[00:00] Yep [00:00] Well, the 4430 is also dual core .. but the addition of mini PCI-e is awesome [00:01] and lets not forget the graphics, do they provide the same type of binary blob as in the x86 world, so it works with the latest kernel and xorg? [00:02] crap, really should have builtin my usb driver for net install.. [00:05] no [00:05] the nVidia driver is not quite so clean [11:07] NCommander: suihkulokki: zumbi: lool: http://bit.ly/cRN5ta <- about tegra2 in case you're interested === JaMaOff is now known as JaMa === JaMa is now known as JaMa|Zzzz [11:31] ssvb: ^ [11:35] armin76: Thanks, interesting, though it's a bit sad that they decided not to include NEON. On some tasks like for example video transcoding, NEON can provide a huge help, and it has nothing to do with floating point. [11:36] armin76: they probably want to offload all this stuff to some dedicated HW accelerators, but the question is how well are they going to be supported in linux [11:39] ssvb: well, you've read what they've said about why neon is missing [11:41] armin76: yes, and it did not sound convincing [11:42] ok :) [11:42] ssvb: btw, if you have any test, i'm happy to do it :) [11:46] armin76: dual-core cortex-a9 - does it survive ltp pthread stress tests? :) [11:50] suihkulokki: gimme link and i'll tel you [11:50] tell* [12:00] armin76: apt-get install ltp [12:07] lol [12:07] apt-get install ltp [12:07] -bash: apt-get: command not found [12:08] :P [12:09] armin76: 'emerge -pv ltp' :-) [12:10] yep :P [12:46] suihkulokki: what command should i use? [13:27] armin76: all the pth* and mmap*, I think you can select them from the menu too === JaMa|Zzzz is now known as JaMa [15:06] http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-live/20100327/ [15:06] now we only need a kernel that actually boots for OMAP :) [15:07] cool ogra! [15:08] to sad it doesnt boot though [15:08] DSS2 patches still missing and aufs is broken .... but we'll get there [15:08] it should boot... is it based off the orignal 500 kernel? [15:08] (dl'ing) [15:09] or wait, does it stop at 'Uncompressing... "? [15:09] no, it stops when trying to mount the squashfs [15:10] but you wont see that since the output defaults to framebuffer (which stays dark without DSS2) [15:10] okay, differenet bug effecting me this morning.. ;) [15:11] the live image needs any kind of working union filesystem [15:11] unionfs isnt enabled in the current kernel build and aufs segfaults ... [15:11] so there is no way to boot the image atm [15:11] but we'll get there [15:12] next kernel upload should make it work [15:12] yeah, DSS2 still needs this enabled for the beagle http://bazaar.launchpad.net/~beagleboard-kernel/%2Bjunk/lucid-ti-omap/annotate/head%3A/patches/beagle/002-beagle-dss2.diff (and it needs to be tweaked for 2.6.34-rc2's dss2 patches amit pulled in) [15:12] right, he is working on that [15:13] we'll have alternate server images soon, these dont need union filesystems and should boot [15:13] my main focus was on the live image for now, server is next [15:14] okay, sounds right... Is there a good way to resync the download without pulling 500mb every new upload? [15:15] use zsync [15:15] there is a zsync file on the download page [15:15] okay, thanks i'll look into zsync.. [15:16] I was -so- picked over by airport security today. I have both the OMAP44xx board and the tegra board in my baggage. They insisted in looking through _everything_ [15:17] in your babbage? :D [15:17] I think one of the ribbon cables to the OMAP4 board was damaged ... so I had a supervisor -sign- that they took it out of a protective case and that they are responsible. [15:17] It's a slow day a the airport, I think they were bored and wanted to search stuff [15:19] hey ogra, not sure if you subscript to the beagle group, but koen posted a new MLO/U-boot combo, so every board needs to be updated.. http://groups.google.com/group/beagleboard/browse_thread/thread/110b417a0ec7b03d [15:20] ah, thanks, i'll look into it next week [15:20] rcn-ee : There's also a new release of UEFI for beagle from last week as well [15:20] we've been testing it @ Smooth-Stone .. it's not bad. [15:21] All this week, I'm looking at polishing the kexec support in the ARM kernel, and trying for a 'soft' bootloader based on a kernel [15:21] from what i've seen the idea looks promising... but we still have a lot of things u-boot set's up for the kenrel (ehci power) [15:22] rcn-ee : Yep, this is why I'm working on a kernel based bootloader [15:22] that way you can get all the drivers up, and use a kernel to boot -another- kernel in a more generic fashion [15:23] with a configuration file present in an ext2/3/4, or fat/vfat filesystem (along with boot kernel and ramdisk) [15:23] well, *if* kexec works [15:23] ogra_cmpc : I'm ---so--- close [15:23] kinda sounds like device tree's. ;) [15:23] rcn-ee : I'm working with Grant .. yea [15:24] Martyn, well, our kexec experiments in ludic (at least for the supported kerenls) broke suspend/resume [15:24] *lucid [15:24] * ogra_cmpc goe afk doing weekendish stuff [15:24] but I have the code working now .. testing it on OMAP3, OMAP4, i.mx51/515, dove, tegra2 and other arm v7 processors I can get my hands on [15:24] ogra_cmpc : Have a good weekend :) [15:25] ogra_cmpc : Yeah, I came across that early. The problem was with the way kexec blows over some registers .. I added a patch that saves the registers first. [15:25] ogra_cmpc : I'm going to submit the patch to arm-kernel list in about two weeks [15:25] and I expect this work to be relevant to lucid+1 [15:26] (next UDS :) ) [15:48] re [15:48] airport wifi FTL === jldugger is now known as pwnguin [18:03] armin76: interesting, thanks [20:46] lool: just got a set of git emails yesterday on pselect support landing on trunk [21:26] rcn-ee: you around? I tried the Lucid stuff, and found it doesn't seem to handle MTD at all. [21:26] That is, the in-repo kernel. [21:28] DanaG, yeap, i think i fixed that one.. (cat /proc/mtd gives the devices) CONFIG_MTD_NAND_OMAP2 needs to be builtin... [21:28] I also got nonworking USB (both EHCI and musb). [21:29] yeap, that also requires 3 modules to be builtin... i sent an email to ubuntu's kernel-team with a patch for that... [21:31] fingers crossed, amit will commit the change later this week https://lists.ubuntu.com/archives/kernel-team/2010-March/009484.html [21:31] What's new between v2.6.33-dl.1/ and v2.6.33.1-dl4/, in a nutshell? (don't want to read however many hundreds of git commits =þ ) [21:32] DanaG, some good news: we have a netbook image.. (although it doesn't boot) http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-live/current/ [21:32] cool, though. [21:33] in a nut shell, i'm trying to figure out why .1 isn't booting on Squeeze & lucid... [21:34] Cross Compiled with 4.3.1, 4.4.1 works... Native built in lucid works (uImage)... But the Squeeze and Lucid Deb packages all fail at boot at uncompressing... [21:34] I'm also wondering about that specialcomp.com LCD board... 480x272 (or whatever the resolution was) doesn't seem to already be in the kernel. [21:34] i think .1-d5 fixed it.. but the beagle won't finish that deb package for another 4 hours.. [21:36] Cool, never heard of their lcd... [21:37] I've also heard of some other companies having bigger LCDs. [21:37] someone would have to implement that.. I haven't connected any lcd thru the lcd interface.. [21:37] (just hdmi/dvi connected ones) [21:39] oh, the other weird thing... the lenny deb package (built with 4.3.x) for .1-dx works just fine, running it on one beagle.. it defintelly looks like a 4.4.x native bug.. [21:41] oh yeah, and I tried suspend on the beagle... it suspended, but didn't seem to give any way to wake it up. =þ [21:43] yeah, not sure what you could wake it up with.. most perhiperals need power from the twlxxxx chip which the kernel and u-boot setup... [21:45] That 3-port-hub-with-ethernet is nice -- in Windows on my host system, it can USB-wake the host. [21:46] hmm, are audio and/or midi gadget enabled now? [21:47] audio output from a wave file worked when i tested it, but all the gadget stuff relies on the usb modules being builtin first.. [21:49] does it work to keep g_ether as a module? [21:49] And can you load more than one gadget-type driver (such as g_ether and g_audio) simultaneously? [21:51] it's possible, but once you do that on boards with only the otg interface, your stuck with using an mmc card as rootfs... [21:53] It's what they do in Angstrom, I've gone in the oposite direction with my debian/ubuntu kernel's as i have the rootfs on a big usb harddrive... [22:08] hmm, I'm also curious whether gadget audio can do multichannel. [22:09] I have a USB surround-sound card that goes all crackly-sputtery in Windows... so I'd like to be able to use the beagleboard as an intermediary. [22:09] windows talks to beagleboard as USB sound card... and then beagleboard talks to real sound card. [22:18] hmm, so does -1-dl3 work? [22:19] nope, thats the one that fails at decompression.. [22:19] i think i solved it in .1-dl5, it should be done in 4 hours.. [22:20] just watch: http://rcn-ee.homeip.net:81/dl/farm/log/BUILDING-2.6.33.1-dl5_1.0-lucid.txt [22:21] i tested a uimage and modules builtin a chroot for that and it worked on lucid, so the deb package variation of it shoudl work... [22:30] Dec 31 16:12:08 beagleboard NetworkManager: device_creator(): /sys/devices/platform/musb_hdrc/gadget/net/usb0: couldn't determine device driver; ignoring... [22:30] hmm, it seems networkmanager doesn't like usb0. Can't use networkmanager to make the shared-out network. [22:45] linux-image-2.6.33-500-omap will be upgraded from version 2.6.33-500.3 to version 2.6.33-500.3. [22:45] weird.