=== prp^2 is now known as prpplague [00:28] just fyi for anyone interested, the call for participation for ELC-E end august 1 - https://events.linuxfoundation.org/events/embedded-linux-conference-europe/cfp [01:52] janimo: well, personally I do not trust his stuff as far as I can throw him, he has been a massive pain with things like not being even able to identify the root device, and he doesn't really know what he is doing === jkridner__ is now known as jkridner [06:41] lilstevie, I don't know whom you are referning to anymore, I saw about 3 3.1 based kernel trees on xda :( === rsalveti` is now known as rsalveti [07:14] lilstevie, do you know anything about the status of the 3.1 jhinta kernel for tf101? It sort of boots here, but hangs on starting network manager [07:14] ^^ that one [07:20] lilstevie, ah ok. INdeed that's the last one I tried. It boots but touchpad at least does not work [07:22] janimo: I have had serious problems with him in the past [07:24] lilstevie, it baffles me that in such a large community there's no 'blessed' git tree where everyone tries to build from an improve on. Instead there are from scratch ports and people posting binaries and zips around on megaupload for others to test [07:25] janimo: welcome to the world of donation hustling === zyga is now known as zyga-afk === zyga-afk is now known as zyga [10:26] I'm running qemu-debootstrap to fetch natty armel. However, often than not, it fails with could not download package, or failed to get release file. The server it's using is a local mirror on our lan (and its working fine). Have anyone else had any experience with this issue? [12:26] sveinse: which error? [13:17] so running into a issue when trying to unpack the 12.04 image for Beagleboard XM. Im running the comand md5sum ubuntu-12.04-r3-minimal-armhf.tar.xz and it gives me errors. [13:19] tar: xz: Cannot exec: No such file or directory I checked and the file is in home. Did the md5sum check and it picked the file right up. [13:21] unpacking ? [13:21] where did you get that image from ? thats definitely not an official one [13:21] (there is nothing to unpack with the official ubuntu images) [13:22] got it from here http://elinux.org/BeagleBoardUbuntu#Demo_Image [13:23] right, not an ubuntu image, ask the creator [13:23] (which is likely robert c. nelson (rcn-ee), judging by the url) [13:24] is there a better image for the beagle board xm you may know of? [13:24] https://wiki.ubuntu.com/ARM/OMAP that links to the official images [13:25] sweet got it. [13:44] xranby_ac100: yea i looked at linaro stuff [13:50] WOO sd card is written and set up. Thanks for the help! [13:51] note that the desktop install requires a monitor for the initial setup [13:51] if you want to use serial you need to use the server image === prpplague is now known as prp^2 [13:53] Im useing the desktop one so ill grab a monitor with dvi. [13:53] * ogra_ sighs, still no sign of life with the ac100 images [13:53] thats not good [13:54] janimo, did you ever look into why we cant use more than 2M for the ac100 initrds ? [13:54] seems the quantal one is 2.6M ... and quantal images hang at the toshiba splash === zyga is now known as zyga-food === Ursinha` is now known as Ursinha [14:57] ogra_, never looked at that, only knew - from you - that there is a limit [14:58] ogra_, is this a know limitation of the bootloader, known by the whole ac100 community? === zyga-food is now known as zyga [15:33] janimo, well, i dont know if its known by anyone, i just see it exploding if i pass a certain size [15:34] ogra_, O [15:34] I mean I'll have a look too :) [15:36] ogra_, what did quantal add that increased the initramfs size? I recall you saying that's the reason we should not add Ubuntu SAUCE as cryptsetup would increase the size too much [15:36] well, i tink binaries just got bigger [15:36] though ... [15:37] it seems dropping the cmdline makes it boot ?!? [15:39] (which is indeed bad since i cant get a working tty without console=tty1) [15:52] ogra_, so droppping the cmdline entirely lets it boot but any cmdline options causes hangs? that is indeed weird [15:52] ogra_, is console=tty1 needed as a workaround for plymouth + 3.1 kernel issues? [15:53] no, its needed for the broken kernel [15:53] afaik marvin24 still waits for a fix from nvidia for that one [15:53] ogra_, but the same thing that chmod -x plymouthd works around? [15:53] you dont get any tty output without it [15:53] so i got it booting but have a black screen [15:57] ogra_, even X does not show? [15:58] no X in the installer :) [16:07] * ogra_ tries with only console=tty1 added === Neko is now known as TheFinder [16:16] janimo, hmm, ok, seems i got it up to a point where i get output and dont hang in the bootloader, but now it panics [16:16] ogra_, ah the tarball installer. But good, progress :) [16:16] "stack-protector: Kernel stack is corrupted in: c06a594c" [16:16] no idea what to do with that ... especially since i cant scroll back now [16:17] your own zImage or quantal deb? [16:17] quantal deb [16:17] a cusotm kernel built is likely needed to debug it [16:17] I can try debugging it [16:17] sigh, i feared that [16:18] well, i have a shiny new fast desktop workstation here ... just finished building it ... i guess i should start trying a cross build one day ;) [16:18] ogra_, it's an excellent time to start cross compiling :) [16:19] haha [16:19] if it is shiny new and fast you may get 2-3 minute kenrel builds [16:19] not really, once we have the calxeda HW cross building will so be last century [16:19] my oldish core duo takes 20 min for a deb, and mucxh less for a barebones zImage [16:19] you will just build with -j2048 and get 30sec kernel builds :) [16:20] you mean in 2015 :) ? [16:20] or do you know somethig I don't [16:20] well, once we have the calxeda HW accessible for devs in the DC [16:20] by that time 32 core Core i9 will kick any ARM server's butt anyway :) [16:20] lol [16:21] well, we worked with Martyn for two years to make sure we dont need to do any bringup work, so we had a guarantee that we can immediately put the HW into the london DC ... [16:21] I used quad armadaxp for some builds and it is fast, but not quite as my laptop. is calxeda supposed to be much faster? [16:22] * ogra_ still doesnt get why that stuff went to lex ... but hey ... no more arm team, ... [16:22] calxeda is supüposed to be much *more* ... not much *faster* :) [16:23] you will simply drown in spare cores ;) [16:23] leave it to libreoffice builds to shame any hardware [16:26] pfft [16:26] if you build a kernel you can have a job for every .c file ... that should be really speedy ... [16:27] for LibO its indeed bound to the slowest step of the build [16:27] memory will be a bottleneck for large c++ builds [16:27] I am not sure how much RAM calxeda hw has [16:27] they have DIMM sockets afaik [16:28] it may not make sense to have more than -j8 if that fills up 2-3Gigs [16:28] like the armadaxp you should be able to just pulg in ram ... though i think it can only address up to two G [16:28] well 32bit still, so limited indeed [16:28] right [16:29] the big advantage of claxeda is that you simply can throw cores at a task if you need them ... [16:29] s/cores/machines/ [16:29] would be good if the kernel could just merge the ram of these into one big chunk ;) [16:41] 4G RAM on calxeda [16:45] Does the 12/04 arm image come with SGX video acceleration on it? [16:49] janimo, ogra_: if you give me something to test, I can try it here [16:50] well, i doubt you would see much more than i on your screen :) [16:50] btw, we could try with a 2 MB zero file in the initrd to check if it is a loader or and expander problem [16:50] ogra_: but I have a serial interface [16:50] Just-in, try the linaro panda images, those have sgx by default [16:50] make sure you pick the right flavour, only some have x11+sgx enabled [16:51] will that work on a Beagle board xm? [16:51] 12.04 has full SGX support [16:51] oh, not on omap3 [16:51] and linaro wont have that either [16:51] thats what i was afraid of. [16:51] TI doesnt release the omap3 SGX drivers built for hardfloat [16:52] should i back down to 11.10 then [16:52] robher_, oh, how do you address the last GiG then ? PAE ? [16:52] not sure how smooth it will run with out SGX [16:53] what do you want to run ? [16:53] so far all it will be is a pc for people to use during breaks till i get a tower in. [16:53] the framebuffer driver is fine for all 2D stuff, its pretty usable even without SGX for normal desktop bits [16:53] im sure they will want to watch youtube and all that. [16:53] LOL [16:53] you surely picked the wrong arch [16:54] im starting to agree [16:54] (no flash for arm linux ...) [16:54] then they have no youtube [16:54] they can live [16:54] right [16:54] ogra_: 1G kernel, 3G user [16:54] Just-in, listen to ogra, I am not uptodate with panda/sgx [16:54] janimo, not panda, beagle :) [16:55] i wonder if Zsnes will run ok. [16:55] i dont think that needs 3D support in any way [16:55] shouldnt [16:55] * ogra_ has to run out and pick up his GF from a 50km away dentist appointment ... [16:56] ouch [16:56] i wonder if other peoples GFs pick their dentist in the town they live in :P [16:56] mine didnt [16:56] heh, so i'm not alone [16:56] i think it so we have to go and get them [16:56] and listen to the drugged up convos on the drive home. [16:57] haha [16:58] aww man i just remebered i have some old usb game pads for the old cable set top boxes i could use for zsnes... [16:58] today is a great day [17:13] ogra_: I've put a 10MB file (full of zeros) into the initrd and it still boots [17:20] a 3MB file full of "randoms" gives an oops on boot ;-) [17:27] marvin24, does a 10Mb initrd fit in the boot partition? [17:27] sure, if it's full of zeros it will be compress very well :-) [17:27] oops -> http://paste.ubuntu.com/1088351/ [17:28] marvin24, ah uncompressed 10M [17:28] what hardwar eyou running that on? [17:29] Just-in: ac100 [17:29] the oops looks like it cannot find the root fs [17:29] well isnt that a neat looking netbook [17:30] dear god that came with 2.1 on it [17:30] sure my dear [17:30] yes, that was an epic fail [17:30] but a nice hw ;-) [17:30] man [17:31] would have liked to see 1gb ram matched with the tegra [17:32] given that most other have 1GB ... [17:33] chromeos would have been better on it [17:33] but linux desktop also works nice with zram enabled [17:33] indeed, a new toshiba with tegra3 and 2Gb of RAM would be an awesome netbook. Too bad there's no market for it [17:33] yeah [17:33] i like my little netbook [17:34] best 700 i ever spent [17:34] * marvin24 only spend 120 ... [17:34] for his 2nd one [17:34] * Just-in has a I7 and nvidia 335m [17:34] nothign better then gaming on a plane ride [17:35] and you call this a netbook? [17:35] time has passes sooo fast ... [17:36] it has a 11 inch screen and no cd drive so.. [17:36] a uber netbook? [17:36] or a heater [17:36] lol [17:37] Shockley doesnt get to hot [17:39] ah, "Initramfs unpacking failed: uncompression error" [17:40] so likely a kernel thing === zyga is now known as zyga-afk [18:05] ogra_, git clone -b packaging-3.1 git://kernel.ubuntu.com/jani/ubuntu-ac100.git [18:05] CROSS_COMPILE=arm-linux-gnueabihf- ARCH=arm make paz00_defconfig [18:06] CROSS_COMPILE=arm-linux-gnueabihf- ARCH=arm make -j 16 zImage [18:06] or whatever, depending how many cores you have :) [18:06] you probably get a zImage in 1 minute [18:11] ogra_: are today's omap4 images working for you? [18:17] stgraber, i havent tested omap4 since the alpha release [18:17] but they should [18:21] ogra_: it's failing with "oem-config/enable doesn't exist" here... I haven't been able to install for a reason or another for over a week now :) (testing with edubuntu though, but shouldn't make a difference) [18:22] starting an install with -d now, hopefully I can figure out what's going on [18:56] ogra_: I think the compressed initrd is either not loaded completely, or the end is overwritten after load [18:57] I tried to increase initrd_phys, but no change [19:04] marvin24: With which bootloader? [19:04] marvin24: If this is under uBoot, try setting "setenv initrd_high 0xffffffff" [19:05] marvin24: If fastboot, no idea. [19:06] fastboot ... [19:06] I ask a u-boot user to try on #ac100 [19:06] If it's just too big, it could be getting truncated in the boot partition? [19:18] no, nvflash would fail [19:19] infinity: but if the kernel uncompresseds the initrd over the compressed initrd, we have a problem [19:19] AFAIK, the mem layout is kernel, compressed initrd, uncompressed initrd [19:19] there is also something tiny before the kernel [19:24] stgraber, erm ... are you sure you are testing a live image ? we dont roll preinstalleds anymore apart from ac100 [19:24] (i dont think live does anything with oem-config by default) [19:27] ogra_: it's definitely a live image ;) [19:27] weird [19:27] ogra_: it's crashing right after hw-detect so I'm testing the new ubiquity as Colin added some debugging in that area of the code [19:28] i wonder where that comes from then ... we used to use oem-config indeed on the preinstalleds [19:28] based on some other bug reports on LP, the error might be completely unrelated to the problem ;) [19:29] heh, hopefully [19:32] well, maybe for you, but I'm not looking forward to debug something where the only error message has nothing to do with the problem :) [19:33] though it's pretty rare to get an "obvious" ubiquity bug these days, they're all weird race conditions triggering a crash 5 minutes after the actual problem happened [19:41] hi [19:41] anybody alive in here? [19:42] stgraber, i'll test omap4 tomorrow, in case you didnt get to the issue yet, i'll take a look as well [19:43] omap4 was simply out of my focus for a few days ... fighting with debian-cd and the ac100 issues [19:47] ogra_, ac100 is a tegra2 thing, right? [19:47] im trying to get debian on eee pad transformer [19:47] is working pretty well as far as you use the console ;) [19:49] Inoperable, lilstevie does the same with ubuntu :) [19:54] ogra_, yeah i know - just got his OLife opened on the console ;) [19:54] and flashing it like 234 time [19:55] i guess i'm gonna fry that flash ram sooner or later [19:56] lilstevie, what is this tegra 2 flash thing? i wanted to beta test ;) [20:07] btw, will ubuntu pack http://nv-tegra.nvidia.com/gitweb/?p=tools/tegrarcm.git;a=summary ? [20:07] it is an open-source nvflash replacement [20:08] Oo. [20:08] Works on all the same devices? [20:08] And does the same thing? [20:08] works on all tegra devices [20:09] only for flashing via usb [20:09] marvin24, i'm on it but it moved down on my TODO [20:09] srwarren pointed me to is two weeks ago [20:09] s/is/it/ [20:09] ogra_: cool, thanks! [20:11] ogra_: Do you maintain anything in Debian? [20:11] infinity, its the free nvflash that exists way longer than nvflash already (internally at nvidia) [20:11] ogra_: I'd either (a) sponsor that for you, or (b) do the Debian maintenance. [20:11] infinity, well, i was wondering about DD :) [20:12] Right, well, you maintaining and me sponsoring it for you is a good step on the road to DDship. [20:12] but essentially i'm doing that since i started at canonical ... [20:13] though being flash-kernel uploader in debian would massively help, stuff starts piling up again already ... [20:58] marvin24, any docs for tegra-rcm? [20:58] i just compiled it and it runs - but i dont know how to use it hehe [21:00] the readme does not says a lot ;) [21:08] well nevermind [21:08] does anybody got a working 3.1 kernel on tegra2?