1167178961 J * mkhl bairral@200-148-41-192.dsl.telesp.net.br 1167180279 Q * mkhl 1167180815 Q * yarihm Ping timeout: 480 seconds 1167181004 Q * dna Quit: Verlassend 1167181932 Q * FireEgl Read error: Connection reset by peer 1167184010 J * Kerberos ~satan@85.138.138.76 1167184011 Q * kaner Ping timeout: 480 seconds 1167184111 J * _dmax ~semaj@81.193.56.198 1167184221 Q * dmax Ping timeout: 480 seconds 1167184226 N * _dmax dmax 1167184268 Q * s0undt3ch Ping timeout: 480 seconds 1167184487 J * kaner kaner@strace.org 1167184573 J * Piet hiddenserv@tor.noreply.org 1167186341 Q * Piet Quit: Piet 1167189156 Q * dmax Ping timeout: 480 seconds 1167189267 J * dmax ~semaj@81.193.57.28 1167189505 J * s0undt3ch ~s0undt3ch@81.193.57.28 1167192842 Q * ntrs Remote host closed the connection 1167198207 J * justinwick ~justinwic@pool-151-201-42-221.pitt.east.verizon.net 1167198216 M * justinwick Hey all! 1167198220 N * justinwick JustinWick 1167198238 M * Bertl welcome JustinWick! 1167198249 M * JustinWick Hey! Any vserver devs lurking about? 1167198273 M * Bertl probably, what's up ... 1167198286 M * JustinWick Well... I have some good news for the vserver community :) 1167198299 M * Bertl let's hear then ... 1167198348 M * JustinWick Eric Tichansky, who runs lylix.net (VPS hosting based on VServer, asterisk, etc) wants to hire me to work on the next gen networking in VServer. His clients need that functionality, and we'd love to make a nice patch that could be selectable within the vserver main distro, according to the next gen roadmap. 1167198376 M * JustinWick But I wanted to chat with people about how they'd like to go about doing this... I really want this to be useful to everyone, and for it to be easy for the devs to continue to support when it works etc... 1167198399 M * Bertl nice, well I guess you found the right place 1167198434 M * Bertl what kind of functionality will be needed (according to Eric)? 1167198453 M * JustinWick Well, the basic idea as I understand it is to make it a lot like the OpenVZ implementation 1167198472 M * Bertl hmm, why not take OVZ then instead? 1167198475 M * JustinWick Every VPS should be secure from other VPSes in terms of things like local loopbacks and IP routing 1167198477 M * JustinWick so... you can't snoop etc 1167198512 M * JustinWick The net code would live on the main host and the virtual host both, but would try and be efficient so it wouldn't be much of a performance hit 1167198544 M * JustinWick (I'll warn you I'm still learning up on these things, I use a vserver for one of my utility servers, hosted @lylix in fact, but I've never directly used the software yet, going to start on that tomorrow etc) 1167198563 M * JustinWick Well, he has a whole stack based on VServer, and OpenVZ has a lot of weirdness in other areas he doesn't like. 1167198608 M * Bertl thing is, there are good reasons why Linux-VServer isn't like OVZ, mainly performance, but also clarity of code 1167198612 M * JustinWick Right right 1167198622 M * JustinWick I'm not saying that this would make it more like OVZ in a code sense 1167198626 M * JustinWick just in a "it can do this" sense. 1167198642 M * JustinWick VServer is lacking in this area, and there's this page: (http://oldwiki.linux-vserver.org/NGNET-Development) 1167198652 M * JustinWick That makes me think *someone* has been working on this, and really does want it to work well. 1167198654 M * Bertl so, simply adding OVZ features (like fully virtualized networking stack) won't cut the mustard :) 1167198693 M * JustinWick Well, I think he likes what's on that page there, the NGNET. 1167198719 M * Bertl I see .. well, I have good and bad news for you ... 1167198726 A * JustinWick listens 1167198772 M * Bertl the good news is, that current Linux-VServer development is already heading towards 'more' functionality (e.g. loopback isolation, ipv6) 1167198788 M * JustinWick Alrighty. 1167198800 M * Bertl and the mainline kernel is _slowly_ heading towards virtualized network stacks 1167198826 M * Bertl the bad news is, we decided not to do NGNET for several reasons 1167198840 M * JustinWick Ok. 1167198844 A * JustinWick listens more 1167198865 M * Bertl one being the fact that all approaches we did, caused noticeable overhead 1167198896 M * Bertl another being that for 99% of all applications, a slightly improved isolation is sufficient 1167198928 M * Bertl nevertheless, I think we can make good use of your time ... 1167198948 M * JustinWick Yes, I'm sure we can find things that will be of mutual benefit. 1167198961 M * Bertl as I already said, mainline is heading towards virtualization 1167198963 M * JustinWick Eric is less concerned with the "NGNET" per-se, or even so much the isolation, as the connectivity 1167198985 M * Bertl this includes the layer 2 isolation you know from OVZ 1167198986 M * JustinWick I'll have to get him to write up exactly what he's concerned with (most of our conversation was by phone so I don't have access to that the same way as I would by an email). 1167198994 M * JustinWick Great! 1167199016 M * Bertl but will also include a layer 3 isolation (as done in Linux-VServer) 1167199049 M * JustinWick OK. He was also talking about wanting to have routing tables unique to each virtual server, but I'm not sure how important that is. 1167199070 M * Bertl one area where you could put your time in would be testing and integrating the layer 2 virtualization into Linux-VServer 1167199093 M * Bertl especially testing will be the time consuming part, IMHO 1167199095 M * JustinWick Alright. I'll have to discuss that with him, but that sounds like that'd be great. 1167199096 M * JustinWick Right. 1167199101 M * JustinWick I was already thinking about how to test this... 1167199121 M * JustinWick I was thinking of some crazy stuff like fuzzing, and also at looking how some of the other folks test their network stacks, like OpenVZ. 1167199140 M * Bertl well, you won't get any details from swsoft there 1167199155 M * JustinWick But I'm sure you also have some great ideas about how to do that, and I think we can get that going, I'll just need to chat w/ Eric as he's the one I ultimately have to satisfy :) 1167199158 M * Bertl but I guess we have some idea how testing could look like 1167199204 M * JustinWick Do you have anything written up about that, or this effort as a whole? 1167199215 M * Bertl okay, so given that Eric is fine with this direction, I'd suggest to start playing around with Linux-VServer and ask if you do not understand the details 1167199251 M * Bertl the layer 2 virtualization? that is currently discussed on lkml and some other mailing lists 1167199319 M * JustinWick Alright. 1167199326 M * Bertl here is an early entry: http://lkml.org/lkml/2006/6/27/372 1167199335 A * JustinWick has a *LOT* of reading to do 1167199406 M * JustinWick I should apologize in advance that... while I'm theoretically qualified to perform this task, most of my practical experience has involved scientific computing, parallel algorithms, data visualization/graphics, and embedded stuff, not kernel hacking. My old boss/friend has some firewire code in the kernel though, so I'm sure he can give me some hints on how to get up to speed, etc. 1167199431 M * JustinWick The basic thing that will sell this to him is, if I work on this, will it be completed sooner at a production-ready level? 1167199485 M * Bertl I'd say so, but of course, no guarantees 1167199498 M * JustinWick Okay, well I can hardly expect anything in writing. 1167199532 M * JustinWick I'll drag him in here sometime in the next day or so and we can figure out how we can help you guys etc. 1167199568 M * Bertl okay, sounds good, what timezone do you live in? 1167199625 M * JustinWick eastern 1167199633 M * JustinWick Where are the main VServer guys, and yourself? 1167199668 M * JustinWick I would also like to apologize about my lack of knowlege of Vserver, I expect to rapidly familiarize myself with it over the next few days... 1167199697 M * Bertl so it's 1:07 for you now, right? 1167199707 M * JustinWick Right. 1167199730 M * Bertl okay, when do you plan to work (hours) 1167199733 M * JustinWick I'm up late because I'm a night owl, and the whole consulting/contracting thing lends itself to such a schedule, but I usually try to be done working by now :) 1167199760 M * JustinWick Well... I'm fairly flexible on that, but I do hate mornings (my next job will invaritably require that, however, but that shouldn't be till February when I move to Boston) 1167199767 M * Bertl okay, so you will typically work before midnight, I guess 1167199786 M * JustinWick Well I can work any time from, say, noon, to 2:00 AM. 1167199794 M * JustinWick Any earlier or later is a little much for me. 1167199811 M * Bertl okay, that sounds good, i.e. it will match 1167199818 M * JustinWick Great! 1167199846 M * JustinWick VServer has been working really well for Eric, and he's happy that he has a chance to give something back while also making things better for his clients. 1167199859 M * Bertl btw, I'm in Austria, and timezone here is CET, it's 7:10 in the morning, but that is okay for me :) 1167199862 M * JustinWick Great. 1167199868 M * JustinWick hah, OK. 1167199873 A * JustinWick wants to go visit NZ 1167199937 M * JustinWick FYI, in case anyone cares, I'm relatively green out from school, I went to Cornell University and did a bachelors in Applied Physics and then a Masters in CS. I've worked on a fair bit of software over the last 6 years, though, including some mission critical software for NASA, etc, so I think there will be a lot of details here but that it won't be insurmountable :) 1167199984 M * Bertl sounds good ... 1167199984 A * JustinWick is looking forward to learning more about the network stack 1167199997 M * Bertl trust me, you will ... 1167200005 M * JustinWick I usually don't touch anything under layer 3 except for troubleshooting networks, but... time to get hands dirty :) 1167200032 M * JustinWick Well my next job looks to be at Raytheon which is a big defense contractor here in the States, and I know they do a lot w/ networking. 1167200058 M * Bertl also linux based? 1167200063 M * JustinWick I'm particularly interested in stochastic networks and modelling, but it'd be great to learn the "basics" before working on anything next-gen 1167200071 M * JustinWick Oh I have no idea what they use there, it's all secret! 1167200092 M * Bertl then it cannot possibly be Linux :) 1167200153 M * JustinWick Well it could be that NSA linux stuff 1167200166 M * JustinWick I guess what I mean is that what they use is often secret, not necessarily the software itself. 1167200172 M * Bertl yeah, but it wouldn't be all secret then :) 1167200216 M * Bertl anyway, welcome to the club (assumed that you find some agreement with Eric) and TTYL 1167200255 M * Bertl I'm off to bed now ... feel free to hang around ... 1167200259 M * JustinWick TThanks. 1167200262 M * JustinWick Sure. 1167200280 M * JustinWick Well Eric seemed to be really gung-ho about doing this, and I have the time before my next job and need some extra $$$ and knowledge 1167200297 M * Bertl ah, and you probably can get some basic Linux-VServer info from the wiki (especially the paper might be interesting) 1167200318 M * Bertl of course, installing the software will help too 1167200332 M * Bertl so, off for now ... have a good one everyone, cya! 1167200342 N * Bertl Bertl_zZ 1167201049 J * FireEgl Atl-NA@adsl-147-90-220.bhm.bellsouth.net 1167204433 J * justinwick_ ~justinwic@pool-151-201-42-221.pitt.east.verizon.net 1167204434 Q * JustinWick Read error: Connection reset by peer 1167205657 J * lylix ~eric@dynamic-acs-24-154-34-43.zoominternet.net 1167206388 Q * lylix Read error: Connection reset by peer 1167207115 Q * FireEgl Remote host closed the connection 1167208738 Q * justinwick_ Quit: justinwick_ 1167210168 Q * m`m`h Ping timeout: 480 seconds 1167214329 J * dna ~naucki@p54BCD76B.dip.t-dialin.net 1167214346 Q * mugwump Quit: leaving 1167215113 J * m`m`h ~simba@deb30.mgts.by 1167215539 J * dna_ ~naucki@p54BCD76B.dip.t-dialin.net 1167215583 Q * m`m`h Quit: õÈÏÖÕ Ñ ÏÔ ×ÁÓ 1167215619 J * m`m`h ~simba@deb30.mgts.by 1167215818 Q * dna Ping timeout: 480 seconds 1167216110 Q * shedi Quit: Leaving 1167216368 Q * ruskie Quit: killed 1167217340 J * ruskie ~ruskie@ruskie.user.oftc.net 1167218492 Q * Aiken Quit: Leaving 1167219639 J * dreamind apwdsl@p548A96C1.dip0.t-ipconnect.de 1167220650 J * lilalinux ~plasma@dslb-084-058-202-242.pools.arcor-ip.net 1167220920 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1167220930 J * ensc ~irc-ensc@p54B4F2F4.dip.t-dialin.net 1167222162 Q * mire Ping timeout: 480 seconds 1167222960 J * mire ~mire@151-167-222-85.adsl.verat.net 1167225018 Q * mcp Read error: Connection reset by peer 1167225018 Q * Hollow Read error: Connection reset by peer 1167225019 J * _mcp ~hightower@wolk-project.de 1167225023 N * _mcp mcp 1167225029 J * Hollow ~hollow@styx.xnull.de 1167225273 J * akaihola ~akaihola@dsl-aur-fed7dd00-50.dhcp.inet.fi 1167225433 M * akaihola Hi, did I understand right that 2.0.2.1 does not support allowing a different set of iptables firewall rules (outgoing traffic in my case) for different virtual hosts and the host system? 1167225592 M * akaihola (my virtual hosts don't have public IPs) 1167225927 Q * mire Ping timeout: 480 seconds 1167226083 M * daniel_hozac depends on what you mean, i guess. 1167226105 M * daniel_hozac there is no set of per-guest iptables rules. 1167226162 M * daniel_hozac you can rather easily emulate that behaviour by creating chains for your guests and making sure they're called from the appropriate place. 1167226168 M * daniel_hozac (all on the host system, of course) 1167226239 M * daniel_hozac there is also a iptables-in-guest userspace implementation, http://www.virtuaserver.com.br/forum/viewtopic.php?p=215 1167226592 J * mire ~mire@161-167-222-85.adsl.verat.net 1167226634 M * akaihola daniel_hozac: hm, I think I have some studying to do on that... 1167226664 M * matti php? 1167226685 M * akaihola daniel_hozac: I'm using shorewall on the host for my firewall rules and was hoping I could just write rules for the IPs of the virtual hosts, but no luck... 1167226693 M * matti You got to be kiddin' me? 1167226704 M * daniel_hozac should work fine. 1167226723 M * daniel_hozac i don't know how shorewall works, but my filtering is completely based on the IP addresses. 1167226749 M * matti abruce is probably not very mental healthy :) 1167226796 M * akaihola daniel_hozac: my policy rejects all outbound traffic from the host to the net, and specifying accept rules for TCP traffic from virtual hosts to the net just doesn't seem to work 1167226848 M * daniel_hozac should be fine, but again, i have no idea how shorewall works. 1167226855 M * daniel_hozac i'm an iptables person :) 1167226865 M * nebuchadnezzar daniel_hozac: hi 1167226889 M * daniel_hozac hello nebuchadnezzar. 1167226919 M * nebuchadnezzar daniel_hozac: in the diet tree, I only found lib_.*la.*.lo 1167226949 M * nebuchadnezzar ho, the .o are in the .libs directory 1167226951 M * akaihola daniel_hozac: shorewall just creates sane iptable rules based on a simple set of configuration files, so maybe I'll look into iptables -L and see what I can do 1167227307 Q * mire Ping timeout: 480 seconds 1167227394 M * nebuchadnezzar daniel_hozac: http://www.asgardr.info/~nebu/vserver/lib_libvserver_la-syscall_getiattr.diff 1167227764 M * dreamind Hi folks ;) 1167227765 M * dreamind hi daniel_hozac 1167228030 J * mire ~mire@156-167-222-85.adsl.verat.net 1167228372 M * daniel_hozac hello dreamind 1167228404 M * dreamind btw, if I didn't say it already, happy christmas everybody ;) 1167228420 M * dreamind a bit late, but hopefully ok ;) 1167228440 M * daniel_hozac you too! 1167228448 M * daniel_hozac nebuchadnezzar: and those are compiled with the exact same options? 1167228452 M * matti daniel_hozac: Thanks! You too! 1167228458 M * matti Ops. 1167228464 M * matti s/daniel_hozac/dreamind/ 1167228465 M * matti ;p 1167228469 M * daniel_hozac i.e. the only difference in configure is --disable-dietlibc? 1167228469 M * dreamind matti: thanks ;) 1167228472 M * matti :)) 1167228474 M * dreamind daniel_hozac: thanks 1167228544 M * nebuchadnezzar daniel_hozac: yes 1167228565 M * daniel_hozac nebuchadnezzar: that's very strange indeed... 1167228570 Q * mire Ping timeout: 480 seconds 1167228570 M * nebuchadnezzar I apply Bertl_zZ' procedure, rm the .o and .lo to force make to rebuild the .o 1167228586 M * daniel_hozac nebuchadnezzar: no difference in the gcc command you ran to create them? 1167228587 M * nebuchadnezzar and use the gcc line the the ouput and add -S 1167228593 M * nebuchadnezzar daniel_hozac: no 1167228598 M * nebuchadnezzar I just verify 1167228603 M * daniel_hozac so the command was the same? 1167228608 M * nebuchadnezzar yes 1167228616 M * daniel_hozac (for the two files) 1167228621 M * daniel_hozac that's... bizarre. 1167228626 M * harry some vserver people @ 23c3? 1167228866 J * Pazzo ~ugelt@dialin-225136.rol.raiffeisen.net 1167228947 M * matti harry: :)))) 1167228955 M * matti What is 23c3? 1167228965 M * matti harry: Merry christmas! 1167229153 M * harry tnx 1167229154 M * harry u 2 1167229170 M * harry http://events.ccc.de/congress/2006 1167229268 M * daniel_hozac harry: phreak`` pinged you again. 1167229322 M * harry yeah, i konw 1167229324 J * mire ~mire@8-166-222-85.adsl.verat.net 1167229326 M * harry got him in pm :) 1167229372 M * matti harry: Oh, 31337 congress. 1167229372 M * matti ;] 1167229477 M * Borg- mattiharry ;) 1167229481 M * Borg- aka mattahari ;) 1167229491 M * matti ;-p 1167229494 M * matti Yes, yes. 1167229498 M * matti We're the same person. 1167229500 M * matti Almost. 1167229501 M * matti ;p 1167229515 M * matti Borg-: Resistance is futile! 1167229516 M * matti ;p 1167229528 M * Borg- exacly ;) 1167229536 M * matti ;) 1167229629 A * harry ponders... 1167229747 M * matti ;-p 1167230049 J * mkhl mkhl@200-153-181-173.dsl.telesp.net.br 1167230979 Q * newz2000 Quit: We are leaving! - quote from Doom with Aliens wad 1167231715 Q * dna_ Quit: Verlassend 1167233808 J * FireEgl ~FireEgl@adsl-147-90-220.bhm.bellsouth.net 1167234429 N * Bertl_zZ Bertl 1167234434 M * Bertl morning folks! 1167234449 M * nebuchadnezzar afternoon Bertl :-) 1167234831 M * matti Morning Borg- 1167234835 M * matti Bertl 1167234835 M * matti ;] 1167234978 M * daniel_hozac nebuchadnezzar: could you upload the two .S files? 1167234988 M * Pazzo moin Bertl! 1167234989 M * nebuchadnezzar sire 1167234991 M * nebuchadnezzar sure 1167235110 M * nebuchadnezzar daniel_hozac: http://www.asgardr.info/~nebu/vserver/lib_libvserver_la-syscall_getiattr-{diet,libc}.S 1167235121 M * daniel_hozac thanks. 1167235362 M * daniel_hozac humm, and glibc is the one that works? 1167235417 M * nebuchadnezzar yes 1167235467 M * daniel_hozac could you paste the exact commands you used to get the .S files somewhere? i just don't get why they'd be different if the command is the same. 1167235523 M * nebuchadnezzar http://paste.linux-vserver.org/826 1167235557 M * nebuchadnezzar It's a cut&past of what is run by make 1167235561 M * nebuchadnezzar both are the same 1167235591 M * Bertl different includes? (config.h) 1167235662 M * daniel_hozac will a couple of defines really make the compiler order the if clauses differently? 1167235685 M * nebuchadnezzar http://www.asgardr.info/~nebu/vserver/config-{diet,libc}.h 1167235710 M * Bertl could be, but doing the same with -E -o .E might help? 1167235734 M * daniel_hozac yeah, i guess. 1167235743 M * Bertl instead of the -S -o .S 1167235749 M * nebuchadnezzar ok 1167235781 M * derjohn what does "15.9+"mean in vserver-stat ?? 200 218 15.9+ 2.5G 10m45s51 2m30s23 1h12m59 gers 1167235812 M * Bertl looks like a bug 1167235834 M * nebuchadnezzar daniel_hozac: http://www.asgardr.info/~nebu/vserver/lib_libvserver_la-syscall_getiattr-{diet,libc}.E 1167235887 M * daniel_hozac humm. 1167235896 M * derjohn strange, after vserver stop/start it#s still: "200 413 15.9+ 1.5G 0m32s80 0m08s47 0m31s99 gers" 1167235896 M * daniel_hozac your glibc build is using traditional syscalls. 1167235932 M * daniel_hozac derjohn: how much do the individual VSZs of the processes in that context add up to? 1167235976 M * derjohn 2 GB + that "15.9 whatever" 1167235982 M * derjohn about 2 GB ... 1167236002 M * derjohn vserver 0.30.210 1167236005 M * daniel_hozac i mean, if you run ps aux in the guest, what does the VSZ column add up to? 1167236023 M * daniel_hozac nebuchadnezzar: could you upload the config.log from the glibc build too? 1167236024 M * derjohn that would assuem i can enter the guest ... i'll try 1167236069 M * daniel_hozac you can't enter the guest? 1167236079 M * daniel_hozac vps should work as well though. 1167236083 M * derjohn after restart I can ... 1167236088 M * nebuchadnezzar http://www.asgardr.info/~nebu/vserver/config-{diet,libc}.log 1167236092 M * derjohn but I have to calc it manually :( 1167236384 M * derjohn 3826920 1167236405 M * daniel_hozac nebuchadnezzar: well, you definitely found a bug. 1167236421 M * derjohn for the log: ps aux |awk '{ print $5 "+"; }' | xargs 1167236423 M * nebuchadnezzar really ? ;-) 1167236425 M * Bertl daniel_hozac: yay :) 1167236470 M * derjohn so the VSZ is about 4 GB right ? 1167236500 M * daniel_hozac nebuchadnezzar: i'll have to fix the configure stuff before we'll have any interesting data. 1167236580 M * nebuchadnezzar daniel_hozac: as you want, I more or less around 1167236645 M * daniel_hozac ensc: is it intentional that the alternative syscalls are only used with diet? 1167236819 J * dreamind_ apwdsl@p548A96C1.dip0.t-ipconnect.de 1167236819 Q * dreamind Read error: Connection reset by peer 1167236859 M * daniel_hozac nebuchadnezzar: quick fix is to add ENSC_SYSCALL_ALTERNATIVE in line 153 of configure.ac and running autoreconf -fi 1167236867 Q * dreamind_ 1167236920 M * nebuchadnezzar daniel_hozac: well, personnaly I do not use NOLEGACY, so it works :-) 1167236958 M * daniel_hozac for the configure problem, i mean :) 1167237228 M * daniel_hozac that should make it less of an apples and oranges comparison. 1167237246 J * _ ~hollow@styx.xnull.de 1167237326 Q * Hollow Write error: connection closed 1167237372 N * _ Hollow 1167237663 M * nebuchadnezzar daniel_hozac: ok, you want me to remake the .S or .E ? 1167237673 M * daniel_hozac both. 1167237677 M * nebuchadnezzar ok 1167237767 J * meandtheshell ~markus@85-125-231-222.dynamic.xdsl-line.inode.at 1167238009 M * Bertl wb meandtheshell! 1167238010 M * nebuchadnezzar daniel_hozac: http://www.asgardr.info/~nebu/vserver/lib_libvserver_la-syscall_getiattr-libc-alt-sysc.{S,E} with ENSC_SYSCALL_ALTERNATIVE 1167238035 M * meandtheshell Bertl: thank you :) 1167238065 J * stefani ~stefani@tsipoor.banerian.org 1167238089 M * Bertl hola stefani! 1167238105 M * stefani salut. 1167238144 M * daniel_hozac nebuchadnezzar: if you compile it normally, does it still work? 1167238162 M * daniel_hozac (because now the asm files are identical) 1167238176 M * nebuchadnezzar arf,I don't even try 1167238392 M * nebuchadnezzar well, I get strange errors 1167238412 M * nebuchadnezzar http://paste.linux-vserver.org/827 1167238430 M * nebuchadnezzar vps afx give me ERR in the CONTEXT column 1167238461 M * daniel_hozac and that's with glibc? 1167238467 M * nebuchadnezzar yes 1167238468 M * daniel_hozac diet still works? 1167238488 M * nebuchadnezzar well, diet still don't works 1167238494 M * nebuchadnezzar I do nothing to diet 1167238496 M * daniel_hozac ah right. 1167238582 M * daniel_hozac Bertl: do you have a simple program that tests the alternative syscalls? 1167238640 J * Sebastian ~sebastian@office.star-hosting.de 1167238711 M * Bertl daniel_hozac: vcmd? 1167238735 M * daniel_hozac ah yeah, i didn't think of that. 1167238754 M * Bertl btw, check that the syscall implementation is the same 1167238760 M * matti Uhm. 1167238766 M * Bertl i.e. same version 1167238771 M * matti What's up people? 1167238778 M * Bertl welcome Sebastian! 1167238794 M * Bertl matti: hmm? 1167238811 M * daniel_hozac nebuchadnezzar: could you build http://vserver.13thfloor.at/Experimental/TOOLS/vcmd-0.07-pre2.tar.bz2 and see if ./vcmd -C get_version works? 1167238821 M * matti Bertl: Nothing. I've just finished my piano lesson ;) 1167238831 M * matti Bertl: So, you know... just a question ;) 1167238837 M * daniel_hozac Bertl: yeah, shiny10 in both. 1167238856 M * Bertl ah, what did you play/practice? 1167238910 M * nebuchadnezzar vserver: Unknown error 4160220576 1167238922 M * matti Bertl: Soon, I'll have a 1st grade in piano (here in UK). 1167238961 M * daniel_hozac nebuchadnezzar: ok, that's interesting indeed. 1167238984 M * matti Bertl: So, just practice - I love Frank Sinatra and his piano compositions. 1167238994 M * daniel_hozac Bertl: do you have any ideas? 1167238998 M * Bertl matti: czerny for now? 1167239005 N * Sebastian Sebastian|aw 1167239025 M * Bertl daniel_hozac: no, no idea ... checking 1167239032 M * matti Bertl: You mean Carl? 1167239089 M * matti Uhm. 1167239093 A * matti need some coffee 1167239128 M * Bertl matti: yep 1167239144 M * matti ;) 1167239149 M * matti Bertl: Coffee? 1167239164 M * Bertl yeah, please ... 1167239173 M * Bertl nebuchadnezzar: could you do the following: 1167239190 M * Bertl - get the http://vserver.13thfloor.at/Experimental/SYSCALL/sc_test.c 1167239197 A * matti hands a mug of delicious coffee to Bertl 1167239203 M * Bertl - get the http://vserver.13thfloor.at/Experimental/SYSCALL/syscall_shiny10.h 1167239213 J * bonbons ~bonbons@83.222.37.103 1167239230 M * Bertl - rename the second one to syscall_new.h 1167239242 M * Bertl - compile it with the following commands: 1167239272 M * Bertl gcc -Wall -O2 -g -o sc_test sc_test.c 1167239281 M * Bertl gcc -Wall -O2 -g -S -o sc_test.S sc_test.c 1167239288 M * Bertl gcc -Wall -O2 -g -E -o sc_test.E sc_test.c 1167239300 M * Bertl and upload the three resulting files somewhere 1167239333 M * nebuchadnezzar I have 2 "../sysdeps/sparc/sparc32/elf/start.S:69: undefined reference to `main'" 1167239388 M * daniel_hozac add an int main() { } too. 1167239405 M * Bertl just make the first one 1167239413 M * Bertl gcc -Wall -O2 -g -c -o sc_test.o sc_test.c 1167239448 J * Piet hiddenserv@tor.noreply.org 1167239517 M * nebuchadnezzar Bertl: http://www.asgardr.info/~nebu/vserver/sc_test* 1167239555 M * Bertl tx 1167239589 M * matti nebuchadnezzar: Nice... A clever trojan I pressume ;p 1167239590 M * matti ;D 1167239623 M * Bertl yeah, and whoever manages to execute it, gets extra points :) 1167239630 M * matti Hyhyh. 1167239649 M * daniel_hozac nebuchadnezzar: just out of curiosity, could you upload the output of objdump -d /lib/libc.so.6 | awk '/:/ { do_print=1 } /^$/ { do_print=0 } do_print==1 { print }' too? 1167239690 Q * mkhl 1167239855 M * Bertl hum hum, looks interesting on sparc32 1167239876 M * nebuchadnezzar daniel_hozac: http://www.asgardr.info/~nebu/vserver/libc6_syscall_dump.asm 1167240073 M * Bertl but nothing wrong with the code, AFAICT 1167240149 M * Bertl and vcmd fails? 1167240231 M * Bertl nebuchadnezzar: could you try ./vcmd -V 1167240260 M * nebuchadnezzar vc_get_version: Unknown error 4160236960 1167240319 M * Bertl ah, I have an idea .. sec 1167240381 M * nebuchadnezzar well, just adding the alternative syscall to configure.ac and it does work anymore :-/ 1167240382 M * daniel_hozac i guess that's expected, from the st %0, [ %0 ] we saw in the dumps. 1167240440 M * Bertl I think the carry test doesn't work on sparc32 1167240498 M * daniel_hozac oh? 1167240509 M * daniel_hozac i.e. carry is always set? 1167240631 M * Bertl nah, I guess the check fails 1167240648 M * Bertl sec, have to think about a proper workaround 1167240721 M * daniel_hozac hmm, how would the check fail? 1167240723 M * Bertl but the addx should work, imho 1167240730 M * daniel_hozac yeah, it looks fine to me. 1167240803 M * Bertl unless the sparc32 reorders it :) 1167240860 M * daniel_hozac hmm? 1167240901 M * Bertl the ta could have a slot there 1167240937 M * Bertl although it doesn't look like, according to the manual 1167240967 M * daniel_hozac gcc doesn't think so either. 1167241411 J * waldi ~waldi@bblank.thinkmo.de 1167241415 M * waldi hi 1167241418 M * daniel_hozac hello 1167241435 M * Bertl hey waldi .. got a minute? 1167241460 M * waldi yeah 1167241475 M * Bertl http://www.asgardr.info/%7Enebu/vserver/sc_test.S 1167241488 M * Bertl that is the result from compiling 1167241498 M * Bertl http://vserver.13thfloor.at/Experimental/SYSCALL/sc_test.c 1167241507 M * Bertl with my syscall wrapper on sparc32 1167241522 M * Bertl do you see anything obviously wrong there? 1167241566 M * Bertl http://paste.linux-vserver.org/828 1167241578 M * Bertl (zero arg syscall to 1100) 1167241614 M * Bertl the syscall return value is in %o0, right? 1167241614 J * shedi ~siggi@inferno.lhi.is 1167241637 M * waldi what is the problem? 1167241650 M * Bertl it seems that we get issues like this: 1167241656 M * Bertl vc_get_version: Unknown error 4160236960 1167241667 M * waldi uh 1167241705 M * waldi i don't know about the syscall interfac on sparc 1167241870 M * Bertl I only have the sparc64 here, where it should be fine 1167242038 Q * michal` Ping timeout: 480 seconds 1167242241 M * Bertl nebuchadnezzar: okay, let's try with something simpler than vcmd 1167242279 M * daniel_hozac like hello world :) 1167242290 M * Bertl http://vserver.13thfloor.at/Experimental/TOOLS/vxid-0.03.tar.bz2 1167242351 M * Bertl nebuchadnezzar: see if ./vxid -V fails too 1167242393 M * Bertl if so, please upload the binary too 1167242411 J * dreamind apwdsl@p548A96C1.dip0.t-ipconnect.de 1167242420 J * michal` ~michal@www.rsbac.org 1167242431 Q * dreamind 1167242434 M * nebuchadnezzar Bertl: vc_get_version: Function not implemented 1167242437 J * dreamind apwdsl@p548A96C1.dip0.t-ipconnect.de 1167242475 M * Bertl uhm 1167242492 M * nebuchadnezzar But, there is one thing I do not understand 1167242521 M * nebuchadnezzar You made me add a line in the configure.ac of my libc util-vserver 1167242530 M * nebuchadnezzar but the problem was for the dietlibc one... 1167242571 M * daniel_hozac yes, but the reason glibc worked was just an unfortunate side-effect of a bug in the configure script. 1167242591 M * daniel_hozac as was made evident by it failing when using the alternative syscalls. 1167242621 M * Bertl nebuchadnezzar: could you copy the syscall_shiny10.h into the vxid dir, as syscall.h 1167242627 M * Bertl and do a make clean && make 1167242669 M * nebuchadnezzar ok, now I get vc_get_version: Unknown error 4160171424 1167242682 M * Bertl okay, good, please upload the binary 1167242727 M * nebuchadnezzar Bertl: http://www.asgardr.info/~nebu/vserver/vxid.o 1167242735 M * Bertl btw, do you have a 64bit toolchain at hand too? 1167242746 M * Bertl that is a sparc64, yes? 1167242751 M * nebuchadnezzar yes 1167242763 M * nebuchadnezzar I thought that everything was 64bits... 1167242775 M * Bertl would be interesting to see if the vcmd works when compiled 64bit 1167242787 M * Bertl nope, check with 'file vcmd' 1167242803 M * Bertl or vxid.o 1167242810 M * nebuchadnezzar yes 32bits 1167242921 M * Bertl okay, the code is obviously wrong, together with the errno checking (as usual) let's see if we can 'mend' this somehow 1167242996 M * Bertl what gcc version do you use atm? 1167243016 M * nebuchadnezzar gcc-4.1 (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) 1167243227 M * nebuchadnezzar ok, off for 1 or 2 hours 1167243235 M * Bertl okay, np 1167243241 M * Bertl thanks! 1167244346 M * eyck are there many bugs in debian etch's linux-image-2.6.18-3-vserver ? 1167244376 M * Bertl kernel should be fine, I guess 1167244443 M * daniel_hozac nah, it has that crash on stop 1167244465 M * eyck yes, that's the one i'm talking about :) 1167244467 M * daniel_hozac it needs to be upgraded to 2.0.3-rc1. 1167244469 M * Bertl hmm, I thought that was fixed? 1167244480 M * daniel_hozac not in Debian :) 1167244482 M * eyck daniel_hozac: thanks. 1167244617 J * jmaier ~jmaier@HSI-KBW-091-089-045-089.hsi2.kabelbw.de 1167244666 M * jmaier hi guys 1167244669 M * daniel_hozac hello 1167244698 M * Bertl jmaier: what about the girls? :) 1167244698 M * jmaier i am quite new to vserver and i was quite lucky how easy it was to use 1167244713 M * jmaier he girls as well =) 1167244754 M * Bertl good to hear, ... but? 1167244771 M * Bertl (the easy to use part) 1167244793 M * jmaier i got a question about the mailing list: on http://linux-vserver.org/Communicate there is a link to a domain which is unknown to my dns : http://list.linux-vserver.org/ 1167244822 M * jmaier or it does not allow http connects, i wanted to subscribe 1167244828 M * daniel_hozac you sure it's a DNS problem? i'm getting connection refused. 1167244840 M * jmaier jer sorry 1167244864 M * jmaier but the list seems used, how can i subscribe 1167244870 M * Bertl guess we have to send a mail to martin, _again_ 1167244988 M * Bertl okay, sent him an email .. 1167244998 M * jmaier the question that i did not find on the list is where is acrual docu about the new networking system? if you dont want to read sourcecode. i read http://oldwiki.linux-vserver.org/NGNET-Testing-HOWTO , but when i downloaded the vnet sources i saw there is no option -x 1167245033 M * eyck girls? where? 1167245040 M * Bertl jmaier: i.c. well, I guess we have to move/rewrite that in the near future 1167245052 M * Bertl eyck: here in the channel, you just have to look closely 1167245060 M * jmaier and i dont know if i still have to patch iptables, i have a 1.3.3 version 1167245061 M * daniel_hozac when the new ngnet appears in 2.3.0? 1167245076 M * Bertl yep 1167245104 M * Bertl jmaier: thing is, the ngnet was a nice idea, but has been mainly obsoleted by now (at least in theory :) 1167245130 M * jmaier ok what is the best way to get a per vserver local "lo" interface? 1167245149 M * daniel_hozac 2.3 has that already, no? 1167245156 M * daniel_hozac in theory, at least. 1167245192 M * jmaier ok, i used the kernel patch that was shipped with ubuntu (had to patch the patch, it was not applicable to the actual sources) 1167245201 M * jmaier dapper 1167245245 M * jmaier so do you think the best way is to use a vanilla kernel and your stable patches for a "production" system? 1167245365 M * daniel_hozac yep. 1167245367 M * Bertl well, the 2.3 branch is devel 1167245402 M * Bertl so, while it is always good to stick to mainline, the stable branches will not give you the 'lo 1167245414 M * jmaier one question: when i download the devel version, can i give the needed capability to the vserver and then configure my net in the vserver (loopback and en0/eth0)? How is the routing done then? like in http://oldwiki.linux-vserver.org/NGNET-Testing-HOWTO ? Which vnet is to use? Do i need to patch my iptables? 1167245439 M * Bertl that is part of the decisions we made, i.e. 1167245450 M * Bertl there is no network virtualization at layer 3 1167245460 M * jmaier ok 1167245469 M * Bertl but, as mainline (kernel.org) is working on layer2 virtualization 1167245492 M * Bertl this feature will get into Linux-VServer as soon as something useable is available 1167245517 M * Bertl the layer 3 isolation we focused on, allows you to 'see' all the stuff properly 1167245531 M * Bertl but it doesn't allow you the setup your networking from inside 1167245966 M * jmaier ok so as i dont rely on lo inside the vservers much (would be nice though), i will use the 2.0.1 version where the patch is maintained by the apt-guys. thanks a lot. i hope in future i can migrate easily to the new setup. 1167246005 M * jmaier Bertl what do you mean by: manline kernel virtualization? the new virtualization layer that was introduced? 1167246021 M * Bertl nah, that is kvm 1167246041 M * Bertl but 2.6.19 already got two new namespaces 1167246052 M * Bertl (uts and ipc) and there will be more 1167246053 M * eyck daniel_hozac: I'm a little bit off with new developments, why 2.0.3-rc1, isn't 2.2.0 the current stable? 1167246113 M * Bertl eyck: we have two stable branches atm 1167246127 M * Bertl 2.0 for < 2.6.19 and 2.2 for 2.6.19+ 1167246330 M * daniel_hozac jmaier: are you referring to kernel-patch-vserver? 1167246357 M * georgek Good afternoon everyone. I just wanted to give a little feedback. A week ago I had first come here with a mysql issue where mysql 4.0.27 would stop responding. I was advised to upgrade to a more recent devel kernel. Since I had upgraded to a 2.6.19.1 kernel and the devel vserver patch (2.3.0.5), I haven't seen the mysql issue reproduce. 1167246388 M * Bertl georgek: that's valuable feedback! thanks a lot! 1167246473 M * georgek Good to hear Bertl. Thanks 1167246526 M * jmaier daniel_hozac: yes i had to change the patch to make it applicable to the actual kernel sources. 1167246544 M * daniel_hozac jmaier: that's a) for Debian sarge kernels, and b) unmaintained (AFAIK). 1167246632 M * jmaier daniel_hozac: i wanted to do it the distribution way to avoid looking for security patches all the time. are they looking for a maintainer? or is patching of the 2.6.15 tree not recommendet? 1167246651 M * daniel_hozac is who looking for a maintainer? 1167246707 M * daniel_hozac kernel-patch-vserver is probably already vulnerable. 1167246720 M * jmaier daniel_hozac: i meant is there a reason for the orphaned package? 1167246729 M * daniel_hozac it has been dropped from Debian etch. 1167246736 M * daniel_hozac in favor of the pre-built binary kernels. 1167246749 M * daniel_hozac (so you could just use that source package if you wanted to build your own) 1167246819 M * jmaier daniel_hozac: i am speaking about ubuntu dapper, which source package should i use? 2.6.15 does AFAIK not contain the needed vserver code 1167246841 M * daniel_hozac Ubuntu is as usual just copying the Debian packages. 1167246864 M * Bertl but yes, it seems that ubuntu has no specific maintainer atm 1167246866 M * daniel_hozac but as they don't do that for the kernel, you don't really have a viable alternative from any official channel, AFAIK. 1167246895 M * daniel_hozac there are a couple of people publishing their own Ubuntu builds. 1167246941 J * justinwick ~justinwic@pool-151-201-42-221.pitt.east.verizon.net 1167246952 M * jmaier daniel_hozac: you mean no alternative from loading using 2.6.17.13 and vs2.0.2.1, and looking for security updates? or using uniklu? 1167246953 M * Bertl welcome justinwick! 1167246955 M * daniel_hozac but having an official Ubuntu maintainer would be a good idea. 1167246965 M * justinwick Hey Bertl. 1167246966 M * justinwick What's up? 1167246982 M * daniel_hozac jmaier: no alternative from any official Ubuntu channel. 1167247000 M * daniel_hozac jmaier: and i'd recommend you use 2.6.19.1-vs2.2.0-rc6. 1167247111 M * jmaier daniel_hozac: as far as i know ubuntu is dapper lts is away from using 2.6.19, so i have to make a decision. Probably i just use the vanilla sources you recommend and patch/recompile it when you update the sources 1167247144 M * daniel_hozac 2.6.17.13-vs2.0.2.1 will not be seeing any updates. 1167247175 M * daniel_hozac if you need to stay <2.6.19, 2.6.18.6-vs2.0.3-rc1 is pretty much your only choice. 1167247185 M * jmaier daniel_hozac: i mean there cannot be a maintainer for a 2.6.15 kernel vserver patch if that kernel version is not recommended by you 1167247199 M * Bertl justinwick: nothing unusual yet, and on your side? 1167247244 M * daniel_hozac jmaier: 2.6.15 only has some really old versions available... not something i'd recommend at all. 1167247245 M * Bertl jmaier: well, a maintainer would probably backport the changes to 2.6.15 1167247336 M * justinwick Bertl: Well, I'm getting a vserver dev machine set up today... it's 2.4 Ghz P4 + 256 megs of RAM, someone I knew got a new machine so this is a freebie, should be good enough, although it only has one ethernet port, not sure if I'll need more to test networking stuff. 1167247369 M * Bertl justinwick: you should be fine with a serial console 1167247479 M * jmaier daniel_hozac: ok, as i think i am not able to backport the new kernel i might use 2.6.19, thank you for enlightment 1167247482 M * justinwick Bertl: I'm planning on doing a "crash corse" on the kernel so I can speak about it intelligently. 1167247502 M * justinwick Bertl: Is this work 2.6 only? 1167247552 A * jmaier leaving, thank you all for great work, cu in future probably here 1167247566 Q * jmaier Quit: BitchX: the original point-and-click interface. 1167247583 M * Bertl justinwick: hmm? 1167247632 M * justinwick Bertl: The networking stuff we discussed, will that work with 2.4 kernel at all, or are we just making it compatable w/ 2.6? 1167247646 M * daniel_hozac 2.4 is dead. 1167247655 M * daniel_hozac or well, on life support through eyck :) 1167247692 J * ard ~ard@gw-cistron.kwaak.net 1167247701 M * eyck 2.4 is very much alive 1167247748 M * daniel_hozac not from a vserver point of view. ;) 1167247749 M * eyck 2.4 is like human race - dying and out of fashion, but there's still nothing better around 1167247771 M * daniel_hozac haha 1167247773 M * eyck not from vserver development point of view. 1167247802 M * justinwick 2.4 is great for people that like stability, I suppose... I'm not a debian kind of guy, so I don't value the idea of "software so old your grandpa used it, and he liked it!" 1167247836 M * eyck well, you just said '2.6 is for kids'. 1167247861 M * daniel_hozac how so? 1167247870 M * daniel_hozac everyone has a grandpa, even your grandpa, no? 1167247880 M * eyck who else is 'people who don't like stability' ? 1167247910 M * Bertl justinwick: yes, all development happens on 2.6 now 1167247913 M * justinwick Well, I guess by "like stability" I mean "are paranoid about stability, likely for a rational liability reason" 1167247931 M * eyck daniel_hozac: so if I'm going for 2.6+vserver then atm I should use 2.2+2.6.19 ? 1167247935 M * justinwick Bertl: Alright. 2.6 seems reasonably mature, I haven't had any debilitating problems with it, so... 1167247964 M * daniel_hozac eyck: i'd say so. 1167247968 M * eyck ok. 1167248098 M * eyck so, as long as we are talking about new and shiny things..., 2.4.34 is out, and so is my bsd34b patchset ;) 1167248140 M * Bertl url? 1167248161 M * eyck out as in 'out in testing' 1167248203 M * eyck http://kuszelas.com/~eyck/bsd/ 1167248375 M * Bertl hmm, is that supposedly empty? 1167248492 M * eyck hmm, vserver patch is clean only-rediffed vs1.2.11-rc1, so there's nothing new on this front. 1167249430 M * Bertl ah, now it is there 1167249490 M * eyck amazing, Fedora? I thought you were a mandrivia man? 1167249530 M * Bertl hmm? 1167249563 M * eyck "GET /~eyck/bsd/34b/34b/027-qemu-enable-2.4.x.herbert.diff HTTP/1.0" 200 3887 "http://kuszelas.com/~eyck/bsd/34b/34b/" "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.0.7) Gecko/20060914 Fedora/1.0.5-0.5.fc5 SeaMonkey/1.0.5" 1167249568 M * eyck Fedora. 1167249569 M * daniel_hozac that's me ;) 1167249579 M * eyck ahh, 1167249603 M * eyck you...you...necrophiliac you, it's not nice to play with dead things. 1167249660 M * daniel_hozac :P 1167249665 M * eyck must be a slow newsday, it's not snowing... 1167249787 M * cehteh any of you kernel hackers knows if a mmaped file will be synced to disk when the program gets killed/aborted without munmap/msync ... or when killed while a msync(..MS_ASYNC..) is scheduled? 1167249953 M * eyck aren't mmaped files bypasing cache? 1167250000 M * cehteh nope 1167250028 M * cehteh at least they are lazy synced 1167250051 M * cehteh i just dont know the semantics when the program gets killed 1167250093 M * cehteh doing a ringbufferv for logging/debugging output .. would definately be nice if that recovers a crash 1167250294 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk 1167250331 Q * Pazzo Quit: Verlassend 1167250750 Q * lilalinux Remote host closed the connection 1167251829 J * Aiken ~james@tooax6-075.dialup.optusnet.com.au 1167252897 Q * Piet Quit: Piet 1167254984 J * justinwick_ ~justinwic@pool-151-201-42-221.pitt.east.verizon.net 1167254985 Q * justinwick Read error: Connection reset by peer 1167255498 Q * bonbons Quit: Leaving 1167257798 Q * sladen Ping timeout: 480 seconds 1167257830 M * eyck interesting...it looks like looking at /proc/net/ip_conntrack locks the whole conntrack hashtable 1167257880 M * eyck great way to dos your host: while (true) do cat /proc/net/ip_conntrack|wc -l;done 1167258075 J * sladen paul@starsky.19inch.net 1167258283 M * cehteh mhm 1167258309 M * cehteh eyck: does just open it lock it? 1167258344 M * cehteh cat /proc/net/ip_conntrack | sleep 100 1167258373 M * hardwire daniel_hozac: you a cow man? 1167258393 M * daniel_hozac a cow man? you mean like 25% cow, 75% human? 1167258400 M * daniel_hozac if so, then yes. 1167258445 M * Bertl LOL 1167258480 M * cehteh more the beefstreak or more the milk kind of cow? :) 1167258493 M * cehteh -r 1167258494 M * daniel_hozac the mad kind. 1167258495 M * daniel_hozac :) 1167258517 M * cehteh mo°o0°oh 1167260543 Q * dreamind Quit: dreamind 1167260647 Q * Blissex Read error: Connection reset by peer 1167261578 Q * justinwick_ Ping timeout: 480 seconds 1167261602 J * gregjojo ~greg@82.247.14.10 1167261745 J * tans ~tan@85.108.235.63 1167261763 M * Bertl welcome gregjojo! tans! 1167261774 M * gregjojo hello! 1167261781 M * tans hi 1167262324 M * daniel_hozac Hollow: i just noticed the vprocunhide initscript doesn't do The Right Thing(tm) on Gentoo. 1167262349 M * daniel_hozac i.e. it missed out on the vserver-wrapper changes we did. 1167262363 Q * eyck Remote host closed the connection 1167262368 M * daniel_hozac i'll fix it up now though, i've moved out that stuff to a separate file so it can be shared. 1167262378 J * eyck ~eyck@nat.nowanet.pl 1167262799 Q * tans Read error: Connection reset by peer 1167263229 Q * FireEgl Remote host closed the connection 1167263235 J * FireEgl ~FireEgl@adsl-147-90-220.bhm.bellsouth.net 1167263640 M * daniel_hozac hmm. 2.0 doesn't have the /proc/virtual/status file? or is that a result of my merge? 1167263680 P * gregjojo 1167263716 M * daniel_hozac it just doesn't have it.