[00:01] the patch introduced a MAYBE definition, sets {v5,v7,thumb}_supported to MAYBE, then uses ifs to turn those true or false [00:01] except it doesn't ever set v7_supported to false on non-v7 [00:02] ah [00:02] so it stays at MAYBE, and later in the code, if (v7_supported) { [00:02] explode_horribly(); [00:02] } [00:03] 14 hours of building to add one printf: totally paid off ^_^ [00:03] MCS [basic] System.Xml.dll [00:03] features: v5: 1, v7: 2, thumb: 1 [00:03] Illegal instruction [00:03] heh, ther wonderful world of arm :) [00:04] -r [00:04] this is why i wish i could make qemu work - i reckon my core i7 would be a pretty fast ARM [00:04] nah [00:04] qemu is always slow [00:05] better invest $120 and buy a beagleboard [00:05] or wait for the pandaboard [00:05] i'm not investing my money in ARM - my concern is for debbuntu [00:05] (dual core 1GHz, 512M similar price) [00:07] ogra_cmpc, i'm trying to work out the most minimal way to fix this. can you sanity-check my idea? [00:07] let me pull up a URL to the file in question [00:09] http://git.debian.org/?p=pkg-mono/packages/mono.git;a=blob;f=mono/mini/mini-arm.c;h=84caef95c1af83bc70cebab8826485a67b89d963;hb=HEAD [00:09] wait, no [00:09] wrong branch [00:10] http://git.debian.org/?p=pkg-mono/packages/mono.git;a=blob;f=mono/mini/mini-arm.c;h=9593f9951510d0dacfc8f5b3a65307476b01a197;hb=refs/heads/master-experimental [00:10] okay, now around the line 545 mark, you see the iffing to assign values to {v5,v7,thumb}_supported [00:11] the defaults here are MAYBE (i.e. all features assumed to be on) [00:12] the assumption is to help qemu, i.e. "if you get no values, just roll with everything on, it'll probably be fine". which given our dicussion earlier, is probably the case [00:13] so why not just insert "v7_supported = FALSE;" between lines 553 and 554? [00:14] sounds ok to me [00:14] test it :) [00:15] yeah, just ran make. see you tomorrow! [00:15] mpoirier: On bug 591941, if I remove the SD card, then reinsert it and restart ./init, it seems to work ok. Might want to explore that a little when you work on that bug. [00:16] Launchpad bug 591941 in linux (Ubuntu Maverick) (and 1 other project) "SDHC card not recognized (affects: 2) (dups: 1) (heat: 171)" [High,In progress] https://launchpad.net/bugs/591941 [00:16] my gut says "apologise to lool", let me check something in the git logs... [00:17] yup, apologies to lool - seems it was meebey that broke it when porting lool's patch from mono 2.4 to mono 2.6. i shall go and poke him [00:23] * mattman_ is away: [00:28] so by the looks of it the resizing routine runs now [00:30] MCS [basic] System.Xml.dll [00:30] features: v5: 1, v7: 0, thumb: 1 [00:30] System.Xml/XmlTextReader.cs(1811,40): warning CS0219: The variable `dummyValue' is assigned but its value is never used [00:30] looks healthy! [00:30] great [00:30] i'll try to push a 2.6.3-3 to experimental tomorrow [00:30] then merges can happen based on that [00:31] sweet [00:47] hang on, main's frozen... so no hurry! === bjf[afk] is now known as bjf [01:31] I'm eyeing that "port glxgears to ES" bit... wanna' try it on libgles1-mesa-drivers-kms (or whatever it is) on my netbook. === bjf is now known as bjf[afk] [05:03] * mozzwald is away: sleeping === bjf[afk] is now known as bjf === bjf is now known as bjf[afk] === hrw|gone is now known as hrw [08:09] morning [08:10] morning [08:12] morning [08:13] morning [08:13] :) [08:14] Who's responsibility is this: http://paste.ubuntu.com/457710/ [08:14] (30: Read-only file system)? [08:15] Surely it shouldn't segfault regardless of filesystem [08:18] I just attempted to mount as rw: http://paste.ubuntu.com/457711/ [08:18] Fail! [08:28] can anyone recommend an embedded board with GbE and SATA that's not constantly overheating like the GuruPlug? [08:29] kai: get guruplug out of case, attach some heatsink? [08:29] my friend made industrial device from sheevaplug [08:44] yeah, got a sheevaplug, thoug I still have one of the old usb only ones [08:45] network throughput is a bit below GbE, though [08:46] but faster then any of my arm devices [08:55] hrw: sure, I was just wondering if it made sense to extend the zoo :) [08:56] change casing of guruplug [09:10] yeah, I guess :) [09:12] or I'll just build my demo with the guruplug and the beagles [09:17] Why, why, why? http://paste.ubuntu.com/457730 [09:21] lag, sudo ?? [09:22] u-boot doesn't have sudo [09:23] no, i was commenting the first paste :) [09:23] is the SD locked ? [09:23] No, it's mounted ro [09:23] yes, thats how it should be [09:23] If I try and mount it rw it does the second paste link [09:24] right, looks like a locked card to me [09:25] ro on the cmdline is normal, mountall will remount it rw after the fsck [09:25] You can't lock a MicroSD card without an SD adapter [09:25] FYI: sudo does the exact same thing [09:26] how do you put a micro SD into the panda without adapter ? [09:27] good to know that pb has normal sd [09:27] Morning. Has anyone tried 10.04 on a OMAP3 based Overo board? [09:28] The kernel boots, but then I see 'init: ureadahead main process (49) terminated with status 5' [09:28] and 'init: procps main process (77) terminated with status 255' [09:28] ogra: No, that problem is on Beagle [09:28] michaelh1, thats both fine ... it should still boot [09:28] The fatload problem is on Panda :) [09:28] and then nothing. I have ttyS2 set up as a console under /etc/init/ttyS2 (but no screen at the moment) [09:28] lag, your u-boot prompt says panda [09:28] ogra: Different issue [09:29] k [09:31] lag, so you said XM works for you ? [09:31] michaelh1: That seems like a ureadahead exit code for when it deson't have data on the first boot [09:32] it definately doesnt for me [09:32] michaelh1: Which I would htink is expected, and ignored by upstart [09:32] Yes [09:32] michaelh1: Did you use /etc/init/ttyS2? [09:32] lag, no screen output at all, so our image doesnt boot [09:32] michaelh1: Cause it should be named .conf [09:32] /etc/init/ttyS2 wont be picked up, /etc/init/ttyS2.conf should be [09:32] ogra: Have you tried the latest kernel? [09:32] lag, its the alpha2 image i'm testing [09:32] lool: having a look now. I ended up with whatever rootstock --serial=ttyS2 generates [09:32] ogra: Wait one, I'll dig mine out for you [09:33] lag, i cant change the kernel on the image [09:33] lag, you said the archive kernel works [09:33] michaelh1: Try pressing enter after boot too [09:33] The one in the git tree? [09:33] michaelh1: One thing which might happen is that you have no network setup [09:33] michaelh1: That's a common issue with rootstock [09:34] Hmm. I supplied --serial=ttyS2 but there's no /etc/init/ttyS2.conf in the image... [09:34] michaelh1: You either need to install ifupdown and create a /etc/network/interfaces with at least lo, or you need ot install network-manager [09:34] Hmm. Might try VERBOSE in /etc/default/rcS as well... [09:34] lool, ?? [09:34] michaelh1: How about you create your rootfs by hand? That might be quicker than looking at rootstock issues [09:34] /etc/network/interfaces is created by rootstock with lo [09:35] michaelh1: apt-get install qemu-kvm-extras-static, qemu-debootstrap --arch=armel maverick /mnt/your-microsd-card [09:35] with a lucid host [09:35] ogra: Point me to the image you're trying to use [09:35] http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/current/ [09:35] michaelh1: You then need to create etc/fstab, etc/hosts, etc/hostname, etc/init/ttyS2.conf and etc/network/interfaces [09:36] Ta. I've created ttyS2.conf on the rootstock image. interfaces does contain lo. Booting... [09:36] lag, works on the C4 but since we dont use a swapfile yet, the desktop runs out of ram ... sincer you said XM works i was hoping we can use the image at least on a 512M board [09:36] lool, rootstock does all that [09:37] michaelh1, did rootstock finish without error ? [09:37] Kind of. I'm using apt-cacher-ng instead of the package cache and rootstock fails due to http_proxy containing a hostname [09:38] This causes apt-get inside qemu to fail on the update. Should still be good though. [09:38] no [09:38] you will only have a half built system [09:38] which rootstock version are you using ... recent versions have a fix for http_proxy [09:39] Ah! I have a login prompt! [09:39] And sudo... [09:40] And no network (due to having no network hardware :) [09:40] ogra: The version that came with 10.04 (0.1.99.3) [09:40] hmm, that should make use of a globally set http_proxy [09:41] how do you set it ? is it in your env ? [09:42] ogra: I usually dont complain about it, but you probably remember I replied a bunch of time here to people who were facing issues with rootstock generated images; it seemed to happen less these days. Another issue for a long while (perhaps still true?) was the version skew between the rootstock trunk bzr branch and the one in Ubuntu; overall I found it painful to deal with various rootstock issues, while I know exactly where manually created ... [09:42] ... rootfs might break; so no offense, but it's easier for me to help people when they are not using rootstock [09:42] ogra: the problem is the DNS resolution. Here goes: [09:42] michaelh1: instead of http_proxy=hostname.domain, you should use http_proxy=http://hostname.domain [09:42] that's more widely accepted [09:42] right [09:43] Yip. I use export http_proxy=http://crucis.local:3142/ [09:43] that looks ok [09:43] rootsotck uses both, /etc/apt/apt.conf.d/proxy or http_proxy [09:43] I think the problem is that crucis.local resolves via mDNS instead of the gateway's DNS proxy [09:43] michaelh1: You need libmdns in your chroot then [09:43] lool, the trunk version is exactly the same as in the archive, i dont know what you are talking about [09:44] lool, and if there are users for which a rootstock rootfs doesnt work, the first thing is to check the log if the build finished [09:44] ogra: for a long while, months ago I agree, there were fixes in bzr and a very different code base than in the archive; that might not be true anymore [09:44] If I set export http_proxy=http://192.168.1.15:3142/ then both sides work fine, but my IP changes depending on wireless vs wired [09:44] instead of fiddling in piecemeal with the broken FS [09:45] michaelh1: Can't you use 127.0.0.1 then? [09:45] michaelh1, can you file a bug, i'll research that [09:45] Oh no it's a real board attached to your laptop, not a chroot [09:45] lool: no, 127.0.0.1 is the qemu host when running [09:45] or qemu running in your laptop, right [09:45] Hmm. [09:45] michaelh1: You could have a separate IP + subnet just for this, but it's a bit heavy [09:46] rootstock should just create an /etc/hosts entry for the host machine during build (and remove the entry at the end) [09:46] I think running dnsmasq will also solve this: gives a real DNS server that both the host and qemu can use [09:46] I mean, you could add another IP to your laptop e.g. 10.0.0.1 and make sure that you have legs for that in any of the network where you need the proxy, but it's heavy [09:46] michaelh1: Yup, that's what libvirt does, but you still need an IP to talk to it [09:47] not with a hosts entry [09:47] orga: It seems that /etc/init/ttyS2.conf wasn't created [09:47] michaelh1, pleaswe file a bug, i'll add that [09:47] orga: my command line was rootstock --keepimage -l ubuntu -p ubuntu -f ubuntu --seed=lxde --serial=ttyS2 [09:48] michaelh1, right /etc/init/ttyS2.conf is created if the build finishes properly, if that failed you should have said that first [09:48] (and as i said above lool should have asked you about that first) [09:48] its a fallout of the missing name resolution, if you file a bug for that i'll fix it [09:48] OK. I'll try again using the IP address and see what happens [09:49] My impression from the log was that the build succeeded. apt-get failed, but it seemed to move on quite happily [09:49] (See http://paste.ubuntu.com/457737/ for the log) [09:50] if apt-get failed and you didng get a success message at the end, the build didnt finish [09:50] i need /home/michaelh/ubu/rootfs/rootstock-201007012009.log [09:53] Hmm. Firefox doesn't like pasting 11000 lines into paste.ubuntu.com :) [09:53] file a bug and attach the log ;) [09:53] ogra: The log is here: http://paste.ubuntu.com/457739/ [09:53] I will do. I'm up and going so it's all good. [09:54] (I quite like rootstock by the way. Nice and simple and readable) [09:55] thanks :) [09:55] lets see if we can make it better for you :) [09:57] (A --version option would be nice) [09:57] wishlist bug please :) [09:57] Ack [10:00] hmm, looking at the code for --serial i think the issue is the equal sign [10:00] "--serial ttyS2" would have worked [10:00] another bug :/ [10:00] Yeah, I saw that setup_serial is run before qemu starts but I thought I'd hammer you with one issue at a time :) [10:00] Will file. [10:01] i didnt have much tile for rootstock in the maverick cycle yet since we're building preinstalled SD images for omap now implementing that ate all my time [10:02] should get better now that we have these images and they only needs fixes [10:02] No worries. My recent history is with Openembedded and the speed (but not size) of Ubuntu+rootstock is refreshing [10:03] michaelh1: thats due to using built packages instead of building them [10:04] well, the speed of the binaries will not be as fast since OE optimizes the built binaries for your HW [10:04] Absolutely. I understand that OE can use binary packages but never got into it. [10:04] Ah, but most things are going to Cortex-A8 these days + runtime specialisation so there's not much difference [10:05] michaelh1: I did builds of images from binary packages in old openzaurus times [10:05] michaelh1: but code was a bit tricky so it took long before cleaner version appeared again in OE [10:05] lag, so wrt XM kernel i guess nobody tested the display (again), the kernel seems to boot fine but wihtout screen output its useless for us [10:06] lp #600539 [10:06] Launchpad bug 600539 in project-rootstock "Running apt-get against a proxy fails during the qemu stage (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/600539 [10:06] (ubot2 is pretty nice) [10:06] ogra: Mine boots with display [10:06] lag, with the archive kernel ? [10:07] I've just downloaded that [10:07] thats weird [10:07] How do I use this binary blob? [10:07] No, I haven't tested it [10:07] Just downloaded it [10:07] dpkg -x the .deb [10:07] Ah [10:07] mkdir tmp; dpkg -x *.deb tmp [10:07] 2 secs, pulling up console [10:07] then it unpacks it in tmp [10:07] k [10:08] I have used Linux before ;) [10:08] you need to run mkimage, it only has the vmlinuz [10:08] ogra: sorry, but how can I mark a bug as wishlist? [10:08] michaelh1, in the importance field [10:09] I'm looking at https://bugs.launchpad.net/project-rootstock/+bug/600540 and can't see how to change the importance [10:09] Launchpad bug 600540 in project-rootstock "Add a --version option (affects: 1) (heat: 6)" [Undecided,New] [10:09] It's down as 'Undecided' and isn't clickable [10:09] * ogra fixes [10:10] Nice. [10:11] ogra: I can only find the .img [10:11] lag, img of what ? [10:12] This is what you linked me to: [10:12] http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/current/ [10:12] lag, of the archive kernel package ? that should be a deb [10:12] lp #600541 (and that's enough) [10:12] Launchpad bug 600541 in project-rootstock "--serial option doesn't work (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/600541 [10:12] Doesn't look like a .deb to me [10:12] lag, oh, i understood you downloaded the archive kernel package :) [10:12] Hence the confusion [10:12] * lag shrugs [10:12] It's what you pointed me to [10:13] you need to bunzip the .img and dd it to SD [10:13] then boot the SD [10:13] note that we dont use serial output by default in the images [10:13] since that breaks the bootsplash [10:14] so before booting it you might want to mount mmcblk0p1 and modify boot.scr [10:14] on your host machine [10:16] lool, i saw you pasting some swapfile creation code in #linaro yesterday, i would like to use that in jasper, how do you determine $SIZE for it ? [10:17] is that a made up value or do you create it based on ram size etc ? [10:19] ogra: See lp:linaro-image-tools [10:19] I didn't work on it myself [10:19] yeah, i got that from your comment, i just thought it makes sense to use the same code [10:19] especially since asac will likely use pieces of jasper [10:22] ah, crap, SWAP_SIZE is a cmdline parameter ... i was hoping for some detection code [10:24] hmm creating the file under /mnt is brave ... it will vanish if the user mounts something in /mnt [10:26] lool, thanks, i'll steal some parts fo it, who owns it ? [10:27] s/fo/of/ [10:30] ogra: I dont know [10:31] ogra: bzr blame [10:39] ogra: Could you please reassign LP #600478 where it belongs? I would say either ubuntu-seeds or ubuntu-cdimage [10:39] Launchpad bug 600478 in netbook-remix "Remove dependencies on dove and imx kernel headers for maverick (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/600478 [10:39] or ubuntu-meta [10:39] or livecd-rootfs [10:39] Actually quite certainly not ubuntu-meta [10:55] ogra: What happens when you try to boot that image? [10:59] Let's try ogra_cpmc [11:06] lool, ist livecd-rootfs and its long on my list [11:06] ogra_cmpc: Could you move the bug so that it's not against an unused upstream project? [11:06] lag, i see it booting the kernel on serial but no output on screen [11:06] lool, will do [11:06] I have a login console on the screen [11:07] GrueMaster: Mind reporting bugs against either Ubuntu packages or ubuntu-cdimage rather than netbook-remix? netbook-remix should not be used anymore [11:07] Although, it won't power a mouse [11:07] GrueMaster: TIA! [11:07] I'm going to try it on mains power in a moment [11:07] lag, did you use the image or did you just rip out the kernel from the image ? [11:07] Just the image [11:08] Nothing more [11:08] Nothing less [11:08] weird [11:08] ;) [11:08] What commandline are you using? [11:08] modified boot.scr ? [11:08] the default one [11:09] Let me try with the default one [11:09] what did you use ? [11:09] setenv bootargs console=ttyS2,115200n8 console=tty2 root=/dev/mmcblk0p2 rootwait ro vram=12M omapfb.mode=dvi:1280x720MR-16@60 fixrtc; mmc init; fatload mmc 0 80300000 uImage; bootm 80300000 [11:10] Did you see the console on the screen? [11:10] hmm, thats not using the initramfs [11:10] no [11:10] The default prints to my screen [11:10] attached to dvi ? [11:11] my monitor doewsnt even power on [11:11] But it says "hub 1-0:1.1: unable to enumerate USB device on port 2 [11:11] " [11:11] HDMI [11:11] aha ! [11:11] But the DVI output yes [11:11] There aren't any USB devices plugged in? [11:11] i'll try hdmi soon [11:12] i have my powered hub plugged in [11:12] It's going nuts [11:12] No login console now [11:12] you shouldnt get a console unless the image resized itself [11:13] that will take about 10min it should tell you though [11:13] I just have lots of 'hub' error messages [11:13] sounds like the udb driver has issues [11:14] * ogra_cmpc finished his breakfast, let me go back to the office and try hdmi [11:14] * lag forgot all about breakfast [11:14] * lag is going now [11:19] ogra_cmpc: How are you getting on? [11:20] nothing on HDMI [11:20] I don't know what to tell you [11:20] Mine works [11:20] With my cmdline I get a login console [11:20] With the default I get kernel messages [11:20] Both on the monitor [11:21] thats really strange [11:22] i see u-boot messages on the serial console up to "Uncompressing Linux... done, booting the kernel." [11:22] +but the monitor never powers on [11:22] Is the monitor on? [11:22] Standby, off or on? [11:23] its the monitor with which i'm typing to you atm :) [11:23] i just switch the input source [11:23] ogra: disable 'quiet'? [11:23] so its on [11:24] hrw, there is no quiet on the image cmdline atm [11:24] our headless images i was told have a console on screen (/me hopes) [11:24] fatload mmc 0:1 0x80000000 uImage [11:24] fatload mmc 0:1 0x81600000 uInitrd [11:24] setenv bootargs vram=12M omapfb.mode=dvi:1280x720MR-16@60 root=/dev/mmcblk0p2 fixrtc [11:24] bootm 0x80000000 0x81600000 [11:24] thats the boot.scr we use in the image, works fine on C4 [11:24] asac: first you need to have that console... [11:24] ogra: bbxm? [11:24] ogra: thats on xm? [11:25] but we dont use a swapfile yet, so C4 goes OOM [11:25] asac, yes [11:25] ah you are fighting netbook ;) [11:25] fun [11:25] ogra: bbxm in ubuntu kernel do not have display iirc [11:25] asac, oem-config works [11:25] didnt get info back on that from plars i remember now [11:25] * asac goes aksing [11:25] asac, just the desktop afterwards makes it die, i'll add swapfile creation to jasper_setup [11:25] plars: did anything came out of the test without --live last night? [11:26] hrw, thats why i'm confused, it works fine for lag [11:26] hrw, same image, same board [11:26] ogra: err. can you please not create swapfiles unless there is a special flag? [11:26] ogra: "but we dont use a swapfile yet, so C4 goes OOM" - So does the XM [11:26] we have our swap file in place ;) [11:26] imo thats too far of magic to put into jasper [11:26] asac, is yours always in the same place ? [11:27] asac, its not, since jasper creates the partition table [11:27] ogra: why? [11:27] hrw: Works for me [11:27] ogra: then jasper should not replace the swap part [11:27] its exactly what jasper needs to do [11:27] if there is a swap entry... jasper could skip that step? [11:27] asac, if i can detect that you have a swapfile in place i can skip it [11:27] ogra: right. you can detect that from fstab [11:28] right [11:28] ogra: also i would really like to keep it optional to create one [11:28] for some images we dont need it [11:28] asac, btw, i looked at your code, i dont think its clever to create the file under /mnt [11:28] ogra: my code? [11:28] you mean linaro-media-create? [11:28] if the user mounts something to /mnt it will be hidden [11:28] yeah [11:28] let me lok at that code [11:29] i agree. where should we put it? [11:29] i'm not sure how others use /mnt but i typically mount manually mounted stuff there by default [11:29] i would just put it in / [11:29] we are redoing parts of it today anyway [11:29] hmm [11:29] lool: any opinion? [11:29] e.g. where to put SWAP.swp file [11:30] we currently have that in mnt ... which feels odd indeed [11:30] I think mnt is a bad idea [11:30] on my systems, I used to put it in /srv and now it's in / [11:30] ok ... lets go for / then [11:30] * asac changes that [11:30] but neither is FHS compliant I'm afraid [11:31] is there any FHS recommendation for swapfiles ? [11:31] In general, swap files are not as good as swap partitions, so we should aim for that if possible [11:31] * ogra never heard of that [11:31] ogra: that's the issue, there is none [11:31] * ogra was planning to usw a swap partition with jasper since i repartition anyway [11:31] ogra: but FHS does list what's supposed to be in / [11:31] ah [11:31] It could in theory be /var/cache/swap too [11:31] yeah [11:31] pushed [11:32] lool: rather /var/cache ;)? (feels a bit nicer to me) [11:32] I think doing it in the form of jasper is not too inelegant, albeit it's important that the user has control on whether or not swap will be created, where, and how large [11:32] sounds sane [11:32] this is all very hard to implement consistently [11:32] well. lets keep it in / [11:32] asac: I would consult with installer folks to get a better opinion [11:33] i think /var doesnt exist on casper etc. [11:33] I think / is good for our current mode of development, but should be revisited [11:33] right [11:33] lets file a bug [11:33] * asac goes for it [11:33] casper is tricky anyway [11:33] since you use an aufs mount [11:34] creating a swapfile will just eat your ram in casper [11:35] lag, XM shouldnt OOM at all it has 512M [11:35] lag, when do you get OOM ? if you start the desktop ? [11:35] The desktop starts automagicly [11:35] You should not create a swap file in casper's tmpfs mode [11:36] I didn't get oom this time [11:36] but casper also supports persistent read-write, and that partition could host a swap file [11:36] lool, well, ubiquity would need to learn about it so it gets copied to the target [11:36]