=== vddlogger is now known as parislogger | ||
=== mayday_jay is now known as mayday_jay-away | ||
=== mayday_jay-away is now known as mayday_jay | ||
=== mayday_jay is now known as mayday_jay-away | ||
=== mayday_jay-away is now known as mayday_jay | ||
=== mayday_jay is now known as mayday_jay-away | ||
=== mayday_jay-away is now known as mayday_jay | ||
=== michaelh1 is now known as michaelh1|away | ||
=== michaelh1|away is now known as michaelh1 | ||
soren | Will Ubuntu's ARM stuff work with a Pandaboard? Everyone seems to talk about Beagleboards, and I have no clue how similar they are. | 09:00 |
---|---|---|
hrw | it does | 09:04 |
soren | hrw: Excellent, thanks. They seem much beefier than Beagleboards and only a bit more expensive. Is there any particular reason by Beagleboards are still the center of attention (which is the impression I get)? | 09:06 |
hrw | beagleboard is easy to buy | 09:10 |
hrw | and if you want to make own hw then omap3 cpu is also easy to buy | 09:11 |
hrw | so companies are buying beagleboards to experiment and then develop own hw | 09:11 |
soren | Funny you should say that. The reason I ask about Pandaboards is exactly because those actually seem to be available, while Beagleboards seemed to be out-of-stock in most places. | 09:12 |
soren | ...but thanks for clarifying. Your other points make lots of sense. | 09:12 |
hrw | I worked for company which went that route ;) | 09:23 |
soren | Ok :) | 09:26 |
infinity | soren: Beagles aren't the center of attention by any means. | 09:42 |
infinity | soren: Pandas are out target proof-of-concept platform for most of what we do right now. | 09:43 |
infinity | soren: (And the omap4 images are built specifically with Pandaboards in mind) | 09:43 |
infinity | s/out target/our target/ | 09:43 |
soren | infinity: I must either not be paying attention or be paying attention to the wrong folks :) | 09:45 |
infinity | soren: To put that in perspective, my desk has Pandas, an i.MX53, an i.MX51, a Toshiba AC100, an LG-P999, but no Beagleboards. :P | 09:45 |
soren | Great. | 09:46 |
ogra_ | panda ftw ! | 09:46 |
* soren commences twiddling thumbs until his shiny new Pandaboard arrives | 09:46 | |
* ogra_ sees that soren finally made it to the bright side of HW :) | 09:47 | |
infinity | soren: We *do* try to make sure the Beagle images continue to function, because there are a TON of them in the community, but our *current* proof-of-concept platforms are the Panda, i.MX53, and the AC100, and the Panda is the only one we fully support on a corporate level as well as community. | 09:47 |
infinity | (Well, for some value of the word "support" that I would refuse to commit to in any capacity more official than an IRC channel) | 09:47 |
infinity | Being a dev board, the word "support" tend to always have "proof-of-concept" floating around it somewhere. ;) | 09:48 |
infinity | s/tend/tends/ | 09:48 |
* ogra_ prefers "reference" :) | 09:48 | |
ogra_ | (sounds less unfinished) | 09:49 |
infinity | ogra_: Well, it's sticky. It's our ref platform for desktop and "mobile", but it's a POC platform for server, since the Panda is pretty effin' obviously not server hardware. ;) | 09:49 |
infinity | It's really not desktop hardare either, to be honest. | 09:49 |
infinity | Mobile is all I'm comfy calling it. | 09:50 |
soren | ogra_: I did play around with an NSLU2 a couple of years ago, but it wore out my patience pretty quickly. I get the impression things have changed ever so slightly since then :) | 09:50 |
ogra_ | indeed | 09:50 |
infinity | Or "a giant cell phone". | 09:50 |
infinity | soren: The Panda's quite nice, modulo a couple of kernel bugs that we've finally got a handle on. | 09:50 |
ogra_ | nuslu2 makes a good NAS :) | 09:50 |
ogra_ | -s | 09:50 |
infinity | soren: And the fact that all your I/O is over USB, which can get frustrating. | 09:50 |
ogra_ | err | 09:50 |
ogra_ | -u indeed | 09:51 |
infinity | soren: But they're speedy, other than that. | 09:51 |
soren | infinity: Oh, no SD card or anything? | 09:51 |
hrw | infinity: anything omap based is mobile | 09:51 |
soren | infinity: (or is that connected over a USB bus?) | 09:51 |
ogra_ | soren, SD is the default boot media | 09:51 |
infinity | soren: The SD reader isn't on USB, but how that would make you happy, I don't know. :P | 09:51 |
infinity | soren: Given that SD isn't exactly "fast". | 09:51 |
ogra_ | and in our preinstalled images also the default rootfs media, but of you want to use the board in production you want USB root | 09:51 |
soren | infinity: Err... Right you are. | 09:52 |
infinity | soren: The ethernet, and any external storage you might attempt, will be over USB2, however. Which isn't AWEFUL, it's just not ideal. | 09:53 |
infinity | soren: Still, they're good little boards, for what they are. And the price point is nice for a dev kit. | 09:53 |
infinity | (As is the price point for an iMX.53, if you're in a collecting mood). | 09:53 |
ogra_ | they surely can cope with the low end atoms speed wise | 09:54 |
infinity | Oh, all the ARM kit on my desk blows my Atom netbook out of the water. | 09:54 |
ogra_ | sadly not IO wise | 09:54 |
infinity | I only still use the Atom because it has a sane SATA controller, and, like, a portably form factor. :P | 09:54 |
soren | The ethernet adaptor is hooked up to the USB bus? | 09:55 |
infinity | (The i.MX53 has sane SATA, the AC100 has a good form factor, if only I could combine the two...) | 09:55 |
infinity | soren: Yes. | 09:55 |
ogra_ | soren, well ... | 09:55 |
soren | Hm. Ok. | 09:55 |
ogra_ | the ethernet adapter *is* the host afaik | 09:55 |
ogra_ | its one chip | 09:55 |
=== michaelh1 is now known as michaelh1|away | ||
infinity | Not that it matters. USB2 is 480Mbps, the ethernet is only 100Mbps. :P | 09:56 |
ogra_ | indeed | 09:56 |
infinity | soren: Wait... I thought you were living in the US these days? | 09:57 |
ogra_ | infinity, i have the ac100 datasheet, you could write a PCIe driver and use SATA SSDs | 09:57 |
infinity | soren: What are you doing awake at this ungodly hour? | 09:57 |
soren | infinity: Good grief, no. | 09:58 |
infinity | soren: Oh, you're in .dk? | 09:58 |
soren | infinity: Yup. Not planning on going anywhere either :) | 09:58 |
* infinity wonders where he got the US idea... | 09:58 | |
soren | infinity: What are *you* doing up at this ungodly hour? Or did you move to a more sensible timezone? | 09:58 |
ogra_ | heh | 09:58 |
* ogra_ was about to ask the same | 09:59 | |
infinity | ogra_: That would be more tempting if I they had PC105 keyboard variants. | 09:59 |
infinity | soren: Long weekend, lots of gin, I dunno. DON'T JUDGE ME. | 09:59 |
infinity | I don't work tomorrow, I'm travelling on Tuesday, and at Plumbers on Wednesday. | 10:00 |
infinity | I'm not sure that sane sleep patterns matter between now and then. | 10:00 |
soren | You're in UTC-5? Or -6? | 10:00 |
infinity | -6 | 10:00 |
soren | Ooh, lunchtime. | 10:01 |
infinity | I don't often eat lunch at 4am. | 10:01 |
infinity | But sure. | 10:01 |
infinity | (Okay, who am I kidding, there's nothing I don't often do at 4am) | 10:01 |
* ogra_ considers breakfast ... | 10:01 | |
=== chrisccoulson_ is now known as chrisccoulson | ||
=== nslu2-log_ is now known as nslu2-log | ||
lilstevie | persia: ping me when you are about | 11:51 |
lilstevie | ogra_: so the ac100 has an unused mini pci-e? | 11:52 |
ogra_ | lilstevie, well, depends on the model, all of them have a PCIe slot, but we only have a driver for the USB part of ot atm | 12:21 |
ogra_ | all non 3G models dont have a header soldered to the socket though | 12:22 |
lilstevie | sounds like what we have on the transformer | 12:22 |
ogra_ | that makes using the slot a bit difficult on the non 3G ones | 12:22 |
lilstevie | the non 3G tf has the port with no header | 12:22 |
lilstevie | and well same deal with the usb part only having a driver | 12:23 |
lilstevie | trimslice has a driver for its pci-e | 12:23 |
lilstevie | wonder if that would help at all | 12:23 |
ogra_ | i doubt its a 1x1 replacement | 12:24 |
ogra_ | so even if they have a driver, that would likely only be a base to build on | 12:24 |
lilstevie | true | 12:24 |
ogra_ | i think there is also a driver in the old harmony kernels | 12:25 |
lilstevie | one good thing is they are all AP50 | 12:25 |
lilstevie | er | 12:25 |
lilstevie | AP20 | 12:25 |
sebjan | /join #linaro | 12:43 |
sebjan | oups | 12:43 |
lilstevie | ogra_: do you have a good test of egl for the ac100? | 12:47 |
ogra_ | not atm, i think janimo did some gles testing on ac100 | 12:48 |
lilstevie | ogra_: hm, what about that quake3 port? | 12:48 |
ogra_ | ask phh, i think he has something like that | 12:49 |
lilstevie | heh ok, | 12:49 |
phh | ok i need to find the url. | 12:49 |
ogra_ | not sure thats usable on a std ubuntu though, i think he adjusted it for the nvidia overlay stuff | 12:49 |
lilstevie | I have been running CrOS kernel and u-boot on the tf | 12:49 |
phh | ogra_: no | 12:49 |
ogra_ | ah, k | 12:50 |
phh | lilstevie: it's pandora's quake3 port | 12:50 |
phh | ogra_: well ok there is one change, i changed the hardcoded :0 to :1 | 12:50 |
lilstevie | which also is compatible with the L4T stuff | 12:50 |
ogra_ | heh, well, that should probably work on all tegras | 12:50 |
phh | lilstevie: http://kotelett.no/ac100/phh/Android2.1/Games/openarena.tar | 12:51 |
lilstevie | phh: ah ok, well I have that | 12:51 |
phh | sed -ie s/:1/:0/g on the binary | 12:51 |
lilstevie | awesome | 12:51 |
lilstevie | kk | 12:51 |
lilstevie | phh: did that libflashplayer.so come from an android flash apk? | 12:52 |
phh | lilstevie: no, it comes from Atrix' L4T | 12:53 |
phh | you can't use android's libflashplayer.so | 12:53 |
lilstevie | ah ok | 12:53 |
phh | well not as is | 12:53 |
lilstevie | yeah I found that out :p | 12:53 |
lilstevie | I attempted with flash 10.3 apk | 12:53 |
phh | and making it work on "standard" linux would be REALLY painful | 12:54 |
lilstevie | heh | 12:54 |
lilstevie | I'm also having a strange problem in oneiric, I don't know if it is my fault though | 12:57 |
lilstevie | :p | 12:57 |
lilstevie | default route is not being set when wifi connects | 12:57 |
ogra_ | works fine here, its usually the duty of NM | 12:58 |
lilstevie | hm | 12:59 |
ogra_ | might be that your driver does not expose the bits and pieces the right way for NM | 12:59 |
lilstevie | it works perfect in natty | 12:59 |
lilstevie | it is the brcmfmac staging driver | 12:59 |
ogra_ | well, might be an NM bug, but i definitely dont see it here | 12:59 |
ogra_ | so i would guess driver ... | 13:00 |
lilstevie | ok | 13:01 |
lilstevie | hmm ok | 13:03 |
lilstevie | so looking at the network manager logs, i have <warn> Failed to add route Invalid input data or parameter | 13:04 |
lilstevie | but that is for setting ipv6 stuff which I don't use cause I don't have | 13:04 |
lilstevie | ogra_: are builds still taking 70 minutes when you build locally on your ac100? | 13:27 |
ogra_ | kernel ? | 13:28 |
ogra_ | i havent built locally for quite some time, and before i always had a configured tree, so only changed files were rebuilt | 13:29 |
lilstevie | heh yeah | 13:29 |
lilstevie | building with debuild cleans the tree every time though :/ | 13:29 |
ogra_ | yeah, i dont do that for tests | 13:30 |
ogra_ | and for binary packages i just use the archive :) | 13:30 |
lilstevie | yeah I am building this not as a test though, trying to seed it into an image | 13:30 |
lilstevie | but also making sure that it builds | 13:31 |
lilstevie | so if I send it to persia that it doesn't get rejected again :p | 13:31 |
lilstevie | I'm not lucky enough to have an archive that I can upload it to :p | 13:32 |
ogra_ | just get more sponsored packages in ... and then apply for upload rights ;) | 13:33 |
lilstevie | heh, well I don't have one yet | 13:33 |
lilstevie | ogra_: though some massive stuff has happened for us with the transformer | 13:35 |
lilstevie | we got a u-boot port | 13:35 |
ogra_ | we have a u-boot port for the ac100 as well | 13:36 |
lilstevie | working well? | 13:36 |
ogra_ | its just not usable without any input device support | 13:36 |
lilstevie | ah yeah, similar space then | 13:36 |
ogra_ | seems to work if you have soldered a serial port on | 13:36 |
lilstevie | trying to figure out how to to get the stupud EC keyboard to work with it | 13:36 |
lilstevie | keyboard gives us some input, just it is random noise | 13:37 |
lilstevie | but I have interaction through boot.scr :p | 13:38 |
=== parislogger is now known as transitlogger | ||
plm | the ubuntu for omap will works in this board too? http://igloocommunity.org/ | 18:24 |
plm | of course is omap.. but my question is about the board.. or just pandaboard with omap4 is supported? | 18:24 |
infinity | The Snowball isn't OMAP. | 18:32 |
plm | infinity: ahh sorry.. is just cortex-A9 | 18:33 |
infinity | Of course, Ubuntu's userspace will run fine on Snowballs, we just don't ship a kernel or installer for them currently. | 18:33 |
plm | infinity: but.. they say that supported by ubuntu | 18:33 |
plm | infinity: but what ubuntu I get to install in snowball? | 18:34 |
infinity | See above. We don't have a Snowball installer. | 18:34 |
plm | infinity: because don't have a ubuntu for ARM. just for omap | 18:34 |
plm | humm | 18:34 |
infinity | All of our packaged software will work fine on them. Just, like I said, no kernel or installer. | 18:35 |
infinity | Linaro's hwpacks should work for Snowball. | 18:35 |
plm | infinity: yes.. option is linaro | 18:35 |
infinity | And you could, for instance, use linaro-image-create with an ubuntu-core tarball as the rootfs and a snowball hwpack, to create a Snowball image. | 18:35 |
plm | humm | 18:36 |
infinity | We're working on trying to make this whole "every board is different, so we can't have a unified installer" business a bit less messy going forward. | 18:37 |
infinity | But building 32 different images isn't the way to go. :P | 18:37 |
plm | infinity: maybe is better use entire linaro.. because has modified kernel and apps to works with board.. like, acell, GPIO, gps and so one | 18:38 |
infinity | It will, ultimately, be about leveraging linaro hwpack type stuff with Ubuntu rootfses, I imagine. | 18:38 |
infinity | Hrm? Other than the kernel, "linaro" is Ubuntu. Well, except that they rebuild binaries to test their toolchain. | 18:38 |
infinity | If you want your userspace to be supported sanely, Ubuntu userspace with a linaro kernel is the way to go. | 18:39 |
infinity | Which is what you get with a linaro hwpack. | 18:39 |
infinity | And an ubuntu tarball. | 18:39 |
plm | infinity: so I use just kernel of linaro and the rest is the same ubuntu used in ubuntu for omap? | 18:40 |
infinity | Yes... There is no "Ubuntu for OMAP". | 18:41 |
infinity | Don't confuse installers with the archive. | 18:41 |
plm | ok | 18:41 |
plm | so for arm :-) | 18:41 |
infinity | Our OMAP images are Ubuntu with an OMAP kernel and an installer that knows how to boot the system. | 18:41 |
plm | ubuntu for arm :-) | 18:41 |
plm | but the packages are ported/recompiled for arm arch.. is not the same use in cisc right | 18:42 |
infinity | The ubuntu/arm packages are just generic armv7, they work on anything. | 18:42 |
infinity | Well, anything v7 and up. | 18:42 |
plm | infinity: are there in "cookbook" or a tutorial for to do that.. linaro kernel + ubuntu? | 18:43 |
infinity | linaro-image-tools takes an argument for a rootfs tarball. You'd just need to point it at an ubuntu-core tarball, and that would be that. | 18:43 |
infinity | I'm sure there's l-i-t documentation somewhere, but other than manpages, I'm not sure where to point you online. | 18:44 |
plm | ok | 18:44 |
plm | infinity: thanks.. that is a good start point :-) | 18:44 |
=== michaelh1|away is now known as michaelh1 | ||
=== michaelh1 is now known as michaelh1|away | ||
=== michaelh1|away is now known as michaelh1 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!