Document getrlimitusage(2), new sysctl, and try to fix kern. and kern.proc. sysctl.3 sections

[kib]

Ping? Could please somebody proof-read the grammar at least?

[olce]

A couple of suggestions and fixes. On the content, I've added a question for KERN_PROC_RLIMIT_USAGE.

lib/libc/gen/sysctl.3

409

411

421

422

482

484

491

532

Sounds a bit strange to me, would rather use this idiom when talking about people, but may be fine here too, a native speaker could confirm.

547

(Would capitalize OID, but not sure about other uses in manpages.)

554

Is a note needed here about quantities that do not lend themselves to reporting some current value, such as RLIMIT_FSIZE and RLIMIT_CORE, and for which -1 is returned?

579

lib/libsys/getrlimitusage.2

33

58

87

[kib]

lib/libc/gen/sysctl.3

532

POSIX uses the formulations like When the identity of a group is associated with a process, .... I changed it to something similar.

554

IMO the description of ENXIO already provides that information.

[kib]

Handle olce' edits.

[olce]

lib/libc/gen/sysctl.3

554

Mmm, but the description of ENXIO applies only to getrlimitusage(), whereas here it's the sysctl (sysctl_kern_proc_rlimit_usage()), which doesn't return an error for resources that internally return ENXIO, but instead returns the value -1 for them (and 0 as the overall return value).

[kib]

Add a sentence about special non-accounted resources.