[cperciva]

The article doesn't mention the most useful of all signals: SIGINFO, aka "please print to stderr your current status". Very useful for tools like dd and tar.

Probably because Linux doesn't implement it. Worst mistake Linus ever made.

Also, it talks about self-pipe but doesn't mention that self-socket is much better since you can't select on a pipe.

[epcoa]

> self-socket is much better since you can't select on a pipe.

This needs further explanation. Why can’t you select on a pipe? You certainly can use select/poll on pipes in general and I’m not sure of any reason in particular they won’t work for the self pipe notification.

Its even right in the original: https://cr.yp.to/docs/selfpipe.html

[cperciva]

Oops, brainfart. Sadly it's too late for me to edit that comment.

Yes, you can select just fine on pipes. What I was thinking of is that recv and send doesn't work on pipes, and asynchronous I/O frameworks typically want to use send/recv rather than write/read because the latter don't have a flags parameter.

[sunshowers]

Thanks for the feedback! As the talk and the post both mentioned, I was focusing on signals that work on all Unix platforms. Within the constraints of a 30 minute talk there must be material left on the cutting room floor. (If I started talking about the specifics of various Unix lineages I could fill up a whole day...)

For most users in the real world, self-pipes are sufficient. This includes mio (Tokio's underlying library)'s portable Unix implementation of wakers (how parts of the system tell other parts to wake up).

[avidiax]

SIGSTOP and SIGCONT are very useful as well.

SIGSTOP is the equivalent of Ctrl-Z in a shell, but you can address it to any process. If you have a server being bogged down, you can stop the offending process temporarily.

SIGCONT undoes SIGSTOP.

The cpulimit tool does this in an automated way so that a process can be limited to use x% of CPU. Nice/renice doesn't keep your CPU from hitting 100% even with an idle priority process, which may be undesirable if it drains battery quickly or makes the cooling fan loud.

[sunshowers]

Note Ctrl-Z is actually SIGTSTP, which is basically "SIGSTOP except the process can install a signal handler for it".

I have a very exciting blog post about debugging a nasty bug with how SIGTSTP works, coming very soon.

[fragmede]

dd prints out status when sent SIGUSR1, but yeah that would be cool if other utilities did that as well off SIGINFO.

[cperciva]

And does ^T map to SIGUSR1? That's the other thing which makes it so useful in BSD.

[saagarjha]

You wouldn’t want it to, because the default behavior for SIGUSR1 is to terminate.

[cperciva]

Exactly. Whereas on BSD hitting ^T is (a) very likely to print useful information, and (b) if it doesn't do that, won't do anything at all.