1414195201 Q * fisted Remote host closed the connection 1414196329 M * Bertl_oO off to bed now ... have a good one everyone! 1414196338 N * Bertl_oO Bertl_zZ 1414200132 J * fisted ~fisted@xdsl-87-78-188-246.netcologne.de 1414225133 Q * zerick Ping timeout: 480 seconds 1414226373 Q * Carpoon Quit: leaving 1414229241 J * Carpoon ~Carpoon@BC248F13.dsl.pool.telekom.hu 1414229245 Q * Carpoon 1414229282 J * Carpoon ~Carpoon@carpoon.hu 1414229833 N * Bertl_zZ Bertl 1414229836 M * Bertl morning folks! 1414232192 Q * ensc Remote host closed the connection 1414232503 J * ensc ~irc-ensc@p54ADCA0B.dip0.t-ipconnect.de 1414233374 Q * Aiken Remote host closed the connection 1414238401 Q * fisted Remote host closed the connection 1414238419 J * fisted ~fisted@xdsl-87-78-191-84.netcologne.de 1414239387 J * bonbons ~bonbons@2001:a18:22d:6e01:c36:b121:e963:b95f 1414240267 Q * wmp Quit: ZNC - http://znc.in 1414240348 J * wmp ~wmp@auburn.sored.pl 1414240383 Q * wmp 1414240513 J * wmp ~wmp@auburn.sored.pl 1414240730 Q * sannes Remote host closed the connection 1414240743 Q * wmp 1414240781 J * sannes ~ace@cm-84.208.138.68.getinternet.no 1414241277 Q * sannes Remote host closed the connection 1414241295 J * sannes ~ace@cm-84.208.138.68.getinternet.no 1414241657 Q * sannes Remote host closed the connection 1414241669 J * sannes ~ace@cm-84.208.138.68.getinternet.no 1414244201 M * Bertl off for now ... bbl 1414244203 N * Bertl Bertl_oO 1414251598 J * Ghislain ~aqueos@adsl1.aqueos.com 1414252153 J * zerick ~eocrospom@190.118.16.131 1414252205 Q * zerick Read error: Connection reset by peer 1414256102 Q * sannes Ping timeout: 480 seconds 1414263660 J * Aiken ~Aiken@d63f.h.jbmb.net 1414270170 Q * fosco_ Remote host closed the connection 1414270172 J * fosco fosco@marx.wirefull.org 1414270808 Q * fosco Remote host closed the connection 1414270814 J * fosco fosco@91.208.40.1 1414271255 Q * daniel_hozac Quit: getting scuzzied 1414273448 J * daniel_hozac ~daniel@h149n2-spaa-a12.ias.bredband.telia.com 1414276497 Q * fback Ping timeout: 480 seconds 1414278448 M * undefined Bertl_oO: well, i patched the kernel so that (make|from)_vtag is the same with or without USER_NS (as we previously talked about not mapping tag in/out/between user namespaces) 1414278464 M * undefined testfs.sh works 1414278476 M * undefined but testme.sh fails on everything but the networking 1414278514 M * undefined chcontext reports that vc_ctx_create is not implemented 1414278527 M * undefined and an strace confirms ENOSYS 1414278625 M * undefined vc_ctx_create is in /proc/kallsyms 1414278708 M * undefined i'm not familiar with the kernel-side syscall implementation (for x86 or any arch), so i'm just looking around the kernel source 1414278743 M * undefined and haven't found any exclusivity between VSERVER and USER_NS that would disable that (or any vserver) syscall when USER_NS is enabled 1414278907 M * Bertl_oO missing capabilities maybe? 1414278908 J * fback fback@red.fback.net 1414278997 M * Bertl_oO the Linux-VServer syscall is blocked when you're missing the necessary capabilities 1414279081 Q * bonbons Quit: Leaving 1414279880 M * undefined okay, i know how to set capabilities with vattribute, and though its "--help" claims there's a "--get" argument, i can't get it to work (or is there an easier way to capabilities of a process, like something through /proc) 1414280316 M * Bertl_oO it won't work either if the Linux-VServer syscall is not available 1414280351 M * Bertl_oO check with /proc//status 1414280371 M * undefined i am wondering if USER_NS is dumping my console root user into a namespace or other config that conflicts with vserver 1414280397 M * Bertl_oO probably, I suspect it removes the Linux-VServer capability 1414280412 M * Bertl_oO i.e. the one required to actually use the syscall switch 1414280438 M * undefined yeah, i was looking at the switch in kernel/vserver/switch.c 1414280452 Q * Ghislain Quit: Leaving. 1414280492 M * undefined vc_get_version and vc_get_vci work because i'm getting valid values with vserver-info 1414280789 M * undefined /proc/$PID/status on 3.14 with userns enabled (and patched) doesn't look any different than a known-good 3.10 (specifically CapInh, CapPrm, CapEff, & CapBnd) 1414280859 M * Bertl_oO you said you get ENOSYS, right? 1414280885 M * undefined yep 1414280911 M * Bertl_oO try to enable Linux-VServer debugging and set the switch debugging part to 255 1414280933 M * undefined yeah, i was just about to recompile with debugging 1414280939 M * Bertl_oO that should show you where the syscall goes wrong and probably why as well 1414280959 M * Bertl_oO unless there is something else going on and the syscall switch isn't even called 1414280968 M * undefined "switch debugging part" = what all vserver syscalls are debugged? 1414280977 M * Bertl_oO (which would provide a clue as well) 1414281004 M * Bertl_oO yes, if you set the bits, you get the debug output 1414281027 M * Bertl_oO for example, check kernel/vserver/switch.c around line 430 1414281037 M * Bertl_oO there is a vxdprintk(VXD_CBIT(switch, 0), 1414281096 M * Bertl_oO which means, when vserver.debug_switch has bit 0 set, then this will turn into a printk 1414281101 M * Bertl_oO (at runtime) 1414281129 M * Bertl_oO but it only affects the Linux-VServer syscall