=== Guest46240 is now known as trelane
trelaneI have followed the instructions on http://www.freepbx.org/trac/wiki/UbuntuServer, and am booting a gumstix, I get the following output and then it hangs. http://pastebin.com/7u5CBQq3.04:11
trelaneI did, then did not copy the boot.scr (because the geometry is for a panda not a gumstix)04:11
trelaneis the output from ls -asl on the fat32 partition04:16
twbtrelane: "Loading u-boot.bin from mmc" is the last message?04:27
twbMost obvious guess is the system is not configured to send boot-time console, and the login screen to the serial port04:28
trelanecould be04:30
trelanein fact, not unlikely04:30
twbNote that it's not always ttyS0 on embedded ARM04:32
twboften it's ACM0 or some other fun thing, I dunno about your device04:32
trelanettyS2 on gumstix04:37
trelaneI didn't set it04:37
trelanegive me a few04:37
twbYou need to do it in two places: first, console=/dev/ttyS2 passed to the kernel, and second you need an /etc/init/ttyS2.conf that is similar to the ones for tty0.conf et al04:37
twbI am speaking in general; I am not an expert re. arm04:38
twbtrelane: out of curiosity, where is the rootfs in your system?  In the MLO/ dir?04:44
trelanesecond ext4 partition on the SD card04:53
trelanepartition 1 has MLO u-boot uImage boot.scr04:54
twbAh, righto04:54
trelanepartition 2 has the ubuntu-base04:54
trelaneextracted as an ext4 partition04:54
trelaneI'm playing with the boot.scr script04:54
trelaneeven if I can't get IT to output to /dev/ttyS2, I _SHOULD_ be able to get the kernel or at least getty to do os04:54
trelanereally I'm not picky04:55
trelaneas long as I get a boot prompt04:55
twbWe'll see if you feel the same AFTER it stops working for some strange reason and you want to debug it ;-)05:01
IamTryingIs there any latest release for Ubuntu? Last-time i had problem to keep Ubuntu inside the "Eee pad transformer" where i need to do manually boot else it does abnormal loop .09:19
twbIamTrying: you have a TF101?09:23
shaolatwb: are you trent?09:23
IamTryingtwb, exactly TF10109:24
shaolahi, iker09:24
twbshaola: I'm also going home in about five minutes, sorry09:24
shaolai got working multitouch in debian armhf09:24
shaolai was writting you an email09:24
twbIamTrying: lilstevie has some pages and code on the xda web forum, which you can use to get oneiric working OK09:24
twbshaola: cool09:24
IamTryingtwb, i think i did that and i end up with this: http://i.imgur.com/AGSzH.png09:25
twbI get that sometimes when trying to shut down / reboot09:26
twbEspecially if I halt/reboot while the dock is connected09:27
twbOr it may be a completely different koops to mine, I can't see there the actual error09:27
twbIamTrying: you are using the OLiFE installer that has the interactive prompting menu stuff?09:27
IamTryingYes for the very first time i had a luck one time and after reboot / shutdown since then i am lost i can not use it anymore.09:28
twbIamTrying: if so, you have to pick the first option (Android + Linux, default to Android) or it has misc bugs09:28
IamTryingAlways using OLiFE yes, thats they best one well done.09:28
IamTryingThere was a problem bug, like you can not auto mate it09:28
twbOLiFE is made by lilstevie, who is here but probably away because he has a girlfriend09:28
IamTryingYou have to press button up or something to boot09:28
twbIamTrying: right09:29
IamTryingYea sure.09:29
shaolawe are talking in #asus-transformer09:29
twbYou have to hit Power+VolDN during the early boot, and then it will prompt to press VolUp which if you do, will boot oneiric09:29
twbshaola: ah, thanks09:29
* twb adds it to autojoin list09:29
IamTryingThe problem i there is no stable release with OLiFE for this? So that i can put one Ubuntu stable now like my Android tablet is "he is dead, useless"09:30
twbIamTrying: this is all alpha-quality early days stuff, sorry09:30
lilsteviethere is no stable09:30
lilstevienot yet09:30
IamTryingWe really need one :)09:30
twbIamTrying: but it is (almost) impossible to brick it completely09:30
lilstevieI am 1 person, who has very very little contributions back09:30
twbIamTrying: even if you mess up the install you should be able to reinstall again with nvflash09:31
lilstevietwb, the only way to brick is hardware damage09:31
twblilstevie: because you have a girlfriend to waste your time  :P09:31
lilstevieAPX is a bootrom mechanism09:31
lilstevietwb, no I mean like nobody else contributes09:31
twbYeah because I'm a lazy bum09:31
lilsteviethe only *real* code contribution has been the network-manager patch for the bcm4329 driver09:32
IamTryingCan we not put Tizen at-least so that we got some options if Ubuntu not then at-least Tizen or other Linux to survive without built-in Android.09:32
twbAnd my old machine is still bricked so I can't mess with my TF easily :-(09:32
shaolalilstevie: i am trying to do something, but i am not that good09:32
twbshaola: even testing and reporting bugs is helpful IMO09:32
twbAnd writing documentation09:32
shaolayes, i know09:33
shaolai try to do my best in debian09:33
shaolareporting bugs, writing for people who knows less than me, i even mantain a simple package09:33
twbPub time, bye09:33
lilstevieyes, even documenting and bug reports are useful09:34
lilsteviebut even that is lax on contributions09:34
shaolalilstevie: are you DD?09:34
shaolay know twb is DD09:34
shaolaDebian Developer09:35
lilstevieI am unaffiliated09:35
shaolaso tell me09:35
shaolawhat do you need?09:35
lilstevieI am a school teacher :p09:35
shaolacongratulations for having the best and the worst profession in the same time09:36
lilstevieyep :p09:36
lilstevieanyway, what I need is people to take some of the load off me for supporting everything09:36
shaolai am not good programming09:37
shaolathat's my problem09:37
lilstevieaka; support for users, documentation, kernel patches09:37
shaolai learned so many years ago with qbasic manual in my home when i was a kid09:37
shaolabut ,... basic is not programing09:37
shaolaanyway, i am kind of busy till mid february09:38
lilstevieheh, I'm not the best programmer, but I have been getting better, well when I am not making stupid mistakes like sizeof(pointer)09:38
lilstevieanyway, as two said before, I have a girlfriend, and she wants to go to the whops09:39
shaolayeah, the first is the first09:39
shaolai don't know if there is an english expresion like that09:39
shaolamaybe too literaly09:39
shaolabut i think you know what i mean09:39
shaolaso, go and enjoy (aghhhh) shopping09:40
Guest9392slangasek: could you please have a look to this bug/patch https://bugs.launchpad.net/ubuntu/+source/multistrap/+bug/874505 TIA11:40
ubot2`Launchpad bug 874505 in multistrap "Native Multistrap oneiric chroots have an error configuring base-files" [Undecided,Confirmed]11:40
=== Guest9392 is now known as zumbi
=== taruti_ is now known as taruti
=== sebjan_ is now known as sebjan
* ogra_ twiddles thumbs watching an ac100 alpha2 testinstall13:12
ogra_hmm, no slideshow on the ac100 installer13:20
Davieyogra_: What do you want, sparkly pony pictures?13:26
ogra_just the std ubiquity slideshow :)13:29
* lilstevie totally thought Daviey said "sparkly porn pictures"13:33
ogra_he meant to say that but typoed :P13:38
=== dduffey_afk is now known as dduffey
jayson_Hi, has anyone got ubuntu working on a qualcomm board15:07
jayson_I have a dragon board with me & its got android15:07
jayson_I can flash android related images onto it using fastboot15:07
ogra_jayson_, the guys in #linaro do afaik15:07
jayson_but I want to wipe out fastboot & try to install uboot15:08
jayson_thanks ogra_ - I'll give it a shot15:08
shadeslayerlilstevie: dude, is there any hope of getting a native ubuntu boot on a device with SBKv2 ?15:37
shadeslayerYou mentioned last time that there might be a way ... haven't heard anything for the past 2 months or so15:37
ogra_GrueMaster, did you get a kbd selection dialog in any of your install tests the last days ?16:09
GrueMasterYes, on omap/omap4/mx5.16:09
GrueMasterI think.  I'm pulling the latest builds now, will test very soon.16:10
lilstevieshadeslayer, yes, but I have been really busy the past month or so16:29
shadeslayerlilstevie: understood, I'll patiently wait, let me know if there's anyway I can help :)16:30
shadeslayerwith testing and stuff16:31
lilsteviewell bed time16:31
shadeslayernight :)16:31
GrueMasterI'm wondering if bug 838200 is u-boot or kernel related.  It is highly annoying, whatever the root cause.17:42
ubot2`Launchpad bug 838200 in u-boot-linaro "No network support on Beagle XM" [High,Confirmed] https://launchpad.net/bugs/83820017:42
GrueMasterogra_: keyboard selection comes up on omap4 desktop.  Could be the images don't like you.  :P18:32
ogra_bah, ok18:32
ogra_might be that something clashes with the new wlan step in ubiquity18:33
ogra_i assume you dont get that on panda18:33
GrueMasterI have all my systems wired.  I could try wireless during install (I usually wait until login to test it).19:27
GrueMasterSigh.  Analog audio is broken again on omap4.19:27
dioxinHi guys, I'm trying to force ubuntu to start with a terminal rather than gnome on a pandaboard, anyideas?19:49
GrueMasterUse the preinstalled server image instead of preinstalled desktop.19:50
dioxinok, but going that route I wasnt able to get gnome running at all19:51
GrueMasterHmm.  There may be a way through one of the /etc/init conf files.  I just don't test that.19:53
infinityWhen you say "a terminal", do you mean a text console, or a GUI session with a terminal application?19:53
dioxintext console19:54
infinityIf the former, just disable lightdm.19:54
infinityOr gdm, on older releases before we switched.19:54
dioxinI've ssh'ed in, stopping lightdm doesnt bring the main display to a console19:55
infinityWhat does it do?19:56
dioxinit gives me a display of recently stopped processes, but doesnt give me commandline access19:56
dioxinok that worked19:58
infinitydioxin: If you add "text" to your kernel command line, that should prohibit fancy graphical stuff from starting on boot.20:00
GrueMasterdioxin: What happens when you move /etc/init/lightdm.conf to a different directory and reboot?20:00
infinity(Alternately, yeah, you can mangle/remove lightdm's config)20:00
dioxinwhere are the kernal command20:02
dioxinwhere are the kernal commands stored, i.e where do I edit them20:03
RoyKhi all. any idea why lshw is dying on me on this pandaboard? http://paste.ubuntu.com/825549/20:03
GrueMasterdioxin: I think you are referring to /boot/boot.script if you want to change the kernel boot parameters.  You will need to run "sudo flash-kernel" after editing this file.20:04
infinitydioxin: sudo sed -i -e 's/splash/text/' /boot/boot.script && sudo flash-kernel20:05
GrueMasterRoyK: What kernel/Ubuntu release?20:05
RoyKGrueMaster: http://paste.ubuntu.com/825556/20:06
infinityYeahp, SIGBUS here too.  Wonder why no one's ever noticed that before. :P20:07
infinityRoyK: Can you file a bug?20:07
infinityRoyK: To be fair, even if fixed, it won't be wildly informative on ARM.  You'll get USB, and that's it.20:08
infinityRoyK: So, lsusb will work fine for you. :P20:08
infinityRoyK: But it should still be fixed to not break.20:08
RoyKhm... so where'll the wifi chipset be connected on this pandaboard? on usb?20:10
infinityIf only you were so lucky.20:10
infinityNo, there are a few devices connected on (usually SoC-specific) busses on all ARM devices, but they're not probable, and until either DeviceTree or ACPI get here, they's also not remarkably easily enumerated.20:11
infinityWalking /sys after you already loaded all the drivers would work, but that's not how lshw traditionally works.20:11
dioxinhmm even if I remove splash and replace it with text , it still boots to a session manager20:13
infinitydioxin: oneiric?20:13
infinityHuh.  That was fixed.  At least, it was here.20:13
GrueMasterdioxin: move /etc/init/lightdm.conf to a different directory.  That works here.20:13
infinity        for ARG in $(cat /proc/cmdline); do20:14
infinity            if [ "$ARG" = "text" ]; then20:14
infinity                plymouth quit || :20:14
infinity                stop20:14
infinity                exit 020:14
infinity            fi20:14
infinity        done20:14
infinitydioxin: Did you change actually "take"?  Do you see a sane command line in /proc/cmdline?20:14
dioxinseems so, cat of cmdline seems normal20:15
infinityAnd by "normal", you mean "contains the word text"?20:16
dioxinro elevator=noop vram=32M mem=456M@0x80000000 mem=512M@0xA0000000 root=UUID=4ad147d9-7e0b-4fb9-af24-940fe60ac01f fixrtc quiet text20:16
infinityHrm, kay.  Then lightdm starting is just plain wrong...20:17
dioxinjust moved lightdm.conf to lightdm.conf.bak20:18
dioxinand I get a cmdline straight away20:18
infinityAs you should, sure.20:19
infinityBut text on the command line should have the same effect.20:19
infinityUnless you don't have the above code in your lightdm.conf20:19
infinityIn which case, I question if you're actually running oneiric. :P20:19
infinity(Note that that was fixed in November in oneiric-proposed, so if you haven't upgraded since the install, you wouldn't have it)20:20
RoyKwhat's the reason for these custom buses?20:21
infinityRoyK: Eh?20:22
GrueMasterCustom busses?  They are actually fairly standard for embedded devices.20:22
infinityRoyK: Every system has custom busses.20:22
infinityRoyK: Even x86 systems.20:22
infinityRoyK: The difference is that x86, powerpc, sparc, and a few others, enumerate said busses in a standard (OF/DT/ACPI) way.20:23
GrueMasterCustom to me implies non-standard.20:23
infinityRoyK: ARM hasn't, traditionally, but things are moving that direction.20:23
infinityGrueMaster: Yeah, I'm going to stick with "custom", even by your definition. ;)20:23
RoyKfor ARM to gain wide support, I guess that's a must20:23
infinityGrueMaster: Stuff hanging right off the SoC isn't in a particularly standard formart.20:24
infinityRoyK: You realise ARM CPUs are in more computers worldwide than any other architecture?20:24
infinityRoyK: I'd say they have "wide support". ;)20:24
infinityRoyK: But yes, we all want better device enumeration.20:24
RoyKinfinity: I know, i know...20:24
RoyKinfinity: but for ARM CPUs to be used in "traditional" computers, you really want device enumeration to work20:25
infinityRoyK: *hand wavy*20:25
infinityRoyK: The only place where it really matters is being able to ship a single kernel that boots all devices.20:25
* RoyK has read a bit about using ARM for HPC, but it seems memory buses on ARM are still slow...20:26
GrueMasterNot that I have time to debate this, but nothing hangs right off the SOC.  The SOC has internal busses, same as Intel with an external south bridge (just in one convenient package).20:26
infinityRoyK: Which is something Microsoft is pushing.  No Win8 logo without ACPI.20:26
infinityGrueMaster: The SoC has internal bridges on the SoC, hence some things hang "right off it".20:26
infinityGrueMaster: That's the definition of an SoC.20:26
infinityGrueMaster: Many devices are right on the SoC, even.20:26
dioxininfinity: apt-get upgrade 'ing now20:27
infinityGrueMaster: And there's no standard way for how it's all put together, or how you discover what's there.20:27
RoyKinfinity: right on, as in "simply memory mapped..."?20:27
infinityRoyK: Yeahp.20:27
infinityRoyK: And nothing wrong with simple mmaped devices (it's the way sparc and powerpc have gotten by for decades), but you need a way to enumerate the devices in a system, since you can't probe that.20:28
infinitysparc and powerpc use device trees, x86 uses ACPI, ARM now has both DT and ACPI implementations, and very few boards supporting either. :P20:28
infinityRoyK: As for memory throughput, things have improved a lot lately.  And products that are "coming soon" are better still.20:29
infinityRoyK: But it's certainly nowhere near the throughput of, say, AMD systems.20:29
infinityRoyK: But hey, Intel can't beat AMD's memory throughput either. ;)20:29
RoyKheh - we use AMD rather a lot for HPC :รพ20:30
RoyKyou really want a good memory bus for 16 cores on a cpu20:30
infinityYeah.  We tend to buy mirrored AMD and Intel systems, and pick and choose depending on workload.20:31
infinityIf throughput is king, go AMD, if math is where it's at, Intel.20:31
dioxinand for portable?20:32
RoyKis really Intel that much better than AMD on fp?20:32
RoyKdioxin: portable HPC? ;)20:32
dioxinRoyK: no reason why not ;)20:33
infinityRoyK: Currently, Intel tends to beat AMD on most fp and int workloads.  But it's got to be pure number crunching and register flipping.  As soon as you start working with large datasets, AMD comes back strong.20:33
RoyKinfinity: the runs we have, usually have 1-4GB datasets per job20:33
RoyKso I guess we can stay with AMD a while...20:34
GrueMasterFor portable, I'd have to push Intel.  I have seen recent models of both Intel & AMD, and the AMD runs a lot hotter.20:34
dioxinGrueMaster: you were supposed to say ARM ;)20:34
GrueMasterWell for battery life and low heat, arm by far.  But until there is a current mass produced Arm netbook that supports Ubuntu, I have to stick with x86 hardware (AC100 doesn't count as it is discontinued and Asus Prime is not supported in our current releases).20:36
infinityYeah, I'm still waiting for a killer ARM netbook. :/20:39
infinityAnd with next-gen ARM cores, we could see ARM "ultrabooks" too.20:39
infinity(ie: fast and small)20:39
infinityBut I won't hold my breath until I actually see hardware.20:40
GrueMastersigh, why is it the platform I care the lease about has working audio ootb?20:42
infinitySpeaking of subarches, I should dig up a spare microSD, and get testing mx53 kernels...20:45
infinitySee if we can't get it bumped to 3.2 after the Alpha.20:45
GrueMasterThat would be nice.  The current kernel won't work on the newer rev boards.20:46
infinityOh, ick, really?20:47
GrueMasterNo usb.  It sits and reenumerates in a constant loop.20:47
infinityThat sounds special.20:47
GrueMasterI'll log it and file a bug, but if the new kernel fixes it, this is moot.20:47
* infinity nods.20:48
infinityDon't waste your time on it until we've upgraded.20:48
GrueMasterMeh.  Typical hw mod.20:48
GrueMasterI may need to file it for A2, so other users don't come whining.20:48
infinityThere's that, yeah.20:49
GrueMasterAlso, the system is horridly slow on SD.  beagleXM is easily 4x faster.20:50
infinityYeah, I use mine on SATA.20:51
infinityHaven't tested a pure SD setup on the mx53 since a day or two after I got it...20:51
GrueMaster[   57.856303] usb 1-1: device descriptor read/64, error -7120:54
GrueMaster[   58.626361] usb 1-1: device not accepting address 8, error -7120:54
GrueMaster[   59.286332] usb 1-1: device not accepting address 9, error -7120:54
GrueMaster[   59.292248] hub 1-0:1.0: unable to enumerate USB device on port 120:54
GrueMasterFrom the new mx53 board.20:54
GrueMaster(and of course we don't enable serial console on our desktop images, so I can't log in).20:55
dioxininfinity: the boot is now fixed after the upgrade21:06
dioxintakes long enough tho!21:06
=== doko_ is now known as doko
GrueMasterinfinity: Here's an interesting (kind of) bug.  Bug 925035 only appears to affect armel.  Not seeing it on armhf images.21:59
ubot2`Launchpad bug 925035 in unity-lens-applications "unity-applications-daemon crashed with SIGSEGV" [Undecided,New] https://launchpad.net/bugs/92503521:59
* RoyK looks at dioxin and wonders if he should renick himself as a pollutant too22:04
infinityGrueMaster: Reliably reproducible?22:05
* dioxin shakes his head at RoyK ..... "pollutant indeed" he muses22:05
GrueMasterBoth bladner and I saw it on mx5, I saw it on omap & omap4.  All armel images.22:05
infinityGrueMaster: That sounds fairly reliable.  Kay. :)22:05
GrueMasterNot sure how reproducible it is.  I'd have to delete the crash report & retry.  Too busy with milestone testing atm.22:06
infinityGrueMaster: Yeah, the anecdotal evidence that several people have seen it is enough for now.22:07
GrueMasterHmmm.  preinstalled-server fails to network via wifi on panda.  (could be my config though).22:09
dioxinI seemed to get that working when I config'ed the wifi via the install process22:10
dioxin(using PandaboardES22:10
GrueMasterdioxin: server or desktop image?22:10
dioxinserver image22:10
GrueMasterHmmm.  Could also be this particular board.22:11
GrueMaster(I have many).22:12
dioxinlucky sod ;)22:12
GrueMasterLucky?  I bought most of them for testing.22:12
GrueMasterThe more I have, the more I can test (to a point).  For example, I can currently test every flavor of Ubuntu Precise on omap4 at one time.  But I only have one beaglexm, and 8 images to test on it, so it is slower.22:14
dioxinI probably shouldnt complain... i have like 5 i5 systems floating around ;) plus others22:17
GrueMasterI also have a lot of other systems.  My office has 5 monitors and 15 systems currently running (2 x86, rest are arm based).  Not counting laptops or systems in my rack cabinet.22:19
dioxinarent they all company supplied?22:19
infinitySome are, but he likes to go overboard. ;)22:20
dioxinnext system I'm likely to go overboard on is RaspberryPi, but only cos per unit they are cheap :D22:21
GrueMasterI like having a home lab.  :P22:21
dioxinI need the time to build mine, components are in every draw I can find!22:22
GrueMasterdioxin: I would hold off.  I hear there is another platform in the works that is armv7 and almost as cheap.22:22
GrueMasterWe won't support raspberry pi.22:22
dioxinI have money to burn on both ;)22:22
dioxin1 pandaboard is like 6 RaspberryPi's I think22:23
dioxinI've heard wind of the armv7, but they arent even close to production yet22:25
RoyKdioxin: isn't 7 the current and 8 the next?22:33
RoyKwith 8 supporting 64bit etc22:33
dioxinI think for RaspberryPi its arm5 or 622:34
GrueMasterSort of.  Armv7 is 32 bit, armv8 will be 64 bit.  There are some things in between that have been announced as well.22:34
GrueMasterRaspberry is Armv6.22:35
GrueMasterKey difference is Thumb2 support, which reduces the size of instructions (I think).22:35
dioxinthats the R-Pi chip22:36
GrueMasterArm versioning prior to v7 is a bit wonky.22:37
dioxinsingle core only as well I think22:37
GrueMasterArmv7 is the first to introduce SMP in the SOC.  I think some vendors have done it with armv6, but only at the instruction set level (i.e they have their own core design).22:38
RoyKGrueMaster: *size* of instructions?22:38
GrueMasterThe Thumb (T32) instruction set provides a subset of the most commonly used 32-bit ARM instructions which have been compressed into 16-bit wide opcodes.22:39
GrueMasterFrom http://www.arm.com/products/processors/technologies/instruction-set-architectures.php22:40
dioxinwill Ubuntu Server ARM fit onto 4GBs SD?22:46
GrueMasterAll of our preinstalled images are designed for 4G SD or larger.22:48
dioxinand is there an image for a minimal file system with either SSH or serial?22:48
GrueMasterIt won't give you much room to really go wild, but it is good for some things.22:48
GrueMasterThe server image is about as minimal for a bootable headless install as we can make.  It is configured via serial console.22:49
desrthello arm people23:36
desrti currently have the quickstart board with linux-linaro-lt-mx5 kernel installed23:36
desrti just upgraded to precise and it seems like the latest version of that kernel is 2.6.3823:37
desrtis there a successor package, or is that the kernel i should be using?23:37
infinitydesrt: That's the current one still.  We're working on getting a 3.2 kernel into shape, but it's not in the archive yet.23:53
desrtcool.  thanks for the info.23:53

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!