daemon: Add -C (--restart-count) option

[otis]

Summary

Add a new option (-C, --restart-count) to specify the maximum
number of times that the controlled process is restarted if
restart (-r) is restarted.

[otis]

And while here, also mandoc -Tlint a bit.

[adrian]

lgtm, whats this addressing? Whats the system behaviour when the count limit is hit?

[otis]

In D44944#1024717, @adrian wrote:

lgtm, whats this addressing? Whats the system behaviour when the count limit is hit?

it is addressing a case when you only want the process supervised to be restarted only certain amount of times (and then not restarted at all). it is a complementary switch to -r (--restart). when the limit is hit, daemon(8) will terminate as usual.

you can then, for example, monitor the processes more easily (like, check for daemon(8) as well as for the supervised proces).

[0mp]

usr.sbin/daemon/daemon.8

73

-C should go before -c.

77

The maximum number of restarts should be documented as well.

81

The final period can be written at the end of this line:

.Fl -restart .

[otis]

Address comments from 0mp

[kevans]

usr.sbin/daemon/daemon.c

60

I think I'm missing someting in the design; why the arbitrary limit here? It seems reasonable, but it's not clear to me that we really care if someone wants to say restart it 512 times then call it good.

349

I'm not sure that it makes sense to have this mode when -C 0 is used -- if they requested restart, then it feels a little weird to have a knob to immediately disable that indirectly. I don't think I strongly object, though.

[otis]

usr.sbin/daemon/daemon.c

60

I think I'm missing someting in the design; why the arbitrary limit here? It seems reasonable, but it's not clear to me that we really care if someone wants to say restart it 512 times then call it good.

there has to be some limit. 512 seemed too much, 32 too less. 128 seemed reasonable to me. anyway, I don't insist on this value, perhaps as little as 16 or 32 would also do.

349

I'm not sure that it makes sense to have this mode when -C 0 is used -- if they requested restart, then it feels a little weird to have a knob to immediately disable that indirectly. I don't think I strongly object, though.

if you call this from a script, let's say, you only need to parametrize one value (the value for -C) while keeping the rest of the command unmodified. that's my imagination of a use case (at least the one I can think of from my experience).

[otis]

rebase

[otis]

Can this be either approved od rejected, please? It is now lingering.