1107734566 J * serving ~serving@213.186.182.181 1107735020 M * Bertl welcome serving! 1107735330 M * Bertl Doener: hmm, maybe we should not increment the num (in your patch) 1107735368 M * Doener hm? you mean fix and fix2? 1107735376 M * Bertl yep 1107735432 M * Doener hm, yeah, "fix" is basically broken... i'm open for suggestions, my experience with numbering schemes is less than limited ;) 1107735502 J * linux-addict ~noname@145.116.6.75 1107736031 M * Bertl hmm, let's see if I udnerstood that code (in tcp_diag) correctly ... 1107736070 M * Bertl there is a function tcpdiag_fill() which adds the information (to be sent to userspace) 1107736090 M * Bertl it is called from tcpdiag_get_exact() and tcpdiag_dump_sock() 1107736095 M * Doener yep, that one fills the information into the skb 1107736141 M * Bertl tcpdiag_get_exact() is the result of a specific (i.e. single) request 1107736161 M * Doener yep 1107736183 M * Doener do you think dcpdiag_fill may be the better function to place a check? 1107736189 M * Bertl where the tcpdiag_dump_sock() is solely called from tcpdiag_dump() as result of the callback 1107736223 M * Doener tcpdiag_get_exact still needs to be fixed i guess... i started with the other one to see if i'm working on the right thing at all *g* 1107736244 M * Bertl now the purpose of the s_num and num is solely to keep track across packets, right? 1107736361 M * Bertl and the num is returned (via cb->args[2]) to continue where we left off ... 1107736421 M * Bertl the num and num_s are never exposed/returned to userspace right? 1107736425 M * Doener yep, that's what i got i like 1107736586 M * Bertl so what stops us from putting a check right in front of the tcpdiag_dump_sock() calls in tcpdiag_dump() which just 'continues' with the stuff done there? 1107736678 M * Doener please rephrase ;) 1107736731 M * Bertl orwe could 'just' _not_ do the tcpdiag_dump_sock() when not appropriate, and leave the loops/checks as is 1107736739 M * Bertl s/orwe/we/ 1107736821 M * Doener then we should care about the call to tcpdiag_dump_reqs, too 1107736846 M * Doener in tcpdiag_dump syn_recv synrecv: 1107736855 M * Doener s/ synrecv// 1107736882 M * Bertl another option would be to 'skip' the unwanted sockets like they aren't there ... 1107736910 M * Bertl i.e. put a simple if (!check) continue; at the begin of each for() 1107736910 M * Doener btw, i just noticed that the first check is wrong. should be "goto next_listen;" not "num++;..." 1107736929 M * Bertl yeah, that was what I meant before ... 1107737001 M * Bertl but we can decide on a per socket basis ... 1107737054 M * Bertl and IMHO we do not need to care about the num counts if we skip ... 1107737082 M * Bertl (because we will skip correctly over unwanted sockets too ... 1107737216 M * Doener hm, not sure if we can leave out the args[3]/[4]=0 part... 1107737228 M * Bertl Doener: I'd opt for a simple if (!check) continue; right after the sk_for_each() ... 1107737301 M * Doener i thought the num check might to cheaper, since you said any use of "current" is pretty expensive... 1107737342 M * Bertl I'll draft up a patch, can you test it reliably? 1107737450 M * Doener i know how ss/netstat output should look on my box, so i could abuse it... getting tw sockets is not a problem either... but not sure what "reliably" means in this context... 1107737476 M * Bertl we have 3 types of sockets to check here, right? 1107737499 M * Doener actually four if you count syn_recv 1107737500 M * Bertl s/types/states/ 1107737551 M * Bertl okay, you should check with _a lot_ of sockets, so that we see if we miss something ... 1107737572 M * Bertl (idea behind that is to have different positions in the hash queues) 1107737633 M * Doener ok, my sister's box is up and running, so i can create of lot of sockets 1107737653 M * Bertl excellent ... 1107737676 M * Doener hm, let's see if anything important is running there, otherwise i could abuse her box, that's way better 8-) 1107737804 M * Bertl okay, here my proposed solution: http://vserver.13thfloor.at/Experimental/RC-1.9.4/delta-tcpdiag-fix01.diff 1107739886 M * Bertl Doener: okay, any status yet, I'm off to bed soon ... 1107739924 M * Doener no, i'm reconfiguring the kernel, box didn't come up with the last one... 1107740320 M * mugwump You guys have seen Xen, right? 1107740342 M * mugwump http://www.cl.cam.ac.uk/Research/SRG/netos/xen/ 1107740348 M * Bertl hey Sam! well, the link on linux-vserver.org might give a hint here ... 1107740396 M * mugwump Oh, true. Duh. Sorry to interrupt you :) 1107740479 Q * are|afk Read error: Operation timed out 1107741030 M * Doener ok, box came up this time... 1107741059 M * Bertl any hint regarding the issue? 1107741130 M * Doener i guess i didn't copy the configuration from that box' but from mine... 1107741211 J * are|afk ~are@dsl-084-056-128-002.arcor-ip.net 1107741712 M * Bertl okay, I'm basically off to bed ... 1107741735 T * Bertl http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.{3,4-rc5}, ng8.13 -- He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime 1107741795 M * Bertl hmm, some means? 1107741799 M * Doener TIME-WAIT 0 0 192.168.0.101:47556 194.25.134.25:pop3 1107741803 M * Doener that one is not in ctx0 1107741823 M * Bertl so you see sockets not in xid=0, right? 1107741836 M * Doener that's without the latest fix btw 1107741847 M * Doener those are all in ctx0 1107741865 M * Doener anyway, latest fix is what matters... 1107741868 M * Bertl ah, okay, so you see some of them, and others not? 1107741873 M * Doener yep 1107741921 M * Bertl okay, let me know how 'rc5' (or the latest fix) works for you ... 1107741929 M * Bertl and have a good night! 1107741934 M * Doener night Bertl! 1107741941 M * Bertl cya folks, off to bed I go ... 1107741956 N * Bertl Bertl_zZ 1107742188 M * Doener that was a short test ;) timewait socket created with nc is not visible in ctx 0, but in ctx 1 1107744173 J * chrish01 ~chrish01@69.90.131.10 1107744888 Q * aba Remote host closed the connection 1107744901 J * aba ~aba@sol.turmzimmer.net 1107746259 J * Sessiz Cruel@81.213.67.57 1107746270 T * Sessiz dsifsdafasdifdiasfþçsadçfçþasdþfçsçþdafþçsdçþfþsdþfdsþfsdfsadfmklsadfklmsadkfmskadmfklsdklfklasklfasd 1107746351 Q * Sessiz autokilled: A user from your address has violated network policy. Please email support@oftc.net if you feel this is in 1107746643 Q * Rusty Quit: 1107746860 T * Doener http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.{3,4-rc5}, ng8.13 -- He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime 1107747424 Q * Doener Ping timeout: 480 seconds 1107747442 J * Doener doener@193.24.208.125 1107749971 Q * locksy Quit: Terminated with extreme prejudice - dircproxy 1.0.5 1107750047 J * locksy ~locksy@202.155.174.16 1107750194 Q * locksy Quit: 1107750231 J * locksy ~locksy@mrtg.sisgroup.com.au 1107750381 Q * locksy Quit: 1107750399 J * locksy ~locksy@mrtg.sisgroup.com.au 1107751139 Q * chrish01 Quit: chrish01 1107751616 J * nox- ~nox@c192193.adsl.hansenet.de 1107751657 J * ntrs ntrs@68.188.50.87 1107751961 Q * nox Ping timeout: 480 seconds 1107751976 N * nox- nox 1107751987 Q * ntrs_ Ping timeout: 480 seconds 1107752832 Q * ntrs Ping timeout: 480 seconds 1107753014 J * ntrs ntrs@Dardeene-68.188.50.87.charter-stl.com 1107755299 P * Pazzo 1107758012 Q * are|afk Quit: Disconnecting 1107763543 M * Doener Bertl_zZ: call me stupid or whatever, i only applied the tcpdiag fix ... 1107764071 J * prae ~prae@ezoffice.mandrakesoft.com 1107764380 Q * prae Quit: Client exiting 1107764758 J * prae ~prae@ezoffice.mandrakesoft.com 1107764815 J * mhepp ~mhepp@83.240.22.205 1107765190 J * BWare ~bware@212.26.196.41 1107766072 J * are|afk ~are@212.185.72.226 1107766075 M * are|afk Hi 1107766081 N * are|afk _are_ 1107768482 Q * logger Ping timeout: 480 seconds 1107768514 J * logger ~rs@84.244.0.15 1107769033 J * rs rs@213.193.3.15 1107769058 M * rs hi there 1107769109 M * Doener hi rs! 1107769167 M * Doener Bertl_zZ: ok, with both patches applied (tcpdiag and sockref) i see tw socks, will try to write some testscripts... 1107769179 M * Doener doing that by hand is pretty annoying *g* 1107775390 Q * serving Read error: Connection reset by peer 1107778623 J * _Plug ~plug@datadot.net 1107778623 Q * Plug Read error: Connection reset by peer 1107778623 N * _Plug Plug 1107778999 J * _Plug ~plug@datadot.net 1107778999 Q * Plug Read error: Connection reset by peer 1107779004 N * _Plug Plug 1107780163 J * DuckMaster ~Duck@dyn-83-157-130-223.ppp.tiscali.fr 1107780582 Q * DuckKing Ping timeout: 480 seconds 1107780597 Q * duckx Ping timeout: 480 seconds 1107780616 J * duckx ~Duck@dyn-83-157-130-223.ppp.tiscali.fr 1107781523 N * Bertl_zZ Bertl 1107781538 M * Bertl morning folks! 1107781542 M * Hollow heya Bertl 1107781548 M * Hollow how's it going? 1107781569 M * Bertl fine thanks, and you? 1107781615 M * Hollow fin too, just spent two hours on fixing the gentoo kernel eclass, to accept -rc patches and the like, to get 1.9.4-rc4 running on 2.6.11-rc3 1107781700 J * mboman ~michael@218.212.230.48 1107781763 M * _are_ i think there had been a chnamespace utility, however, seems I don't have it, can anyone give me a link? 1107781883 M * Bertl vnamespace 1107781908 M * _are_ yes, that one :-) 1107782039 M * Bertl it's part of util-vserver (0.30.19x) 1107782094 M * _are_ yes, i have it 1107782131 M * _are_ i just messed up the name, got missleaded by chcontext, chattr, chown, chgroup <-> vtop, vps 1107782361 Q * mhepp Quit: mhepp caught signal: Autobus error 1107782417 M * Bertl _are_: np 1107782431 M * Bertl rs: morning! 1107782470 M * rs morning Bertl! 1107782474 M * rs hwo are you ? 1107782499 M * Bertl fine thanks, and you? 1107782564 M * rs fine too :) 1107782578 M * rs I should have more time to do some vserver stuff this week :) 1107782579 M * Bertl what about your project? 1107782632 M * Bertl well, we had to do a lot of 1.9.4-rc work last week, so no big deal ... 1107782635 M * rs it works fine, we just still have the percistant context bug that prevent some vserver to reboot from time to time, except that it's really stable 1107782662 M * rs about this bug, did you manage to fix it ? 1107782672 M * Bertl (hmm, actually meant the one which required your attention last week ;) 1107782686 A * Bertl is still lagging a little (network issues) 1107782714 M * Bertl rs: we fixed some ref counting related issues in 1.9.4-rc5, so I would give that a try ... 1107782770 M * rs ok, so maybe 1.9.5 will fix this 1107782776 M * rs it's a good neas 1107782778 M * rs news 1107784859 J * nish ~nish@220.224.11.12 1107784870 M * nish evening Bertl :) 1107785038 Q * _are_ Ping timeout: 480 seconds 1107785073 M * Hollow rc5 is in gentoo portage btw 1107785149 M * Bertl evening nish! 1107785270 J * Pazzo ~thomas@adsl136-175.aknet.it 1107785342 M * Bertl welcome Pazzo! 1107785441 M * Pazzo hi bertl! 1107785496 M * Pazzo when do you expect will kernel people release 2.6.11? within the next two weeks? 1107785529 M * Pazzo I've had a bad kernel panic on a running 2.6.9-1.9.3 yesterday :-( 1107785570 M * Pazzo and I can't tell you why - it's not one of my host, I did just set it up and have no access to the box 1107785622 M * Bertl hmm, 2.6.11-rc3 is there ... so I guess it will be soon ... 1107785682 M * Pazzo 2.6.11-rc2-vs1.9.4-rc4 is running very well - but I don't want to migrate every host to migrate once again in two weeks... 1107785722 M * matti Bertl: :-)))))) 1107786071 J * _are_ ~are@212.185.72.226 1107786766 M * Bertl welcome matti! 1107786786 M * Bertl wb _are_! 1107786932 M * _are_ hi, replacing a masquerading router seems to be bad for network connections. ;) 1107787676 N * nish nish_dinner 1107789837 Q * Pazzo Read error: Operation timed out 1107790185 M * meebey has someone samba in a vserver? seems I have a problem now 1107790210 M * meebey I remember oplocks can cause problem, how do I need to disable them in samba? there are several ways 1107790419 Q * Snow-Man Remote host closed the connection 1107790452 J * Snow-Man ~sfrost@66.92.160.6 1107790459 N * nish_dinner nish 1107790763 Q * Snow-Man Remote host closed the connection 1107790831 J * Snow-Man ~sfrost@snowman.net 1107790865 M * Bertl kernel oplocks = yes 1107790885 M * Bertl change that to no ;) 1107790970 M * meebey thanks 1107790977 M * meebey I found something with oplevels and so on 1107791144 Q * Snow-Man Remote host closed the connection 1107791198 Q * monrad Quit: Leaving 1107791242 Q * _are_ Ping timeout: 480 seconds 1107791665 J * Snow-Man ~sfrost@66.92.160.6 1107791968 Q * Snow-Man Remote host closed the connection 1107792541 Q * mboman Quit: One day I'll get that peer and reset HIS connection! 1107793946 M * Bertl k, dinnertime, back later ... 1107793954 N * Bertl Bertl_oO 1107795426 N * Bertl_oO Bertl 1107795446 M * Doener wb Bertl 1107795461 M * Bertl hey Doener! 1107795464 M * Doener any thoughts about today's wiki changes? 1107795509 M * Bertl they have been taken care of ... 1107795619 M * Loki|muh website slow as hell? 1107795636 M * Doener hm, no? 1107795649 M * Loki|muh ah now it works 1107795701 M * meebey aaahh 1107795702 M * meebey Fatal: Label "2.4.29-vs1.2.10p" is too long 1107795707 M * meebey mhmmhmhmmm 1107795733 M * Doener ah yeah, i love that one, too :) 1107795740 A * Bertl .o( who uses lables anyway? ) 1107795753 M * meebey Bertl: users with more than one kernel 1107795775 M * meebey I could remove all dots 1107795781 M * meebey 2429-vs1210p 1107795789 M * Doener just linux and linux-old would confuse me at lot when testing kernels on the real box... 1107795795 M * Bertl aha, I have about 1300 kernels here, so what exactly should I do with the labels? 1107795808 M * meebey Bertl: how you know which kernel is which? 1107795838 M * Bertl by the path and of course ny the name 1107795843 M * meebey grub would solve this, but the servers have lilo 1107795852 M * Bertl of course I do use grub ... 1107795860 M * meebey nice for you :) 1107795894 M * meebey Added 2.4.29vs1.2.10p 1107795895 M * Bertl it's currently the only bootloader able to give me a boot shell over serial ... 1107795898 M * meebey that works for now 1107795912 M * meebey Bertl: AFAIK lilo with patches can do that too 1107795923 M * Bertl the lilo-shell? 1107795963 M * meebey now idea, I only heard about it 1107795984 M * meebey serial support, about shell no idea 1107795994 M * meebey is the boot parameter prompt not enough? :) 1107796039 M * Bertl well, I can boot a kernel from the grub shell, even if I put it in the wrong path ... 1107796108 J * Snow-Man ~sfrost@snowman.net 1107796192 M * Bertl wb Snow-Man! 1107796264 M * Bertl Doener: so 1.9.4-rc5 does work for you as expected? 1107796316 M * Doener didn't test with rc5, but with 1.9.4-rc4 + tcpdiag + sockref2, which should be the same in that regard, right? 1107796356 M * Bertl yes, guess so, delta is at the usual place ... 1107796396 M * Snow-Man Bertl: heya. 1107796430 M * Bertl how are you? 1107796491 M * Loki|muh Bertl: hmmm, important Changes between rc4 and rc5? A changelog would be nice ;) 1107796520 M * Bertl - socket refcounting changed/improved/fixed 1107796545 M * Bertl - SCHED_PRIO without hard scheduler config fixed 1107796583 M * Loki|muh hmmm, am I too blind to see it? 1107796619 M * Bertl - missing clr_vx_info() regarding mm added (Doener) 1107796641 M * Bertl - missing checks in tcpdiag added (Doener) 1107796663 M * Bertl - warning on 64bit archs fixed 1107796685 M * Bertl that's it ... 1107796743 M * prae See'ya 1107796774 M * Bertl hmm, cya2 ;) 1107796789 M * prae :) 1107796836 Q * prae Quit: Client exiting 1107797053 M * Bertl Loki|muh: the changelog or the patch? 1107797075 M * Loki|muh the changelog 1107797091 M * Bertl it doesn't exist yet, well except for my lines I pasted here ... 1107797103 M * Loki|muh ah, okay 1107797634 Q * nish Read error: Operation timed out 1107798639 J * nish ~nish@220.224.48.158 1107798647 M * Bertl wb nish! 1107798691 M * nish thanx Bertl :) 1107798698 J * Borg_Nr1 ~borg_@80.136.114.185 1107798744 Q * linux-addict Remote host closed the connection 1107798919 P * Borg_Nr1 1107800612 Q * click Ping timeout: 480 seconds 1107802780 M * Doener Bertl: hm, the last wiki mails didn't contain an uri anymore, is that intentional? 1107802807 M * Bertl nope, I'm in the progress of fixing it ;) 1107802818 M * Doener ok :) 1107802826 M * Bertl just didn't find anything to change yet ;) 1107802839 M * Bertl (i.e. useful changes in the wiki ;) 1107802884 M * Bertl if you have something to change, please do so ;) 1107802982 M * Bertl ah, found something ... 1107803181 M * Bertl okidokili, seems to work ;) 1107803242 M * Bertl okay, I'm off to bed now .. too tired ... 1107803265 N * Bertl Bertl_zZ 1107803316 J * prae ~prae@sherpadown.net 1107804439 J * mhepp ~mhepp@r72s22p13.home.nbox.cz 1107804455 M * mhepp hello everybody! 1107804575 N * Doener Doener|gone 1107805552 Q * nish Read error: Connection reset by peer 1107806359 Q * prae Quit: Pwet 1107808744 Q * hansi33 Remote host closed the connection 1107809176 J * are|lunch ~are@dsl-084-056-128-002.arcor-ip.net 1107810341 J * click click@dsl-84-161.aal.tiscali.no 1107810549 Q * mhepp Quit: mhepp caught signal: Autobus error 1107810839 J * lars||| ~ledj@213-156-52-105.fastres.net 1107811043 Q * lars||| Quit: 1107814757 J * nish ~nish@220.224.35.17 1107816240 Q * Plug Remote host closed the connection 1107818518 Q * nish Quit: Leaving