xcheckrestart: reports docker/podman processes

[usagi-flow]

It would seem like xcheckrestart reports processes which run in a container, but are not/never were installed on the host system:

$ xbps-query -s postgre (reports nothing)

$ xcheckrestart:

22684 /usr/local/bin/postgres ()
22691 /usr/local/bin/postgres ()
22998 /usr/local/bin/postgres ()
22999 /usr/local/bin/postgres ()
23007 /usr/local/bin/postgres ()
23008 /usr/local/bin/postgres ()
23016 /usr/local/bin/postgres ()
23017 /usr/local/bin/postgres ()
23018 /usr/local/bin/postgres ()
23027 /usr/local/bin/postgres ()
23028 /usr/local/bin/postgres ()
23048 /usr/local/bin/postgres ()
23049 /usr/local/bin/postgres ()
23050 /usr/local/bin/postgres ()
23441 /usr/local/bin/postgres ()
23442 /usr/local/bin/postgres ()
2667 /usr/local/bin/postgres ()
28263 /usr/local/bin/postgres ()
32512 /usr/local/bin/postgres ()
4953 /usr/local/bin/postgres ()
4955 /usr/local/bin/postgres ()
5031 /usr/local/bin/postgres ()
5032 /usr/local/bin/postgres ()
5033 /usr/local/bin/postgres ()
5949 /usr/local/bin/postgres ()
751 /usr/local/bin/postgres ()
752 /usr/local/bin/postgres ()

Idea

Maybe we could inspect /proc/<pid>/cgroup for each process before printing them out, and avoid printing them if the cgroup paths are not "/", as in this example:
$ cat /proc/752/cgroup:

12:name=systemd:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
11:pids:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
10:net_prio:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
9:perf_event:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
8:net_cls:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
7:freezer:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
6:devices:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
5:memory:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
4:blkio:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
3:cpuacct:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
2:cpu:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
1:cpuset:/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67db
0::/docker/2c4f7ce9f2c04ed0892c8bfe8d68e23804243a23f9834dd742ad3addf8ed67d

On the host, I would expect non-containerized processes to output something like this:
$ cat /proc/1/cgroup:

12:name=systemd:/
11:pids:/
10:net_prio:/
9:perf_event:/
8:net_cls:/
7:freezer:/
6:devices:/
5:memory:/
4:blkio:/
3:cpuacct:/
2:cpu:/
1:cpuset:/
0::/

The downside might be a performance hit...

PS: Thank you for this wonderful collection of tools!

[leahneukirchen]

I think a better check is test /proc/$pid/root -ef /