[00:23] Hello Everyone, I recently got a donate board from TI. It's the LogicPD OmapL138EVM board. Would anyone happen to know if this board supports ubuntu11arm? [00:35] what processor is it? Arm11? [00:36] We only support Armv7 (Cortex A8 & Cortex A9). [00:37] howdy how GrueMaster [00:37] An-iSociaL: Hey. Get your kernel up? [00:37] GrueMaster: internet says ARM926 core [00:37] negative :( [00:38] not down yet though, still trudging [00:38] Aww, sad face. Good luck. [00:39] I'm EOD. Time for Skyrim. [00:39] laters [00:40] I think OmapL138--> A8, i believe [00:41] Oh whoops it has an ARM926EJ-S + C6748 VLIW DSP [00:42] any chance of seeing ubuntu on this system? i miss aptitude :( [00:42] bge0: debian armel then [00:43] ty! ill look into it [00:43] and howdy to you too twb, sorry i didnt see you there [00:43] An-iSociaL: whatever [00:43] lol [00:43] hmph [00:47] whatre you upto? [01:13] twb: I'm looking up debian armtel, but there seems to only be support for iop32x, ixp4xx, kirkwood and orion5x [01:15] http://wiki.debian.org/ArmEabiPort [01:16] You're not going to get turnkey installation, but it certainly seems to support OMAP1/2/3 boards [01:16] yea I figured it wouldn't be quite that easy [01:18] But if you aren't attached to the board I would suggest getting something armv7 flavoured, since everybody is switching to those [01:20] Well it's a pretty nice board though. i would feel bad scrapping it. It has a good onboard dsp too [01:22] Shrug [01:26] Well, I guess I might just stick with the linux port that the board came with [01:26] ty though === Jack87 is now known as Jack87|Away [01:34] An-iSociaL: that unable to open /dev/fb0 is indicating that the fb patch from the chromeos code review didn't take [01:35] aha [01:37] well im not ready to give in just yet [01:39] An-iSociaL: http://www.cactuar.net/uploads/Xoom_Debian/xoomkernel-gnu.patch over https://github.com/Tiamat%E2%80%93AOSP/Tiamat-Xoom [01:40] ah [01:40] ill have to manually patch that in [01:41] the AOSP tiamat is old [01:41] but im using his latest version [01:41] the only kernel i could get to ocmpile properly [01:41] compile* [01:45] right now i cant seem to generate another working initrd [01:51] :/ [01:51] i keep ending up with the installer [01:56] there.. === ericm-Zzz is now known as ericm [01:57] An-iSociaL: you want that installer "oem-config" so you can set up a user and shit [01:59] yea i got it, i think i was doing something ignorant by trying to use relative path for initrd output from mkinitramfs [01:59] did the initrd installer steps manually, format, extract, write fstab, touch oem-config so itll run === Jack87|Away is now known as Jack87 [02:00] got a nice bootup, only kernel module to fail was btfrs with unknown symbol _global_offset_table_ which is caused by my cross compiler [02:01] video failed, as if im shocked, it flashed a couple times trying to switch but never did [02:01] so, on to the patch [02:14] did you already update this patch lilstevie ? [02:25] not my patch [02:27] Sorry for interrupting but are there any team that works with Nokia N900? apachelogger, mpoirier, rbelem, someone else? [02:54] ok all patched up [02:54] booting [02:56] whatup! ubuntu is booted with xdesktop on my Motorola Xoom! [02:57] tyvm lilstevie [02:59] now next problem [02:59] there's no touch lol [03:00] An-iSociaL: What is the touchscreen hardware? [03:01] err, i actually dont know [03:01] Can you execute lsusb? :) [03:03] i would love to [03:03] but [03:03] no keyboard [03:03] lol [03:03] let me get adbd installed onto it, then i can use a usb terminal [03:03] the only touchscreen enabled in a default config is CONFIG_TOUCHSCREEN_QUANTUM_OBP [03:04] one sec on the lsusb [03:07] http://pastebin.com/iRgkqPnF [03:07] if youre the nerdy type like me, theres my dmesg log [03:09] Offtopic: Why 2.6.36? (I just started reading IRC) [03:12] An-iSociaL: ok, now that patch had my touchscreen patch [03:12] its just what was available [03:12] int_ua: because thats what exists [03:12] hm lilstevie... [03:12] i did all the sections, and reverted one change cuz it broke something [03:12] which one? [03:12] so, it should work then [03:13] the one you linked [03:13] http://www.cactuar.net/uploads/Xoom_Debian/xoomkernel-gnu.patch [03:13] yeah, which change did you revert [03:13] lilstevie: I asked because the only kernel I can boot on N900 is 2.6.35. And I don't know the reason yet. [03:13] just a define in fb_tegra which was being used by dc.c [03:13] also, YMMV with the touch patch [03:14] revert it [03:14] mmk lemme figure out where that is in here... [03:14] ubuntu from natty on, includes handling in evdev for what my patch does, that patch was done because debian doesn't have the same evdev handling [03:15] although it should just be ignored [03:15] touchscreen/qtouch_obp_ts.c [03:15] ? [03:15] yep [03:15] ok and away i go [03:17] bbs [03:17] lunch [03:17] k ttyiab [03:18] unpatched, compiled, flashed and now we're booting [03:20] ugh [03:21] still no touch and adbd didnt start [03:21] lets try this again [03:24] attempted fix on adbd [03:24] and im gonna have a smoke [03:27] i have touch... [03:27] its just... not quite right [03:29] and adbd didnt start again :/ [03:29] is there another way to get a shell over usb? [03:37] ok the touch device doesnt seem to register clicks [03:37] it thinks theyre mouseovers [03:38] and still no adbd [03:38] this is going to be difficult [03:40] ok, let me eat [03:40] I will gen a patch of it [03:40] oh cool [03:41] dont choke cuz you think im waiting tho.. enjoy it [03:41] ive been working on this for a solid week and this is the furthest ive been able to get [03:42] heh [03:42] ask twb I don't rush for anyone [03:42] lol cool [03:42] btw twb known bug exists on 32GB tfs with u-boot [03:43] sad face [03:43] twb: for some reason it will only rawread, and not read the partition [03:43] I haven't tried your latest .38 with asus bl yet either, because all these 360 games came out last week [03:43] even though it can locate it [03:44] hah [03:44] I want a 360 so i can play fable3 [03:44] and new games [03:45] cause my ps3 is 3.55-MFW [03:45] I have a 360 because 1) I've given up on the games industry getting FOSSilized anytime this decade; and 2) it's turn-key. I don't need to give a shit about h/w drivers or windows licensing or so on [03:46] lol [03:46] I also want to get a slim, and glitch attack it [03:46] linux + games on the 360 [03:46] if slim is the new black ones, they have really noisy PSU fans [03:47] indeed twb, you can just turn it on and play, no thinking about shit [03:47] i enjoy systems like that [03:47] twb: and the old ones were really loud with the exhaust fans [03:47] I would like to play online but then I have to give MS money, and the menus inside teh 360 become much more annoying [03:47] lilstevie: it's a fan inside the power brick [03:47] lilstevie: smaller than the one inside the main case on the white units, i.e. noisier [03:48] lilstevie: I have both units [03:48] heh [03:48] I don't really care [03:48] I guess if you have the volume of the game up it's not noticable [03:49] I usually mute it and read a book via synthesizer [03:49] I would be a lot happier with steam, I think, if ICBF dealing with the non-turn-key-ness [03:51] lol [03:53] Having to shuffle CDs is a pain, and I don't enjoy dealing with the monkeys at the game store [03:57] wow finally [03:57] http://pastebin.com/CuwfZqxW [03:57] lsusb [03:57] very short list [03:58] i somehow am pretty sure thats wrong [03:58] asin incorrect, not misreported [04:00] hm [04:00] corresponds with dmesg. So it must be connected somewhere else. lspci? lshw? [04:01] lemme run those out [04:01] im cheating and putting in startup scripts since i cant type lol [04:02] yeah, they don't connect via usb [04:04] if i dont get something workable tonight ill go buy a usb adapter tomorrow so i can plugin a keyboard+mouse [04:04] just for testing [04:16] lspci is empty [04:16] ofc [04:16] heres a neat empty lshw [04:16] nothing connects via pci [04:16] near* [04:17] http://pastebin.com/qJxwEgNf [04:17] yeah, arm connectivity doesn't really show up [04:17] seems like nearly nothing is working [04:17] ah [04:17] mmk [04:17] most things connect via i2c and i2s [04:20] ive got alot of work ahead of me here [04:20] milestone 1 has passed [04:20] heh [04:20] it doesn't go any quicker either [04:21] I have been working on the tf101 for months [04:23] hehehe [04:23] well anything gained too easily we esteem too lightly [04:23] or somethin like that [04:26] once i get this done my next goal is adual kernel [04:28] do a pendrive type linux, put it on my evo, run it side by side with android and use the hdmi out for a full desktop [04:28] then use a bluetooth keyboard n mouse [04:28] tiny pc [04:30] heh good luck with the hdmi part [04:31] shouldnt be so difficult, the hdmi under normal linux just shows up as fb1 [04:31] normal android* [04:31] there's just no virtual terminal devices to hook into [04:36] yeah, you would think so [04:36] but at least on my tegra device, it doesn;t really work [04:36] doesn't* [04:38] hehe [04:38] well its gonna be awhile off before i embark on that [04:38] right now im aiming at full ubuntu [04:38] then xbmc media center === Jack87 is now known as Jack87|Away [04:42] you will need accel for that [04:42] and that isn't going to happen with the kernel you are running [04:42] yea need to get the nvidia tegra drivers to function [04:43] likely try out the android 4.0 kernel for the xbmc build [04:43] it should be much more updated [04:45] still may not help [04:45] the android driver interface is different to the L4T interface [04:47] hmm [04:47] only problem is the damn stingray stuff [04:47] as i recall when i was on kernel.org, they have ventana in 3.0 [04:48] that doesn't mean anything really [04:48] itll help alot [04:48] nearly every device has a different pinmux [05:00] hmm [05:00] i think its crashing [05:11] yes! i have adb shell! [05:11] now i can do somehting [05:12] heh [05:24] so uhm, hows that lunch there? [05:26] eaten [05:26] got a link to the kernel you use [05:26] like its git [05:26] http://code.tiamat-dev.com/tegra2/tiamat-xoom/commits/tiamat [05:27] git://code.tiamat-dev.com/tegra2/tiamat-xoom.git [05:27] thats the one i cloned [05:34] yeah looking at the touch driver now [05:35] patch out input_set_capability(ts->input_dev, EV_KEY, BTN_2); and input_set_abs_params(ts->input_dev, ABS_HAT0X, [05:36] and any other HAT* [05:36] which is input_set_abs_params(ts->input_dev, ABS_HAT0Y, [05:36] mmk [05:36] and that should do it [05:39] sweet [05:39] thanks [05:39] we'll know in a couple mins [05:41] fingers crossed [05:47] and we're booting... [05:48] switched cross compilers too, the newer one was playing tricks with my kernel modules [05:48] thats a fail [05:48] heh yeah [05:48] awh damn [05:48] ok, have your brought up wifi yet? [05:48] cant get it to come up [05:49] unable to mount rootfs on unknown block [05:49] ok, insmod firmware_path=/path/to/firmware.bin nvram_path=/path/to/nvram.txt iface_name=wlan0 [05:49] wth [05:49] and use wpa_supplicant [05:49] kernel panic [05:57] ok, well you have adb [05:57] so you can transfer a file [05:57] find the evtest deb [05:57] just got the wireless module loaded... [05:58] its not coming up anyway [05:58] yeah, you need to use wpa_supplicant [05:58] https://launchpad.net/ubuntu/oneiric/armel/evtest [05:58] that? [05:58] yep [05:59] figure out which input event is the touchscreen [05:59] (it will be in the dmesg and the Xorg log [05:59] ) [06:00] ok cool [06:01] im goping nuts over here, this thing isnt really setup all that well [06:01] going* [06:01] can i paste here? [06:01] like 2 long lines [06:02] k [06:02] dpkg: error: parsing file '/var/lib/dpkg/available' near line 3214 package 'language-pack-es': [06:02] `Pre-Depends' field, invalid package name `d`kg': character ``' not allowed (only letters, digits and characters `-+._') [06:02] its jacked [06:02] hm [06:02] odd [06:02] you can always manually extract anyway [06:03] it is an ar archive [06:06] Event: time 172.655713, type 1 (Key), code 330 (Touch), value 0 [06:06] thats me picking my finger up [06:06] ok touch the screen a few times [06:07] move it around a little [06:07] do a two finger touch [06:07] then pastie the result [06:08] http://pastebin.com/6236hXgC [06:09] hmm [06:09] the X/Y seem out of bounds [06:09] one finger center of screen on then off [06:11] http://pastebin.com/9tx9b2BZ [06:11] bout as close as i can guess [06:14] you still have btn_2 and the hat stuff in the driver [06:15] its there, just commented [06:15] it isn't :) [06:15] not in what im using atm [06:15] if the set_abs_params and set_bit was out it wouldn't show up [06:15] the compile from the changes somehow caused a kernel panic [06:16] so i reflashed an older kernel image [06:16] maybe something went sideways, lemme try again [06:18] with set_bit for those parts evdev generally tries to treat the tsp as a touchpad [06:19] attempting to boot [06:20] panic [06:20] try going back to that other toolchain [06:20] k [06:20] it may be the change in toolchain that screwed it [06:21] recompiling [06:22] hey i really appreciate the help [06:23] np, I don't mind [06:24] rebooting [06:25] and panic [06:25] thats so weird [06:25] hm [06:26] let me try something else [06:26] I helped liv2 with debian too :p [06:26] aha [06:26] mr smart guy [06:27] i still dont know how he did it [06:27] my xoom no matter what kernel panics if i try to boot from normal boot partition [06:28] if i try to use the sdcard [06:28] i have to go through recovery to get access to it [06:29] im trying your changes with a modded config, compiling it now, gonna take a bit longer as a lot more things are enabled [06:29] yeah he had to go through recoveery too [06:29] kk [06:29] ah, see now in every tutorial i found recovery was never mentioned [06:30] it was like ooh magic it works [06:31] heh [06:34] building new initrd [06:34] hehk [06:36] and we're booting again [06:37] cool [06:37] with the patch? [06:37] aha no panic [06:37] need to stick with this config... [06:37] yep w/patch [06:37] ok cool [06:38] im gonna need to put this somewhere [06:38] like for people to use [06:38] heh [06:38] git it [06:39] now how is the touch going, or no x yet [06:39] sounds like a plan, i would like to host a kernel binary too for people like me who go nuts [06:39] yeah [06:39] ok i just tried to click continue [06:39] well thats what I do [06:39] it registered a down but not an up [06:39] http://forum.xda-developers.com/showthread.php?t=1191141 [06:40] ah youre a member too :) [06:42] nice work [06:45] ugh i lost adb [06:46] hm, don't know how you lost adb [06:46] for me it is something I just cannot kill [06:48] its hacked in... [06:48] comes up n says cannot bind tcp:5037 [06:49] hmm [06:49] and you dont have qtouch in your kernel tree [06:49] must be specific [06:49] no, I have a different panel [06:49] my devices have atmel mxt touch panels [06:50] and there isn't a "universal arm tree" so to speak [06:51] hm [06:51] https://github.com/LIV2/LIV2-Xoom-GNU/blob/tiamat_stock_gpu/drivers/input/touchscreen/qtouch_obp_ts.c [06:51] liv2's kernel still has those HAT0X/Y [06:52] wonder if he did any fixes to his [06:52] no I left them behind [06:53] main thing is getting rid of btn2 [06:53] but I see no point in keeping hat0x/y [06:54] yea neither do it, its made no difference [06:55] could this be as simple as a config in ubunbtu itself/ [06:55] ? [06:55] nah it really shouldn't make a difference at this point [06:55] mmk [06:56] everything left are valid touch coords [06:56] and evdev ignores those that it doesn't use [06:56] actually can you evtest again [06:57] no adb [06:57] bleh [06:57] yea, quite the pickle [06:58] can i put adbd into the initrd? [06:58] possibly, not tested it myself [06:58] and really, that shouldn't affect anything [06:58] must be a usb thing then [07:03] probably [07:11] alright, well im stumped and rather tired [07:11] so im going to take a break for tonight [07:16] k [08:15] Hi === xruud_ is now known as ruudt [09:44] hi [09:48] can anyone give me a good howto to crosscompile for arm7 technology? thanks in advance [09:55] guys, whenever i finish the oem-config for the 11.10 desktop image on the pandaboard i briefly see lightdm and then the screen flickers and switches off [09:55] anyone seen this [09:55] ? [09:55] i've tried a few times [09:56] not here [09:56] what kind of panda is that ? 4430 or 4460 ? [09:56] i know the 4460 still hat temperature issues with the 11.10 kernel [09:57] well, it's Rev A1 if that says anything [09:59] thats 4430 says google [09:59] well, that should definitely work [10:03] maybe it's my damn monitor [10:03] what would be the easiest way to develop programs for a minimal ubuntu installation? Given that I need internet and show images [10:03] got the chance to try another ? [10:03] ogra - i can try my TV :) [10:03] yeah, do that [10:04] brendand: what happens if you wait a few minutes? Just to rule out temperature issues? [10:04] might be an issue in the EDID parsing [10:04] ruudt, the 4430 shouldnt have any temp issues [10:04] for the 4460 i know it for sure that we have them though [10:04] ow ok, sorry, I was guessing [10:05] has any of you any programming experience for linux? [10:05] ruudt, we have the ubuntu-core images that you can use as a base for rolling specific minimal rootfses [10:05] I mean, I just installed quickly, but it is not really what I need [10:06] orgra_ I have an installation. But need to build a custom app [10:06] ah [10:06] well, quickly would be a good start if you actually want gtk and python apps [10:06] getting a working installation took several days of trying though :( [10:06] not sure where it stands with things like C [10:06] ouch [10:06] what board is that ? [10:06] something we support ? [10:07] beagleboard xm rev c [10:07] ah, yeah, revC is a constant headdache [10:07] damn, is it that bad? [10:07] will hopefully be better in 12.04 [10:08] won't help me now [10:08] well, we didnt have one in the team to test on it until very recently before release [10:08] and it came out after kernel freeze and feature freeze [10:08] Can I, with quickly then also develop programs that don't rely on X to display the images? [10:08] you might, i have never used quickly to be honest [10:09] What do you use? [10:09] shell, C, whatever else suits the usecase [10:09] ogra - yeah, TV works perfectly [10:09] bl**dy pos samsung monitor :/ [10:09] (for quick GUI apps i usually take python/gtk and later re-implem,ent in C/gtk) [10:10] brendand, yeah, blame the monitor ... [10:10] although - TV doesn't have DVI so it was a HDMI -> HDMI connection [10:10] whereas monitor doesn't have HDMI so it's HDMI -> DVI [10:10] so it might be that and not EDID [10:10] that shouldnt make it break though [10:11] well, i don't know. also the v-alignment is off on the monitor [10:11] there is about 4 inches on the LHS with just static in it [10:11] did you try the auto adjustment of the monitor ? [10:11] but not on the TV [10:11] ogra - yeah [10:11] also look if there is an overscan setting in the menu [10:12] and switch that off [10:12] though that shouldnt lead to no pic at all [10:12] only to a cut off pic [10:14] no overscan setting i can see [10:14] i'll just have to commandeer the telly for now [10:15] well, you could wait for robclark to get up, he might know the monitor models that still have EDID issues [10:15] (i thought they were solved, but apparently they arent) [10:15] ogra_ - i'm pretty sure i had EDID issues with this one before [10:15] how about using hdmi on the monitor [10:15] doesn't have it [10:17] brendand: When you were connecting to the DVI monitor, did you use the other HDMI port? (The one marked hdmi-dvi) [10:17] Or DVI-D, on some boards. Whatever. The one on the right. :P [10:19] infinity - ah. i'm using the one that worked for me before. the one on the left [10:20] brendand: The one on the left is HDMI, the one on the right is HDMI-DVI. In theory. I don't use the one on the right, cause everything here is straight HDMI. [10:21] infinity - well. i no longer have a problem [10:21] that works perfectly [10:21] hah [10:22] so to sum up, you must use the right one if you connect to a dvi monitor [10:22] pebkac [10:22] the one that is labeled DVI, yes :) [10:22] ruudt - since oneiric, yes [10:22] ogra_: Pfft, like anyone reads silscreening (except me, apparently). [10:23] heh [10:23] s/sil/silk/ [10:23] in fact i never even looked at the labels [10:23] just used the one that worked [10:23] When infinity asked which connector I was looking that up, because on beagleboard you'd best connect to dvi as well :P [10:24] yeah, my beagleboard just has a DVI-D connector [10:24] But I'm still learning about everything here. My next hurdle; which development environemnt to use. But despite the advices I have not yet found one [10:25] didnt you say you wanted to do web development ? [10:26] ogra_: nope, but I do need to access the internet [10:26] ah, i thought you wanted to serve images through the net [10:27] Yes, but that is not web development? To me that is anything that runs from a server and is accessed by clients [10:28] I already did the web development part; I have a server ready that will supply all data my application will need [10:29] Did that while waiting for the BB to ship [10:32] Hmm, the minimal install that I got running doesn't even have man... [10:33] apt-get install man-db manpages [10:33] and watch your system get slow on upgrades :P [10:33] And manpages-dev, if you're doing C programming. [10:34] well, at least I can read up on the things that ARE installed. [10:35] http://manpages.ubuntu.com/ also works, if you prefer not to grind dpkg runs to a halt with man-db triggers. :P [10:36] ++ [10:37] (But I'm with you, all my machines have man-db installed... Not being able to use man confuses me) [10:41] well, I'll just continue installing man-db and then figuring out what to develop the programs with. Just wish it where more clear what to use... [10:58] re [11:08] i guess it's well known that suspending a pandaboard doesn't really compute? [12:23] hi all [12:23] it's early in the morning and i haven't had my coffee yet, so please excuse what may be a very stupid question [12:24] the kernel package that comes on the i.mx53 install image is called linux-linaro-lt-mx5 [12:24] but that package doesn't appear to be in the archive [12:25] it's not like i strictly need a kernel since boot is handled slightly specially here anyway, but i'd like to have some modules (installed properly) at least [12:30] it is in universe [12:30] see? no coffee. [12:30] ogra_: thanks :) [12:30] yw :) [12:31] * desrt blames debootstrap for that [12:32] it is one of the "community supported" images [12:32] like the ac100 one [12:32] seems to be working quite nicely, fwiw [12:32] the userspace on the image leaves me a bit confused though [12:32] well, the installation is quite slow [12:32] it appears to have an installer that installs to itself, or something [12:32] and then leaves the system in a somewhat broken state after [12:33] that would be a bug [12:33] ogra_: it's the microsd that's slow. i just plugged in an intel sata drive and it's about 5-10x faster now [12:33] it is a preinstalled image, on first boot it expands the rootfs partition to the full size of the SD [12:33] ah [12:33] then it reboots into oem-config to set up user, hostname, kbd etc etcd [12:34] ya. i did all that and now i boot to an empty desktop [12:34] (which on the desktop images requires X, so make sure to have a monitor attached) [12:34] in any case, it's all a bit much [12:34] * desrt just wants something he can ssh to [12:34] did it properly finish actually ? [12:34] i read about this bug with the window disappearing... [12:34] to be honest, i don't know if it did or not [12:35] you should get an aptd/dpkg UI that shows you the removal of all unneeded packages [12:35] i was careful not to select autologin [12:35] well, i'm just debootstrapping onto this sata drive now [12:35] if you didnt have that (and if ubiquity is still installed) you know that oem-config failed [12:35] use ubuntu-core ;) [12:36] i couldn't find an imx image for ubuntu core/server [12:36] tar xvfz is faster than debootstrap ;) [12:36] seemed to only have omap [12:36] ohhh [12:36] for that [12:36] there are no board specific ubuntu-core images [12:36] meh. already done :) [12:36] its all userspace [12:36] completely unconfigured [12:36] (but largely the same you get from debootstrap) [12:37] any board-specific userspace config i should know about before i try to boot back into this thing? [12:38] no idea, ask janimo, he owns the mx5 images [12:38] janimo: sup? :) [12:38] i dont actually know what bootloader config is usually done there [12:39] the image seems to have a fat partition on it with the kernel, initramfs and some sort of boot description file [12:39] (text) [12:39] looks pretty straight-forward [12:40] probably it needs the microsd to load the kernel image at least, then i can use the sata as the rootfs [12:40] likely a u-boot script file, thats not text [12:40] (boot.scr) [12:40] it has a header [12:40] * desrt takes a peek [12:40] you shoudl eb able to get away with a vfat for booting [12:40] * desrt notes that the SD card is hilariously slow [12:41] and then switch to rootfs on sata by root=UUID=.... on cmdline [12:41] so there is a boot.scr here [12:42] but boot.txt is the one i was looking at earlier [12:42] right, that should have a 64byte hearder [12:42] and is generated using mkimage [12:42] likely from boot.txt [12:42] it appears to be plain text... [12:42] .scr ? [12:42] .txt [12:42] yeah [12:42] oh. i see what you're saying [12:42] thats the input file [12:42] .txt gets compiled into .scr [12:42] right [12:43] mkimage adds a header [12:43] and somehow the .scr is 'blessed'? [12:43] yep [12:43] u-boot looks for it on boot [12:43] cool [12:43] lemme guess -- this 4MB /dev/mmcblk0p1 partition is uboot [12:43] not sure, thats where janimo comes into play [12:44] i dont have such a board and never had one in my hands [12:44] it's gotta be -- the kernel doesn't seem to fancy mounting it [12:44] so i dont know exactly how it boots [12:44] i think you've helped me put together a pretty clear image for myself [12:44] yeah, we tag the boot partitions with a label that udev ignores [12:44] so you dont have it mounted on your desktop install ;) [12:45] it has dos partition type 0xfa ("Non-FS Data") [12:45] (the kernel should see it in dmesg though, it should just nott be automounted) [12:45] 0xda, rather [12:45] ah, yeah, that coudl eb raw u-boot [12:45] *be [12:47] * desrt wonders why this uboot wasn't given ext4 support [12:47] ask upstream :) [12:47] oh. srsly? [12:47] it has ext2 [12:47] that's a pretty good reason, then :) [12:48] ext4 would rock though, you wouldnt need a separate boot partition [12:48] desrt, the mx53 ubuntu image is very slow indeed. No idea why yet, something to fix for 12.04 [12:48] but i think it bloats the codee to much or something [12:48] desrt, the linaro image for same board was much snappier iirc [12:48] and not only because if does not do an install [12:48] lies ! [12:48] :P [12:49] janimo: it seems to be mostly the IO that is slow -- and only to the SD card [12:49] we use the same kernel as linaro so I am puzzled why this happens, but all of us who tested mx53 late in the oneiric cycle [12:49] saw it was very slow [12:50] ogra_, shouldn't an ext2 driver see an ext4 partition especially in read only mode as uboot is? [12:50] just ignoring the extra features [12:50] no, i think you need ext3 support to see ext4 [12:51] ok [12:51] not sure though [12:51] maybe it is only vice versa then [12:54] you can definitely see ext3 from an ext2 driver [12:54] there the only difference should be the journal [12:55] ndec, so you are sure about the model ? (you didnt say anything in the mail) [12:56] ogra_: sorry. i forgot that ... do you have a serial number or any other number on the board? [12:56] to be honest i haven't even seen such a board myself ;-) [12:56] on the board or on the shell ? [12:56] whatever... [12:56] anything... [12:56] there is a barcode glued underneath the sd slot [12:57] T0000001266 [12:57] seems to be the serial # ... its on the box too [12:58] * desrt discovers flash-kernel.conf [12:58] ndec, then i also have TEVM4460G-22-20-01 [12:58] sounds a useful discovery... [12:58] not really [12:58] ogra_: that one looks useful! [12:58] once we switch to flash-kernel-ng there wont be a config file anymore [12:58] my first comment was for desrt ;-) [12:58] yeah [12:59] argh... [12:59] argh ? because of new flash-kernel ? [12:59] the failure situation for that file missing is a bit sad [12:59] flash-kernel is all in all a bit sad :P [12:59] yes, because we are doing tricky things with flash-kernel.conf ;-) [13:00] it makes you cry like onions if you look to long at the code [13:00] in fact i would recommend rubbing onions in your eyes instead of looking at the code :) ... if you have the choice [13:00] hurts less [13:00] :) [13:01] looking at the code led me to discover UBOOT_PART [13:01] since it does something like this: [13:01] yeah, thats an ugly wart [13:01] mount ${UBOOT_PART} ${tmpfile} [13:01] and doesn't consider that ${UBOOT_PART} may be unset :p [13:01] the new flash-kernel has a proper DB wiht an entry per board arch [13:01] neat [13:01] that will be way cleaner [13:01] still hackish but lots better [13:02] ogra_: i am amazed, but we do have an internal DB where I could find your reference that you gave me ... TEVM4460G-22-20-01  -> OMAP ES1.1 GP,    1.5GHz, PWR ES2.0, AUD ES1.1  [13:02] ndec, to be honest the device doesnt look much different from the one i already have ... [13:02] chrome metal frame etc [13:03] NIC built in the wrong way round again :) (which i only discovered after i had the cable attached indeed) [13:03] etc etc [13:03] gonna try a reboot now [13:03] good luck :) [13:05] WIN [13:05] works on first try [13:05] failed to set root password or add self to sudoers file [13:06] * desrt hands head in shame [13:06] ogra_: haha I had a look at flash kernel, proceeded to rip my eyes out, then think about doing my own platform specific hack till flash-kernel-ng [13:07] janimo: know offhand what serial parameters you configured uboot for? [13:30] re [14:04] ndec, see pm :) [14:05] An-iSociaL, if you need help with ac100 md5sum mangling, i'm around now [14:09] hmm can anybody help me with my ubuntu-arm and crosscompiling? ;) [14:09] better try #linaro [14:09] ok thx [14:20] do all armv7l platforms follow the same layout for /proc/cpuinfo? [14:21] the kernel does === ericm|ubuntu is now known as ericm-Zzz [14:23] GrueMaster: http://people.canonical.com/~ppisati/linux-image-2.6.35-903-omap4_2.6.35-903.27~mlockfix_armel.deb [14:23] GrueMaster: this should fix the [14:23] GrueMaster: "Make sure the stack guard page does not split the stack on mlock ... FAIL" [14:24] GrueMaster: you saw in test-kernel.py [14:24] GrueMaster: can you verify it? [14:56] GrueMaster: and while here [14:56] GrueMaster: https://launchpadlibrarian.net/85419190/test-kernel-security.log [14:56] GrueMaster: on one line it says: "PR_SET_SECCOMP works ... FAIL" but below that [14:57] GrueMaster: "CONFIG_SECCOMP enabled ... (skipped: not available on ARM) ok" [15:59] afternoon all [16:00] ogra_, if you mean the tarball installer, thanks but i figured it out [16:00] well, you hacked it to be broken afterwards i guess [16:00] :) [16:01] hm? [16:01] the proper way is to roll a cpio.gz archive with a file carrying the tarball md5 of your modified tarball [16:01] and then cat'ing that to the end of the initrd [16:01] ahhh [16:01] if you hack the script you might get unexpected results [16:02] and unlike GrueMaster claims we dont do "werid things" with the md5 at build time :) [16:02] i actually performed the role of the installer script myself, did the formatting, extracting, getting uuid, purging intaller [16:02] we just put the file in place before rolling the initrd [16:03] and the fstab [16:03] well, that will miss a bunch of configuration [16:03] unless you also made oem-config work [16:03] (which does the actual configuration in a proper way) [16:03] im pretty sure i did, oem-config did run after touching it [16:03] ah, cool [16:03] then you shoudl be fine [16:04] the code we use to add the md5 is in the livecd-rootfs source btw [16:04] for future reference [16:04] will have to go back to the installer if i intend to distribute [16:04] nice [16:07] noted it [16:23] heh [16:23] the way I do it is with the omap image [16:24] sup lilstevie [16:24] not much just working on my app [16:24] oh cool. whatcha building? [16:25] an app for the tf101 ubuntu/android system [16:25] android app for managing it [16:25] nice [16:26] im just watchin tv trying to get in the mood to do something [16:33] If I just want to run an arm precise image in qemu-arm to verify a bugfix, what's the quickest way to create an image? Do I download one of the preinstalled images from cdimages.u.c? Is there a disk image i can wget? [16:35] uhm [16:35] the only image i currently know of is for the ac100 [16:35] likely wouldnt do what you want [16:36] fastest way otherwise would be debootstrap [16:37] and then d/l a kernel from somewhere? [16:38] hallyn, use ubuntu-core and look in the linaro docs about vexpress with qemu [16:39] or just use an arm chroot, if you dont need a full VM [16:40] ok, thx. (i don't think a chroot'll work as it's ecryptfs-related, but will try) [16:41] ah, yeah [16:41] wiki.ubuntu.com/Core [16:41] could be difficult working with lower level stuff like filesystems in a chroot [16:41] note that there isnt a user/pw or any configuration [16:41] you will need to do that yourself [16:42] qq ogra_, what cross compiler do you use? [16:42] ok, thanks guys [16:42] An-iSociaL, i compile native [16:42] ahh [16:43] obviously that would be better but how much better? [16:43] depends what you compile really :) [16:43] specifically for kernel builds [16:44] i usually build packages ... for that native is always better [16:44] though there are already a bunch you can actually cross build nowadays [16:44] for kernel test builds i would actually recommend cross [16:44] and use the gcc-cross in the archive [16:45] aha [16:45] sweet [16:45] i think using the android cross compiler's been jacking my stuff up [16:46] kernel modules for sure [16:46] makes me wonder if its jacked up kernel modules wth has it done to the kernel itself [16:47] there should be a package gcc-armel-cross or some such [16:48] (ask hrw, he maintains it, i can never remember the actual name) [16:48] i saw it when i was extracting [16:51] gcc-arm-linux-gnueabi [16:53] bah, not even remotely a -cross in the name :P [16:53] hehe [16:55] cool cool thanks [16:55] grabbin it now [16:55] perhaps this will solve a couple little bugs [16:56] getting closer to being in the mood to do something... yay... [17:00] cu [18:33] anyone have any experience getting python 2.7 to compile on a arm platform? (gumstix) [21:35] hey guys [21:35] any reason why unity run very slow on a pandaboard [21:35] might it be because im not using a swap? === kc4zvw_ is now known as kc4zvw