[09:05] cooloney: I updated the bug with another defconfig [09:05] cooloney: with this defconfig I got the musb working with maverick kernel, now just need to identify what needs to be changed to make it work by default [09:06] rsalveti: awesome [09:06] rsalveti: let me check [09:08] cooloney: but is a little painful to recompile it, install and test :-) [09:08] rsalveti: is that defconfig the omap3_beagle_defconfig? [09:08] rsalveti: yeah, i need to compile it in our server [09:08] cooloney: http://bazaar.launchpad.net/~beagleboard-kernel/+junk/2.6.35-devel/annotate/head:/patches/defconfig [09:09] cooloney: I tried one time merging the usb and musb stuff, but didn't get it to work [09:09] so it must be something else [09:10] rsalveti: yeah, i also tried that, set the usb and musb config as the same as those in beagle_defconfig [09:10] but doesnot help [09:12] cooloney: you can try to trace the code, with debug activated, from the version that works [09:12] and then compare with the version that doesn't work [09:12] the state tracing is easy to see inside the code [09:26] cooloney: yep, for some reason the otg state changes from one config file to the other at the point you put the cable in [09:27] the path to set it to a_idle is simple, the question is why the checking failed and it got into that path [09:32] rsalveti: exactly, looks like there is no power on the OTG port [09:32] rsalveti: so it never change to the right a_host mode [09:33] and failed to enable roothub to enumerate USB device [09:33] yep [09:33] rsalveti: i am also comparing the defconfig from you with our Maverick config [09:38] cooloney: ok, I'm now looking again what I need to have for the zippy 2 support [09:38] our kernel is missing lots of patches, it seems === kmargar is now known as markos_ [09:50] NCommander: http://paste.ubuntu.com/466880 [09:51] rsalveti: could you please help me to post the dmesg which is from the OTG working Maverick kernel? [09:52] rsalveti: i think you cross build it with the working defconfig [09:54] cooloney: just the dmesg? [09:54] cooloney: just a sec, need to change the uImage [09:54] rsalveti: oh, the uImage is the better. hehe [09:54] rsalveti: np, man [10:15] cooloney: http://rsalveti.net/tmp/test-kernel.tar.bz2 [10:15] cooloney: you can find the uImage, uInitrd, modules and dmesg.txt [10:15] from the working maverick kernel [10:16] cooloney: if you still didn't compile it [10:17] rsalveti: thanks a lot, man [10:20] rsalveti: i saw this in the dmesg [10:20] musb_hdrc musb_hdrc: MUSB HDRC host driver [10:20] [ 0.933654] musb_hdrc musb_hdrc: new USB bus registered, assigned bus number 2 [10:21] [ 0.934417] hub 2-0:1.0: USB hub found [10:21] [ 0.934448] hub 2-0:1.0: 1 port detected [10:21] cooloney: yep, the otg recognized that it should act as host, and activated the host modue [10:21] that's we want [10:24] cooloney: yep [10:26] cooloney: and with this kernel I'm unable to reproduce the stack we have with lucid [10:26] so it seems to be working better [10:36] hi rsalveti... I tryied your new release of rootstock from bzr ( rev 104) but I have the segmentation fault with qemu again :-( [10:36] furibondox: more than one time? [10:36] yes [10:36] argh [10:37] every time I run rootstock I have the segmentation fault [10:37] furibondox: what packages are you giving by default? [10:37] ubuntu-minimal,openssh-server [10:37] furibondox: did you run as root or as user? [10:37] as root [10:37] rsalveti: yeah, for the otg oops, i am pretty sure we need to build int twl4030_usb [10:37] from ubuntu lucid server [10:38] furibondox: what ubuntu version are you running? [10:38] 10.04 [10:38] rsalveti: and the OTG host failure is another problem. [10:38] cooloney: yep, sure [10:39] furibondox: that's really weird, as I saw it break just a few times when running as root and installing a small package set [10:39] furibondox: does it also break when you just request ubuntu-minimal? [10:40] that will avoid the apt-get install phase [10:40] it breaks in the second stage [10:40] furibondox: yeah, I know, and probably when trying to give apt-get install on the openssh-server [10:40] furibondox: do you have the log? [10:40] just after python-support [10:40] furibondox: you can paste it at paste.ubuntu.com and set me the link [10:41] ok [10:42] * rsalveti needs to update rootstock to run only as user mode emulation when running as root [10:42] that would avoid this kind of weird full emulation stuff, until we get qemu fixed upstream [10:42] http://paste.ubuntu.com/466907/ [10:43] that is the log generated [10:45] this is the output of the error in the console: http://paste.ubuntu.com/466909/ [10:48] furibondox: what's the whole command line you passed to rootstock? going to check here [10:49] ./rootstock --fqdn NGP --imagesize 2G --seed openssh-server --copy-package-cache --restore-package-cache [10:55] furibondox: why do you want to copy and restore at the same time? [10:55] furibondox: you can just copy one time and later you can use restore if you need it [10:55] going to test here, but it should work fine [10:57] just to be sure that if some packages are not already present they will be cached [11:03] furibondox: yeah, but it shouldn't make much difference for a basic image, but ok :-) [11:03] yes I know... but it was just to be sure ;-) [11:07] furibondox: please, try giving a user and a password argument [11:07] than the oem stuff is not going to be installed by default [11:07] and you'll probably be able to get your image without seg faults [11:08] --login ubuntu --password ubuntu [11:08] something like that [11:08] ok i try [11:18] rsalveti: now it seems to stop earlier... [11:18] start: Unable to connect to Upstart: Failed to connect to socket /com/ubuntu/upstart: Connection refused [11:18] just a moment that I paste the log... [11:21] http://paste.ubuntu.com/466920/ [11:22] and this is the output of the console: http://paste.ubuntu.com/466921/ [11:28] furibondox: yep, but it did finish correctly [11:30] mmm I thought that error compromised the generated image [11:47] rsalveti: do you know where can i find the initrd.gz after i build a kernel package which contains zImage [11:52] rsalveti: the error appear again using username/password http://paste.ubuntu.com/466941/ [11:54] furibondox: this log is telling that rootstock is installing the oem stuff, and that only gets installed when you don't put the user/passwd [11:54] I: Enabling firstboot configuration [11:54] ^ [11:54] this is the full log generated http://paste.ubuntu.com/466943/ [11:55] I'm going to lunch [11:55] see you later... bye [11:56] furibondox: still, it seems that didn't got the user/passwd [11:56] if [ ! "${NEWUSER}" ] || [ ! "${PASSWD}" ];then install oem stuff [11:56] furibondox: see ya [11:57] cooloney: you mean, inside the package? [11:58] rsalveti: i don't think i can find it in the package [12:30] re [12:30] rsalveti: now I try with this command line: [12:30] ./rootstock -f NGP --username spot --password gilbarco --seed openssh-server,ubuntu-minimal -i 2G --serial ttyS2 --doswap --swapsize 256 --script /root/setup.sh --restore-package-cache [12:48] same error: ./rootstock: line 360: 967 Segmentation fault qemu-system-arm $QEMUOPTS -append "${APPEND}" > $QEMUFIFO 2>&1 === XorA|gone is now known as XorA [13:15] furibondox: --login not --username [13:15] * rsalveti still needs to change how rootstock parse it's arguments [13:19] ogra_cmpc: can you tell me what offset I need to loop mount the correct partition in the 20100719 daily image? [13:20] my naive calculation based on fdisk -l output seems to have been incorrect [13:25] ogra_cmpc: n/m, sorted it; works better when I use sectors as units [13:30] slangasek: Perhaps you can kpartx it? [13:31] lool: nah, sorted [13:31] yeah, sectors helps [13:32] * ogra has some livecd-rootfs code ready to shove off 2h buildtime just waiting for a lamont signoff [13:32] that might help a bit with building test images [13:33] slangasek, btw, i wouldnt mind if we started building base images, would that help you ? [13:34] (in parallel to the netbook ones) [13:34] don't see that it would help [13:34] k [13:35] hrw: hi! I tried to apt-get upgrade the gcc-4.4-arm-linux-gnueabi package, and get an error because the libmpfr4 'is not installable' (I have a i386 workstation). Any idea what is going on? [13:35] rsalveti: argh! sorry... now I try with --login [13:35] furibondox: yep, np, there's a bug on the way it's parsing the arguments [13:36] normally it should display you an error when you give a wrong arguments [13:36] i hope that now it will work... [13:50] rsalveti: great! it generates the image!! [13:50] furibondox: cool [13:50] I've another question for you, is possibile to ADD another repository to the default? [13:51] -m option only replace the mirror or add a new mirror? === kmargar is now known as markos_ [14:05] furibondox: yep, --extra-mirror [14:06] great! [14:07] I'll try [14:30] sebjan: hmm... [14:31] sebjan: 4.4.4-7ubuntu1~ppa2hrw1 version? [14:31] sebjan: and pastebin me whole installation output [14:32] hrw: ok, let me do that... [14:45] persia: MMC hang fix patch is: http://gitorious.org/~robclark/pandaboard/robclarks-kernel-omap4/commit/6c8d7209371b5969ea6769382693019ef6cb5eb5 [14:45] (I don't know if it is in ubuntu kernel yet) [14:48] hrw: here it is, and ensuring the proper version (which is the one I see with an apt-cache show) http://ubuntu.pastebin.com/YqiGpzBa [14:50] sebjan: strange [14:50] apt-cache policy libmpfr4 tells that this is in main [14:50] sebjan: and try with aptitude - it has better dependency solving [14:52] 15 minutes for resizing an 8gb sd card, seems ok [14:52] just missing a progress bar :-) [14:53] yeah [14:54] zumbi: hey [14:54] XorA: hey ... so, when and where does the drinking take place :) [14:56] hrw: not better - http://ubuntu.pastebin.com/06wwXfmS [14:56] hrw: is there a simple way to clean-up all the previous packages I installed (I may have been playing with dpkg when there packages were first available, and did not upgrade since then) [14:58] sebjan: run aptitude, limit to .*-armel-cross and remove all of them [15:03] hrw: apt-cache search or policy does not find libmpfr4 for me... [15:03] Marex: hi [15:03] sebjan: maverick? [15:03] hrw: no I am on lucid [15:03] zumbi: don't tell me you're in prague too [15:04] Marex: no, i am not :) [15:04] zumbi: pity :) [15:04] Marex: did you get a new toy? [15:04] sebjan: http://people.canonical.com/~hrw/ubuntu-maverick-armel-cross-compilers/ [15:04] zumbi: what new toy ? [15:04] zumbi: you mean the one I was speaking last week ? yea [15:04] Marex: the one we were talking about.. [15:05] sebjan: I do not provide packages for mythripk: ping [15:05] zumbi: and Eric gave me an akita and made me (forcibly) a new maintainer for that [15:05] I mean Marex: akita? what's that? [15:05] omg.. akita [15:05] zumbi: there are so many of them I barely recall which one was that :) [15:05] zumbi: sharp zaurus sl-c1000 [15:05] hrw: thanks :) [15:05] zumbi: pile of crap hardware loaded with crap software [15:05] zumbi: pxa272/416MHz 64mb ram, 128mb flash [15:06] hrw: argh, so I would need to upgrade to Maverick now? It used to work with the previous version I had :'( [15:06] antique [15:06] sebjan: you can use chroot [15:07] hrw: right, I'll try with chroot then. Thanks! [15:07] sebjan: I was aksed about lucid packages before but rejected - it takes too much time to build maverick ones [15:09] and my goal is to have package which will build cross compiler. but that would require maverick packages [15:10] hrw: np, I shall have a maverick chroot already, just need to upgrade it :) [15:10] ;D [15:12] Marex: uhm... well.. sounds like fun :) [15:16] zumbi: yea ... not really ... I'm slowly reaching the point where I'll have so many machines I won't be capable of maintaining them all [15:17] Marex: know that... [15:17] Marex: have spare not needed palmos device? [15:17] hrw: I have a palm lifedrive here ... it was still running palmos yesterday [15:17] hrw: why ? [15:18] just asked to know do you still play with them [15:18] hrw: yea, I maintain them in mainline [15:18] hrw: but I'm moving to newer hardware ... pxa3xx stuff and soon iMX [15:19] imx3 or imx5? [15:20] or maybe imx6? :D [15:20] hrw: I dunno ... lemme check [15:21] i.MX25 [15:21] robclark, persia: this mmc fix (6c8d7209371) is not yet into the ubuntu kernel, but is part of the patch set that I have provided to cooloney today [15:23] sebjan: Thanks for chasing that down. [15:23] Marex: which device? [15:24] zumbi: something that still haven't left the manufacturing it seems :D [15:24] zumbi: I don't know the details yet ... I just know this board is in planning [15:25] Marex: but imx25 is old [15:25] well, i guess it has its use [15:25] sebjan: yeah, thx, i'm looking at your branch [15:26] persia: no pb, I just saw a commit pass by, and I care of getting all the necessary ones in our kernel :) Feel free to notify/ping me when there is a doubt [15:29] zumbi: I wasnt looking into imx yet ... I will when I have the board [15:30] Marex: imx5, i had a NDA datasheet of 5000 pages.. fun! :) [15:39] rsalveti: how do you think we firstly built-in twl4030_usb in kernel to kill the oops [15:39] rsalveti: and mark the OTG config bug as fixed, and file another bug to let us work on the host mode issue [15:40] cooloney: didn't you change the config file to let it as built-in already? [15:40] the kernel you posted on the bug [15:40] cooloney: also, the bug was reported for lucid, so we still need to check at lucid if that fixes the problem [15:41] rsalveti: yeah, i did that for testing, the patches are not in our official kernel tree [15:41] i am going to do that, [15:41] i think we need built-in that, and enable OTG === JaMa is now known as JaMa|off [15:45] cooloney: yep, but I tested your testing kernel at lucid and it didn't work [15:45] the ops was gone, but as a side effect the host mode don't work anymore for otg [15:45] so it actually doesn't fix the bug [15:46] we can't say we're fixing something when we're removing features :-) [15:46] cooloney: but I agree we should create a different bug [15:47] cooloney: I can do that from my beagle [15:48] rsalveti: i agree, we need a different bug. [15:51] cooloney: will create it soon [15:51] let me just boot maverick again [15:52] but for the oops, we can still mark that Fix, since when we got oops, the host mode does not work at all [15:52] we can point it out to the new bug in OTG config launchpad log [16:06] cooloney: yeah, but some times it does work [16:06] most of the time when you add the cable later on [16:07] so just changing the config file will kill this "feature" [16:07] we go from something that works sometimes, to something that doesn't work at all [16:12] yeah, so new bug for the host mode missing [16:12] rsalveti: and the old one is a oops fixing. [16:12] now it looks to me like some timing issue. [16:13] cooloney: just opening the bug right now [16:13] rsalveti: ok [16:19] I hope that one day libelf-dev libelf0g-dev conflict/transition will be done [16:24] NCommander, cn you bump https://edge.launchpad.net/ubuntu/+source/update-notifier/0.103ubuntu1/+build/1883485 a bit higher ? [16:24] (so i can test the image build speedup (its bocking atm)) [16:24] *blocking === fta_ is now known as fta [16:32] cooloney: bug 608312 [16:32] Launchpad bug 608312 in linux (Ubuntu) "Usb host mode on OTG doesn't work on Maverick with BeagleBoard (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/608312 [16:34] rsalveti: thanks a lot, man [16:37] np === fta_ is now known as fta === XorA is now known as XorA|gone [18:20] packages.ubuntu.com/lucid as it stands doesn't include any notation of arm which packages are built for arm ... is there another webpage or command line tool where one can see a list of what is built for arm? [19:40] tgall_foo: do you want something like http://ports.ubuntu.com/ubuntu-ports/pool/main/? [19:40] there are lots of arm packages over there [19:45] rsalveti, was hoping there might be something like packages.ubuntu.com that listed arm packages instead of having to pull down the individual package text files from like : http://ports.ubuntu.com/dists/maverick/main/binary-armel/ [19:46] tgall_foo: oh, ok [19:46] you can try from the repository itself, but I don't think it's on-line in any other way [19:47] at least I never saw it [19:48] tgall_foo: the packages files are easily parsed using grep-dctrl [19:49] ukleinek, yeah I'm probably being picky .. it's not like text files are hard to deal with === kmargar is now known as markos_ === JamieBen1ett is now known as JamieBennett === fta_ is now known as fta === JaMa|off is now known as JaMa