| Eric_Elliott | My first IRC use is now. Where or how do I request a current module be included in Ubu kernel to replace module of 2006? | 01:29 |
|---|---|---|
| spO | hi | 03:09 |
| spO | i did a go clone git://kernel.ubuntu.com/ubuntu/ubuntu-jaunty.git , git checkout Ubuntu-2.6.28-15.48 , git fetch , but then when i do a make menuconfig the top of the menu says configuration for Kernel 2.6.28-10 | 03:10 |
| mjg59 | The build scripts set EXTRAVERSION | 03:10 |
| mjg59 | The source just has whatever upstream defined | 03:11 |
| spO | can you please speak english | 03:11 |
| spO | the extraversion means that it is what is used? | 03:12 |
| spO | or the source is what is used>? | 03:12 |
| mjg59 | EXTRAVERSION is the makefile variable that contains the bit after .28 | 03:13 |
| mjg59 | The source code defines that as one thing. The Ubuntu build system overrides that definition. | 03:13 |
| spO | the build system uses extraversion | 03:14 |
| spO | the source code says it is version 10 | 03:15 |
| spO | err | 03:15 |
| spO | the source code says it is 15, but the extraversion in the makefile says it is 10 | 03:15 |
| mjg59 | Yes | 03:15 |
| spO | so the build system will use the makefile that says it is 10 and you will get a version 10 kernel? | 03:15 |
| mjg59 | No | 03:15 |
| spO | even though the source is 15? | 03:15 |
| spO | you will get a version 15 kernel then? | 03:16 |
| mjg59 | The Ubuntu build system will pass EXTRAVERSION=-15.48 when you use it | 03:16 |
| spO | because that is what the source is | 03:16 |
| spO | what about any extra options that comes with the new version -- 15 | 03:16 |
| mjg59 | What are you trying to do? | 03:16 |
| spO | nothing special | 03:17 |
| spO | just a custom kernel | 03:17 |
| mjg59 | Then ignore the version numbers | 03:17 |
| spO | to make my system super duper faster | 03:17 |
| mjg59 | You're going to be ABI incompatible anyway | 03:17 |
| mjg59 | It doesn't matter | 03:17 |
| mjg59 | Also, the time taken to compile the kernel will dwarf any speed gains you get before you have to recompile again because there's a security update | 03:17 |
| spO | why woudl it be ABI incompatible? | 03:18 |
| mjg59 | Because you'll have made a change that makes it ABI incompatible | 03:18 |
| spO | because security patches/updates are general? | 03:19 |
| spO | i mean too general | 03:19 |
| spO | ie , thye cannot be applied to any custom kernel, in general | 03:19 |
| mjg59 | Kernel security updates replace the entire kernel | 03:19 |
| spO | what does ABI incompatible mean , i mean how is that useful? | 03:19 |
| mjg59 | The only reason to care about the version number is if you want to carry on using the Ubuntu module packages | 03:20 |
| mjg59 | And you can't use the Ubuntu module packages if you've made any change to the config that affects the kernel's internal data structures | 03:20 |
| spO | of course you could build all your modules again, all the modules everytime you make a custom kernel, right? | 03:21 |
| mjg59 | Yes | 03:21 |
| mjg59 | But in that case you don't need to care about the ABI string | 03:21 |
| mjg59 | (which is the -15.48 bit) | 03:21 |
| spO | using modules instead of having it inside the kernel , does that make the system run faster, especially if your system does not load some modules (ie, they are not automatically loaded) | 03:22 |
| spO | auto loading of modules is usually have people use modules anyways, i guess | 03:22 |
| mjg59 | It saves RAM | 03:23 |
| mjg59 | So, basically, just do what you're doing anyway | 03:23 |
| mjg59 | The different numbers make no difference | 03:23 |
| spO | how does a s ystem handle a kernel, is the whole of it loaded into ram? | 03:24 |
| mjg59 | Yes | 03:24 |
| mjg59 | And that ram can't be swapped | 03:24 |
| spO | the best reason to make a custom kernel is proabbyl because you can change the processor family from generic to core 2 | 03:37 |
| spO | i guess | 03:37 |
| spO | so without ABI support, i cannot do an "Update Manager" that is in gnome? | 04:05 |
| spO | or is that mostly somethign different? | 04:05 |
| * hughhalf yawns | 10:01 | |
| mdz | how badly screwed am I? | 10:57 |
| mdz | [ 6.065259] b43-phy0 ERROR: FOUND UNSUPPORTED PHY (Analog 8, Type 4, Revision 4) | 10:57 |
| mdz | hmm, seems to be in bcmwl-modaliases, I'll try that | 11:02 |
| manjo | chroot, kernelbuilds, etc are written up in the wikipage https://wiki.ubuntu.com/KernelTeam/KernelMaintenanceStarter | 11:45 |
| === You're now known as ubuntulog | ||
| Q-FUNK | greetings! is the ability to 'cat' the .config enabled on ubuntu kernels? | 14:26 |
| Q-FUNK | one bug I reported (bug #396286) appears like it might be ubuntu-specific and I'd like to compare ubuntu's config with the one from someone else for whom vanilla 2.6.31-rc4 works fine. | 14:26 |
| ubot3 | Malone bug 396286 in linux "kernel 2.6.31-generic oops after loading initramfs" [Undecided,New] https://launchpad.net/bugs/396286 | 14:26 |
| makke | Q-FUNK: cat /boot/config-`uname -r` there is no /proc/config.gz | 14:31 |
| Q-FUNK | makke: that only provides a partial config, because of config splitter | 14:31 |
| Q-FUNK | that's why /proc/config.gz would be useful | 14:32 |
| ogasawara | JFo: I don't have a specific one generated right now, but you can always take a look at the list of bugs in launchpad | 14:46 |
| JFo | ok cool | 14:46 |
| ogasawara | JFo: https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bugs | 14:46 |
| ogasawara | JFo: and then you can use the Advanced search to narrow down that list to for ex just the New bugs | 14:46 |
| JFo | grazie | 14:46 |
| JFo | excellent | 14:46 |
| JFo | :) | 14:46 |
| ogasawara | JFo: let me know if you have questions. We're sprinting this week so if you don't hear back from me, shoot me an email | 14:47 |
| Q-FUNK | ogasawara: is there any info missing for bug #396286 before the team can process the bug? | 14:47 |
| ubot3 | Malone bug 396286 in linux "kernel 2.6.31-generic oops after loading initramfs" [Undecided,New] https://launchpad.net/bugs/396286 | 14:47 |
| JFo | no problem ogasawara | 14:48 |
| JFo | thanks again | 14:48 |
| JFo | forgot about the sprint | 14:48 |
| ogasawara | Q-FUNK: you mentioned another person being able to boot the vanilla kernel . . . | 14:54 |
| ogasawara | Q-FUNK: care to give the 2.6.31-rc5 mainline kernel build a quick try | 14:54 |
| ogasawara | Q-FUNK: http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.31-rc5/ | 14:54 |
| Q-FUNK | ogasawara: thanks for the URL. could the generic URL for the kernel team's PPA be added to the bug filing instructions on LP? | 14:55 |
| Q-FUNK | yes, I can try it | 14:55 |
| ogasawara | Q-FUNK: you also mentioned wanting to compare the Ubuntu kconfig | 14:56 |
| Q-FUNK | yes, indeed | 14:56 |
| ogasawara | Q-FUNK: take a look at /boot/config-2.6.31-4-generic | 14:56 |
| Q-FUNK | ogasawara: that seems to be an incomplete config, due to config splitter. | 14:57 |
| ogasawara | Q-FUNK: what filesystem are you running? | 15:12 |
| Q-FUNK | ogasawara: ext3 | 15:39 |
| Q-FUNK | is there any repo whre I can find e.g. 2.6.30-10 packages that were released earlier? | 15:47 |
| Q-FUNK | I'd like to try an incremental upgrade of standard 2.6.30 and 2.6.31 packages until I can spot the exact release that broke it | 15:47 |
| ogasawara | Q-FUNK: lets try to do a rough bisect | 16:09 |
| Q-FUNK | ogasawara: ok. can you guide me? | 16:09 |
| ogasawara | Q-FUNK: yup, just grabbing the links | 16:09 |
| ogasawara | Q-FUNK: so you mentioned the ubuntu 2.6.30-2 kernel worked | 16:10 |
| ogasawara | Q-FUNK: so lets have you first confirm the mainline 2.6.30 was working | 16:10 |
| ogasawara | Q-FUNK: http://kernel.ubuntu.com/~kernel-ppa/mainline/?C=M;O=D | 16:10 |
| ogasawara | Q-FUNK: that's all of the mainline kernel builds we've done | 16:10 |
| ogasawara | Q-FUNK: so test http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.30/ | 16:11 |
| Q-FUNK | ogasawara: btw, on a really old bug, someone narrowed-down the cause: bug #241307. the bug is still valid on more recent kernels as well. | 16:12 |
| ubot3 | Malone bug 241307 in linux "kernel oops during bootup in LTSP" [High,In progress] https://launchpad.net/bugs/241307 | 16:12 |
| ogasawara | Q-FUNK: looks like apw is already assigned to that one | 16:13 |
| Q-FUNK | ogasawara: he probably has his hands full tracking 2.6.31 issues for Karmic already, but since I haven't heard from him since adding that info, I thought that I'd ping back. | 16:14 |
| Q-FUNK | btw, seems I was momentarily disconnected. if you have pasted any bisect instructions, I might have missed them | 16:16 |
| ogasawara | [16:10:13] <ogasawara> Q-FUNK: so you mentioned the ubuntu 2.6.30-2 kernel worked | 16:19 |
| ogasawara | [16:10:32] <ogasawara> Q-FUNK: so lets have you first confirm the mainline 2.6.30 was working | 16:19 |
| ogasawara | [16:10:37] <ogasawara> Q-FUNK: http://kernel.ubuntu.com/~kernel-ppa/mainline/?C=M;O=D | 16:19 |
| ogasawara | [16:10:53] <ogasawara> Q-FUNK: that's all of the mainline kernel builds we've done | 16:19 |
| ogasawara | [16:11:29] <ogasawara> Q-FUNK: so test http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.30/ | 16:19 |
| Q-FUNK | ok | 16:19 |
| Q-FUNK | aren't standard ubuntu releases equivalent to mainline releases and rc releases? | 16:20 |
| ogasawara | Q-FUNK: the Ubuntu kernels are based on mainline kernels, but with some additional patches on top | 16:20 |
| Q-FUNK | ok | 16:20 |
| Q-FUNK | so I should try all of the 2.6.30 mainline or just the last one? | 16:21 |
| ogasawara | Q-FUNK: just try 2.6.30 first. I just want to establish a working mainline kernel for us to bisect from | 16:22 |
| Q-FUNK | ok | 16:22 |
| ogasawara | Q-FUNK: after that test 2.6.31-rc1 | 16:24 |
| Q-FUNK | ok | 16:25 |
| Q-FUNK | 31-rc1 wasn't built for i386 | 16:26 |
| Q-FUNK | I'll use -rc1fix1 instead | 16:28 |
| Q-FUNK | /~kernel-ppa/mainline/v2.6.31-rc1-fix1 | 16:29 |
| gnarl | Q-FUNK, Yes, this was specially build instead of the automatic one | 16:44 |
| === elmo_ is now known as elmo | ||
| === Keybuk_ is now known as Keybuk | ||
| === akgraner_ is now known as akgraner | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!