**** BEGIN LOGGING AT Thu May 13 02:59:56 2010 May 13 06:37:15 Stskeeps: hey :) May 13 06:37:36 can somebody point me at uds channel for this topic for tomorrow? May 13 06:51:52 tmzt: they are actually about to get started, but are scattered in different rooms May 13 06:52:12 tmzt: see: http://summit.ubuntu.com/uds-m/track/ubuntu_on_arm/ May 13 06:52:47 tmzt: typically, the channels take the form of #ubuntu-uds-ROOMNAME, where the room name can be found in that link May 13 06:54:30 tmzt: so for example, the ASoC discussion will be in mahogony, so #ubuntu-uds-mahogony. However since it's live, you would just have 1 way communication on irc to ask questions. To know what's going on, you will want to listen to the icecast stream May 13 06:57:28 thanks May 13 06:58:14 what timezone in the panel clock? May 13 06:58:50 oh, got it May 13 06:58:55 it's later than I thought May 13 07:01:20 Is there an all-discussions stream? May 13 07:01:27 Or are there multiple things at the same time? May 13 07:01:50 And do they create transcripts? May 13 07:02:27 everything is in one room? May 13 07:02:50 there seem to be two talks (X/dpkg for ex) at the same time May 13 07:02:52 in one track May 13 07:03:12 oh, the room is under the title May 13 07:03:18 I greatly prefer transcripts. May 13 07:03:34 It takes me 5 minutes to read something that could take 45 minutes to listen to. May 13 07:08:49 http://www.lczajkowski.com/2010/05/04/uds-m-remote-participation/ May 13 07:09:06 oh, and I want glx_ext_texture_from_pixmap May 13 07:09:24 don't we all May 13 07:09:24 :P May 13 07:10:34 That's 99% of my Linux 3D stuff I use regularly: compiz. May 13 07:10:45 Or how about reviving that old XGL... on top of GL ES? May 13 07:10:50 hi May 13 07:11:16 doesn't gallium have some kind of es emulation now? May 13 07:11:22 What timezone is uds in, anyway? May 13 07:11:36 Oh, and fglrx claims to have some GL_ES strings.... wonder how you'd use that. May 13 07:11:39 DanaG: european one May 13 07:11:46 DanaG: 09:11 here May 13 07:11:50 Probably it's as xvba-video was: not public. May 13 07:12:23 s/was/is/ May 13 07:12:37 it's a state tracker May 13 07:18:05 anybody who can comment on why recent ubuntu releases are limited to armv6mmu and if that is changing with the new branching/repo strategy? May 13 07:18:48 tmzt: armv6? lucid is armv7a May 13 07:19:14 right, so more generally why is this required? May 13 07:19:30 how many packages of the full repo actually require neon or vfp May 13 07:19:45 tmzt: All of main was rebuilt for the v7t2 transition May 13 07:19:50 So all of them :-) May 13 07:20:11 http://sourceware.org/redboot/ May 13 07:20:14 er May 13 07:20:19 wrong link May 13 07:20:21 https://wiki.ubuntu.com/Specs/ARMSoftbootLoader May 13 07:20:21 tmzt: The point is to take advantage of v7 and thumb2 instructions May 13 07:20:26 I'm aware of the requirement, I'm asking for what the justification might have been May 13 07:20:34 That reminds me of the Angstrom "kexecboot" loader on Zaurus. May 13 07:20:38 since only certain packages should benefit from those May 13 07:20:44 tmzt: get as much as possible from cpu? May 13 07:20:52 tmzt: No, everything benefits from it May 13 07:20:56 * DanaG scrolls down and sees that mentioned. May 13 07:21:04 tmzt: thumb2 is smaller and fits better in cache May 13 07:21:07 this has made projects such as rhobuntu and mer difficult to support a broad range of hardware with vanilla ubuntu repos May 13 07:21:10 the more code in cache, the faster May 13 07:21:14 "that" being "kexecboot" May 13 07:21:20 so all packages are now thumb2 code? May 13 07:21:34 tmzt: Sure, but if we kept v5 only, we'd be irrelevant for v7 projects May 13 07:21:46 tmzt: thumb2 is the compiled code May 13 07:21:58 it's instructions May 13 07:22:09 I can understand that for, say, gstreamer codecs May 13 07:22:20 tmzt: It really benefits everything May 13 07:22:33 tmzt: It's not like neon May 13 07:22:42 tmzt: what you say would typically apply for neon May 13 07:23:07 since most packages dont take advantage of it, and the toolchain doesn't use much of it when it can May 13 07:23:50 tmzt: But when you tell gcc/gas that they should generate thumb2, it makes a big difference because they will generate 16-bits opcodes instead of 32-bits ones May 13 07:23:56 how much of an improvement is thumb2 over thumb? May 13 07:23:58 In the end, all the little libraries and programs fit better in cache May 13 07:24:07 UEFI ARM? That's an odd idea. May 13 07:24:08 tmzt: It's mostly an improvement in code coverage May 13 07:24:20 tmzt: it wasn't really possible to use thumb for everything May 13 07:24:34 but the compiler would fall back to arm when thumb didn't support the needed instruction? May 13 07:26:10 Thu May 13 00:26:10 PDT 2010 May 13 07:26:15 time for me to go to bed. =þ May 13 07:27:43 Anyway, I suppose the eventual blueprints will be enough of a "transcript" (or rather, will distill down the overall points). May 13 07:28:21 tmzt: Yes, you'd have to switch back and forth May 13 07:28:43 Or stream in stereo. May 13 07:28:51 Left side in one room, right in the other. May 13 07:28:57 =þ May 13 07:29:46 http://www.arm.com/about/newsroom/24978.php May 13 07:29:48 Nifty. May 13 07:29:54 UEFI including ARM. May 13 07:30:01 Now... we just need Ubuntu to support UEFI! May 13 07:31:03 Actually, I can imagine HP, for example, liking that.... they already use UEFI on their business laptops. May 13 07:34:56 cf. https://wiki.ubuntu.com/FoundationsTeam/Specs/MaverickUefiSupport May 13 07:35:05 which was discussed the first day ;) May 13 07:35:18 Sweet. May 13 07:36:04 I've messed with UEFI on my laptop... first-gen Elitebook. Has somewhat broken memory map. May 13 07:36:06 is there any public discussion of qualcomm kernel-qcm? May 13 09:32:41 has anyone tried to build mysql-server from the source package (karmic) on arm? version mysql-dfsg-5.1_5.1.37-1ubuntu5.1, here's the pbuilder log: http://pastebin.com/Y7fwHrV7 May 13 09:33:14 no matter if I build it using pbuilder or without, I get the same result May 13 09:33:55 I'll file a bug report, but I thought I'd ask here first just in case anyone else has seen this before May 13 09:35:22 markos_: no idea - I am new member of ubuntu/arm team May 13 09:36:23 hrw-uds: congrats then :) May 13 09:36:30 thx May 13 10:16:56 NCommander, around ? May 13 10:18:35 ogra: yeah, I'm in the basement May 13 10:18:43 ogra: where are you? May 13 10:18:53 NCommander, in the UI heads session May 13 10:19:00 NCommander, so i talked to colin May 13 10:19:17 ogra: (I've been having some thoughts about doing hd-install, and I'm not sure this is the proper solution) May 13 10:19:36 NCommander, after a bit of going back and forth we came to the conclusion that actually the best solution is an OEM image May 13 10:19:42 ogra: ? May 13 10:20:04 its trivial to change livecd-rootfs to omit casper and instead add oem-config.-gtk May 13 10:20:21 ogra: didn't I suggest that earlier? May 13 10:20:38 roll a live image thats going to a real FS on antimony May 13 10:20:39 ogra: its not a bad idea expect then we need a way to resize the DD'ed SD card image May 13 10:21:29 nah, leave it to the user to add more partitions, we can put gparted in May 13 10:21:30 ogra: I thought your objection to doing oem-config in place of ubiquity was a speed issue and such May 13 10:21:37 it is May 13 10:21:38 ogra: ugh, that stinks of HACK May 13 10:21:44 but we need to fix that anyway May 13 10:21:56 ogra: fair enough. if a live image is acceptable, this works for me. May 13 10:22:05 well, its not a live image May 13 10:22:15 ogra: since its trivial to drop ubiquity (we just need to change the second package installation pass) May 13 10:22:21 its essentially a rootfs on a artitioned image May 13 10:22:25 ogra: well, from the perspective of image building :-) May 13 10:22:26 *partitioned May 13 10:22:30 er May 13 10:22:31 hrm May 13 10:22:32 actually May 13 10:22:38 tihs is a new type of image now that I think of it May 13 10:22:43 since we'll have no squashfs May 13 10:22:47 right May 13 10:22:56 we only use the squashfs as a base May 13 10:23:15 ogra: I rather just have it dump a tarball and let antimony learn how to deal with it; it will greatly reduce build time May 13 10:23:20 but do the content copying that ubiquity would usualy do during image build May 13 10:23:22 squashing an image takes ~1 h May 13 10:23:41 yeah, true, we can add such a function May 13 10:23:44 ogra: /etc/fstab, /etc/passwd, /etc/group May 13 10:23:49 f-k-i will have to be run May 13 10:23:54 the only thing that concerns me a bit is the actual image size May 13 10:24:11 you dont need fstab beyond proc May 13 10:24:17 ogra: we might want to look at using livehelper instead, since this looks like a massive pain to hack onto livecd-rootfs May 13 10:24:20 oh May 13 10:24:21 fair enough May 13 10:24:25 passwd and group are set up by oem-config May 13 10:24:32 ogra: but then there won't be an fstab entry for / May 13 10:24:40 you dont need one May 13 10:24:42 ogra: I thought oem-config needs an account added to run at boot time May 13 10:24:50 it doesnt May 13 10:24:56 casper makes one for the livecd environment, and the installer adds an ubuntu:ubuntu user for first boot May 13 10:24:57 it adds one actually :) May 13 10:25:05 thats the live session May 13 10:25:13 oem-config doesnt need any user May 13 10:25:23 ogra: then what user does GDM log in as? May 13 10:25:32 so i'd like to build two kinds of images ... May 13 10:25:48 there is no gdm running, oem-config is the DM May 13 10:25:55 DM? May 13 10:26:03 display manager May 13 10:26:16 ogra: the dm is started up by GDM, which is set to autologin on first boot with oem-config May 13 10:26:21 an oem-config image boots into a stripped down ubiquity May 13 10:26:29 ad the DM May 13 10:26:43 instead of GDM May 13 10:26:44 I've got to be missing something May 13 10:26:48 But that's fine May 13 10:26:48 yes :) May 13 10:27:01 ogra: er, ubiquity launchs GDM though in all cases though ... May 13 10:27:06 try a rootstock build and install oem-config-gtk in it May 13 10:27:11 ogra: I'll take your word for it :-) May 13 10:27:16 ubiquity doesnt touch gdm May 13 10:27:29 so May 13 10:27:37 i'd like us to build two images May 13 10:27:38 livecd-rootfs needs to grow a new function mode May 13 10:27:45 starting out with minimal one May 13 10:27:54 +a May 13 10:28:06 ogra: minimal? like just a basic command line system? May 13 10:28:12 which is only a basic debootstrap+oem-config May 13 10:28:17 right May 13 10:28:26 we'll need a seed for that May 13 10:28:38 ogra: we can't use tasks in PPAs May 13 10:28:42 (limitation of Launchpad) May 13 10:28:44 and the other one should be a stripped down netbook May 13 10:28:58 thats why i want to use a seed :) May 13 10:29:04 so we can have a metapackage May 13 10:29:09 ogra: right, but we'll have to make sure we keep the metapackages up to date May 13 10:29:12 :-) May 13 10:29:22 ah good, we're on the same page, we're just cross-typing each other May 13 10:29:26 they live in the PPA, shouldnt be a prob May 13 10:29:37 right, we did this for Moblin May 13 10:29:42 * NCommander shivers May 13 10:29:55 so what i'd like us to come up with is some idea for the size issue May 13 10:29:58 ogra: hrm, we can't put stuff on an ext4 image though May 13 10:30:05 i dont want to end up with 3G images May 13 10:30:07 At least, not if we use antimony May 13 10:30:17 we can use ext3 May 13 10:30:22 Slow, but doable May 13 10:30:23 no pob with that May 13 10:30:25 e2tools is VERY slow :-/' May 13 10:30:28 nope, pretty fast May 13 10:30:38 It is? Not the last time I used it May 13 10:30:40 right, we will use gene2fs May 13 10:30:45 OH May 13 10:30:46 That works May 13 10:30:50 indeed May 13 10:30:54 and doesnt need root May 13 10:30:59 Right, so d-cd needs to grow a new ext2/3 backend May 13 10:31:05 needs to tinker with the journal afterwards though May 13 10:31:10 I already have code on a local branch for that May 13 10:31:16 so it gets converted to ext3 May 13 10:31:27 ogra: how do we do that though, we can't loopback mount as non-root May 13 10:31:45 why would you loopback mount anything ? May 13 10:31:56 ogra: to twiddle the journal with tune2fs May 13 10:32:04 no need for that May 13 10:32:05 unless you another way to get journals enabled May 13 10:32:13 tune2fs can work on .img files May 13 10:32:16 directly May 13 10:32:19 didn't know that May 13 10:32:21 Nifty May 13 10:32:47 ogra: the lack of ext4 is a bit disappointing, but not a show stopper May 13 10:32:55 ext3 is fine May 13 10:33:14 ogra: so the images will have two partitions. vfat /boot, and / ext3, right? May 13 10:33:32 i was planning to implement the same setup in rootstock to overcome the usage of root May 13 10:33:48 so i can just copy/paste the PoC from there into antimony May 13 10:33:59 three May 13 10:34:03 ogra: we can expand the partition map on first boot now that I think about it. OEM had to do that on the Dell mini's once May 13 10:34:08 err May 13 10:34:09 two May 13 10:34:19 We just have to look at how they did it, and then make the ext filesystem grow May 13 10:34:19 but no vfat /boot May 13 10:34:30 ogra: ? May 13 10:34:40 dpkg doesnt cope with vfat /boot May 13 10:34:48 ugh May 13 10:34:55 we will do something similar to imx51 May 13 10:35:01 ogra: so flash-kernel voodoo? May 13 10:35:01 yeah May 13 10:35:07 Ok, so I see how this will come together May 13 10:35:13 a raw hidden partition that pretends to be a flash May 13 10:35:34 d-cd will grow ext2/3 image support, that image is passed into {post-}boot scripts, and then we handle it like we do on imx51 and dove May 13 10:35:37 flash-kernel will then mount that and cp uImage/uInitrd to it May 13 10:36:35 that code from oem ... does that automatically grow the partition ? May 13 10:36:48 ogra: it had to, but I don't know the full story behind it May 13 10:37:01 ogra: but its a solved problem, we just need to figure out how they solved it May 13 10:37:04 can you try to get your hands on that code and test it ? May 13 10:37:21 ogra: let me see if I can figure out who solved it originally, this was almost a year ago May 13 10:37:24 i'll look into livecd-rootfs and the build scripts May 13 10:37:57 (or feel free to do the d-cd changes and give them to me) May 13 10:38:06 so we have something to get started from May 13 10:38:16 ogra: I can handle the build scripts/d-cd if you want, I have local infrastructure May 13 10:38:26 fine then May 13 10:39:00 i'll put an action point list together and write up an explanation of what we'll do to david May 13 10:39:56 ogra: right, so I'll make d-cd grow ext support, you'll make livecd handle making a pre-installed image May 13 10:40:17 or do you want me to handle the entirity of the image stack? May 13 10:40:19 right, and you try to get your hands on the grow partition code May 13 10:40:26 (this is work I enjoy doing so I don't mind taking it) May 13 10:41:31 let me make up the bulletpoint list and we look who does what then (by mail), if you want to start already, start with d-cd May 13 10:42:21 ogra: execellent, I'm just fixing up my local build infrastructure here now May 13 13:55:29 what 's the dbfifo use for ? May 13 13:59:05 armel-rootfs-201005131535.tgz ? May 13 13:59:20 i cannot find the img file May 13 13:59:41 did you tell rootstock to actually build an img ? May 13 13:59:50 by default it only builds a tgz May 13 14:00:00 you need to use --keepimage May 13 14:00:44 but how to convert armel-rootfs-201005131535.tgz to an img file ? May 13 14:01:14 its described on the RootfsFromScratch wikipage (see channel topic) May 13 14:16:27 dd if=/dev/zero of=ubuntu-arm.img bs=1MB count=0 seek=1024 May 13 14:17:06 what size i should to set ,i don't want to waste the space May 13 14:35:31 dd if=/dev/zero of=myimage.img bs=1M count=1024 May 13 14:35:40 mkfs.ext3 myimage.img May 13 14:35:51 mount myimage.img /mnt -o loop May 13 14:36:06 sudo tar xf tarball.tgz into /mnt May 13 20:15:13 hello May 13 20:15:31 am trying to run ubuntu in my iphone May 13 20:15:55 is there any one tried to do something like that? May 13 21:05:53 waleed: rhobuntu May 13 21:05:55 what phone? May 13 21:06:09 Iphone May 13 21:06:26 iphone 2g May 13 21:06:31 with the new android kernel support it should be possible May 13 21:06:48 I can try to help you in #htc-linux but I'm not in #iphonelinux or whatever May 14 00:58:17 hey DanaG, i've been playing around with doing something different to build the SGX modules (so far it works with the '02).. (instead of cross building the modules later..) http://bazaar.launchpad.net/~beagleboard-kernel/+junk/2.6.34-devel/files/head:/patches/sgx/ May 14 01:05:13 ah, where do those apply? May 14 01:07:25 i basicly dumped the gpl modules from the SDK then dumped/wired them into a staging directory... May 14 01:09:19 ah, cool. May 14 01:09:37 right now i'm building them against 2.6.34-rc7 but they should work on any kernel source with a staging tree.. now it's just make uimage, make modules.. copy over.. module loads.. ;) May 14 01:09:41 Then what's left to build manually? Just userspace? Or wait... those parts would be binary, anway. May 14 01:10:03 ah. I was hoping to have it be debs -- like packaged fglrx or nvidia. May 14 01:10:23 yeap. deb's included. ;) May 14 01:10:37 then you just need to run this user script on x86 http://bazaar.launchpad.net/~beagleboard-kernel/+junk/2.6.34-devel/annotate/head:/create_sgx_package.sh May 14 01:10:49 i'll tar up a package with an install script... May 14 01:10:55 Sweet. May 14 01:11:00 Though, I'll probably need a bigger SD card. May 14 01:11:26 My 2GB card is... oh, 87% full. May 14 01:11:33 i use a 2Gb with an extra 1GB for the SDK Demos... May 14 01:11:56 have you done a "sudo apt-get clean" to remove any deb's in /var/cache.. May 14 01:13:10 yeah... just did it; only freed 17 megs. May 14 01:13:40 dang.. you must have isntalled one of the *-desktop's.. May 14 01:16:19 I'm not sure... lemme' check. May 14 01:26:39 /dev/mmcblk0p2 1.9G 1.5G 243M 87% / May 14 01:27:08 http://pastebin.com/saRgbwgF May 14 01:27:27 dpkg -l May 14 01:31:01 yeah a little bit of everything, like any normal system... May 14 01:33:53 I do have an 8GB Cruzer I could use... but it's damn slow (slow as in, 2 megabytes per second write.) **** ENDING LOGGING AT Fri May 14 02:59:56 2010