sysbeep: Adjust interface to take a duration as a sbt

[imp]

Summary

Change the 'period' argument to 'duration' and change its type to
sbintime_t so we can more easily express different durations.

Sponsored by: Netflix

Test Plan

This is a 'deckchairs' sort of change to prepare for settable bells in vt(4).
I've test booted this kernel and nothing bad happened, but more testing is likely needed.

[tsoome]

Thanks for doing this:)

[• hselasky]

sys/kern/kern_cons.c

673

Hi,

Could this function generate a devd event, with pitch and duration, for implementing via a real audio device?

--HPS

[glebius]

Thanks for taking a constructive approach to the heated discussion! I'm no expert in this part of kernel, so please treat my "Accepted" more as "Liked" rather than "Approved" :)

[mav]

sys/kern/kern_cons.c

685

I am curios whether up to 50% variation of the duration due to C_PREL(1) will be too noticeable or even annoying. May be we should be more strict here.

[mav]

sys/kern/kern_cons.c

684–686

Here should better be beeping = 1 now to not overflow. I don't see a point to expand beeping to spintime_t.

[imp]

sys/kern/kern_cons.c

684–686

Yea, nobody has a beep more than 1s, but it really is a bool and has been logically back to 386BSD, even though bool didn't exist then :)...

685

Likely. I think C_PREL(4) or 5 is in order...

[mav]

sys/kern/kern_cons.c

684–686

I see couple SBT_1S above, that would already overflow.

[imp]

mav's comments