[00:31] i'm trying to boot maverick on my beagleboard [00:31] its stuck on 'Uncompressing Linux... done, booting kernel' [00:31] You had a C3, right? [00:31] -xm [00:31] Ah, sorry. My poor memory. [00:33] do you want boot log? or got other ideas? [00:33] I think you need someone who has an XM. [00:36] persia, any idea who that is? [00:37] There's been a few folks here who talked about them in the past. I'm not sure if any of them have them now, or would be able to troubleshoot with you. [00:39] daurnimator: First, what image are you running? [00:40] Also, do you have a monitor attached? [00:40] was trying ubuntu-netbook-10.10-preinstalled-netbook-armel+omap.img. yes I have a monitor attached. [00:42] Hmm. You should have video on the monitor. [00:43] It should show text output indicating that it is enlarging the root file system. [00:44] that I do not. [00:45] if it helps, D6 is flashing twice at a regular rate.. [00:46] ogra: ping [00:46] ogra_ac: ping [00:48] GrueMaster, you have an -xm? [00:48] yes. [00:48] Not hooked up currently, though. [00:48] GrueMaster, advice/helpful tips then? [00:49] First, I would check your downloads with md5sum. Compare it with the MD5SUMS file online. [00:49] well the gzip archive extracted correctly... [00:50] What size sd card are you using? [00:51] ok. the md5 is correct [00:51] sd card is 8gb [00:51] Ok. I have tested 4, 8, and 16G, so that should be ok. [00:52] Try rebooting. [00:52] Is this a production XM? [00:52] yes [00:52] ok. [00:52] as I sit here I haven't touched it, D6 is flashing twice regulary, and D7 occasionally turns switches on/off [00:52] I feel like it could be doing something? [00:53] How did you write to SD. [00:53] how long does the rootfs take? [00:53] ? [00:53] On 8G, about 5-10 minutes max. [00:53] zcat | sudo dd [00:54] Try gunzip < > ;sudo dd bs=4M if= [00:54] I have seen a few issues using zcat pipes. [00:55] i've got to go out in a sec, so I might just leave it for now, see if something has happended by the time I get back [00:55] will be back in a couple of hourss [00:55] Reset it first. YOu should see something on the screen almost immediately. [00:55] ~10 seconds after reset. === rsalveti` is now known as rsalveti [01:10] GrueMaster, nope [01:11] pastebin [01:11] uh, I meant http://pastey.net/141810 [01:11] GrueMaster, ^^ hope that helps, its where it is now [01:11] Hmm. Try "sudo dd bs=1024 if=/dev/zero of=" to wipe your SD & reflash. [01:11] i'm off for for a bit now [01:12] You can also try adding serial console info to the kernel parameters. WIll explain when you return. [01:57] GrueMaster, back for 10 mins, I previously had alot larger block size: what wil change? [02:00] For this sort of thing, the only difference between different blocksizes for dd is how the caching works. I almost always use 1M. [02:02] Very large bs (where zcat is the max) tends to be more likely to get some unexpected corruption. Very small bs (anything less than the eraseblock size on the flash) tends to cause the FTL to be overworked, and may waste space on the SD. [02:06] persia, so whats the value I should be using? 1024? 1M? 4M? [02:06] What's the eraseblock size on your flash? Use some integer multiple of that. [02:06] prpplague: hi dave, got your email. [02:07] persia, how would I find that one out? >.< [02:07] 1M and 4M tend to be integer multiple for most eraseblock sizes. 4M is a bit faster. 1M is a bit safer (but breaks for 2M eraseblock sizes, although these are very rare) [02:08] You'd have to find it in the documentation for your SD. In practice, I don't know of any good ways to discover this based on the hardware. [02:14] cooloney: hey bud [03:05] ahoy ! thank you for providing the community with 10.10 for the beagleboard! (https://wiki.ubuntu.com/ARM/OMAPMaverickInstall) [03:05] now i am having problems =D usb keyboard and mouse aren't being recognized! [03:25] i have returned :D [03:25] * daurnimator pokes GrueMaster [03:25] Oww. [03:25] Sorry, had stepped out to start dinner. [03:26] Ok, so lets try with serial console to see where it is hanging up at. Reset and press -c to stop autoboot. [03:27] Let me know when you have a uboot prompt. [03:29] now. [03:29] one sec (phone call) [03:29] ok, you're second is up :P [03:29] Heh. Family. [03:29] ahh [03:29] hope everything is ok [03:30] At any rate, type "mmc init" [03:30] Jut my son wanting a new monitor. [03:30] just [03:30] ahh [03:30] and then? [03:30] fatload mmc 0:1 0x80000000 uImage [03:30] fatload mmc 0:1 0x81600000 uInitrd [03:30] setenv bootargs vram=12M omapfb.mode=dvi:1280x720MR-16@60 root=/dev/mmcblk0p2 fixrtc console=ttyS2,115200 console=tty1 [03:30] bootm 0x80000000 0x81600000 [03:31] That should load the kernel & initrd and boot with a lot of stuff on the serial console. [03:33] mouse-_, You've connected them to a powered hub? [03:34] persia: Not getting that far. Should be able to boot to oem-config w/o mouse/keyboard. [03:34] Currently nothing on display. [03:35] GrueMaster, "mouse-_" != "daurnimator" :) [03:35] Oh. Oops. Didn't see that person. [03:35] persia, i'm on dc power now [03:36] GrueMaster, its doing things :) [03:36] daurnimator: Anything on the serial console? [03:36] cool [03:36] currently on setting up swap, this will take a moment ... [03:37] Anything on the monitor or is it serial console only? [03:38] nothing on monitor [03:38] On the monitor, is it HDMI or DVI? [03:38] using hdmi=> dvi cable [03:39] Hmm. Should work. Make sure monitor has DVI input selected. [03:39] What res can the monitor handle? [03:40] up to 1920x1080 [03:41] So it should be able to handle the 1280x720 res. [03:41] Depends. Some displays can't handle some non-native resolutions. [03:42] Might pass 1920x1080 with setenv to see if it makes a difference. [03:42] I'd have to look, but I believe the 1280x720-16@60 is vesa compliant. [03:42] It is: I'm just not convinced arbitrary displays are VESA compliant :) [03:43] The're *supposed* to be so though. [03:43] True. [03:43] I have to go fix dinner. Cooking part is done, now need hands on attention. Sorry. [03:45] persia: no, this is the beagleboard XM. the board comes with 4 usb ports =D [03:46] mouse-_, what did you do to get it working? mines not [03:46] the XM is so badly supported right now :( [03:46] mouse-_, Ah. I don't have an XM, but on my C4, I needed to supply additional power through a hub to get the keyboard and mouse to work. [03:47] mouse-_, Just needs more folk looking at it: I'm certain it *can* be supported, but unless a number of XM users use Ubuntu and help, it's unlikely the rest of us can wave a magic wand :) [03:47] :) [03:48] ok, so the board rebooted now [03:48] still no display [03:48] mouse-_, Your display just worked? [03:48] the display does just work ! [03:48] this is 10.10 [03:49] i'm guessing, maybe someone forgot to compile usb support for it, being that it was just released 8 days ago? lol. i am trying 10.04 now [03:49] Hmm. Are there different variants of the XM? [03:49] but the setup_sdcard.sh crashed my kernel! ha! [03:49] mouse-_, 10.04 is very likely not to support the XM: I don't believe they were widely available in April. [03:49] here is my serial output: http://pastey.net/141812 [03:49] ahhh, crap, persia. yeah, you're right [03:50] check this out, [ 50.032255] EXT4-fs warning (device mmcblk0p2): ext4_clear_journal_err: Marking fs in need of filesystem check. [ 50.032276] BUG: unable to handle kernel paging request at 027c0000 [03:50] ha! [03:50] ext4 bug. [03:51] or something else to do with it. what a nightmare [03:52] XM is only supported in Maverick. Also, I just finished booting mine again. I am at the oem-config. Mouse & keyboard work. Try bottom USB slots. [03:52] Gone again. [04:28] so yeah, it seems to be booting now [04:28] but theres nothing on the display [04:35] Even hardcoding the proper resolution? [04:35] What happens if you send Ctrl+Alt+F1? Should get you a VC. [05:46] hello [05:47] Yesterday I've installed canonical maveric prebuild image on BB C4. The speed is too slow for normal work - is it OK? [05:51] and my UCB mouse connected to unpowerd USB hub @ EHCI port is blinking 1 per 2 seconds and not working properly! Althought in Angstrom image it worked perfect. [05:56] Sonny_8bit__, I found I needed to use a powered USB hub to get mouse and keyboard working the way I wanted. [05:56] I agree the speed is very slow: I find it nearly unusable. [05:57] But the C4 doesn't meet minimum Ubuntu specifications, so I'm not that surprised. [05:57] where's a bottleneck? [05:57] Speed might be a bit better if you try uninstalling ubuntu-netbook and try installing lubuntu-desktop [05:58] Three things: there's not enough RAM (minimum is 384), the CPU is kinda slow, and the graphics drivers aren't installed by default. [05:58] ubuntu netbook @ Geode LX with 500MHz @ 256 MB DDR worked much better! [05:59] may be the problem is low speed of SD-card read-write? [05:59] Sonny_8bit__: Will 10.10 even RUN on a geode? [05:59] That doesN't help. [05:59] You might try instlaling powervr-omap3: I believe that will give you better graphics support. [05:59] i try only 8.xx [06:00] anyway winXP @ Geode works fine enough [06:00] of course w/out 3D [06:02] Martyn, At least some docs I find claim the Geode LX is i686 compatible, so it ought run Maverick (but I expect it to have performance similar to omap3, albeit likely with accelerated X without installing a driver package) [06:02] ok today i will buy new powerd usb hub and try again. and also try to install Maverick to Gedoe [06:02] *Geode [06:03] Do try installing powervr-omap3: it should at least make video performance a *lot* faster, and offloading that from the already overwhelmed CPU can't hurt. [06:04] ok thank you. hope this helps. [06:05] Also i want to ask how to boot Linux from the ext. USB HDD? [06:05] Sonny_8bit__: for c4 sd card and lack of ram are the main problem [06:05] Sonny_8bit__, Just be prepared for performance on the Geode to be just as slow as the C4. [06:05] My understanding is that the C4 *can't* boot from USB. [06:05] ubuntu in general eats lots of ram [06:05] You could put a kernel on NAND, and then boot with rootfs on USB if you like. [06:05] lots of daemons all around [06:06] persia: yes, it can [06:06] rsalveti, We prefer to say that Ubuntu avoids the user to perform manual actions to deal with environmental concerns :) [06:06] :) [06:06] but as c4 has nand, you could just write the correct boot arg and then use the rootfs at your usb [06:06] rsalveti, Really? Loading the bootloader from USB? I thought that was a ROM issue. [06:07] persia: nops, you can even with panda [06:07] uart, usb, nand and sd [06:07] can you give detailed steps how to boot from USB? [06:07] may be @ wiki page [06:07] yeah, need to create the wiki page [06:08] what I've done with my panda rootfs is to run the pre-installed, run the first boot and the setup (user account and etc) [06:08] i should buy panda instead of BB [06:08] then I mounted the sd card at my host pc, copied the files (careful with file owners and permission) to my usb disk [06:09] rsalveti, So, how? And if so, why don't we have a regular install? [06:09] then you just need to change the boot args from the first partition [06:09] to use root=/dev/sda1 for example, instead of root=UUID.... [06:10] persia: http://elinux.org/BeagleBoardRecovery see usb [06:10] i will try to do the same [06:12] Sonny_8bit__: be sure you're using a powered hub [06:14] will be out now, ping me in a few hours in case it doesn't work for you [06:14] * rsalveti out === hrw|gone is now known as hrw [07:58] aloha [08:18] hey guys whatever happened to ubuntu netbook remix for arm? [08:21] what is supposed to go in /etc/apt/sources.list for armel maverick (on cortex, this is a different machine) === JaMa|Zzz is now known as JaMa|Wrk [08:42] got it it's on ports [08:53] dommod, Some time back the netbook team decided to play with us, and stopped being a remix. It's one of the standard images (and likely the most popular) [08:54] tmzt_, The software-sources tool is supposed to do the right thing. If you don't have a local mirror and don't want anything fancy, just deb http://ports.ubuntu.com/ubuntu-ports maverick main restricted universe multiverse (repeat for maverick-security and maverick-updates) [08:55] what's the different with ubuntu-ports ? [08:55] like I'm missing xxd without it [08:55] ok [08:56] Nothing should be different with ubuntu-ports except architecture. If you find something, please complain loudly and file lots of bugs. [08:58] of course xxd is already installed I knew that [09:08] persia: oh ok. did it ever got released? i can't find any information.. [09:32] ogra: i tried to fresh install the 10.10 release on my es2.1 cleanly. [09:32] ogra: and then upgrade to latest kernel and alsa-lib, alsa-utils things in -proposed [09:32] ogra: still fail to get sound work, there is only dummy device [09:39] cooloney: I think one of the files have been placed in the incorrect location [09:40] cooloney: Where is the omap4 file on your system? [09:47] lag, cooloney: ogra said yesterday he has an update for this [09:47] lag: do you mean the omap4 for alsa? [09:47] yes [09:48] find / -name omap4 [09:49] Where did it end up? [09:49] lrg and lag, it looks like there is no such omap4 fiel in $ ls /usr/share/alsa/init/ [09:49] 00main default hda help info test [09:49] Where did it go then? [09:49] no idea [09:49] Issue that command [09:50] do you guys know the latest version of alsa-lib or alsa-util which including the fixing? [09:50] alsa-utils (1.0.23-3ubuntu1) natty [09:51] It was on the bug [09:51] alsa-utils 1.0.23-2ubuntu3.2 [09:51] that's mine [09:51] libasound2 1.0.23-1ubuntu2.1 [09:51] Then you still have the error [09:52] i still don't have the sound [09:52] kernel is the latest one [09:52] 903.16 [09:54] But alsa-utils are not the latest and the omap4 file is missing [09:54] It's not surprising that your sound doesn't work :) [09:55] i think it is not in the repo for upgrade [09:57] dommod, It's just part of the normal release. Check the ports/ directories on cdimage.ubuntu.com [09:57] cooloney, Did you enable -proposed? [09:57] persia: yeah, [09:58] persia: thanks, i will check it out.. [09:58] Hrm. [09:58] dommod, http://cdimage.ubuntu.com/ubuntu-netbook/ports/releases/maverick/ [09:59] persia: deb http://ports.ubuntu.com/ubuntu-ports maverick-proposed restricted main universe multiverse [09:59] that's in my source.list [10:00] just finished one apt-get update. nothing to upgrade now. [10:00] lag: why we need to use this natty version 'alsa-utils (1.0.23-3ubuntu1) natty'? [10:01] Because that's the one the fix is in [10:01] Look at the bug [10:01] lag, Was that not submitted for SRU? [10:02] What? The alsa-utils fix? [10:04] Yes. [10:04] No idea [10:04] I've had no input WRT alsa-utils [10:05] The only information I have is on the bug report [10:05] bug 637947 [10:05] Launchpad bug 637947 in linux-ti-omap4 (Ubuntu Maverick) (and 3 other projects) "no sound devices on current ES2.0 boards (affects: 2) (dups: 1) (heat: 125)" [High,Fix committed] https://launchpad.net/bugs/637947 [10:05] I thought someone was working on an SRU. Odd that it's not there. [10:05] Don't know dude [10:05] ogra: Is your man [10:05] Probably, indeed. [10:06] lag: i am using the latest one of alsa-utils, i think [10:06] alsa-utils" versions published in Ubuntu [10:06] Natty (1.0.23-3ubuntu1): main/sound [10:06] Maverick-proposed (1.0.23-2ubuntu3.2): main/sound [10:07] it is from Maverick-proposed [10:07] That looks like the latest version, yes [10:07] You said you were using 1.0.23-2ubuntu3.2 a moment ago [10:08] 3.2 looks to have the right patch in it. [10:09] persia: Where did you look to find that out? [10:09] https://launchpad.net/ubuntu/+source/alsa-utils/ [10:09] Expand the details of 1.0.23-2ubuntu3.2 [10:10] cooloney, lag, i missed to also patch the Makefile (assuming there was a .install file) so omap4 isnt installed at all [10:10] Ah [10:10] fix is already in the queue [10:10] Ah, implementation bug. Now it all makes sense. [10:10] That'll do it [10:10] -v through 3.3? [10:10] and i'm still waiting for it before starting to tinker with the pulse stuff [10:11] ogra: ok, got it. [11:28] http://marcin.juszkiewicz.com.pl/2010/10/19/how-to-cross-compile-arm-kernel-under-ubuntu-10-10/ [11:31] hrw: Looks good [11:31] tested before posting ;D [11:32] I'll bookmark it in case I forget how :) [11:32] mkey [11:33] hrw: Will you be at HDS? [11:33] lag: HDS? [11:34] Wow, that was all kinds of wrong [11:34] UDS = Ubuntu Developer Summit (I didn't mean this either) [11:34] I meant UDS [11:34] Ah [11:34] he has to be, he subscribed to my specs ! [11:34] UDS? I will be [11:34] What is wrong with me today [11:34] You have UDS on the brain [11:35] It was crossed between UDS and UHS [11:35] But missed the U and put both U and H [11:35] UH [11:35] :) [11:35] HU [11:35] :) [11:35] lag: I have to be at uds to get my ea1 pandaboard from JamieBennett's hands [11:35] isn't UHS usually in September, anyway? [11:35] EA1? [11:35] Or ES1? [11:35] lag: EA1 [11:35] EA1 = ES2.1 [11:36] EA1 = es2.0 [11:36] persia: Yeah, I was there [11:36] Are they the same? [11:36] ogra_ac: ea1 with es2.1 are TI internal [11:36] its the same as 2.1 [11:36] What does ES and EA mean? [11:36] lag: ES is silicon version. EA1 is pandaboard early adopters [11:36] ES -> Engineering sample [11:36] lag: ES = Extra Special EA = Extra Awesome [11:37] ogra_ac: es2.0 has 200MHz mem instead of 400MHz on es2.1 [11:37] Lol @ JamieBennett [11:37] or sth like that [11:37] hrw, afaik only es2.1 go to the public [11:37] ogra_ac: Where's mine? [11:38] I still don't have one [11:38] * lag jumps up and down! [11:38] hrw, there are two versions of the 2.1 [11:38] ogra_ac: A1 boards are es2.1 but 4weeks at digikey [11:38] lag, i guess still at pandaboard.org [11:38] anyway it is not es1.0 [11:38] well, the 2.0 had probs [11:39] at least the 6 layer one [11:39] which iirc is the one with the black PCB [11:39] there was an 8 layer with green PCB that was close to what the 2.1 now is [11:40] fi you have a green one and are sure ist es2.0 then you are fine === amitk is now known as amitk-afk [11:40] ogra_ac: JamieBennett has my board and board is working. thats all I need so far [11:45] hrw, well, if the compiler doesnt die mid way in a build all is fine i guess :) [11:46] ogra_ac: worksforme [11:46] (and if it boots with the final image) [11:47] ogra_ac: I know that ubuntu people assumes that cross compilation is source of all evil etc but maybe it is time to forget that? [11:47] ?? [11:47] ;D [11:47] did i refer to cross compilation anywhere ? [11:47] we even have a cross howto link in our topic ;) [11:48] hey i'm trying to install ubuntu on my nslu2 and it's killing ssh with an OOM message when it goes to detect the disks [11:48] saboteur: 9.10? [11:48] ops.. 9.04? [11:48] hrw, 9.10 had no support for nslu2 [11:48] ah [11:48] this is the link - https://wiki.ubuntu.com/ARM/NSLU2 [11:48] snap [11:49] jaunty ? [11:49] * ogra_ac is happy that jaunty is EOL next week [11:49] saboteur: I would go for Debian instead of such old release [11:49] yeah, better take debian [11:49] ogra_ac: looks like I need to write newer version of amitk's post one day [11:49] ok, i'll give it a go :) [11:50] ogra_ac: as his points to deprecated repos [11:50] you know how long writing the flash takes though, lol [11:50] saboteur: nslu2 has such small amount of flash that nearly no one uses it for rootfs [11:50] this device has to only run mrtg so its sufficient [11:50] ok [11:51] hrw: actually not rootfs, i just meant the installer image [11:51] You really want to be installing Debian, not Ubuntu. Don't worry about the Ubuntu installer image working. [11:51] it seemed to take about an hour to run through upslug2 [11:52] you should mark that page as being broken, although if that software is EOL soon there's probably little point [11:53] I'm planning to remove that page in about 2 weeks (just because I'll be at a conference in a week) [11:53] ah :) [11:53] it runs nicely on sheevas and stuff though yeah? [11:53] Install it if you like, but if you're the sort active enough to come ask for support, we'll do our best to point you at the best way to support your device. [11:53] 9.04 does. [11:53] 9.10 won't. Needs Coretex-A8 or Coretex-A9 [11:54] ah [11:54] hrw, feel free to fix the topic, its open for editing [11:54] I sympathise with you: I've a few machines that would be running Ubuntu ARM if it weren't for that. Still, Debian works. [11:55] cool, well thanks for the pointers [11:56] If you want to run Ubuntu, Sharp, Genesi, and Toshiba have produced hardware that works. [11:56] And TI just announced the Panda board, which fits in a Bamboo case to make a complete solution. [11:57] (extra bonus: the TI kernel is in-archive and fully supported). [11:57] and the panda is about the same price as the slug [11:57] persia: 9.10 needs arm1136 only. 10.04 needs cortex-Ax [11:57] yeah, 9.10 still was v6 iirc [11:58] I need to finally restart my sheevaplug... so far it is fastest arm here [11:58] until you get your panda ;) [11:59] ogra_ac: yep. then only clock will be faster [12:01] Advertised panda clocks are still slower than my QNAP, but would expect panda+bamboo would be perceived as faster for most purposes. [12:03] persia: I hope that bamboo will be released soon [12:05] I'm not that fussed. I'm not expecting to be in the first round of HW releases anyway (given how many other people will do more useful things than have another desktop to use for IRC), so I'm convinced bamboo will be available by the time I have a need for one. [12:06] (and it is my secret hope that someone will release an omap4 consumer device first) [12:06] (well, maybe not so secret) [12:07] just for interests sake and for nerd factor, this will be monitoring my ups voltages etc as well as link utilization and solar inverter output/temps [12:07] persia: anyway even naked pandaboard will be better then my current beagleboard c3 setup [12:07] what a way to kill time [12:08] saboteur, solar inverter management is critical: I'm not sure that's any sort of waste. [12:08] hrw, Absolutely. [12:08] :) [12:09] later, i'm out! [12:09] thanks again [12:09] persia: not in power meaning - much less cables [12:10] Why less cables? Don't you still need KVM+E? [12:11] persia: beagleboard c3 + psu + usb hub + usbwifi + usbhdd + usbsd + usbmouse + usbkeyboard + usbhub + 2nd psu + ethernet cable + hdmi cable + serial cable [12:12] Yeah, that's too many. [12:12] persia: now it will be: pandaboard + hdmi + usbmouse + usbkeyboard + usbhdd + serial cable [12:12] network via wl1271 wifi [12:12] ah.. one usbhub will be needed [12:12] but power from same psu as PB [12:13] pfft, serial cable [12:13] ogra_ac: 7 serial ports in desktop needs use [12:13] nothing to see or do on the serial console [12:14] ogra_ac: tell it to my alix.1c machine [12:14] you manage all of u-boot through /boot/boot.script and flash-kernel [12:14] (only talking of the panda ) [12:14] I have "Serial Graphics Adapter" even in alix.1c ;D [12:14] heh [12:14] Serial consoles are useful (for servers without IPMI [12:14] weird HW [12:15] hrw: you should. the cross-compile landscape has been changing too rapidly for me to keep up === amitk-afk is now known as amitk [12:15] persia, not on panda though [12:15] ogra, Sure, but panda is a desktop, not a server. [12:15] ogra_ac: serial cable or ethernet cable... I prefer serial one [12:15] Anyone building an omap4 server ought add IPMI, really. [12:15] panda can be a server too [12:15] persia: omap4 is not good for server [12:16] Sure. So can a Mac Mini (I have several as servers). Doesn't make them server designs. [12:16] hrw, Why not? [12:16] and i still wouldnt recommend fussing around on the serial console [12:16] no ethernet, no storage controller [12:16] hrw, I thought there was talk about a way to get storage working. ethernet USB isn't that bad. [12:16] persia: kirkwood with armv7a would be lovely server chip [12:17] persia: 110MB/s via sata contra 30MB/s via usb? [12:17] hrw, well, you havent subscribed to https://blueprints.edge.launchpad.net/ubuntu/+spec/other-arm-n-public-panda-ppa-build-cluster :) [12:17] 800Mbps via 1GbE contra 40Mbps via usb-ethernet? [12:17] i'll show you how well a server it is ;) [12:17] hrw, kirkwood doesn't do that. I think the successor is called "dove", but that's rumour and speculation. [12:17] no, its called armada in public i think [12:17] persia: kirkwood is nice armv5 chip for server use [12:18] (but, yes, I have a kirkwood in a case with SATA and like it very much indeed) [12:18] hrw, Even my 1.2GHz kirkwood isn't as fast as some folk would like for some workloads. Compute-intensive servers do better with multi-core chips. [12:18] persia: so with omap4 transferring of 2GB of data from sata drive over ethernet would make usb controller "overheat" not giving any performance [12:19] persia: sure. but for NAS it is ideal [12:19] Who said anything about USB for storage. Aren't there some huge number of IO pins? [12:19] persia: want to connect sata over gpio lines? [12:19] Oh, for NAS, yes, kirkwood is sufficient. I'm not sure about ideal, as I'd like more computation for some things. [12:19] hrw, Why not? [12:20] persia, oh, btw, do you think we need a spec for omap3 kernel version ? [12:20] ogra, How much do you really want that in a separate source? [12:20] persia: some time ago prpplague told that omap4 gpio is not able to do 33MHz [12:20] i,e, the possible swithc to linaro in case the kernel team doesnt want it in a separate branch [12:20] * persia kinda wants to move another kernel *into* the mainline source [12:20] ugh [12:20] apw, what do you think ? [12:20] persia: coreboot developer wants to make x86 hw debugger on pandaboard base [12:20] hrw, when it comes to the detailed technique of how to make SATA work, prpplague is the person I look towards to come up with a good answer. [12:21] persia: indeed [12:21] ogra, So, it's my feeling that if you think you need the flexibility, you're expecting to be breaking rules we have in place for very good reasons. [12:22] persia, well, it affects the kernel team more than us i think [12:22] hmm. looks like pandaboard will be able to get 320GB sata disk + enclosure [12:22] but i'll happily switch to linaros omap3 kernel instead of having a separate branch [12:23] ogra, Yes, but also only affects them when you want to break the rules. [12:23] If we can follow the rules, it's not much extra trouble. [12:23] ogra_ac: less maintaince for team? [12:23] persia, as long as they build arm from teh main tree they are always affected [12:23] its about the delay [12:24] What delay? [12:24] and the testing they have to do [12:24] builds get massively slowed down due to arm builds [12:24] Yeah, but it's not a bad thing. [12:24] The alternative is what happens to the powerpc -meta packages, which commonly don't get uploaded for SRUs for *weeks* [12:24] for them it is [12:24] ogra_ac: grab 32 pandas from TI, make a cluster and run buildd? [12:25] It's not that long, and it will be less long with better buildds. [12:25] * persia agrees with hrw on this [12:25] Anyway, there's a spec on kernel flavours: maybe mention it there. [12:25] hrw, see above, thats what i plan [12:25] It's really a kernel team decision unless there's some special reason someone else wants to have an external source. [12:25] hrw, wont help though [12:25] ogra_ac: it is a matter of grabbing 2U/4U case, mount pandas inside + ethernet switch + storage and put in rack somewhere [12:26] you cant easily cluster package builds [12:26] hrw, right, thats what we plan ... still [12:26] ogra, Depends how you do it, there are ways. [12:26] persia, no supported ones that ubuntu has [12:26] ogra_ac: "make a cluster" as not an exactly cluster - just a bunch of hw in common case [12:26] in case you refer to beowulf like technology [12:27] hrw, right, one package will still build on one panda [12:27] ogra_ac: and one panda will be faster then one babbage [12:28] sure [12:28] and 32 pandas will be faster then 8 babbages (or how many of them are in use now) [12:28] but not as fast as an intel multicodre server [12:28] not for asingle package [12:28] ogra_ac: blame TI, STE etc for not giving you 4GHz 8core armv7a boards [12:29] its still 2.5h vs 15min or so [12:29] but building 200 packages will take 8h instead of 8 days? [12:29] i'm not talking about 200 packages [12:29] only about one of them [12:29] Isn't panda multicore? [12:30] dual, yep [12:30] Aren't some of the other buildds only dual? [12:30] ogra_ac: so s/pandaboard/arm realview express/ [12:30] realview express has 4core A9 on board [12:30] hrw, get me some into the DC [12:30] afaik thats something only linaro gets :P [12:30] costs more then 10 pandas anyway [12:30] ogra, But you'd probably make the kernel team happier by being able to drop -versatile than by dropping -omap, really. At least -omap has real users. [12:31] persia, i want to have both dropped [12:31] I think that ubuntu/arm may drop versatile - it is only for qemu anyway [12:31] persia, and with rsalveti's qemu spec we wont need versatile anyway [12:31] hrw, "only" is broad ;) === JamieBen1ett is now known as JamieBennett [12:32] its used by many tools we provide [12:32] but there is a replacement [12:32] which we'd like in the archive for natty [12:32] ogra, Anyway, I don't think it needs a separate spec. There's a couple specs in that area already. [12:32] k [12:32] ogra_ac: iirc someone in Linaro works on qemu to make it emulate beagleboard, there is spec for qemu-a9 [12:33] * ogra_ac just remembers several UDSes we had before where the title of the spec didnt match our expectations at all [12:33] hrw, There's a number of folk who rely on qemu working for armel for their regular work: we can't make that go away. [12:33] Sure. Spec titles almost never match content. [12:33] hrw, yeah, we have one for packaging (or pulling in the existing qemu-omap package) it [12:34] But it's not worth a spec for a few people to argue one way or the other, nor for a 5 minute session where everyone agrees. There are other places that talk about what comes out of "linux", and adding new source is trivial enough to not need a spec. [12:35] maintaining a complete branch is [12:35] which is one of the options [12:36] How is "maintaining a complete branch" different from "adding a new source package" except finding someone to do it? [12:36] I presume that if you want it that much, you already have a way to find someone to do it. [12:37] If not, I recommend finding someone to do it before pushing for it to exist [12:37] (note that using the linaro tree is another way, because then you're expecting linaro to do it) [12:39] ogra: hi! [12:40] ogra: i am just testing the audio stuff now... and i am confused. is that supposed to work? it looks like i am missing some files after clean 10.10 install, + apt-get upgrade (for -updates) and apt-get upgrade for (-proposed), with no reboot in between [12:48] ndec, There's another update to alsa-utils in the queue. [12:49] persia: argh... i missed this info... is the .deb available anywhere? [12:51] I don't think so: I think the source is still awaiting approval. [12:54] ndec, and after that the pulse profile needs to be added additionally, buut i first want proof that the alsa fix works [12:55] (pulse profile is needed to expose hdmi additionally, without it, alnalog should fully work though) [13:06] ogra_ac, sorry i am struggling to understand the question [13:06] (lacking context i suspect) [13:07] apw, i would like to not have omap3 built from the main tree one way or the other in natty [13:07] ogra_ac, omap3 is already not build from the master branch in natty :) [13:07] apw, and i would like to discuss the possibilites at UDS [13:07] oh ! [13:07] that saves all probs then, you already took action [13:07] See. No spec required. [13:07] and interestingly we are discussing that at the 'version and flavours' session [13:08] great, i'll attend [13:08] apw, Just for fair warning, I'm planning to raise ppc at that session. [13:08] ie, where are we gettign omap3 from for natty, are we using linaro, basing on linaro, doing our own thing [13:08] persia, what might you raise on it, as yes thats also on my radar [13:08] and its best to not be blind sided :) [13:08] if linaro provides all functionallity we have in maverick i'd be happy with that [13:09] apw, Mostly just to make sure it's discussed. I'm not going to foreguess conclusions. [13:09] also with taking their binaires directly so kernel team has it off their sholders [13:09] apw, That it's on your radar means you are unlikely to be blindsided. [13:09] ogra_ac, sounds liek we are thinking the same [13:09] great ;) [13:10] persia, well i am worried about powerpc yes, but i am interested in what you are going to ask [13:10] apw: link to spec can you? [13:10] * persia goes to -kernel to be even close to topic [13:12] hrw, its a skeleton at the moment, but it'll be this one: https://blueprints.edge.launchpad.net/ubuntu/+spec/hardware-kernel-n-version-and-flavours [13:13] subscribed [13:13] apw: catch Linaro kernel people to attend [13:13] morning [13:27] ogra_ac: nice, I also believe we can go with linaro's kernel for omap3 [13:27] :) [13:27] at least I believe it's going to be well supported by the next cycle [13:58] Hello. Where can I find the logs from #ubuntu-arm? [13:58] irclogs.ubuntu.com [14:21] GrueMaster, you have a2 or a3? === JaMa|Wrk is now known as JaMa|Off [14:27] ogra, does https://blueprints.edge.launchpad.net/ubuntu/+spec/hardware-arm-n-omap-edid-autodetection need a UDS session ? [14:28] ogra, NCommander, you have a few other-arm-natty- blueprints which seem to be about h/w detection [14:28] would they belong on the hardware- track ? [14:28] which ones ? [14:28] apw: this, probably not [14:28] the edid stuff is rsalveti [14:29] https://blueprints.edge.launchpad.net/ubuntu/+spec/other-arm-natty-improved-subarch-detection [14:29] thats for the installer and dpkg iits [14:29] (there are others similar i seem to remember) [14:29] *iirc [14:29] * ogra_ac doesnt remember having any spec related to hw detection [14:29] there's a discussion about integrating the same stuff with omap 4 and omap 3, but don't know if it actually needs a session [14:30] ogra_ac, well does that stop it being h/w i wonder? [14:30] apw, its about implementation in userspace [14:30] rsalveti, well i am in your hands, i ask because its an old blueprint being continued, if you need a session say yes and i'll accept it :) [14:31] apw, its a blueprint nobody worked on last cycle [14:31] ogra_ac, but by that approach we'd have nothing on the cloud track cause we do all our work on our laptops ? no ? [14:31] well, rsalveti invested some days but to late [14:31] ogra_ac, no complaints, just checking if it needs one ... the slots are filling and if it should be in the mix i need to tick and if not i need to cross :) [14:31] apw, yeah, doesnt that show the sillyness of the new approach :P [14:32] ogra_ac, well indeed. and why i wondered if it was a h/w related issue as its about detecting and handling specific hw [14:32] other-arm-n-arm-thin-clients and other-arm-n-public-panda-ppa-build-cluster are my specs [14:32] (the two in other) [14:32] i don't really mind, if its hardware- then i can accept it ... otherwise its someone elses problem :) [14:32] apw: I'm ok for not having a session for it [14:32] rsalveti, if you are sure ... i'll bosh it from the pending queue [14:32] apw: cool, thanks [14:32] apw, given how hard it is to rename it ... [14:33] actually rename is pretty easy, but annoying for everyone who has the link [14:33] the ppa cluster is more an IS thing and an overall solution [14:33] yeah i concur the other two sound like something other than hardware- [14:33] only a very small part is HW [14:34] and the thin clients would be more for a server (but surely in no way a could) issue [14:36] apw, feel free to rename the subarch thing ... [14:36] NCommander is on vacation atm so he wont change it [14:38] ogra_ac, i am more advising, i have no investment one way or the other on which track its on [14:38] hmm, k [14:38] to my eye its more h/w than other, but it matters little to the overall goals [14:39] do you know who is scheduling other ? [14:39] jcastro i belive [14:40] k [14:41] https://blueprints.edge.launchpad.net/ubuntu/+spec/other-arm-n-userland-subarch-detection [14:42] ogra_ac, that one should probabally be disponsitioned the same way as the previous one [14:42] if the other moves it should to, or not together [14:42] hmm [14:42] ogra_ac: rsalveti: anyone tried the kubuntu netbook edition? just read about this yesterday. i don't even know if it's using efl or ogl/ogles [14:43] ndec, its using a lot of GL stuff if available and no, we dont test it, ask the community [14:43] ogra_ac: ok! [14:44] ogra_ac: by the way, i just realized that I don't have the icon on the desktop. I rebooted with desktop, not netbook, and I can't find the icon. is that normal? [14:44] ndec, they are desparately looking for someone to sponsore them omap4 HW btw ;) [14:44] ogra_ac: argh... it's in the menu [14:44] heh [14:44] ogra_ac: who is 'they'? [14:44] kubuntu community [14:45] ah. ok [14:46] well, they will have to write their UI to use gles first ;-) [14:46] hard to do if you dont have gles HW ;) [14:46] good thing is that qt can use the eges backend [14:46] and this is tested by nokia in general [14:46] and meego [14:46] yep [14:47] so kubuntu *might* work [14:47] i know the images we have dont though [14:47] theyy went untested for nearly the whole cycle [15:01] ogra_ac: i'd like to rebuild alsa-utils latest patch to test it. if I get the source for x.y.z, and the deb diff, I just need to apply the diff and rebuild, right? [15:02] yep [15:02] ubuntu3.2 should have the complete fix [15:03] ogra_ac: you mean ubuntu3.3, right? [15:03] err, indeed [15:04] it hasnt been accepted yet, i might have to change it again ... though that shouldnt change functionallity [15:15] <_dash__> hi all [15:15] <_dash__> I am trying to use ubuntu on pandaboard and require some help [15:15] sure ;) [15:16] <_dash__> how to get source of ubuntu for pandaboard ? [15:16] i told you in the #pandaboard channel already [15:16] <_dash__> I want some git repo or tarball ... is it available ? [15:16] you can only build it package by package [15:16] no [15:16] it's not like OE [15:17] thats not how a packaging system based distro works [15:17] <_dash__> then how to get source ? [15:17] you can grab the source for all the packages [15:17] you pull it from ports.ubuntu.com [15:17] <_dash__> of current stable release ? [15:17] yes [15:19] <_dash__> u mean /dists/maverick/main/source/Sources.bz2 ? [15:20] no [15:21] there are 17000 packages under pool/ [15:22] honestly, what do you plan to achieve [15:22] <_dash__> ok I see main, multiverse, restricted, universe ... [15:23] <_dash__> I just want to get sources of ubuntu, cross compile it and run on panda board [15:23] <_dash__> I don't want pre-builts [15:23] but why ? WHATS WRONG WITH TEH BINARIES WE PROVIDE [15:23] OOPS [15:23] sorry for the caps [15:23] you're probably not going to be able to cross-compile everything [15:23] right [15:23] and yes, why do you need to build everything? [15:24] add neon? hf? [15:24] _dash__, So, there's something like 12G of source, and it takes something like a few weeks to compile it, given a farm of fast machines. [15:24] <_dash__> I may require to plug-in plug out or port some of the drivers [15:25] so you want to rebuild kernel bits [15:25] <_dash__> Basiccaly i want bsp port (u-boot, kernel atleast) [15:25] or compile additional modules [15:25] <_dash__> yes [15:25] http://cdimage.ubuntu.com/daily/current/source/ has a good chunk of it (the sources needed for the Ubuntu Desktop flavour), if you want a smaller bite. [15:25] then you don't need to build everything by hand [15:25] <_dash__> ogra_ac: yes [15:25] _dash__, no need to recompile the rest of the distro for that [15:25] you just need to hack some packages [15:25] right [15:25] like the ones you want to change [15:25] <_dash__> yes [15:26] <_dash__> So how do i get BSP port of it ? [15:26] there is no BSP [15:26] apt-get source on the panda can get you the sources [15:26] <_dash__> ok u-boot and kernel sources ? [15:26] i.e. apt-get source linux-ti-omap4 will get you the kernel package source [15:27] u-boot-linaro-omap4 is the u-boot source iirc [15:27] x-loader-omap4 is the MLO source [15:27] <_dash__> How do i get it to host machine ? [15:27] <_dash__> so that I can cross compile kerel and U-boot port [15:28] _dash__: you can also use your own panda to build your packages if you need [15:28] scp ? [15:28] right, the panda is really fast enough to build everything on it [15:28] <_dash__> And does ubuntu recomends any specific toolchain ? [15:28] _dash__: http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-maverick.git;a=shortlog;h=refs/heads/ti-omap4 [15:28] _dash__: we have our own cross toolchain [15:28] yes, you should use the ubuntu toolchain [15:28] packaged by hrw [15:29] <_dash__> how do i get that toolchain ? [15:29] apt-get install build-essential on the panda [15:29] or for cross builds ask hrw [15:29] apt-get install gcc-arm-linux-gnueabi for cross [15:29] but I believe the packages are only available for maverick [15:30] yeah [15:30] <_dash__> ok so this will give me tooclchain [15:35] <_dash__> toolchain problem is solved. [15:35] <_dash__> how to get kernel and u-boot and xloader sources ? [15:36] * ogra_ac told you above [15:36] <_dash__> I got link for kernel ... but which hash ? [15:36] rsalveti, hi, if you are there right now, let me ask.. did a kernel ever exist on maverick for fsl-imx51 or? [15:36] _dash__: head? [15:36] was it removed almost instantly? [15:36] we would love to see whatever development it got to before it stopped, post-Lucid [15:36] Neko: Lucid was it [15:36] Neko: amitk is your guy [15:37] Neko: the same branch was carried forward in the maverick tree, but not kernel version changes, etc. [15:37] <_dash__> rsalveti: is it available on git ? [15:37] _dash__: what exactly? [15:37] _dash__, you will trash the ubuntu install, you should use packages [15:37] (and rebuild them) [15:37] amitk, same question [15:38] Neko, no changes post lucid [15:38] Neko: already answered [15:38] <_dash__> rsalveti: Kernel source for cross compilatin ?? [15:38] _dash__, use the kernel package and xdeb from the cross toolchain to build a deb [15:39] <_dash__> 6931849710b21c0688a24646b15ff9db782b21d9 this hash ? [15:39] _dash__: the git kernel source is the same thing as the source package [15:39] _dash__: http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-maverick.git;a=shortlog;h=refs/heads/ti-omap4 (if you break anything, you get to keep both pieces) :0 [15:39] :) [15:40] _dash__: hrw posted on his blog how to cross build the kernel today [15:40] <_dash__> Please share blog link [15:40] * hrw ends for today [15:40] hrw, how about updating the channel topic with that ? [15:40] http://marcin.juszkiewicz.com.pl/2010/10/19/how-to-cross-compile-arm-kernel-under-ubuntu-10-10/ [15:40] ogra_ac: feel free [15:40] bah, way to long [15:41] that will likely trash it [15:41] * ogra_ac tires === ogra_ac changed the topic of #ubuntu-arm to: Ubuntu ARM Discussion & Development | https://wiki.ubuntu.com/ARM | Want to Submit a Bug? https://bugs.launchpad.net/ubuntu/+filebug | Build a rootfs from scratch: https://wiki.ubuntu.com/ARM/RootfsFromScratch | wanna cross build ? http://marcin.juszkiewicz.com.pl/2010/10/19/how-to-cross-compile-arm-kernel-under-ubuntu-10-10/ | Maverick is out ! http://cdimage.ubuntu.com/ubuntu-netbook/ [15:41] ogra_ac: http://42.pl/u/2u8U [15:41] this one is shorter [15:41] nah, seems it worked ;) === hrw is now known as hrw|gone [15:41] ogra_ac: please, get the shorter one === ogra_ac changed the topic of #ubuntu-arm to: Ubuntu ARM Discussion & Development | https://wiki.ubuntu.com/ARM | Want to Submit a Bug? https://bugs.launchpad.net/ubuntu/+filebug | Build a rootfs from scratch: https://wiki.ubuntu.com/ARM/RootfsFromScratch | wanna cross build ? http://42.pl/u/2u8U | Maverick is out ! http://cdimage.ubuntu.com/ubuntu-netbook/ [15:42] done :) [15:42] thanks a lot :-) [15:42] <_dash__> Ok kernel problem solved [15:44] just make sure you build a proper .deb and install it the right way [15:44] else you will trash the setup [15:45] <_dash__> yes thanks [15:46] launchpad should have a tiny url service :D === Tscheesy_ is now known as Tscheesy [16:19] Neko: https://bugs.launchpad.net/ubuntu/+filebug :-) [16:21] :D [16:28] question: I remember for Jaunty/Karmic etc. there was a console font that was used from the installer that replaced the horrible VGA one, and it was LIKE Terminus, but it was a little more curvy.. [16:28] does anyone know what on earth that font was? [16:34] also: is prelink safe on armel? [16:36] daurnimator: I have a prerelease A board, but it should be nearly identical. I don't remember any mention of changes that would affect video. Looking at the SRM, it looks like there were changes to power OVP on A1 and Camera and Memeory changes for A2. [16:45] GrueMaster, apparantly on the a3 there is now a pin to turn on the hdmi out [16:45] which isn't in the ubuntu kernel [16:45] Interesting. I'll relook at the documentation. [16:45] btw, anyone have a link to the .deb to offload some things to the dsp? [16:46] i remember someone linking it recently [16:48] The really odd part of this change is that it would make the XM incompatible with the Beagleboard I would think, which is not the intention. [16:50] GrueMaster, it shoudn't... there's already an is_omap36xx check that it's buried under.. and the change doesn't afffect the A2 as it isn't connected.. [16:51] hm, yeah, was going to look at this a3 issue later today [16:51] rcn-ee_: so, do we need an x-loader/kernel update for a3? [16:51] rcn-ee_: Have you been following daurnimator's problem? He has an A2 and is not getting video out. I have an A rev, and mine works fine here. [16:51] rsalveti, ping mpoirier we talked about it this morning and he has my patch. ;) [16:51] rcn-ee_: cool :-) [16:52] wait, what? There is a change to the uboot/xloader for Beagle, a week after release? [16:52] Yeap GrueMaster talked with daurnimator on #beagle this morning, he tested out the patch i was workign on that enables dvi on the A3... kernel's are a building and then images will get rebuilt for lucid/maverick.. [16:52] x-loader change not required.. [16:53] cool, then it's just the kernel [16:53] too bad they changed in a way we can't boot the pre-released image on a3 anymore =\ [16:53] cool *moves to here again* [16:53] ok. Had me worried. But still w/o patch, system won't boot. We'll need to come up with some documentation on how to update kernel, which will need to be done before first boot. [16:54] rsalveti, sadly there was a note in the A2 schematic that i never bothered to read.. ;) [16:54] GrueMaster: yup, like we do for blaze [16:55] on the A2, the pin was not connected to make it work, on the a3 it got connected to a new unused pin.. ;) [16:55] will create the same uImage but with the patch included [16:55] then the user can replace and install the system [16:55] rsalveti: Blaze is slightly different in that it isn't widely sold. [16:55] ogra_ac: ping [16:55] later on he can update the kernel and be fine with it [16:55] rcn-ee_: yeah [16:56] rcn-ee_: When will the XM A3 SRM be posted to beagleboard.org? [16:56] what sort of speed is expected in the end (in terms of general use) [16:57] it should be up today, i'd pin jkridner as he usually does that.. [16:57] daurnimator: The speed is decently usable. Not as good as a dual core panda, but good for most work. [16:58] GrueMaster: I just uploaded the files to S3.... [16:58] creating the links shortly. [16:58] daurnimator: Don't expect a lot of speed when doing massive compiles like openoffice. [16:58] daurnimator, if your not using the mpurate variable you can get more speed by bumping that to 800Mhz (1Ghz will be avialbe in the next kernel) [16:58] jkridner|work: Excellent. [16:59] GrueMaster, I've got it all booted now :P [16:59] Excellent! [16:59] bit slower than expected, especially dpkg stuff [16:59] cool! [16:59] yeah, blame sd card performance [16:59] hence wondering if theres anything that can be dsp-offloaded [16:59] or sd in general [17:00] jkridner|work: Is this change at the board trace level or can it be made in the field? [17:00] or very slow if using btrfs with default maverick dpkg.. [17:00] i'm using an 8gb sdhc card. should be very fast. otherwise all defaults [17:00] in any case, I really should be in bed; asleep. [17:01] jkridner|work: rcn-ee_: when the a3 was released to public? [17:01] first customer got them friday... [17:01] daurnimator: With SDHC, the higher the class of SD, the faster it is. [17:01] hm, ok, so no that long ago [17:01] so everyone who got a shipment notice this last week, got them.. [17:02] ooh really? thats when I got mine >.< I'm quite suprised I'm in the first lot... [17:02] rsalveti: ogra_ac: Should we update the release notes with instructions on how to update the image prior to first boot? If not, we could see a lot of bug reports. [17:03] GrueMaster, or just people like me on irc, getting utterly lost :P [17:03] GrueMaster: sure [17:03] will create the uImage and request someone to test, like daurnimator :-) [17:03] i put a notice on elinux ubuntu page, but https://wiki.ubuntu.com/ARM/OMAPMaverickInstall needs a warning.. ;) [17:03] then we can update it [17:03] rcn-ee_: yeah, will update with the proper uImage link there once I build it [17:03] notes for newbies: use a power adaptor; use a serial cable, and make sure it works. [17:04] yeah, I'm happy to test anything you guys care to throw my way [17:04] cool [17:04] * daurnimator departs [17:04] BeagleBoard-xM Rev A3 files are uploaded to http://beagleboard.org/hardware/design [17:05] rsalveti, btw patch is here: http://bazaar.launchpad.net/~beagleboard-kernel/%2Bjunk/2.6-stable/annotate/head:/patches/beagle/0001-omap-beagle-use-GPIO2-on-the-xM-A3-to-turn-DVI-on.patch it won't apply cleanly to ubuntu kernel, but it should be in that omap36xx if.. [17:06] rcn-ee_: yeah, cool, thanks for the link [17:53] GrueMaster, ?? [17:53] ogra_ac: I tried to build alsa-utils with the patch from the upload queue... still the file omap4 file isn't there: http://paste.ubuntu.com/516366/ [17:54] ndec, you patched the makefiles ? [17:55] ogra_ac: What are you asking?? [17:55] GrueMaster, upgarding images prior to first boot ? [17:55] ECONTEXT [17:55] See above. TI modified BeagleXM to require a GPIO to enable video out. [17:56] ogra_ac: with xm A3 [17:56] All shipping XM boards have this mod. [17:56] fun [17:56] yeah [17:56] ogra_ac: I downloaded the .diff from the upload queue, and applied [17:56] and we have a fix ? [17:56] yup [17:56] ndec, check if the makefiles in the init dir get patched to install oamp4 [17:56] I'm just building a new uImage and mpoirier will probably forward this to the kernel team [17:56] *omap [17:57] We'll need to test against a non-modified board to make sure it doesn't break C4 & XM A-A2. [17:57] rsalveti, GrueMaster so once we have the file, we indeed should provide info [17:57] yup [17:57] A3 shipped as of last Friday. [17:57] thats really bad timing [17:57] yeah [17:57] given we will all be travelling [17:57] for default there's no console, and the screen doesn't work [17:58] not good :-) [17:58] ogra_ac: grep omap alsactl/init/Makefile --> empty [17:58] kinda.. there was a note in the A2 schematic that i and other didn't notice.. ;) [17:58] rcn-ee_: A2 or A3? [17:58] ndec, thats weird, my patch definitely patches Makefile.in and .am [17:58] the A2 schematic had a note that it would be changing.. ;) [17:59] Ah. [17:59] rcn-ee_: I just didn't expect a3 so soon [18:02] glad that at least we don't need to change x-loader/u-boot for it [18:04] on the plus side for power managment users, invert that pin and you disable the tfp410 chip to save some juice on the a3.. [18:05] ogra_ac: it's weird, when I run dpkg-buildpackage, it does not apply the patch serie [18:08] rcn-ee_: nice [18:09] ogra_ac: quilt.make is commented out in debian/rules. [18:09] ogra_ac: i need to go home now. let's talk about this tomorrow. [18:14] hello, i am currently @ beagleboard under maverick. pls help me to install powervr-omap3 [18:15] Sonny_8bit: https://wiki.ubuntu.com/ARM/OMAP [18:15] thanks [18:16] after installing powered usb-hub mouse is working fine [18:16] cool [18:17] also, make sure your user is also at the video group [18:17] tested this morning and it seems the user is not at the video group by default [18:17] and this is required to access the pvr device [18:19] ok, i am starting... [18:22] guys where are the libegl1-sgx-omap4 etc. packages kept? [18:22] system says sonny@sonny-beagleboard:/$ sudo apt-get install libegl1-sgx-omap3 libgles1-sgx-omap3 libgles2-sgx-omap3 Чтение списков пакетов... Готово Построение дерева зависимостей Чтение информации о состоянии... Готово E: Не удалось найти пакет libegl1-sgx-omap3 E: Не удалось найти пакет libgles1-s [18:22] sorry i use russian localisation [18:23] great, my UTF-8 works :P [18:23] Neko, in the omap4 PPA [18:23] it cannot find packages [18:23] Sonny_8bit, do you have multiverse enabled ? [18:23] URL? [18:23] and universe too [18:23] launchpad URL I mean [18:24] I can't search for it it says no results [18:24] Neko, just click on the TI icon on your panda [18:24] oh, wait :P [18:24] I don't have a panda [18:24] hmm i dont know exactly 'cause i am new in linux [18:24] get one ;) [18:24] I have a beagle, but I'm not doing this on that :D [18:24] the idea is we match the mx51 gl libs to the same kind of installation process [18:24] you wouldnt get the icon on a beagle ;) [18:25] https://edge.launchpad.net/~tiomap-dev/+archive/release [18:25] Neko, ^^^ [18:25] next time search LP yourself [18:26] I did! [18:26] it said no results [18:26] nothing.. [18:26] tiomap-dev is what I should have searched for it seems :D [18:26] Sonny_8bit: you can activate at the package manager or by editing /etc/apt/sources.list [18:27] you'll see that the multiverse and universe repos are commented [18:27] searching for omap4 gets me 350 results [18:27] ok - BB is too slow i am waiting for package manger task switch :) [18:27] ogra_ac: the default user is not part of the video group, the one used to access the pvr driver [18:27] ogra_ac: in this case how we fix that? [18:28] rsalveti, well, the video group isnt used by the system, not sure [18:28] in fact none of the groups are anymore [18:28] apart from adm and admin probably [18:28] yup, true [18:29] i wonder how fglrx and nvidia do it [18:29] ubuntu adm dialout cdrom plugdev lpadmin admin sambashare [18:29] default groups [18:29] am i nee to run synaptic package manager or ubuntu software center? [18:29] *need [18:29] Sonny_8bit, either will work [18:30] Sonny_8bit: see http://www.omappedia.org/wiki/Ubuntu_PPA [18:30] Sonny_8bit, in software center the sources entries can be found under edit [18:30] the Enable Universe Repositories on your PandaBoard [18:30] rsalveti, thats for omap4 though [18:30] should be similar, but you should enable universe and multiverse [18:30] yeah [18:30] ogra_ac: I know, just to display the picture and how it should look like [18:30] yep [18:31] hrm ... so much about USB audio works ... [18:31] * ogra_ac just bought a USB soundblaster stick [18:31] ogra_ac: still not fixed? [18:31] oh, sorry [18:31] was thinking about panda audio :-) [18:31] rsalveti, no, thats for my ac100 [18:32] i bought a lot of USB crap today ;) [18:32] hahah [18:32] 16G usb keys that completely vanish in the socket [18:32] pretty cool things [18:32] and a 1.8" HDD [18:32] nice [18:33] yeah, really great for traveling [18:37] ogra, show show show.. I would love a USB key like that [18:53] tsk [18:55] so my new shiny 1.8" disk comes preformatted ... [18:56] with NTFS [18:56] * ogra_ac shakes his head [19:00] Cool. Should work great on your Windows 7 system. :P [19:08] Hello, I'm new to linux arm, but not so new to linux (using and tweaking a debian lenny from command line only for 2 years) [19:08] I try to run ubuntu on a HTC Desire android smartphone, using http://nexusonehacks.net/nexus-one-hacks/how-to-install-ubuntu-on-your-nexus-oneandroid/ [19:09] I installed an Ubuntu ARM img so far, but I can tell things are broken in the OS : dpkg is broken for example : [19:09] "failed to exec dpkg-deb to extract control information: No such file or directory" [19:10] /usr/bin/apt-get doesn't exist neither [19:10] aptitude reinstall dpkg is not possible, since it itself uses dpkg and give me the same error than above [19:10] So here is my question : how can I reinstall an Ubuntu ARM image ? [19:11] The current image is tweaker for the phone, so I'd like to overwrite/reinstall things, without clearing the vhole install [19:11] First, is your phone cpu arm v7? [19:11] GrueMaster, Yes [19:12] Jere is what I get from inside the ARM image : [19:12] root@localhost:/# uname -a [19:12] Linux localhost 2.6.32.15-gf9c0527 #1 PREEMPT Fri Jul 23 17:26:53 CST 2010 armv7l GNU/Linux [19:12] root@localhost:/# cat /etc/lsb-release [19:12] DISTRIB_ID=Ubuntu [19:12] DISTRIB_RELEASE=9.10 [19:12] DISTRIB_CODENAME=karmic [19:12] DISTRIB_DESCRIPTION="Ubuntu 9.10" [19:12] root@localhost:/# [19:12] As to Ubuntu Images, they are created for specific platforms (Beagleboard, BeagleXM, Panda, etc). [19:12] Karmic was ArmV6+VFP iirc. [19:13] And how is packages management performed ? Since each image has a specific arch that is bound to a phone ? [19:14] (if you've links, I 'm asking you the man here :s) [19:15] I mean : atm, I use a lenny, that has repositories and packages for i386 or amd64, but nothing for armForSpecificPhone [19:16] They are just armel under Ubuntu. Jaunty was Armv5, Karmic armV6+vfp, lucid+ is armv7. [19:17] I'm assuming that somewhere down the road (around Ubuntu-S), we will be supporting Armv8 or some such. [19:22] GrueMaster, Ok :) Thanks for info. So if dkpg, and so far base system, is broken, I need to shift from the "place ubuntu.zip on your SD card" black box method to https://wiki.ubuntu.com/ARM/RootfsFromScratch [19:22] Through android asb shell, I should be able to run this ? [19:22] That would probably be the best approach. [19:23] (adb) [19:23] Ok, thanks :) [19:23] I don't know about the android part. I only test Ubuntu on supported hardware. === JaMa|Off is now known as JaMa [20:40] Bye [20:43] rcn-ee_: at http://bazaar.launchpad.net/~beagleboard-kernel/%2Bjunk/2.6-stable/annotate/head:/patches/beagle/0001-omap-beagle-a2-3-fix-dvi-reset-gpio.patch do we actually need to set up GPIO 129? [20:44] on A2 it goes to DVI_UP, but pin 4 is disabled (from the schematics) [20:44] and on A3 this gpio is nc [20:45] I know that setting up gpio 170 for xM is wrong, but not sure if we actually need to set up gpio 129 [20:45] rsalveti, that first gpio should be done, but it for maverick it's not needed.. (can't think of a good sru reason, since it works..) [20:45] should make no difference, just to see if I should apply the entire patch [20:46] rcn-ee_: but from the schematics, does it make any sense to set up gpio 129 while pin 4 lifted to disable [20:46] on page 9, from a2 schematics [20:47] loading.. ;) mouse died.. [20:48] on maverick we're already setting beagle_dvi_device.reset_gpio = 129; [20:48] at beagle_display_ini [20:50] then you'r safe.. i diff'ed between angstrom (which wasn't having an issue) that was one big thing i was missing in my tree but it doesn't affect the dvi issue people where having.. (i need to rename that patch..) [20:51] rcn-ee_: I know, was just reviewing other changes [20:51] to make sure we're fine with all the needed changes [20:51] okay... yeap as long as you set it somewhere else it should be fine.. [20:52] gpio 129 pin.. [20:53] rcn-ee_: but if pin 4 is lifted to disable, does it make any difference? [20:53] just curious, can be wrong because I'm not a hardware engineer [20:54] for the A2? with pin 4 lifted, that pin should go high on TFP410 which sets it in non power save mode... [20:55] rcn-ee_: this is the patch we're using: http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-maverick.git;a=commitdiff;h=8a1f3f2ed8b35c03606b90eecccda612aa48d32a;hp=a9296f7cc96073c74d06b2ea5e300e620f856177 [20:56] rcn-ee_: yeah, but my question is, does it make any difference to set gpio_129 when the pin 4 is lifted and the pin is high anyway [21:01] anyway, good that we're not setting gpio 170 at xM [21:01] ah sorry i see it.. yeah it looks like the 129 gpio is not needed, since it's not hocked up on the A2 and the A3 uses a differnet pin.. [21:01] that's why it's worked on my board without that patch.. [21:01] rcn-ee_: yeah [21:02] Back, I'm running a debian lenny am64, and plan to create a ubuntu maverick rootstock. Debian lenny's bpo debootstrap doesn't suport ubuntu maverick (no script), so I plan to run a maverick live cd (if it exists) and create the rootstock from here. [21:02] Does it sounds possible ? [21:02] rcn-ee_: so the only needed change is to avoid using gpio 170 at xM while setting up dvi [21:02] pwork, lenny's qemu might be too old.. [21:02] it very likely is [21:03] yeah avoid the gpio 170 on the xm and set GPIO2 high to enable the TFP410.. [21:03] using a livecd will be slow but work [21:03] because this will probably affect some other stuff [21:04] rcn-ee: Indeed, but I just realize that I have an unstable debian chroot installed on it. Maybe I can debootstrap from this unstable chroot, to create a ubuntu maverick chroot. It would be a chroot inside another chroot, don't know if it's reasonable [21:04] yeap, it might.. for maverick i'd only touch the xM and add the GPIO2 bit, tested on my xM A [21:04] yeah, if you're using qemu be sure that the version is newer enough [21:04] rcn-ee_: yup, porting the patch right now [21:04] rcn-ee, ogra : Thanks. Slow is not a problem. I'm myself quite slow too :p [21:05] heh [21:05] pwork, the unstable chroot should work, rootstock with lucid/maverick really stress's out qemu.. ;) [21:05] persia, hi... this "autodetecting hardware in X that is not properly detectable" discussion [21:05] markos will be there but he won't be able to see that discussion as we think arm stuff is more important [21:05] rcn-ee_: we also need to check upstream later, still pulling the linux-omap tree [21:05] but I would love to find input on braindead stuff like how xorg-video-vesa and vga and other completely non-arm video drivers are installed by default [21:05] Neko, eh should have subscribed properly [21:06] ogra "subscribed properly"? :D [21:06] Neko, the scheduling software takes attendance into account [21:06] they conflict [21:06] to the spec [21:06] right [21:06] Yeap, i was looking at tmlind's master, it's missing the xM check in that location, so we need to wait for that.. [21:06] the SW cares for them to not conflict if you are subscribed to both [21:06] we had to sub late, sorry [21:06] but, persia has an efikamx... :) [21:06] I'm surely missing some point, but I won't use qemu : this rootstock image is intended to run on an android phone, not on a PC emulator (I believe qemu is an ARm emulator for PC :s) [21:07] is he subbed to both specs ? [21:07] the schedule is still dynamically changing [21:07] pwork: yeah, but rootstock uses qemu [21:07] to generate the image [21:07] unless you're on arm already [21:07] Ah ok [21:08] rsalveti, : Tip: In my country, 'salveta' is a brand of water with bubbles. Could be of some interest for you [21:08] :D [21:08] pwork: haha, which country? [21:09] sounds italian [21:09] (guessing) [21:09] markos_ see ogra's comments above.... [21:09] salveti is an italian name [21:09] subscribe and they may not conflict? [21:09] rsalveti, France [21:09] Neko, yeah, with a bit of luck ... [21:10] Neko, if you want to make sure, you need to make your participation essential ... but that should be reserved for people actually being essential for the spec [21:14] it is more an informational thing === ian_brasil___ is now known as ian_brasil === ian_brasil_ is now known as ian_brasil [21:36] persia: ping [21:46] Neko: yeah, I'll subscribe to it as well [21:51] kmargar, http://summit.ubuntu.com/uds-n/track/performance/ Thursday, top [21:53] moment === kmargar is now known as markos_ [21:54] those go without saying : [21:54] :) === JaMa is now known as JaMa|Zzz [23:01] what a couple on uds :)