**** BEGIN LOGGING AT Thu Aug 26 02:59:57 2010 Aug 26 06:49:56 Hi Aug 26 06:52:28 Hi Aug 26 06:53:08 mouse and kayboard not working Aug 26 06:53:26 writing ubuntu-10.04-netbook-armel+omap.img on sd and cheng boot.scr Aug 26 06:53:27 to folowing: Aug 26 06:53:38 fatload mmc 0:1 0x80000000 /casper/uImage Aug 26 06:53:39 fatload mmc 0:1 0x81600000 /casper/uInitrd Aug 26 06:53:39 setenv bootargs quiet splash vram=12M omapfb.mode=dvi: Aug 26 06:53:39 1280x720MR-16@60 fixrtc file=/cdrom/preseed/ubuntu-netbook.seed -- Aug 26 06:53:39 boot=casper only-ubiquity nocompcache mpurate=720 console=tty0 Aug 26 06:53:39 console=ttyS2,115200n8 Aug 26 06:53:41 bootm 0x80000000 0x81600000 Aug 26 06:53:43 booting beagleboard and see this error: Aug 26 06:53:45 hub 1-0:1.0: unable to enumerate USB device on port 2 Aug 26 06:53:47 My mouse and keyboard not working, how to resolve this problem ? Aug 26 06:53:49 tanks Aug 26 06:54:19 On C4 beagleboard Aug 26 06:55:50 ? Aug 26 06:55:51 ? Aug 26 07:21:39 morning Aug 26 07:23:01 say, anyone have experience with webcams on a beage? Aug 26 07:33:59 try #beagle Aug 26 07:34:28 lamp_: have you connected to the ehci host port or the otg port? Aug 26 07:34:50 lamp_: and are they connected directly or through a usb hub? (A hub is required) Aug 26 07:35:07 ehci Aug 26 07:35:25 connected directly Aug 26 07:40:04 lamp_: directly won't work, you need to connect it through an externally powered hub Aug 26 07:48:53 even otg port Aug 26 07:49:46 flash usb working on directly connect Aug 26 07:51:07 flash usb working on directly connect to ehci Aug 26 07:51:46 sell externally powered hub ? Aug 26 07:52:17 ? Aug 26 08:01:05 Are you sure that buying a externally powered hub The 'hub 1-0:1.0: unable to enumerate USB device on port 2' problem is solved? Aug 26 08:04:05 amitk: can you use codesourcery 2010q1 cross compiler to build our maverick kernel package for omap3 Aug 26 08:16:09 cooloney: I haven't tried it. I stick to old CS toolchains in general :) Aug 26 08:16:35 amitk: sorry for bothering, i just tried that, it works. Aug 26 08:38:13 amitk: Eh you should dogfood or cross-compiler packages!! Aug 26 08:45:58 * ogra thinks we need more builders ... 50 package in queue and all builders blocked :( Aug 26 08:47:59 lool: I'm running maverick on the laptop, so I do a bit of dog-fooding there. I wish we had lucid packages though Aug 26 08:49:02 amitk: they don't install on lucid? Aug 26 08:49:22 lool: they didn't last I tried (a few weeks ago) Aug 26 08:49:50 amitk: I'd be curious whether that's still the case Aug 26 08:50:11 * amitk re-adds hrw's repo to sources.list Aug 26 08:50:45 amitk: why not /etc/apt/sources.list.d/hrw-cross-compilers.list instead? Aug 26 08:50:56 easier to enable/disable Aug 26 08:51:09 hrw: that's what I'll do, much longer to type though ;) Aug 26 08:51:16 ;d Aug 26 08:52:58 hrw: is there a meta package? Aug 26 08:53:25 amitk: nope Aug 26 08:53:42 amitk: for kernel you only need gcc-4.4-arm-linux-gnueabi I think Aug 26 08:55:56 hrw: lool: http://pastebin.ubuntu.com/483859/ (needs libmpfr4) Aug 26 08:57:50 they are for *maverick* Aug 26 08:59:28 hrw: Would it be possible to build them under lucid? Aug 26 08:59:32 hrw: that's what you told me the last time, but I repeated the experiment because lool asked me to :) Aug 26 08:59:53 amitk: Given that the toolchain is relatively self contained, I was hoping these would be installable on lucid Aug 26 08:59:55 * amitk feels that lucid being LTS should be a target Aug 26 09:00:49 lool: some backports from maverick would be needed probably Aug 26 09:00:50 amitk: You could most probably install libmpfr4 from maverick though ;-) Aug 26 09:00:57 hrw: Yeah Aug 26 09:01:10 hrw: How far are you from a toolchain package for the archive? Aug 26 09:01:25 * ogra tries to pronounce libmpfr4 Aug 26 09:01:30 ... and fails :P Aug 26 09:03:22 lool: need to solve stage3 problems Aug 26 09:03:36 ~curse ubuntu for not using sysroot Aug 26 09:04:09 Don't do native for linux, it's heavy :-( Aug 26 09:04:25 ? Aug 26 09:04:57 lool, arent all our linux packages native anyway ? Aug 26 09:04:58 90+ MB Aug 26 09:05:40 ogra: it is libmp-4-french people ;) Aug 26 09:05:42 * ogra thought the packaing tree merge happens on a git level Aug 26 09:05:49 amitk, lol Aug 26 09:06:02 amitk: you use amd64? Aug 26 09:06:06 hrw: yes Aug 26 09:07:54 good. deboostrap lucid in progress Aug 26 09:11:51 hrw: thanks! Let me know when to test Aug 26 09:24:26 ok Aug 26 09:31:16 amitk: one more thing - you will get maverick gcc not lucid **** BEGIN LOGGING AT Thu Aug 26 09:33:24 2010 Aug 26 09:40:43 hrw: that is fine Aug 26 09:44:36 anyway - first lucid one Aug 26 09:44:53 I am curious what will it bring and how will look Aug 26 09:48:48 where can I find content of hrw-cross-compilers.list ? :) Aug 26 09:51:12 ynezz: At the top of http://people.canonical.com/~hrw/ubuntu-maverick-armel-cross-compilers/ Aug 26 09:51:16 deb http://people.canonical.com/~hrw/ubuntu-maverick-armel-cross-compilers/ ./ Aug 26 09:51:19 etc. Aug 26 09:51:23 Currently maverick-only Aug 26 10:06:08 XorA: hi Aug 26 10:06:11 oops Aug 26 10:06:14 -> linaro Aug 26 10:21:54 rsalveti: ping Aug 26 10:27:32 amitk: lucid require too many backports to build maverick cross toolchain. better grab maverick one and add all needed maverick libs Aug 26 10:34:54 hrw@canonical? what a change :) Aug 26 10:35:22 ynezz: its 4 months now Aug 26 10:35:51 ah, didn't noticed Aug 26 10:41:50 does it mean, that ubuntu is switching from native to cross? :) Aug 26 10:41:56 no Aug 26 10:42:09 Very much not! Aug 26 10:42:14 it means that developers can do cross builds if needed Aug 26 10:42:22 ok Aug 26 10:42:29 ubuntu will never switch from native to cross Aug 26 10:47:18 lool, bug 624568 FYI Aug 26 10:47:20 Launchpad bug 624568 in busybox (Ubuntu) "building busybox without -marm on armel makes several binaries unusable (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/624568 Aug 26 10:51:14 asac: hi Aug 26 10:52:09 hey XorA ;) Aug 26 10:52:36 XorA: so the maverick stack is now more or less ready to start working on things Aug 26 10:52:57 XorA: you can intall mutter and then just install Aug 26 10:52:58 libclutter-eglx-es20-1.0-0 Aug 26 10:53:06 to flip to egl Aug 26 10:53:35 XorA: to implement the extension we talked to about just do a apt-get source libclutter-eglx-es20-1.0-0 ... and then look at the texture_pixmap function in the gles tree Aug 26 10:53:36 ynezz: cross builds are option for developers to not wait 3 days for kernel rebuild but have it done in short time Aug 26 10:53:40 etc Aug 26 10:54:04 XorA: -> #linaro ;) Aug 26 11:11:58 NCommander, please bump https://edge.launchpad.net/ubuntu/+source/u-boot-linaro/2010.06-695-gbd23130-linaro-0ubuntu1/+build/1934915 to $very_high Aug 26 11:12:35 * persia encourages asking random buildd admins for such things in #ubuntu-devel: tends to avoid timezone limitations Aug 26 11:12:54 i *like* to nag NCommander directly :P Aug 26 11:12:55 (as in "Could a buildd admin please ...") Aug 26 11:13:06 Yeah, but doesn't help if he's not around. Aug 26 11:13:10 i usually fall back to that if he's not around Aug 26 11:13:32 I'm 90% certain that's the case now, although I may be mistaken. Aug 26 11:17:43 ogra: If I can get your XM to do this: http://paste.ubuntu.com/483904/ Aug 26 11:17:48 ogra: Can you start working on it? Aug 26 11:17:51 (again) Aug 26 11:19:20 lag, i'll try soon, i'm currently a bit busy aith panda Aug 26 11:19:24 *with Aug 26 11:19:35 lag, thats with todays image ? Aug 26 11:19:43 ogra: No Aug 26 11:20:05 ogra: Today's image provides: http://paste.ubuntu.com/483908/ Aug 26 11:20:42 grmpf Aug 26 11:20:48 doesnt resize, doesnt reboot Aug 26 11:20:59 Correct Aug 26 11:21:07 i uploaded a fix for the reboot issue yesterday Aug 26 11:21:08 So, back to my original question Aug 26 11:21:17 * ogra takes a look what happened to the busybox upload Aug 26 11:21:22 ogra: No Aug 26 11:21:30 ogra: It has nothing to do with it Aug 26 11:21:45 ? Aug 26 11:21:47 I believe the daily image is still ro Aug 26 11:22:00 it should still reboot Aug 26 11:22:39 it cant be ro if you get through to oem-config Aug 26 11:22:54 since thats only enabled by touching a certain file on the FS Aug 26 11:23:20 line 327 in http://paste.ubuntu.com/483908/ Aug 26 11:23:25 Well tell me what the differences betweek the two pastes I gave you then? Aug 26 11:24:03 between* Aug 26 11:24:14 weird i see the initrd script being executed twice in the latter one Aug 26 11:24:36 lag, the first one reboots fine Aug 26 11:24:45 ogra: Correct Aug 26 11:24:54 which image is that ? Aug 26 11:25:06 The second one is the daily build Aug 26 11:25:26 The first one is the daily image with my kernel and initrd Aug 26 11:25:26 and the first one that works ? Aug 26 11:25:32 oho ! Aug 26 11:25:41 same image ? Aug 26 11:25:45 but different kernel ? Aug 26 11:25:45 Yeah Aug 26 11:25:49 And initrd Aug 26 11:26:01 so i was right saying the broken reboot is a kernel issue ;) Aug 26 11:26:07 and you already fixed it ! Aug 26 11:26:17 Seemingly Aug 26 11:26:23 perfect Aug 26 11:26:24 Where is Busybox? Aug 26 11:26:36 whatever you did, i want it in our kernel then :) Aug 26 11:26:47 its the shell that runs in initrd Aug 26 11:26:58 So it must be using my Busybox? Aug 26 11:27:07 your chroot should have two packages, busybox-static and busybox-initramfs Aug 26 11:27:19 * lag looks Aug 26 11:27:26 can you check which version your used inside the chroot ? Aug 26 11:27:31 *you Aug 26 11:27:41 might be that my -marm fix is moot Aug 26 11:27:56 How do I check? Aug 26 11:28:04 dpkg -l|grep busybox Aug 26 11:28:17 ii busybox-initramfs 1:1.15.3-1ubuntu1 Standalone shell setup for initramfs Aug 26 11:28:23 1:1.15.3-1ubuntu3 has the -marm fix Aug 26 11:28:30 hmm Aug 26 11:28:47 1:1.15.3-1ubuntu1 is the oldest one we had in maverick Aug 26 11:29:07 could you try upgrading ? that version used to work all the time Aug 26 11:29:27 1:1.15.3-1ubuntu2 stopped working, 1:1.15.3-1ubuntu3 was supposed to fix that Aug 26 11:29:57 1:1.15.3-1ubuntu4 is the current one (with another fix on top of mine) Aug 26 11:40:29 gah, todays daily still doesnt reboot :/ Aug 26 11:40:33 at least on omap4 Aug 26 11:45:52 ogra: I know, I told you that! Aug 26 11:46:02 yes Aug 26 11:52:05 f*ck Aug 26 11:52:12 * ogra thinks he found the issue Aug 26 11:53:13 111m355 Aug 26 11:53:27 damn :) Aug 26 11:53:45 have to change that now :) Aug 26 11:55:13 it is screenlock passwd Aug 26 11:55:25 but yeap, i need to change Aug 26 12:23:34 ogra: And ... Aug 26 12:24:23 lag, fix is uploaded Aug 26 12:24:42 ogra: Which was? Aug 26 12:24:51 ogra: And why did it work with my kernel/initrd? Aug 26 12:24:58 if e2fsck finds a wrong last mount time (because of rtc skew or something) it tears down the whole resize script Aug 26 12:25:17 Where is that? Jasper? Aug 26 12:25:21 the last line in the resize script sets that reboot mark for the next script Aug 26 12:25:24 yeah Aug 26 12:26:12 no idea why your kernel initrd worked though Aug 26 12:26:22 since it doesnt seem to be busybox at all Aug 26 12:26:50 what fixes does your kernel include ? Aug 26 12:27:16 anything related to ext2/3 ? Aug 26 12:41:06 ogra: I'll show you Aug 26 12:48:14 http://kernel.ubuntu.com/git?p=lag/ubuntu-maverick.git;a=commitdiff;h=fa324d1893d06e157a2f93040a20f1d490c3c834;hp=978e830c47ca5de5824ddf3ba9f7d3571da765a7 Aug 26 12:48:21 http://kernel.ubuntu.com/git?p=lag/ubuntu-maverick.git;a=commitdiff;h=27dafb99b153bab4bf7f061889775761cf7c2caa;hp=fa324d1893d06e157a2f93040a20f1d490c3c834 Aug 26 12:48:26 http://kernel.ubuntu.com/git?p=lag/ubuntu-maverick.git;a=commitdiff;h=12f88410bcaab12078b27e766a42b12a7d4bd2b5;hp=27dafb99b153bab4bf7f061889775761cf7c2ca Aug 26 12:48:31 ogra: -^ Aug 26 12:48:39 yep, looking Aug 26 12:49:36 ah, mmc fixes Aug 26 12:49:50 Correct o mondo Aug 26 12:50:08 might be that thats related, i.e. if the mmc is bad e2fsck will exit 1 Aug 26 12:50:17 same symptom, different cause Aug 26 12:50:45 lets wait until jasper is in the archive and i rolled a new image Aug 26 12:50:54 (which might take eternally) Aug 26 12:51:21 ogra: Are you saying that there is no massive rush to place these patches into our kernel? Aug 26 12:51:23 we need more buildds Aug 26 12:51:38 lag, there is, but no way to test them quickly in the archive Aug 26 12:51:40 ogra: That's find by me - they're not even in the linux-omap kernel yet Aug 26 12:51:45 Nah, just faster buildds. Aug 26 12:51:49 lag, https://edge.launchpad.net/builders Aug 26 12:51:55 more just means more parallel builds of superceded stuff Aug 26 12:52:03 lag, we're 60 package (or 10-12h) behind on armel Aug 26 12:52:27 (thats a guesstimate, could also only be 8-9h) Aug 26 12:53:07 lag, where do they come from initially ? rcn-ee ? Aug 26 12:53:23 ogra: Yes, they are Robert's patches Aug 26 12:53:27 cool Aug 26 12:53:30 add them ! Aug 26 12:53:55 They have been ack'ed in linux-omap, but they're not in the tree yet Aug 26 12:54:07 pfft, lets be ahead :) Aug 26 12:55:48 When will development continue on the XM? Aug 26 12:57:33 as soon as we can boot it Aug 26 12:57:40 which your patch should fix Aug 26 12:57:47 i want working images for beta Aug 26 13:20:55 lag, dont you need to mention a bug # in pull requests (seeing your mail to the kernel ML) Aug 26 13:22:14 ogra: Although this does fix one of the bugs, it is not directly associated with one Aug 26 13:22:24 ah Aug 26 13:22:38 * ogra lears something new every day :) Aug 26 13:23:12 * ogra also learned that the new tires for his new car will cost him about 1000€ :( Aug 26 13:23:47 its only rubber ! damned Aug 26 13:24:23 Get steel wheels: more sparks, less frequent replacement Aug 26 13:24:53 yeah, looks surely more shiny to do 270Km/h with these ... though i fear the traction wont be as good :) Aug 26 13:25:21 Yeah, well. Aug 26 13:27:08 ogra: costs of using porsche? Aug 26 13:27:59 hrw, costs of buying porsche with horribly wide tires and knowing they need replacement Aug 26 13:28:46 ogra: Have you tested the daily build on the Panda ES1.0 today? Aug 26 13:28:48 i didnt know there is only one set thats allowed though Aug 26 13:29:03 lag, yes, thats what i did above to find the jasper issue Aug 26 13:29:07 ogra: sucks a bit Aug 26 13:29:12 well Aug 26 13:29:17 Hmm Aug 26 13:29:19 That's not good Aug 26 13:29:25 hrw, i asked for it, i get it :) Aug 26 13:29:29 yep Aug 26 13:29:42 * hrw rebuilds gcc-4.4 again to check for regressions Aug 26 13:30:27 lag, whats not good ? Aug 26 13:32:11 ogra: I recieved my, replacement Panda today Aug 26 13:32:19 ogra: My old one's USB was borked Aug 26 13:32:42 ogra: This one doesn't seem to want play nice with HDMI Aug 26 13:32:56 ah, well, your broken monitor again Aug 26 13:33:02 ogra: Nope Aug 26 13:33:04 is that es1.0 ? Aug 26 13:33:08 or 2.0 Aug 26 13:33:10 ogra: Different one Aug 26 13:33:14 ES1.0 Aug 26 13:33:17 yes, but which version Aug 26 13:33:19 ah Aug 26 13:33:22 I haven't plugged my ES2.0 in yet Aug 26 13:33:30 There are more than one? Aug 26 13:33:45 mine is on the way, just had a call from fedex customs Aug 26 13:34:44 How do you tell which version is which? Aug 26 13:34:58 one is painted black, one isnt Aug 26 13:35:17 wasn't black ones es2.0? Aug 26 13:35:24 right Aug 26 13:36:43 You asked me which version of ES1.0 I had Aug 26 13:39:26 ?? Aug 26 13:39:28 i didnt Aug 26 13:39:41 is that es1.0 ? Aug 26 13:39:41 or 2.0 Aug 26 13:40:45 Hmm my beagle doesn't see NAND anymore; I think this was a known bug and got fixed recently; can someone confirm that latest kernels have NAND? Aug 26 13:41:07 * ogra isnt sure that was uploaded to the archive yet Aug 26 13:41:14 i know the fix is committed Aug 26 13:41:25 lool, mpoirier was working on that Aug 26 13:42:08 lool: I committed the fix last week. Aug 26 13:51:15 mpoirier: c08fa0be3ddeaca289b0646c8d087fc3820a7f3f right? Aug 26 13:51:42 lool: hold on, I'll double check. Aug 26 13:56:07 lool: yes that is correct. Aug 26 13:57:15 mpoirier: thanks Aug 26 13:57:18 lool, jcrigby, btw, there is a few patches we carry that are not upstream yet, you might want to pull these in from the ubuntu tree Aug 26 13:57:44 (additionally to the NAND ones) Aug 26 14:07:40 Well mpoirier's patch doesn't apply in Linaro 2.6.35; presumably we got it from linux-omap Aug 26 14:08:14 * ogra_cmpc thought it was our own developent Aug 26 14:08:37 It's a workaround for the 2.6.35 situation, but upstream had already fixed it differently -- that's how I read it at least Aug 26 14:08:58 mmmmk Aug 26 14:09:02 mpoirier: f450d86790ebf72ac93c7ea5addd6fa278aae64c upstream I think Aug 26 14:09:27 well in linux-omap; checking linus now Aug 26 14:09:49 f450d86790ebf72ac93c7ea5addd6fa278aae64c in linus tree Aug 26 14:09:56 ogra: hi. back to my question on neon. should we append -neon on the package name or on the version? Aug 26 14:10:08 package name, definitely Aug 26 14:10:35 if we cant do runtime detection with hwcaps put it in the name Aug 26 14:10:47 lool: yes indeed, this is exactly what is written in my commit. Aug 26 14:11:01 persia: thanks! Aug 26 14:11:04 ogra: Aug 26 14:11:06 ES1.0 Aug 26 14:11:06 yes, but which version Aug 26 14:11:14 mpoirier: Eh indeed Aug 26 14:11:20 lag, wrong context :P Aug 26 14:11:22 mpoirier: Sorry, got confused by ogra Aug 26 14:11:27 :) Aug 26 14:12:06 ogra: Who is developing the x-loader for ES2.0 Aug 26 14:12:12 TI Aug 26 14:12:26 lag, did you try with the linaro uboot ? Aug 26 14:12:43 ogra: Yea Aug 26 14:12:51 didnt work either i guess Aug 26 14:12:59 ogra: It's okay, rsalveti Is going to sort me out with new binaries Aug 26 14:13:10 well, we need new packages too Aug 26 14:13:24 ogra: sure, as we need a new kernel Aug 26 14:13:42 ogra: it'd be good to wait sakoman get it working with es2 Aug 26 14:13:42 i thought the current kernel can run on both Aug 26 14:13:59 only the one we'll get with the next commit cant Aug 26 14:14:29 (thats how i understood it) Aug 26 14:15:17 ogra_cmpc: lag: http://rsalveti.net/pub/ubuntu/kernel/maverick/es2/ Aug 26 14:15:48 sweeeet, my es2 just arrived :D Aug 26 14:15:58 davidm: ^ :-) Aug 26 14:16:10 lucky guy Aug 26 14:17:28 rsalveti, good, I knew it was on a truck heading to you Aug 26 14:18:51 rsalveti: which kernel source did you build your kernel from? Aug 26 14:19:18 ndec: lag: ogra: http://kernel.ubuntu.com/git?p=rsalveti/ubuntu-maverick.git;a=shortlog;h=refs/heads/rsalveti-ti-omap4-es2 Aug 26 14:19:26 I basically applied the patches from Aug 26 14:19:26 http://gitorious.org/pandaboard/kernel-omap4/commits/L24.8_panda_es2.0 Aug 26 14:19:31 on top of our kernel Aug 26 14:19:40 also some extra display patches Aug 26 14:20:54 rsalveti: ok.. sebjan has sent a more official patch set to cooloney. not sure if you're aware of that. our patchset is here http://dev.omapzoom.org/?p=integration/kernel-ubuntu.git;a=shortlog;h=refs/heads/for-ubuntu-2.6.35. it should be same content as the initial patchset but with many cleaning in the patches Aug 26 14:21:28 ndec: thx Aug 26 14:21:31 rsalveti: and this one is more likely to become the 'official' branch in maverick/ti-omap Aug 26 14:21:33 rsalveti: i am working on it Aug 26 14:21:37 Where's my email? Aug 26 14:21:40 cooloney: thx for? Aug 26 14:21:47 ndec: sure, but this is the 2.6.35 tree, that we're waiting cooloney do review and etc, but thanks for pointing that Aug 26 14:22:01 ndec: will this tree work with es1 and es2? Aug 26 14:22:01 lag: i just got it today. Aug 26 14:22:03 or just es2? Aug 26 14:22:22 cooloney: oh, so this is the "final" tree? Aug 26 14:22:23 cooloney: sebjan said he'd sent it to both of us? Aug 26 14:22:28 cooloney: Are you doing it then? Aug 26 14:22:31 ogra: lag: i see you are discussing linaro uboot. are you planning to use this one instead of the current one? Aug 26 14:22:48 ndec, already switced, yes Aug 26 14:23:00 rsalveti: it depends on what 'final' means ;-) Aug 26 14:23:04 awesome, with all latest hdmi fixes Aug 26 14:23:22 rsalveti: yes. including support for multiple FBs Aug 26 14:23:38 ndec: I mean the point that cooloney can just review and get the tree :-) Aug 26 14:23:40 ogra_cmpc: oops... didn't know that. how does that work? Aug 26 14:24:05 rsalveti: cool... i thought that final means that there was no more issue with OMAP4... Aug 26 14:24:14 ndec, lacking an es2 i cant tell yet but anything will be better than the 1.1.4 one Aug 26 14:24:30 ndec: it's working fine at the moment Aug 26 14:24:33 lag: sebjan doesn't finished it Aug 26 14:24:37 based on sakoman's work Aug 26 14:24:43 ogra_cmpc: so the panda uboot support was merged in the mainline uboot? who did that? Aug 26 14:24:45 ndec, 1.1.4 has a broken vfat driver Aug 26 14:24:53 ndec, sakoman Aug 26 14:24:56 cooloney: ndec just said you've been sent the email? Aug 26 14:25:13 lag: no, i didn't sent email Aug 26 14:25:28 pm215: Hey there Aug 26 14:25:29 lag, ndec only pointed to the work branch of sebjan Aug 26 14:25:39 without saying its done Aug 26 14:25:39 pm215: ^P/^N or /win 2 to switch windows? Aug 26 14:25:46 lool: hello. I see I'm now in both channels with a hopelessly confusing UI :-) Aug 26 14:26:16 ndec: cooloney: should this kernel be compatible with both es1 and es2? Aug 26 14:26:25 the new one, based on 2.6.35 Aug 26 14:26:30 pm215, Alt+2 might work also Aug 26 14:26:44 rsalveti: sebjan said it's not ready for both es1 and es2. Aug 26 14:26:51 but he is working on it. i think Aug 26 14:26:58 cooloney: cool Aug 26 14:27:03 i dont think es1 support is wanted Aug 26 14:27:07 rsalveti: only 2.0 for now. sebjan is trying to see if it can work on es1 as well.. Aug 26 14:27:15 at least thats what i understood in the call Aug 26 14:27:16 ogra_cmpc: sure, but just wanted to confirm :-) Aug 26 14:27:26 ndec: What's the difference between "http://dev.omapzoom.org/?p=integration/kernel-ubuntu.git" and "git://dev.omapzoom.org/pub/scm/integration/kernel-omap4.git" Aug 26 14:27:30 ^P/^N//win/alt+2> none of those do anything, I'm afraid. Never mind. Aug 26 14:27:33 i've already prepared a omap4 branch based on sebjan's branch Aug 26 14:27:47 lag and rsalveti http://kernel.ubuntu.com/git?p=roc/ubuntu-maverick.git;a=shortlog;h=refs/heads/ti-omap4 Aug 26 14:27:53 cooloney: on current sebjan's tree? Aug 26 14:28:01 pm215: ctrl-1/ctrl-2? /win 2 might be keeping you in the same window, try /win 1 Aug 26 14:28:09 cooloney: awesome! Aug 26 14:28:14 rsalveti: sebjan is still working on it. so we need wait for a while Aug 26 14:28:16 * rsalveti preparing the build machine Aug 26 14:28:36 lag: kernel-omap4.git is TI BSP team tree, kernel-ubuntu.git is our tree (sebjan's tree) that we use as a staging area between TI BSP and Ubuntu tree Aug 26 14:28:41 rsalveti: i cross compiled it on my local machine Aug 26 14:28:46 ctrl-1/ctrl-2> nope, no effect. /win 1 says "Unknown command: WIN" Aug 26 14:28:52 pm215: If you want to see a quote of the day /qu erat demonstrandum Aug 26 14:29:05 the L24.9 gonna some issue about ASOC codec driver Aug 26 14:29:22 hm, ok Aug 26 14:29:23 so currently i disabled that ASOC codec drivers Aug 26 14:29:27 lag: our tree has some decent level of cleaning (400 patches removed, 3700 checkpatch errors removed) Aug 26 14:29:32 This joke works better in French, with "/qui est là" Aug 26 14:29:34 prpplague: didn't you do some work on that? Aug 26 14:30:14 lool: even in French I don't understand the joke... Aug 26 14:30:18 ogra_cmpc: another topic, how is today's image? Aug 26 14:30:26 rsalveti, still bad Aug 26 14:30:28 just download them Aug 26 14:30:34 ogra: what is broken? Aug 26 14:30:49 jasper, oem-config etc Aug 26 14:30:52 want to test the gles stuff with sgx and efl Aug 26 14:30:58 and latest clutter and etc Aug 26 14:31:02 ndec: Excellent news Aug 26 14:31:13 ndec: Well, IRC clients have an easter egg; you type /qui est là, and it shows you a quote Aug 26 14:31:16 lag: well... thx sebjan... Aug 26 14:31:17 ogra: ouch hehe :-) Aug 26 14:31:25 ndec: I already have Aug 26 14:31:29 ndec: Yesterday :) Aug 26 14:32:13 Apparently, this joke only worked back in the days I was at school Aug 26 14:32:18 rsalveti, the build queue was stuck for most of the day so the fixes arent built yet Aug 26 14:32:30 or in case of oem-config not even uploaded Aug 26 14:32:32 Many easter eggs got polished out as more folks use the software, unfortunately. Aug 26 14:33:00 That one is quite hard to polish out Aug 26 14:33:08 (now I have a sane UI with the unfortunate effect of having to use two nicks...) Aug 26 14:35:22 ogra: hm, ok Aug 26 14:37:27 lool, hey ! Aug 26 14:37:38 lool, where is the bzr commit for your flaash-kernel change = Aug 26 14:37:40 ? Aug 26 14:40:17 Hmm I thought this was using the package imports Aug 26 14:40:22 lag: cool, so you're pushing the xM mmc fixes :-) Aug 26 14:40:32 Attempting to Aug 26 14:40:39 lag: I'm using it already for days, and it's working nicely Aug 26 14:40:43 lool, it has a tree mentioned in debian/control :) Aug 26 14:41:09 I'll fix that Aug 26 14:41:13 thanks Aug 26 14:41:25 its my working tree for debian merges too Aug 26 14:41:27 rsalveti: Yes, I tested them this morning Aug 26 14:41:34 so it would be good to keep it consistent Aug 26 14:42:06 lag: nice Aug 26 14:42:40 Grmpf, import-dsc doesn't work with native packages Aug 26 14:43:13 lool, just push your changes to the tree, what are you doing ? Aug 26 14:44:14 I'm using the canonical way of importing a dsc into a branch! Aug 26 14:44:32 sigh, cant you just leave it as it is and commit your changes ? Aug 26 14:44:48 ogra: Dude, I did already Aug 26 14:44:56 ah, then its fine :) Aug 26 14:45:05 I don't understand why you care how I do it Aug 26 14:45:17 i dont, as long as it ends up in the right tree :) Aug 26 14:46:05 i thought you were fiddling with trees here ... sorry Aug 26 14:46:28 and thanks for the fix :) Aug 26 14:46:36 * ogra strikes it from his TODO Aug 26 14:48:02 lool, btw, does linaro care about kirkwood ? (there are some pending debian changes for flash-kernel i didnt actually plan to merge unless someone really needs them) Aug 26 14:48:39 i got the 2.6.35 kernel boots on my panda now Aug 26 14:48:42 http://pastebin.ubuntu.com/483996/ Aug 26 14:49:12 cooloney, that becomes intresting if line 8 changes ;) Aug 26 14:49:18 cooloney: cool, at your es1? Aug 26 14:49:52 too bad, i don't have the es2 HW Aug 26 14:50:01 rsalveti: yeah, mine is ES1 Aug 26 14:50:10 cooloney: nice, I will test on my es2 Aug 26 14:50:44 great, gonna sleep now Aug 26 14:50:57 cooloney: see ya! Aug 26 14:51:12 ogra: still using 512MB on panda? I thought that 1GB was running already Aug 26 14:51:43 hrw, not yet Aug 26 14:57:24 help ? Aug 26 14:57:54 whoops, can't drive my irc client still :-) Aug 26 15:02:48 lool, so why are the updates of my falsh-kernel branches all failing now ? Aug 26 15:02:59 what the heck did you do ? Aug 26 15:04:02 GRRRR !!!! Aug 26 15:07:04 ogra: if you are brave, you could try porting the 1gb patch from es2 branch.. Aug 26 15:07:26 http://gitorious.org/~robclark/pandaboard/robclarks-x-loader/commits/omap4_panda_es2.0-1gb Aug 26 15:08:06 robclark, why porting ? shouldnt i be able to just build that one ? Aug 26 15:08:23 well.. depends on how aligned theory is with fact ;-) Aug 26 15:08:26 we dont plan to support es1 anyway Aug 26 15:08:30 in *theory* it should work Aug 26 15:08:55 i'll try as soon as my es2 arrives, if its good i'll just rebase the package on that one ;) Aug 26 15:12:48 robclark: nice, will try this one Aug 26 15:13:31 I've been using it for a couple weeks on es2.. Aug 26 15:13:33 ogra: I did what I noted in the changelog: moved to the package import branches... Aug 26 15:13:51 and in theory same x-loader should work on es1... but I haven't tested that myself, so who knows Aug 26 15:13:52 lool, well, and i have to rebade about ten development branches now Aug 26 15:13:56 *rebase Aug 26 15:14:02 ogra: "rebase", why rebase? Aug 26 15:14:05 it's the same branch Aug 26 15:14:22 I replaced the package import branch with what was ~ubuntu-core-dev/flash-kernel/ubuntu Aug 26 15:14:43 i have to touch them still, you just wiped out teams work branch Aug 26 15:15:01 even though i asked you not to Aug 26 15:15:14 thats not nice ... but i'll point my branches to the new one Aug 26 15:15:22 ogra: I don't understand what you're speaking about Aug 26 15:15:30 You can merge them as you could before Aug 26 15:15:55 the master branch of my (and likely others) dev brances is gone from LP Aug 26 15:16:52 ogra: It's not, it just moved Aug 26 15:17:08 ogra@osiris:~/Devel/branches/flash-kernel/ubuntu$ bzr pull Aug 26 15:17:08 Using saved parent location: bzr+ssh://bazaar.launchpad.net/~ubuntu-core-dev/flash-kernel/ubuntu/ Aug 26 15:17:08 bzr: ERROR: Not a branch: "bzr+ssh://bazaar.launchpad.net/~ubuntu-core-dev/flash-kernel/ubuntu/". Aug 26 15:17:27 bzr pull --remember lp:ubuntu/flash-kernel Aug 26 15:18:14 lool, yes, i have to do that on all my branches now, at least note the new location in your changelog entry next time so people using it know where to pull from Aug 26 15:18:29 ogra: it's the standard location... Aug 26 15:18:36 that was a totally unnecessary change that causes extra work Aug 26 15:18:39 ogra: Wouldn't I do that, you would continue using the old location Aug 26 15:18:49 that was my plan yes Aug 26 15:18:56 What's unnecessary is having two branches, and this will prevent people from not committing before upload Aug 26 15:19:41 ... Aug 26 15:20:21 It doesn't prevent it. It just does the right thing anyway. Aug 26 15:20:34 ogra: There, I've pushed lp:~ubuntu-core-dev/flash-kernel/ubuntu again and marked it as abandoned Aug 26 15:20:43 But it breaks if there are unuploaded commited changes in the branch. Aug 26 15:20:44 ogra: please don't push to this "old" branch anymore, thanks Aug 26 15:21:15 (which is a feature, not an impediment: stuff ought be uploaded) Aug 26 15:29:33 ogra: Look, it sounds like you're pissed off; I'm trying to do the right thing by avoiding the mistake I made in the future; I thought it was obvious how to adjust your local branches for that, but it apparently wasn't as easy as I thought it was, so I've pushed something back will should help a bit, but wont force you into the right branch; do you understand why it's better to move to the new branch? Aug 26 15:30:18 lool, so will you run around and move all of cjwatsons branches too (i know he, like me prefers to use branches the old way) ? Aug 26 15:30:53 So, this doesn't feel like constructive criticism (although I really like they way you've raised it early and publically) Aug 26 15:31:19 Would you both agree that in future it's best to coordinate between common uploaders of a package before migrating a branch to the UDD home? Aug 26 15:31:23 and can you understand that people dont like to be "forced" intop a new way in the middle of the busiest time before a freeze where a branch might be used atm for fixes ? Aug 26 15:31:52 Quite frankly, I didn't expect that there were any pending merges; I did look for pending merge requests and didn't see any Aug 26 15:32:06 Would I have seen any, I wouldn't have moved before dealing with these Aug 26 15:32:09 lool, i would have moved the branch with the merge in N, its just not cool to do it if someone asks you not to Aug 26 15:32:36 hey all you early developers for the Panda, i'm trying to finalize the features for the Bamboo accessory board for the panda Aug 26 15:32:36 but it happened now, so lets forget about it Aug 26 15:32:48 if any has hardware requests for additions to the bamboo, now is the time Aug 26 15:33:42 ogra: So the additional work it creates is this --remember lp:ubuntu/flash-kernel thing? Aug 26 15:34:05 I dont want to defer dealing with problems when I can Aug 26 15:34:10 lool, that too, worse is that you simply ignored my request Aug 26 15:36:37 ogra: where did you request what? Aug 26 15:36:47 prpplague, do you have a list of stuff thats in already ? so we can see what might be missing ? Aug 26 15:37:35 lool, just push your changes to the tree, what are you doing ? Aug 26 15:37:37 I'm using the canonical way of importing a dsc into a branch! Aug 26 15:37:37 sigh, cant you just leave it as it is and commit your changes ? Aug 26 15:37:43 ogra: accessory board that provides: second sd/mmc, 2 user leds, 2 user buttons, reset button, power led, battery backed RTC, built in usb->rs232 for console power, 2 additional USB host ports, and an abs plastic case Aug 26 15:37:49 ogra: creating a wiki page now Aug 26 15:37:56 ogra: Dude, I did already Aug 26 15:37:56 ah, then its fine :) Aug 26 15:37:56 I don't understand why you care how I do it Aug 26 15:37:56 i dont, as long as it ends up in the right tree :) Aug 26 15:37:56 i thought you were fiddling with trees here ... sorry Aug 26 15:37:57 and thanks for the fix :) Aug 26 15:38:31 prpplague, nothing strikes me on first sight Aug 26 15:38:57 prpplague: yep, sounds ok already :-) Aug 26 15:43:50 prpplague, Any chance of eSATA? Aug 26 15:44:07 I was going to ask... Aug 26 15:44:09 persia, heh or inflatable helicopters Aug 26 15:44:27 persia: still working on that request Aug 26 15:44:45 * persia is *much* more interested in eSATA than inflatable helicopters. One can glue the abs case to the baloon easily enough, but ... Aug 26 15:44:55 prpplague, heh, OK :) Aug 26 15:45:19 heh Aug 26 15:45:39 http://www.elinux.org/Panda_Bambo Aug 26 15:47:13 * persia is all sorts of excited: the ABS case is the best feature Aug 26 15:51:53 i've updated the page with the color and dimensions of the case Aug 26 15:51:57 * GrueMaster assumes the case design includes externally accessable buttons? Aug 26 15:51:59 feel free to leave comments there Aug 26 15:52:04 GrueMaster: yea Aug 26 15:52:27 Nice. Aug 26 15:52:28 GrueMaster: front panel will have the power led, 2 buttons, 2 leds, the sd/mmc slot Aug 26 15:52:39 GrueMaster: and the two addtional usb host ports Aug 26 15:52:48 My main concern about the case is that it be stackable with other stuff: bare boards are hard to feel comfortable about sticking on a shelf with other stuff and attaching to a KVM. Aug 26 15:52:50 Very nice. Aug 26 15:52:57 GrueMaster: back will be the back of the panda, with a usb slave port for the console Aug 26 15:53:15 persia: yea the case is stackable Aug 26 15:53:31 Excellent. Aug 26 15:53:54 * GrueMaster might have to sneak the credit card away from the wife again soon. Aug 26 15:54:16 estimated cost will be $55.00 Aug 26 15:54:32 (plus or minus $5) Aug 26 15:54:52 Excellent. That fits into my monthly turn & burn budget. Aug 26 15:55:21 (i.e. not something I need to wait until I have the funds for). Aug 26 15:58:23 prpplague: nice, not that expensive Aug 26 15:58:46 prpplague: when will be available? Aug 26 15:58:47 y Aug 26 16:01:59 hrw: should be available on the same day panda's become available Aug 26 16:03:53 nice! Aug 26 16:20:10 lunch time! Aug 26 16:43:45 Gah, seeing python issues with package updates. Haven't updated libc6 yet, so there is hope this isn't a bug. Aug 26 17:18:40 rsalveti, it wont work completely, the .1 build i'm doing should only fix the reboot issue Aug 26 17:18:56 ogra: hm, ok Aug 26 17:18:58 rsalveti, oem-config/ubiquity was still not uploaded so it will fail Aug 26 17:19:23 Still? I thought the fix went in last week? Aug 26 17:19:41 yes, into the branch Aug 26 17:20:24 i didnt want to interfere with the ongoing work on the installer team so i waited for them to upload (the tree might have half breeded code in there) Aug 26 17:22:23 s/on/of/ Aug 26 19:13:26 hi Aug 26 19:14:10 what is the difference between Aug 26 19:14:13 http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100826.1/maverick-preinstalled-netbook-armel+omap4.img.gz Aug 26 19:14:25 http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100826.1/maverick-preinstalled-netbook-armel+omap.img.gz Aug 26 19:15:50 default kernel, u-boot and x_loader basically Aug 26 19:18:22 ok Aug 26 19:18:49 dcordes: omap4 is for panda/blaze, omap is for Beagle Aug 26 19:19:09 rsalveti: so there are no differences in the userspace binary compilation ? Aug 26 19:19:21 (should be omap3 in my opinion - just to keep from confusing people with older platforms). Aug 26 19:19:28 None. Aug 26 19:19:33 agree Aug 26 19:19:34 ok guys thanks Aug 26 19:19:59 yeah that would make sense to change omap into omap3 Aug 26 19:20:03 +1 Aug 26 19:20:42 after people loved ubuntu on the htc hd2 I am going to create a new 'release' Aug 26 19:21:18 cool Aug 26 19:23:49 just to make sure everyone knows(i announced earlier today) i'm taking feedback and comments on the features for the bamboo board - http://www.elinux.org/Panda_Bambo Aug 26 19:25:11 prpplague: love the case ;) Aug 26 19:25:24 dcordes: hehe, i get 50/50 on that Aug 26 19:25:47 the case is ugly Aug 26 19:31:42 prpplague: despite from having that classical box, what is it ? Aug 26 19:32:56 dcordes: accessory board for the panda Aug 26 19:33:46 prpplague: is it bamboo or bambo ? Aug 26 19:33:53 (page title says bambo) Aug 26 19:34:32 yea someone just pointed that out Aug 26 19:34:37 * prpplague movies the page Aug 26 19:36:02 you should add a small intro. bamboo is an expansion board for the [[panda]] device Aug 26 19:36:06 something like that Aug 26 19:37:51 downloading that maverick netbook rootfs with 20kB/s Aug 26 19:37:53 * dcordes sighs Aug 26 19:38:05 university is charging too much for such slow dormitory net Aug 26 19:44:06 dcordes: will do, i just assumed the folks who have a panda would understand Aug 26 19:44:14 GrueMaster, the omap vs omap4 naming was chosen with the idea in mind that we might have a single omap image at some point once linaros unification work for kernel and u-boot is done Aug 26 19:44:37 GrueMaster, you might have noticed that we use the same scheme for the kernel packages Aug 26 19:44:39 My only point is that we are not supporting omap2 hw. Aug 26 19:45:04 afaik thats on linaros plans too (i might misremember though) Aug 26 19:45:16 at least kernel wise Aug 26 19:45:17 aha Aug 26 19:45:35 ogra_cmpc: how is the rootstock coming along ? Aug 26 19:45:57 dcordes: I'm using it right now, working fine :-) Aug 26 19:46:02 dcordes, i gave it to rsalveti and since it made a quamntum jump in quality ;) Aug 26 19:46:12 :P Aug 26 19:46:25 it is because I joined the bugtracking system Aug 26 19:46:26 just need some ui fixes Aug 26 19:48:08 ogra_cmpc: I'm down to the last 27 package updates to bring A3 current. Still booting into X with gdm & netbook-launcher-efl so far. Most critical path packages are updated now. Aug 26 19:48:25 whats missing ? Aug 26 19:48:38 * ogra_cmpc doesnt get why we have that issue Aug 26 19:49:11 GrueMaster: cool, good to know Aug 26 19:49:22 worrying though Aug 26 19:49:43 * ogra_cmpc would perfer a clear pointer to a package thats broken Aug 26 19:50:28 Woohoo. leann posted kernel with XM fix. Aug 26 19:50:33 yeah Aug 26 19:51:37 sadly i assumed she would just take the branch and upload it, but she cherrypicked ... so the NAND fix is still waiting until after beta Aug 26 19:52:00 * ogra_cmpc didnt think about that when discussing the freeze exception :( Aug 26 19:52:28 I thought that was already in. Aug 26 19:52:39 in the tree, yes Aug 26 19:52:43 ogra_cmpc: what nand fix? Aug 26 19:52:52 for beagle Aug 26 19:52:54 hm, so isn't it released already? Aug 26 19:52:57 rsalveti, the one that makes mtd work again Aug 26 19:53:08 i didnt see it in any upload yet Aug 26 19:53:19 mine works, I even get i/o errors Aug 26 19:53:28 and tohdays only has the cherry picked mmc fix Aug 26 19:53:32 * rsalveti looks at the kernel tree Aug 26 19:54:05 I think it is already there. I'm looking at today's image on beagle and seeing /dev/mtd* Aug 26 19:54:28 oh, hmm, then i might be blind and have missed it in the changelog Aug 26 19:54:51 That's highly possible. :P Aug 26 19:55:53 * ogra_cmpc wonders if he has a misbehaving proxy .... i saw the 20100826.1 image on cdimage a few mins ago, now it seems to be gone again Aug 26 19:56:49 ogra_cmpc: bug 608266 Aug 26 19:56:52 fix released already Aug 26 19:56:53 Launchpad bug 608266 in linux (Ubuntu Maverick) (and 1 other project) "[regression] no more /dev/mtdblock devices on omap3 in maverick (affects: 1) (heat: 104)" [Medium,Fix released] https://launchpad.net/bugs/608266 Aug 26 19:56:58 ah, ik Aug 26 19:57:05 with 2.6.35-16.22 Aug 26 19:57:08 :-) Aug 26 19:57:10 consider me officially blind then Aug 26 19:57:18 I even told you that I was getting i/o errors Aug 26 19:57:19 :) Aug 26 19:57:32 :-) Aug 26 19:57:33 right, i thought that was with your own kernel Aug 26 19:57:38 Helps not to look through a half full beer glass. :p Aug 26 19:57:45 i'm never sure what you use over there :P Aug 26 19:57:49 haha :-) Aug 26 19:57:50 true Aug 26 19:58:01 I try to use our kernel with possible fixes Aug 26 19:58:06 GrueMaster, i had my last beer in prague :) Aug 26 19:58:11 so we can get it pushed later Aug 26 19:58:18 ouch Aug 26 19:58:19 GrueMaster: I can offer a 1/3 full whine glass if somebody is interested Aug 26 19:58:21 this saddens me. Aug 26 19:58:25 2 Aug 26 19:58:27 * dcordes always ready to help the community. Aug 26 19:59:01 no, its healthy, keep my liver happy and makes me bear more in orlando ;) Aug 26 19:59:22 dcordes, you whine into a glass ? Aug 26 19:59:35 now *thats* saddening ;) Aug 26 19:59:57 * GrueMaster only whines about an empty glass. Aug 26 19:59:59 * dcordes everybody lol for the funniest typo 2010 Aug 26 20:00:20 that got me so down I need to refill Aug 26 20:00:26 that year isnt over ... and my bad humor isnt either ;) Aug 26 20:01:55 hmm, where is the .1 image gone .... looking above dcordes saw it too (according to the url) Aug 26 20:03:20 I'm pulling it ok. Aug 26 20:03:36 i get a 404 for the dir Aug 26 20:03:41 http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100826.1/maverick-preinstalled-netbook-armel+omap4.img.gz.zsync Aug 26 20:04:01 odd Aug 26 20:04:17 i started a zsync upstairs, not sure it still runs Aug 26 20:05:42 yeah, seems to be done Aug 26 20:06:29 ogra_cmpc: .1 image ? Aug 26 20:06:43 * dcordes is downloading http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100826.1/maverick-preinstalled-netbook-armel+omap4.img.gz Aug 26 20:07:26 dcordes, right, but http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100826.1/ gets me a 404 suddenly Aug 26 20:07:39 and http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/ doesnt have a .1 Aug 26 20:07:57 my download worked fine as well Aug 26 20:08:26 oh, there is is now Aug 26 20:08:31 irritating Aug 26 20:09:05 * ogra guesses our sync process from the builder to the webserver is somehow strange Aug 26 20:13:20 ogra: try to request the url 10 times Aug 26 20:13:25 then it works Aug 26 20:13:34 well, it seems stable now Aug 26 20:13:39 I always have weird problems with that server *_* Aug 26 20:13:52 err, no, it doesnt, its gone again Aug 26 20:14:02 * ogra checks the server itself Aug 26 20:14:33 here it works but weirdo style as always Aug 26 20:14:48 don't understand the underlying network mechanisms well enough to say what it is Aug 26 20:14:49 yeah, its definitely on the server Aug 26 20:15:06 all I know is sometimes the sites there don't show up at a ll Aug 26 20:15:34 well, it would be ok if it was just delayed, that on/off stuff is weird Aug 26 20:15:35 but when I keep hitting enter in the browser's url bar it works Aug 26 20:15:36 :) Aug 26 20:15:56 I think when I just let it time out it will also go 404 ... Aug 26 20:25:08 rsalveti: I think I found a fix for the panda u-boot reset command issue Aug 26 20:25:35 sakoman: nice, what was the issue? Aug 26 20:26:14 OMAP4 seems to want a different bit written to trigger the reset Aug 26 20:27:02 hm, makes sense Aug 26 20:27:35 TI u-boot was using bit 1 (i.e. 0x02), but I think bit 0 (0x01) is right Aug 26 20:28:07 it also uses a different address than OMAP3, but I already took care of that Aug 26 20:28:47 cool Aug 26 20:29:02 so while 0x02 worked for all earlier OMAP3s, OMAP36XX/37XX need 0x04, and OMAP4 needs 0x01 Aug 26 20:29:21 got it Aug 26 20:29:22 so I need restructure the code/headers a bit Aug 26 20:29:50 I'll revise the patch I posted previously to also take care of OMAP4 Aug 26 20:31:17 rsalveti: little things like this take way too much time, what with building & testig on multiple boards and reviewing multiple 3000 page TRMs! Aug 26 20:31:40 ouch :-) Aug 26 20:37:47 how does the current preinstalled image decide if it is starting on beagleboard or beagleboard xm? Aug 26 20:40:57 suihkulokki: the omap image should work on both Aug 26 20:41:16 not with today's image, but that's going to be fixed for tomorrow Aug 26 20:43:10 rsalveti: I know it support both :) my question is what does it _do_ to figure out which one it is running on Aug 26 20:43:53 suihkulokki: http://kernel.ubuntu.com/git?p=rsalveti/ubuntu-maverick.git;a=blobdiff;f=arch/arm/mach-omap2/board-omap3beagle.c;h=623c352f14695a3a01545db9a1307adc5d11e21d;hp=5501f310a9d0c34df4516d6efd0f5e5e25ca2960;hb=52e9cdbb825eae0f3f75550adacebdc36303700b;hpb=978e830c47ca5de5824ddf3ba9f7d3571da765a7 Aug 26 20:44:02 an example Aug 26 20:44:46 then you can also check by the cpu type Aug 26 20:45:16 ok, so there is a gpio to read. thanks. Aug 26 20:47:21 yup Aug 26 20:47:28 Hmmm. I'm thinking ureadahead is the cause of our current issues. Testing that theory now. Aug 26 20:52:59 GrueMaster: hm, it shouldn't, unless we got a new release Aug 26 20:53:16 Since Alpha 3, yes. Aug 26 20:53:28 you can disable it, mv /etc/init/ureadahead.conf /etc/init/ureadahead.disabled Aug 26 20:53:32 doesn't make a differente Aug 26 20:53:38 I mean, a major release Aug 26 20:53:44 *difference Aug 26 20:54:19 I updated that, rebooted, then updated all of network manager packages, rebooted. Not I am getting corrupt filesystem and hangs, but not getting past uInitrd. Aug 26 20:54:33 ouch Aug 26 20:56:53 rsalveti, i think parts of ureadahead start in initrd Aug 26 20:57:15 hm, I'm not sure Aug 26 20:57:32 * ogra_cmpc neither ... thats why i said i think :) Aug 26 20:57:33 disabling it from init is enough to get it removed and avoid oom Aug 26 20:57:37 * ogra_cmpc goes to check Aug 26 20:57:53 so probably doesn't run inside uinitrd Aug 26 20:57:56 but :-) Aug 26 20:57:58 please check Aug 26 20:59:03 rsalveti, right, only from init Aug 26 20:59:34 It does. I get "init: ureadahead main process (208) terminated with status 5" before it mounts rootfs. Aug 26 20:59:44 thats fine Aug 26 21:01:46 GrueMaster, that doesnt mean a thing about the initrd though, there could be a process that hands over data after the initrd for example Aug 26 21:02:25 thats why i rather look at the code to make sure :) Aug 26 21:06:16 rsalveti: I push a revised patch that fixes both 37XX and OMAP4 u-boot reset command issues: Aug 26 21:06:20 http://www.sakoman.com/cgi-bin/gitweb.cgi?p=u-boot.git;a=commitdiff;h=56473fba8010c5def9ed778da4bcb4455d265b54 Aug 26 21:06:54 GrueMaster: ogra_cmpc: exit 5 means that the trace failed, for some reason Aug 26 21:07:19 rsalveti, according to Keybuk thaqts because its MMC Aug 26 21:07:52 i asked him about it a few months ago, he said we shouldnt worry about it, it properly exits Aug 26 21:08:21 sakoman: cool :-) Aug 26 21:08:38 ogra_cmpc: hm... Aug 26 21:08:43 rsalveti, i thinki we should talk to him soon so he can explain the possible benefits for us Aug 26 21:08:49 i dont really see any Aug 26 21:08:50 if any Aug 26 21:08:58 snap :) Aug 26 21:09:11 ogra_cmpc: sure, we first need to know when he is going to release the new major release Aug 26 21:09:19 ureahead 2 I guess Aug 26 21:09:29 likely not during beta freeze Aug 26 21:09:41 but probably for maverick Aug 26 21:09:47 thats what he said Aug 26 21:09:51 hm, ok Aug 26 21:09:59 but sure, lets talk to him later Aug 26 21:10:16 the question is if it will work any better then :) Aug 26 21:10:20 So, ureadahead is good because RAM is faster than flash is faster than {e,}MMC-mitgated flash. Aug 26 21:10:41 persia, is it ? you still need to read from the SD Aug 26 21:10:44 persia: doesn't make a difference for beagle Aug 26 21:10:46 So getting stuff into the page cache in advance always makes boot faster (and ureadahead doesn't read anything that isn't read anyway) Aug 26 21:10:59 sd is so slow that it really doesn't make a difference Aug 26 21:11:05 ogra, The point is that you don't have to *wait* on reading from the SD. Aug 26 21:11:06 the seek is not the issue here Aug 26 21:11:07 persia, only if things end up in the cache in time Aug 26 21:11:08 as a normal disk Aug 26 21:11:29 rsalveti, that's because the beagle doesn't meet the minimum ram requirements (384MB) Aug 26 21:11:38 384?! Aug 26 21:11:42 Yep. Aug 26 21:11:43 yeah Aug 26 21:11:55 funny number Aug 26 21:11:59 Been 384 since Breezy or so. Aug 26 21:12:03 3x128. Aug 26 21:12:09 384M is the minimum the x86 livecd works in Aug 26 21:12:12 I know, but still funny Aug 26 21:12:43 persia: ogra_cmpc: I'd like to test at a valid xm (512mb) to see if changes anything Aug 26 21:12:48 right Aug 26 21:12:51 and compare the bootchart Aug 26 21:12:53 i really doubt it Aug 26 21:12:57 2 Aug 26 21:13:22 the IO is to slow to gain any benefit form it is my impression Aug 26 21:13:27 Well, any suggestions on where to go? Aug 26 21:13:46 outside into the sun ? Aug 26 21:13:55 It helps best when the IO is slow, because it means *not* spending time without full IO bandwidth in use. Aug 26 21:14:06 lol Aug 26 21:14:15 persia: doesn't help much Aug 26 21:14:25 persia, how so if your bandwith is saturated all the time anyway Aug 26 21:14:28 Mind you, it's Keybuk's code competing with Keybuk's code: ureadahead is only advantageous when upstart fails to use all available IO. Aug 26 21:14:28 helps a *lot* with normal disks, because the seek time Aug 26 21:14:31 not much from sd Aug 26 21:15:08 at least didn't show any difference on my bootchart when caching 64 mb Aug 26 21:15:16 even more Aug 26 21:15:27 but, still waiting for a xm test Aug 26 21:15:28 ogra, If upstart can saturate the IO without ureadahead, then it makes no difference, but Keybuk isn't trying to saturate IO with upstart because he assumes ureadahead will do that (ureadahead is specifically designed to saturate the IO) Aug 26 21:15:28 rsalveti: you guys have a preference where you want to have a pandaboard revision in sysfs? Aug 26 21:15:32 can test on panda later Aug 26 21:15:45 prpplague: hm... interesting question Aug 26 21:16:00 ogra: GrueMaster: ^ Aug 26 21:16:19 I have nrp. Aug 26 21:16:21 no idea, really, i personally dont have one Aug 26 21:16:33 as long as i know where to look in the end :) Aug 26 21:16:58 prpplague: my question is more with which device do you think of pluging this file into? Aug 26 21:17:04 Where does this information live for other boards. Let's strive towards some consistency throughout the industry. Aug 26 21:17:10 yeah Aug 26 21:17:13 I would suggest looking at existing systems for consistency. Aug 26 21:17:18 proc can be anywhere, but sysfs you need to use a valid device Aug 26 21:17:23 persia: yea that was my question as well, but i can't seem to find any examples Aug 26 21:17:31 probably no examples Aug 26 21:17:36 * rsalveti never saw it Aug 26 21:18:01 likely, else we woldnt parse /proc/cpuinfo for hardware detection Aug 26 21:18:21 (on all armel systems) Aug 26 21:18:43 prpplague: is it related with gpios like beagle? Aug 26 21:19:55 rsalveti: yea Aug 26 21:20:09 rsalveti: same type of config Aug 26 21:20:54 I think that there isn't a standard place (checked 3 arches just now for a variety of HW). I think most folks just enumerate the devices, and don't actually discuss which board is providing those devices. Aug 26 21:20:55 prpplague, currently all tools we use parse the Hardware line of /proc/cpuinfo Aug 26 21:21:02 prpplague: I'm not seeing any specific standard compared to multiple systems I have here. Aug 26 21:21:03 Then we autodetect what we can do based on the devices. Aug 26 21:21:20 So that board mapping becomes fuzzy, based on the set of peripherals exposed. Aug 26 21:21:21 thats what i thought , but i was told by "someone" (not sure who) that canonical wanted a sysfs entry Aug 26 21:21:42 i wouldnt mind one to set a new standard Aug 26 21:21:47 * persia wonders what they were thinking Aug 26 21:21:48 prpplague: hm, for gpio there is the omap-gpio... Aug 26 21:22:07 the /proc/cpuinfo parsing can be really tricky if you miss a tab or space or so Aug 26 21:22:08 maybe we can probe it if we can have access to the gpio values from userspace Aug 26 21:22:23 * rsalveti prefers not touching proc/cpuinfo Aug 26 21:22:25 ogra, Why? I'd rather map devices, so that if someone creates $random_board with the same SoC as a panda, and the same peripheral devices, it gets the same support, without lying about itself. Aug 26 21:22:28 ogra_cmpc: might have to generate a new class device Aug 26 21:22:48 prpplague: could be Aug 26 21:23:07 or letting userspace to decide by probing the gpios Aug 26 21:23:10 /sys/class/ is where it'd belong, but I'd be unhappy to see Ubuntu use it. Aug 26 21:23:17 persia, sure, but having a file thats easier to parse with a predefined entry might make thiings easier Aug 26 21:24:02 ogra, For all the same reasons that I argue against all the embedded development practices that make things easier short-term for folks building one-off solutions, I reject that entirely. Aug 26 22:33:51 GrueMaster: what issue did you have after upgrading the packages? Aug 26 22:34:01 I updated my panda and now it doesn't boot anymore Aug 26 22:34:08 seems like an upstart issue Aug 26 22:41:20 Yep, that's what I am seeing. Aug 26 22:41:40 I updated upstart before ureadahead. Aug 26 22:42:17 The fact that it doesn't appear to be going beyond mounting root is what puzzles me. Aug 26 22:42:41 GrueMaster: it mounts the rootfs here Aug 26 22:42:49 tries to start postfix and hangs Aug 26 22:43:05 disable postfix and now I can only see the hang :-) Aug 26 22:43:28 I'm creating another minimal rootfs with rootstock to see if I didn't mess with anything Aug 26 22:43:32 try downgrading it Aug 26 22:43:33 I had an old rootfs there Aug 26 22:44:33 (either chroot into the SD on an x86 or dpkg -x the older upstart into /mountpoint/of/SD Aug 26 22:44:35 ) Aug 26 22:44:38 yep Aug 26 22:44:43 and debug upstart Aug 26 22:44:50 oh.. how I love it Aug 26 22:44:51 yesah Aug 26 22:45:21 well, lets bug Keybok if its really upstart, he is usually helpful (if he gets online :P) Aug 26 22:45:46 yeah, but first we need to identify if our problem is upstart Aug 26 22:46:06 well, if downgarding the package fixes it ... Aug 26 22:46:19 I have a definite advantage here. Babbage w/ two SD slots. Aug 26 22:46:20 yep Aug 26 22:46:32 lots and lots of issues on our current image :-( Aug 26 22:46:43 yes :( Aug 26 22:47:06 it's going to take a while to be able to test efl with sgx and stuff asac asked us to do Aug 26 22:47:11 * ogra_cmpc goes back to his midnight dinner Aug 26 22:47:18 see ya Aug 26 22:52:32 rsalveti: if you send me the cmdline for rootstock, I can build my own image for the XM here. Much faster as I have my own mirror server. Aug 26 22:54:03 GrueMaster: I'm just finishing putting everything into my sd card, then will dd from it Aug 26 22:54:06 so others can test Aug 26 22:54:29 but if you want, this is how I'm generating it: ./rootstock --fqdn beaglexm-maverick --login ubuntu --password ubuntu --dist maverick --serial ttyS2 --components "main universe multiverse" --seed linux-image-omap Aug 26 22:54:36 use rootstock upstream Aug 26 22:54:44 rsalveti / ogra_cmpc / GrueMaster : http://paste.ubuntu.com/484226/ Aug 26 22:54:53 pandaboard revision reporting Aug 26 22:55:03 it reports it as part of the boot up Aug 26 22:55:18 and has the board revision available under /sys/kernel/pandaboard/board_revision Aug 26 22:55:24 any issues with that? Aug 26 22:57:13 I'd prefer something more generic, like /sys/kernel/board/revision Aug 26 22:57:24 and then inside you'll have pandaboard: 0.1 Aug 26 22:57:25 for example Aug 26 22:57:31 because then we can later use that for beagle Aug 26 23:01:05 rsalveti: Thanks. Aug 26 23:01:17 understood, making the change now Aug 26 23:01:26 prpplague: I have to agree with rsalveti on that one. Aug 26 23:01:35 np Aug 26 23:04:28 two line paste Aug 26 23:04:30 PandaBoard Revision: 001 Aug 26 23:04:51 oops Aug 26 23:04:53 PandaBoard Revision: 001 Aug 26 23:04:58 well darn Aug 26 23:05:30 GrueMaster / rsalveti http://paste.ubuntu.com/484228/ Aug 26 23:05:52 nice. Aug 26 23:06:19 question though, do you want it more human readable, or more parse friendly Aug 26 23:06:20 now to get other arch/platforms to follow suit. Aug 26 23:07:05 i.e. should i just report revision without referencing pandaboard? Aug 26 23:07:32 no, platform/revision is great info. Aug 26 23:08:13 okie dokie Aug 26 23:08:22 * prpplague gets ready to send patch to L24.9 team Aug 26 23:08:31 This way we can differentiate between systems (i.e. Blaze/Panda or Beagle/BeagleXM/Gumstix, etc). Aug 26 23:11:19 I'm assuming this info is gathered through probing the system & using a lookup table of some sort? Aug 26 23:13:26 argh, my system is trashing... very slow Aug 26 23:13:29 opening the link Aug 26 23:13:39 prpplague: cool :-) Aug 26 23:14:20 GrueMaster: 3-bit value on three gpios Aug 26 23:14:48 ok Aug 26 23:14:52 prpplague: if we have a separator like : I guess it's ok for parsers Aug 26 23:15:09 name : version Aug 26 23:15:14 rsalveti: thats what i was thinking Aug 26 23:15:15 or something like that Aug 26 23:16:01 rsalveti: i was thinking Aug 26 23:16:21 rsalveti: something like this Aug 26 23:16:34 boardname - revision : revisionnumber Aug 26 23:16:49 PandaBoard - revision : 001 Aug 26 23:17:13 is it that too much? Aug 26 23:17:22 just leave it Aug 26 23:17:28 PandaBoard : 001 Aug 26 23:18:46 hm Aug 26 23:19:08 don't know, most of the time just the revision should be ok Aug 26 23:19:11 * GrueMaster finds nothing wrong with the pastebin version. Aug 26 23:19:15 like beagleboard: c4 Aug 26 23:19:24 pandaboard: es1 Aug 26 23:19:27 an example Aug 26 23:19:30 yea Aug 26 23:19:43 ok, do the KISS thing Aug 26 23:20:27 :-) Aug 26 23:20:28 always Aug 26 23:22:23 rsalveti: Back to the image issues, I am really having doubts as to what could be breaking the system. I created a manifest of the A3 image and compared it with 20100809 (first failing image). Upstart & ureadahead are unchanged. Aug 26 23:22:43 hm... Aug 26 23:22:52 maybe a dependency between the upstart services Aug 26 23:23:02 waiting for something that doesn't get started Aug 26 23:23:10 debugging upstart would be the way to go Aug 26 23:23:15 testing new theory. We now have a new oem-config apparently. Will flash 20100809 image and update oem-config on it. Aug 26 23:24:24 Two issues that I have seen; latest images fail oem-config and updating from A3 fails to boot. I doubt there is any relation to the two. Aug 26 23:26:06 too many variables and not enough fast test systems. Aug 26 23:26:50 now I can see 20100826.1 Aug 26 23:26:54 will also download it Aug 26 23:26:54 * GrueMaster has spent a frustrating 2.5 weeks on this issue. Aug 26 23:26:59 :-( Aug 26 23:27:06 sd card performance is frustrating Aug 26 23:27:14 Yes. Very. Aug 26 23:27:22 Even on my desktop. Aug 26 23:29:33 GrueMaster: 152K/s of download is also frustrating Aug 26 23:29:55 yes. Which is why I have my own mirror. Aug 26 23:30:38 but this is for the image files Aug 26 23:30:45 I have a cache server around Aug 26 23:30:47 1Gb network throughout the house also helps Aug 26 23:30:50 for normal stuff Aug 26 23:30:59 for sure Aug 26 23:31:10 but an sd card with 3, 4m/s doesn't help Aug 26 23:31:11 haha Aug 26 23:31:19 I mirror both. And retain dailys until final. Aug 26 23:31:52 Not that arm dailys take much room given the pool churn. Aug 26 23:34:04 yep, will start mirroring the image stuff Aug 26 23:34:25 the problem is that my stupid internet provider limits the amount of gb per month Aug 26 23:34:34 I can download only 80gb I guess Aug 26 23:34:46 with a 10m/s link Aug 26 23:35:04 so I avoid doing mirror and downloading only what's needed Aug 26 23:35:24 ouch. Some in the US are doing the same. Fortunately, my ISP has been around since early '90s and has a fat pipe. Aug 26 23:35:41 that's nice Aug 26 23:36:04 with my previous provider I didn't have this issue, but doesn't reach my current city Aug 26 23:36:21 so I had to change to the stupid one Aug 26 23:37:13 I'm currently limited by DSL speeds. Which out here in the sticks is ~6.5M/768k Aug 26 23:37:26 * GrueMaster wishes he could get FIOS. Aug 26 23:37:48 mine is currently 10/1 m/s Aug 26 23:37:58 nice Aug 26 23:44:41 rsalveti, What blocks getting newest rootstock upstream into Ubuntu? Aug 26 23:44:56 persia: a new release? :-) Aug 26 23:45:18 persia: just 2 simple bug fix, still need to get some work at the ui to create a new release Aug 26 23:45:19 You don't happen to know any upstream rootstock folk that might feel like rolling a release, do you? Aug 26 23:45:47 persia: it works fine with our version, but with upstream you'll always get the initrd and vmlinuz outside the rootfs Aug 26 23:45:50 Ah, makes sense. Best to get the FFe filed *before* beta-release, as otherwise getting it into maverick will be sticky. Aug 26 23:45:54 that's why I said GrueMaster to try it Aug 26 23:45:58 easier to run mkimage later Aug 26 23:46:14 persia: can you buy me some time? :P Aug 26 23:47:00 Unfortunately, time and tuits are the two things of which I currently have shortages :( Aug 26 23:47:04 I'd love to do it, but have more important issues do solve atm Aug 26 23:47:10 yup :-( Aug 26 23:54:07 rsalveti: Running now (got sidetracked by lack of lunch). Aug 26 23:54:33 GrueMaster: with latest minimal image it boots... Aug 26 23:54:44 let me start the upstart debugging session Aug 27 00:02:22 Looks like rootstock finished ok. Aug 27 00:02:34 :-) Aug 27 00:03:28 oops. All three of my microSD cards are preoccupied atm. hrm. Aug 27 00:07:27 Gahhh. Brb. Need to start dinner (roast). Aug 27 00:26:47 Anyone good at picking the right place to drop casts for qreal != double issues? Seems like the outstanding bit for koffice. Aug 27 00:30:15 argh, nothing works today Aug 27 00:32:57 Ok, roast is now...roasting. Aug 27 00:32:58 Which part of nothing is most annoying at the moment? Aug 27 00:33:55 haha, like boards not booting? :-) Aug 27 00:34:57 I get that all the time. Aug 27 00:35:09 Oh yeah. that one. This is why we need large volumes of consumer-grade hardware (although I think I once heard a story about a chicken and an egg...) Aug 27 00:35:41 I just file bugs and let the engineers fix it. Oh, wait... Aug 27 00:36:08 ... Aug 27 00:36:26 :P Aug 27 01:02:08 that's weird, I can't see 20100826.1 anymore Aug 27 01:02:10 wtf Aug 27 01:02:17 some times it works, some times it doesn't Aug 27 01:02:21 persia: any clue? Aug 27 01:02:58 transparent proxies should be removed, and everyone should use IPv6 with endpoint verification. Aug 27 01:03:35 argh Aug 27 01:03:56 and the weird thing is that my current wget is still working! Aug 27 01:04:03 but I can't browse this link Aug 27 01:04:11 clean cache, different browsers Aug 27 01:04:12 just can't Aug 27 01:04:21 Once you get a hit that works, it will be fine. But stale/erroneous cache data needs to wait for timeout to pass. Aug 27 01:04:40 * persia forgets how to trick those offhand Aug 27 01:04:42 yep Aug 27 01:10:00 GrueMaster: weird, it hangs completely Aug 27 01:10:07 the kernel, serial, everything Aug 27 01:11:45 rsalveti, Alternate thought: tell me which file you seek, and I can probably give you a deep URL (which may not be cached) Aug 27 01:12:57 persia: now I'm able to download it! haha Aug 27 01:13:04 persia: for the fs I'm still looking for it Aug 27 01:13:37 * persia fails at parsing Aug 27 01:14:47 nevermind :-) Aug 27 01:14:55 * rsalveti needs a lot more coffee Aug 27 01:15:31 Or a good night's sleep... Aug 27 01:17:25 well, that'd help, for sure Aug 27 01:21:09 rsalveti: you work so late and hard, man Aug 27 01:21:23 rsalveti: do you have chance to test the 2.6.35 on your es2? Aug 27 01:21:27 cooloney: well... nothing better to do haha :-) Aug 27 01:21:41 cooloney: not yet, too much problems during the day Aug 27 01:21:47 cooloney: do you have a binary? Aug 27 01:21:53 I can easily test Aug 27 01:23:45 rsalveti: i got an update patch from sebjan, will try to build again. Aug 27 01:24:00 rsalveti: if i am ready, i will post the url for you, thx Aug 27 01:24:07 cooloney: cool, just ping me when you're done Aug 27 01:34:31 hm, installed network-manager and rebooted, now it hangs Aug 27 01:35:25 Excellent! Uninstall network-manager (from a cross-chroot mounted on something else), and see if that unhangs. Aug 27 01:35:48 that's what I'm doing now Aug 27 01:36:09 but got many other packages installed as a consequence, but let's se Aug 27 01:36:10 see Aug 27 01:36:46 That's OK. We can dig through dpkg.log, and find the responsible party. Aug 27 01:55:27 GrueMaster: persia: seems to be dbus, will test more Aug 27 01:55:49 That's what ogra was finding Aug 27 01:56:01 yep, will try to identify why Aug 27 01:56:21 Sorry, had to take a break. Odd that dbus is giving you guys grief. I had updated that a couple of days ago and it has worked fine. Aug 27 01:59:20 yep, it's dbus :-) Aug 27 01:59:27 initct dbus start -> hangs Aug 27 01:59:32 *initctl Aug 27 01:59:43 or some dependency from upstart Aug 27 01:59:46 Excellent! Can you get a strace? Aug 27 01:59:49 checking again Aug 27 01:59:57 will try Aug 27 02:02:39 Is it dbus or network-manager? Like I said, I updated dbus a couple of days ago while slowly churning through the 400+ package updates. Didn't hang until network-manager update. Aug 27 02:03:02 network-manager Aug 27 02:03:04 But oddly it doesn't appear to start anything after mounting root. Aug 27 02:03:10 removed the dbus dependency and it started fine Aug 27 02:03:17 not networkmanager got the hang Aug 27 02:03:35 It's obviously dbus if "initct dbus start -> hangs" Aug 27 02:03:55 Might only happen in combination with other stuff, but strace should tell us. Aug 27 02:04:05 nops, remove the dbus trigger from networkmanager Aug 27 02:04:10 then started by hand Aug 27 02:04:13 and it hang Aug 27 02:04:14 stack trace (rather than syscall trace) would be better, but hard to get for a hang. Aug 27 02:04:17 so it's probably nm Aug 27 02:04:26 * persia suspects two bugs, rather than one Aug 27 02:04:28 let me start nm with strace or logs Aug 27 02:04:38 could be related with dbus also, on how it's using it Aug 27 02:04:52 but would say it's closer with nm Aug 27 02:04:53 let me check Aug 27 02:04:55 1 sec Aug 27 02:05:39 Please strace the initctl hang first: that's lower level. it7s not worth masking it in nm if that makes it hard to find the dbus part. Aug 27 02:06:00 one thing at a time Aug 27 02:06:07 I want to first find the binary that's causing this Aug 27 02:06:17 then we can trace that Aug 27 02:06:52 Honestly, strace the symptom. Aug 27 02:07:01 if you set strace to follow, you'll track over *many* binaries. Aug 27 02:07:08 then dpkg -S will tell you the package. Aug 27 02:07:13 Saves time doing it twive. Aug 27 02:08:17 but for that I'd have to trace upstart Aug 27 02:08:24 and not initctl start dbus Aug 27 02:08:43 because when dbus is started, upstart tells nm to start Aug 27 02:08:46 and that hangs Aug 27 02:09:31 hrm. Having power & serial console on XM is not fun. Aug 27 02:09:51 Right. strace -f -p ... Aug 27 02:10:17 Then run initctl, and wait for the hang. Aug 27 02:10:26 then check the strace to find out what hung it. Aug 27 02:10:37 then dpkg -S to find the offending binary. Aug 27 02:15:16 rsalveti: Did you need to use a null modem with your XM? Aug 27 02:15:43 GrueMaster: nops Aug 27 02:15:48 usb-serial only Aug 27 02:15:56 same for panda Aug 27 02:16:04 Ok. Aug 27 02:17:34 Oops. Forgot xloader & uboot. That would explain things. Aug 27 02:21:17 GrueMaster: :-) Aug 27 02:23:23 I only appear to be getting garbage on serial Aug 27 02:24:14 Same xloader & uboot as on the daily for beagle, right? Aug 27 02:24:16 speed? Aug 27 02:24:34 GrueMaster: yep Aug 27 02:25:31 Serial works fine on the beagle. Aug 27 02:25:34 Not on XM Aug 27 02:26:39 * GrueMaster was worried that the exterior renovations done with the dremel may have fouled serial usb adapter. Aug 27 02:30:38 * persia wishes koffice would build faster. Aug 27 02:30:45 sigh. Time to fix dinner. Back later. Aug 27 02:30:49 GrueMaster, You still have spare buildd cycles about? Aug 27 02:30:51 Ah, not now :) Aug 27 02:37:33 persia: GrueMaster: interesting, the hang happens when NM tries to activate the usb0 interface Aug 27 02:37:50 if I give ifconfig usb0 up before initializing NM everything goes well Aug 27 02:38:16 but if I try to load it directly, it'll hang when trying to activate the usb0 Aug 27 02:45:34 What's the last syscall before it hangs? Aug 27 02:59:36 a sendmsg to a socket Aug 27 02:59:39 could be dbus **** ENDING LOGGING AT Fri Aug 27 02:59:57 2010