02:10.16 | *** join/#maemo-ssu LaoLang_cool (~LaoLang_c@14.151.29.219) |
02:28.30 | *** join/#maemo-ssu amiconn_ (amiconn@rockbox/developer/amiconn) |
03:04.06 | *** join/#maemo-ssu DocScrutinizer05 (~HaleBopp@openmoko/engineers/joerg) |
03:22.26 | *** join/#maemo-ssu M13 (~Miranda@83.149.34.197) |
04:25.22 | *** join/#maemo-ssu _rd (~rd@p57B4B639.dip0.t-ipconnect.de) |
05:17.11 | *** join/#maemo-ssu dhbiker (~dhbiker@95.87.145.172) |
06:12.16 | *** join/#maemo-ssu M4rtinK (~M4rtinK@ip-86-49-81-87.net.upcbroadband.cz) |
06:41.51 | *** join/#maemo-ssu amiconn (amiconn@rockbox/developer/amiconn) |
06:57.41 | *** join/#maemo-ssu LauRoman (~LauRoman@5-14-92-227.residential.rdsnet.ro) |
07:34.54 | *** join/#maemo-ssu andre__ (~andre@ip-62-24-76-119.net.upcbroadband.cz) |
07:34.54 | *** join/#maemo-ssu andre__ (~andre@wikimedia/aklapper) |
07:40.52 | *** join/#maemo-ssu sunny_s_ (~sunny_s@86.57.254.135) |
07:47.33 | *** join/#maemo-ssu M13 (~Miranda@83.149.35.212) |
08:04.01 | *** join/#maemo-ssu futpib (~futpib@89.106.197.102) |
08:06.51 | *** join/#maemo-ssu wmarone_ (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
08:10.30 | *** join/#maemo-ssu wmarone (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
08:16.42 | *** join/#maemo-ssu wmarone_ (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
08:22.56 | *** join/#maemo-ssu wmarone (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
08:26.02 | *** join/#maemo-ssu wmarone_ (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
08:28.51 | *** join/#maemo-ssu wmarone__ (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
08:34.34 | *** join/#maemo-ssu wmarone (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
08:39.46 | *** join/#maemo-ssu Pali (~pali@Maemo/community/contributor/Pali) |
08:40.36 | *** join/#maemo-ssu wmarone (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
08:42.43 | *** join/#maemo-ssu wmarone_ (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
08:52.32 | *** join/#maemo-ssu wmarone__ (~wmarone@c-67-174-151-253.hsd1.ca.comcast.net) |
09:01.15 | *** join/#maemo-ssu Martix_ (~martix@static-84-242-103-180.net.upcbroadband.cz) |
09:50.58 | *** join/#maemo-ssu Pali (~pali@Maemo/community/contributor/Pali) |
10:12.57 | *** join/#maemo-ssu Pali (~pali@Maemo/community/contributor/Pali) |
10:15.36 | *** join/#maemo-ssu Martix_ (~martix@static-84-242-103-180.net.upcbroadband.cz) |
10:23.43 | *** join/#maemo-ssu lizardo (lizardo@nat/indt/x-wvytgaypdpomyepi) |
11:31.11 | *** join/#maemo-ssu mnzaki (~mnzaki@41.46.178.191) |
11:55.51 | *** join/#maemo-ssu macmaN (~chezburge@138.167.190.90.dyn.estpak.ee) |
11:59.50 | *** join/#maemo-ssu tg (~irc@eu.tgbit.net) |
12:13.11 | *** join/#maemo-ssu BCMM (~BCMM@unaffiliated/bcmm) |
12:45.47 | *** join/#maemo-ssu AndrewX192 (~andrew@131.191.116.183) |
12:45.47 | *** join/#maemo-ssu AndrewX192 (~andrew@unaffiliated/andrewx192) |
12:46.15 | *** join/#maemo-ssu tg (~irc@eu.tgbit.net) |
12:48.32 | *** join/#maemo-ssu arcean (~Tomek@apn-46-76-189-169.dynamic.gprs.plus.pl) |
12:50.27 | Pali | merlin1991: ping |
13:54.50 | *** join/#maemo-ssu NIN101 (~NIN@p57B9E147.dip0.t-ipconnect.de) |
14:00.54 | Pali | this merge request https://gitorious.org/community-ssu/modest/merge_requests/13 for modest is there quite a long |
14:01.08 | Pali | I'm going to merge it into master (cssu-devel) |
14:54.48 | merlin1991 | Pali: yeah, go ahead |
14:54.59 | merlin1991 | I hope I'll finally get to work on the -t release today/tomorrow |
14:55.17 | Pali | merged and pushed to cssu-devel |
14:55.38 | kerio | yay :D |
14:56.31 | merlin1991 | currently updating servers from squeeze to wheezy |
14:56.41 | merlin1991 | I kinda fear touching merlin1991.at though |
14:56.49 | kerio | nay D: |
15:01.37 | merlin1991 | blindly running dist-upgrade is a scare thing without proper backups :D |
15:01.46 | merlin1991 | s/scare/scary/ |
15:39.46 | *** join/#maemo-ssu asem (~chatzilla@197.150.168.252) |
15:59.42 | merlin1991 | makes backups |
16:25.25 | kerio | backups are good! |
16:40.39 | Pali | and sometimes you find that you have backup in backup in backup in backup... |
16:40.55 | Pali | and trying to find out why you have full HDD :D |
16:41.26 | kerio | i solve the backup problem by not having anything important |
16:59.34 | *** join/#maemo-ssu Sicelo (Sicelo@gets.high.on.ircspeed.net) |
17:14.21 | *** join/#maemo-ssu BCMM (~BCMM@unaffiliated/bcmm) |
17:25.54 | *** join/#maemo-ssu M4rtinK (~M4rtinK@ip-86-49-81-87.net.upcbroadband.cz) |
17:33.10 | *** join/#maemo-ssu M13 (~MirandaLS@170.133-224-87.telenet.ru) |
18:08.12 | Pali | it looks like my change to rcS-late working fine. it detecting maemo installed on SD card and eMMC and also correct filesystems |
18:08.32 | Pali | and has support for static /etc/fstab |
18:38.57 | *** join/#maemo-ssu M4rtinK (~M4rtinK@46.227.13.46) |
18:43.07 | kerio | Pali: does it still break horribly if you use LABEL= or UUID=? |
18:44.44 | Pali | ah, I forgot this |
18:45.31 | kerio | and are you still using your crappy version of fsck fs type autodetect? |
18:47.02 | Pali | yes and it is not crappy |
18:47.24 | kerio | adds completely unnecessary complexity |
18:47.26 | Pali | I really do not want to call ext fsck untility on fat/ntfs partittion |
18:47.48 | kerio | it won't do anything except complain that it can't find any superblock |
18:47.59 | kerio | actually, hold on |
18:48.03 | kerio | YOUR version will do that |
18:48.10 | Pali | kerio, can you write your fstab with label or uuid? |
18:48.17 | kerio | my fstab uses LABEL |
18:48.23 | kerio | oh you mean pastebin it somewhere? |
18:48.26 | Pali | yes |
18:48.39 | *** join/#maemo-ssu ruskie (ruskie@sourcemage/mage/ruskie) |
18:48.45 | kerio | Pali: it's *your* version that will blindly call fsck.$type_in_fstab |
18:48.47 | Pali | blkid will transtale it to /dev/device |
18:48.48 | kerio | fsck will use blkid |
18:50.17 | kerio | no, seriously, are you really trying to do this to avoid calling e2fsck with a "wrong" argv[0] that WON'T BE USED AT ALL BY IT? |
18:51.22 | kerio | Pali: http://fpaste.org/10910/67952673/ |
18:51.52 | Pali | kerio, ok I can change code which will call generic "fsck" version |
18:52.31 | Pali | but what happen if: you have ext4 in fstab and partition is extš? |
18:52.37 | Pali | ext3? |
18:53.14 | Pali | strace told me that generic fsck using data from /etc/fstab |
18:53.45 | Pali | so this *can* damage/convert filesystem |
18:53.53 | kerio | `fsck LABEL=lolol` correctly picked up my vfat partition |
18:54.02 | kerio | and that's nowhere in fstab |
18:54.57 | kerio | and regardless |
18:55.39 | kerio | dosfsck just said "logical sector size is zero" on my ext4 partition |
18:55.55 | kerio | and e2fsck complained about "The superblock could not be read or does not describe a correct ext2 |
18:55.56 | kerio | filesystem. If the device is valid and it really contains an ext2 |
18:55.56 | kerio | filesystem (and not swap or ufs or something else), then the superblock |
18:55.56 | kerio | is corrupt" |
18:56.02 | kerio | and did nothing |
18:56.15 | Pali | are you called it with -p or -y? |
18:56.18 | *** join/#maemo-ssu _rd (~rd@p57B4B639.dip0.t-ipconnect.de) |
18:56.32 | kerio | if a fsck tool can't figure out if what's ahead of it /is/ a file system of the correct type, what the hell can it do? |
18:56.53 | kerio | Pali: yep |
18:57.16 | Pali | filesystem detection is overwritten by /etc/fstab |
18:57.25 | kerio | so that's even better |
18:57.35 | Pali | but we using HAL/blkid for detection |
18:58.02 | kerio | we're also using blkid to make fstab |
18:58.05 | Pali | so if you specify BAD filesytem in fstab for SD card, then it still works |
18:58.18 | Pali | now you can have static fstab |
18:58.30 | kerio | and at that point, it becomes your responsibility |
19:00.58 | DocScrutinizer05 | wtf are you doing? |
19:01.23 | Pali | ok, I will extract only device line from /etc/fstab for /home and call generic version fo fsck on it |
19:01.39 | kerio | Pali: omg omg omg omg wait |
19:01.40 | kerio | fsck /home |
19:01.51 | kerio | i just tested it, it works |
19:02.18 | Pali | $ fsck /home |
19:02.22 | Pali | this not working |
19:02.30 | Pali | fsck.ext2: Is a directory while trying to open /home/ |
19:02.36 | kerio | wut |
19:02.50 | kerio | no, /home |
19:02.58 | Pali | fsck /home called fsck.ext2 |
19:03.09 | Pali | why?? I have ext3 on /home |
19:03.37 | kerio | user@kerio900:~$ sudo fsck -n /home |
19:03.37 | kerio | fsck 1.41.3.maemo0 (12-Oct-2008) |
19:03.37 | kerio | e2fsck 1.41.3.maemo0 (12-Oct-2008) |
19:03.37 | kerio | Warning! /dev/mmcblk0p2 is mounted. |
19:03.37 | kerio | Warning: skipping journal recovery because doing a read-only filesystem check. |
19:03.38 | kerio | N900_optfs: clean, 35771/643376 files, 257958/2572288 blocks |
19:03.56 | Pali | ok, it looks like if filesystem is unknown it calling fsck.ext2 |
19:04.03 | kerio | seems so |
19:04.17 | kerio | and stop calling it fsck.ext2 like it's something different from fsck.ext3 or e2fsck |
19:04.30 | kerio | does it work on your end without the trailing slash? |
19:04.54 | Pali | http://pastebin.com/xaDjhU9R |
19:05.08 | kerio | yeah, yeah, i got that too |
19:05.20 | kerio | the mountpoint is "/home", though |
19:05.21 | kerio | not /home/ |
19:05.45 | Pali | yes, working now: http://pastebin.com/PpthS5mE |
19:06.00 | Pali | detected as ext3 |
19:06.01 | Pali | ok |
19:06.47 | kerio | DocScrutinizer05: i'm trying to turn a 'fsck.`grep /home /etc/fstab | cut -d" " -f3` `grep /home /etc/fstab | cut -d" " -f2' into a 'fsck /home' in your future... rcS-late? preinit? where is this? |
19:07.01 | Pali | kerio, ok, then I only call "fsck -p /home" |
19:07.02 | kerio | rcS-late |
19:07.05 | Pali | rcS-late |
19:07.11 | kerio | Pali: use -a, -p isn't standard |
19:07.15 | kerio | it's the same for e2fsck, though |
19:07.47 | kerio | ...and if we assume that /home must be on extfs, or something whose fsck supports -p, we can just use -p i guess |
19:07.51 | Pali | for vfat too? |
19:07.59 | DocScrutinizer05 | > >For most types all the mount program has to do is issue a simple mount(2) system call, and no detailed knowledge of the filesystem type is required. For a few types how- ever (like nfs, nfs4, cifs, smbfs, ncpfs) ad hoc code is necessary. The nfs, nfs4, cifs, smbfs, and ncpfs filesystems have a separate mount program. In order to make it possible to treat all types in a uniform way, mount will |
19:08.01 | DocScrutinizer05 | execute the program /sbin/mount.TYPE (if that exists) when called with type TYPE. Since various versions of the smbmount program have different calling conventions, /sbin/mount.smbfs may have to be a shell script that sets up the desired call.<< |
19:08.29 | kerio | Pali: dosfsck doesn't have a -p option i think |
19:08.35 | Pali | hahahaha, we have messybox version of mount |
19:08.45 | Pali | and there autodetection not working properly :D |
19:08.47 | kerio | oh nvm, it still understands -p |
19:09.09 | kerio | i think that -a and -p are completely synonymous |
19:09.21 | DocScrutinizer05 | > >If no -t option is given, or if the auto type is specified, mount will try to guess the desired type. Mount uses the blkid library for guessing the filesystem type; if that does not turn up anything that looks familiar, mount will try to read the file /etc/filesystems, or, if that does not exist, /proc/filesystems. All of the filesystem types listed there will be tried, except for those that are |
19:09.22 | DocScrutinizer05 | labeled "nodev" (e.g., devpts, proc and nfs). If /etc/filesystems ends in a line with a single * only, mount will read /proc/filesystems afterwards.<< |
19:10.07 | kerio | DocScrutinizer05: fsck, not mount |
19:10.23 | Pali | DocScrutinizer05: this is not tru for our stupid version of mount provided by busybox |
19:10.23 | DocScrutinizer05 | who the fuck cares about fsck? |
19:10.35 | kerio | has strong opinions regarding fsck |
19:10.45 | Pali | it really cannot detect ext4 filesystem |
19:10.46 | kerio | the new rcS-late will support automatic preening on boot |
19:10.58 | kerio | Pali: e2fsck is a single binary ;_; |
19:11.37 | kerio | ext4 doesn't even exist as a concept, in e2fsprogs, it's just a set of default options to mke2fs |
19:11.55 | DocScrutinizer05 | fsck can't detect *anything* since you're *usually* dealing with a _broken_ fs when it comes to fsck, and a broken fs is impossible to reliably detect it's type |
19:12.18 | kerio | DocScrutinizer05: if you can't detect its type, then it's too borked to fix already |
19:12.26 | kerio | often, it's only slightly borked |
19:12.27 | DocScrutinizer05 | BS |
19:12.38 | kerio | missing extfs superblock? |
19:12.49 | kerio | like, missing extfs superblock and missing all the superblock backups |
19:12.59 | DocScrutinizer05 | pff, so one single block got overwritten? |
19:13.15 | kerio | various blocks |
19:13.27 | DocScrutinizer05 | you can not reliably detect a broken fs' type |
19:13.32 | kerio | not reliably |
19:13.48 | DocScrutinizer05 | that's a basic truth |
19:13.53 | kerio | but if the broken fs is not too broken to be unfixable, you can often detect it |
19:14.03 | kerio | and regardless, fsck uses /etc/fstab data too |
19:14.07 | DocScrutinizer05 | that's also why fsck never gets executed automatically in fix-all mode |
19:14.09 | Pali | DocScrutinizer05: now we tested that generic fsck using /etc/fstab and blkid for detecting filesytem |
19:14.32 | Pali | and my scripts used fstab/blkid too |
19:14.52 | Pali | so there is really no need to use long scripts for that... |
19:15.23 | kerio | trusts Pali, but trusts the e2fsprogs authors more |
19:15.36 | DocScrutinizer05 | shakes head on what it might be you guys try to do here, and wanders off |
19:15.43 | kerio | are they also the kernel guys? |
19:16.29 | kerio | DocScrutinizer05: more configurability in rcS-late regarding the mounting of /home and the generation of fstab |
19:16.49 | kerio | including (Pali: optional?) automatic preening |
19:16.59 | DocScrutinizer05 | so wtf has fsck to do with that?? |
19:17.13 | kerio | aka fsck in "do a summary check and only fix broken stuff if you're certain you can fix it" mode |
19:17.17 | Pali | optionaly call fsck before mounting |
19:17.33 | Pali | ext* filesystems need to be checked every N mounts |
19:17.54 | DocScrutinizer05 | ooh you are sure about thatß |
19:17.59 | DocScrutinizer05 | ? |
19:18.10 | kerio | always uses -C 0 -i 0 when mke2fsing |
19:18.15 | Pali | and in nokia original rcS-late was check: if mounting /home failed then try to run fsck |
19:18.41 | Pali | DocScrutinizer05: this is default behaviour of mkfs.ext3 |
19:18.49 | Pali | do you think that is incorrect? |
19:18.57 | kerio | mkfs.ext4 -C -1 -i 0 /dev/mmcblk0p1 #YOLO |
19:19.22 | DocScrutinizer05 | yes, and it was rather idiotic since it didn't inform user, not to mention any sane thing like opening console to show diagnostic output and even ask for user interaction to fsck |
19:19.27 | Pali | why every desktop linux distribution using that default behaviour too when installing |
19:20.48 | kerio | DocScrutinizer05: my laziness regarding the preparation or the acquisition of food is winning over my hunger, i think i'm going to follow your diet method this evening |
19:22.09 | *** join/#maemo-ssu xes (~xes@unaffiliated/xes) |
19:56.02 | *** join/#maemo-ssu _rd (~rd@p57B4B639.dip0.t-ipconnect.de) |
20:03.06 | *** join/#maemo-ssu Vlad_on_the_road (~Vlad_on_t@ip-66.net-82-216-1.versailles2.rev.numericable.fr) |
20:28.45 | Pali | kerio, now when I enabled auto fsck on my /home (n900) it show me very bad message |
20:28.56 | Pali | filesystem is corrupted |
20:28.58 | kerio | :c |
20:29.08 | kerio | is it corrupted? |
20:29.23 | Pali | I backed up (via dd) full partition to pc (via usb) |
20:29.37 | Pali | and started calling fsck -y on home device |
20:29.48 | Pali | fsck wrote a lot of lot of errors |
20:29.53 | Pali | and *crashed* |
20:29.58 | kerio | neat |
20:30.07 | Pali | and now fsck crashing always |
20:30.35 | Pali | but (before calling fsck) partition is/was mountable and maemo worked |
20:30.50 | kerio | btw, we should upgrade e2fsprogs |
20:31.09 | Pali | not going to extract data from backup on pc and I will call mkfs.ext3 on /home |
20:31.25 | kerio | Pali: again, you keep saying "mkfs.ext3" as if to say that it changes things |
20:31.56 | Pali | I will create clean fileystem |
20:32.03 | kerio | Pali: call mkfs.ext2 on it, too |
20:32.05 | Pali | and then copy files extracted from dump |
20:32.19 | Pali | seems there is no way to normal recovery |
20:32.32 | kerio | you could just copy the files from the backup you made before tinkering with filesystems |
20:32.56 | kerio | :> |
20:33.11 | Pali | I have dd backup before fsck was called |
20:33.23 | kerio | call maemo's fscks on it |
20:33.45 | Pali | I called fsck on emmc device |
20:34.20 | Pali | rather create new clean filesytem |
20:34.27 | kerio | to test, i mean |
20:34.29 | Pali | and restore data |
20:34.33 | *** join/#maemo-ssu _rd (~rd@p57B4B639.dip0.t-ipconnect.de) |
20:35.20 | kerio | Pali: anyway, if there was a difference, they would've put a way to specify the filesystem features in `e2fsck` |
20:36.21 | kerio | most likely, we have a crap version of e2fsprogs |
20:37.37 | Pali | hm, now called fsck -n on desktop on that dd backup and it wrote only small number of Inode XYZ was part of the orphaned inode list |
20:37.58 | Pali | this looks like on maemo is really crap version of e2fsprogs |
20:38.07 | kerio | "crap" = old |
20:38.34 | Pali | freemangordon: it looks like we need new version of e2fsprogs in cssu... |
20:39.18 | kerio | i don't think e2fsprogs has particular library requirements |
20:39.37 | kerio | it *might* be as simple as grabbing the latest debian source package and compiling it |
20:51.41 | *** join/#maemo-ssu Woody14619b (~Woody@2620:4:4000:11:cde8:f755:5dd8:8bb0) |
20:58.02 | Pali | kerio: on desktop fsck fixed that dump without any problems (it removed only 10 inodes which was orphaned) |
20:58.18 | Pali | should now I dd that dump back to n900 eMMC? |
20:58.26 | kerio | just put the files back |
20:58.31 | kerio | make a new fs from the desktop |
20:58.47 | Pali | ok |
21:08.38 | *** join/#maemo-ssu amiconn (amiconn@rockbox/developer/amiconn) |
21:19.13 | ShadowJK | Pali; re."fs was mountable before calling fsck", I've discovered it often behaves like that on flash |
21:19.59 | Pali | ShadowJK, fsck in check mode show me error that manual fix is needed |
21:20.19 | Pali | so I called fsck -y from backupmenu |
21:20.30 | Pali | and this totally damaged filesystem... |
21:20.36 | ShadowJK | yes. |
21:21.03 | Pali | I called desktop version of fsck on cloned dd image |
21:21.32 | Pali | and on desktop (new) fsck version fixed filesystem without problem |
21:22.09 | Pali | ShadowJK: so in maemo we have old version of fsck which cannot recover errors in ext3 |
21:22.46 | ShadowJK | I think it just gets confused by the way flash corrupts data |
21:23.08 | kerio | still, a newer e2fsck might get confused less |
21:23.21 | kerio | ShadowJK: it's not like it wasn't corrupt to begin with, though |
21:23.22 | kerio | is it |
21:23.49 | kerio | you just didn't notice |
21:23.50 | ShadowJK | Yeah but the kernel seems to often deal with it better |
21:38.12 | *** join/#maemo-ssu nox- (noident@freebsd/developer/nox) |
22:01.46 | *** join/#maemo-ssu arcean (~arcean@apn-77-112-67-219.dynamic.gprs.plus.pl) |
22:06.57 | Pali | kerio: I pushed ke-recv and upstart packages to cssu-devel |
22:07.10 | Pali | source code is on gitorious in my cloned repos |
22:07.30 | kerio | i'll test them tomorrow morning before uni, if i remember to do so |
22:07.35 | Pali | you can test static fstab, fsck on non ext3 parts, ... |
22:07.49 | kerio | alright |
22:08.11 | Pali | now rcS-late generating correct fstab for maemo copied to eMMC |
22:30.02 | *** join/#maemo-ssu M4rtinK (~M4rtinK@46.227.13.46) |
23:11.47 | *** join/#maemo-ssu M4rtinK (~M4rtinK@ip-86-49-81-87.net.upcbroadband.cz) |