=== chuck_ is now known as zul === Amaranth_ is now known as Amaranth === Amaranth_ is now known as Amaranth === XorA|gone is now known as XorA === ogra is now known as Guest42038 === Guest42038 is now known as ogra_ [12:38] are the daily images for omap4 working? [12:38] should, yes [12:38] * ogra_ hasnt tested any graphical ones recently, only headless [12:40] ogra_: ah, that's the part I'm worried about, linaro images apparently don't turn on the display [12:41] the ubuntu ubuntu-netbook ones do [12:41] just use these [12:41] Yeah, I'm getting the one from http://cdimage.ubuntu.com/ubuntu-netbook/daily-preinstalled/current/ now [12:43] that should be fine, if you have probs with it, the alpha-3 one definitely works and was tested [12:43] (see topic for url) [14:20] for those interested, there is a an ARM porting Jam on #linaro this time every week; it's not too busy but if anyone knows of particularly pesky ARM only bugs or if anyone needs help with ARM tools/porting, please come and ask [14:20] davidgiluk, yes, we all get the gcal notifications ;) [14:21] ogra_: I assumed those in this channel weren't all on these lists [14:21] davidgiluk, but nobody will fix bug 739374 for me anyway :P [14:21] Launchpad bug 739374 in eglibc "eglibc newer than 2.12.1 in natty results in alignment errors, SIGLILL and segfaults on tegra2 systems" [Undecided,New] https://launchpad.net/bugs/739374 [14:21] * davidgiluk looks [14:23] ogra_: Youch, that's a nasty errata [14:23] yes [14:23] and no real fix for libc [14:23] that indeed wont help to get nvidia on board with linaro ever [14:24] (or with ubuntu) [14:24] ogra_: Is the details of that errata public? [14:25] well, the kernel code as well as bionic are public ... not sure what else you need [14:25] i doubt nvidia released anything with more details [14:25] ogra_: So what's the problem exactly - bit 20 always reads in one way or do you get junk out? [14:26] i cant boot :) [14:26] nor can i chroot [14:27] so its a bit hard to debug at all [14:27] chrooting immediately segfaults, booting dies very early with alignment errors and SIGILLs [14:28] currently i use the natty libc pinned in apt to have a working system at all [14:28] and i have no idea how to debug such early userspace issues [14:30] ogra_: I'd probably start with a statically build busybox [14:31] right, if i hadnt beta freeze tomorrow i would invest that time :) [14:31] * ogra_ uses his tegra as main work machine so its a bit hard to tinker with it [14:32] given the feedback from peter it seems like a pretty desparate issue anyway [14:35] ogra_: I mailed Gary King of Nvidia who did that patch to ask if he can give any more detail on the failure mode [14:36] sweet ! why didnt i think of that ! [14:36] did you? [14:37] no i didnt :) [14:38] * davidgiluk admits to knowing very little about TLS, but it strikes me if the failure is anything other than returning random junk then it might be possible to align the allocation - still painful, but it might actually work [14:38] well, you actually need to know you are on a tegra etc [14:38] that will likely be hard to do inside libc [14:39] ogra_: Nah [14:39] ogra_: Messy, not hard [14:39] well, messy yeah [14:39] :() [14:39] bionic just bends the headers at build time [14:41] ogra: I'm fairly sure I saw something about some other hardware having specific TLS memory, so perhaps it's not too unusual [14:42] hmm [14:42] [14:42] ogra_: Well, the bad news is that his email address bounces [14:43] bah [14:58] ogra_: I've mailed a random other Nvidia person I found in a commit log :-) [14:58] heh [14:58] * ogra_ crosses fingers [14:59] ... and goes back to bang his head against bug 736111 [14:59] Launchpad bug 736111 in ubiquity "oem-config does not respect the console on serial tty" [High,New] https://launchpad.net/bugs/736111 === prpplague^2 is now known as prpplague [17:26] Are there any rootfs tars of Maverick available other than the pre-install image? I tried building one with rootstock, but I'm continually running into issues of it hanging during the process. === renato is now known as renatofilho [17:52] amelim: a) use a recent qemu-linaro, if you're not running natty, you can use the ~linaro-maintainers/tools PPA and install qemu-user-static from there b) there are alternatives solutions to rootstock I can hint at if it doesn't help [18:02] Is the preinstalled image supposed to turn on the screen before the first-run unpacking is complete or does it wait until after? [18:03] wondering if I should give this time or dig out my serial adapter [18:04] natty ? [18:04] yeah [18:04] do you use DVI ? [18:04] no, HDMI plugged in to the first port [18:05] the current image doesnt support HDMI, only DVI with fixed 720p [18:05] *headdesk* [18:05] new kernel with a fix that works like in maverick was just uploaded today [18:05] should be in by the weekend [18:05] until then you need to use DVI [18:07] i guess i need to check and see if they got the regulator patch included [18:07] I wouldn't have the proper adapter until the weekend anyway [18:07] I don't even own a DVI cable :P [18:15] Amaranth: The "DVI" port uses the same signalling as the HDMI port (except audio). [18:15] So the same cable should just work. [18:16] GrueMaster: So I can plug an HDMI cable in to it and it'll show something, just locked to 720p? [18:16] Amaranth: http://pandaboard.org/content/resources/references [18:17] yes. [18:17] That's how I test here. [18:17] Amaranth: there is a table there describing the different combinations [18:18] ok, I guess it needs a reset to know to probe for it, hope it was done with the first boot extracting :P [18:19] ooh I have an offset plymouth [18:24] lool: I'm on a lucid distro, if I install qemu-user-static will I need to remove any other qemu packages for rootstock to run correctly? [18:24] amelim: We renamed various packages around since lucid; what you'll be getting is an equivalent set of features, but provided by different package names [18:24] amelim: You might have to upgrade other packages as well, but essentially it's different package names with same contents but newer versions [18:25] and eventually, when you upgrade to some later Ubuntu versions, you'll have the same package names provided by Ubuntu [18:25] (these are just backports) [18:25] hmm, it seems to have rebooted and now plymouth is showing correctly, this seems promising [18:25] ahhhh, ok [18:27] yay it booted, thanks GrueMaster [18:27] np. [18:27] now fix compiz ! [18:27] :) [18:27] Are you trying the netbook image or the headless image? [18:27] GrueMaster: netbook [18:28] ogra_: Well, I have this code that compiles and I'm pretty sure does the right things but GLES is apparently broken in mesa for sandy bridge so I haven't been able to run it [18:28] Thus getting real hardware going [18:28] There is a bug in jasper that breaks video output after first boot. [18:28] GrueMaster, hmm, how so [18:28] i dropped all special casing for video stuff recently [18:29] it hands through the cmdline directly now [18:29] ogra_: Your fix hasn't landed in an image yet. [18:29] hmm, it should be in headless [18:29] As the current image is 20110321. [18:29] Not in netbook [18:29] no, indeed [18:29] blame kdelibs [18:29] crap [18:30] Oh, I blame KDE for everything, don't worry [18:30] heh [18:30] and glib too [18:30] first glib was out of sync, now kdelibs held up a lot of subsequent stuff (like compiz) [18:30] How do you add icons to either the launcher or places in unity-2d? [18:31] we should drop these crappy graphical images and just resort to headless everywhere :P [18:31] And how does kdelibs hold up compiz? [18:31] heh, compiz build depends on them [18:31] no idea why [18:31] compiz pulls in _everything_ [18:31] Very odd. [18:32] kdecompat plugin (plasma stuff), kde4-window-decorator, and the kconfig compizconfig backend [18:32] And people say they don't work with other environments...sheesh. :P [18:32] If you need to make sure you have -dev packages for every possible GUI app you ever want to build just apt-get build-dep compiz [18:33] So, back to my unity-2d question... [18:33] no idea [18:34] places should be connected to nautilus places [18:34] somehow [18:35] and that gets updated...how? [18:35] so adding one in nautiluxs might show up in the launcher [18:35] Isn't nautalis used in the une-efl? [18:36] hmm, this doesn't look like 720p [18:36] looks more like 640x480 [18:37] Amaranth: Yea, that's the jasper bug. [18:37] You can fix it by editing the /boot/boot.script. [18:38] Don't I have to extract that from boot.scr then regenerate it? [18:38] add "omapfb.mode=dvi:1280x1024MR-32@60 omapdss.def_disp=dvi" to the bootargs, then you need to rerun mkimage on the file. [18:39] The /boot/boot.script is a text version of the u-boot script. [18:39] might as well enable the serial console while I'm in there, in case I screw it up :) [18:40] It is in the root partition. After editing it, rerun "mkimage -A arm -O linux -T script -d mnt/boot/boot.script boot.scr". Then umount the second partition, and copy the boot.scr to the fitst partition. [18:40] huh ? [18:40] sudo flash-kernel [18:40] Kind of a pita, but the next image should fix this. [18:41] no need to run any mkimage [18:41] ogra_: That is assuming you are already running. oem-config has major issues on 640x480. [18:41] (on the running system) [18:41] I am running [18:41] ah, k [18:41] oem-config has zero issues if you know alt-button1 moves the window around :) [18:41] My steps can be done on the desktop system. [18:41] * ogra_ didnt know oem-config misbehaves [18:42] oem-config windows are designed for *x600 screen height. [18:42] Amaranth, well, just edit /boot/boot.script and run sudo flash-kernel afterwards then [18:42] on next reboot the new config will be used [18:46] hrm, says something is using /dev/mmcblk0p1 so it can't unmount it [18:46] but as far as I can tell the mount point is one flash-kernel is creating [18:47] But it seems to be writing things properly, just not completing umount afterward so I guess it's reboot time [18:49] dang, mode not supported, guess it's a good thing I put that serialtty bit in there [18:50] well, you can always use GrueMaster's method above with your desktop PC [18:51] that's no fun :) [18:51] No one ever said my job was fun. :P [18:59] so with the next kernel update none of this will be needed, right? [18:59] yes [19:00] will behave exactly like in maverick [19:02] Ok, figured out how to hack in the ti addons icon into unity-2d. The method is anything but user friendly. :( [19:02] But it looks good. [19:03] oh, for favorites ? [19:03] thats trivial, there is a gconf key [19:04] That's what I was asking about. [19:04] Thanks. [19:04] no, you asked about adding an icon to the launcher [19:04] or to places [19:04] nothing indicated you wanted a new favorite [19:04] And the launcher currently has it's own list of favorites. [19:04] jasper should take care post beta [19:04] yeah [19:05] same goes for hiding behavior [19:05] sigh. [19:06] and the method doesnt need to be user friendly, no user should use it ;) [19:07] No user would ever want to add their own favorite app to the favorites? [19:08] You're kidding, right? [19:08] not programmatically [19:09] drag and drop is implemented in the latest upload afaik [19:09] Nope. I'm running unity-2d daily. [19:09] For Maverick. [19:09] users will never have to fiddle with the gconf system default [19:09] maverick ... [19:09] Well, until D&D works... [19:09] it makes use of unity features for that afaik [19:10] so needs natty [19:10] The unity-2d is from the ppa. [19:10] but not the unity :) [19:10] it uses the dash backend afaik [19:12] ogra_: I thikn you can close multiple bugs with LP: #123, #456 [19:13] lool, but that wouldnt give me the ascii art :) [19:13] it would give you a different one [19:14] yeah [19:14] i'll use it next time or teach upstream about using dch === Jack87|Away is now known as Jack87 === Jack87 is now known as Jack87|Away