> On June 4, 2014, 6:47 p.m., Ian Downes wrote: > > 3rdparty/libprocess/3rdparty/stout/include/stout/os/setns.hpp, line 41 > > <https://reviews.apache.org/r/22160/diff/3/?file=602054#file602054line41> > > > > This is true for all namespaces *except* PID namespaces where setns() > > means that *children* of the re-associated thread will be in the specified > > namespace.
Disabled PID namespace for this function for now since its semantics is different from others. > On June 4, 2014, 6:47 p.m., Ian Downes wrote: > > 3rdparty/libprocess/3rdparty/stout/include/stout/os/setns.hpp, line 64 > > <https://reviews.apache.org/r/22160/diff/3/?file=602054#file602054line64> > > > > Why are we always using syscalls directly and not setns(2) when it's > > available? > > > > When would __NR_setns be defined and not SYS_setns? There is no way that I can check if setns exists or not. Removed __NR_setns check. > On June 4, 2014, 6:47 p.m., Ian Downes wrote: > > 3rdparty/libprocess/3rdparty/stout/include/stout/os/setns.hpp, line 71 > > <https://reviews.apache.org/r/22160/diff/3/?file=602054#file602054line71> > > > > What happens on 32 bit systems? The code won't compile. I think it's ok for now. If we do have a case for x86_32, we can add it later. > On June 4, 2014, 6:47 p.m., Ian Downes wrote: > > 3rdparty/libprocess/3rdparty/stout/include/stout/os/setns.hpp, line 69 > > <https://reviews.apache.org/r/22160/diff/3/?file=602054#file602054line69> > > > > Can you comment on which versions of glibc are old? Added comments. > On June 4, 2014, 6:47 p.m., Ian Downes wrote: > > 3rdparty/libprocess/3rdparty/stout/include/stout/os/setns.hpp, line 88 > > <https://reviews.apache.org/r/22160/diff/3/?file=602054#file602054line88> > > > > What does it mean to associate a *thread* to another pid's namespace? > > This implies different threads in a process could be in different > > namespaces so which one do you associate to? > > > > Perhaps we only support associating all threads of the process to the > > target namespace. Added a check so that if a process has multiple threads, this function will fail. - Jie ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/22160/#review44736 ----------------------------------------------------------- On June 2, 2014, 11:20 p.m., Jie Yu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/22160/ > ----------------------------------------------------------- > > (Updated June 2, 2014, 11:20 p.m.) > > > Review request for mesos, Ben Mahler, Ian Downes, and Vinod Kone. > > > Bugs: MESOS-1450 > https://issues.apache.org/jira/browse/MESOS-1450 > > > Repository: mesos-git > > > Description > ------- > > See summary. > > > Diffs > ----- > > 3rdparty/libprocess/3rdparty/Makefile.am 1474bbc > 3rdparty/libprocess/3rdparty/stout/Makefile.am 8f32a66 > 3rdparty/libprocess/3rdparty/stout/include/stout/os/setns.hpp PRE-CREATION > 3rdparty/libprocess/3rdparty/stout/tests/os/setns_tests.cpp PRE-CREATION > > Diff: https://reviews.apache.org/r/22160/diff/ > > > Testing > ------- > > make check > sudo make check > > > Thanks, > > Jie Yu > >
