----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/20816/#review41770 -----------------------------------------------------------
src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75362> static int _call? src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75354> Kill the extra space before '>': static_cast<const lambda::function<int()>*> src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75364> static? src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75368> Se we are moving away from process group now? Comments would be nice. src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75358> Check cgroups existence failed ...? src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75361> Do you need this LOG here? Will the caller of this function print the same message? src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75360> Failed to create freezer cgroup ...? src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75366> Can you move this down: pid_t pid = ::clone(...); if (pid == -1) { ... } src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75369> _call is a little confusing. Maybe call it childMain, or something similar? Also, it does not read well since we have: childFunction blockedChildFunction child func, _call What about the following: static int childMain(void* args) { const lambda::function<int()>* func = (...) args; (*func)(); } static int _childMain( const lambda::function<int()>& childFunction, int pipes[2]); lambda::function<> func = lambda::bind(&_childMain, childFunction, pipes); ::clone( childMain, ... static_cast<void*>(&func)) src/slave/containerizer/linux_launcher.cpp <https://reviews.apache.org/r/20816/#comment75367> Add a comment about SIGCHLD here - Jie Yu On April 29, 2014, 6:25 p.m., Ian Downes wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/20816/ > ----------------------------------------------------------- > > (Updated April 29, 2014, 6:25 p.m.) > > > Review request for mesos, Chi Zhang, Jie Yu, and Vinod Kone. > > > Repository: mesos-git > > > Description > ------- > > The launcher can now support clone'ing processes with namespace flags. > > > Diffs > ----- > > src/slave/containerizer/linux_launcher.hpp > 8f96c6996e0d06c37c29233f5e02f002db743dfb > src/slave/containerizer/linux_launcher.cpp > 530e0bd64d71bad761a2eab3d6e2f2179a167b4b > > Diff: https://reviews.apache.org/r/20816/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Ian Downes > >
