=== Berstrom is now known as Martyn [08:39] plars: ping [08:40] saeed, ~3am where he lives ... [08:40] ogra: hi [08:40] hey [08:41] have you installed lucid over dove x0? [08:41] i'm not one of the lucky guys to have an X0 [08:41] I see [08:42] and i think NCommander is travelling, plars and GrueMaster ate in later timezones [08:42] *are [11:54] asac: ping [12:59] saeed: hi [13:02] asac: have you installed lucid on dove? [13:05] Hello [13:06] does somebody has some experienc an i.mx51 bbg board combined with gstreamer? [13:07] A lot of people here played with babbage boards, but with varying versions; less so with the gstreamer codecs [13:07] its the latest evk board [13:08] installed with the demo image from freescale itself [13:08] Problem is that I want to stream some rtp (mpeg ts, with divx) to this board [13:08] and use the gpu to decode [13:09] but it kindof hangs [13:09] I mean [13:09] things seem ok [13:09] it's detected [13:09] it does some things but then just "hangs" [13:09] i see no video [13:10] command to receive I use is: gst-launch-0.10 -v gstrtpbin name=rtpbin udpsrc caps="application/x-rtp, media=(string)video, clock-rate=(int)90000, encoding-name=(string)MP2T" port=1234 ! rtpbin.recv_rtp_sink_0 rtpbin. ! rtpmp2tdepay ! decodebin2 ! xvimagesink [13:10] do I choose a wrong output device or something? [13:10] do I miss an option? [13:10] on a normal pc on my network, it just works like that [13:10] and localy also [13:12] wv: Sorry, no idea about the hangs [13:12] You don't have any buffer though [13:12] You might want to add some ! queue elements in your pipeline [13:14] what are queue elements? [13:14] You could also try in software without the gpu decoder [13:14] I mean, I'm not a gstreamer specialist [13:15] wv: literally buffers [13:15] this pipeline is also constructed with some help from the gstreamer channel :) [13:15] wv: gst-inspect-0.10 queue will give you some doc [13:15] lool: ok [13:15] lool: I already tryed with vlc and it works ok, despite the lack of hardware acceleration [13:15] wv: but I'm not sure it's the problem; it's just usual to have at least one queue to have sufficient data for the decoders to work [13:16] wv: Try with gstreamer without hardware acceleration [13:16] wv: Or try the pipeline on your PC [13:16] wv: FSL's gst codecs require patching to gstreamer [13:16] and it might be that they either hang with some input, or that the gst stack misbehaves [13:17] I would validate my pipeline in pure software decoding mode, and then try with FSL's stack and optimized elements [13:20] lool: how can i test it without the fsl gst codecs? [13:21] it just loads them as I run this command [13:21] or do I uninstall them? [13:21] wv: Try on your Ubuntu host? [13:22] wv: You could just "mv" the elements away to try it on the board; you'll need replacement elements such as the ffmpeg divx decoder [13:23] I tryed it on my host and it worked (streamed to 127.0.0.1) [13:24] lool: also tried to another networked pc and worked also [13:24] wv: Ok; so try without the gpu codecs; look for the gstreamer plugins which provide these and move them away, see if the software decoders work [13:24] If it works with software decoders, you can basically point at the fsl plugins [13:25] If it hangs with software decoders, you should look into the fsl gstreamer patches, perhaps rebuilding an unmodified gstreamer [13:27] owkey [13:28] well, I'll first have a look to add some "queue" [13:28] * ogra lols about lool's creative long patch names [13:28] * lool uses git format-patch's default [13:28] long sentences with dashes ? [13:29] Yes [13:29] git format-patch takes the first line of the commit, just as shortlog does [13:29] And changes spaces to dashes [13:29] ah [13:30] ogra: Since you mention it, at your earliest convenience please test this new qemu-kvm upload; I wonder whether the GCC atomics changes break anything and would like some additional testing before submitting upstream [13:30] I usually submit to upstream first, but this change is quite intrusive [13:30] is that working around the glib issues or do i still need to export the gslice stuff ? [13:31] or are you just referring to -static here ? [13:31] This is unrelated, it's for the armel build [13:31] But it changes locking on all architectures [13:32] ok [13:32] given that i'm madly testing rootstock atm i'll have it with the next upload and will notice any issues i guess [13:34] * ogra curses compiz broken focus behavior in lucid [13:40] hi all [13:42] amitk, am I right that for changing board ID (understand assinging Netwalker's specific ID) I would need to modify RedBoot in NOR? [13:42] Sleep-Walker: redboot would need to pass the correct board id to the kernel, yes [13:43] whats the current id you see in /proc/cpuinfo ? [13:43] Sleep-Walker: what is it you're trying to do? [13:43] (the field is "Revision:" iirc) [13:44] I'm preparing that wiki page and read Netwalker's code [13:44] amitk, I'd like to have support for Netwalker in upstream [13:45] I'm trying to make some plan of pushing all neesary stuff [13:45] saeed: the smoke testing? let me ask whats the progress on that [13:46] (I know that large part of code need to be rewritten or modified a lot) [13:46] saeed: i was fighting with getting stuff done for feature freeze (which was today) ... so i didnt get to it personally yet [13:47] but GrueMaster and plars wanted to check it ... and NCommander afaik [13:47] Sleep-Walker: as I said the other day, try booting the stuff that is upstream with a modified bootloader. You'll find it probably boots [13:47] yeah [13:48] but there is more work left than support for i.MX51 [13:48] wifi, optical point, touch buttons,... [13:49] Sleep-Walker: one small step at a time is my motto. Personally, I want to concentrate on the drivers _on_ the SoC first. Not the external peripherals. [13:49] and there is lots of things there too [13:50] I'd rather try kexec instead of bricking my new toy for now [13:51] I'm currently not writing code itself, just creating plan, status page etc [13:52] asac: I've seen few issues with lucid installation, Iet me know when we can work on it [13:53] saeed: what issues did you see? [13:53] saeed: which image did you try? [13:53] http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-live/20100215/ ? [13:53] (i dont remember if i gave you the desktop image) [13:54] ogra: do you know whats up with the image builders? [13:54] they are down? [13:54] no logs for me [13:54] ah wait [13:54] looking at wrong logs i think [13:54] http://people.canonical.com/~ubuntu-archive/livefs-build-logs/lucid/ubuntu-netbook-imx51/ [13:54] has logs [13:55] buildd breakage fallout [13:55] today we should get images again [13:58] ogra: i dont even say a fatal error in the current log [13:58] see [13:58] it produces squashfs [13:58] and then its gone [13:58] huh ? [13:58] The following packages have unmet dependencies: [13:58] update-notifier: Depends: update-notifier-common (= 0.94) but 0.95 is to be installed [13:59] from http://people.canonical.com/~ubuntu-archive/livefs-build-logs/lucid/ubuntu-netbook-imx51/20100218/ [13:59] not sure what you look at :) [13:59] http://people.canonical.com/~ubuntu-archive/livefs-build-logs/lucid/ubuntu-dove/current/livecd-20100126-armel.out [13:59] ok [13:59] now i see [14:00] looking at netbook might help [14:00] yeah :) [14:04] hi, what does "Unhandled fault: external abort on non-linefetch (0x1018) at 0x4033e000" mean? [14:07] saeed: I tried the patched kernel yesterday and was able to boot [14:07] I did hit some problems when I built it into an image, it was complaining about my squashfs, so something went wrong when rebuilding it I suspect. [14:08] saeed: I'd like to try with the image once it's built in, but I'm traveling today, GrueMaster should be up in about 2 hours and would be able to test it though === dmart_ is now known as dmart === dmart_ is now known as dmart === dmart is now known as Guest45491 [16:19] saeed: I'm up now. What would you like to know? [16:23] GrueMaster: hi [16:23] I've tried to install lucid from live image on dove [16:24] dove x0 [16:24] The most recent image I have is 20100215, and it doesn't boot on X0. [16:24] It gets as far as uncompressing the kernel, then nothing. [16:25] ok, you should take this img http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-live/current/lucid-netbook-armel+dove.img [16:25] Oddly enough, I get the same problem with a known working Karmic image. [16:25] then take the image and initrd from http://people.canonical.com/~mcasadevall/dove/x0_casper/ [16:27] the kernel image from that link should work on x0 [16:29] GrueMaster: new dove x0 boards (rev 1.4) needs patches that still not merged to the lucid dove kernel [16:29] I planned on trying that later today. [16:30] the live images boots fine [16:30] but I started to see issues when installed it on sata hdd [16:30] the installation process doesn't complete gacefully [16:31] I mean there was no message the it completed, and some error message regarding ubiquity showed up [16:31] Any ideas as to why Karmic fails? Same kernel issue? [16:32] yes [16:33] eric uploaded kernel package that suppots dove x0 board at http://people.canonical.com/~ycmiao/dove-x0test/ [16:33] Ok, I'll look into them. Thanks. [17:34] Wee upstream u-boot gets support of imx51 with fec [17:35] for babbage actually (evk) [17:35] lool: \o/! [17:36] funny, wolfgang told me he doesnt want to merge the git tree when i pointed him to it yesterday [17:37] They just selected a new imx custodian on the u-boot list and the port was proposed immediatly thereafter [17:38] http://lists.denx.de/pipermail/u-boot/2010-February/067329.html [17:38] http://lists.denx.de/pipermail/u-boot/2010-February/067475.html [17:38] (Stefano Babic proposed himself as the custodian) [17:39] oh, funny [17:39] its not the freescale tree [17:39] It's from freescale though [17:39] but not the git tree [17:39] somehow that defeats the purpose [17:40] and that's the fec fix http://lists.denx.de/pipermail/u-boot/2010-February/067329.html [17:40] err [17:40] 08 ? [17:40] freescale has everyting on 09 [17:40] all 08 patches were dropped and rewritten [17:42] Funny, the new custodian gets pointed at our source package http://lists.denx.de/pipermail/u-boot/2010-February/067656.html [17:44] oh my [17:44] and that source is really messy [17:44] a mix of 08 and 09 patches [19:59] NCommander: ooo built, however it fails as http://www.openoffice.org/issues/show_bug.cgi?id=105359 [19:59] OpenOffice.org bug 105359 in udk "bridges:arm failure to start OOo - terminate called after throwing an instance of 'com::sun::star::ucb::InteractiveAugmentedIOException'" [Defect,Started: ] [20:00] armin76: there's already a bug for this, I just wanted to make sure it wasn't an Ubuntu specific bug; I need to do some testing but its possible this is a binutils issue [20:01] y [20:01] you have your answer now :) [20:01] armin76: thanks. [20:01] yw [20:02] openoffice-3.1.1: Wed Feb 17 02:12:23 2010: 1 day, 11 hours, 42 minutes, 56 seconds [20:22] apw: Thanks for udebs, linux-source, and meta, you rock :-) [20:53] even meta ? [20:53] wow