1391473510 Q * AndrewLee Ping timeout: 480 seconds
1391474054 Q * _BovineSpongiformEncephalitis_ Ping timeout: 480 seconds
1391474298 J * AndrewLee ~andrew@210.240.39.201
1391475602 Q * fisted Remote host closed the connection
1391475626 J * fisted ~fisted@xdsl-87-78-232-242.netcologne.de
1391477017 M * Bertl off to bed now ... have a good one everyone!
1391477035 N * Bertl Bertl_zZ
1391478483 J * undefined ~undefined@00011a48.user.oftc.net
1391479566 J * _BovineSpongiformEncephalitis_ ~MadCow@0001c2a3.user.oftc.net
1391479622 Q * _BovineSpongiformEncephalitis_ 
1391481284 N * l0kit Guest29
1391481289 J * l0kit ~1oxT@0001b54e.user.oftc.net
1391481688 Q * Guest29 Ping timeout: 480 seconds
1391497436 J * Ghislain ~aqueos@adsl1.aqueos.com
1391503005 Q * yang Read error: Connection reset by peer
1391503008 J * yang yang@irs.si
1391505517 Q * jrklein Remote host closed the connection
1391505519 J * jrklein ~osx@proxy.dnihost.net
1391505739 Q * ircuser-1_ Read error: Operation timed out
1391505919 J * ircuser-1_ ~ircuser-1@35.222-62-69.ftth.swbr.surewest.net
1391510974 Q * Aiken Ping timeout: 480 seconds
1391515951 Q * ircuser-1_ Ping timeout: 480 seconds
1391515959 Q * ircuser-1__ Ping timeout: 480 seconds
1391517966 Q * BlackPanx Remote host closed the connection
1391518441 J * ircuser-1 ~ircuser-1@35.222-62-69.ftth.swbr.surewest.net
1391518801 Q * fisted Remote host closed the connection
1391518822 J * fisted ~fisted@xdsl-78-35-86-112.netcologne.de
1391522450 P * undefined 
1391526875 J * _BovineSpongiformEncephalitis_ ~MadCow@0001c2a3.user.oftc.net
1391530915 J * zerick ~eocrospom@190.114.248.34
1391531011 Q * g_en__ Read error: Connection reset by peer
1391531460 N * Bertl_zZ Bertl
1391531468 M * Bertl morning folks!
1391532038 M * fback morning, let it be ;)
1391532477 M * Ghislain hello
1391532484 M * Ghislain is it me or freenode was down ?
1391532704 M * Bertl they seem to have some issues
1391532724 M * fback Bertl: has anybody reported any problems with 3.13 patch and netns?
1391532848 M * Bertl not that I know of
1391532878 M * fback our system admins are afraid of filling bug reports :>
1391532892 M * fback they always write to me 
1391532950 M * fback this two-liner seems to crash the kernel with vserver patch:
1391532960 M * fback ip netns add test
1391532967 M * fback lsof /var/run/netns/test
1391533254 J * alpha_one_x86 ~kvirc@200.119.211.77
1391533343 Q * l0kit Ping timeout: 480 seconds
1391533422 J * l0kit ~1oxT@0001b54e.user.oftc.net
1391534200 Q * zerick Ping timeout: 480 seconds
1391535387 J * zerick ~eocrospom@190.187.21.53
1391535414 Q * yang Remote host closed the connection
1391535416 J * yang yang@irs.si
1391535660 Q * zerick Max SendQ exceeded
1391535689 J * zerick ~eocrospom@190.187.21.53
1391536871 J * ensc|w ~ensc@www-old.sigma-chemnitz.de
1391539745 Q * _BovineSpongiformEncephalitis_ Ping timeout: 480 seconds
1391540521 M * clopez what's the point of the alternative syscalls on util-vserver ? (flag --enable/--disable-alternative-syscalls on ./configure)
1391540847 M * clopez there are many warnings in lib/syscall-alternative.h saying "warning syscall arch $arch not tested yet"  ($arch=[arm hppa ia64 m32r mips s390 sh sh64 v850])
1391540915 M * clopez Is there any benefit/difference by enabling or disabling this alternative syscalls? Bertl?
1391541413 M * Bertl this option controls how util-vserver invokes the Linux-VServer syscall
1391541449 M * Bertl which can be via custom implementation (we did that for several archs back then) or via the libc syscall wrapper (which is the default nowadays)
1391541663 M * clopez The default is to use the alternative syscalls
1391541718 M * clopez https://github.com/linux-vserver/util-vserver/blob/master/m4/ensc_syscall.m4#L32
1391541781 M * Bertl then it still uses those macros as default :)
1391541795 M * Bertl but check with daniel_hozac for details
1391541933 M * clopez this macros fail to build on armhf  https://buildd.debian.org/status/fetch.php?pkg=util-vserver&arch=armhf&ver=0.30.216-pre3038-1&stamp=1390172426
1391541959 M * clopez and i'm a bit worried about the other archs where it declares that the syscalls were not tested
1391541996 M * clopez should better I use the libc syscall wrapper?
1391542025 M * Bertl hmm, it doesn't seem to fail
1391542046 M * clopez src/rpm-fake.c:712:1: error: r7 cannot be used in asm here
1391542092 M * Bertl but that's in rpm-fake.c
1391542104 M * Bertl i.e. might be related to the syscall wrapper
1391542124 M * Bertl do you have the asm code or the processed C file generated?
1391542132 M * clopez I tried passing the flag -fomit-frame-pointer to gcc as suggested here https://polarssl.org/kb/development/arm-thumb-error-r7-cannot-be-used-in-asm-here but it didn't helped
1391542184 M * clopez yes..
1391542186 M * clopez https://github.com/linux-vserver/util-vserver/commit/981064d039d0e7a5fd66b5369f2a0390728ca4be#diff-184922e36acdc19de1d40bce8f788228
1391542209 M * clopez that's the assembly part that is included for armhf
1391542259 M * Bertl yeah, but that is fine AFAICT
1391542278 M * Bertl what I mean is, the syscall macro is _included_ in the rpm-fake.c code
1391542292 M * Bertl it gets expanded by the preprocessor 
1391542298 M * Bertl and then compiled by gcc
1391542342 Q * thierryp Remote host closed the connection
1391542343 M * Bertl now to actually check what happens I need to see the code after the preprocessor and the assembler version 
1391542354 M * Bertl i.e. -E and -S
1391542399 M * Bertl so, check what command is used to compile src/rpm-fake.c
1391542415 M * Bertl then run it once with -E added (which will only do the preprocessor stage)
1391542426 M * Bertl and then with -S added (which will stop before assembly)
1391542465 M * clopez with -E added : http://people.igalia.com/clopez/vcontext_armhf_preprocesed.txt
1391542482 M * clopez inline static int pivot_root(const char * new_root, const char * put_old) { long ret; register long __ret __asm__ ("r0");register long __sc_id __asm__ ("r7") = (long)((0 +218)); register long __sc_a1 __asm__ ("r0") = (long)(new_root);register long __sc_a2 __asm__ ("r1") = (long)(put_old); __asm__ __volatile__ ("swi $0x0" : "=r" (__ret) : "r" (__sc_id) ,"r"(__sc_a1),"r"(__sc_a2) : "memory" ); ret = __ret; if (((unsigned long)(ret) >= (unsigned long)(-(125)))) { do
1391542482 M * clopez  { (*__errno_location ()) = (-ret); } while(0); ret = -1; } return (int)(ret); }
1391542512 J * thierryp ~thierry@home.parmentelat.net
1391542615 M * Bertl okay, looks good so far, and the assembler version?
1391542788 M * clopez with -E and -S: http://people.igalia.com/clopez/vcontext_armhf_preprocesed_S.txt
1391542856 J * _zerick_ ~eocrospom@190.187.21.53
1391542905 M * Bertl remove the -E, i.e. only S
1391542997 M * clopez the it fails to do anything useful
1391542999 M * clopez src/vcontext.c: In function ‘main’:
1391542999 M * clopez src/vcontext.c:516:1: error: r7 cannot be used in asm here
1391543044 M * clopez and passing -fomit-frame-pointer to gcc changes nothing
1391543107 M * Bertl and if you add the -fomit-frame-pointer ?
1391543117 M * Bertl (directly to this command)
1391543149 M * clopez i'm trying with this command:
1391543192 M * clopez http://paste.debian.net/plain/80160
1391543245 M * Bertl what gcc version is that?
1391543249 M * clopez simplifying it to: gcc -DHAVE_CONFIG_H -I.  -I ./lib -I ./ensc_wrappers -D_GNU_SOURCE -D_REENTRANT -DNDEBUG -fomit-frame-pointer -S src/vcontext.c
1391543254 M * clopez gives the same error
1391543264 M * clopez gcc version 4.6.3 (Debian 4.6.3-14)
1391543284 M * Bertl so they probably changed something in gcc then
1391543287 M * clopez i tried also with gcc-4.7 .. same error
1391543321 M * Bertl yeah, I guess older gcc works fine, anyway, this won't be fixable by simple adjustments, as gcc doesn't seem to allow this register anymore
1391543355 M * Bertl but it is required for the syscall, so your only option with this toolchain is to disable the alternative syscall macros
1391543397 M * clopez yes, I'll do that
1391543432 M * clopez and for the other archs is any difference between using the alt syscalls or the libc ones? which one do you recommend?
1391543494 M * Bertl back then, we had to do those macros as libc wasn't able to get it right
1391543514 M * Bertl when the compiled util-vserver works, then there is no problem/difference to the macros
1391543573 M * Bertl it is like using this or that crypt library ... whatever works is fine
1391543595 M * clopez i see
1391543601 M * clopez thanks for the help :)
1391543632 M * Bertl no problem!
1391544039 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f
1391550699 Q * jrklein Ping timeout: 480 seconds
1391552865 J * jrklein ~osx@proxy.dnihost.net
1391553246 Q * _zerick_ Quit: Saliendo
1391555251 Q * thierryp Remote host closed the connection
1391555319 Q * alpha_one_x86 Quit: KVIrc KVIrc Aria 4.3.1, revision: 6250, sources date: 20120701, built on: 2013-12-17 11:56:03 UTC http://www.kvirc.net/
1391556243 Q * Ghislain Quit: Leaving.
1391557087 J * thierryp ~thierry@home.parmentelat.net