And what can I Say?!?!?!?! i need to be there physically for that so i cant come to UDS at all it's certainly an interesting conundrum hmm, actually not a problem slangasek, do you know the kernel config option that turns on kprobes? I seem to not have that, but some of our kernel patches disable CONFIG_KPROBES=y on !ARM mterry: not off-hand; do you know why you need it turned on? slangasek, the whole point of kprobes is for performance profiling, right? mterry: yes mterry: "performance profiling" :) slangasek, OK. :) But yeah, I need kprobes turned on, even if we don't turn it on for other arches. What I mean is, I want to know if this config option is somehow in our config set, or not set at all (and I'm missing it) if you want to use kprobes for performance profiling, you generally have to enable the kprobes kconfig option slangasek, oh really? That's fine then, then I will do that but there is a conditional in our kernel package so you can turn it off for architectures that don't support it (it uses that config option) mterry: yes, please set up the kprobes config option regardless - it's only a performance optimization if you're actually going to use it :) otherwise it's pointless slangasek, right, of course. I will do so. Thanks if you're not actually going to use it, having it turned off in the kernel package means you won't get the performance optimization but it should be present for boot time profiling slangasek, OK, k sigh, why cant i get my second NIC on my laptop to work on my second VM ... *sigh* cjwatson: I guess you didn't see I got bug #746275? :) Launchpad bug 746275 in ubiquity (Ubuntu Natty) "Can't access the web during install with ubiquity" [Medium,Incomplete] https://launchpad.net/bugs/746275 do you have any thoughts on it? slangasek, is the bootchart stuff supposed to work with the new kernel ? slangasek, i have a system here that doesnt finish booting, it ends up at a black screen slangasek, is there a special kernel needed to make that work ? ev: could you attach /var/log/syslog and /var/log/installer/debug? ogra: it's working fine here, but there's no reason it shouldn't work with the oneiric kernels - please file a bug slangasek, i'm fine with reporting the issue, i just wanted to know if its because i use lucid kernels on lucid :) ogra: it'll be kernel, sure ogra: nope, I just updated it to latest upstream (from 2.6.37) cjwatson: sure thing ogra, lucid ? ! i'm running maverick thats why i am in an IRC channel with lts in the name ;) well, yes, I did say "lucid" earlier, but yes I mean the hardware ah, ok well, it also is possible that my nfs shares are to blame, i can see the luks prompt and all in my kern.log but i cant read the content so there might be some issue there ... /dev/pts is mounted? it is (that is indeed a kernel bug but the kernel can't do anything about it) yeah, its a weird issue indeed bah ah, bah, you have the old nfs mount code ... right ok, I'll fix that soon in maverick so i get the whole system in front of me i'll poke around and see what i can find no hurry ... its just a bit annyoing that everythimg seems to get stuck *everything do you know what the kernel command line was? no, though i see the luks stuff in there i'll try to find something similar on a running system hmm, i see the UUID twice, is that normal ? I doubt it seems all fie systems use different values could you put /var/log/installer/debug on paste.u.c or something? gah, it doesnt show anything without the extra UUID being on there what doesn't? the prompt that asks to type something it's supposed to be hidden but I can turn it off in the UI if you like no, its just the prompt of the luks-encrypted partition i mean that one *the* one ok, the content of the command line gets pasted from the kernel message i still only see "begin: Running head for script /usr/share/finish-install/finish-install" it never gets to "end: Back to main menu" and no errors occur (my test system is a headless server running a kvm guest though) and i cant find any additional details in /var/log/installer/debug either well, the problem here isn't the kernel I'm seeing a similar thing here it's in code called by the kernel before the initramfs runs so is this a kernel regression or is there something deeper that can be fixed ? I don't know, that's why I'm asking you to log the debug messages I just wanted to ensure we didn't have a real problem here i cant access any logs apart from kern.log and dmesg and thats the content of /var/log/kern.log /var/log/boot is empty? no, but /var/log/dmesg is what i pasted above (and yes, empty, but thats expected) well, that won't give you much ... and no boot.log in /var/log/ what hardware is this? my laptop intel atom netbook you were very quick to say "no boot.log" its running maverick I mean is it an imx51 or something similar? yup not running armel but i dont think thats special here imx51 that is ok what I'm trying to tell you is that /var/log/boot on an imx51 is empty because I didn't have time to implement that in the default kernel build understood did the initramfs ever run? i have an initramfs on the system if that makes a difference yes yeah, you have to press enter in order for it to actually try anything I mean, that's how it's supposed to work ah, ok this looks to me like something to file against plymouth well, i would guess there is more stuff that doesnt work if there is no luks or do you have that on lucid too ? I have it on lucid what you're seeing is due to it not being there ok well, its a maverick build on maverick i think at least it didnt have any maverick stuff added in config ok, i have to leave for a bit now ... but filing a bug if the kernel logs reveal anything and i'll also look at the initramfs setup, just to make sure i dont overlook anything s/kernel logs/kernel logs for me/ *if they