=== rsalveti` is now known as rsalveti === XorA|gone is now known as XorA [10:23] lool, pingaling [10:27] lool, any idea about bug 1026780 would be appreciated [10:27] Launchpad bug 1026780 in flash-kernel "3.0~rc.4ubuntu4 doesn't honor bootargs from /boot/boot.script anymore" [High,New] https://launchpad.net/bugs/1026780 [10:29] * ogra_ curses, I/O error on an omap3 desktop install ... [10:29] (with a brandnew SD and USB key) [10:31] gar ! [10:32] ppisati, squashfs errors all over the place with my beagle revA [10:32] (XM) [10:32] * ogra_ hasnt seen them in a while [10:40] cool [10:40] ppisati, hmm, i wonder if bug 746137 might be at fault here (we dont have the mechanism anymore to apply that sysctl.d file in quantal) [10:40] so my rev a is not broken [10:40] Launchpad bug 746137 in jasper-initramfs "Page allocation failure on Pandaboard and Beagle XM" [Undecided,Fix released] https://launchpad.net/bugs/746137 [10:40] LOL !!! [10:40] do you remember my email to you and ming's? [10:40] vm.min_free_kbytes = 8192 ... [10:41] thats the workaround [10:41] that was a sysctl [10:41] did we drop it in live? [10:42] btw, i'm out for lunch shortly [10:42] we dont have the hackish jasper stuff anymore [10:42] well, it's a sysctl [10:43] and no replacement mechanism that could apply that hack on a subarch specific basis [10:43] well, i guess it's too late to work on it now for alpha3 [10:43] omap3 wont be there for A3 ( skaetr decision it seems) [10:43] *skaet [10:44] so something to look at for post milestone work [10:45] besides you, who has a beagle? [10:46] there are lots in the DC we should share out ... apart from that ... we could only randomly ping here or in #beagle or blog about testing images [10:46] not sure, infinity might have one too ... [10:46] probably no XM though [10:47] ogra_: Commented in the bug itself, I think you had already found the root cause though [10:47] lool, right, my prob is that even if the user changes the scripts, padckage upgrades will overwrite the changes [10:48] * ogra_ reads the comment [10:49] I don't think flash-kernel generates the boot script anymore; any generated stuff should be temporary files and removed at the end of the run [10:49] then it installs boot.scr into the firmware partition [10:50] well, there is definitely a mkimage_script() function [10:51] called for the generic type [10:53] seems it updates on every run actually [10:59] ogra_: the oom bug can be fixed by replacing the GFP_ATOMIC by GFP_KERNEL in the relevant kmallocs [10:59] this worked for our rt2x00 wifi driver [10:59] and may work for other usb devices as well [11:01] mainline commit: 8821102 rt2x00: Use GFP_KERNEL for rx buffer allocation on USB devices [11:09] ogra_: mkimage_script just takes a script text file and makes an u-boot image out of it, but it's reading the boot script text from BOOTSCRIPTS_DIR and generating boot.scr in $tmpdir [11:31] lool, right, so it updates on every flash-kernel run [11:32] i.e. boot.scr is always regenerated [11:32] ppisati, see marvin24's comment above [11:53] infinity, thanks for volunteering me to fix issues around mono over a month ago on the cross-arm ml and not telling me. Any more of this and you won't be my second favourite canadian [11:53] * janimo must admit he stopped following that list shortly after it started [12:34] ogra_: to me it looks like a different problem [12:34] ogra_: we experience squashfs problems [12:34] ogra_: i mean, corruption [12:34] i don't know if the two are related [12:35] the smsc driver issue eats all your ram [12:39] so i think it is very very likely that it causes the squashfs issues [12:39] easy to test though, i will try once i have time for omap3 again [12:43] let me cut&paste what i found back then [12:45] ogra_: http://paste.ubuntu.com/1108129/ [12:45] marvin24: http://paste.ubuntu.com/1108129/ [12:46] what i saw was the same squashfs/memory corruption you found now [12:46] but only on the rev a [12:46] on the rev c everything is fine [12:46] i probably already thrashed the system i used back then [12:46] but i can recreate it [12:46] anyway [12:46] well, the different revs use different types of RAM [12:47] problem with memory allocation in a nic driver != squashfs corruption [12:47] could be [12:47] is so :) [12:47] that's why i would like someone with a rev c do repeat my test [12:47] and the turbo mode of the smsc chip will eat all your ram [12:47] until it falls over with an oops [12:48] ok, but here is different [12:48] my transfer of files over the net was ok [12:48] but md5 locally was screwed [12:48] i could mount the squashfs on rev c [12:48] its super trivial to echo the proper value into the sysfs and test :) i'm just busy fixing omap4 for A3 atm [12:48] but NOT on rev a [12:48] etcetc [12:49] (or to force a proper module option to enforce turbo mode off) [14:34] infinity, FYI I am not sure how easy it usually is to bring up a full desktop from ubuntu-core chroot, but for me it failed both today and last time you first mentioned it [14:34] maybe it's due to quantal xubuntu/lubuntu not being installable? It failed in dbus and other coreish packages [14:35] oh well, I hear rsalveti will make linaro rootfs tarballs of quantal and then I won't have to worry much about it :) [14:35] janimo, ubuntu-core is just deboiotstrap+apt [14:36] if that wouldnt work, no images would work at all [14:36] ogra_, yes, I know. Too bad it's not always straightforward to apt-get install a desktop from that [14:36] indeed that's what I find frustrating and confusing as well [14:36] http://omappedia.org/wiki/OMAP_Ubuntu_Core is a pretty good read [14:36] and hence I want tarballs :) [14:37] sudo apt-get install ubuntu-desktop should definitely work, else we would notice in all other builds [14:38] ogra_, I guess the steps of bind mounting /proc & co in the chroot are not that unimportant [14:38] i dont think anyone checked xubuntu or lubuntu on arm [14:38] that I did not do, even though there were warnings [14:38] bind mounting /proc ? [14:38] I used the chroot method qith qemu-arm-static [14:38] making an armhf rootfs on x86 is my goal [14:39] right, no bind mounting for proc or sysfs needed [14:39] bottom of the page you linked to [14:39] you need to bind mount /dev though [14:39] for m in `echo 'sys dev proc'`; do sudo mount /$m ./$m -o bind; done [14:39] on that page [14:39] right [14:39] well, it wouldnt fail for sure, but i would do it differently :) [14:40] yeah, TMTOWTDI is what annoys me with these things, as opposed to one simple command :) [14:40] you *definitely* *always* need mounted /proc and sysfs ... [14:40] if you work in a chroot [14:40] dev only for cretain things like update-initrtamfs [14:40] *certain [14:41] that is what I usually did when doing images, but now I just copy pasted whatever infinity said on the assumtion that He Knows What He's Talking About (TM) [14:41] your dbus errors are very likely due to missing /proc [14:41] well dbus fails to init in the postinst hook [14:41] ok, let's try again [14:41] he should just have pointed you to the panda page [14:41] what was the problem with rootstock again? [14:42] i was actually planning to copy over bits from there to the ubuntu wikipage ... biut then i changed teams and now its just sitting at the bottom of my TODO [14:42] basically I need something like that but very simple, something that linaro-media-create coudl consume [14:42] janimo, that we dropped support for it :) [14:42] janimo, but your boss took over, ask him ;) [14:42] make-ubuntu-rootfs quantal armhf --extras=ubuntu-desktop, xmoto [14:42] done [14:42] yeah, rootstock at its best :) [14:43] ogra_, I know , Chris pinged me with a config he uses with live-build [14:43] still why there is no simple tool for such a common task is beyond me [14:45] janimo, well, there is again... now that rootstock was revived [14:58] jcrigby: I'm looking at bug 927781 now, and do you know why we didn't reviewed/applied the patch described at the bug? [14:58] Launchpad bug 927781 in u-boot-linaro "PXELINUX implementation doesn't respect dhcp ConfigFile or PathPrefix values" [Undecided,In progress] https://launchpad.net/bugs/927781 [14:58] might still be useful for Daviey I guess [14:59] Daviey: jcrigby also applied a few PXE patches on the latest u-boot-linaro: http://git.linaro.org/gitweb?p=boot/u-boot-linaro-stable.git;a=summary [14:59] don't know how useful they can be for you at this point, but jcrigby should be updating the package at quantal soon as well [14:59] rsalveti: ooo, will check that out in a bit.. thanks!! [14:59] rsalveti: do you know if local boot will appear ? :) [15:01] Daviey: what you mean exactly? [15:06] rsalveti: PXELINUX config that says.. boot from local disk. [15:06] rsalveti: sorry, OTP right now [15:07] Daviey: I think so [15:07] rsalveti: I appreciate there is complexity there [15:07] pfft, rsalveti eats complexity for breakfast [15:08] :P [15:08] /nick complexity [15:08] rsalveti, btw, A3 is on thu. any news about the PVR driver ? [15:08] (else i need to release note it) [15:08] ogra_: one thing we'd like to check as well, at least with panda, is that if we can enable booting the kernel from zimage and from the usb disk [15:09] booting the kernel from zimage ? [15:09] you mean kexec ? [15:09] instead of uImage from first SD partition [15:09] ah ! [15:09] so we could from u-boot, look at the zimage file at the usb disk, and load that instead [15:10] ogra_: for pvr it'd be good to get it after A3, it's not in a good shape still, and we'd need to update the xorg driver, pvr driver and probably xorg-server as well [15:11] but I believe we'll have to go with the current version anyway, at least to have something [15:11] and ppisati works on a new kernel too that might need adjust ments [15:11] from past experience, waiting a new driver from TI can take ages :-) [15:11] TI/SGX :-) [15:11] no prob, i can release note it ... its just that its properly integrated into ubuntu-drivers-common now and everyone will get it offered right after install [15:11] ogra_: yeah, what reminds me I need to sync with him about the 3.5 kernel [15:11] ++ [15:12] rsalveti, re the pxe magic patch. I think I lost track of that bug sometime ago. No reason. [15:12] Daviey: is bug 927781 still relevant for you? [15:12] Launchpad bug 927781 in u-boot-linaro "PXELINUX implementation doesn't respect dhcp ConfigFile or PathPrefix values" [High,In progress] https://launchpad.net/bugs/927781 [15:12] ppisati: ping [15:13] rsalveti, i think there was another similar bug from rbasak btw [15:13] rsalveti: pong [15:14] and indeed i cant find it :/ [15:14] ppisati: I'm curious about how you might be maintaining the kernel for omap4 now [15:14] are you planning on switching to 3.5? [15:14] rsalveti: yes [15:14] rsalveti: wait [15:14] ppisati: are you tracking and doing the forward-porting by yourself? [15:15] rsalveti: yes [15:15] rsalveti: so far i ported around ~500 patches [15:15] rsalveti: of 1700 [15:15] so i'm still in the middle of it [15:16] oh [15:16] would probably be a sane idea to sync with andy green as well [15:16] wow, is it really that many ? [15:16] rsalveti: I'm not sure.. can i get back to you? [15:16] (OTP) [15:16] as he'd probably be doing the same quite soon [15:16] * ogra_ thoguht there were only powermgmt and framebuffer issues [15:16] Daviey: lazy bastard, need to learn how to multitask ;P [15:17] Daviey: but yeah, sure :-) [15:17] i'm surprised that results in 1700 patches [15:17] that's just... *huge* [15:17] yep [15:18] jcrigby: would be nice to review the current open bugs for u-boot-linaro [15:18] we could probably have a few other interesting ones [15:18] and probaly a few that are already fixed as well [15:18] rsalveti: TI told me they will skip 3.5 entirely [15:19] ogra_: another thing we need to change is to start using uEnv.txt instead of boot.scr [15:19] ogra_: i tried cherry-picking some stuff [15:19] ogra_: but it was impossible in the end [15:19] ogra_: the more i advanced [15:19] ouch [15:19] ppisati: hm, then why we're forwardporting to 3.5? [15:19] ogra_: the more new patches relied on some that didn't were suppoed to be there [15:19] ogra_: etcetc [15:19] so it can be maintained by the kernel team as well? [15:19] rsalveti: because our master is 3.5 [15:20] rsalveti: and we want it to be a rebased topic branch [15:20] rsalveti: as we did since natty [15:20] oh god, I see a lot of pain there [15:20] rsalveti, yeah, i feared the switch to uEnv.txt [15:20] especially since we would have to hack up flash-kernel a lot again [15:20] anyway, i'll shot andy an email to tell him that i'm doing it [15:21] btw, so far so good [15:21] boards boot, etcetc [15:21] so many kittens... [15:21] ppisati: cool, yeah, please always keep him in the loop [15:21] (i'm still working on the flash-kernel 3.0 fallout in many areas, even though we switched before A1) [15:21] as I know he's crazy sometimes and also work heavily on forward-porting stuff [15:21] "everytime you use something different from traceroute, Van Jacobson kills a kitten" [15:22] today's facebook status :) [15:22] pfft, ubuntu hasnt used it in years :) [15:22] dont be reluctatnt to progess ! [15:30] janimo: Oh, I "volunteered" you as a possible person for mono issues and then got distracted by other things and forgot to talk to you about it, apparently. [15:31] janimo: Would you be interested in making the hardfloat port actually work, instead of being the icky softfloat-on-hard thing that it currently is? [15:32] infinity, I was pinged by directhex today and started looking at it [15:32] not sure how much work it is or whether I'll be able to do it but I am setting up the build tree, writing my first C# hello world, etc [15:33] I have a simple app calling floor() from libm which misbehaves as the minimal testcase now [15:44] janimo: Well, there's already "hard float" support in the upstream code, sort of, in bits. But turning it on doesn't work. :P [15:45] janimo: And, of course, the current build doesn't work for other reasons (ie: can't link to C libraries usefully). [15:50] * ogra_ sighs ... [15:51] i find the fact that my install tests now take 1.5h vs 10min pretty annoying [15:51] but i guess i cant have everything [15:53] ogra_: Yeah, but on the bright side, fewer images to test. :P [15:53] well, testing 3 desktop images simultaneously i was done in 10min in the past :) [15:54] now the only fast one is ac100 [15:55] gar and omap4 doesnt init the display again :( [15:56] infinity, ah so the situation with mono is different than what I was presented :) [15:56] janimo: What were you presented with? [15:56] as in, no hard float abi at all, everything else works :) [15:56] janimo: Well, that's sort of true. But only because we're not enabling the HF ABI bits, because that appears to work even less. [15:57] which would have been just a matter of writing new code - not necessarily easy but also not mono style wild goose chase [15:57] I thought the work needed is fairly well localized in the tree [15:57] It might be. I'm not sure. :P [15:58] so our current build does not work on armhf either because it cannot link to C libs? [15:58] My goal was just to make it build enough to unblock all the dep-waits, and then I ran out of time to care about it working properly. [15:58] janimo: No, it "works". Well, it also "links". What I meant was that it can't correctly call into them, because, well, it's not HF. [15:58] janimo: So, any float calls explode, obviously. [15:59] infinity, ah ok, so just the FFI is broken then as expected. phew :) [15:59] janimo: It works just enough to be used as a build-dep to make the rest of the archive build, it works for pure mono code (but it's not HF), it breaks for linking to other languages in a floaty fashion. [16:00] janimo: Working towards making the p/invoke tests pass would probably be enough to say "hey, it works". [16:00] wow, omap4 display wasnt broken, its just that plymouth bug and that casper takes over 10min to boot [16:01] alright then. I'll be on it this week along with other tasks, see if I can understand the issues enough that there is a chance they don't get cut off from wheezy - the reason I got pinged today [16:01] so i sit in front of a black screen for 10min [16:01] I saw the general call to action in february but I hoped people who actually care about mono will solve it, upstream included [16:01] * ogra_ really expects a lot of complaints from the community about the switch to live images [16:01] they didnt get even remotely faster [16:02] ogra_, I would take them even if they were slightlu slower just for the amount of maintenance work being cut [16:02] ogra_: Live images woudn't be faster, they only allow an easier way to install on faster media. [16:02] then any optimization in live will translate here too [16:02] janimo, well, its adds 1h20min tzo the install process [16:03] also, easier wa to install is something that should be synonimous with ubuntu [16:03] ogra_: Slow install that leads to faster runtime doesn't hurt my feelings any. [16:03] the current preinstalled are all but easy [16:03] ogra_: You forget that most people install once or twice, not over and over again like you. :P [16:03] the copying step takes about 1h ... booting alone takes over 10min thanks to casper etc [16:03] infinity, agreed ansd i dont want to go back [16:04] but we will definitely get complaints [16:04] Yep, we will. [16:04] But we always do. [16:04] heh, true [16:04] Our community is small but remakably vocal. :P [16:04] remarkably, too. [16:04] ++ [16:04] Oh, if only you had someone dedicated to testing these images. [16:04] GrueMaster: Then you'd be the one complaining. ;) [16:04] Well...yea! [16:05] But at least my bitchiness can be mellowed with beer. [16:05] GrueMaster, we will ... i'm in boston next week to take care for that [16:05] at least for the automated side [16:05] s/automated/automatable/ [16:06] Good luck with that. [16:07] Server side should be 100% automatable. I had a start on some good ideas & started laying the groundwork, but was cut short. [16:07] Not sure how you would automate the desktop stuff. [16:07] GrueMaster, arm images are built identically to x86 now ... the only bit i have to solve is the booting [16:07] (thats for server as well as desktop) [16:08] That part is easy. Use usb boot to boot a custom kernel/initrd that pulls an image from the network and flashes the SD. [16:08] Yeah, they already do that. [16:09] So the hard part is automating the desktop install. On x86 it is done through libvirt. [16:15] * janimo is looking forward to the ac100 live installer :) [16:15] janimo, that wont happen [16:15] uness you want to do it (and write a ton of partman reciepes to actually make it work) [16:15] so all that nice talk above and turns out you are still evil [16:16] lol [16:16] i'm just lazy [16:16] better than evil for sure :) [16:16] no, seriosly, before adding a live image for ac100 i would rather create a minimal server one for it :) [16:16] so that people dont have to remove unity :) [16:17] which apparently every second seems to do [16:17] I'd rather brick all the ac100s, pretend Tegra2 never happened, and will a new ARM netbook into existence. [16:17] ogra_, why not make the default image xubuntu then or something [16:18] infinity, so bribe KEybuk that his team finally releases something arm based ;) [16:18] infinity, the omap3 sucks more than tegra2 :) [16:18] janimo, because i dont want to invest more time than it takes to keep tehm installable [16:18] janimo: omap3 has neon, at least. tegra2 is the source of so many headaches. [16:18] as a product, not as something to deal with internally [16:19] janimo: If it wasn't for tegra2, I could just say "armv7 == neon" (as so many upstreams mistakenly do), and carry on. [16:19] I thought marvell too had some neon-less armv7? [16:19] Not any that we care about, if they do. [16:19] but agreed, stupid thing from nvidia to hope their GPU is a good enough neon replacement [16:20] Tegra3 is, by all accounts, quite lovely. [16:20] I'm kinda hoping the Trimslice guys upgrade. [16:20] I'd even consider buying a T3 trimslice. [16:20] Marvell, possibly Calxeda. Not sure about the Samsung SOC. [16:20] infinity, I'd settle for a tegra3 tablet where they max out the RAM [16:58] come on pandaboard, pedal faster ... [16:59] * ogra_ is still in the same install ... [16:59] and its still removing packages ... [17:16] Hence why I had 7 pandas running tests. :) [17:19] I wish there was a way to attach a few SATA interfaces to my panda... [17:19] RoyK: USB-Sata? [17:19] aka slow-as-hell-disk-io [17:20] Not true sata for sure. [17:20] so I'm using a good-old pc mobo for this :) [17:20] RoyK come on, at least you can create users with the gui now after your bug from precise was finally fixed :) [17:20] yeah - saw it [17:20] But a SSD in a usb case is decently quick. [17:20] haven't test yet, though [17:21] GrueMaster: SSD on USB is like a ferrari with tractor wheels [17:22] Heh. [17:26] except i'd pay good money for a ferrari with tractor wheels [17:27] just to pay more money to get proper tires ? [17:28] ogra_: If you have to ask why, this isn't for you. :P [17:28] heh [17:31] say, an Intel 520 SSD connected on USB2 - whoooaahhrghghhzz! [17:38] RoyK: you can do sata via the GPMC, but it takes some work [17:44] RoyK, you mean usb3 :P [17:45] scientes: afaik the samba doesn't have usb3 [17:45] erm [17:45] panda [17:46] prpplague: if that's the same thing as running the SD card, it's limited to about 32MB/s, which isn't very amusing [17:46] it really just needs sata [17:47] but then again - not an issue - I thought I could use a panda for a homeserver, but found I couldn't, so I use a fat pc instead :D [17:48] RoyK, i use a sheeveplug w/ sd, but a hard driver with sata would be nicee [17:49] RoyK: no, i get about about 120MByte/s transfers [17:57] prpplague: hm... good :) But I guess my RAID-6 on seven drives may live better on a fat box ;) [17:58] RoyK: indeed, but your commend was to attach a few sata to the panda [17:59] yep - and that was the original question - then I found I needed more space and ... [18:27] I have a beagleboard C3. I have successfully installed Ubuntu-arm on it with Xfce desktop envorinment. I have been using it with a dvi monitor. Now I wanted to switch to svideo. I have edited uenv.txt in boot partition. I have set the display as tv and mode as 640x480@60. Unfortunately, I can not see the desktop with my display. [18:27] I can see the kernel booting up phase [18:28] But not the desktop, any ideas on how to find a cure for the case? [18:30] I can see lightdm is running under the list of top [18:38] I have trid to sudo dpkg-reconfigure lightdm but no chance :( [18:39] I have manually stopped and restarted lightdm through rs232 link but still no luck :( [18:40] I am using ubuntu 12.04 by the way. [18:41] Kernel 32.13 [18:41] Kernel 3çü2.13 [18:41] Kernel 3.2.13 [18:45] There is a xorg.conf file under boot partition table. I have deleted it so that xorg should reconfigure itself. But still the same :( [18:47] kunguz: "kernel 3.2.13"? So, not an Ubuntu kernel? [18:50] infinity: that's what I end up when I type in uname -a [18:50] kunguz: Right, which means it's not an Ubuntu kernel. [18:50] infinity: I did all the updates with sudo apt-get update and sudo apt-get upgrade [18:51] If you got that image from anywhere other than ubuntu.com, you might want to talk to the people who made it. [18:51] Perhaps in #beagle [18:51] Linux omap 3.2.13-x7 is the beginnig line what uname returns [18:51] Yes, see above. [18:51] (Where did you download the image from?) [18:52] infinity: this is where I found the image: http://rcn-ee.net/deb/rootfs/precise/ubuntu-12.04-r4-minimal-armhf-2012-07-16.tar.xz [18:52] infinity: so I need some image without a desktop env. [18:53] infinity: any suggestions? [18:53] infinity: I do not want to run gnome on a 256 mb ram device :D [18:53] kunguz: The server image from http://cdimage.ubuntu.com/releases/12.04/release/ [18:54] kunguz: Not that I'm saying that the Ubuntu kernel would treat you any better than rcn-ee's, but at least I know how it works. :P [18:54] kunguz: But you could also talk to the folks in #beagle (rcn-ee hangs out there too) about rcn-ee's image. Either way. [18:54] Our kernel comes with security updates and such, though. [18:54] infinity: thanks, I will try my chances over there [18:57] kunguz, that is out of date too for mainline 3.2, the current 3.2 stable release is 3.2.23 [18:58] which is the base of the ubuntu 3.2 kernel (they just fake the version numbers in a confusing way) [18:58] scientes: I have all the repository in my update list. When I give sudo apt-get update and sudo apt-get upgrade, it should upgrade itself, right? [18:58] No, cause you don't have any Ubuntu kernel packages installed. [18:59] infinity: and the way to install them is? [18:59] You could try installing linux-omap, but rcn's kernel being a higher version, flash-kernel will be a bit confused, unless you move his out of the way. [19:00] infinity: sudo apt-get install linux-omap? [19:00] kunguz: Yeah. But, read above. May not do what you think it should because of the "newer" kernel already there. [19:01] infinity: I am giving it a try to see if anything goes better :) === zyga is now known as zyga-afk === Ursinha` is now known as Ursinha === Ursinha is now known as Guest69114 [19:06] there was a loss of internet connection, now I am back. I hope I did not miss any suggestions [19:08] Ok, now my beagleboard C3 is installing linux-omap. It will take a while [19:09] it is great that beagleboard provides usb-otg connection so that I can share my internet with the beagleboard. [19:15] Now I am rebooting my beagleboard [19:17] although I installed the new kernel, it appears that it boots to old kernel [19:18] how do I make it boot into the new kernel? [19:20] there is no grub :) so I guess I have to change something in uEnv.txt? [19:22] any directions might help :( [19:24] Guys, I really need help. Can anyone at least comment? [19:28] how do the location of the kernel image? so that I can boot using something like bootm 0x80300000 [19:29] kunguz: Like I said, flash-kernel would have done the right thing if the other (newer-versioned) kernel weren't there in /boot. [19:29] Well, this is assuming that in all other respects, that's a "normal" Ubuntu image. [19:29] And I really can't say, cause I don't use rcn-ee's images. [19:30] infinity: thanks anyway [19:34] infinity: I am going to reinstall ubuntu with a server-image. [19:35] infinity: preinstalled-server image is without a desktop and it is light compared to the ones with desktop? [19:36] kunguz: Yeah. Ignore the size of the image itself, that's mostly just a package pool on the image, the actual software installed is very minimal. [19:40] 600 mb is quite large, I hope this solves my problem. [19:41] if you need to trip space after install you ran remove /usr/share/doc/* [21:47] desktop is broken with last update. login screen is all i get. [21:47] hi ppl [21:47] anybody here with a galaxy tab 2? [21:55] pandaboard ES (omap4)