1160352192 P * michael 1160355018 Q * azazel Quit: Client exiting 1160360056 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1160360066 J * ensc ~irc-ensc@p54B4D51C.dip.t-dialin.net 1160362334 J * laomiao ~laomiao@HSE-Toronto-ppp306428.sympatico.ca 1160362338 P * laomiao 1160364434 N * root_ mz 1160364723 J * pinskia ~pinskia@adsl-68-123-29-99.dsl.pltn13.pacbell.net 1160364726 P * pinskia Leaving 1160367761 Q * meebey Ping timeout: 480 seconds 1160367776 Q * nib-nico Ping timeout: 480 seconds 1160367782 Q * m4z Ping timeout: 480 seconds 1160368041 Q * kir Ping timeout: 480 seconds 1160368339 J * kir ~kir@swsoft-mipt-nat.sw.ru 1160369329 N * transaci1 transacid 1160371456 Q * _node Ping timeout: 480 seconds 1160371595 J * marl_ ~matt@84.92.193.226 1160371595 Q * marl Read error: Connection reset by peer 1160371714 J * nammie ~nam@S0106001195551ff0.va.shawcable.net 1160372146 Q * Nam Ping timeout: 480 seconds 1160373484 J * pinskia ~pinskia@adsl-68-123-29-99.dsl.pltn13.pacbell.net 1160373574 P * pinskia Leaving 1160374649 J * meandtheshell ~markus@85-125-231-154.dynamic.xdsl-line.inode.at 1160377178 Q * Aiken Remote host closed the connection 1160377972 Q * mz Ping timeout: 480 seconds 1160378551 J * mz ~root@tor-irc.dnsbl.oftc.net 1160379251 Q * gex- Quit: Lost terminal 1160379658 J * lilalinux ~plasma@h1-gw.of.net-lab.net 1160380227 J * dna_ ~naucki@p54BCEF69.dip.t-dialin.net 1160380593 Q * _are_ Ping timeout: 480 seconds 1160380740 M * matti Morning. 1160382058 J * m4z ~m4z@bastard-operator.from-hell.net 1160382793 J * FireEgl FireEgl@Sebastian.Atlantica.US 1160383421 J * _are_ ~are@stgt-d9be64d7.pool.mediaWays.net 1160384307 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr 1160384622 M * h01ger http://linux-vserver.org is down 1160384675 Q * h01ger Remote host closed the connection 1160384678 M * Loki|muh I second that 1160384689 J * h01ger ~holger@socket.layer-acht.org 1160384711 M * daniel_hozac hmm, so it seems... 1160384933 Q * shedi Quit: Leaving 1160385001 J * flob ~flob@c169053.adsl.hansenet.de 1160385026 M * flob i there, is the page down? 1160385033 M * flob hi not i ... 1160385053 M * Loki|muh it is 1160385061 M * flob :( thx 1160385169 Q * Hunger Ping timeout: 480 seconds 1160385303 M * daniel_hozac does it work now? 1160385315 M * Loki|muh yes 1160385318 M * Loki|muh :) 1160385329 M * Loki|muh what was the problem? 1160385386 M * daniel_hozac seems to be the Gentoo-baselayout problem, the start scripts weren't starting the services but are waiting for a cat of a pipe to finish... 1160385422 M * Loki|muh hmm, there are still problems with gentoo-guests? thats bad :( 1160385466 M * daniel_hozac i don't know if it's the same problem, but that problem is new ;) 1160385673 J * User57 ~User57@85.106.129.159 1160385679 M * daniel_hozac can't be the same problem, i don't think... it's got the wrong baselayout installed. 1160385714 Q * User57 1160385721 M * daniel_hozac (wrong meaning correct ;)) 1160386361 J * s0undt3ch_ ~s0undt3ch@bl7-254-36.dsl.telepac.pt 1160386505 Q * s0undt3ch Ping timeout: 480 seconds 1160386505 N * s0undt3ch_ s0undt3ch 1160386892 J * shedi ~siggi@dsl-149-109-85.hive.is 1160386896 Q * cehteh Ping timeout: 480 seconds 1160386946 J * cehteh ~ct@pipapo.org 1160387317 M * meandtheshell daniel_hozac: I was just wondering why harry's Linux-VServer + grsecurity patch is not placed at http://ftp.linux-vserver.org/pub/kernel/ ? Why is that? 1160387414 M * meandtheshell sure - I know it's in harry's personal directory and on the wikis mainsite but not in the ftp directory (would be gread since I love to use my command line client) 1160387456 M * meandtheshell daniel_hozac: could you upload it? 1160387573 M * daniel_hozac meandtheshell: i'd like to put the entire people space on FTP, but i think we need to discuss how to make that happen... 1160387640 M * meandtheshell daniel_hozac: I see - of course that would be best :) 1160387649 M * daniel_hozac seems like a bind mount per person might not scale very well. 1160387671 M * meandtheshell hm ... 1160387695 A * meandtheshell hasn't any clue how the infrastructure is maintained 1160387841 M * daniel_hozac well, we all have our home directories in /home, and then ~/public_html is the people space. that would have to be somehow made available in /var/ftp as well... 1160387875 M * daniel_hozac i suppose putting the actual directories in /var/ftp and then symlinks from ~/public_html might do it. 1160387932 M * daniel_hozac i'd like to discuss it with everyone else before i go around doing things though ;) 1160388086 M * meandtheshell daniel_hozac: sure - having all the ../people stuff available with ftp eases things a lot - not a lot effort to do but a lot more comfort for CLI ftp people :) 1160388876 J * Hunger Hunger.hu@213.163.11.138 1160390156 J * s0undt3ch_ ~s0undt3ch@bl7-245-32.dsl.telepac.pt 1160390285 Q * s0undt3ch Ping timeout: 480 seconds 1160390285 N * s0undt3ch_ s0undt3ch 1160393294 Q * flob Quit: Ciao! 1160395701 M * gdm hi, a quick question about limits... 1160395720 M * gdm we run a webserver adn it stopped working at a certain point this morning - just apache 1160395731 M * gdm and when i look at the limits: 1160395732 M * gdm gdm@kompost:~$ cat /proc/virtual/2/limit 1160395732 M * gdm PROC: 9 238 750 0 1160395732 M * gdm VM: 7450 1047965 1048576 3 1160395759 M * gdm so i am wondering if actually it just hit the limit and the kernel killed it. is this a possibility? 1160395837 M * daniel_hozac well, hitting the VM limit causes ENOMEM, IIRC. 1160395854 M * gdm what is ENOMEM? 1160395859 M * daniel_hozac out of memory. 1160395902 M * gdm so would it kill the webserver in that vserver if that was the cause? or one of the causes? 1160395917 M * gdm or should i worry about getting cracked in someway, do you think? 1160395944 M * daniel_hozac no, hopefully your webserver handles ENOMEM gracefully and exits. 1160396229 Q * FireEgl Read error: No route to host 1160396268 N * Bertl_zZ Bertl 1160396272 M * Bertl morning folks! 1160396390 M * daniel_hozac morning Bertl! 1160396602 M * daniel_hozac how are you doing today? 1160396648 M * gdm daniel_hozac: thanks for your help. i think that is what happened 1160396824 M * Bertl daniel_hozac: well, still headache, a little dizzy, but it seems to get better (knocking on wood) 1160396893 M * daniel_hozac sounds good. 1160396911 M * daniel_hozac (the getting better part, i mean) 1160396972 J * FireEgl FireEgl@Sebastian.Atlantica.US 1160397007 M * Bertl here is a funny ;) patch: http://vserver.13thfloor.at/Experimental/delta-smon-fix02.diff 1160397127 M * Bertl and here is the latest userspace version of the dump tool 1160397129 M * Bertl http://vserver.13thfloor.at/Experimental/TOOLS/vdump-0.03.tar.bz2 1160397340 M * Bertl ah, before I forget: 1160397342 M * Bertl http://www.linuxworldexpo.de/vkf_06.php?ID=122 1160397370 M * daniel_hozac cool! 1160397593 M * Bertl yeah, a really good placement IMHO too :) 1160397836 M * daniel_hozac yeah. 1160397995 M * meandtheshell and after a coffee break too - that's always good mv /coffee/super_cow_attention /audience/people :) 1160398345 M * nayco_work Hello, all :-)! 1160398357 M * daniel_hozac hi nayco_work 1160399822 M * FaUl ahuhu 1160399835 M * FaUl how do I mount some filesystem into a vserver without restarting it? 1160399999 M * daniel_hozac vnamespace -e mount... 1160400093 M * essobi_ ~WEEE! 1160400438 M * FaUl daniel_hozac: how do i get it into /etc/mtab of the new vserver then? 1160400566 M * FaUl daniel_hozac: i need something like vnamespace -e jabber mount /mnt/crypto/jabber /vserver/jabber/mnt/crypto, but if I do that it isn't in the vservers fstab 1160401058 M * daniel_hozac cd /vservers/jabber; vnamespace -e jabber secure-mount --chroot /mnt/crypto/jabber /vservers/jabber/mnt/crypto 1160401137 M * FaUl ah 1160401182 M * daniel_hozac oh, you'll have to specify the full path to secure-mount, it should be in /lib/util-vserver or so. 1160401302 M * FaUl mh 1160401376 M * FaUl vnamespace -e jabber /usr/lib/util-vserver/secure-mount --bind --chroot /mnt/crypto/jabber/ /mnt/crypto/ 1160401380 M * FaUl mount: special device /mnt/crypto/jabber/ does not exist 1160401413 M * daniel_hozac i take it /mnt/crypto is something you've mounted after starting the guest? 1160401432 M * FaUl maybe 1160401486 M * FaUl it's cryptomount which I mounted while booting, i dont know exactly if the jabber-vserver was started then 1160401543 M * FaUl is this importent? i realy need to get into that vnamespace-stuff 1160401614 M * daniel_hozac yes, if it's mounted after you'll have to mount it inside the namespace as well, to make it available inside the namespace as well. 1160401631 M * daniel_hozac ... +sense +english 1160401633 M * FaUl how can i mount this? loop-mount, again? 1160401653 M * FaUl /dev/loop0 /mnt/crypto ext2 rw 0 0 1160401663 M * FaUl this is what the host sayes 1160401665 M * daniel_hozac yep, but i don't know if that'll cause problems. Bertl? 1160401686 M * FaUl no way to get around this without restarting the guest? 1160401707 M * daniel_hozac not that i know of. 1160401716 M * FaUl ok 1160401728 M * FaUl then I'll try to announce some jabberserver-restart 1160401762 M * FaUl ;-) 1160402021 M * Bertl hmm, what do you try to do? 1160402046 M * FaUl Bertl: i got some jabber-vserver and the admin would like the jabber-user-database on cryptofs 1160402082 M * Bertl okay, there is dm-crypt for that (if you mean 'encrypted ext2/3') 1160402104 M * Bertl and what was the problem with such an approach? 1160402113 M * FaUl the host already has cryptofs and i try to mount one directory of this crypt-container into the vserver without restarting it 1160402138 M * Bertl cryptofs is a filesystem or what? 1160402144 M * FaUl i use loopaes because i consider it more secure 1160402165 M * Bertl so you have a cryptoloop with a normal filesystem mounted 1160402166 M * FaUl it is ext2 in a loop-aes-container 1160402169 M * FaUl yes 1160402172 M * FaUl exactly 1160402190 M * Bertl well, that's pretty insecure, so bind mounting should work 1160402206 M * FaUl the problem seems to be that it is mounted after the vserver is started 1160402210 M * FaUl why is it insecure? 1160402234 M * Bertl once the crypto loop is set up, everybody can access it 1160402242 M * FaUl yes 1160402284 M * Bertl you can also mount it from scratch (new loop device) inside the guest 1160402314 M * FaUl it's about how you define the attack-secenario - in our case it is if someone from "team gruen" takes the server they don't get the data ;-) 1160402339 M * Bertl i.c. okay, that should work, if swap is encrypted too :) 1160402368 M * FaUl Swap: 0 0 0 1160402374 M * FaUl it should no problem so ;-) 1160402379 M * Bertl okay 1160402411 M * Bertl so try to unmount it outside, and mount it again inside then? 1160402456 M * FaUl Bertl: we got on the host /mnt/crypto with directorys jabber, mail, kasse and so on 1160402463 M * FaUl so unmounting on the host is not an option 1160402490 M * Bertl but that wasn't there before the guest(s) were started 1160402540 M * FaUl well, i restarted mail and kasse as some seconds downtime does not harm here, but in the jabber case it disconnects every user which is not to nice 1160402581 M * FaUl the jabber server is now for at least 10 days online and the idea to crypt stuff inside developed an hour ago ;-) 1160402615 M * Bertl i.c. well, in this case a sheduled restart is the only option I see too 1160402656 M * FaUl ok, well then 1160402665 M * FaUl i'll urge the jabber-admin to announce a downtime ;-) 1160402853 J * simon00 ~simon00@host154-158-static.47-85-b.business.telecomitalia.it 1160402859 M * simon00 hi all 1160402898 M * simon00 someone can help me with bind inside a vserver? 1160402981 M * FaUl simon00: maybe 1160402989 M * simon00 :-) 1160403048 M * simon00 i get an "operation Not permitted, ensure that capset kernel module is loaded" 1160403054 M * simon00 workaround? 1160403166 M * Bertl bind as in named? 1160403175 M * simon00 yes, bind is named 1160403188 M * FaUl ah, now i got it, too ;-) 1160403191 M * Bertl well, three options there 1160403198 M * simon00 yes.... 1160403205 M * Bertl - recompile bind with --disable-linux-caps (secure) 1160403217 M * Bertl - switch to a devel Linux-VServer kernel (secure) 1160403235 M * Bertl - give the require capability (CAP_SYS_RESOURCE IIRC) to the guest (insecure) 1160403251 M * simon00 as this is an internal machine 1160403264 M * simon00 i go for the 3rd solution 1160403266 M * simon00 but.... 1160403282 M * Bertl yup? 1160403297 M * simon00 i tried tu put the line S_CAPS="CAP_SYS_RESOURCE" in .conf 1160403309 M * simon00 in /etc/vservers 1160403315 M * simon00 but nothing happens 1160403323 M * Bertl if you have a .conf, then you should update ASAP :) 1160403344 M * Bertl the config is directory based since more than a year now 1160403349 M * simon00 i create a .conf same name of the virtual server 1160403369 M * simon00 the path is right? 1160403392 M * Bertl nope, it would be: /etc/vservers//bcapabilities 1160403427 M * Bertl see here for details of the config: http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1160403440 M * simon00 i have to create a file named bcapabilities 1160403598 J * St7 ~ooo3000@tor-irc.dnsbl.oftc.net 1160403604 M * St7 hi all 1160403624 M * St7 hi alll 1160403630 M * simon00 hi 1160403642 M * simon00 thank you bertl 1160403645 M * St7 please tell me my IP 1160403645 M * simon00 solved 1160403691 M * St7 any please 1160403709 M * St7 please tell me my IP 1160403755 M * mnemoc you are tor-ed 1160403794 M * Bertl St7: I dedicate the following ip to you: 127.23.42.97 1160403814 M * Bertl St7: you can use it wherever you like :) 1160403837 M * St7 TY 1160403890 P * St7 1160403955 Q * shedi Quit: Leaving 1160404067 J * Piet hiddenserv@tor.noreply.org 1160405422 Q * doener Ping timeout: 480 seconds 1160405460 J * doener ~doener@host.magicwars.de 1160406159 J * meebey meebey@booster.qnetp.net 1160406186 M * Bertl wb doener! welcome Piet! meebey! 1160406214 M * doener irssi just died an interesting death... 1160406238 M * doener both, screen and irssi were hanging in sys_write trying to write to stderr... 1160406422 M * Borg- perl... 1160406494 Q * meebey Remote host closed the connection 1160406568 M * Bertl doener: interesting ... 1160406586 M * Bertl doener: I hope you didn't spend too much time on the scheduling monitor stuff 1160406614 M * doener I tried to feed some data to rrd, but failed... 1160406644 M * doener maybe 3-4 hours in total, most of the time reading up on stuff that I didn't use in the end 1160406646 M * Bertl the latest vdump can decode the events quite fine (so that should work) 1160406673 M * Bertl if you want to work/test with it, it could benefit from a realtime guesstimation 1160406720 M * Bertl http://paste.linux-vserver.org/474 1160406776 M * Bertl and of course, deltas for activation/deactivation cycles and such would be nice to have too 1160406793 M * Bertl (assuming that you still want to work on that :) 1160406862 M * doener after giving up on rrd and discovering that uts and ipc namespaces made it into linus' tree, I thought it might be worth to port 2.1.1 to .19-rc1 1160406888 M * Bertl yeah, that would be interesting too 1160406938 M * doener [addr,histno] action [context,?] 1160406941 M * Bertl (read: no problem with that, the current vdump should suffice for the testing I plan to do) 1160406946 M * doener (data format) 1160406949 M * doener is that right? 1160406960 M * Bertl [cpu,jiffies,sequence] 1160406975 M * doener ah, there's another comma 1160406979 M * Bertl then for tasks: pid [context,state] 1160407021 M * doener what's a guesstimation? 1160407033 M * Bertl well, you see the sync lines? 1160407039 M * Bertl [0,ffffb914] sync 1160339737.389.999.999 1160407060 M * Bertl they basically tie the jiffie value (ffffb914) to the xtime 1160407061 A * doener .oO( What's a battle? What's a diorama? ) 1160407065 J * meebey meebey@booster.qnetp.net 1160407089 P * simon00 1160407110 M * Bertl so, we only have jiffies for the other events, but some interpolated xtime would be nice to have 1160407126 M * Bertl (that's the purpose of the sync events after all) 1160407201 Q * brc_ Ping timeout: 480 seconds 1160407301 M * doener ah! not a guesstimation in realtime but of realtime! :) 1160407322 M * Bertl yep, that's it :) 1160407682 M * h01ger do i read http://oldwiki.linux-vserver.org/HowTo+Read+ProcFS correctly, that in /proc/virtual//limit the values for RSS, VM, VML, ANON and SHM need to be multiplied by $pagesize to give meaningful (to humans) values? 1160407718 M * Bertl yep, if you do not consider pages meaningful 1160407753 M * h01ger Bertl, even worse: i dont consider people not to be human, who consider pages meaningful :-D 1160407757 M * h01ger Bertl, thanks 1160407765 M * h01ger s/not// 1160407778 M * Bertl oh :( 1160407780 M * daniel_hozac how are pages not meaningful? 1160407795 A * h01ger used to think in megabytes, not pages 1160407800 M * h01ger "is used" even 1160407864 M * h01ger micah, do i read 'VSERVERS="$vservers"' in vserver_resources correctly, that a.) $vservers is from the plugin-configfile and b.) that its a space-seperated list, i.e. "VSERVERONE TWO THREE"? 1160408047 M * micah h01ger: yes to both questions 1160408159 M * h01ger cool. thought so, but its nice to have this confirmed :) 1160408202 M * h01ger micah, i will refactor the code a bit and then i think its really ready 1160408224 M * micah yeah I think I lost that comment that made that clear 1160408677 J * stefani ~stefani@tsipoor.banerian.org 1160408690 M * Bertl welcome stefani! 1160408696 M * stefani hola 1160408978 Q * FireEgl Ping timeout: 480 seconds 1160409633 J * lilalinux_ ~plasma@dslb-084-058-197-049.pools.arcor-ip.net 1160409852 Q * lilalinux Ping timeout: 480 seconds 1160411995 Q * Loki|muh Remote host closed the connection 1160411997 J * Loki|muh loki@satanix.de 1160412813 Q * _are_ Ping timeout: 480 seconds 1160414481 J * boss ~boss@ANancy-152-1-19-178.w83-194.abo.wanadoo.fr 1160414587 Q * boss Quit: BitchX by any other name is still BitchX 1160414659 J * brc_ ~bruce@201.19.171.203 1160414714 M * lilalinux_ is /etc/passwd something special in vserver? 1160414793 M * Bertl nope 1160414798 M * Bertl why do you think so? 1160414857 Q * prae Quit: Quitte 1160415210 M * lilalinux_ Bertl: we have an old server here. the former owner wrote some strange backup scripts, that need to rewrite /etc/passwd. dont ask me why, i don't know. however the script is bailing out when trying to write to that file 1160415760 M * Bertl maybe strange permissions? 1160415906 M * Bertl I mean, usually /etc/passwd is not allowed for anybody except root 1160415913 M * Bertl (to be written to) 1160416022 A * matled found out that the immutable flag (dunno how I enabled it on many files for one vserver) makes some programs behave very strange 1160416207 M * Bertl well, there was a long lasting bug on debian (i.e. the maintainer got the flags wrong) which caused -i and especially -t flags to be everywhere (a year ago or so) 1160416245 M * Bertl and it's not unexpected that both, the immutable flag and the immutable-linkage-invert, give strange results when applied to files 1160417839 Q * lilalinux_ Remote host closed the connection 1160418527 J * id23 ~id@p50814414.dip0.t-ipconnect.de 1160418539 M * id23 greetings #vserver 1160418685 M * Bertl greetings id23! 1160418944 M * id23 Hi Bertl 1160419001 Q * Piet Read error: Connection reset by peer 1160419101 M * h01ger micah, you might want to look at the plugins in svn, i consider the code ready and am testing it now 1160419115 A * h01ger is prepared to collect ugly pieces :) 1160419161 J * Piet hiddenserv@tor.noreply.org 1160419237 J * bonbons ~bonbons@83.222.36.111 1160419378 Q * Piet Remote host closed the connection 1160419512 J * Piet hiddenserv@tor.noreply.org 1160419754 Q * Piet Remote host closed the connection 1160419794 J * Piet hiddenserv@tor.noreply.org 1160420467 M * ex i'm trying to mount to vserver filesystem (squashfs) mounted in main host (loop), it it possible? 1160420519 M * ex vnamespace with mount --bind dont work, it looks like it's mounting normal part. on which is squashfs mounted 1160420535 M * daniel_hozac was the squashfs mounted when the guest was started? 1160420550 M * ex no 1160420555 M * daniel_hozac that'd be why then. 1160420560 M * ex when i'm using fstab, it's working 1160420571 M * daniel_hozac as it should. 1160420588 M * ex so this setup will not work? 1160420590 M * ex :( 1160420593 M * daniel_hozac what setup? 1160420607 M * ex squashfs mounted after vserver start 1160420618 M * daniel_hozac then you'd have to mount it inside the guest's namespace from the start. 1160420646 M * daniel_hozac (i.e. don't mount it on the host at all) 1160420694 M * ex i can mount squashfs into vserver, but with several vservers i have to do it several times (and i guess squashfs take some extra memory) 1160420708 M * ex so i prefer to mount it once on main server 1160420720 M * daniel_hozac so why aren't you mounting it from fstab, if that's possible? 1160420758 M * ex it's a option :) 1160420766 M * ex i'm using squashfs for gentoo portage 1160420789 M * ex so when i upgrade it (not often) i've to restart vserver 1160420890 M * daniel_hozac ah. 1160421073 M * ex anyway, thx for info 1160422068 M * doener daniel_hozac: don't you get some kind of bind mount if you mount the same thing multiple times? or is squashfs special? 1160422128 M * daniel_hozac you do? i don't really know how that works at all. 1160422183 M * doener IIRC you used to get "already mounted" errors "back then", when you tried to mount the same fs twice, as that's no longer the case... 1160422208 M * doener but let me take a look at the source 1160422256 M * Bertl yes, usually the superblock is already there 1160422268 M * Bertl so it is just mapped into the 'new' namespace 1160422341 Q * hardwire Ping timeout: 480 seconds 1160423014 J * hardwire ~hardwire@89-208-58-66.gci.net 1160424263 M * daniel_hozac Bertl: heh, took me a while to get the debug_cmd.h report. you're missing the semicolons at the end of the structs ;) 1160424294 M * daniel_hozac (in the #ifdef CONFIG_COMPAT) 1160424366 M * Bertl ah, well, patch? 1160424402 M * Bertl haven't tested with all archs yet, have to submit it to PLM anyways, but was waiting for your cow patches ... 1160424418 M * Bertl btw, did I miss an url for the updated cow features? 1160424476 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-debug_cmd-fix01.diff 1160424490 M * daniel_hozac hmm, i sent it right after you went to bed, IIRC, http://people.linux-vserver.org/~dhozac/p/k/delta-cow-feat05.diff 1160424538 M * Bertl ah, so I simply missed it then! tx! 1160425258 M * Bertl hmm, I get a warning from gcc regarding new_dentry, but I do not see what causes that, did you encounter that one too? 1160425266 M * daniel_hozac hmm. 1160425282 M * Bertl fs/namei.c: In function `cow_break_link': 1160425282 M * Bertl fs/namei.c:2703: warning: `new_dentry' might be used uninitialized in this function 1160425310 M * daniel_hozac i guess it's the return causing it? 1160425353 M * doener Bertl: from cow-feat5? 1160425364 M * Bertl ah, yes, we can jump over the assignment actually 1160425380 M * Bertl so it is probably the best to initialize it to zero 1160425411 M * daniel_hozac but then err shouldn't be 0. 1160425436 M * daniel_hozac err isn't set to 0 until the function has finished, and new_dentry has been dget'ed. 1160425517 M * Bertl yes, indeed, but the assignment should not cause to change the compiled code anyways 1160425536 M * daniel_hozac true. 1160425634 M * Bertl 1902693 49154 50672 2002519 1e8e57 fs/built-in.o 1160425639 M * Bertl 1902677 49154 50672 2002503 1e8e47 fs/built-in.o 1160425656 M * Bertl so it actually seems to benefit gcc here :) 1160425664 M * daniel_hozac hehe, indeed. 1160425723 M * doener hm, as long as no false assumption is the reason for the difference ;) 1160425863 M * Bertl daniel_hozac: hmm, the allow_cow argument ... 1160425876 M * daniel_hozac yes, it's rather ugly. 1160425885 M * Bertl (for chown_common) why is it 0 for sys_fchown()? 1160425912 M * daniel_hozac because then the file descriptor wouldn't point to the file anymore. 1160425940 M * Bertl hmm, shouldn't a change to file->f_dentry fix that? 1160425974 M * daniel_hozac ... i have no idea :) 1160425982 M * daniel_hozac what happens to mapped regions of the file in that case? 1160426017 M * Bertl hmm .. interesting thought 1160426019 M * daniel_hozac i suppose it doesn't really matter, as the file will only be opened for reading anyhow. 1160426060 M * doener Bertl: which gcc version do you have? 1160426074 M * daniel_hozac i figured it was easier to just not break links on f*, as it seemed problematic. 1160426084 M * Bertl 3.3.5/3.3.6 here 1160426129 M * doener 4.1.2 20061007 -- doesn't warn for new_dentry anymore 1160426148 M * doener funny as the concensus on lkml seems to be that gcc produces more and more useless noise 1160426158 M * daniel_hozac hehe. 1160426173 M * daniel_hozac is the kernel compiled with -Wall by default, or do i have to enable it somehow? 1160426229 M * doener it's on by default 1160426238 M * daniel_hozac 4.1.1 20060525, no warning here either. 1160426249 M * doener make V=1 will use the old mode that shows the full commandlines 1160426264 M * doener and make -s will show warnings/errors only 1160426286 M * daniel_hozac thanks! 1160426287 M * Bertl btw, can anybody tell me what 'make headers_check' means when it says: make: unifdef: Command not found 1160426302 M * daniel_hozac means you're missing some command? :) 1160426317 M * daniel_hozac http://freshmeat.net/projects/unifdef/ 1160426326 M * Bertl that really does exist? 1160426337 M * daniel_hozac sure looks like it. 1160426355 M * Bertl fascinating ... tx 1160426457 Q * bonbons Quit: Leaving 1160427044 M * doener http://people.linux-vserver.org/~doener/cow-fix.diff 1160427106 M * doener 21867 176 0 22043 561b fs/namei.o 1160427109 M * doener 21841 176 0 22017 5601 fs/namei.o 1160427112 M * daniel_hozac that is a lot nicer. 1160427121 M * Bertl yep, looks good to me too, tx 1160427348 J * shedi ~siggi@inferno.lhi.is 1160427629 M * doener daniel_hozac: btw, do you know "make help"? 1160427829 M * doener that was kind of a huge discovery back when we tried to get cross-compiles to work (because we discovered "defconfig" that way) 1160428105 M * daniel_hozac nice. 1160429620 T * Bertl http://linux-vserver.org/ <- new and shiny | latest stable 2.02.1, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.1.1-rc38, stable+grsec 2.0.2.1 | util-vserver-0.30.211 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1160429868 J * DreamerC_ ~dreamerc@61-217-226-183.dynamic.hinet.net 1160430055 Q * mire_ Quit: Leaving 1160430062 M * doener hm, reverse rebasing with git is a lot less painful... ie. not rebasing my vserver tree (with just the single big patch applied) on 2.6.19-rc1, but rebasing linus' tree on the vserver tree :) 1160430102 M * Bertl well, that's the way I would go anyway 1160430127 J * Aiken ~james@tooax6-174.dialup.optusnet.com.au 1160430133 M * Bertl welcome Aiken! 1160430152 M * doener well, "usually" you'd have a lot of small patches that would get rebased, not that single huge beast 1160430185 M * Bertl well you could check in all the delta's no? 1160430214 M * Bertl but even if you do that, I doubt that the other way round wouldn't be more efficient 1160430241 M * doener I'm just playing around with git, I'm not going to keep that tree, it will probably be FUBAR in a few hours ;) 1160430278 Q * DreamerC Ping timeout: 480 seconds 1160430319 P * stefani I'm Parting (the water) 1160430342 M * doener I'm not sure about that... especially when more stuff is moved into mainline, dropping the vserver patches might be easier to add "reverse" deltas to the git tree.... and the history will become messy 1160430383 M * Bertl well, we already separated out uts and pid (not ipc yet) 1160430407 M * Bertl after all. I expected uts and pid to be the first few namespaces 1160430436 M * doener separated out? 1160430462 M * Bertl the vs_pid.h stuff is basically in a separate patch 1160430485 M * Bertl (the split patches reflect this, not updated yet, though) 1160430494 M * Bertl similar is true for the utsname 1160431623 Q * id23 Remote host closed the connection 1160432650 M * doener hm, giving up on rebase... 1160432861 Q * Piet Remote host closed the connection 1160432937 J * Piet hiddenserv@tor.noreply.org 1160433106 Q * meandtheshell Quit: exit (0); 1160433110 J * mire ~mire@76-167-222-85.adsl.verat.net 1160433267 Q * DreamerC_ Quit: leaving 1160433284 J * DreamerC ~dreamerc@61-217-226-183.dynamic.hinet.net 1160433989 J * _are_ ~are@62.112.159.81 1160434117 Q * dna_ Quit: Verlassend 1160435161 Q * cryptronic Ping timeout: 480 seconds 1160435211 Q * BeLu_zZZz Ping timeout: 480 seconds 1160436516 Q * mz Remote host closed the connection 1160437644 M * mugwump hmm, why does `vserver xxx stop' not always run /etc/init.d/rc 6 1160437655 M * mugwump (on an old util-vserver, but still... confusing) 1160437666 J * mz ~root@tor-irc.dnsbl.oftc.net 1160437719 M * mugwump the vservers have no difference in /etc/vservers that I can see 1160437730 M * mugwump I looked for initstyle, but neither has it set 1160437736 M * doener what does it run in the other case? 1160437740 M * mugwump both have an executable /etc/init.d/rc 1160437743 M * mugwump it runs nothing 1160437758 M * doener got a "vserver --debug xxx stop" log? 1160437798 M * doener namespace enabled vserver with /vservers/xxx/etc/init.d/rc not being available on the host? 1160437851 M * mugwump is there a nopaste in the house? 1160437864 M * doener paste.linux-vserver.org 1160437896 M * mugwump http://paste.linux-vserver.org/475 1160438013 M * doener does this file exist on the host? /etc/vservers/srsvapp1/vdir//etc/init.d/rc 1160438027 M * mugwump yes 1160438029 M * mugwump and +x 1160438188 M * doener hm, dynamic context 1160438243 M * doener is the xid correct? 1160438250 M * doener 49185 1160438274 M * Bertl don't do that :) 1160438345 M * doener Bertl: reverse rebasing causes lots of non-vserver related merge failures 1160438359 M * Bertl interesting ...