=== ian_brasil___ is now known as ian_brasil [04:43] Say, any new news on getting dkms for powervr? [04:43] Beagleboard, specifically. [04:45] one of the linaro kernel guys pinged me about it last week.. (his dvi display wasn't working) but he has some questions on my sgx module patches.. [04:46] It really sucks that TI is so uptight about their graphics... they should take an example from nvidia, of all people. [04:46] It's also confusing... for an Ubuntu beagleboard, do I need psp-sdk? [04:47] oh yeah, and thanks for the new setup script -- setup_sdcard.sh [04:48] Though, one improvement: it needs to "inhibit" automount as gparted does, or sometimes you get "Can't reread partition table". [04:48] umm.. TI can't do anything about it.. By licenseing the core from imgtek (first mistake), TI has some rights, but imgtek makes the rules.. [04:48] Or TI should at least make the build script not phail. [04:48] And even when inhibited, the error still sometimes happens. [04:49] do you know anyway to in inhibit, other then going into gconf and turning it off? [04:49] It's probably a dbus call. [04:50] or udisks --inhibit [04:53] interesting, i'll give that a try tomorrow.. hopefully in time for the 10.04.1 image, thanks DanaG [04:54] ah, and when I tried to deploy btrfs, I got lack of fsck.btrfs. [04:56] yeap, i've seen that too... i even added btrfs-tools to the default roostock image, thinking that would help, but no.. (even alpha-3) [04:56] Argh, stupid wireshark... [04:57] I try to type "usb0" in capture interface box, I get: [04:57] ub0s [04:57] And wlan0 is ub0s [04:57] er [04:57] wan0l [04:57] And eth0 is eh0t [04:58] Note: I think the udisks call to inhibit runs only as long as the udisks command is running. [04:58] So, you may have to background it, then kill %something later. [04:59] yeap, that's the way it looks in the man page.. [05:03] DanaG, what's weird, the package btrfs-tools contains 'fsck.btrfs' http://packages.ubuntu.com/lucid/i386/btrfs-tools/filelist kinda weird it isn't found at boot.. [05:03] It must not end up in the initramfs, then? [05:04] ARGH, stupid EEM. [05:04] I see packets of type 0xbc59 from ff:ff:d6:e3:b3:a0 [05:04] that's not a valid MAC address! [05:08] argh, and TI's mirrors are dog-slow... [05:08] 160 KB/sec on a connection that can do 500-600. [06:00] Wait, maybe it's wifi fail, or Charter fail. === freeflyi1g is now known as freeflying [07:58] does anyone still meet this issue with Maverick on Beagle? [07:58] [ 2.557128] mmc0: error -110 whilst initialising SD card [08:00] cooloney: known bug that mathieu is working on [08:02] cooloney: Are you hitting it? [08:03] cooloney: I guess this was fixed with latest update [08:04] cooloney: rsalveti: amitk: bug 591941 [08:04] Launchpad bug 591941 in linux (Ubuntu Maverick) (and 1 other project) "SDHC card not recognized (affects: 2) (dups: 1) (heat: 70)" [High,In progress] https://launchpad.net/bugs/591941 [08:04] amitk: yeah, i guess so [08:04] lag: i just hit that [08:04] rsalveti: i tried latest maverick kernel in our tree [08:05] rsalveti: lag: I don't consider that revert a 'fix'. It is a workaround at best :) [08:05] Mathieu and Robert are working on it [08:06] yeah, true, they are still working on it [08:06] ok, understood, thanks, bros [08:06] amitk: Agreed, hence the present tense of my previous statement :) [08:06] i just wanna test the kexec things on maverick [08:06] :) [08:06] the bug that got fixed is the mtd one [08:07] cooloney: Either do this: [08:07] CONFIG_SND_TIMER=m [08:07] CONFIG_SND_PCM=m [08:07] # CONFIG_SND_SOC is not set [08:07] or turn off CONFIG_PREEMPT [08:08] well, time to get some sleep [08:09] rsalveti: good night [08:10] lag: so that's config patch from Mathieu? [08:11] cooloney: ? [08:13] lag: i saw some discussion about the CPU_IDLE in the LP thread [08:13] but failed to find the CONFIG_SND_PCM stuff [08:14] cooloney: Ignore CPU_IDLE [08:14] I just turn off CONFIG_PREEMPT [08:14] cooloney: Wait one [08:15] lag: ah, i saw that, rcn-ee posted in his message [08:15] lag: ok, let me disable PREEMPT [08:15] cooloney: Do you know where it is? [08:16] Kernel Features -> Preemption Model [08:16] Change it from Desktop to Server [08:17] cooloney: Do you have a working kernel uImage that USB is proven to work with? [08:17] cooloney: In fact the .deb would be more helpful [08:19] lag: oh, haven't try USB much this time [08:19] Would you mind? [08:20] lag: could you try low-latency desktop for the pre-empt model. I noticed that Nokia ship with that enabled so their driver is tested in that environment [08:20] lag: need i test it on omap4 panda or omap3 beagle? [08:21] amitk: I've just looked, and that's what we have enabled at the moment - when did that happen? [08:21] lag: I did ask Mathieu to try that out in response to his patch on the list [08:22] cooloney: 4 [08:22] so that doesn't work either? [08:22] amitk: I haven't tried [08:22] ok [08:22] * amitk is off to meet Tony [08:22] amitk: I'll do so now though (neither of my boards work properly, so I may as well) [08:23] amitk: Say hi from me :) [08:23] sure [08:26] lag: will test it later. [09:33] Hey npitre [09:34] Hey lag. [09:34] MOrning GrueMaster [09:34] How's Oxford? [09:34] Same as Oregon, just older. [09:35] GrueMaster: I was there over the weekend Saturday was a bit drizzly, but Sunday was really nice [09:35] You should have stopped by. I got here Saturday. [09:36] GrueMaster: I was at a festival - I was in no state to be meeting colleagues :) [09:37] GrueMaster: I should mention that it was a Cider Festival :) [09:37] More the reason to find me. [09:37] GrueMaster: :) [09:38] So, what can I do you for? [09:38] Nothing really. Just saying hey. [09:38] GrueMaster: Ah, good morning [09:38] GrueMaster: What's the plan for today? [09:38] Did you say you needed to test something on panda? [09:39] No, I need a replacement [09:39] I'm in the CoP QA meeting, but also have my omap hw. [09:40] But I would need an email from David before I could leave you with my es1. [09:40] GrueMaster: Well if you have some spare cycles, feel free to test this: http://people.canonical.com/~ljones/usb-stopped-panda/ [09:40] The USB on my Panda has stopped working [09:40] I think it's hardware [09:40] ok, I'll pull it now and try it later this morning. [09:41] Excellent, thanks [09:42] Did you try rebooting with Alpha 3 image? [09:44] I've tried Alpha 3, yes [09:45] I hope this isn't another hardware killer patch from the kernel team. [09:47] GrueMaster: It's more likely to be a userland issue [09:48] =:-p [09:48] ogra: Have you noticed the reason for the image build failures? It appears that we are hanging on unity missing a package that exists in a ppa (from what I could find). [09:49] lag: PLBKAC [09:49] GrueMaster, have you noticed that tonights images have built ? :) [09:50] No, missed that. Last email Ihad said fail. Will pull immediately (but I believe oem-config is still bad). [09:50] the package was stuck in universe, pitti fixed that yesterday [09:50] ah. [09:50] * ogra is just done with pulling, gunzipping already [09:50] GrueMaster: DYJMTUY? [09:51] ? [09:51] My sentiments exactly [09:51] Did You Just Make That Up Yourself? [09:52] Problem Lies Between Keyboard and Chair. [09:53] PLBKACPU [10:04] SD I/O is slowwwww. [10:10] NCommander, could you put http://qa.ubuntu.com/reports/team-assigned/canonical-mobile-assigned-bug-tasks.html on the meeting agenda ? [10:10] (as a standing item) [10:13] ogra: can you add it yourself? my internet connection is really crappy, and the wiki VERY slow [10:16] ogra: what's the usual window manager when UNE 2D is runnign? [10:16] metacity [10:16] ogra: and what draws the bar at the top? [10:16] with maximus attached [10:16] (and what starts metacity) [10:17] hrm [10:17] thats gnome-panel [10:17] gnome-session starts metacity [10:17] ogra: seems I have an issue with gnome-session then :-/ [10:17] NCommander: panda is still updating. Hopefully done soon. [10:17] ogra: what tells gnome-session then to start UNE or normal GNOME? env variable? [10:18] gconf [10:18] * DanaG still wants texture_from_pixmap on ARM... =/ [10:18] well, gdm actually [10:18] gconf defines whats running in the session [10:18] And I wish that "panda" weren't so impossible to find online. [10:18] DanaG, they are not sold yet [10:18] ogra: ok, that makes sense. I'm trying to trace why I'm having a launcher failure [10:19] NCommander, which session do you try ? [10:19] ogra: both UNE and UNE 2D [10:19] UNE will just hang hard [10:19] I get netbook-launcher-efl to pop up, but nothing else [10:19] dont try that [10:19] UNE starts 2D [10:19] it doesnt [10:19] I don't even mean to buy... I just mean to find pictures! [10:19] if it does thats a bug you should talk to the desktop team about [10:19] Unless it's under NDA. :( [10:21] GrueMaster, no change with oem-config in todays image :( [10:21] Literally all I can find for "omap4 panda" are the kernel patches for "add support for omap4 panda". [10:21] fail. [10:22] i'm out of debugging ides :/ [10:22] *ideas [10:23] ogra: on what? [10:23] Say, if I were to take the GL ES apps source from the TI SDK, could I use it with the Mesa GL ES? [10:26] NCommander, for the oem-config issues [10:26] ogra: I'll see if I can tweak with it a bit this week. [10:26] ogra: ouch. I can confirm however that using the normal UNE session and not UNE 2D still causes the 2D launcher to stop [10:27] ogra: can you please poke the casper bug for me? [10:27] NCommander, if the papaerwork on the bug is done :) [10:27] (remember i asked you about an additional task) [10:28] ogra: bah, I thought you were going to add that :-P [10:28] right, thats the reason why i asked you :P [10:29] ogra: ok, so starting the 2D launcher causes gnome panel to startup in desktop configuration, and not "normally" for UNE 2D [10:29] ogra: and the 3D session starts the 2D session san panel [10:29] yes, thats a known bug [10:29] W.T.F. [10:29] the 3D session shouldnt start a launcher at all [10:29] only the two unity panels [10:29] ogra: is the 2D session known to be broken? [10:29] but they wont start up if you dont have proper GL support [10:30] I don't think it should be showing a normal gnome-panel [10:30] ogra: ok, so starting the 2D launcher causes gnome panel to startup in desktop configuration, and not "normally" for UNE 2D [10:30] yes, thats a known bug [10:30] argh, stupid PowerVR... not only are their drivers closed-source... even their freaking demos are closed-source. [10:30] * NCommander hits his head [10:30] ogra: so what's the fix :-)? [10:30] So you can't compile any GL ES stuff for non-PowerVR stuff. [10:30] NCommander, oem-config needs to work again so i can move on :P [10:30] Or at least, there's no GL ES glxgears! [10:31] Anyway, off to bed I go nowz. [10:31] NCommander, we have a -settings package but no settings in it yet [10:31] ogra: but this worked in lucid! [10:31] to set these up i need a working session [10:31] GNOME changed a lot [10:31] .oO(we should have used KNE) [10:31] * NCommander ducks [10:31] and the setting package we shared with the 3D launcher is gone [10:32] because of unity [10:32] ogra: paperwork done :-P [10:32] thanks [10:32] i'll upload before the meeting [10:35] * GrueMaster votes +1 for KNE [10:55] hmm [10:55] seems /var/lib/dbus/machine-id exists and is wrong on the images [10:56] so dbus doesnt start and it tears down oem-config with it [10:56] * ogra tries to prove that theory now [10:56] * GrueMaster is still waiting for Alpha 3 -> current updates to finish. === XorA|gone is now known as XorA [11:04] ogra: I'll look at ubuntu-netbook-2d-settings sometime this week if you want, since I need to ASAP for Marvell :-/ [11:04] i'll do it as soon as i can get a working session [11:04] ogra: fair enough [11:04] its not that it blocks anything from working [11:05] * NCommander must admit he feels exhausted today [11:05] * ogra guesses you will enjoy the team call this week :P [11:06] hmm, so there is no such dbus file in the plain images [11:06] i wonder why dbus doesnt start then [11:07] Did you compare with A3? [11:08] ogra: maybe it missed its bus to work [11:08] * NCommander runs [11:09] haha [11:09] GrueMaster, no, nothing to compare here [11:33] ogra_cmpc: bug 605488 [11:33] Launchpad bug 605488 in linux-ti-omap4 (Ubuntu Maverick) (and 1 other project) "BUG: scheduling while atomic: mmcqd/46/0x00000002 (affects: 1) (heat: 132)" [High,In progress] https://launchpad.net/bugs/605488 [11:33] ogra: How big should the swap file be? [11:36] GrueMaster: ? [11:38] sawp file? [11:38] swap [11:39] I think swap is currently turned off. [11:39] ogra said this bug only appears when swap is on [11:39] Oh, that one. Make a swap file == ram size. [11:40] k [11:40] Well ... [11:40] Only 256MB works [11:40] Shall I use that? [11:40] I guess. [11:40] Or the actual 512MB [11:40] k [11:40] Panda? [11:40] Beagle [11:41] bs? [11:42] so, you are checking an omap4 bug on an omap3 system? [11:43] GrueMaster: So it is! [11:43] Actually - I'm not [11:43] I am on Panda [11:44] (things get confusing when you're working on both boards at the same time) [11:44] then my previous reply counts. [11:44] (I am doing the correct thing) :) [11:44] Which was? [11:44] 512MB? [11:44] yes [11:44] np [11:44] Thanks [11:57] GrueMaster: How can I reproduce this bug? [12:04] lag, just reboot, the messages about swapping should show up in dmesg [12:06] yep. Should see it as soon as the system mounts swap. [12:14] ogra_cmpc: GrueMaster: Should I still be able to see it whilst using a rootstock rootfs? [12:18] I don't know. [12:19] since its a kernel message the rootfs shouldnt matter [12:19] http://paste.ubuntu.com/479343/ [12:20] though it might only happen if the system actually uses swap [12:20] Well here's what I do get [12:20] ogra_cmpc: Should I write an application which uses lots of memory? [12:33] Lanch une. [12:33] launch [12:40] lag, looks o so far, just start a desktop session and use firefox for a while or something [12:41] ogra: I can't [12:41] ogra: No USB [13:58] GrueMaster: ping [14:02] pong [14:02] * persia starts a meeting in #ubuntu-meeting [14:36] ogra: any change to sponsor me at bug 619026? [14:36] Launchpad bug 619026 in pybootchartgui (Ubuntu) "pybootchartgui gives ZeroDivisionError when parsing the tgz file generated by a beagleboard (affects: 1) (heat: 8)" [Undecided,In progress] https://launchpad.net/bugs/619026 [14:36] was testing ureadhead yesterday but found out that I couldn't generate the bootchart file on beagleboard [14:37] well, you can copy the tgz over to an x86 machine and generate the png there [14:37] and about ureadahead I'll just disable it if you run on systems with less than 512, doesn't change a thing at the bootchart [14:37] ogra: that's what I did [14:37] ogra: but the bug is inside pybootchartgui [14:38] it's not using the mmcblk device while generating the bootchart [14:38] ah, right [14:38] i'll sponsor it after the meeting [14:41] thanks [14:55] ogra: /sys/devices/system/cpu/ showscpu0 & cpu1. Checking further. [14:55] Kind of slow, still updating (from yesterday). [14:55] are you sure it just doesnt scroll off the screen or so ? [14:56] ogra: ever heard of the Sharp PC-Z1 NetWalker ? [14:56] what scroll off the screen? I have an HDMI monitor here. [14:56] rsavoye: I think persia has one. [14:56] rsalveti, persia had one for a while [14:56] err [14:57] rsavoye, indeed [14:57] it ships with 9.04, I but I heard you can;t upgrade due to driver problems [14:57] I used to have one, but I dropped it in the sink. It's a lovely machine, and I am tempted to get another almost every week. [14:57] I just want something that's stable :-) [14:57] A few folks have looked at the kernel patches, but untangling what needs to be applied to current kernels (some parts are upstream) seems yet to be resolved. [14:58] figures... [14:58] There's no git tree available, so it's just a huge monolithic patch to dig through. [14:59] somebody said they were running 10.04 on one, but just never upgraded the kernel [15:01] Interesting. Jaunty was armV5 (iirc). Lucid+ is armV7. [15:01] 800Mhz sounds nice too [15:02] ogra: On the panda, /sys/devices/system/cpu/cpu[01]/online are both 1. I assume this means they should be on. [15:04] GrueMaster, It's i.mx51x, so it can do v7, but who knows how stable the userspace/kernel interface is in that configuration :) [15:06] rsavoye, My critique of the netwalker is only that I don't like the keyboard that much (the 'a' key is in the wrong place, to my mind), and that the screen doesn't lay flat (or convert into tablet mode). Beyond that, it's lovely (even running jaunty), and there is some upstream kernel support, so you might join the several folks trying to get newer kernels there. [15:06] Yes, I know. I was refering to the images. [15:06] persia: I have an old Sharp 3200, same teeny keyboard [15:07] The way the super-special-sharp-installer for those works, one has to make a mess of the images anyway, which makes it less important. [15:07] Does the raw Jaunty image work? [15:07] as much as I'd love to jump into the kernel patches, I need to finish my current ARM hacking first [15:07] rsavoye, Not at all. the SL-C3200 has a *much* better keyboard. [15:07] persia: oh no! Course I was considering using it as a build slave mostly [15:07] GrueMaster, nope, missing kernel patches [15:07] (and is ergonomically better overall: pity that form-factor wasn't used in the more recent devices) [15:08] Oh, as a buildd slave, just install pbuilder on top of the sharp remix, and have at it. Nothing blocking there. no need to fiddle the kernel. [15:18] Hmm, interesting. /proc/cpuinfo only shows one proc, but top shows both (one is barely awake). === bjf[afk] is now known as bjf [15:18] It does seem to be passing the load back & forth. [15:20] thats how its supposed to work [15:20] htop is better than top btw [15:21] I just found it odd that /proc/cpuinfo only showed one proc. [15:23] are you sure its not scrolled offscreen ? [15:27] Is what scrolled of screen? [15:29] the cpuinfo output [15:30] It only fills a quarter of the screen. [15:30] if that [15:32] ok, now I see it. Very weird. [15:45] ogra: How are rootstock's commands compiled? (which toolchain) [15:49] lag: a little elf writes down the list of commands and then spends straw into binary code [15:49] prpplague: Okay, I'll cut to the chase [15:50] prpplague: My cross-compiler is a little out of date and I need another [15:50] * prpplague isn't familiar with rootstock operations [15:51] lag: i use OE and codesourcery for my cross-compilers [15:52] prpplague: I'm trying to cross-compile a little program to use up lots of memory - but if I compile it dynamically I receive "file not found" [15:53] prpplague: If I try it statically I receive "illegal instruction" [15:53] lag: that sounds odd [15:53] lag: mind if i give it a try? [15:53] I have chopped it (all the way) down to a hello world - same result [15:54] Do you want me to send you the binary? [15:55] lag: sorry i ment try compliling the code [15:56] prpplague: That's my point - it does the same for hello world [15:56] It must be the compiler - the code is sound [15:56] lag: what is command line you are using to cross-compile the source? [15:57] arm-unknown-linux-gnu-gcc <-static> -o use-memory-arm use-memory.c [15:58] lag, rootstock is a script [15:58] lag: after you compile it, issue the command: file use-memory-arm [15:59] Done that [15:59] lag: that should give you some basic info about how the binary was compiled [15:59] I did it already [15:59] lag: all that info looks good? [15:59] I have to go (meeting) will you be here in a bit? [15:59] Yeah [15:59] lag: yea all day [15:59] Sort of === Termana_ is now known as Termana [16:09] ogra: upgrading from Alpha 3 to current seems to fail somewhere. Just rebooted (finally), and it is locked up. Will dive into, butthought you would want to know. [16:10] GrueMaster, awesome, then its not oem-configs fault as i suspected [16:10] asac: I am having trouble getting to a point where I can test banshee for you. Will let you know when I can get past these issues. [16:11] GrueMaster, just dont upgrade, apt-get install banshee on a plain A3 [16:11] ogra: X isn't even starting. [16:11] GrueMaster: ok. let me know [16:11] it should only upgrade banshee related packages then [16:11] thanks for the update [16:11] GrueMaster, try startx on a tty [16:11] see what it tells you [16:12] I wonder if I can debug thumb2 code on an i.mx51x... [16:13] ogra: Total lockup. need to modify kernel cmdline to get serial out first. No prompt available. [16:13] wow [16:14] Not even getting numlock or capslock. [16:14] thats ES1 ? [16:14] yes. [16:14] smells like kernel [16:15] mount the boot partition, there are .bak files of kernel and initramfs [16:15] No. Had to reboot a couple times during the upgrade process (thought it had locked up) New kernel came up fine. [16:15] copy them around so that you boot the .bak files [16:16] I know the process, but I doubt it is the kernel. [16:16] well, do you see the same on an omap3 board ? [16:17] I haven't had a chance to try this on omap3. The time it takes is painful. [16:24] rsavoye, Should be able to do so: the core supports it [16:25] cool, I've tempted to buy one so I can stop sharing a C4, which is insanely slow [16:26] persia: since you have one, is the keyboard worth it, or is the tablet one better ? [16:26] ogra: old kernel locks. Now trying with serial console. [16:28] rsavoye, Depends on what you want to do with it. The tablet one is uninteresting to me (I don't want only a tablet at that size/weight). I complain about the keyboard, but I'm seriously considering getting another one. [16:29] for me it'd mostly be an ARM build slave, I don't like netbooks generally [16:29] the only other difference is the mmc size, but I could add a bigger one myself [16:30] That would involve soldering :) [16:31] ouch. one has 4GB, the other 8GB [16:32] maybe an external USB drive... [16:33] GrueMaster: i'd really like to see some info on how to get the ubuntu builds working with distcc or ice to improve native compiles [16:34] GrueMaster: you guys got anyone working on that? [16:34] The only one that did that I know of is NCommander, but he's probably out atm. [16:35] GrueMaster: yea i asked, he only used ice once to do a couple of small compiles [16:35] rsavoye, As a buildd slave, you *definitely* want an external drive, preferably rotary (otherwise you go through a heap of flash) [16:35] You could always setup an NFS share. [16:35] USB2 drives are cheap these days, or I could just mount space off another machine [16:36] * prpplague has been experimenting with a fpga design that emulates nand flash but uses an external SATA drive [16:36] I use NFS on my XM board, which unfortunately is useless still for doing builds :-( [16:36] somehow I think I shouldn't hold my breath for the XM kernel fixes... [16:36] I have a NFS share for builds and nfs-root, but not enough systems to try this with. [16:37] but it's impossible to run GDB on a C$ when somebody else is running linpack and whetstone tests... [16:37] rsavoye: they should be out soon. [16:37] GrueMaster: this week ? :-) [16:37] I'm in a crunch for time unfortunately [16:37] I don't know, I only test them. [16:37] I' [16:37] Srry. [16:37] GrueMaster: other fun items on my todo list, research multi-framebuffer support in ubuntu [16:37] l'll gladly test them too :-) [16:38] prpplague: Same here (if someone will send me a blaze). [16:38] prpplague: Hola [16:38] GrueMaster: or panda [16:38] lag: hey [16:38] prpplague: Where were we? [16:38]