[12:25] infinity: l-r-m was breaking d-i, so I've uploaded what I think should be a fix for that [12:25] hard to test without building new images though [12:25] heh [12:28] Keybuk: line 7 of 10-devfs.rules looks suspicious; missing %? [12:29] yup [12:30] Keybuk: could you replace awk '{print $2}' with sed 's/^[^ ] * //; s/ .*//' in /lib/udev/scsi-devfs please? we don't have awk in d-i === Kamion -> bed [12:31] can you mail that one to me? === dilinger [i=dilinger@mouth.voxel.net] has joined #ubuntu-boot === pitti [n=pitti@ubuntu/member/pitti] has joined #ubuntu-boot === HiddenWolf [n=HiddenWo@136.6.dynamic.phpg.net] has joined #ubuntu-boot === fabbione tests root=/dev/disk/by-uuid/* [10:44] cat /proc/cmdline [10:44] root=/dev/disk/by-uuid/4d216a51-2317-47a4-872f-125a6d035d08 ro quiet splash [10:44] it seems to work [10:45] but there is another side effect we didn't think when writing the specs [10:45] root is not enough [10:45] we need to possibly detect other partitions and swap [10:45] otherwise even if we get root, we might miss the rest of the system [10:45] that defeats the purpose of the spec [10:46] bah Keybuck isn't here [10:49] But other partitions should be in /etc/fstab... [10:49] (by uuid, perhaps, but they should be there anyway) [10:54] the problem is that swap doesn't have a uuid [10:54] fstab supports the UUID= and LABEL= syntaxes [10:54] you can use LABEL= for swap [10:54] and the thing is that yes, the other partitions needs to have uuid too [10:54] fail to do so, we are in trouble [10:54] But multiple swap partitions can have the same label. [10:54] mm [10:54] like a / on ext3 = ok, /home on xfs = nok [10:55] I thought uuid-on-swap was something we were fixing? === infinity thought he heard something about this during the breezy cycle. [10:55] uuid is a property of the fs [10:55] Yes, and? [10:55] swap isn't exactly a fs [10:55] mkswap could do uuid magic. [10:55] swap isn't an FS, but it still has magic. [10:55] Not much, but some. [10:56] And losing another few bytes for a UUID isn't the end of the world. [10:57] no, but you would introduce tons of other problems [10:57] think of the resume from hibernate that clear the swap or so [10:57] your approach would assume that swap isn't used by anyone else [10:57] what if the user dual boot breezy/dapper [10:58] both clean the swaps in 2 different ways [10:58] bang [10:58] Yeah, ew. [10:58] Hrmph. [10:58] Of course, if this sort of breakage isn't done some day, it can never be done. Catch-22. [10:58] yes i agree [10:58] I'd love SOME way to uniquely identify swa ppartitions. [10:58] there is :) [10:59] partition IDs [10:59] Har. [10:59] but still [10:59] you don't know if that swap is the one you are allowed to use [10:59] so it still the suck [10:59] a possible solution would be to generate a uuid [11:00] and assign it to the swap as LABEL [11:00] that might do [11:00] Oh wel, for all the noise people make about the dangers of shared swap, using the wrong swap, etc, etc, it seems to be a reasonably uncommonly-tripped bug. [11:00] People THINK about it far more than they actually get bitten by it. [11:00] i would rather avoid to be the one crossburned for that [11:00] And yes, we could generate a uuid-like label. Would be ugly in fdisk output, but that's not the end of the world. [11:01] i think actually.. that going for LABEL=uuid is a much better approach [11:01] because it might allow us to do it for all kind of FS [11:01] since LABEL is at partition level [11:01] and not FS [11:02] Do we ge LABELs on all partition tables, or just MSDOS? [11:02] I get them on PowerPC. Yay. [11:02] no idea [11:02] Well, MacPPC anyway. [11:02] but i guess we do [11:03] what fs? [11:03] /dev/hda7 Apple_UNIX_SVR2 Debian 55398400 @ 2098368 ( 26.4G) Linux native [11:03] /dev/hda8 Apple_UNIX_SVR2 swap 1136576 @ 57496768 (555.0M) Linux swap [11:03] Uhh, ext3, not that that matters. [11:04] no, it shouldn't [11:04] that's why i was cross checking ;) [11:04] Not sure if all partition tables have label-like metadata that gets translated, but at least the three release arches would be okay. [11:05] I don't recall if BSD disklabels and slices do it. [11:05] no idea [11:06] (That was one motivation for UUID, though, is that since it's not at the partition table level, we're not subject to weird architectures and their weird partitioning schemes) [11:06] that's why we can use uuid over label protocols [11:07] all of this assuming we can stick such a long label [11:07] Look for uuid, fall back to label, if we have neither, cry? [11:07] No, we can't have labels that long they'd need to be a bit less unique. [11:07] DOS labels are... 12 chars?... 16?... I don't recall.. But short. [11:08] crap [11:08] i suggest we postpone the spec for dapper+1 [11:08] this is getting already too hairy === Keybuk [n=scott@descent.netsplit.com] has joined #ubuntu-boot [11:16] hey Keybuk [11:17] Keybuk: in about 10 minutes you want to check the irc log for this chan (last 30 minutes discussion) [11:17] about probe for root fs [11:19] ok [11:24] infinity: DOS labels are 11 characters, according to 'head debian-installer/installer/build/config/x86.cfg' [11:40] *giggle* [11:40] for those following the "Dude? Where's My DMA?" bug [11:41] it cleared itself up for mdz all by itsself [11:41] with BenC saying the only change to piix was a tweak to the modular IDE driver patch that couldn't POSSIBLY have caused it [11:41] upstream IDE subsystem maintainer has identified the patch causes problems for us [11:41] it's the (you guessed it) modular IDE driver patch! [11:58] Keybuk: let me know whenabouts you're uploading udev with the changes I mentioned on IRC last night; I'll need to rebuild d-i with those to get the live CD working [11:59] yup [11:59] will be this morning [11:59] just ploughing through the things everyone sent me overnight for my attention this morning [11:59] ta [12:03] i wonder if we have dpkg-query in d-i somewhere... === fabbione adds another item in the things to check [12:08] no [12:08] no plans to add it either [12:09] you can look through /var/lib/dpkg/status in d-i, but obviously that will only list udebs [12:09] if you want to inspect /target, chroot to it [12:09] infinity: what's the general stance on useful-things-for-debugging in initramfs [12:09] like lsmod, udevmonitor, etc. [12:09] Kamion: yeah the issue is that we can't trust what's in /target ;) [12:10] Kamion: that's why i am looking at several different options [12:10] fabbione: then you can't trust /target/var/lib/dpkg/status either ... [12:10] Kamion: i know. that's why there will be 2 checks available. one that uses status and one that will gather pkgs info from the files installed on the system [12:11] but the latter is way slower than the former and prone to more false positives [12:11] as it stands now i am implementing all small shell functions do to almost atomic operations that requires less tools as possible [12:12] discarding one in favour of another it's easy [12:13] fabbione: you asked me to check the irc log, I assume about the /dev/disk stuff? [12:16] Keybuk: yes [12:17] I'm not sure I understand the problem ... the root filesystem knows where the other partitions are [12:17] Keybuk : I wouldn't be against a DEBUG=yes flag in initramfs.conf that optionally copied in some debug stuff, and could be seen by other hooks (like udevs) to copy debug stuff for them... [12:18] Keybuk: no it doesn't... it does if the disk doesn't change name.. [12:18] Keybuk : Not sure we want debug bloat in all initramfs images, hence the flag springs to mind as sane. [12:18] fabbione: the disk is named [12:18] it has a serial number [12:18] Keybuk: but we are using uuid exactly because we want to be able to move it around [12:18] you could use /dev/disk/by-id/usb-_-part2 [12:18] (inside the root filesystem) [12:19] then use root=/dev/disk/by-uuid or whatever to find the root filesystem itself [12:19] that won't work either.. [12:19] why not? [12:19] 0 lrwxrwxrwx 1 root root 9 2005-12-02 09:41 scsi-0ATA_WDC_WD1200JB-00CLinux_ATA-SCSI_simulator -> ../../sda [12:19] <- [12:19] this is an IDE disk in a USB external box [12:20] right? [12:20] i could remove it from there [12:20] stick it in my desktop [12:20] that would change to ATA [12:20] sorry [12:20] ata-* [12:20] use partition labels then [12:20] and /dev/disk/by-label [12:21] Keybuk: label can only be 11 chars [12:21] there is a high possibility of clashing [12:21] the reason why we did chose uuid was exactly because of its meaning [12:21] pretty low probabibility [12:21] what's wrong with using /dev/disk/by-uuid/ for the other filesystems too (apart from swap)? [12:21] Kamion: not all FS support uuid [12:21] Keybuk: quite high if you have two parallel installations done with the same installer [12:21] most the FS we care about support uuid [12:21] for those that don't generate a random label [12:21] root on ext3 and home on fancy fs [12:21] you are doomed [12:22] that gives you a 1 in 269561249468963094528 chance of clash [12:22] we don't support home on "fancy fs" [12:22] Keybuk: yes we do. we still provide the option for manual partitioning [12:22] and iirc not all the selectable FS have uuid [12:22] so? just tell them "uh, no, go back and fix this" [12:23] or for those FS that don't support uuid, generate a randomly unique LABEL and force it upon them [12:24] so you would end up with a fstab that will look like: /dev/disk/by-uuid/$longnumber / etc... [12:24] and other lines with LABEL= [12:24] that will be quite inconsistent [12:24] just use /dev/disk/by-uuid/$uuid and /dev/disk/by-label/$label [12:24] that's not inconsistent [12:24] Keybuk: what happens if 2 lables are clashing? [12:25] labels even [12:25] at udev level.. does it go foobar? or it continues like if nothing happen [12:26] just makes a link for one of them [12:26] and the changes of two labels clashing is incredibly remote [12:26] when we have 250 billion billion installations, we can worry [12:27] with 11 chars is a bit less than that [12:27] and that's just assuming we only stick to alphanumerics [12:27] no, that's 11 alpanumeric chars [12:27] right [12:27] DOS is 11 8-bit chars, so it's a number so high my computer can't calculate it [12:28] in reality people do kinda want labels to be human-readable ;) [12:28] or at least PRINTABLE [12:28] indeed [12:28] i sort of agree with Kamion [12:28] that's why I stuck to alphanumerics [12:29] hell, if you just limit yourself to upper case and numbers, you're still over a billion billion possible labels [12:30] i was wondering if the first 11 digit of md5sum something would do [12:30] like md5sum the output of the / uuid with the date in nanosecond [12:30] that would probably be almost unique [12:31] why not just pack the / uuid with the partition number or date? [12:31] the / uuid is 37 hex characters [12:32] actually, 33 and 4 -s [12:32] yeah [12:32] you could certainly use much of that [12:32] and there are shorter uuid standards too [12:32] Can we convert base16 to base36? :) [12:32] the md5 was to scramble numbers completely [12:33] anyway, this isn't a SKY FALLING problem [12:33] no [12:33] there's plenty of ways to identify the partition we want to put the swap on [12:34] Keybuk: other than LABEL no... [12:34] there are :p [12:34] we could even go as far as tricks [12:35] we could improve /dev/disk/by-path to create a tree of directories, and then figure out which one the root filesystem is on, then use ../part2 if we REEEEALLLY wanted [12:35] but that's a bit silly [12:35] actually, we probably wouldn't need to improve it at all [12:35] just figure out which /dev/disk/by-path is the root fs [12:35] the problem isn't root [12:35] that assumes you want swap on the same disk, of course [12:36] we can't assume :( [12:36] we would be able to do so only in one condition [12:36] getting Kamion to kill manual partitioning [12:36] till users are able to do partitions as they like [12:37] we are kind of doomed [12:37] OR [12:37] the possibility is to add all this chunk of code only if they use our recipe's [12:37] via guided partitioning === fabbione eyes the clock and ponders food [12:38] or just make the manual partitioning generate the random label for every new partition [12:38] then if the user changes it to something more readable, which might clash (like "SWAP") that's their problem [12:38] and the bug can be closed with something like [12:39] "This appears to be your foot that you shot, with your gun" [12:39] LOL [12:39] no really.. we can do it only if they use our Guided stuff [12:39] no, really, we can do it in manual [12:39] we can't ensure that the user doesn't have 2 disks [12:39] and stick half system on one [12:39] and swap on the other [12:39] we don't need to [12:40] ok let me eat and think about it [12:40] just use by-uuid for everything that has it, and by-label for things that don't [12:40] otherwise i will soon be grumpy ;) === jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-boot [01:20] Keybuk: could you put the /dev/discs/ support back into scsi-devfs, please? d-i needs it [01:21] and the Debian 076 package has it [01:24] "back into" ? [01:24] is it not there? [01:24] I just used the scsi-devfs script from upstream [01:27] Kamion: the hooks i need for that md5sum stuff are in the rescue-udeb right? [01:28] Keybuk: it's not there, no [01:28] fabbione: rescue-mode [01:28] Kamion: thanks [01:28] fabbione: see yaboot-installer for (at present) the only example of another udeb using it [01:28] meh, I'll have to see what Debian patches they have for that [01:28] *mutters and grumbles about the installer using obsolete, unsupported and broken scripts* :p === ..[topic/#ubuntu-boot:Kamion] : known: fb drivers loaded, oss drivers loaded, no network plugging, alsa rules not reloaded, mtab not updated, /dev/pts not mounted, scsi-devfs || fixed: notify-reboot, vio_type segfault, no modules loaded, hal/pcmcia rules not reloaded, nfs root fails, sata root fails, pnp devices not loaded, init stop/start, grepmap bitching [01:29] it's gradually getting fixed, but the live CD doesn't have parted_devices so disk-detect falls over === ..[topic/#ubuntu-boot:Keybuk] : known: oss drivers loaded, no network plugging, alsa rules not reloaded, mtab not updated, /dev/pts not mounted, scsi-devfs || fixed: notify-reboot, vio_type segfault, no modules loaded, hal/pcmcia rules not reloaded, nfs root fails, sata root fails, pnp devices not loaded, init stop/start, grepmap bitching, fb drivers loaded [01:29] I've got a patch heading into parted upstream (I hope) to avoid some more of the problems [01:30] *nods* [01:31] getting the installer to eventually search /sys for drives and cds would be much sweeter [01:31] it'll just use parted to walk drives [01:32] hmm, scsi-devfs.sh seems to be in extra/ and in the .diff.gz; does it come from upstream anyway? [01:32] Kamion: it looks pretty simple the README looks clueful [01:32] "and" the README.. [01:32] it comes from upstream, I guess Marco's copied it and modified it for his own nefarious purposes [01:32] oh, meh, it's completely different [01:33] it's been in Debian since 0.019-3; I imagine it predates upstream's [01:33] fabbione: thanks, I wrote the README ;) [01:34] Kamion: that might be why i was able to understand it :) [01:36] I'll just use a non-SATA/SCSI system for testing for now [03:11] *giggle* at kay [03:12] after spending most of yesterday evening arguing that there was no reason to run udevd in initramfs, but just place as /sbin/hotplug [03:12] he's now decided we were right after all === HWolf [n=HiddenWo@136.77.dynamic.phpg.net] has joined #ubuntu-boot [03:31] Keybuk: ^5! === pitti [n=pitti@ubuntu/member/pitti] has joined #ubuntu-boot === pitti [n=pitti@195.227.105.180] has joined #ubuntu-boot [04:05] hmm [04:05] it seems the ide subsystem could be going away :) [04:05] huh? [04:05] Huh? [04:05] libata has got parallel ide support now [04:05] and that seems to be strongly favoured to replace much of the ide subsystem [04:06] pretty much how sata is handled [04:06] s/away/replaced [04:12] Keybuk: I asked about that the other day, and I thought ben said ide-generic would still be needed for a while. === pitti [n=pitti@ubuntu/member/pitti] has joined #ubuntu-boot [04:17] you didn't dig enough === HWolf [n=HiddenWo@136.169.dynamic.phpg.net] has joined #ubuntu-boot === jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-boot === pitti [n=pitti@ubuntu/member/pitti] has left #ubuntu-boot [] [06:27] right, I think that's all the pile-o-udev bugs solved [06:28] Kamion: what installer issues do you have (currently) ? [06:32] Keybuk: scsi-devfs lack of /dev/discs, live filesystem isn't building, casper needs to be updated to do something initramfs-ish instead of pivot_root [06:32] I've been concentrating on the live CD rather than on the install CD so it's entirely possible that there are install CD bugs I don't know about yet [06:33] I'm rolling new install CDs now and will test them shortly [06:33] I'm looking into the devfs thing now, I think I may just drop that stuff in wholesale from Debian [06:40] Keybuk: could you take a look at debian #342057, BusLogic not loaded because !sysfs (seems needed for vmware testing of initramfs-tools). [06:40] makx: I think we fixed that in the Ubuntu kernel. [06:41] jbailey: you added sysfs hooks? :) [06:41] Ben or Fabio did, I think. [06:41] I try to stay on this side of userspace. =) [06:41] why's that not upstream :-P [06:42] - [scsi/BusLogic] Add MODULE_DEVICE_TABLE [06:42] it probably is upstream by now [06:42] -- Ben Collins Sat, 12 Nov 2005 23:34:38 -0500 [06:42] From the 12th to now, it's hard to guess. [06:42] I know there's alot queued to go upstream. [06:44] a quick grep in git-commits-head shows it's not yet landed for linus. [06:45] jbailey: anyway thanks for the hint, good news. :) [06:49] makx: Or perhaps check with BenC in #ubuntu-kernel [06:50] jbailey: for now i'll include it in d-kernel, but yes. === jbailey_ [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-boot [07:19] Kamion: wow, I see what you mean about these scsi-devfs scripts being totally different [07:19] they aren't even related [07:21] right, I don't think they share ancestry [07:21] the upstream one looks far more complete [07:22] it does a lot of stuff I don't care about ;-) [07:22] what do you actually care about/ [07:23] do you need all the /dev/bus/*, /dev/scsi/*, etc. [07:23] or do you just want /dev/cdroms and /dev/discs ? [07:23] /dev/scsi/ is good, I don't especially care about /dev/bus/ personally [07:23] I don't think I actually NEED /dev/scsi/ though [07:24] I'm trying to work out how to make this work [07:24] we plug all the disks at once, so get_next_number returns 0 for every single disk [07:27] what I'm wondering is ... [07:27] if I throw all of this away [07:27] and instead replace it with a small "enumerate disc-like objects" helper that just makes /dev/discs and /dev/cdroms [07:27] would that be sufficient for the installer? [07:27] or does it really need /dev/ide and /dev/scsi ? [07:30] no, it doesn't need those [07:30] I use them occasionally but I can easily adjust [07:30] so it really just relies on /dev/discs and /dev/cdroms ? [07:33] yep [07:34] how about /dev/floppies? [07:50] /dev/floppy [07:50] floppy-retriever looks for /dev/floppy/0 [07:54] does it matter if there are gaps in the numbering, btw? [07:54] /dev/floppy/0 has to be thus, for the rest probably not [07:55] how about for cdroms and discs? [07:56] "for the rest probably not" [07:56] ah, sorry === HiddenWolf [n=HiddenWo@136.169.dynamic.phpg.net] has joined #ubuntu-boot [07:58] Kamion: any chance of seeing my parted /sys/block patches making it into parted [07:58] ? [07:58] for ubuntu's installer [07:58] dilinger: have they gone upstream / to Debian? [07:58] I've got parted back into sync now, would be kind of a shame to branch again [07:59] I can if need be though, file me a bug [07:59] Kamion: they have gone upstream; I have no idea debian's status w/ them [07:59] Debian's tracking upstream pretty closely [07:59] upstream seems to want me to assign copyright to the FSF, but doesn't seem to be able to put me in contact w/ whomever i need to talk to about that [07:59] *shrug* [07:59] file a bug and I can keep an eye on it [07:59] I've done FSF copyright assignment; you can probably just mail their clerk [08:00] Kamion: file a bug in which bts? [08:00] Ubuntu [08:00] bugzilla? malone? [08:01] bugzilla [08:01] http://www.fsf.org/licensing/assigning.html - the maintainer is supposed to send you the questionnaire though [08:03] well, it's svenl who's expressed interest in the patch, but i guess he wanted otavio to mail me or something === dilinger shrugs [08:19] ok, submitted [08:20] thanks [08:47] Kamion: http://people.ubuntu.com/~scott/storage_enum.sh [08:49] I'll take your word for it on the implementation, but the intent looks plausible [08:50] that should actually fix our current problems too, as it should reliably never use the same number twice [09:33] * Update installer startup script to use "cp -au" like the init script, [09:33] instead of "cp -af". [09:33] Keybuk: please revert that - busybox cp doesn't have -u [09:33] d'oh [09:33] tsk [09:33] busybox sucks [09:33] it has a -f that doesn't behave like the real one [09:33] and then doesn't have -u [09:33] and in any case I'm sure it was the "cannot be opened" trap I hit [09:40] reverted that bit === jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-boot