Messages in this thread | | | Date | Tue, 25 Nov 2014 18:17:32 +0100 | From | Oleg Nesterov <> | Subject | Re: [PATCH 1/2] exit: reparent: fix/update the comments in zap_pid_ns_processes() |
| |
forgot to mention...
On 11/25, Oleg Nesterov wrote: > > On 11/24, Eric W. Biederman wrote: > > > > - We won't wait for an injected process in a pid namespace, > > or a processes debugged with gdb to be reaped before the pid > > init process exits if we don't wait. > > Yes, and I do not see why this is bad, but this is off-topic. > > Again, lets discuss this in another thread. This patch doesn't try to > document the desired semantics, it only tries to explain why zap_pid_ns > _has_ to wait until EXIT_DEAD (and in fact EXIT_ZOMBIE) tasks go away.
Just in case, as for "processes debugged with gdb" I obviously agree, but this has nothing to do with EXIT_DEAD threads. zap_pid_ns_processes() will sleep in sys_wait4() until debugger detaches the tracee and reaps it due to ignored SIGCHLD.
But again, this is off-topic now.
> Eric. I'd really ask you to take another look. But in any case: thanks for > looking at this.
Yes, please.
Oleg.
| |