Teach sysctl about the new optional suffix after IK to specify precision. Update input as well. Add IK to the manual (it was missing completely).
ClosedPublic
Actions

Authored by imp on Jul 24 2015, 5:24 AM.

Details

Reviewers

wblock
loos
darius-dons.net.au

Group Reviewers

manpages

Commits

rS285994: Teach sysctl about the new optional suffix after IK to specify

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Passed

Unit

No Test Coverage

Event Timeline

imp updated this revision to Diff 7246.Jul 24 2015, 5:24 AM

imp retitled this revision from to Teach sysctl about the new optional suffix after IK to specify precision. Update input as well. Add IK to the manual (it was missing completely)..

imp updated this object.

imp edited the test plan for this revision. (Show Details)

Herald added a reviewer: manpages. · View Herald TranscriptJul 24 2015, 5:24 AM

Herald added a subscriber: imp. · View Herald Transcript

imp added reviewers: loos, darius-dons.net.au.Jul 24 2015, 5:25 AM

probably ok but could do with an example or better comments

sbin/sysctl/sysctl.c
708	so this int isn't just rounded to a granularity, it's actually scaled to an arbitrary amount.. do all the caller know what to do with this? Is there just one caller?
930	and the caller will never accidentally put anything in there > 9?
935	Am I right in assuming that this is the ultimate destination for the scaled number above.. do we have temperature sensors that have 3 digits of accuracy? or is this for use with "1" as a scaling factor?
share/man/man9/sysctl.9
327	this is not very clear.. an example would be nice.

imp added inline comments.Jul 24 2015, 1:46 PM

sbin/sysctl/sysctl.c
708	There's only one caller (well, two, the input path and the output path). The number after the K is the number of digits after the decimal point reported, as an integer.
930	No.
935	Yes. Right now temperatures are reported only in deciKelvin. This allows one to report in any SI-sub units you want up to nanoKelvin (though as an integer, and for temperatures in every-day life, microKelvin is likely the limit because 2^31 tops out at 2100K).
share/man/man9/sysctl.9
327	There's no other examples in this rather terse paragraph. I am open to better wording though.

wblock added a subscriber: wblock.Jul 24 2015, 9:31 PM

wblock added inline comments.

share/man/man9/sysctl.9

327

Table/list form along with expanded description of the IK[n] format from imp in IRC. Using .Cm for the formats is cribbed from printf.1. Insert/replace between "Currently used format names are:" and ".It Fa descr":

.Pp
.Bl -tag -width "S,TYPE" -compact -offset indent
.It Cm N
node
.It Cm A
.Li "char *"
.It Cm I
.Li "int"
.It Cm IK Ns Op Ar n
temperature in Kelvin, multiplied by an optional single digit power of
ten scaling factor: 1 (default) gives deciKelvin, 0 gives Kelvin, 3 gives
milliKelvin
.It Cm IU
.Li "unsigned int"
.It Cm L
.Li "long"
.It Cm LU
.Li "unsigned long"
.It Cm Q
.Li "quad_t"
.It Cm QU
.Li "u_quad_t"
.It Cm "S,TYPE"
.Li "struct TYPE"
structures
.El

wblock's updates. Some better comments. Less repetition, I hope.

Updating D3181: Teach sysctl about the new optional suffix after IK to specify

precision. Update input as well. Add IK to the manual (it was missing
completely).

wblock added inline comments.Jul 24 2015, 11:03 PM

share/man/man9/sysctl.9
321–342	I meant to leave the "currently used formats line" before this one, but looking at it now, it can be simpler and better: Current format names: