1213574582 Q * chigital Ping timeout: 480 seconds
1213574953 M * LuckyLuke ok I got a backtrace for h2ext crashing
1213575094 M * LuckyLuke now I need to get some sleep :) will leave the backtrace url ( http://pastebin.com/m791f5990 ) in query to daniel_hozac and get back here tomorrow.
1213575254 M * Bertl seems like mmap() return -1 here
1213575283 M * LuckyLuke is that cool? :)
1213575287 M * Bertl and the mmap failure check is wrong
1213575315 M * Bertl it shouldn't be !mapping, but something like if (mapping == (void *)-1)
1213575336 M * Bertl (line 150, h2ext.c)
1213575363 M * Bertl after you change that, you will get the actual error message instead of the segfault
1213575430 M * Bertl the interesting part is, what makes mmap fail :)
1213575482 M * LuckyLuke could be some grsec protection? 
1213575496 M * Bertl we'll se, once we get the error
1213575499 M * Bertl *see
1213575525 M * LuckyLuke it logs when it kills something, but of course it leaves to the process to correctly manage function return values :)
1213575536 M * LuckyLuke gimme a minute
1213575660 M * LuckyLuke raupenschlepper ~ # /usr/local/lib/util-vserver/h2ext -d /usr/local/lib/util-vserver/defaults/h2ext.desc -q /gentoo-20080515-vserver-stage4.tgz
1213575664 M * LuckyLuke mmap(): No such device
1213575666 M * LuckyLuke unknown format
1213575678 M * Bertl ENODEV The underlying filesystem of the specified file does not support memory mapping.
1213575689 M * Bertl what filesystem do you use?
1213575739 M * LuckyLuke ext3, but if that's in a chroot there are additional restrictions
1213575757 M * Bertl so back to grsec? :)
1213575828 M * LuckyLuke with a meaningful error message, yes :) gonna check the exact config. but if that's the problem, it would be arch-independent. nobody else uses your vserver+grsec patches other than me? :D
1213575860 M * LuckyLuke (btw, does h2ext work in a chroot even if manually invoked as above?)
1213575861 M * Bertl well, a bunch of folks use them, but probably not with gentoo, as it looks like
1213575877 M * Bertl also those folks might use them with a proper chroot config
1213575900 M * Bertl (or with most of grsec disabled :)
1213576023 M * LuckyLuke I would just like to know what is a "proper chroot config" :)
1213576042 M * Bertl you have to ask those grsec folks then, I'm not using it :)
1213576156 M * LuckyLuke :)
1213576175 M * LuckyLuke anyway I disabled everything I could disable via sysctl and it still gives that error.
1213576204 M * LuckyLuke isn't it possible that it simply doesn't like this .tgz file? it's not one of the "official" ones from the gentoo devs since they do not build hppa versions
1213576268 M * LuckyLuke mmm as far as I can tell from the source, it's just mmaping a readonly opened file.
1213576318 M * Bertl I think it just tries to look into the file (to decide what kind of archive that is)
1213576339 M * Bertl an mmap of that file should work just fine, so that sounds like an issue with the kernel 
1213576353 M * Bertl (again grsec comes to my mind)
1213576367 M * Bertl but it might be an architecture specific mmap issue
1213576429 M * LuckyLuke a non-grsec kernel is already building and I'll be able to reboot tomorrow. But a generic mmap bug seems improbable: this system is in production since early 2007 with lots of software on it (mysql, postgres, apache-2.2, mod_php+suhosin, mod_perl, python... to name a few)
1213576463 M * LuckyLuke afaik lots of them use mmap to work
1213576508 M * Bertl so we'll know tomorrow then .. thanks for locating the h2ext issue
1213576524 M * LuckyLuke should I report the !mapping thing somewhere? 
1213576536 M * LuckyLuke thanks to you for the help.
1213576580 M * Bertl well, I think daniel_hozac will read up on it .. and fix it, but you can also send a patch (With the changes) to him/savannah .. and you're welcome! 
1213576589 M * LuckyLuke ok, I'm leaving a note in query 
1213576604 M * LuckyLuke as I already pasted him the pastebin URLs
1213576809 Q * mire Quit: Leaving
1213577311 M * LuckyLuke good night, see you tomorrow.
1213577463 P * mugwump 
1213577645 M * LuckyLuke I'm back for a last note: I tried strace on h2ext and seems that the file descriptor gets lost? it got fd=3 in line 148 (EopenD), used fd=3 in 149 (Efstat) but then it becomes fd=0 at the mmap in line 150
1213577670 M * Bertl gcc issue?
1213577681 M * LuckyLuke open("/gentoo-20080515-vserver-stage4.tgz", O_RDONLY|O_LARGEFILE) = 3
1213577681 M * LuckyLuke fstat64(3, {st_mode=0, st_size=0, ...}) = 0
1213577681 M * LuckyLuke mmap2(NULL, 4096, PROT_READ, MAP_SHARED, 0, 0) = -1 ENODEV (No such device)
1213577719 M * LuckyLuke the same happens before when it reads the .desc file, but after failing to mmap() it just read()'s something
1213577762 M * LuckyLuke don't know. it's gcc (GCC) 4.1.2 (Gentoo 4.1.2 p1.0.2)
1213577779 M * LuckyLuke definitely going to sleep now. bye.
1213577786 M * Bertl k, cya
1213578463 J * doener_ ~doener@i577BBE05.versanet.de
1213578567 Q * doener Ping timeout: 480 seconds
1213581850 M * Bertl okay, off to bed now ... have a good one everyone!
1213581857 N * Bertl Bertl_zZ
1213584101 J * derjohn_foo ~aj@p5B23E6DF.dip.t-dialin.net
1213584546 Q * derjohn_mob Ping timeout: 480 seconds
1213585062 Q * dowdle Remote host closed the connection
1213587511 Q * fatgoose Quit: fatgoose
1213589297 J * ktwilight_ ~ktwilight@113.126-66-87.adsl-dyn.isp.belgacom.be
1213589415 J * squat_ ~squat@85-10-210-61.clients.your-server.de
1213589436 J * xdr_ ~xdr@34-173-96-87.cust.blixtvik.se
1213589441 J * ace_ ace@har.sagt.no
1213589442 J * bXi_ bluepunk@irssi.co.uk
1213589462 J * cccp_ ~spamfail@lulzmeister.de
1213589462 J * tokkee_ tokkee@ssh.faui2k3.org
1213589464 Q * sannes charon.oftc.net kinetic.oftc.net
1213589464 Q * pmenier_off charon.oftc.net kinetic.oftc.net
1213589464 Q * ktwilight charon.oftc.net kinetic.oftc.net
1213589464 Q * DLange charon.oftc.net kinetic.oftc.net
1213589464 Q * cccp charon.oftc.net kinetic.oftc.net
1213589464 Q * wibble charon.oftc.net kinetic.oftc.net
1213589464 Q * _eSa|_ charon.oftc.net kinetic.oftc.net
1213589464 Q * Medivh charon.oftc.net kinetic.oftc.net
1213589464 Q * m_o_d charon.oftc.net kinetic.oftc.net
1213589464 Q * franck34 charon.oftc.net kinetic.oftc.net
1213589464 Q * Genghis charon.oftc.net kinetic.oftc.net
1213589464 Q * tokkee charon.oftc.net kinetic.oftc.net
1213589464 Q * xdr charon.oftc.net kinetic.oftc.net
1213589464 Q * bXi charon.oftc.net kinetic.oftc.net
1213589464 Q * squat charon.oftc.net kinetic.oftc.net
1213589464 J * wibble_ wibble@vortex.ukshells.co.uk
1213589464 N * squat_ squat
1213589466 J * franck34 ~franck34@sd-10138.dedibox.fr
1213589470 J * m_o_d ~m_o_d@host-80.54.30.252.ltv.pl
1213589473 J * DLange ~DLange@dlange.user.oftc.net
1213589475 J * Medivh ck@dolphin.serverbox.de
1213589476 J * NetNuttt ~NetNuttt@adsl-065-006-153-049.sip.asm.bellsouth.net
1213589508 J * pmenier_off ~pmenier@ACaen-152-1-44-99.w83-115.abo.wanadoo.fr
1213589715 Q * NetNuttt Remote host closed the connection
1213590111 J * |eSa|| ~kvirc@ip-87-238-2-45.static.adsl.cheapnet.it
1213590284 J * Genghis ~Genghis@got.debian-inside.com
1213590312 N * Genghis Guest160
1213590765 Q * balbir Ping timeout: 480 seconds
1213591815 Q * FireEgl Read error: No route to host
1213592509 J * FireEgl ~FireEgl@adsl-220-224-144.bhm.bellsouth.net
1213593748 Q * pmenier_off Quit: Konversation terminated!
1213593969 N * Guest160 Genghis
1213594002 N * Genghis Guest165
1213596139 Q * larsivi Quit: Konversation terminated!
1213597096 J * balbir ~balbir@59.145.136.1
1213597659 N * Guest165 Genghis
1213597692 N * Genghis Guest171
1213598111 J * ntrs_ ~ntrs@77.29.76.61
1213600656 N * Bertl_zZ Bertl
1213600660 M * Bertl morning folks!
1213600875 J * bfremon ~ben@lns-bzn-29-82-248-229-3.adsl.proxad.net
1213600896 N * DoberMann[ZZZzzz] DoberMann
1213601086 J * larsivi ~larsivi@85.221.53.194
1213601182 J * jsambrook ~jsambrook@aelfric.plus.com
1213601240 P * jsambrook 
1213601349 N * Guest171 Genghis
1213601382 N * Genghis Guest177
1213601452 J * jsambrook ~jsambrook@aelfric.plus.com
1213601623 A * LuckyLuke 'morning
1213602170 J * dna_ ~dna@156-200-dsl.kielnet.net
1213602722 J * ktwilight ~ktwilight@87.66.207.9
1213603072 Q * ktwilight_ Ping timeout: 480 seconds
1213604047 Q * derjohn_foo Ping timeout: 480 seconds
1213604280 Q * ntrs_ Ping timeout: 480 seconds
1213604456 J * meandtheshell1 ~sa@d86-33-47-39.cust.tele2.at
1213605039 N * Guest177 Genghis
1213605072 N * Genghis Guest189
1213605287 J * ktwilight_ ~ktwilight@19.64-66-87.adsl-dyn.isp.belgacom.be
1213605288 Q * Aiken Read error: Connection reset by peer
1213605294 J * Aiken ~james@ppp59-167-113-120.lns3.bne4.internode.on.net
1213605682 Q * ktwilight Ping timeout: 480 seconds
1213606885 Q * Aiken Quit: Leaving
1213607026 J * kir ~kir@swsoft-msk-nat.sw.ru
1213607280 M * LuckyLuke Bertl: it's official, h2ext looses its file descriptors even under a non-grsec kernel :)
1213607331 J * chigital ~chigital@91.90.144.102
1213607412 M * Bertl LuckyLuke: so probably a gcc issue then
1213607436 M * Bertl (gcc or toolchain)
1213607546 M * LuckyLuke it's definitely not in the ebuild since I'm using the copy I compiled myself from the source
1213607552 Q * eSa| Ping timeout: 480 seconds
1213607559 J * esa bip@ip-87-238-2-45.static.adsl.cheapnet.it
1213607576 Q * |eSa|| Ping timeout: 480 seconds
1213607582 Q * esa` Ping timeout: 480 seconds
1213607908 M * LuckyLuke EopenD returns a valid fd, then it gets passed to Efstat and then to mmap. But when it gets to mmap() it's ==0.
1213607965 M * LuckyLuke line 149 is Efstat(fd, &st); so it should be passed as a copy and Efstat() shouldn't be able to modify it
1213607975 M * LuckyLuke is that correct?
1213608024 M * LuckyLuke Efstat() itself is defined in ensc_wrappers/wrappers-stat.hc
1213608119 J * eSa| ~kvirc@ip-87-238-2-45.static.adsl.cheapnet.it
1213608133 J * |eSa|| ~kvirc@ip-87-238-2-45.static.adsl.cheapnet.it
1213608170 M * chigital hi
1213608183 M * chigital can anyone help with openvcp
1213608372 J * pmenier ~pme@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr
1213608532 M * Bertl LuckyLuke: yep, I see no problem with the source
1213608632 M * LuckyLuke I put what I know in http://pastebin.com/d233b9b38
1213608643 M * LuckyLuke but I don't know how to proceed from that.
1213608729 N * Guest189 Genghis
1213608762 N * Genghis Guest194
1213608802 J * yarihm ~yarihm@104-43-239-77-pool.cable.fcom.ch
1213608853 M * Bertl LuckyLuke: get gcc 3.3 or a more recent 4.x version and recompile
1213608870 M * Bertl LuckyLuke: maybe try without the optimization flags
1213608901 M * Bertl also, don't trust strace, double check that it actually is the fd which is lost
1213608928 M * LuckyLuke I'm actually recompiling with -O0 -ggdb (and without -march)
1213608928 M * Bertl (i.e. put a printf debug statement there to check it right before, or add it to the error message)
1213609643 M * LuckyLuke Bertl: you were right, and strace was wrong. Seems it's them map() actually failing: http://pastebin.com/deccf6d8
1213609659 M * LuckyLuke maybe a dietlibc issue?
1213609707 M * Bertl could be, is it statically compiled?
1213609734 M * Bertl (check with 'file')
1213609742 M * LuckyLuke it is
1213609751 M * LuckyLuke configure says it's using dietlibc
1213609761 M * LuckyLuke maybe it's worth trying with glibc?
1213609781 M * Bertl well, glibc is _known_ to be broken for the chroot stuff
1213609812 M * Bertl so, yes, you can try if that fixes this tool, but no, that's not a solution :)
1213609848 M * LuckyLuke it should help narrow the h2ext issue, I'm trying.
1213610217 J * pisco ~pisco@tor.noreply.org
1213610437 M * LuckyLuke it works with glibc.
1213610475 M * Bertl what dietlibc version do you use?
1213610532 M * LuckyLuke 0.30
1213610553 M * LuckyLuke shouldn't perror(s) print s AND errno?
1213610604 M * Bertl no, it prints the errno as readable message
1213610626 M * LuckyLuke yes, I mean it should print errno in one way or the other.
1213610639 M * Bertl read up on this and check if the patches mentioned did make it into your dietlibc 
1213610643 M * Bertl http://osdir.com/ml/lib.dietlibc/2006-03/msg00012.html
1213610654 M * LuckyLuke when linked to dietlibc mmap() fails and reports "mmap(): No such device"
1213610671 M * LuckyLuke 'kay
1213610767 M * Bertl if not, try dietlibc 0.31, and if that doesn't work with the mmap either, double check with recent debian patches (maybe you can adapt one)
1213610905 M * LuckyLuke I'll let it emerge dietlibc-0.31_pre20070612 while I check if and when that patch got adopted
1213611043 M * Bertl pre? well ... 0.31 was released last year, IIRC
1213611099 M * Bertl and it seems to be in the vanilla 0.31 sources, AFAICT
1213611108 M * LuckyLuke that's what is in portage, they are somewhat late :)
1213611127 M * LuckyLuke anyway I checked the source and the patch is already included. Now checking if it was also in the 0.30 I was using
1213611154 Q * meandtheshell1 Quit: Leaving.
1213611168 Q * chigital Quit: Leaving
1213611465 M * LuckyLuke and it was in 0.30 too. boh. trying to recompile util-vserver with 0.31_preX, after that I'll try a newer version
1213611490 Q * infowolfe Ping timeout: 480 seconds
1213612419 N * Guest194 Genghis
1213612452 N * Genghis Guest201
1213613069 J * Aiken ~james@ppp59-167-113-120.lns3.bne4.internode.on.net
1213613320 J * infowolfe ~infowolfe@c-67-160-167-96.hsd1.or.comcast.net
1213614217 N * tokkee_ tokkee
1213614671 Q * larsivi Quit: Konversation terminated!
1213615079 Q * Mojo1978 Remote host closed the connection
1213616109 N * Guest201 Genghis
1213616142 N * Genghis Guest207
1213616350 M * daniel_hozac LuckyLuke: you might want to see if other syscalls with 5+ arguments (e.g. select(2)) work using dietlibc.
1213616416 Q * ensc Ping timeout: 480 seconds
1213616463 M * daniel_hozac Bertl: re pid-stuff, i thought we decided we'd (try to) embrace pid namespaces? i did some testing with that, but couldn't get the right semantics... i'll do some more prodding later.
1213616555 M * LuckyLuke daniel_hozac: is there anything I can use to test? meanwhile I'm compiling dietlibc-0.31 (not _preX)
1213616808 M * daniel_hozac Bertl: entering is the biggest problem. i think the easiest route is to have userspace call clone(2) after vc_enter_space to get new pids.
1213616840 M * daniel_hozac LuckyLuke: i don't think we have a select test... but dietlibc should have its own testsuite, i think.
1213616868 M * daniel_hozac strace the select test, and see if the 5th argument is non-zero.
1213616961 M * daniel_hozac i suppose it would also be interesting to look at the disassembly of the mmap call.
1213617094 M * LuckyLuke straced dietlibc select.c test and got:
1213617095 M * LuckyLuke _newselect(1, [], NULL, NULL, {3, 0})   = 0 (Timeout)
1213617095 M * LuckyLuke exit(0)                                 = ?
1213617125 M * LuckyLuke seems nonzero.
1213617127 M * daniel_hozac that looks fine.
1213617177 M * Bertl daniel_hozac: hey, did you get the 'bugfix' for the mmap error check?
1213617186 M * daniel_hozac yeah.
1213617192 J * larsivi ~larsivi@169.80-202-217.nextgentel.com
1213617198 M * daniel_hozac i fixed it in all 5 places in util-vserver.
1213617204 M * daniel_hozac thanks
1213617212 M * Bertl ok, great! thanks!
1213617233 J * loddafnir ~mike@193.170.48.107
1213617385 M * LuckyLuke that's great. will we find it in the next release?
1213617403 M * daniel_hozac of course.
1213617429 M * LuckyLuke ok. now I hope to sort this thing out...
1213617495 M * Bertl LuckyLuke: so, do we know by now, if the proper fd is passed to the mmap call or not?
1213617510 M * Bertl (userspace and kernelspace that is)
1213617924 M * LuckyLuke fd seems ok in "printfspace" :) I did this: http://pastebin.com/d62db2110
1213617976 M * daniel_hozac could you upload an objdump -d h2ext (specifically, __unified_syscall* and process_file)?
1213618287 M * LuckyLuke daniel_hozac: here it is the whole objdump https://srv4.lm-net.it/pub/h2ext_objdump.bz2
1213618352 Q * pisco Ping timeout: 480 seconds
1213618438 M * daniel_hozac that doesn't look like dietlibc.
1213618467 Q * padde Read error: Connection reset by peer
1213618472 J * padde ~padde@patrick-nagel.net
1213618472 Q * bzed Remote host closed the connection
1213618478 J * bzed_ ~bzed@devel.recluse.de
1213618484 N * bzed_ bzed
1213618693 M * daniel_hozac hmm, my bad. apparently mmap is all kinds of funkiness in dietlibc.
1213618820 M * LuckyLuke :)
1213618830 M * LuckyLuke that's diet 0.31_presomething
1213618840 M * LuckyLuke I did not yet succeeded in building 0.31 final
1213618916 M * daniel_hozac why's that?
1213618944 M * LuckyLuke trying with gentoo portage and it hangs with an error. before spending time in that, I'm retrying compiling by hand from the official source.
1213618976 M * LuckyLuke invalid operand in parisc/somewhere
1213619000 M * LuckyLuke ok, the same with 'make' on a clean tree:
1213619001 M * LuckyLuke parisc/clone.S: Assembler messages:parisc/clone.S:40: Error: Invalid operands make: *** [bin-parisc/clone.o] Error 1
1213619084 M * daniel_hozac try changing that line to iitlbp%r0,(%sr0,%r0)
1213619191 M * daniel_hozac (with a tab, of course :))
1213619191 M * LuckyLuke cool, that works.
1213619431 Q * Aiken Remote host closed the connection
1213619637 M * LuckyLuke daniel_hozac: is that something that upstream is aware of or should I report that somewhere?
1213619677 M * Bertl report, you are probably the first person to compile it on parisc :)
1213619799 N * Guest207 Genghis
1213619832 N * Genghis Guest210
1213620047 M * LuckyLuke lol,
1213620062 M * LuckyLuke and you fixed parisc assembly on the fly?
1213620125 M * Bertl well, that was daniel_hozac this time, no? :)
1213620170 J * I`pklbsEQ ~hollow@proteus.croup.de
1213620182 J * infowolfe_ ~infowolfe@c-67-160-167-96.hsd1.or.comcast.net
1213620244 J * nenolod_ ~nenolod@ip70-189-74-62.ok.ok.cox.net
1213620329 Q * Hollow Remote host closed the connection
1213620329 Q * nenolod Quit: <arika> this sh scripting is the language of the future
1213620329 Q * infowolfe Read error: Connection reset by peer
1213620345 N * nenolod_ nenolod
1213620348 N * I`pklbsEQ Hollow
1213620361 Q * FireEgl Read error: Connection reset by peer
1213621189 J * FireEgl FireEgl@adsl-220-224-144.bhm.bellsouth.net
1213622837 Q * |eSa|| Ping timeout: 480 seconds
1213622841 Q * eSa| Ping timeout: 480 seconds
1213622910 J * dna__ ~dna@156-200-dsl.kielnet.net
1213623071 M * LuckyLuke re dietlibc-0.31, I made a bug report in gentoo with the clone.S patch and sent the url to upstream. See http://bugs.gentoo.org/show_bug.cgi?id=227429
1213623317 Q * dna_ Ping timeout: 480 seconds
1213623489 N * Guest210 Genghis
1213623522 N * Genghis Guest223
1213623931 J * yarihm_ ~yarihm@dhcp-204-97-vpn.unizh.ch
1213624374 Q * yarihm Ping timeout: 480 seconds
1213624473 J * mauro__ ~yarihm@104-43-239-77-pool.cable.fcom.ch
1213624630 M * LuckyLuke re util-vserver, I re-did the checks with dietlibc-0.31 and got the same problems: http://pastebin.com/d2d40add7
1213624911 Q * yarihm_ Ping timeout: 480 seconds
1213625265 Q * nkukard Quit: Leaving
1213625740 J * nkukard ~nkukard@196.212.73.74
1213625873 M * daniel_hozac LuckyLuke: does the mmap_test in dietlibc succeed?
1213625956 Q * mauro__ Quit: Leaving
1213625962 J * yarihm ~yarihm@104-43-239-77-pool.cable.fcom.ch
1213626002 Q * bfremon Remote host closed the connection
1213626279 J * eSa| ~kvirc@ip-87-238-2-45.static.adsl.cheapnet.it
1213626281 J * |eSa|| ~kvirc@ip-87-238-2-45.static.adsl.cheapnet.it
1213626281 J * _eSa|_ ~kvirc@ip-87-238-2-45.static.adsl.cheapnet.it
1213626660 M * LuckyLuke daniel_hozac: it exits with 0 without any output so I think it's ok
1213626685 Q * Hunger Read error: Connection reset by peer
1213626723 J * Hunger ~Hunger.hu@213.163.11.138
1213626773 J * bfremon ben@lns-bzn-29-82-248-229-3.adsl.proxad.net
1213626827 M * daniel_hozac try with -D_FILE_OFFSET_BITS=64.
1213627057 M * LuckyLuke in util-vserver?
1213627076 M * daniel_hozac no, the mmap_test in dietlibc.
1213627082 M * LuckyLuke ok
1213627179 N * Guest223 Genghis
1213627200 M * LuckyLuke ta-daaan:
1213627201 M * LuckyLuke # ./mmap_test ; echo $?mmap returned error: No such device1
1213627212 N * Genghis Guest228
1213627245 M * daniel_hozac great.
1213627253 M * LuckyLuke this time it files just like h2ext :)
1213627262 M * LuckyLuke s/files/fails/
1213627307 M * LuckyLuke daniel_hozac: so the problem is in dietlibc when using 64bit file offsets? is that the "large file support"?
1213627316 M * daniel_hozac yeah.
1213627323 M * daniel_hozac looks easy enough to fix.
1213627337 M * LuckyLuke cool
1213627348 M * LuckyLuke I'll test it on amd64 this evening, just in case.
1213627359 M * daniel_hozac it's a PA-RISC thing.
1213627387 M * LuckyLuke ok
1213627477 M * LuckyLuke now what? do you plan / are you able to fix it or should I contact upstream or someone else?
1213627609 M * daniel_hozac try http://people.linux-vserver.org/~dhozac/p/m/dietlibc-parisc-mmap2.patch
1213627741 M * Bertl sometimes I think, in regard to dietlibc we _are_ upstream :)
1213627755 M * LuckyLuke ah, lol :) - I'm recompiling with that patch.
1213628127 M * LuckyLuke it still bombs
1213628155 J * dowdle ~dowdle@scott.coe.montana.edu
1213628171 M * LuckyLuke I'll recheck to see if there isn't any ccache or similar that's fooling me, but I don't think so
1213628228 M * Bertl note: you need to rebuild and install dietlibc, and _then_ rebuild util-vserver (and install)
1213628427 M * LuckyLuke I didn't actually try util-vserver, I tried the mmap_test
1213628447 M * LuckyLuke I know I have to rebuild it since it's statically linked. I clean everything between every try to be sure.
1213628481 M * daniel_hozac do an objdump on mmap_test, see if __mmap2 branches to __unified_syscall6.
1213628486 M * LuckyLuke I emptied ccache and am rebuilding dietlibc but it shouldn't be causing any problem.
1213628582 M * LuckyLuke if this gets long or if someone wants/needs an hppa to test dietlibc builds and similar things, keep in mind I can provide accounts. And full virtual servers if we get linux-verver running :)
1213628593 Q * bfremon Remote host closed the connection
1213628616 M * daniel_hozac i've got two lying around... :)
1213628633 M * LuckyLuke so you're ok, as far as they aren't too old :D
1213628681 M * daniel_hozac 80 MHz, 32 MiB RAM, 2 GB drive... ;)
1213628684 M * LuckyLuke I wonder how much it takes to compile dietlibc on a 712/60, I used to use that machine
1213628687 M * LuckyLuke d'oh
1213628761 M * LuckyLuke 2x875MHz 64bit, 8GiB RAM, 2x18GiB drives (in mirroring, with much more storage available in lan)
1213628789 M * LuckyLuke that'll actually go in production next week, I'm still using the C3600 now (1x554MHz 64bit, 2GiB RAM)
1213628811 J * cryptronic ~oli@p54A3B41B.dip0.t-ipconnect.de
1213628824 Q * balbir Ping timeout: 480 seconds
1213628950 M * LuckyLuke there was a guy in USA giving away an hppa for free, or just pieces (like the ram, EDO ram not SDRAM DIMMs), if you're interested
1213628968 J * bfremon ben@lns-bzn-29-82-248-229-3.adsl.proxad.net
1213629039 M * LuckyLuke it was a C240 iirc (PA-8200 236MHz, 12 EDO SIMMs, four pci, three uw-scsi)
1213629198 M * LuckyLuke ok the mmap_test fails even after cleaning everything and applying your last patch.
1213629214 M * Bertl check with objdump as suggested
1213629216 M * LuckyLuke that is, it fails with -D_FILE_OFFSET_BITS=64 and is ok without it
1213629240 J * ntrs ~ntrs@77.29.195.229
1213629286 M * Bertl daniel_hozac: mmap64?
1213629300 M * LuckyLuke 000107d4 <__mmap2>: 107d4:       e8 1f 16 7d     b,l 10318 <__unified_syscall>,r0
1213629311 M * LuckyLuke it jumps to unified_syscall without the '6'
1213629415 M * Bertl LuckyLuke: and you're positive that the patch is in?
1213629492 M * daniel_hozac i.e. you have parisc/mmap2.S?
1213629533 M * daniel_hozac Bertl: seems that way...
1213629788 M * LuckyLuke Bertl: the patch is in, I added it in my ebuild and checked that it applied it. 
1213629997 Q * bfremon Remote host closed the connection
1213630002 M * Bertl LuckyLuke: that dump above, was that from dietlibc itself?
1213630057 M * Bertl (the mmap2>: 107d4: .. )
1213630093 M * LuckyLuke it was from dietlibc-0.31/test/mmap_test.c
1213630147 M * daniel_hozac do you see it compiling parisc/mmap2.S?
1213630221 M * LuckyLuke http://pastebin.com/d450edc <<<--- mmap_test
1213630241 M * LuckyLuke daniel_hozac: I can try rebuilding dietlibc and look for it.
1213630253 M * Bertl could you objdump the following sections for me?
1213630276 M * Bertl /usr/lib/dietlibc/lib-parisc/libc.a: mmap.o, mmap2.o, and mmap64.o ?
1213630292 M * Bertl (or wherever the dietlibc is located on your system)
1213630321 M * Bertl please use objdump -t and objdump -d
1213630342 M * LuckyLuke ok, meanwhile I'm make'ing dietlibc and dumping stdout/err of make
1213630371 M * Bertl we should see if it got compiled in from the objdump -t :)
1213630430 M * Bertl (something like objdump -t /path/to/libc.a | grep mmap)
1213630523 M * LuckyLuke here are those three sections from objdump -t
1213630524 M * LuckyLuke http://pastebin.com/d28ea6eef
1213630571 Q * ntrs Ping timeout: 480 seconds
1213630619 M * LuckyLuke and objdump -d: http://pastebin.com/d3e449f97
1213630658 M * daniel_hozac your headers do have __NR_mmap2 defined, right?
1213630691 M * LuckyLuke and from make's stdout: gcc -D__dietlibc__ -I. -isystem include -pipe -nostdinc -Os -fomit-frame-pointer -Wall -W -Wchar-subscripts -Wmissing-prototypes -Wmissing-declarations -Wno-switch -Wno-unused -Wredundant-decls -c parisc/mmap2.S -o bin-parisc/mmap2.o
1213630711 M * LuckyLuke daniel_hozac: what headers? kernel headers? 
1213630771 M * daniel_hozac is that the only mmap2 reference in the output?
1213630784 M * daniel_hozac yes, kernel headers.
1213630785 M * LuckyLuke asm/unistd.h:#define __NR_mmap2             (__NR_Linux + 89)
1213630817 M * LuckyLuke that's from linux-headers-2.6.23-r3
1213630822 M * LuckyLuke installed on the system
1213630860 M * LuckyLuke no, there's another very long line 
1213630869 N * Guest228 Genghis
1213630879 M * LuckyLuke ar cru bin-parisc/dietlibc.a lots-of-.o-files
1213630888 M * LuckyLuke I think it's the linker?
1213630902 N * Genghis Guest233
1213630927 M * daniel_hozac it's building the static library.
1213630948 M * daniel_hozac that is very strange though...
1213630961 M * LuckyLuke where's the strangeness?
1213630982 J * ntrs ~ntrs@77.29.195.229
1213631036 M * daniel_hozac that it's using __unified_syscall instead of __unified_syscall6.
1213631045 M * LuckyLuke ah.
1213631081 M * LuckyLuke I got fsck'ed up by a "six"? :)
1213631092 M * LuckyLuke (or lack thereof)
1213631144 M * sid3windr six is pretty sexy
1213631157 M * sid3windr (http://images.google.com/images?q=battlestar%20galactica%20six&ie=UTF-8&oe=utf-8&client=mozilla&rls=org.debian:en-US:unofficial&um=1&sa=N&tab=wi)
1213631227 M * LuckyLuke I must agree with you
1213631273 M * LuckyLuke send that girl to me and I'll stop caring about vserver :)
1213631277 M * sid3windr :>
1213631585 Q * ntrs Ping timeout: 480 seconds
1213631869 M * LuckyLuke at least I can reboot into grsec without fearing that it is the problem :)
1213632233 J * fatgoose ~samuel@76-10-149-199.dsl.teksavvy.com
1213632573 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1213633088 J * |4G|Qbick`` ~qbick@83.2.24.254
1213633092 P * |4G|Qbick`` 
1213633162 J * sylv ~timely@ae.teipir.gr
1213633169 P * sylv 
1213633856 J * TheSeer ~theseer@border.office.nonfood.de
1213633859 M * TheSeer heya :)
1213633880 M * TheSeer as usual when i join.. i got a bizare problem ;)
1213633906 M * TheSeer vserver c2p build -m yum --hostname  --context 1121 --interface c2p0=eth0:x.x.x.x/27 -- -d centos5
1213633915 M * TheSeer -> ERROR: Can not find configuration for the distribution 'rh';
1213633920 M * TheSeer ouhm...?
1213634125 M * daniel_hozac you're lacking a hostname.
1213634151 M * TheSeer *sigh*
1213634207 M * TheSeer gnarf.. that was it..
1213634214 M * TheSeer *sham*
1213634215 M * TheSeer e
1213634247 M * daniel_hozac :)
1213634275 M * TheSeer sometimes, one should simply quit and head home...
1213634307 J * jsambroo1 ~jsambrook@aelfric.plus.com
1213634388 Q * jsambroo1 
1213634559 N * Guest233 Genghis
1213634592 N * Genghis Guest242
1213635327 A * Wonka could not really resist...
1213635328 M * Wonka http://qdb.us/183563
1213635538 M * TheSeer hehe
1213635577 N * pmenier pmenier_off
1213636228 J * balbir ~balbir@122.167.176.8
1213636387 M * LuckyLuke :)
1213636563 M * Bertl LuckyLuke: try to change the __mmap2 in daniel_hozac's patch to mmap2
1213636593 M * Bertl (or the other way round :)
1213636631 J * meandtheshell ~sa@d86-33-47-39.cust.tele2.at
1213636643 M * Bertl I'm off for now .. but I'll be back later .. 
1213636648 N * Bertl Bertl_oO
1213636866 M * LuckyLuke Bertl_oO: I got both:
1213636866 M * LuckyLuke +syscall6(mmap2, __mmap2);
1213637297 Q * balbir Remote host closed the connection
1213637320 J * bfremon ben@lns-bzn-29-82-248-229-3.adsl.proxad.net
1213637915 N * DoberMann DoberMann[PullA]
1213638102 Q * bfremon Ping timeout: 480 seconds
1213638249 N * Guest242 Genghis
1213638282 N * Genghis Guest249
1213638371 Q * bonbons Quit: Leaving
1213638489 Q * kir Quit: Leaving.
1213639154 Q * TheSeer Quit: Client exiting
1213639595 Q * phedny_ Remote host closed the connection
1213639598 N * phedny Guest255
1213639604 J * phedny ~mark@2a02:348:35:5a26::1
1213639629 M * blathijs Hmm, is there a git tree for vserver?
1213639757 J * mfv ~mfv@adsl-ull-254-46.47-151.net24.it
1213639891 M * daniel_hozac no.
1213639941 M * blathijs Is there any versioning system used?
1213639989 M * daniel_hozac for the kernel, it's just patches.
1213640227 M * blathijs different patches that get merged together for a release or something?
1213640258 M * blathijs Sounds a bit fragile. Any particular reason for not using git (considering that the kernel sources already come in git..)?
1213640273 M * blathijs Or just "it works, so why change it?" :-)
1213640429 M * daniel_hozac it's worked for years, learning git would take at least as many years... :)
1213640487 M * blathijs Git has a touch of voodo, that's certainly true :-)
1213641311 J * balbir ~balbir@122.167.176.8
1213641939 N * Guest249 Genghis
1213641973 N * Genghis Guest268
1213642011 M * LuckyLuke daniel_hozac: what about Bertl_oO's idea to change the underscores? 
1213642026 M * daniel_hozac AFAICT, they're correct.
1213642064 M * daniel_hozac you could try commenting the #ifdef temporarily.
1213642113 Q * yarihm Quit: Leaving
1213642299 N * DoberMann[PullA] DoberMann
1213642835 M * LuckyLuke daniel_hozac: tried, the mmap_test still fails
1213642937 J * ntrs ~ntrs@77.29.71.198
1213642951 Q * eyck Quit: leaving
1213643050 M * LuckyLuke time to go now. later.
1213643442 Q * nenolod Ping timeout: 480 seconds
1213643472 J * eyck 3zNglSZ6@89.186.2.57
1213643969 J * nenolod ~nenolod@ip70-189-74-62.ok.ok.cox.net
1213644002 Q * mfv Ping timeout: 480 seconds
1213644470 J * fatgoose_ ~samuel@76-10-149-199.dsl.teksavvy.com
1213644741 Q * fatgoose_ Remote host closed the connection
1213644771 Q * fatgoose Read error: No route to host
1213644772 J * fatgoose ~samuel@76-10-149-199.dsl.teksavvy.com
1213645629 N * Guest268 Genghis
1213645662 N * Genghis Guest272
1213648177 M * blathijs Is there any reason that the latest stable release has only half of the kernel options described on the wiki?
1213648192 M * blathijs Or do I have to enable other stuff to see those extra options
1213648193 M * blathijs ?
1213648290 M * daniel_hozac they're all tagged with the version they're visible in
1213648319 M * daniel_hozac if your version should have an option, it most likely depends on something else you disabled.
1213648417 M * daniel_hozac (menuconfig allows you to search for options using /..., where it shows where it's located, what its dependencies are, etc.)
1213648686 J * Aiken ~james@ppp59-167-113-120.lns3.bne4.internode.on.net
1213648852 Q * meandtheshell Quit: Leaving.
1213649193 N * DoberMann DoberMann[ZZZzzz]
1213649298 M * blathijs Does the 
1213649308 M * blathijs Does the "Prevent invalid userland pointer dereference" grsec option cause a problem with vserver?
1213649318 M * daniel_hozac i doubt it.
1213649319 N * Guest272 Genghis
1213649331 M * daniel_hozac but i don't know, i've never used grsec.
1213649352 N * Genghis Guest277
1213649407 M * blathijs It says it commonly causes problems with non-hardware assisted virtualization
1213649417 M * blathijs though it probably applies more to VMWare etc
1213649452 M * blathijs I'll just enable it and see what happens :-)
1213649504 M * blathijs Ah, a lot of options were depending on EXPERIMENTAL, which I had disabled
1213650171 J * bfremon ben@lns-bzn-29-82-248-229-3.adsl.proxad.net
1213651717 M * Bertl_oO okay, off to bed now ... have a good one everyone!
1213651723 N * Bertl_oO Bertl_zZ
1213651848 Q * cryptronic Quit: Leaving.
1213653009 N * Guest277 Genghis
1213653042 N * Genghis Guest291
1213653413 J * derjohn_mob ~aj@e180212076.adsl.alicedsl.de
1213653580 Q * bfremon Quit: Leaving.
1213653767 Q * openblast Remote host closed the connection
1213653835 Q * ntrs Ping timeout: 480 seconds
1213653911 Q * dna__ Ping timeout: 480 seconds
1213654017 Q * fatgoose Ping timeout: 480 seconds
1213656699 N * Guest291 Genghis
1213656732 N * Genghis Guest297
1213656823 Q * larsivi Quit: Konversation terminated!
1213656941 Q * dowdle Remote host closed the connection
1213660085 Q * puck Quit: Coyote finally caught me
1213660389 N * Guest297 Genghis
1213660422 N * Genghis Guest303