[03:36] Hey all. [03:36] Anyone have experience writing gigabit ethernet drivers? === tmzt_dg2root is now known as tmzt_g2root [03:57] rbelem and Tscheesy: We need GCC fixed before anything happens on Qt/ARM for Natty. === hrw|gone is now known as hrw [08:03] morning [09:44] can anyone comment how to build kernel Headers for kernel source ? As I am using diff kernel tree than ubuntu and wanted to build pvr kernel modules [09:46] make headers_install ARCH=arm INSTALL_HDR_PATH=/custom [10:00] https://wiki.ubuntu.com/KernelGitGuide i think [10:02] ah, no, its actually https://help.ubuntu.com/community/Kernel/Compile [10:03] https://wiki.ubuntu.com/KernelTeam/GitKernelBuild might also help [10:19] how to enable loopback device [10:20] i cant see ifconfig -a lo interface [10:25] ifconfig lo 127.0.0.1 up [10:25] ? [10:25] or "ifup lo" if you have ifupdown in your filesystem === ogra_ac_ is now known as ogra_ac === dmart is now known as Guest16033 === Guest16033 is now known as dmart__ [13:35] grmpf [13:35] merging shadow provides to become really painful === dmart__ is now known as dmart [14:46] hello, i have some problems with the ubuntu 10.10 installation on the beagleboard xm with sd card. i tried the two methods on webpage to flash the sd card - https://wiki.ubuntu.com/ARM/OMAPMaverickInstall but i cant see the card on my linux pc after i put the files on the card. then i boot windows and format the card. any help what i did wrong ? [14:57] GrueMaster: for https://wiki.ubuntu.com/ARM/OMAPMaverickInstall, should there be a "sync" after the zcat >device / dd of=device step? [14:58] rOxx: are you sure you used the correct device node? what did you do to write the image to the card exactly? [14:58] after the zcat and dd step i got no message [14:59] i tried this two methods but nothing works [14:59] rOxx: Try again, but after the zcat or dd step, type: [14:59] sync [14:59] ...and wait for the prompt to return [14:59] Possibly not all of the data has hit the card? [14:59] @dmart, yes im sure that is used the correct device. the name was mmcblk0p1 - a sd card with 8 gb [15:00] thats definitely the wrong device name [15:00] oh, yes, it must be the whole-device node: /dev/mmcblk0 [15:00] right, the above is a partition name [15:01] sorry is used /dev/mmcblk0p1 [15:02] im back in few minutes. i want to test "sync" in console after flashing [15:02] and im now in windows :;-( === kmargar is now known as markos_ [15:26] dmart: sync could help. I have a script that I use to zero then flash my SD card. The last step is for it to pop up a notification window letting me know it is done. [15:27] writing to the device instead of the partition too [15:27] And Idon't use the zcat method. I have found it to be less reliable than "gunzip ; dd" personally. [15:28] huh ? [15:28] there shouldnt be any difference [15:28] there surely isnt techincally [15:31] Not sure why. May be because with dd youcan write a block at a time, whereas the zcat method writes in byte mode. [15:37] I have seen issues when using dd if I don't specify "bs=1024" or at least have"bs=" set to a sector or cluster divisable number. [15:40] i´m back, but the sync command doenst help me [15:42] here are the details [15:42] bastian@bastian-Inspiron-1520:~/Downloads$ sudo dd bs=4M if=ubuntu-netbook-10.10-preinstalled-netbook-armel+omap.img of=/dev/mmcblk0p1 [15:43] 499+1 Datensätze ein [15:43] 499+1 Datensätze aus [15:43] 2094403584 Bytes (2,1 GB) kopiert, 228,692 s, 9,2 MB/s [15:43] after that is used the sync command. after 1 second the prompt come back [15:44] but after that i cant see the sd card [15:46] gparted does not start after this process, when i disconnect the sd card gparted starts, the only way i find out to get access to the card was a windows system.... [15:48] you are still writing to the wrong device with dd [15:49] /dev/mmcblk0 is what you want [15:49] *not* /dev/mmcblk0p1 [15:49] thats a partition [15:49] rOxx: Linux will not know about the new partition layout that was written to the card right away ... what happens after you remove and reinsert the card? [15:49] dmart, he writes a partitioned image into a partition [15:49] that wont work ;) [15:50] that's also true [15:50] dd ... of=/dev/mmcblk0 [15:50] thx ogra, i will test without p1 [15:50] (not ... /dev/mmcblk0p1) [15:50] im back in few minutes. thanks for your help [15:56] GrueMaster, did you put the suggestion for gunzip/dd on the install instructions ? [15:57] yes, a while ago. Others were reporting on this channel that they were having issues, butthat the other method worked fine. [15:57] can you please add a warning that his eats hilarious amounts of diskspace and should really only be used if zcat doesnt work [15:58] with emphasis on *only* [16:02] GrueMaster: I edited the wiki to add syncs; feel free to check it's OK [16:03] Also, xM A3 / OMAP4 Blaze instructions talk about mounting partitions -- the user will need to remove/reinsert the card before this will work (or blockdev --rereadpt ) ... otherwise the kernel will use the old partition layout [16:04] ...I think [16:04] shouldnt [16:04] ogra_ac: do you know why zcat doesn't work? What are the symptoms? [16:05] dmart, i have never seen zcat fail [16:05] on various systems [16:05] dmart, GrueMaster claims its faulty [16:05] Could it have been a forgot-to-sync problem maybe? [16:06] Unless the kernel or SDIO layer is actually broken, there should _really_ be no difference between zcat and dd :O [16:06] right [16:06] thats why i think as well [16:06] I have been able to make zcat fail reliably on some sd cards. Remember, I do this more often on a wider spread of card types & speeds. [16:06] weird [16:07] I can even get the dd method to fail if I don't specify "bs=" and have it divisable by 512. [16:08] sounds like a kernel bug on your side [16:08] Besides, my method also allows you to mount the image w/o writing it to an SD card. [16:08] sure [16:09] but it is one step more, complexifies the instructions and eats a lot of diskspace [16:09] ogra_ac: I have seen this on Lucid & maverick now, both x86 & amd64 kernels. I have also seen this on babbage. [16:09] (and its not a slow setp to gunzip 2G) [16:10] * ogra_ac humms .... why oh why did i ever touch shadow [16:11] that package went from looking like a sync to be an easy merge ... until all patches stopped applying [16:11] If you feel there is a serious problem with my additional instructions (that people should use if zcat fails, according to the wiki as I added it), then by all means remove it. [16:12] no, leave it in [16:12] i just want them to be aware that it takes 10min extra and eats their disk [16:12] 10 minutes? Try 2 on my system with a Class 2 SD card. [16:12] and should really make no difference [16:13] i dont get why it does [16:13] I agree that it *should not* make a difference. What I have seen in my daily testing tells me otherwise. [16:13] i was talking about the extra gunzip above [16:13] nothing to do with SD cards [16:14] I'm talking about the whole procedure (including the gunzip). [16:14] the dd as well as zcat should take the same time [16:17] Read http://irclogs.ubuntu.com/2010/11/05/%23ubuntu-arm.html about 3/4 down. [16:19] you mean the guy who had a hardcoded line in NAND to only load his kernel ? [16:20] If it were just me having issues with zcat, I would do my own thing and move on. [16:20] * ogra_ac doesnt see how that relates to zcat [16:20] apart from you telling him that it does [16:20] zcat didn't work for me, dd did perfectly [16:20] without any further comment, yep [16:22] Do I really needto pour through all the IRC logs to back myself up? I thought this was Linux, where there are multiple solutions to a single problem. [16:22] well [16:22] Again, if you have a problem with my alternate method being on the wiki, remove it. [16:22] I'm done with this discussion, because it really isn't productive. [16:23] i just asked you to add a warning ... [16:23] why do you feel attacked [16:24] it should simply be clear that this is just a fallback method and that it uses up a lot more resources and time [16:24] I can't remember what happened when I used zcat but it didn't work as nice as dd, which have never failed me :) [16:24] * ogra_ac would love to know how it failed [16:25] given that there is no technical difference between the two [16:26] i can imagine that something goes wrong due to sudo quoting or some such [16:26] I could probably try it out again but not time atm for it [16:26] but its weird that two techinally equal methods can have different results [16:26] IIRC the problem was that it didn't boot properly [16:26] yeah, i didnt mean to push you to it ;) [16:27] weird [16:27] pretty sure I did the command both with sudo and root [16:27] but I could have done that wrong :) [16:28] well, try it once if you need to anyway and note down whats wrong, i'm surely intrested in hearing about it [16:28] btw, does your cloak indicate what i think ? :) [16:29] (xbmc i mean) [16:30] hehe, it does :) [16:30] (if you think I'm a developer of the app?) [16:30] have you seen https://blueprints.launchpad.net/ubuntu/+spec/multimedia-arm-n-set-top-box ? [16:31] we're planning to provide a settop box install (no image but you should be easily able to build one with it) [16:31] i was wondering why xbmc didnt make it into ubuntu yet [16:31] Haven't seen, cool! [16:32] Well I know one of our devs where with the media center thing for ubuntu [16:32] seems there is a PPA but nobody ever uploaded it to the archive [16:32] And reasons then was mostly license problems (and that it doesn't run without proper rendering) [16:32] ah [16:32] I think the biggest issue is that we have our own ffmpeg which is abit iffy [16:32] well, rendering should be fine with the pandaboard [16:32] we have drivers and codecs [16:32] (gstreamer only though) [16:33] ah [16:33] yeah, ffmpeg is hard if its out of sync [16:33] we are going to move it out of xbmc at some point, probably make it an addon. That way the only codec we provide out of the box is hw accel stuff which should be fine from a license point of view [16:34] It is possible to build against external though, think angstrom does this [16:34] debian do atleast, but its causing some problems at points [16:34] ah, cool [16:34] probably something to investigate for ricardo then [16:34] A big goal for us with eden (next release) is to be perfectly ready for repo inclusion [16:35] cool ! [16:35] since camelot (and abit before) we have worked lots with fedora and such and should have most fixed [16:35] before it was a mess with lots of internal libraries (xbox herritage :) ) [16:35] ocra_ac and dmart thx for your help. now it works. mmcblk0 was the answer :-) [16:35] heh, yeah [16:35] rOxx, good to hear :) [16:36] are you working with that blueprint ogra_ac? [16:36] the config with der seriell ouput works and now im bootet ubuntu 10.10 on beagle. all works fine [16:37] topfs2, nope, rsalveti-afk is [16:37] (on vacation this week) [16:37] topfs2, i'll only test BT remotes (for the fun of it) [16:38] ah cool. Well I have a panda and trying to get xbmc to run awesome on it and considering we have openmax support already hopefully it will work out on panda also [16:38] well, its all gstreamer currently [16:38] does xbmc handle that ? [16:39] (in ubuntu that is, i dont think anyone at TI has put work into OMX beyond gstreamer) [16:40] nah we don't have gstreamer support but thats also something we have been investigating, atleast use the decode of it [16:40] that would make it really easy [16:40] but I really hope omx works properly outside it, otherwise I'm gonna ping them like crazy :) [16:40] well, i know that they only focused on gst integration [16:41] so you might have to work out solutions on your own [16:41] personally I like gst but the timing and syncing of it when displaying isn't as good as it could but for decode its great [16:41] i dont think i ever recorded with it [16:42] I meant playing thogh :) [16:43] hmm, i usually dont have issues with totem, but then thats a very simple app [16:43] timing/syncing wise i mean [16:44] I actually have quite a lot (on desktop) [16:44] perhaps its just my computer [16:44] on panda its fine for me [16:58] speaking of BT remotes... does Playstation3 remote counts? [16:59] hrw, sure [16:59] i dont really expect issues with BT remotes [16:59] it was just a good reason to buy one [16:59] or more [17:00] ogra_ac: ps3 remote is not BT HID but rather BT serial something [17:00] i think there is an app for it in ubuntu [17:01] yeah we have a special thing for it for xbmc [17:01] ;D [17:01] anyremote [17:02] its to bad there isn't a great remote out there :) [17:02] Remote control daemon for applications using Bluetooth, IrDA or WiFi [17:02] that might do it [17:03] * hrw -> out [17:03] have a nice rest of day === hrw is now known as hrw|gone [17:03] what bluetooth remote are you trying ogra_ac ? [17:03] none yet [17:03] i will have to buy a few soon [17:45] ogra_ac: hi. any idea why package cpuburn from universe is not built for armel, and as such not available in the archive? the debian/control explicitely mentions armel as a target arch [17:47] is there a way to cause dma_map_single to only choose from a very small region of memory? say 0xf0000000 -> 0xf1000000? [17:50] ogra: and the armel .deb is available in debian === njpatel is now known as njpatel|away [18:58] is there a screen package for armel? [19:06] yes [19:10] can you direct me to it? [19:12] found it, nvm. [19:12] I was going to say ... [19:12] apt-get install screen _should_ work [19:12] even if it does install byobu for whatever reason [19:13] Martyn: I was trying to grab it for an offline install because my embedded system currently doesn't have network connectivity [19:14] https://launchpad.net/ubuntu/+source/screen/4.0.3-14ubuntu1.2/+build/1922902/+files/screen_4.0.3-14ubuntu1.2_armel.deb is where I ended up finding it [19:14] it has some dependencies === zyga_ is now known as zyga [19:37] howdy! after using ubuntu for several years and doing arm coding for also about as long - i just thought it might be a good idea to combine the both. So, the obvious question... whta might be a good point to start? [19:54] LetoThe2nd: What would you like to do? [19:54] If you just want to use an ARM based linux device, I reccomend the beagleBoard as a good starting place. [19:55] There's also the upcoming PandaBoard from TI, which is a dual core ARM platform that will be supported in Natty [19:56] Martyn: hehe, the "just using" point is long passed - i have an openrd here running debian and at work i maintain an embedded platform, for which i build the environment with ptxdist. :-) [19:58] but its more like being an end user at the moment, and i think i'd like to get more into the inner workings. I have reasonably good c knowledge and a little kernel experience (what it is you need to port a bsp to a closely related platform), but at the moment i'm more like: ok, i have a few building blocks and if i use them in order, it works. but what lives outside of "using existing blocks in given order"? [20:02] Martyn: heh, pandaboard is already supported in maverick :P [20:03] Well, yes.