=== JaMa is now known as JaMa|Zzzz === JaMa|Zzzz is now known as JaMa|Wrk === hrw|gone is now known as hrw [08:23] morning [08:29] hrw, hi [08:38] hi [08:39] if a bug like https://bugs.launchpad.net/ubuntu/+source/linux-ti-omap/+bug/584920 has been fixed, how can i get a the new, fixed installer image? [08:39] Launchpad bug 584920 in linux-ti-omap (Ubuntu) "netinstall fails, it has no network driver for moschip (affects: 1) (heat: 6)" [Medium,Fix committed] [08:39] a.. i mean, the :) [08:40] neure, you have to wait for a debian-installer respin, that will likely happen for 10.04.1 (end of july) [08:40] the fix includes all USB NIC modules now, was uploaded this week [08:43] could i somehow would the installer myself? [08:44] like, i should be able to try the fix, right? [08:44] you need plar's script to install a kernel into a live image [08:44] s/install/update [08:44] thats almost mumbo jumbo to me .. :) [08:45] ok that makes a bit more sense [08:48] of course would be lovely if someone did that for me .. :) [08:49] but.. where do i get the script and where do i get the new kernel? [08:49] and where do i get suitable live image? [08:53] http://people.canonical.com/~amitk/update-image-kernel-x86.sh [08:53] http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-live/current/lucid-netbook-armel+omap.img [09:02] amitk, that wont help for netinstall though [09:03] also the live image is http://cdimage.ubuntu.com/ports/releases/lucid/release/ubuntu-10.04-netbook-armel+omap.img [09:03] (though its not really live, it just fires up an installer under X) [09:04] ogra: no, it won't help with netinstall [09:06] neure, the "live" image has the moschip driver included no need to update the kernel there, its just missing in the netinstall [09:06] asac, hi [09:06] asac, did plars talk with you about image identity/timestamps? [09:08] so.. [09:08] is there some way for me to have installed with the netinstaller? [09:15] neure, netinstaller works fine ait asix based NICs if you can get such a device it will definately just work [09:15] s/ait/with/ [09:16] asix based? [09:18] neure: any reason you can't use live image? [09:20] i dont know exactly what live image means in this case [09:20] perhaps i could, i havent tried [09:21] is it something that i could consider as 'preinstalled' ? [09:34] neure, no, they are all installer images, the "live" image simply installs a netbook system to an USB key [09:34] preinstalled images will come with the next release [09:43] no [09:43] i can not use live image [09:44] at least the server installer failed to access usb stick on my C2 beagleboard [09:44] i tried two different powered USB hubs and two different USB keys [09:45] weird [09:45] so i assume all ubuntu installer images have the same issue: they can not access usb discs on this setup [09:45] if you had some ideas, i could investigate further, but im really not a linux expert in any way [09:45] neure: anything other on usb was visible? [09:45] * ogra doesnt have a C2 so cant tell [09:45] keyboard and ethernet were on the same usb [09:46] same hub [09:46] keyboard worked fine [09:46] ethernet wasnt working as it lacked the drivers [09:47] angstrom works on the otg with powered hub and it can access usb discs and everything works [09:47] so i wish you guys also supported otg :) [09:48] but of course im not sure if this was really usb issue or if for example my installer image was bad [09:48] otg will come with maverick [09:48] there are MD5 files on the download page [09:48] i was able to proceed with the installation to usb with a usb stick i had preformatted on pc [09:48] but then it failed complaining some package was corrupt [09:49] i suppose i would need to read the image back from the sdcard and then check md5? [09:50] if you didnt keep it on your PC [10:08] well im suspecting the image might have been corrupted while it was transferred to sdcard [10:08] image on pc could be ok [11:27] lool: ok, added note === XorA|gone is now known as XorA [11:28] lool: intermediate stages for gcc/eglibc: do we make subpackages for them (gcc-cross-stage1, gcc-cross-stage2) or other? [11:31] lool: 'build binutils with -sysroot support' is already done in packaging (--with-sysroot=/) since 2.20.51.20100407-0ubuntu1/lucid (08 Apr 2010) [11:45] hrw: I'm not sure I understand the first question [11:46] hrw: In the end, you will have a meta-package called e.g. cross-toolchain which will build-dep on the gcc source and the binutils source, you'll have to call the binutils rules and the gcc rules to achieve building a full set of cross-compiler packages [11:46] ok, ignore first question now [11:48] lool: started binutils part [11:49] hrw: You experienced the process of creating them in the emdebian way, right [11:50] hrw: You could see how the various steps created more packages, which you then installed and allowed you to complete the next one [11:50] hrw: Here, the hardest part is getting away without dpkg -i [11:50] hrw: You're going to have to build all packages from a single meta-package build [11:50] I know [11:50] dpkg -i wontbe allowed [11:51] basically "dpkg-source -x cross-toolchain-for-arm.dsc; cd cross-toolchain-for-arm; debuild -b -uc -us" [11:52] hrw: In the end, we will upload a cross-toolchain-armel.dsc, and it will give binutils-arm.deb, gcc-arm.deb etc. [11:53] hrw: The hard part is that while you have the full binutils sources and gcc sources and their packaging, you need to call into their packaging to reuse their logic as much as possible (as to give us cross packages which look identical to the non-cross ones) [11:53] hrw: and you may not install them in /usr during the build, since the build runs as non-root [11:54] yep [11:55] hrw: it is likely that you indeed have to build the intermediate stages, I highly suspect you will have to use --sysroot to achieve building in a subtree [11:58] will get to it in one moment [11:58] now working on binutils [12:07] grmbl [12:07] * ogra starts to hate x-loader [12:11] * amitk gives ogra redboot :) [12:11] heh, i wish i could use it on omap4 :P [12:12] redboot... argh [12:12] i can at least get it to build :P [12:13] that huge mess fo hard links between git trees is nearly unresolvable to get a natively building package [12:14] ogra: working on the omap4 x-loader? [12:14] yeah [12:14] how did you guess ? :) [12:14] ogra: someone probably needs to a) pay sakoman to incorporate that work into his or b) merge WTBU trees with his [12:15] ogra: I had to deal with that mess for Angstrom [12:15] XorA: and got it working? [12:15] well, i got the headers from u-boot in my tree now but x-load-arm.h seems to have broken assembler now [12:15] its just massively messy [12:16] and the biggest fun is that all i'm doing atm is throw away work because x-loader is going away anyway [12:16] hrw: for omap3 which is same tree I ended up just mirroring a tarball of the u-boot tree and extracting that next to the x-loader git repo [12:16] but currently we need it to get the image building for omap4 going [12:17] XorA, heh, i did a giant patch with the headers ... [12:17] ogra: I started that way then got real bored, I wasnt being paid for that work :-) [12:17] but nontheless there are assembler errors in x-load-arm.h [12:18] ogra: maybe they use gcc 2.95 for building [12:18] oh, assembling the patch didnt take much ... just a small script with some find loginc to search for dangling symlinks [12:18] *logic [12:18] try fixing the arch to armv5 [12:18] hmm [12:19] * ogra tries [12:19] I vageuly remember armv7a needs and old gcc to actually compile that file [12:20] nope [12:21] ogra: what toolchain are you using? [12:22] lucid atm [13:06] * hrw → lunch [13:11] ogra: do you know why I wouldn't get a password prompt after the login prompt? [13:11] i return back to the login prompt on entering the user name (nfs root) === rgreening_ is now known as rgreening [13:42] amitk, not really, check auth.log (with init=/bin/bash) [14:14] x-loader-omap4_L24.6-p1git20100520-0ubuntu3_armel.deb [14:14] yippie [14:18] phew, what a painful exercise [14:21] Do we need a separate loader? [14:21] I thought it was supposed to be backwards-compatible? [14:22] lool, we need MLO for each single bpard until the header is added to u-boot [14:22] and even then it needs to be a binary header per board since they are wired up differently [14:23] panda and blaze definately are different here, as well as XM and beagle [14:24] s/beagle/C*/ [14:26] lool, though for omap4 i was told the boundary to block zero in vfat for MLO has fallen, so we can just ship all MLOs that exist and you can boot different boards with the same image through just renaming i.e. MLO-blaze to MLO [14:26] thats already a big win [14:26] (or with the binary header you can rename u-boot-blaze.bin to u-boot.bin) [14:32] re [14:41] is there public doc on the blaze? [14:42] lool: I found this: http://svtronics.com/market_omap [14:43] * hrw is building cross-binutils for all arches now using debuild [14:43] vstehle: thanks [14:44] lool: No schematics it seems [14:44] vstehle: Oh right, I've actualy seen it [14:45] lool: You need some specific info about the blaze? [14:45] Was just curious what it was, but I see now, thanks [14:46] zyga: we started to discuss that [14:59] 1 file changed, 52 insertions(+), 13 deletions(-) [14:59] whole changes for binutils [14:59] zyga: depends on what the imjage building concepts are etc. [15:00] hrw: you found lp:ubuntu/binutils I guess? [15:00] lool: I took latest maverick package but will provide diff against that branch [15:00] asac, I think that even thought there are several bits of infrastructure involved having timestamp and image type inside the image is very important [15:03] what are the bootloaders usually used on ARM? [15:04] james_w, u-boot, redboot, apex are some we used before [15:04] there sadly are plenty [15:04] james_w: uboot, redboot, apex, blob and few others [15:04] ogra: all of them worth having support for in image building tools? [15:04] james_w, nope [15:04] ogra: which would you pick to have support for? [15:04] afaik jcrigby is doing some work to unify on u-boot [15:05] james_w: u-boot [15:05] u-boot though i like redboot too, but we dont use it anymore [15:05] ok, that's what we will start with then [15:05] thanks ogra, hrw [15:05] * ogra is just rolling u-boot-omap4 packages [15:17] zyga: yes. i fully agree for images, but from what i understood plars wanted a way to have a way to describe what version an install is on [15:18] e.g. even after dist-upgrading other packages [15:56] asac: ar9170 [16:05] WTF [16:05] # CONFIG_CFG80211_WEXT is not set [16:05] bonkers! [16:05] useless! [16:05] please fix, quickly, total showstopper [16:06] xorAxAx, i guess you rather mean amitk [16:07] * amitk looks up [16:07] amitk, he's right [16:09] mpoirier: ^^^ could you please look at this as a Lucid SRU? [16:09] amitk: you mean the discussion above ? [16:10] yes, enabling CONFIG_CFG80211_WEXT [16:10] amitk: Ok, I'll open a pub and fix it. [16:11] * ogra will go and have a beer at mpoirier's pub :) [16:11] heh [16:12] * amitk notes that it is beer o'clock here [16:12] here too but i'm out of beer :( [16:16] keep working until new beer arrives :-) [16:56] have a nice weekend everybody === hrw is now known as hrw|gone [17:07] XorA, hmm [17:08] XorA, did you ever try your mkcard script with a real SD cardreader ? [17:08] and on a non english system :) [17:11] ogra: non english no, but the version in OE should work for that if the patches I got from french speakers were correct [17:11] ogra: and it certainly works on read SD card readers [17:12] well, it does explode on mmcblk0p1 :) [17:12] since you have the p there [17:12] and "Disk" doesnt show up in german localized fdisk -l output [17:13] i had to modify it like that for my system http://paste.ubuntu.com/440936/ [17:14] ogra: thats an old version of script, the one in OE is more robust [17:14] ah [17:14] i took the one from your blog [17:14] I thought I had updated all the pages to point to OE git [17:15] http://www.xora.org.uk/2009/08/14/omap3-sd-booting/ [17:15] http://cgit.openembedded.net/cgit.cgi/openembedded/tree/contrib/angstrom/omap3-mkcard.sh [17:15] well, it works now, i just needed something quick [17:15] ogra: top of that post tells you to go to OE git :-) [17:16] heh, i'm blind [17:16] :-) [17:16] that version even runs on the target platform :-) [17:16] grr, my MLO doesnt work :( [17:25] NCommander, the driver for the otg port in lucid doesnt work [17:26] ogra: ugh [17:26] use an usb nic [17:26] ogra: I don't have one, or an ethernet jack I can plug it into [17:29] ogra: why is it broken? [17:34] because its broken ... no idea the patch didnt work [17:37] ogra: ugh, ok. What package is x-loader in? (google is failing me) [17:37] x-loader-omap [17:39] Texas Instruments X-Loader 1.4.3 (Mar 24 2010 - 21:16:42) [17:39] W00t [17:39] it boots from MMC! === XorA is now known as XorA|gone [18:04] phew [18:04] * ogra dputs x-loader-omap4 [18:04] and cllas it a day [18:04] *calls [18:28] ogra: \o/ [18:53] sebjan: ping? [20:16] what does the USR0 blinking mean in my lucid install? [20:21] ah, heartbeat === rsalveti_ is now known as rsalveti [21:08] how to overclock the beagle to 600 mhz without using memory pokes in uboot? === sbambrough is now known as sbambrough-afk [23:16] http://www.engadget.com/2010/05/28/lenovo-kills-skylight-os-in-favor-of-android-u1-hybrid-and-skyl/ [23:16] hmm... does that mean the PRODUCT is dead... or just the OS? [23:17] yeah, product is canned. Dang. [23:18] They could've just taken the "skylight" and put ubuntu on it... and called it a full OS (which it is!).