| lamont | initramfs-tools: Depends: lvm2 (>= 2.01.04-5) but 2.00.32-1 is to be installed | 12:58 |
|---|---|---|
| === lamont glares at jbailey | ||
| lamont | hrm.. my bad | 01:00 |
| jbailey | lamont: If you can find an earlier version that's sure to work, I'll relax the dependancy. | 01:01 |
| lamont | nah - I just forgot that I have hoary pinned high and breezy pinned low in the real root on that machine | 01:01 |
| jbailey | *lol* | 01:01 |
| === lamont crosses digits, praying that the ACPI code in -7.11 is love | ||
| lamont | RAMDISK: Compressed image found at block 0 | 01:03 |
| lamont | Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0) | 01:03 |
| lamont | hrm... that's not love | 01:03 |
| lamont | ACPI: Looking for DSDT in initrd... not found! | 01:04 |
| lamont | not found! | 01:04 |
| lamont | wonder if that matters | 01:04 |
| lamont | sigh. serious kernel hacking is going to be required to figure this one out, I fear | 01:04 |
| jbailey | lamont: Are you using mkinitrd or mkinitramfs ? | 01:07 |
| lamont | jbailey: dpkg -i | 01:17 |
| lamont | note also that -6.10 doesn't boot either | 01:18 |
| jbailey | 'kay, so it's not the new dsdt patch, good. =) | 01:32 |
| lamont | jbailey: well, -6.10 didn't boot because interrupts don't happen (damn acpi patch..) | 01:48 |
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| === doko [n=doko___@dsl-084-059-069-234.arcor-ip.net] has joined #ubuntu-kernel | ||
| fabbione | mdz: there was no code change in unionfs since your fix | 06:01 |
| fabbione | what regression do you have? | 06:01 |
| fabbione | bah no changes have been done to VFS layer or unionfs | 06:25 |
| fabbione | and i did tell you in /msg once that unionfs is oopsorama | 06:25 |
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| fabbione | BenC: you here? | 06:39 |
| fabbione | i think i know what's wrong.. but that's unionfs error | 06:44 |
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| === Seveas [n=seveas@seveas.demon.nl] has joined #ubuntu-kernel | ||
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| === one2 [n=me@cpe-66-87-4-181.ut.sprintbbd.net] has joined #ubuntu-kernel | ||
| === one2 [n=me@cpe-66-87-4-181.ut.sprintbbd.net] has left #ubuntu-kernel ["Leaving"] | ||
| === fs [i=fs@213.178.77.98] has joined #ubuntu-kernel | ||
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| infinity | Wow, who do I have to blame for Hibernate working even worse than it did before? | 04:20 |
| mjg59 | infinity: Comparing what with what? | 04:20 |
| infinity | Well, before it used to boot, do the whole "reading state form swap (progress bar)" thing, then hang. I just tested with the new kernel, and it doesn't even seem to bother trying. | 04:21 |
| infinity | It just booted normally and fscked my dirty partitions, no attempt reo resume. | 04:21 |
| infinity | s/reo/to/ | 04:21 |
| === infinity stares at his fingers. | ||
| infinity | So, it's not like things got "worse", per se, since I couldn't suspend (either to disk or ram) before, it just seems to try less hard now. :) | 04:22 |
| === infinity decides to try suspend to ram instead, to see how bad that is. | ||
| mjg59 | infinity: Uhm. What kernel version worked? | 04:24 |
| mjg59 | Or, at least, what kernel version tried harder? | 04:25 |
| mjg59 | My suspicion would be that it's the initrd/initramfs issue that's biting you | 04:25 |
| mjg59 | s/issue/transition/ | 04:25 |
| jbailey | mjg59: How so? | 04:26 |
| mjg59 | jbailey: If there are no messages at all, then the kernel isn't being told to resume | 04:26 |
| jbailey | infinity: Is RESUME= set in your /etc/mkinitramfs/initramfs.conf ? | 04:27 |
| === infinity 's head explodes. | ||
| infinity | Dude, suspend to RAM now works. | 04:27 |
| jbailey | infinity: And also please look at what $(cat /sys/power/resume) gives you | 04:28 |
| mjg59 | infinity: Rock. What hardware? | 04:28 |
| infinity | mjg59 : Brand new SATA T43. | 04:28 |
| mjg59 | infinity: Hurrah | 04:28 |
| infinity | jbailey : RSUME= is commented out. | 04:28 |
| mjg59 | infinity: That'll be your problem, then | 04:28 |
| jbailey | Yup | 04:29 |
| mjg59 | infinity: (New kernel has extreme amounts of love) | 04:29 |
| infinity | Hrm, but that's a regression, mkinitrd didn't need manual love to find the resume partition. | 04:29 |
| mjg59 | Indeed | 04:29 |
| mjg59 | infinity: Is this an upgrade or a clean install? | 04:29 |
| jbailey | infinity: It's a long-fixed regression. | 04:29 |
| infinity | Upgrade from older mkinitramfs versions. | 04:30 |
| infinity | But I've not touched that conffile, so it should have been updated if it's changed. | 04:30 |
| jbailey | It populates it at first install. | 04:30 |
| infinity | And the initrd itself has been regenerated with 0.22 | 04:30 |
| jbailey | It can't change it after that. | 04:30 |
| infinity | Oh, it's not a conffile?... Feh. | 04:30 |
| jbailey | is. | 04:30 |
| jbailey | (lagging phone) | 04:30 |
| === infinity purges and reinstalls. | ||
| Mithrandir | jbailey: perhaps use ucf and then upgrade if you can? | 04:31 |
| infinity | Right, can't purge, kernels want it there. | 04:31 |
| jbailey | Mithrandir: I don't know ucf at all. | 04:31 |
| jbailey | infinity: --force-depends | 04:31 |
| === infinity goes to manually rerun the postinst. | ||
| jbailey | Then reinstall it. | 04:31 |
| jbailey | Or manually rerun, that'll do too. | 04:31 |
| mjg59 | I'm sure there's something broken with acpipnp | 04:32 |
| === mjg59 digs further | ||
| infinity | I assume I'll need to regenrate my initrd after I reinstall this? :) | 04:32 |
| jbailey | infinity: Yes. =) | 04:33 |
| infinity | jbailey : Hrm, now it's not commented out, but it's still an empty string. Was the postinst supposed to put my swap partition in there? | 04:33 |
| jbailey | Yes. | 04:33 |
| jbailey | It works for all the cases I've tried, do you mind troubleshooting it? | 04:33 |
| infinity | It didn't. | 04:33 |
| infinity | Oh, that's cause I don't HAVE aswap partition anymore, cause it still has the bloody suspend magic. | 04:34 |
| === infinity reruns mkswap and swapon, THEN tries the postinst. | ||
| infinity | <grumble> | 04:34 |
| infinity | That worked better. | 04:35 |
| jbailey | Oh good. =) | 04:36 |
| infinity | Still, that seems a bit unintuitive. What happens if I reparition and my swap moves around? | 04:36 |
| infinity | Isn't there a way for us to figure this stuff out on the fly in a reasonably sane fashion? | 04:36 |
| mjg59 | infinity: We can't mount the root filesystem on resume | 04:36 |
| infinity | Not even readonly? | 04:36 |
| mjg59 | Arguably what we /could/ do would be to rewrite the grub config on the fly on suspend | 04:37 |
| mjg59 | No - it'll still replay the journal | 04:37 |
| infinity | Ahh, hrm. | 04:37 |
| infinity | Yes, rewriting the grub config would work too. | 04:37 |
| Mithrandir | that's fairly evil. | 04:38 |
| jbailey | Or have grub probe the swap partitions to detect which ones. | 04:38 |
| jbailey | +have sigs in them. | 04:38 |
| infinity | So is hardcoding the partition we think you resume from. | 04:38 |
| jbailey | infinity: Wlel, we hardcode the partition you're resuming *to* as well, right? =) | 04:38 |
| infinity | jbailey : That's the elegant solution, but suffers from the inability to have two different supended OSs at once. | 04:38 |
| Mithrandir | jbailey: less evil, I'd say. | 04:39 |
| jbailey | If you move shit around during a suspend, your system will break, no sympathy. | 04:39 |
| Mithrandir | infinity: do we care about that use case? You'll have to be careful not to have any overlapping mounts, for instance. | 04:39 |
| infinity | jbailey : Erm, but does anything actually read the initramfs config on suspend and pick the "right" partition?... When I had nothing in there, the kernel still happily suspended to my swap partition. | 04:40 |
| jbailey | Right, that's the default it falls back to. | 04:40 |
| jbailey | I wish it wouldn't. | 04:40 |
| infinity | Mithrandir : Well, I don't know how many people have multiple Linuxes installed at once, but I dual-boot between Win32 and Ubuntu using suspend-to-disk for both, cause it's much faster than full reboots. | 04:40 |
| jbailey | I guess I could detect an image if it's not set. | 04:40 |
| jbailey | phear | 04:40 |
| infinity | So, for a Linux hacker with Fedora, SuSE and Ubuntu installed, I suppose it's a valid use-case to do OS-swapping using suspend-to-disk and three swap partitions. | 04:41 |
| infinity | jbailey : Well, since the kernel plays russian roulette on suspend if that value isn't set, there's no reason you can't do the same on reboot. | 04:42 |
| infinity | jbailey : And the multi-OS thing is probably enough of a corner case that you could just disable the autoprobing feature through a config option, for the power user. | 04:42 |
| === infinity goes to retry suspend-to-disk with his new initrd. | ||
| jbailey | So if resume isn't set, i should check every partition on the system for an image? | 04:45 |
| infinity | Oh well, one out of 2 ain't bad. | 04:48 |
| infinity | With initramfs properly configured, it attemps to resume and fails. | 04:48 |
| infinity | Reading image (progress ... done) ... ACPI: PCI interrupt for XX:XX disabled (repeat several times) ... system hangs. | 04:49 |
| infinity | mjg59 ^ | 04:49 |
| mjg59 | infinity: Ok | 04:49 |
| mjg59 | infinity: I'll be uploading a new acpi-support soon. We'll see if that helps | 04:50 |
| infinity | jbailey : Maybe. Or maybe just not care, I dunno. I'm undecided, now that I know mine still doesn't work anyway. :) | 04:50 |
| infinity | mjg59 : Mmkay. At least suspend-to-ram works for me for the first time. Shame about hibernate being broken. Also odd, since I figured hinbernate kinda worked everywhere. :) | 04:50 |
| jbailey | mjg59: Is it possible for now to just tell the kernel not to attempt to guess if the resume partition hasn't been set? | 04:54 |
| jbailey | mjg59: All it will result in is a clobbered swap file. | 04:54 |
| mjg59 | jbailey: Not without removing that code from the kernel | 04:54 |
| mjg59 | (Which would be easy enough) | 04:54 |
| jbailey | Right, I'm thinking the gratuitous application of #if 0 #endif to a section.\ | 04:55 |
| infinity | That's certainly the safest approach. | 04:55 |
| infinity | Will it just refuse to suspend at all in that case? | 04:55 |
| infinity | Cause that would be fine by me. | 04:55 |
| mjg59 | Yeah | 04:55 |
| mjg59 | Well, it can be made to | 04:55 |
| infinity | "Can't suspend, no partition selected as suspend target." | 04:56 |
| infinity | That's by far the safest way around this whole issue. | 04:56 |
| jbailey | Hey STR work on my laptop with the new kernel | 05:04 |
| jbailey | Sweet. | 05:04 |
| jbailey | HAvn't had that since 2.6.10 =) | 05:05 |
| infinity | Join my l33t STR club, yo. | 05:11 |
| infinity | Though it's never worked on this laptop until today. | 05:11 |
| infinity | Now I just need to get the Xorg ATi X300 driver to a) support PCIe, and b) STOP FUCKING CRASHING EVERY COUPLE OF HOURS, and I may like my new computer. | 05:12 |
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| mjg59 | infinity: Yeah, I have the same issue on an X300 | 05:36 |
| mjg59 | Daniel's supposed to be pulling CVS radeon soon | 05:36 |
| infinity | I suspect A) and B) are closely related. | 05:37 |
| infinity | Though, while CVS may help B), it won't help A). | 05:37 |
| infinity | PCIe isn't going to make breezy unless someone offers some hacker big bags of cash to make it happen. | 05:37 |
| infinity | mjg59 : If you tell me your crashing X300 is AGP, that will put my mind at ease that I may see a fix for the crashes before I see PCIe support. | 05:38 |
| infinity | Cause the latter won't be soon. | 05:38 |
| mjg59 | infinity: It's PCIe | 05:52 |
| infinity | Damn you. | 06:07 |
| infinity | You could have just lied to make me feel better. | 06:07 |
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| === doko [n=doko@dsl-084-059-069-234.arcor-ip.net] has joined #ubuntu-kernel | ||
| === doko [n=doko@dsl-084-059-069-234.arcor-ip.net] has joined #ubuntu-kernel | ||
| lamont-away | Setting up udev (0.060-1ubuntu6) ... | 08:42 |
| lamont-away | udev requires a mounted sysfs, not started. | 08:42 |
| lamont-away | interesting... | 08:42 |
| infinity | Interesting that you don't have a mounted sysfs, or that you do and udev's too dumb to notice? :) | 08:43 |
| lamont-away | iz in a chroot | 08:45 |
| lamont-away | and with a 2.6.12-1-hppa64 kernel | 08:46 |
| lamont-away | I should really upgrade to 2.6.12-7.11 | 08:48 |
| === BenC [n=bcollins@ip68-107-227-12.hr.hr.cox.net] has joined #ubuntu-kernel | ||
| === xxx_ [n=xxx@220.235.201.15] has joined #ubuntu-kernel | ||
| === xxx_ [n=xxx@220.235.201.15] has left #ubuntu-kernel ["Leaving"] | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!