4.5 landed in ubuntu (gcc-4.4 is on a way) Sep 22 13:45:46 hey you didnt post that to #ubuntu-motu (but in every other dev channel) Sep 22 13:46:21 rsalveti, ok.. I have no idea how to do that though. :) Sep 22 13:46:30 ogra: I am not on ubuntu-motu Sep 22 13:46:35 'd Sep 22 13:46:40 but all universe developers are Sep 22 13:46:44 ;) Sep 22 13:47:23 jo-erlend: it's easy, we're just discussing if we're able to have this change to our kernel Sep 22 13:47:31 but first we need to understand why it's needed Sep 22 13:48:22 oh, ok. Can you tell me how to do it? :) Sep 22 13:49:19 jo-erlend: you'll need to recompile the kernel, not that hard to do it, but takes a while Sep 22 13:49:41 especially on omap3 Sep 22 13:49:41 if you want, I can easily generate a new uImage for you Sep 22 13:49:54 rsalveti, that'd be nice. Thanks. Sep 22 13:49:57 based on the kernel package I sent you Sep 22 13:50:05 perfect. Sep 22 13:50:28 ndec, lol, you commented on the wrong bug :) Sep 22 13:50:48 * ogra was wondering which steve you are talking to in the systemtap bug Sep 22 14:00:55 ogra, sebjan: I will prepare a u-boot patch for the panda pinmux corrections Sep 22 14:01:19 sakoman, yeah, i saw the bug comment, thanks a loit Sep 22 14:01:21 *lot Sep 22 14:01:41 sakoman: thanks! Sep 22 14:01:49 I was told that omap4 wasn't a priority, so I've kept all omap4 stuff on the back burner Sep 22 14:01:55 huh ? Sep 22 14:01:57 morning/afternoon/evening all Sep 22 14:02:05 its our highes prio in ubuntu Sep 22 14:02:09 really? Sep 22 14:02:10 *highest Sep 22 14:02:15 ogra: from a linaro perpective Sep 22 14:02:21 omap4 is the fuzz :D Sep 22 14:02:22 we're not linaro here :) Sep 22 14:02:30 <- ubuntu.arm Sep 22 14:02:38 understood, but my direction comes from linaro ;-) Sep 22 14:02:48 bah, we should also pay you then :P Sep 22 14:02:56 haha :-) Sep 22 14:02:56 works for me Sep 22 14:03:35 ogra: patch is basically ready, but there is one pad I need to research a bit more Sep 22 14:03:49 ok Sep 22 14:03:52 the panda x-load sets a few pins more than once Sep 22 14:04:02 and not the same each time :-) Sep 22 14:04:03 we'll need to get it into the package too which will add some delay Sep 22 14:04:32 so I need to look at the schematic and see if that is just a screwup or might be intentional Sep 22 14:04:42 sakoman: cool, thanks for working on it :-) Sep 22 14:04:47 ++ Sep 22 14:04:51 yeah Sep 22 14:05:12 you might notice that I have a few other omap patches in my omap4-dev branch Sep 22 14:05:29 they are still being tested and may be tweaked a bit Sep 22 14:05:29 useful for us ? Sep 22 14:05:38 some of them Sep 22 14:05:50 (if we need an upload anyway i would indeed like them in the package) Sep 22 14:06:02 oh, and note that we use linaros u-boot in ubuntu :) Sep 22 14:06:43 yep, to have all these new patches we would need it first to go to linaro's tree Sep 22 14:06:53 then a new release and then a package update Sep 22 14:07:11 I was just giving a "heads up" that new omap4 stuff was pending Sep 22 14:07:24 ogra: can I remove a comment from launchpad... Sep 22 14:07:28 sakoman: cool, will take a look at the tree :-) Sep 22 14:07:35 ndec, nope Sep 22 14:07:50 mistype above -- omap4-exp brnach is the right one Sep 22 14:07:56 branch :-) Sep 22 14:08:18 ok, i dont know where jcrigby pulls from currently though Sep 22 14:08:22 for omap4 Sep 22 14:09:47 I don't wither :-) Sep 22 14:10:03 either! jees I can't type this morning! Sep 22 14:10:17 * ogra hands sakoman some good brazilian coffee Sep 22 14:10:18 7am is too early Sep 22 14:10:27 u-boot stable is just -rc1 plus a few patches for vexpress and mx51 Sep 22 14:10:34 sakoman: coffe man Sep 22 14:10:42 omap is just upstream Sep 22 14:10:52 I'm drinking a brazilian/guatamalan blend (home roast) Sep 22 14:11:14 jcrigby, omap4 :) Sep 22 14:11:52 jcrigby, we'll need a fix for bug 645167 in your package Sep 22 14:11:53 Launchpad bug 645167 in u-boot-linaro (Ubuntu Maverick) (and 2 other projects) "omap4: u-boot overrides pin-muxing settings from x-loader (affects: 1) (heat: 10)" [High,New] https://launchpad.net/bugs/645167 Sep 22 14:12:47 I see that Sep 22 14:15:06 ogra: lool: we tested with systemtap 1.3 from experimental, our scripts compile again... the problem is related to a kernel trace API change that requires systemtap update. even on x86... Sep 22 14:15:46 ndec, can you note that on the bug (especially the successfull test) Sep 22 14:15:50 i'll care for the rest Sep 22 14:16:01 ogra: it's done already! Sep 22 14:16:23 oh, cool, my mail is slow today :) Sep 22 14:16:36 ogra: only tested on OMAP4, but given the problem we have it's very likely that x86 has the same problem Sep 22 14:17:00 ogra: cool ;-) can you please let me know when you pull it in maverick? Sep 22 14:17:10 sakoman: cool, some interesting patches Sep 22 14:17:17 sakoman: new mmc driver, hm Sep 22 14:17:35 saveenv for SDP4430 Sep 22 14:18:56 could work for panda on sd card too Sep 22 14:19:07 but would require raw partition Sep 22 14:19:22 got it Sep 22 14:19:48 right, i guess its mainly for the blaze eMMC Sep 22 14:19:57 yup Sep 22 14:19:59 we plan to support that Sep 22 14:20:07 so we shoudl get these patches too Sep 22 14:24:24 Hi, I'm following https: / / wiki.ubuntu.com / ARM / BeagleServerInstall Sep 22 14:24:33 I do the dd command, but nothing is copied to my sdcard Sep 22 14:25:14 mmcblk0 what is? Sep 22 14:29:44 * ogra curses .... Sep 22 14:29:53 damned, the panda installs to fast Sep 22 14:30:30 i was testing an oem-config fix and indeed i turned away the second where i should have watched now i need to start over Sep 22 14:30:42 (the preparation takes longer than the install actually) Sep 22 14:31:04 ndec, please build slower boards ! Sep 22 14:31:41 ogra: well... this should be easier than the opposite ;-) Sep 22 14:31:56 fredim, you need to adjust the command to match your device name indeed Sep 22 14:35:28 fredim, some MMC controllers name them sdb or sdc or so, some (the real MMC controllers) name them mmcblk Sep 22 14:35:54 easiest is to watch dmesg when you plug in the card Sep 22 14:36:03 that should show the name Sep 22 14:36:14 argh! git meltdown :-( lost some of my patches in omap4-dev :-( Sep 22 14:38:05 ogra, rsalveti: while I recover, you can review the pinmux patch here: Sep 22 14:38:08 http://www.sakoman.com/cgi-bin/gitweb.cgi?p=u-boot.git;a=commitdiff;h=56b18c5248a1478b6d6263e428ada02af174df77 Sep 22 14:38:13 ouch Sep 22 14:38:13 sure Sep 22 14:39:39 sakoman: You might be able to recover them if you can dig into the git repo Sep 22 14:40:02 lool: yeah, doing that now Sep 22 14:40:09 searching for files modified recently in the objects dir for instance Sep 22 14:40:20 otherwise will need to recreate them Sep 22 14:40:42 1 - Plug SDcard in my notebook and use "sudo dd if = / part / to / img / file of = / dev/mmcblk0" Sep 22 14:40:46 2 - take sdcard, place in BeagleBoard Sep 22 14:41:09 lool: I build with OE, so I have a copy of the most recent state of the tree. I would just need to recreate the patch sequence Sep 22 14:41:12 plug in the keyboard and monitor BeagleBoard, then connect the power supply Sep 22 14:41:36 correct? Sep 22 14:43:27 fredim, /path/to/image/file needs to be a real path, /dev/mmcblk0 needs to match the SD card device name in your laptop Sep 22 14:45:47 / Dev/mmcblk0 symbolics Sep 22 14:45:50 / Media / disk files sdcard Sep 22 14:46:03 sakoman: cool, at least it's setting the same values as x-loader Sep 22 14:46:15 3 directories are mounted Sep 22 14:46:18 sebjan: could you also check at http://www.sakoman.com/cgi-bin/gitweb.cgi?p=u-boot.git;a=commitdiff;h=56b18c5248a1478b6d6263e428ada02af174df77 ? Sep 22 14:46:28 ndec: ^ Sep 22 14:46:32 should fix the pin mux at u-boot Sep 22 14:46:48 rsalveti: that was the intent, right? :-) Sep 22 14:46:53 you recommend me some good howto? Sep 22 14:47:14 fredim, so you are sure your device is /dev/mmcblk0 ? Sep 22 14:47:18 yup :-) Sep 22 14:47:44 yes Sep 22 14:48:09 and what is the path to the img file you downloaded ? Sep 22 14:48:14 starting from / Sep 22 14:49:18 sakoman: your patch should apply on top of uboot-linaro-stable which is used by ogra for panda, right? Sep 22 14:49:36 thats the plan, yeah Sep 22 14:49:44 ndec: I would expect so Sep 22 14:50:06 /home/ubuntu-10.04-server-armel+omap.img Sep 22 14:50:07 sakoman: ok. we will give this a try. sebjan will confirm if it works. Sep 22 14:50:26 sudo dd if=/home/ubuntu-10.04-server-armel+omap.img of=/dev/mmcblk0 Sep 22 14:50:29 ndec: thanks! I couldn't test becasue my panda is broken at the moment Sep 22 14:50:46 but nothing is copied to SDcard Sep 22 14:50:55 sakoman: ouch, your es2? Sep 22 14:51:14 rsalveti: yes -- the serial port connector broke :-( Sep 22 14:51:29 I need to do some surgery later today Sep 22 14:51:34 sakoman: what do you mean at the moment ;-) Sep 22 14:51:40 ouch, not good for a u-boot/x-loader developer Sep 22 14:51:53 rsalveti: indeed Sep 22 14:52:03 sakoman: the problem with the wrong pin mux was only affecting bluetooth, not sure if you have the firmware for this. do you? Sep 22 14:52:12 no, I don't Sep 22 14:52:30 so all the more reason for you to test :-) Sep 22 14:52:36 jo-erlend: could you test http://people.canonical.com/~rsalveti/kernel/uImage.igep2-v3 ? Sep 22 14:53:00 sakoman: rsalveti: so we should cherry pick this patch on top of current uboot-linaro-stable, right? this is a better test, no/ Sep 22 14:53:00 jo-erlend: i just removed CONFIG_FIXED_PHY from our config Sep 22 14:53:16 ndec: yep Sep 22 14:53:41 rsalveti: do you build it native or cross in general? Sep 22 14:54:14 currently I'm building most of my things in a native environment Sep 22 14:54:26 my panda es2 6 layer is doing a good work, with an usb disk :-) Sep 22 14:55:26 rsalveti: cool. root FS on SD and you build on USB? or root FS on USB? Sep 22 14:55:31 fredim, are you sure ? the command is definitely ok Sep 22 14:55:39 rsalveti: USB flash stick or USB external disk? Sep 22 14:55:40 ndec: rootfs on usb Sep 22 14:55:45 external disk :-) Sep 22 14:55:51 fredim, oh, you need to make sure the card is not mounted before running dd Sep 22 14:56:03 100gb, 7200 rpm Sep 22 14:56:26 rsalveti: do you have scripts/instructions to install daily .img on USB? it shouldn't be too hard, and we were about to write such instructions... but if they exist already ;-) Sep 22 14:56:37 So, after successfully copying. what is the next path? Sep 22 14:56:41 rsalveti: how long for kernel build? full .deb? Sep 22 14:57:08 ndec, 2.5h Sep 22 14:57:15 ndec: usually 2 hours for omap4 and 3 for omap3 Sep 22 14:57:28 less modules I guess Sep 22 14:57:40 ndec, mount SD, mount USB disk, cp -a ? Sep 22 14:57:45 rsalveti: 2 hours for omap4 means that USB is faster than SD, since we are in the 2.5+ hours range Sep 22 14:57:59 ndec, then add the UUID of the USB disk to boot.scr Sep 22 14:58:12 ndec: currently I'm not using the daily img, just a very simple rootfs built with rootstock Sep 22 14:58:13 and modify /boot/boot.script after booting and run flash-kernel Sep 22 14:58:16 without X11 and stuff Sep 22 14:58:16 ogra: remember I said instructions would be simple ;-) Sep 22 14:58:22 to be just a build machine Sep 22 14:58:27 ndec, that were all instructions :) Sep 22 14:58:48 ogra: can we run the installer from USB. e.g. cp -a everything before running the installer, and then add root=/dev/sda, and let it go? Sep 22 14:58:54 neno Sep 22 14:58:57 ndec, no Sep 22 14:59:07 you need to do the install on SD Sep 22 14:59:13 then do the above Sep 22 14:59:13 ogra: why? Sep 22 14:59:23 because its designed like that currently Sep 22 14:59:24 ogra.. So, after successfully copying. what is the next path? Sep 22 14:59:34 fredim, boot your beagle Sep 22 15:00:18 ogra: if we keep the boot partition on SD, flash-kernel should work as expected. the resize should be able to resize the USB disk too, no. then you should retrieve the UID of the USB and put this in boot.scr. what is missing? Sep 22 15:00:34 sdcard shot the notebook, I put in BeagleBoard, correct? Sep 22 15:00:38 ndec, if we wanted to install to USB we shouldnt have gone with a preinstalled image ... i'll try to either add such functionallity to the exiting preinstalled image in natty or just make the ubuntu installer work Sep 22 15:00:40 orbarron, sdcard shot the notebook, I put in BeagleBoard, correct? Sep 22 15:00:54 sry... orbarron Sep 22 15:01:04 fredim, right Sep 22 15:01:35 sakoman: looking at your patch, we had not identified a misalignment on PAD1_FREF_CLK3_REQ, but on PAD0_FREF_CLK4_OUT ? Can you please confirm? Sep 22 15:01:40 I connect usb BeagleBoard -> notebook? Sep 22 15:01:49 ogra: well... good point... Sep 22 15:01:52 ndec, the prob is that what you want to do is exactly what ubiquity does, i wouldnt maintain a preinstalled image for that (but for the cost of a 10x slower installation) Sep 22 15:02:15 ndec, we can discuss live images with installer for panda at UDS Sep 22 15:02:32 one way or the other i would like to see USB rootfs support too Sep 22 15:02:37 ogra: by the way what triggers the resizer? are you parsing root= and if there is no UUID, you resize and update boot.scr? Sep 22 15:02:41 ogra, I turn BeagleBoard on the notebook via USB cable? Sep 22 15:02:42 its just not clear yet which way Sep 22 15:02:52 ndec, right Sep 22 15:03:19 ogra, or just turn on my monitor and keyboard in BeagleBoard? Sep 22 15:03:25 fredim, depends how your beagle is configured, if you use the lucid image you will likely need a serial console to tell it the right boot commands Sep 22 15:03:33 sebjan: it's different because u-boot is setting it for LED Sep 22 15:03:43 lucid relied on the preinstalled bootloader on the beagle Sep 22 15:04:14 ndec, though boot.scr is updated from a different script than the resizing Sep 22 15:04:42 ndec, we have one srcipt only for configuration, the other one has the dangerous pieces (i.e. resizing) Sep 22 15:04:44 ogra, theoretically BeagleBoard sdcard should ride with, but in my dmesg shows nothing. Sep 22 15:05:00 ogra: ok.. sometimes I boot manually and overide the bootargs and use mmblk0p2... and I could see the resizer was run again! Sep 22 15:05:12 fredim, well, if it boots you should see something on the screen thats attached to the board Sep 22 15:05:28 ndec, then you didnt run the oem-config bit Sep 22 15:05:43 rsalveti, I just have to replace uImage, nothing else? Sep 22 15:05:45 ndec, that removes the resizer (jasper-initramfs) at the end Sep 22 15:05:49 ok Sep 22 15:05:50 thanks Sep 22 15:05:54 rsalveti: ok, for both pins? Sep 22 15:06:13 ogra: no the installer was run only once. but later on when I want to enable console, I change bootargs by hand, and this is when I used mmcblk... Sep 22 15:06:55 ndec, well, it is definitely gone if you rean the configuration tool (the graphical bit) Sep 22 15:06:58 sebjan: yep, FREF_CLK4_REQ is the first led and FREF_CLK4_OUT is the second one Sep 22 15:07:06 there is no way it could start Sep 22 15:07:17 since its not there anymore :) Sep 22 15:08:07 ogra: well, not sure what happens, but if I install the cdimage, then reboot and change root= to mmcblk0p2 by hand, then reboot it fails since UUID has changed. Sep 22 15:08:26 ndec, if you see it after running the graphical installer up to the point where it drops you into a desktop then something is definitely screwed up during your installation Sep 22 15:09:11 jo-erlend: please let me know if the newer uImage works for you, because then we can forward the change to our kernel Sep 22 15:09:17 the last bit of the graphical installer removes jasper-initramfs and re-rolls your initrd Sep 22 15:10:22 if you see it resizing something *after* that (while using the newly created initrd), then we have a bug Sep 22 15:15:45 ogra: can you try this on your side? i am pretty sure that I did this. i don't have a setup ready right now. the only thing I did was booting with setenv bootargs console=ttyO2,115200n8 ro elevator=noop vram=32M mem=460M@0x80000000 mem=256M@0xA0000000 root=/dev/mmcblk0p2 fixrtc, and booting without uinitrd. Sep 22 15:16:14 rsalveti, the first partition keeps getting wiped. I copied your new uImage and the boot.ini file, but it won't boot. Do I still need uInitrd? Sep 22 15:16:49 ndec, i need a working install for that, i'm currently testing oem-config patches, but i guarantee you that there is no tecnical possible way that any part of the resizing tool remains after a successful install Sep 22 15:16:59 jo-erlend: yep, you just needed to overwrite your uImage with this one Sep 22 15:17:15 and you still need to use the uInitrd Sep 22 15:17:27 well, there was nothing to overwrite. All contents of the first partition is deleted every time I reboot the board. Sep 22 15:17:41 jo-erlend: ouch Sep 22 15:18:27 that's definitely a bug, not sure where Sep 22 15:19:05 jo-erlend: Are you rebooting or are you letting the resize process reboot the system? Sep 22 15:19:31 GrueMaster, I'm rebooting. Sep 22 15:20:12 if you still have jasper installed at your system it could be that it's affecting you somehow Sep 22 15:20:15 Let the script run it's course. Part of what it does is copy off the first partition to ram, reformat for proper CHS allignment, and move everything back. Sep 22 15:20:48 ndec, without uInitrd ? Sep 22 15:21:02 ndec, the resizing happens *from* uInitrd Sep 22 15:21:19 its not possible that you saw any resizing at all without using it Sep 22 15:22:18 ogra: ok.. let me do this again. and I will let you know. i am not 100% sure of what I did Sep 22 15:22:35 ndec, in any case the resizer lives inside the uInitrd :) Sep 22 15:23:35 since we keep a backup of the old initrd around it could possibly happen if you use the uInitrd.bak file for booting Sep 22 15:23:46 but surely not without Sep 22 15:25:25 GrueMaster, that might be the case. After the first boot, I had to manually resize rootfs. Sep 22 15:26:48 rsalveti, I still have no network. It sais "usb0: no such device". I tried setting up eth0, but that gave me the same results. Sep 22 15:26:53 jo-erlend: I would suggest you let it boot and monitor the console. It should reboot in ~10 minutes depending on the size & speed of your SD card. Sep 22 15:27:09 and CPU :) Sep 22 15:27:24 ogra: Same as beagle. Sep 22 15:27:24 pnada install just fly by nowadays Sep 22 15:27:27 jo-erlend: please paste me your lsmod output Sep 22 15:27:51 (apart from tehg SWAP creation) Sep 22 15:28:39 rsalveti, I cannot paste from it since I have no network. Is there anything in particular you're looking for? Sep 22 15:28:42 ogra: Is there any way we can have the swap creaded during image build? It is a zero filled file, should compress to almost nothing. Sep 22 15:29:09 GrueMaster, yes, but it will require some heavy changes to the build scripts ... i'm scared ... Sep 22 15:29:15 jo-erlend_nb: try loading module smc911x Sep 22 15:29:30 I've seen the build scripts. Fear is justified. Sep 22 15:29:31 see if it makes any difference Sep 22 15:29:48 rsalveti, just "modprobe smc911x"? Sep 22 15:29:55 jo-erlend_nb: yep Sep 22 15:29:58 GrueMaster, i'm trying to leverage if its better to just live with 2-3min for SWAP creation or risk the build system Sep 22 15:30:09 its very late for such a change Sep 22 15:30:12 rsalveti, no change. Sep 22 15:31:21 jo-erlend_nb: ifconfig -a? Sep 22 15:31:29 that's weird Sep 22 15:31:34 cat /proc/net/dev Sep 22 15:31:44 (lower level than ifconfig) Sep 22 15:33:20 ogra, that only gives me lo Sep 22 15:33:36 then your kernel definitly doesnt know about other interfaces Sep 22 15:38:55 jo-erlend_nb: check dmesg | grep -i smsc9 , see if your kernel says something about it Sep 22 15:39:58 rsalveti, smsc9 or smc9? Sep 22 15:40:48 well. Neither returned any results. Sep 22 15:44:39 rsalveti, ogra, jcrigby: I think I've recovered from my git meltdown, omap4-exp branch should be OK now: Sep 22 15:44:44 http://www.sakoman.com/cgi-bin/gitweb.cgi?p=u-boot.git;a=shortlog;h=refs/heads/omap4-exp Sep 22 15:44:55 feedback/test results most welcome Sep 22 15:45:38 jo-erlend_nb: http://people.canonical.com/~rsalveti/kernel/uImage.igep2-v4 new kernel, with the same changes from linaro Sep 22 15:46:11 sakoman, thanks Sep 22 15:46:46 jcrigby: I'm sure there will be tweaks this week as I spend some time testing and tracking down a few issues Sep 22 15:46:50 sakoman: cool, expansion stuff is now in Sep 22 15:47:22 rsalveti, is there sru info for the panda pinmux problem? Sep 22 15:47:28 rsalveti: it is in, but performance needs improvement and there is an i2c error message that shows up on beagle but not overo Sep 22 15:48:05 rsalveti: I'll be investigating this week, but let me know if you find fixes :-) Sep 22 15:48:31 everybody who is intrested (and jcrigby indeed) i just switched the oma3 images to linaros u-boot as well Sep 22 15:48:37 *omap3 Sep 22 15:48:37 sakoman, latest git looks much better, I was wondering why no new mmc for beagle Sep 22 15:48:51 so we sue it on all images now Sep 22 15:48:53 *use Sep 22 15:49:04 ogra, so now when I break u-boot everything breaks:) Sep 22 15:49:18 jcrigby, we'll be the first to complain, yeah :) Sep 22 15:49:19 jcrigby: bt and wlan doesn't work with current u-boot for omap4, so that should fill the sru Sep 22 15:49:24 jcrigby: something went wonky while I was pushing the pinmux patch Sep 22 15:49:59 ogra: cool Sep 22 15:50:12 test image should be ready soon Sep 22 15:50:22 (its building since a while already) Sep 22 15:50:50 ogra: will this be 20100922.1? Sep 22 15:51:09 or a private test image? Sep 22 15:51:29 jo-erlend_nb: smc911x is the wrong module, smsc911x should be the correct one =\ Sep 22 15:51:53 but the latest uImage should have the correct one as built-in Sep 22 15:52:10 if it works for you than we need to understand why the module wasn't loaded by default Sep 22 15:53:18 GrueMaster, 20100922.1 (i only rebuilt omap3, omap4 will just be copied from 20100922) Sep 22 15:54:59 Ok. I had just finished pulling 0922. Shouldn't take too much to pull with zsync. Sep 22 16:00:42 rsalveti, jcrigby, ogra: let me know whe you are happy with the testing of the pinmux patch and I will submit upstream Sep 22 16:00:52 it would be good to get the fix in this release Sep 22 16:01:18 (v2010-09) Sep 22 16:02:55 sure Sep 22 16:03:23 but I guess we're depending on ndec and sebjan, as we don't have the bt and wlan firmware Sep 22 16:03:49 ndec, so fyi, i tried what you proposed (with and without initrd and also by only switching root= from UUID to the device name) and i do not get any resie here Sep 22 16:04:00 *resize Sep 22 16:04:31 you must have done something really weird or the install didnt finish properly before you made that change Sep 22 16:05:01 ogra: ok... i am sure I have seen the UID change at some point. I had to generate a new boot.scr with the new UID which I took when mounting the SD on my pc. Sep 22 16:05:41 ogra: just installed unity on my laptop... and i am lost... Sep 22 16:05:42 well, as i said above there is no techinically possible way resize runs again after a successful install run Sep 22 16:05:45 rsalveti: We should test it anyways to make sure it doesn't break something else. Sep 22 16:05:57 ndec, i havent tried it since UDS :) Sep 22 16:06:28 ndec, i think you are supposed to use the search bar a lot with it Sep 22 16:06:31 GrueMaster: sure, I just mean that we're unable to test if it actually fixed what we wanted :-) Sep 22 16:06:42 agreed. Sep 22 16:08:12 ogra: http://paste.ubuntu.com/498575/ Sep 22 16:08:23 I believe this should be enough to get the correct headers for the dkms module Sep 22 16:08:32 yeah Sep 22 16:08:32 and I added linaro headers as an alternative Sep 22 16:08:40 right, thats how it should be Sep 22 16:08:45 will push & test Sep 22 16:08:49 let you know about the result Sep 22 16:08:57 then I believe we should be fine Sep 22 16:09:38 ok Sep 22 16:10:03 i'll put uploading them to the archive and filing the bugs on my TODO for tomorrow Sep 22 16:10:14 cool Sep 22 16:10:18 with luck we'Re ready by weekend :) Sep 22 16:10:44 nice! Sep 22 16:10:58 there are still bits to solve with the sw-center stuff, if thats done you can make up some thoughts about a nice icon for the driver install stuff :) Sep 22 16:11:09 we'll need something on the desktop Sep 22 16:11:26 hm, ok Sep 22 16:11:27 oh, and one of the postinsts needs some hack to remove icon and .desktop file Sep 22 16:11:32 (in your packages) Sep 22 16:11:49 ndec, see PM Sep 22 16:12:01 sure, that's fine Sep 22 16:12:20 rsalveti, sakoman, jcrigby: I just tested the patch, on top of the ubuntu u-boot package, and it fixes the BT issue! Sep 22 16:12:34 sebjan, thanks Sep 22 16:12:38 perfect ! Sep 22 16:12:53 sebjan: awesome, thanks! Sep 22 16:13:22 rsalveti, bug #607250 Sep 22 16:13:24 Launchpad bug 607250 in linux-linaro (Ubuntu) (and 2 other projects) "omapdss: VDDA_DAC regulator on IGEPv2 (affects: 2) (dups: 1) (heat: 24)" [Undecided,New] https://launchpad.net/bugs/607250 Sep 22 16:13:47 I need to grab that from ubuntu once it gets there right? Sep 22 16:14:04 rsalveti, grab you latest patch I mean Sep 22 16:14:07 jcrigby: yep Sep 22 16:14:10 jcrigby, it should be in some tree already Sep 22 16:14:19 there was a pull request on the kernel ML Sep 22 16:14:22 you can find it at my tree, the one from the pull request Sep 22 16:14:44 ogra, rsalveti, I have link to your tree Sep 22 16:14:45 http://kernel.ubuntu.com/git?p=rsalveti/ubuntu-maverick.git;a=shortlog;h=refs/heads/fix-igepv2-display Sep 22 16:16:06 jcrigby: do you have an igepv2 around? Sep 22 16:16:15 if so then you could also help testing them Sep 22 16:16:29 rsalveti, no but some other linaro guys do Sep 22 16:16:41 otherwise I'd wait for an answer from our kernel team Sep 22 16:17:12 ok, once I see it in a released ubuntu kernel I'll cherry pick it from there Sep 22 16:17:21 jcrigby: hm, ok, would be nice to test them on top of linaro's kernel Sep 22 16:18:29 rsalveti, I'll see if torez can test it, she has an igep Sep 22 16:18:51 nice Sep 22 16:19:39 sebjan: thanks for confirming your test results! I will submit the patch to upstream u-boot list today Sep 22 16:20:18 sakoman: great, thanks! Sep 22 16:21:47 * rsalveti out for lunch Sep 22 16:22:07 GrueMaster, http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100922.1/ is there Sep 22 16:23:46 Ok, pulling. Sep 22 16:28:37 mpoirier, looking at bug 644828, shouldnt we keep the sound work on bug 637947 ? Sep 22 16:28:39 Launchpad bug 644828 in linux-ti-omap4 (Ubuntu) "BUG: scheduling while atomic: alsa-source/2058/0x00000002 (affects: 1) (heat: 8)" [Undecided,New] https://launchpad.net/bugs/644828 Sep 22 16:28:40 Launchpad bug 637947 in linux-ti-omap4 (Ubuntu Maverick) (and 1 other project) "no sound devices on current ES2.0 boards (affects: 1) (heat: 432)" [High,Confirmed] https://launchpad.net/bugs/637947 Sep 22 16:28:46 sebjan: I just posted the patch to the u-boot list. if you could respond with a "Tested-by" it would help move it into mainline more quickly! Sep 22 16:28:50 it really seems we fragment a lot here Sep 22 16:29:45 ogra: I'm not *actively* working on 644828. Sep 22 16:30:09 mpoirier, but you research the sound problems on panda Sep 22 16:30:17 ogra: i'm concentrating on 637947. Fixing that one will most likely help fixing the other one. Sep 22 16:30:39 yes, that's what I'm doing 12hrs a day. Sep 22 16:31:50 mpoirier, right, i was just wondering if it wouldnt be easier to keep all sound prob info in one bug, i'm not doubting your workhours :) Sep 22 16:32:17 it fells a bit scattered to me Sep 22 16:32:21 sakoman: sure, can you CC me please? Sep 22 16:33:05 ogra: We need several bugs as it is extremely complex to address bugs within other bugs. Sep 22 16:33:16 ok Sep 22 16:33:20 ogra: as long as we all know sound on panda is taking care of. Sep 22 16:34:05 ogra: to the best of my knowledge, there is no way in lp to file bugs under and umbrella subject, that could be 'sound' in our case. Sep 22 16:34:26 ogra: /s/and/an/g Sep 22 16:35:37 you could add a tag but that wont help much Sep 22 16:35:53 what i usually do is just add a comment with a link to the other bugs Sep 22 16:38:24 sebjan: and about that weird issue you saw with our x-loader? the one that could be a compiler issue, any time to debug it? Sep 22 16:39:01 rsalveti: it's the same issue. We were suspecting an issue in x-loader Sep 22 16:39:16 rsalveti: ... since this is where I expected the pinmux to be done... Sep 22 16:39:37 hey all is there anyway to get the dailybuild for 20100914? Sep 22 16:39:40 hm, ok Sep 22 16:40:43 sebjan: done Sep 22 16:40:56 orbarron, nope, thats gone Sep 22 16:41:11 orbarron, why would you want such an old build ? Sep 22 16:42:57 looking for an environment that will be tested thoroughly... will all omap4 stuff :D Sep 22 16:42:59 orbarron: I have a copy for omap4 only. Sep 22 16:43:30 * orbarron has a copy but thought the dailybuilds would be archive somewhere :D Sep 22 16:43:40 But I wouldn't go as far as saying "thoroughly tested". Sep 22 16:43:49 * orbarron did not want to load up a build on another server Sep 22 16:43:49 orbarron, daylies dont recive much testing and such an old daily will have a lot more bugs than a recent one Sep 22 16:44:15 orbarron, use zsync to keep them up to date :) Sep 22 16:45:19 orbarron, the 14 build is way worse than anything recent Sep 22 16:45:56 iirc, oem-config was borked on it pretty badly. Sep 22 16:46:08 no, that was the one after 14 Sep 22 16:46:13 or two of them Sep 22 16:46:23 after that it improved daily Sep 22 16:46:27 mpoirier: are you into making a ubuntu-arm-sound group? Sep 22 16:46:38 actually todays is pretty close to being good Sep 22 16:46:48 (not perfect, but good) Sep 22 16:46:53 * orbarron is looking for an environment that I can send new users to that will 1. work 2. is available to public 3. is tested with MM, GFX, etc.. so far I believe that will be 20100914? Plz correct me if I am wrong Sep 22 16:46:56 rlameiro: what a good idea ! Sep 22 16:47:16 orbarron, dailies are movimg targets Sep 22 16:47:21 mpoirier: i would glady help on what i can Sep 22 16:47:41 orbarron, and we dont test them much, wait for RC Sep 22 16:47:41 rlameiro: but I was wondering if it couldn't be something like linux-omap-sound... Sep 22 16:47:43 I have an idea for a project that will greatly benefit from it Sep 22 16:48:01 opinion anyone ? Sep 22 16:48:07 well, i just have one card to test... Sep 22 16:48:26 orbarron, http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100922.1/ this one is really good, but it still has bugs Sep 22 16:48:40 orbarron, and it will be wiped in two days Sep 22 16:48:41 if someone is into send me more caqrds i will hapily receive them :D Sep 22 16:49:05 mpoirier, i wouldnt make it arch specific Sep 22 16:49:16 once you have the bugs fixed it will be dead Sep 22 16:49:27 while the team moves on to new HW and has new sound bugs Sep 22 16:50:18 ogra is correct - we will work with other processor, ubuntu-sound it is. Sep 22 16:50:36 i think ubuntu-sound exists Sep 22 16:50:39 isnt there already a grup like that?> Sep 22 16:50:45 i thought you wanted something arm specific Sep 22 16:50:50 crimsums group or something? Sep 22 16:50:58 persia usually knows details about such groups Sep 22 16:51:07 ubuntu-sound now exists - go for it. Sep 22 16:51:14 hmm Sep 22 16:51:22 you might clash with an existing group Sep 22 16:51:31 which is probably differently named Sep 22 16:51:36 well, the thing its that arm sound is far away from the other archs.. Sep 22 16:51:42 right Sep 22 16:51:49 ok, i'll fix. Sep 22 16:51:50 thats why it shoudl have arm in its name Sep 22 16:52:11 arm is ok, mmaybe not omap just... Sep 22 16:52:12 and we should find out what that general sound group is :) Sep 22 16:52:27 but on the description explain that for now is omap oriented Sep 22 16:52:34 to work together, since they mainly maintain the packages (alsa, pulse, jack etc) Sep 22 16:52:58 mpoirier: when ogra said to dont lock to archs I think he meant ARM sub archs, like tegra omap, imx5 etc Sep 22 16:52:58 how do I delete my own group ? Sep 22 16:53:09 rlameiro, yeah Sep 22 16:53:55 mpoirier, #launchpad might know Sep 22 16:54:51 if I install lucid, can I still run a "rootstock" for maverick ? Sep 22 16:55:41 devilhorns: You will need to pull rootstock manually. The lucid version has issues iirc. Sep 22 16:55:41 devilhorns: i did it Sep 22 16:55:57 GrueMaster, ok, but it can be done right ? Sep 22 16:56:01 yes. Sep 22 16:56:03 just need to put -dist maverick Sep 22 16:56:15 *--dist maverick Sep 22 16:56:16 ok, looks like I may need todo a reinstall then :/ Sep 22 16:57:22 GrueMaster, thanks Sep 22 17:00:11 #ubuntu-arm-sound has been created Sep 22 17:07:34 persia: ping Sep 22 17:08:29 rlameiro: It is 2am where he's at. Sep 22 17:08:40 ahhh, ok Sep 22 17:08:50 australia? Sep 22 17:09:13 well, someone in here knows how to put an ubuntu bot on a ubuntu related channel? Sep 22 17:09:22 Tokyo. Sep 22 17:09:36 mpoirier, err, an irc channel ? Sep 22 17:10:11 mpoirier, a) we have policies for that, you need to talk to the IRC council ... and b) it would be better to keep the IRC discussion in one place (here) Sep 22 17:10:51 devilhorns, the board powers from USB :-) Sep 22 17:11:18 davidm, gotcha :) just figuring that out now w/ ogra :) Sep 22 17:11:58 so I pretty much have everything then ... except the powered usb hub, which I can proally pick one up @ a local store here Sep 22 17:12:22 devilhorns, yea, I think most folks with a beagle use a powered hub Sep 22 17:12:56 davidm, yea :) he mentioned the power drain from keyboard when using the board via usb-mini ... so I'll have to run out to store today/tonight and grab a powered hub Sep 22 17:12:58 not a big deal Sep 22 17:17:11 devilhorns: Eventually you might want to get a cable like this: https://specialcomp.com/beagleboard/DSCN0401_s.JPG Sep 22 17:17:13 * ogra calls it a day Sep 22 17:17:32 g'night ogra :) thanks again Sep 22 17:17:33 It costs $8. Or you can make one. Real simple. Sep 22 17:17:54 GrueMaster, not really into making my own cables ... they always shoot sparks :P Sep 22 17:18:03 heh. Sep 22 17:20:22 If you get a powered hub, make sure the PS is at least 2A. I have one that is 1000mA and it is just not enough. Sep 22 17:21:05 ahhh good to know, thanks :) Sep 22 17:21:44 And if you can find a powered hub with ethernet, it makes life even simpler. Sep 22 17:22:59 yea, that certainly would make life easier :) Sep 22 17:23:05 For the serial port, I am using a serial cable that came with a motherboard with an internal port header. Sep 22 17:23:55 At any rate, here is a link for a supplier that has good accessories. https://specialcomp.com/beagleboard/order.htm Sep 22 17:24:20 gotcha ... yea I may have one of those floating around here somewhere ... lord knows I have enough MBs laying around here from previous integration work Sep 22 17:24:53 Mine came from an old 486 before it found the e-Recycling bin. Sep 22 17:25:01 hahaha Sep 22 17:36:44 ogra, what exactly do you need from me re the permissions problem, I can get you the passwd stuff, but do you need anything else? Sep 22 17:36:50 I can just drop you the entire archive, or a portion? Sep 22 17:43:47 rsalveti: Have you tried the mini-usb port on the ES2 yet? Sep 22 17:49:14 Nevermind. It isn't even enabled in the kernel. Sep 22 17:49:23 * GrueMaster files a bug. Sep 22 17:50:59 oh neat maverick is frozen frozen?? Sep 22 17:51:49 devilhorns: cool, got your board? Sep 22 17:52:23 rsalveti, yea :) missing some items to actually make it work, but I'll go out tonight and grab em from a local place Sep 22 17:53:05 rsalveti: Can you mark bug 645420 as confirmed? thanks. Sep 22 17:53:08 Launchpad bug 645420 in linux-ti-omap4 (Ubuntu) "OTG port not enabled for OMAP4 (affects: 1) (heat: 8)" [Low,New] https://launchpad.net/bugs/645420 Sep 22 17:53:08 rsalveti, rcn-ee, have you guys built rootstock images with desktops and had oem-config-gtk run correctly, /var/lib/gdm permissions correct etc.? is it just me that is seeing this?? Sep 22 17:58:24 argh, my notification system is still broken :-) Sep 22 17:58:40 devilhorns: cool, now you can test it on arm :-) Sep 22 17:58:48 GrueMaster: let me check Sep 22 17:59:26 Neko: last time I tested I used the console oem-config Sep 22 17:59:55 Neko: what is the issue you're facing? Sep 22 18:00:29 rsalveti, yea, when I get the rest of the stuff I need to be able to run the board :) Sep 22 18:00:59 I can i askt for the same IP address, but still have dhcp activated? for instance, only gets a different address if the default cant be given by thr router? Sep 22 18:01:34 devilhorns: Make sure you use a >=4G SD card. Image is 2.2G atm. Sep 22 18:01:52 GrueMaster, indeed :) Sep 22 18:02:22 * rlameiro spelling sucks Sep 22 18:02:38 Anything I can tell you before you head to the store? Want to make sure you have a proper shopping list. Sep 22 18:03:06 don't think so ... only thing I am really missing is the usb powered hub Sep 22 18:03:13 rlameiro: I have my DHCP server set to assign the same IP based on mac. Sep 22 18:03:19 buy a cheap usb sound device:D Sep 22 18:03:30 that small thingies that cost 3 usd Sep 22 18:03:34 or something Sep 22 18:03:49 GrueMaster: my router isnt so advanced :D Sep 22 18:03:53 rlameiro: Why? Sound works on beagle. Sep 22 18:04:15 eh, not really needed in my case ... not going to be using it for any games or anything :) just need it to build the efl arm packages really Sep 22 18:04:16 GrueMaster: testing :D Sep 22 18:04:17 rlameiro: You mean it isn't running linux? Sep 22 18:05:09 GrueMaster: i almost sureit runs linux, but i dont want to fiddle with it, it haves a telnet connection... Sep 22 18:05:20 devilhorns, well, you could play quake, we have it in the archive and rsalveti just packaged the GL drivers *g* Sep 22 18:06:06 for sure you could use it for games ;-) Sep 22 18:06:07 Or just run with the aalib drivers. :P Sep 22 18:06:37 ogra_ac2, lmao ... who has time for games ? :) Sep 22 18:06:53 heh Sep 22 18:07:21 ogra_ac2: bought another ac100? :D Sep 22 18:07:43 armin76, no, the wlan disconnects often Sep 22 18:07:51 More likely broke another. :P Sep 22 18:07:52 heh Sep 22 18:07:54 * GrueMaster hides. Sep 22 18:08:02 and that android client just attaches numbers it seems Sep 22 18:08:26 GrueMaster, nah, i'm waiting for someone to hack teh bootloader Sep 22 18:08:35 i wont risk that again Sep 22 18:08:43 though it seems they are close Sep 22 18:09:16 (someone managed to pull the partitions down via nvflash) Sep 22 18:09:57 replace with omap4? :D Sep 22 18:10:02 Does Ubuntu on OMAP have a project registered on launchpad.net -- if so what's the URL... (Somehow I think https://launchpad.net/ubuntu/+source/linux-ti-omap is not the right URL) Sep 22 18:10:33 jayabharath, no we dont have a project Sep 22 18:10:58 well, we do, its ubuntu :) Sep 22 18:11:06 Mmm.. I see. But, bugs can be logged at https://bugs.launchpad.net/ubuntu/+source/linux-ti-omap ?? Sep 22 18:11:23 depends what the bug is Sep 22 18:11:41 the bugs should be targetd for teh specific package it is in Sep 22 18:11:45 * jayabharath is new to launchpad and its confusing... Sep 22 18:12:13 just use ubuntu-bug to file them Sep 22 18:12:15 how will me average joe know where to log omap kernel bugs, bugs on omap ppa pacakges... and also on ubuntu packages... please advice Sep 22 18:12:27 they wont Sep 22 18:12:46 the average joe will get an automatic popup if there is an oops Sep 22 18:12:53 which walks you through Sep 22 18:13:07 That presumes board has not reset or network is functional.. :) Sep 22 18:13:11 a developer will use: ubuntu-bug linux Sep 22 18:13:14 for omap3 Sep 22 18:13:27 and ubuntu-bug linux-ti-omap4 for panda Sep 22 18:13:52 even without network it will collect crash info in /var/crash Sep 22 18:14:16 and tell syou wnt to do with it Sep 22 18:14:29 great... Sep 22 18:14:37 * jayabharath looks for a wiki page that tells this Sep 22 18:15:02 i think there is one Sep 22 18:15:15 persia ususally has the urls up his sleeve Sep 22 18:15:23 * ogra_ac2 searches Sep 22 18:15:39 no worries.. I will search for it.. I am sure there is something on ubuntu.co Sep 22 18:15:40 m Sep 22 18:16:40 help.ubuntu.com/community/ReportingBugs Sep 22 18:17:16 * jayabharath bows to ogra_ac2 and thanks him Sep 22 18:18:48 heh, just thank google :) Sep 22 18:46:48 bug 628029 Sep 22 18:46:49 Launchpad bug 628029 in linux-ti-omap4 (Ubuntu Maverick) (and 1 other project) "[maverick] panda ES1.0 does not suspend on beta image (affects: 1) (heat: 167)" [High,Confirmed] https://launchpad.net/bugs/628029 Sep 22 18:47:32 lag: ogra: ndec: this bug was created for ES1, but does any of you know if CONFIG_PM works for ES2? Sep 22 18:48:58 rsalveti: I would just try it. Worst case is that it fails (I hope). Sep 22 18:49:36 I think there was mention of work needed to be done for suspend to work properly and it was waiting on ES2.1. Sep 22 18:49:36 yep, building a new kernel Sep 22 18:49:51 hm Sep 22 19:12:17 hy all Sep 22 19:12:26 i can install ubuntu arm on my samsung omnia i900? Sep 22 19:14:41 It would take a lot of work, but you can try. Bear in mind that Ubuntu is currently working on supporting Arm Netbooks, not cell phones. Sep 22 19:16:18 but Sep 22 19:16:26 ubuntu for cell phones exists? Sep 22 19:16:34 not that I know of. Sep 22 19:28:24 pirearadu: i900? jaunty (9.04) may work if you have kernel Sep 22 19:28:50 yea Sep 22 19:28:52 i900 Sep 22 19:29:02 what kernel hrw? Sep 22 19:29:17 :D Sep 22 19:29:33 pirearadu: ubuntu does not support i900. if you want to run ubuntu on it you need to have kernel which will run on it. Sep 22 19:29:49 pirearadu: if this is still black magic for you then forget about ubuntu on i900 Sep 22 19:29:52 i know that Sep 22 19:30:16 but Sep 22 19:30:21 what kernel i need? Sep 22 19:30:38 the one which works on i900 - thats simple Sep 22 19:30:52 how to get it and from where is other thing Sep 22 19:32:16 we should probably also stop suggesting jaunty, given its only supoprted for another 4 weeks Sep 22 19:33:08 ogra_ac: or generally: if you do not have (list of all 4 supported boards) then you out of luck? Sep 22 19:33:32 nah, then rootstock wouldnt make sense Sep 22 19:33:46 but jaunty will be EOL very soon Sep 22 21:44:32 NCommander, For qreal issues: isn't it best to instantiate for literals and declare correctly for members? I'd think casting ought be avoided (potential for precision loss, etc.) Sep 22 22:02:14 well, out of the channel :D Sep 22 22:02:25 persia: the idea was to make a group really Sep 22 22:02:28 not a channel Sep 22 22:04:07 Why does it need a group? What does it do that isn't achieved by ubuntu-audio and ubuntu-arm? Sep 22 22:04:35 How does this group increase collaboration and cooperation? Sep 22 22:04:37 well, that was the disvution Sep 22 22:04:45 discusion Sep 22 22:06:06 i think that the problem is that, for now, there are some bugs arm related Sep 22 22:06:50 and i dont think there are to much people interested in fix them Sep 22 22:07:22 I'll agree with that, which is why I'm against a new team or channel. Sep 22 22:07:58 Since there are only a few people currently interested, it makes sense to have the discussion in a broader area, so that knowledgeable folks who aren't especially interested can help share their expertise. Sep 22 22:08:03 so, better inter communication beetween ubuntu-sound and ubuntu-arm? Sep 22 22:09:03 i shal tag the bugs to ubuntu-sound also then Sep 22 22:09:05 I'd rather "more overlap between ubuntu-audio and ubuntu-arm", but better communication may help if that fails. Sep 22 22:09:36 Please don't. Sep 22 22:10:15 See https://wiki.ubuntu.com/DebuggingSoundProblems for how to triage the sound bugs. Sep 22 22:11:00 And be aware that the goal of that team is to address everything in the kernel, so ALSA just works, and everything on top of that just works. Sep 22 22:11:44 We may still have some issues with stuff like mad or pd, but it's more likely those will end up being solved by the ARM team or flavour teams, rather than the audio team. Sep 22 22:18:01 persia: but i am not using an ubuntu kernel, i am using rcn-ee one Sep 22 22:19:41 Ubuntu can only support Ubuntu kernels. To the best of my knowledge, rcn-ee regularly pushes patches into Ubuntu. Sep 22 22:20:06 So, if you want to solve this in an Ubuntu context, the target goal should be to get it working with an Ubuntu kernel. Sep 22 22:21:21 persia: is there an easy way to install an ubuntu kernel inside my board? Sep 22 22:21:39 Make sure flash-kernel does the right thing for you. Sep 22 22:21:40 for instance if i install it where is the uImage and uInitrd Sep 22 22:21:50 Then apt-get install linux-omap (or whatever) Sep 22 22:22:08 flash-kernel? Sep 22 22:22:11 Yep. Sep 22 22:22:18 never heard of it.. Sep 22 22:22:20 rlameiro: you can easily install it by giving apt-get or downloading the deb and giving dpkg -i Sep 22 22:22:34 flash-kernel will replace your uImage and uInitrd with the new one Sep 22 22:22:36 rsalveti, Do you know this works for IGEPv2? Sep 22 22:22:40 at the first partition Sep 22 22:22:56 * persia is fairly certain flash-kernel is device-specific Sep 22 22:23:01 i have them on my fat partition Sep 22 22:23:03 should work, I believe, but nobody tested I guess Sep 22 22:23:10 ok Sep 22 22:23:13 here i go Sep 22 22:23:17 testing freenzy Sep 22 22:23:24 if doesn't work, fill a bug :-) Sep 22 22:23:36 rlameiro, If you want to check the script first: install the flash-kernel package, and look at /usr/sbin/flash-kernel Sep 22 22:23:40 I should make a backup :D Sep 22 22:23:47 otherwise install your package, generate the uImage and uInitrd with mkimage and copy them to the first partition Sep 22 22:23:47 (shouldn't that really be in /sbin, rather than /usr/sbin ?) Sep 22 22:27:22 weird Sep 22 22:27:40 flash-kernel isnt neither in sbin or /usr/bin ... Sep 22 22:28:23 apt-get install flash-kernel Sep 22 22:28:51 rsalveti, Maybe rootstock should encourage people to include flash-kernel? Sep 22 22:28:56 persia: i did that already :D Sep 22 22:29:16 And you don't have /usr/sbin/flash-kernel? Sep 22 22:29:19 * persia is very confused Sep 22 22:29:20 no Sep 22 22:29:22 depends, as our current support at flash-kernel is not that good Sep 22 22:29:38 oops Sep 22 22:29:38 rsalveti, Shouldn't we fix that :) Sep 22 22:29:39 sorry Sep 22 22:29:40 and generally rootstock should be recommended for hardwares that we don't officialy support Sep 22 22:29:46 i didnt look at usr/sbin Sep 22 22:29:51 just /usr/bin Sep 22 22:30:10 its a binary Sep 22 22:30:31 rsalveti, Especially as ARM kernels converge, we'll find that the set of supported hardware grows wider than is well-understood. Sep 22 22:31:03 Specifically, the set of supported hardware is the set of hardware supported by kernels in Ubuntu plus any bugs anyone wants to fix :) Sep 22 22:31:59 rlameiro, From what I've heard, you probably want to install linux-linaro-omap rather than linux-omap for IGEPv2, but you might try each, and see which you like better. Sep 22 22:35:06 we have bugs on both Sep 22 22:35:17 Not surprising :) Sep 22 22:35:35 the only difference is that at linaro's the ethernet is working Sep 22 22:35:56 I sent the patches to fix the display, but not applied yet Sep 22 22:36:04 Really? Maybe I should try that: might make the internet on my board work. Sep 22 22:36:24 persia: do you have an igepv2? Sep 22 22:36:38 * persia decides to stop IRCing until there is enough awakeness to not mistype "ethernet" as "internet" Sep 22 22:36:55 rsalveti, Nope: beagleboard Sep 22 22:36:58 C4 Sep 22 22:37:19 but the ethernet fix at linaro's is related with igepv2 Sep 22 22:37:43 rsalveti: there is comming more work from igepv2 side Sep 22 22:37:54 they are to launch their expansion board Sep 22 22:38:04 I'd like to get one for testing, but needs to find a way to buy it Sep 22 22:38:07 LCD/VGA/compositevideo etc etc Sep 22 22:38:27 well, ask them to send 2 or 3 to canonical Sep 22 22:38:34 could be :-) Sep 22 22:38:49 well, true is that it is on their interest Sep 22 22:39:01 sure, will send an email later :-) Sep 22 22:39:20 * rsalveti out for while, football game :-) Sep 22 22:44:47 persia: flash kernel outputs "unsuported platform.... Sep 22 22:45:35 rlameiro, It's a shell script: tell it how to store the kernel for your board, and file a bug :) Sep 22 22:46:02 (this is how platforms gret supported) Sep 22 22:46:42 well, i wish i had the kwoledge Sep 22 22:47:01 Which part of the shell script don't you understand? Sep 22 22:47:12 just started reading it Sep 22 22:47:39 I'm certain you'll be able to figure out most of it, and I suspect that asking here will cause others to help you. Sep 22 22:51:00 persia: http://pastebin.ubuntu.com/498796/ Sep 22 22:51:08 maybe its in here on this case Sep 22 22:51:18 i need to add IGEPv2 to it Sep 22 22:51:45 actually the igepv2 will be under omap.. in that script there should be a "beagle" case Sep 22 22:52:48 http://pastebin.ubuntu.com/498798/ Sep 22 22:52:55 it check for proc cpu info Sep 22 22:52:56 there should be "OMAP3 Beagle Board" check (from cat /proc/cpuinfo) Sep 22 22:52:58 this is mine Sep 22 22:53:17 and in hardware it is not the omap but igep :D Sep 22 22:53:31 that is my cpuinfo cat http://pastebin.ubuntu.com/498798/ Sep 22 22:53:51 yeap mine is that too... i forget where in flash-iamge it happens.. Sep 22 22:54:19 but from what i see it copy it to the NAND, is that right? Sep 22 22:54:27 because i dont want to mess with the nand Sep 22 22:54:29 ... Sep 22 22:54:47 i prefer to copy it by hand to the partition Sep 22 22:54:48 in the beagle case (lucid defaulted to nand) and with maverick mmc is first i think... Sep 22 22:55:09 i'd really use mmc, less likely for bricks by new users... Sep 22 22:55:36 well I dont know shell script, so i will not mess with it.. Sep 22 22:55:43 i think i will copy it by hand Sep 22 22:55:49 If you have a fat partition, you can create /etc/kernel-img.conf file and add UBOOT_PART= entry. Sep 22 22:56:43 rlameiro, here you go.. add the IGEVP2 machine id to this list: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/maverick/flash-kernel/maverick/annotate/head:/flash-kernel#L306 Sep 22 22:57:03 Sorry, /etc/flash-kernel.conf is the file. Sep 22 22:57:40 It contains two lines, one for UBOOT_PART= and one for root=. Sep 22 22:58:07 beagle routine Sep 22 22:58:10 http://pastebin.ubuntu.com/498802/ Sep 22 22:59:56 If you set UBOOT_PART, that routine should just work. The trick would be in ensuring that IGEPv2 boards trigger that routine. Sep 22 23:00:02 rlameiro, this http://pastebin.com/2RWdavzp on top of https://code.launchpad.net/~ubuntu-branches/ubuntu/maverick/flash-kernel/maverick Sep 22 23:00:12 The way flash kernel works now, if you are using our kernel and have /etc/flash-kernel.conf created, the script should work fine. It looks for specific hardware based on grep Hardware /proc/cpuinfo. Failing that, it uses uname -r. Sep 22 23:00:52 ok Sep 22 23:01:06 i will do that in a moment, girl needs atention... Sep 22 23:01:09 rlameiro, So you want to add to the massive case statement near where it says "OMAP3 Beagle Board" an entry for your machine, which calls check_subarch "omap" Sep 22 23:02:43 rcn-ee, That alone isn't enough: one should also add something around line 668 Sep 22 23:03:20 oh right... Sep 22 23:04:48 http://pastebin.com/JUD5ZQpJ ;) Sep 22 23:06:29 http://paste.ubuntu.com/498807/ Sep 22 23:06:38 i almost think for maverick +1 maybe the nand usage in that omap_flash_kernel could be dropped and just default to mmc? Sep 22 23:06:40 rcn-ee: Heh. Same idea. Sep 22 23:07:02 Either one. Now get someone to test it, and get it attached to a bug :) Sep 22 23:07:32 if rlameiro system is still up, he could manually patch it.. and then try running it. .;) Sep 22 23:08:01 rcn-ee, My personal preference is to boot from NAND, and have MMC available for secondary boots for testing, etc. Mind you, I use USB storage anyway. Sep 22 23:08:34 rcn-ee: i came back Sep 22 23:08:37 yeah, it's nice for your own embedded systems, we got tons of rma's with incorrect nand on the beagles. ;) Sep 22 23:08:51 going to open second ssh and edit it with nano Sep 22 23:09:05 rcn-ee, Oughtn't one be able to just boot off MMC and recreate NAND to work around that? Sep 22 23:09:19 rlameiro, Rather than using nano, could you use patch? Sep 22 23:09:22 where do i found my partitions Sep 22 23:09:44 persia: well i can try, but i am not experienced Sep 22 23:09:45 `cd /usr/sbin; sudo patch < /tmp/flash-kernel-patch;` Sep 22 23:09:57 Just download one of the patches before you do that. Sep 22 23:10:04 so i edit a different version and then patch it? Sep 22 23:10:04 yeap... that's exactly how i fixed most of the returns and sent back to the customers. ;) Sep 22 23:11:09 so only this patch? Sep 22 23:11:09 rlameiro, You shouldn't have to edit anything at all. Sep 22 23:11:14 http://paste.ubuntu.com/498807/ Sep 22 23:11:29 yeap, that one is good.. Sep 22 23:14:29 /usr/sbin/flash-kernel: 671: Syntax error: word unexpected (expecting ")") Sep 22 23:15:10 Oops. Add a | before \ on line 25 of patch. Sep 22 23:15:47 missing a pipe lol Sep 22 23:15:47 actually 671, needs a space.. like; "OMAP3 Beagle Board" | "IGEP v2 board") Sep 22 23:17:10 btw, has anyone tried using rootstock on maverick x86 to build a maverick dist with any xfce4 based packages? it worked this afternoon for me, so maybe the old qemu bug has fixed it self? Sep 22 23:17:14 why the "\"? its to tell that the line continues? Sep 22 23:17:39 yes. Makes it more readable. Sep 22 23:18:25 so it stays "|\" or "| \" ? Sep 22 23:18:36 second option Sep 22 23:18:42 space is good. Sep 22 23:20:07 Reversed (or previously applied) patch detected! Assume -R? [n] -R Sep 22 23:20:07 Apply anyway? [n] y Sep 22 23:20:07 Hunk #1 FAILED at 149. Sep 22 23:20:07 Hunk #2 FAILED at 319. Sep 22 23:20:07 Hunk #3 FAILED at 665. Sep 22 23:20:08 3 out of 3 hunks FAILED -- saving rejects to file flash-kernel.rej Sep 22 23:20:19 do i need to make a diff? Sep 22 23:20:43 No. Did you try to edit the patch and apply it to a clean flash-kernel? Sep 22 23:20:52 nope Sep 22 23:21:00 to the alreday patched one Sep 22 23:21:16 Ok. edit the flash-kernel script directly. Fix line 670. Sep 22 23:22:08 You can only apply a patch once. Sep 22 23:22:42 is there a way to revert a patch? Sep 22 23:22:51 -R (with same patch) Sep 22 23:23:07 ok Sep 22 23:23:20 sudo patch -R < Sep 22 23:27:25 U-Boot partition not found, using Nand. Sep 22 23:27:25 The Kernel doesn't fit in flash. Sep 22 23:27:37 so i need to add the partition to the file? Sep 22 23:28:31 assuming it's mmc add UBOOT_PART=/dev/mmcblk0p1 to /etc/flash-kernel.conf Sep 22 23:28:46 Yes. create /etc/flash-kernel.conf with two lines: Sep 22 23:28:46 UBOOT_PART= Sep 22 23:28:47 root= Sep 22 23:29:45 GrueMaster, looking thru flash-kernel, what uses root=? Sep 22 23:30:23 or is the boot.scr creater script pulling that.. Sep 22 23:30:33 might be. Sep 22 23:30:45 It could also be a hold over. Sep 22 23:31:01 it's new in my book, that's why i wondering.. ;) Sep 22 23:31:11 flash-kernel is something I think was pulled from Debian. Sep 22 23:31:58 so what do i put on the root? Sep 22 23:32:01 yeap i remember it well specially when it first came into lucid... (quit touching my flash..) Sep 22 23:32:06 /dev/mmcblk0p2? Sep 22 23:32:18 yeah, that looks good.. Sep 22 23:32:52 fingers cross Sep 22 23:33:03 lol Sep 22 23:33:06 same error Sep 22 23:33:13 do i need to reboot? Sep 22 23:33:15 no... Sep 22 23:33:18 is it the same line? Sep 22 23:33:51 where? on the flash-kernel.conf? Sep 22 23:33:55 no separated files Sep 22 23:34:08 line 671/2 Sep 22 23:36:12 rcn-ee: its like this Sep 22 23:37:53 rcn-ee: i just made into one line and the output is the same Sep 22 23:38:24 i actually think it's line 671... make sure there is spaces: "OMAP3 Beagle Board" | "IGEP v2 board" Sep 22 23:39:37 yes ther is Sep 22 23:39:42 check this line Sep 22 23:40:05 omap_flash_kernel() { Sep 22 23:40:05 if [ -n "${UBOOT_PART}" ]; then Sep 22 23:40:05 echo "Using u-boot partition: ${UBOOT_PART}" >&2 Sep 22 23:40:19 do i need to pass it as an argument ? Sep 22 23:40:37 it pulls that from /etc/flash-kernel.conf Sep 22 23:40:58 like flash-kernel -n /dev/mmcblk0p1? Sep 22 23:41:18 where? i dont see where does it loads the variable Sep 22 23:41:32 flash-kernel runs ". /etc/flash-kernel.conf" the variable gets loaded... Sep 22 23:41:59 line 290/295 Sep 22 23:42:11 config="/etc/flash-kernel.conf" Sep 22 23:43:11 lol Sep 22 23:43:27 i was actually editing kernel-info.conmf Sep 22 23:43:27 lol Sep 22 23:43:37 kernel-img Sep 22 23:44:37 yay Sep 22 23:44:55 rcn-ee: so now reboot? Sep 22 23:45:46 maybe? did it print out everything okay, and is a new uImage in /boot/(mmc1) Sep 22 23:48:45 well it booted Sep 22 23:48:51 but now i dont have network Sep 22 23:49:16 yep uname-a confirms Sep 22 23:49:21 those should have loaded.. i think ubuntu se them as built -in... Sep 22 23:49:30 Sorry, flash-kernel doesn't change that. Sep 22 23:49:42 :P Sep 22 23:49:43 its the linaro kernel 2.6.35.1005-linaro-omap Sep 22 23:49:55 GrueMaster: but the kernel does :D Sep 22 23:51:04 cant do lspci also Sep 22 23:51:11 At any rate, if you can file a bug on launchpad.net against flash-kernel, I'll attach my (updated) patch and we can roll it in. Sep 22 23:51:13 found my notes.. the igepv2 needs the smsc911x as built-in... Sep 22 23:51:35 o mannnnnn Sep 22 23:51:45 so i need to reverse this? Sep 22 23:51:57 No. (not flash kernel at least). Sep 22 23:52:04 yeah Sep 22 23:52:05 something about u-boot setting up the pins and ussing it, therefor it can't be used setup as a module? Sep 22 23:52:09 but my kernell Sep 22 23:52:25 But if you file a bug on it, I can get my patch rolled into the package. Sep 22 23:52:32 rlameiro, you were using my kernel before right? Sep 22 23:52:40 pcilib cannot open /proc/bus/pci Sep 22 23:52:47 yes rcn-ee Sep 22 23:53:18 GrueMaster: i would like to do so, no network on the device :D Sep 22 23:53:37 first check in your /mm1 there might bee a "uImage_old/bak" "uInitrd_old/bak" Sep 22 23:54:05 rcn-ee: flash-kernel said it make a backup Sep 22 23:54:08 flash-kernel should have moved them to uI*.bak Sep 22 23:54:15 i also made a backup before doing this :D Sep 22 23:54:32 2nd you can bypass flash-kernel with adding "FLASH_KERNEL_SKIP=yes" to /etc/flash-kernel.conf (untill upstream has the kernel fix) Sep 22 23:55:00 Or alternatively, you can type "flash-kernel " (i.e. flash-kernel 2.6.35-22-omap) Sep 22 23:56:00 what is the command to rename on the cli? Sep 22 23:56:05 That will move whatever test kernel you are using onto the uboot partition (as long as the kernel is installed in /boot). Sep 22 23:56:10 i just 'mv' to rename... Sep 22 23:56:15 Use mv Sep 22 23:56:19 ok Sep 22 23:58:58 i think it should be the other way around Sep 22 23:59:03 it boot the same kernel Sep 22 23:59:18 and maybe i lost the bak... Sep 22 23:59:37 if you did mv uImage uImage.bak you did.. Sep 22 23:59:56 does your wifi work? Sep 22 23:59:59 no Sep 23 00:00:05 i did first the back Sep 23 00:00:16 but it actually asked something about permissions Sep 23 00:00:20 i just press enter Sep 23 00:00:37 my wifi doesnt work, i need to instal the libertas drivers Sep 23 00:02:46 well, maybe using sudo mv helps... duh!!! Sep 23 00:03:44 so GrueMaster i file the bug and you post the patch? Sep 23 00:03:55 Yes, please. Sep 23 00:07:36 lp #645659 Sep 23 00:07:36 Launchpad bug 645659 in flash-kernel (Ubuntu) "flash-kernel doesn't work with the IGEPv2 board (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/645659 Sep 23 00:08:02 do you need some of my files? Sep 23 00:08:15 i can send them to you GrueMaster Sep 23 00:08:20 dropbox :D Sep 23 00:08:26 "cat /proc/cpuinfo" should be enough.. Sep 23 00:08:35 Nope, this is enough. Now to work my magic. Sep 23 00:09:10 Yes, add /proc/cpuinfo to the comments. Sep 23 00:10:49 Then I can assign it to ogra. Might even attach the patch before he notices. :P Sep 23 00:11:17 you should attach it Sep 23 00:11:28 persia: not if you cast up to double. I should have made that clear :-/ Sep 23 00:12:30 GrueMaster: why wit for ogra ? Sep 23 00:12:51 testing the annoyance factor. Guess I'll play nice. :P Sep 23 00:13:37 Patch attached. If you could now comment that the patch was tested ok, we should be done here. Sep 23 00:15:08 GrueMaster: the patch has an error Sep 23 00:15:22 it misses 2 spaces around a "|" Sep 23 00:15:25 This is the fixed one. I tested it already. Sep 23 00:15:44 case "$machine" in Sep 23 00:15:44 - "OMAP3 Beagle Board") Sep 23 00:15:44 + "OMAP3 Beagle Board"|"IGEP v2 board") Sep 23 00:15:59 I changed it on mine Sep 23 00:16:10 maybe they arent needed, it was rcn-ee idea Sep 23 00:16:22 Hmm. Will check. Sep 23 00:16:29 NCommander, In terms of code clarity, for all architectures, why would you want to do that? Isn't it cleaner to rely on system qreal to be as expected? Sep 23 00:18:15 rlameiro: Ok, fixed (again). Sep 23 00:18:30 rcn-ee: so what is needed to put the kernel working with my hardware? Sep 23 00:18:49 i cant debug some audio issues withou network Sep 23 00:19:16 and i cant file bugs against the kernel/alsa against your kernel... Sep 23 00:19:33 persia: I have received a new XM board. It works and audio is currently in the same state as the beagle. Sep 23 00:19:35 * rlameiro looks at a nice paradox.... Sep 23 00:20:26 rlameiro, according to my notes, i had to move CONFIG_SMC911x=m to y and CONFIG_SMSC911X=m to y.. http://bazaar.launchpad.net/~beagleboard-kernel/%2Bjunk/2.6-stable/revision/90 (and haven't changed that config since..) Sep 23 00:20:48 GrueMaster: http://launchpadlibrarian.net/56291674/IGEP-add.patch Sep 23 00:21:00 it appended the second patch to the other Sep 23 00:21:05 rlameiro: You can always file a bug using apport-cli --save= and it will save the bug report to disk for transferring to another system. Then you can sneaker-net it to your desktop/laptop. Sep 23 00:21:25 GrueMaster: not that, i want my ssh Sep 23 00:21:44 and also test some configs with GUI over ssh -X Sep 23 00:21:49 grr. Sep 23 00:21:50 no network, no gain Sep 23 00:22:53 rcn-ee: so its dependent on the kernel team? Sep 23 00:23:01 keyboard issues. Accidently hit >> instead of > when creating the patch file. Sep 23 00:23:14 phat phingers. Sep 23 00:24:08 hihi Sep 23 00:24:18 fixed now Sep 23 00:24:21 i checked it Sep 23 00:24:46 Well, prepare for a lot of lp-spam. Sorry about that. Sep 23 00:25:04 yeap, they'll need a bug, i'd ping lag or mpoirier to help get it pushed.. Sep 23 00:25:05 its on a different mail :D Sep 23 00:25:28 rcn-ee: so i need to file another bug? Sep 23 00:25:30 lol Sep 23 00:26:48 yeap.. ;) that's why i have my external tree, i don't really wait for anyone.. Sep 23 00:27:18 well, i guess i can do the flash-kernel, file the bug withh apport-cli and then reveret the kernel Sep 23 00:27:52 GrueMaster, Thanks for the confirmation, as unfortunate as the result is. Sep 23 00:28:27 persia: The only bit is the kernel tweaking for enabling the right channels. Sep 23 00:28:35 and one more board down.. ;) so who wants to do the devkit8000, it needs some dss2 fixes.. Sep 23 00:28:36 I mean unmuting. Sep 23 00:28:42 rlameiro, Which kernel are you using? Have you tried both linux-omap and linux-linaro-omap? Sep 23 00:28:50 just linaro Sep 23 00:28:59 should i try the not linaro one? Sep 23 00:29:01 GrueMaster, Should be easy. Up for writing a patch? :) Sep 23 00:29:23 Not tonight. I have a headache. :P Sep 23 00:29:23 rlameiro, Since there are two kernels that ought boot on your board, if you try both, you'll know if one works better than the other for you :) Sep 23 00:29:46 Is this a kernel patch for alsa? Sep 23 00:30:05 rcn-ee, In terms of moving from a module to a built-in, shouldn't udev be loading the modules at boot? I know there's an issue with some storage modules, but those aside. Sep 23 00:30:22 GrueMaster, Would be ASoC, but ends up passing through the ALSA tree. Sep 23 00:31:01 persia, i can't find the link... but i believe it has something to do with u-boot setting up and tieing to the device.. so once the kernel loads... well's it's kinda already setup... Sep 23 00:31:09 Hrm. will have to look at the code. Sep 23 00:31:11 rcn-ee, If you're pushing enough patches to the kernel, you could ask for commit access, rather than maintaining an external tree :p Sep 23 00:31:52 rcn-ee, Ah, Ugh. That's an annoying situation :( My concern is that we might not want it built-in for some omap kernels: can you think of a case where we might not need it? Sep 23 00:32:00 well right now only two little ones are mine, the rest are other beagle users and that big TI sgx thing which i don't have copyright. ;) Sep 23 00:32:30 Having copyright isn't important, if you have license to redistribute. If you don't have license to redistribute, please don't :) Sep 23 00:32:43 (and we are in the omap merge for 2.6.37 i'm submitting those tomorrow..) Sep 23 00:33:06 persia, well what i meant, they are GPL 2.. it's just they are TI's so there is no point in my pushing.. Sep 23 00:34:03 Depends on to where you push :) If you were trying to maintain a perfect Ubuntu kernel, you might find it useful to push to Ubuntu. For pushing to linux-omap proper, indeed, better to just encourage the authors to push there. Sep 23 00:35:35 well the important bits i do ping the ubuntu guys. ;) some of it is expermental and too risky for ubuntu.. (like dspbridge stuff that just went into 2.6.36) Sep 23 00:37:08 persia`404 error W: Failed to fetch http://ports.ubuntu.com/ubuntu-ports/pool/main/l/linux/linux-image-2.6.35-22-omap_2.6.35-22.32_armel.deb Sep 23 00:37:08 404 Not Found Sep 23 00:37:36 cant test it Sep 23 00:41:41 http://ports.ubuntu.com/ubuntu-ports/pool/main/l/linux/linux-image-2.6.35-22-omap_2.6.35-22.33_armel.deb replaces it. Sep 23 00:41:44 Run apt-get update :p Sep 23 00:44:02 well toolate Sep 23 00:44:08 filing for the linaro one :D Sep 23 00:44:16 how do i see the kernel PID? Sep 23 00:49:10 kernel doesn't have a PID. Sep 23 00:49:30 even 0? Sep 23 00:50:03 PID 1 is init, which is usually upstart. Sep 23 00:50:14 I think 0 is undefined, but I don't really understand the specifics. Sep 23 00:50:20 ok Sep 23 00:50:27 always learning :D Sep 23 00:51:54 Yep :) Sep 23 00:55:03 * GrueMaster heads out for a bit. Sep 23 01:01:36 rcn-ee: lp #645684 Sep 23 01:01:38 Launchpad bug 645684 in linux-linaro (Ubuntu) "linux-linaro kernel, doesnt boot with support for the LAN interface (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/645684 Sep 23 01:01:44 another bug :D Sep 23 01:02:03 looking now for the linux-omap one :D Sep 23 01:19:38 rlameiro: this bug is a duplicate of 618734 Sep 23 01:19:44 bug 618734 Sep 23 01:19:45 Launchpad bug 618734 in linux-linaro (Ubuntu Maverick) (and 2 other projects) "[omap/igep] No eth0 with Linaro kernel (affects: 2) (dups: 1) (heat: 108)" [High,Fix released] https://launchpad.net/bugs/618734 Sep 23 01:19:51 this should be fixed with latest linaro's kernel Sep 23 01:20:24 ok, linux omap is going to have another Sep 23 01:20:44 or should i not report this? Sep 23 01:21:56 rsalveti, Except rlameiro just tested with the latest linaro kernel, and it didn't work... Sep 23 01:22:06 persia: not the latest Sep 23 01:22:24 rlameiro, Probably not worth having two bugs: you can mark that bug "Also affects" if you can reproduce in the other kernel. Sep 23 01:22:32 rsalveti, The "latest" isn't in the repos? Sep 23 01:22:37 2.6.35-1006.11 should be the latest, not 2.6.35-1005.10 Sep 23 01:22:51 rlameiro, Would you try .11 :) Sep 23 01:22:56 well, maybe i should do an update before installing that Sep 23 01:23:01 yes i will Sep 23 01:23:04 well, the bug was closed by Launchpad Janitor, so I think the deb should be at our repo Sep 23 01:23:15 but the last linux omap is not working either Sep 23 01:23:22 and this I am sure is the last Sep 23 01:23:43 rcn-ee: for me is weird that changing CONFIG_SMC911x=m to y and CONFIG_SMSC911X=m to y fixes the problem Sep 23 01:23:51 it should work the same way as module and built-in Sep 23 01:25:04 Well, unless uboot does odd things. Sep 23 01:25:26 probably a kernel bug Sep 23 01:25:27 But, yeah, the kernel has a bug, and ought be able to do bring-up whether u-boot did or not. Sep 23 01:25:34 Most certainly a kernel bug. Sep 23 01:25:48 rcn-ee's kernel brings the interface up Sep 23 01:26:00 because the config is set as built-in Sep 23 01:26:01 rlameiro, No it doesn't: it assumes it's up. Sep 23 01:26:17 but I first wanted to understand it better before submitting a patch Sep 23 01:26:38 rsalveti, You're going to end up buying an IGEPv2 :) Sep 23 01:26:53 persia: I'm just sending an email requesting one :-) Sep 23 01:27:25 with linaro and linux-omap kernel, lspci brings up an error Sep 23 01:27:30 and thanks for the flash-kernel patch :-) Sep 23 01:29:24 i didnt made the patch :D Sep 23 01:29:33 it was rcn-ee GrueMaster and persia Sep 23 01:29:49 i just tested it realtime... almost Sep 23 01:30:01 * persia didn't do it :p Sep 23 01:30:26 lp #645689 Sep 23 01:30:28 Launchpad bug 645689 in linux (Ubuntu) "linux-omap cant bring up eth0 on igepv2 board, no network (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/645689 Sep 23 01:30:54 well, no a break before testing the last linaro kernel Sep 23 01:31:03 OK. That should be a duplicate to 618734, which should have both limux-linaro and linux-omap tasks. Sep 23 01:31:55 the kernel source is the same ? Sep 23 01:33:13 mine has more info :D Sep 23 01:36:32 rsalveti: sorry, but the latest version is linux-image-2.6.35-1005-linaro-omap Sep 23 01:36:37 .10 not .11 Sep 23 01:37:10 hm, that's weird, we should be able to find 2.6.35-1006.11 Sep 23 01:37:24 let me find the sources from launchpad Sep 23 01:37:43 the .11 is not supported by canonical Sep 23 01:37:55 it is there, but not supported Sep 23 01:38:07 1006.11 Sep 23 01:38:15 yeap Sep 23 01:38:24 well, i will ty it anyway Sep 23 01:38:28 try Sep 23 01:38:33 It's supported by Ubuntu, and by Linaro, so it ought be safe. Sep 23 01:38:56 https://edge.launchpad.net/ubuntu/maverick/armel/linux-image-2.6.35-1006-linaro-omap/2.6.35-1006.12 Sep 23 01:38:59 lol Sep 23 01:39:09 .12 already. heh Sep 23 01:39:10 now it appears as supported Sep 23 01:39:17 http://launchpadlibrarian.net/56232546/linux-image-2.6.35-1006-linaro-omap_2.6.35-1006.12_armel.deb Sep 23 01:39:23 forget it... problably some synaptic stuff Sep 23 01:39:34 yep Sep 23 01:41:33 rsalveti: sorry about crapy feedback... synaptics deceived me.... Sep 23 01:41:34 :( Sep 23 01:41:48 Take care with terminology :) Sep 23 01:43:33 rsalveti: one thing i did found interestinf is the size of the images compared to the image of rcn-ee's kernel Sep 23 01:43:48 config definitions Sep 23 01:43:50 images from ubuntu where quite smaller Sep 23 01:48:10 ok, it works on the new one Sep 23 01:48:22 at least networking Sep 23 01:48:45 cool Sep 23 01:49:02 for display fixes they are still waiting the kernel team to review and apply Sep 23 01:52:23 rsalveti: do i need to confirm its working? installing it from the repo? Sep 23 01:52:46 rlameiro: if you want to test, I have a deb for you Sep 23 01:52:53 i tested it :D Sep 23 01:52:58 its working Sep 23 01:53:19 network, didnt tried display, dont have hdmi tv here Sep 23 01:53:29 http://people.canonical.com/~rsalveti/kernel/linux-image-2.6.35.4+_2.6.35.4+-43_armel.deb should have the display fixes Sep 23 01:53:37 http://people.canonical.com/~rsalveti/kernel/uImage.igep2-v4 is this kernel with the config changes Sep 23 01:53:56 so if you install my deb and use this uImage, you should be able to have network and display working Sep 23 01:54:13 well, for now i dont need display Sep 23 01:54:29 so you're fine already with linaro's kernel :-) Sep 23 01:54:33 i dont have an WM so its ok for me, what i want is audio working :D Sep 23 01:54:51 then it's another issue Sep 23 01:55:13 yes I know, one at a time Sep 23 01:55:40 :-) Sep 23 01:58:50 rsalveti: lp 645689 Sep 23 01:58:51 Launchpad bug 645689 in linux (Ubuntu) "linux-omap cant bring up eth0 on igepv2 board, no network (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/645689 Sep 23 02:02:41 nice Sep 23 02:02:57 I'd just like to understand it better, like why using it as module doesn't work Sep 23 02:05:23 From what rcn-ee said earlier, I had the impression that uboot is doing bring-up on the device, and hands it off already configured. Sep 23 02:05:39 So by building it in, it's easy to notice it's there. Sep 23 02:06:24 My (blind) guess is that there isn't an entry in the module detection routine, so there's no coldplug event, so udev doesn't know to load the module. Sep 23 02:06:44 That said, I suspect there also aren't the necessary module mappings so that udev would be able to load the right module on detection. Sep 23 02:06:56 well, igep is configured to have the device ready for netboot Sep 23 02:07:18 maybe it stays configured Sep 23 02:07:56 hm, ok, but then driver needs to be initialized correctly Sep 23 02:07:59 Separately, if the code isn't written to be hotplug-compatible, modload isn't going to be able to bring it up from scratch (even with coldplug), because the module initialisation routines will be missing the right bits. Sep 23 02:08:22 we have platform data for it at the board file, and the kernel should be able to request the module Sep 23 02:08:29 rsalveti, Yep. Precisely. Cleaning up the initialisation and reporting should enable coldplug. Once coldplug works, hotplug is trivial. Sep 23 02:09:02 still prefer to debug it further than changing the config file Sep 23 02:09:22 but, have no board ( Sep 23 02:09:23 :-( Sep 23 02:09:25 rlameiro, the size difference might be from all the omap4 stuff i have enabled to.. The sgx and dspbridge modules do add a little extra Sep 23 02:09:47 hmm Sep 23 02:10:27 rsalveti, Isn't there a way to pass a boot option to get extra debug information in dmesg from coldplug initialisation? Would that be enough to understand the nature of the issue? Sep 23 02:12:53 not sure, would first check if we have everything write at the board file Sep 23 02:13:03 that's where the platform_data is stored Sep 23 02:14:38 and afaik the ethernet chip is deployed with the usb hub Sep 23 02:14:40 like we have for xM Sep 23 02:23:49 hm, LAN9221i is just an ethernet controller Sep 23 02:24:31 rlameiro: where do you get your u-boot? Sep 23 02:25:52 stock Sep 23 02:26:10 rlameiro: do you have the source's link? Sep 23 02:26:12 do you want the version? Sep 23 02:27:02 rsalveti: you need to register at their site Sep 23 02:27:04 http://www.igep.es/index.php?option=com_content&view=article&id=46&Itemid=55 Sep 23 02:27:29 there should be a link on the left menu where you can go to support or downloads section Sep 23 02:27:38 there are all the sources that igep use Sep 23 02:27:46 kernels, modules etc Sep 23 02:29:13 hm, ok, thanks Sep 23 02:29:51 even if i give you the direct link, you will need to be registered... Sep 23 02:29:54 sorry Sep 23 02:31:30 argh, still need to wait for the administrator to approve my account Sep 23 02:48:36 cooloney: hey, morning! :-) Sep 23 02:48:43 maybe you can help me Sep 23 02:49:12 I believe that the smsc911x module is not automatically loaded simply because this is a platform_driver Sep 23 02:49:27 it's not connected to any pci or usb Sep 23 02:50:10 what we have at the board file is the platform_data, that's correct registered for igepv2 Sep 23 02:50:44 but I don't think we have an automatic way to find and load the proper modules while booting the kernel Sep 23 02:50:54 that's why it works since boot when you have it as built-in Sep 23 02:51:38 then we have a platform_driver_register that matches with a platform_data, the hardware is then initialized correctly Sep 23 02:52:01 rlameiro: persia: ^ Sep 23 02:52:23 so loading the module should do the same, so it's expected to work Sep 23 02:52:33 well, so the solution is the driver to be builtin? Sep 23 02:52:52 but then we have another issue currently with ubuntu's kernel Sep 23 02:52:58 how do you load a module without detecting it? Sep 23 02:53:02 that amitk pointed me: http://www.spinics.net/lists/netdev/msg140340.html Sep 23 02:53:12 and that's why we have to remove CONFIG_FIXED_PHY Sep 23 02:53:34 so, removing CONFIG_FIXED_PHY and loading the kernel by hand should be enough to have the ethernet chip working Sep 23 02:53:48 rlameiro: scripts, but should be avoid Sep 23 02:54:16 I could be wrong, but I don't know a way to detect the platform devices Sep 23 02:54:41 all upstream config files put it as built-in Sep 23 02:55:02 rsalveti: built-in FIXED_PHY? Sep 23 02:55:45 cooloney: currently we have it, and as a side effect we can't make smsc911x to work Sep 23 02:55:53 cooloney: http://www.spinics.net/lists/netdev/msg140340.html Sep 23 02:56:16 rsalveti: yeah, i removed the FIXED_PHY in fsl-imx51 ubuntu kernel before Sep 23 02:56:22 cooloney: a question, is there a way to auto detect the platform devices and trigger user space events so udev could load the proper modules? Sep 23 02:56:24 due to the conflict Sep 23 02:56:48 or in this case we should always put it as built-in Sep 23 02:57:10 cooloney: cool, we're missing this currently Sep 23 02:57:20 will request the linaro's patch to be applied to our tree Sep 23 02:59:41 rsalveti: i think for our hardware which contains smsc911x needs to built-in smsc911x driver and remove that FIXED_PHY **** ENDING LOGGING AT Thu Sep 23 02:59:58 2010