[CRIU] [PATCH v3 45/55] ns: Add usernsd signal handler

[Kirill Tkhai]

On 26.04.2017 10:47, Andrei Vagin wrote:
> On Tue, Apr 25, 2017 at 03:14:33PM +0300, Kirill Tkhai wrote:
>>
>> On 25.04.2017 08:34, Andrei Vagin wrote:
>>> Could you explain in each commit message why we need these changes and
>>> what they do?
>>
>> ns: Add usernsd signal handler
>>
>> Handler to catch exited children of usernsd.
>> This will be used in next patches to watch
>> for pid_ns helpers exit.
>>
>> v3: pr_err() -> pr_info()
>> v2: New
>>
>> Signed-off-by: Kirill Tkhai <ktkhai at virtuozzo.com>
>>
>>>
>>> On Mon, Apr 10, 2017 at 11:22:33AM +0300, Kirill Tkhai wrote:
>>>> v3: pr_err() -> pr_info()
>>>> v2: New
>>>>
>>>> Signed-off-by: Kirill Tkhai <ktkhai at virtuozzo.com>
>>>> ---
>>>>  criu/namespaces.c |   34 ++++++++++++++++++++++++++++++++++
>>>>  1 file changed, 34 insertions(+)
>>>>
>>>> diff --git a/criu/namespaces.c b/criu/namespaces.c
>>>> index 3562f9a2a..8bc876465 100644
>>>> --- a/criu/namespaces.c
>>>> +++ b/criu/namespaces.c
>>>> @@ -14,6 +14,7 @@
>>>>  #include <limits.h>
>>>>  #include <errno.h>
>>>>  #include <sys/ioctl.h>
>>>> +#include <sys/ptrace.h>
>>>>  
>>>>  #include "page.h"
>>>>  #include "rst-malloc.h"
>>>> @@ -24,6 +25,7 @@
>>>>  #include "mount.h"
>>>>  #include "pstree.h"
>>>>  #include "namespaces.h"
>>>> +#include "restore.h"
>>>>  #include "net.h"
>>>>  #include "cgroup.h"
>>>>  #include "kerndat.h"
>>>> @@ -1508,6 +1510,33 @@ static void unsc_msg_pid_fd(struct unsc_msg *um, pid_t *pid, int *fd)
>>>>  	}
>>>>  }
>>>>  
>>>> +static void usernsd_handler(int signal, siginfo_t *siginfo, void *data)
>>>> +{
>>>> +	pid_t pid = siginfo->si_pid;
>>>> +	int status;
>>>> +	int exit;
>>>> +
>>>> +	if (siginfo->si_code == CLD_TRAPPED && siginfo->si_status == SIGCHLD) {
>>>> +		/* The usernsd is ptraced. Allow it to handle SIGCHLD */
>>>> +		ptrace(PTRACE_CONT, siginfo->si_pid, 0, SIGCHLD);
>>>
>>> Who ptraces usernsd? Signals can be merged, so you have to user wait4 to
>>> get all pids?
> 
> Pls, tell something about this ^^^^

Ah, I did't see this, sorry.

I though, when we're tracing criu, we're tracing all its processes. Than, why we do
the following in sigchld_handler()?

        if (!current && siginfo->si_code == CLD_TRAPPED &&
                                siginfo->si_status == SIGCHLD) {
                /* The root task is ptraced. Allow it to handle SIGCHLD */
                ptrace(PTRACE_CONT, siginfo->si_pid, 0, SIGCHLD);
                return;
        }
 
>>>> +		return;
>>>> +	}
>>>> +
>>>> +	while (pid) {
>>>> +		pid = waitpid(-1, &status, WNOHANG);
>>>> +		if (pid <= 0)
>>>> +			return;
>>>> +
>>>> +		exit = WIFEXITED(status);
>>>> +		status = exit ? WEXITSTATUS(status) : WTERMSIG(status);
>>>> +		if (status) {
>>>> +			futex_abort_and_wake(&task_entries->nr_in_progress);
>>>> +			pr_err("%d finished abnormal\n", pid);
>>>> +		} else
>>>> +			pr_info("%d exited normally\n", pid);
>>>> +	}
>>>> +}
>>>> +
>>>>  static int usernsd(int sk)
>>>>  {
>>>>  	struct sockaddr_un addr;
>>>> @@ -1537,6 +1566,11 @@ static int usernsd(int sk)
>>>>  		return -1;
>>>>  	}
>>>>  
>>>> +	if (criu_signals_setup(usernsd_handler) < 0) {
>>>> +		pr_err("Can't setup handler\n");
>>>> +		return -1;
>>>> +	}
>>>> +
>>>>  	while (1) {
>>>>  		struct unsc_msg um;
>>>>  		static char msg[MAX_UNSFD_MSG_SIZE];
>>>>