fixme 1.22 KB
Newer Older
Lennart Poettering's avatar
Lennart Poettering committed
1
2
3
4
5

- mounts might take time, need to be executed asynchronously in a forked process?

- need gc for active jobs that nothing cares for

Lennart Poettering's avatar
Lennart Poettering committed
6
- need gc for units that are not referenced anymore
Lennart Poettering's avatar
Lennart Poettering committed
7

Lennart Poettering's avatar
Lennart Poettering committed
8
- refreshing of units (i.e. reload config files)
Lennart Poettering's avatar
Lennart Poettering committed
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37

- dbusification

- cgroupsification

- udev hookup

- snapshot system

- hookup with battery status, suspend/resume

- add a job mode which shuts everything down but what is requested

- implement priv dropping

- allow to disable priv dropping for hook scripts from fragment

- implement timer

- implement mount/automount

- more process attributes: chroot, cpu affinity, scheduling

- create session/pgroup for child processes

- fs namespaces

- templating/instances

Lennart Poettering's avatar
Lennart Poettering committed
38
- verify fragment data after loading: refuse cycles on yourself, service units contradicting, more than one Start executable, ...
Lennart Poettering's avatar
Lennart Poettering committed
39

40
- automatically delete stale unix sockets
41
42

- .socket needs to be notified not only by .service state changes, but also unsuccessful start jobs
43
44
45
46
47

- we probably cannot use glibc's syslog() for logging, since it
  presumably uses the logging socket in blocking mode which might
  trigger a deadlock if syslog does not process the socket anymore
  (maybe because it is restarted) and the socket buffer is full.