Page MenuHomeFreeBSD
Differential D53845

kqueue(9): document f_copy
ClosedPublic
Actions

Authored by imp on Nov 20 2025, 4:13 PM.
Tags
None
Referenced Files
Unknown Object (File)
Sat, Jan 31, 12:34 AM
Unknown Object (File)
Fri, Jan 16, 2:35 PM
Unknown Object (File)
Fri, Jan 16, 10:43 AM
Unknown Object (File)
Fri, Jan 16, 7:38 AM
Unknown Object (File)
Fri, Jan 16, 1:25 AM
Unknown Object (File)
Thu, Jan 15, 7:19 PM
Unknown Object (File)
Thu, Jan 15, 5:59 PM
Unknown Object (File)
Jan 9 2026, 10:40 PM
View All 34 Files
Subscribers
pauamma_gundo.com
ziaee

Details

Reviewers
kib
Commits
rG618347747494: kqueue(9): document f_copy
Summary

f_copy controls whether and how a knote is inherited by the child
process.

Sponsored by: Netflix

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

imp created this revision.Nov 20 2025, 4:13 PM
Herald added a subscriber: ziaee. · View Herald TranscriptNov 20 2025, 4:13 PM
imp requested review of this revision.Nov 20 2025, 4:13 PM
Harbormaster completed remote builds in B68771: Diff 166858.Nov 20 2025, 4:13 PM
pauamma_gundo.com added a subscriber: pauamma_gundo.com.Nov 20 2025, 4:49 PM
pauamma_gundo.com added inline comments.
share/man/man9/kqueue.9
208

Did you mean "knote_triv_copy" here?

kib added inline comments.Nov 20 2025, 4:58 PM
share/man/man9/kqueue.9
214

I do not understand the second half of this sentence (WRT error), and think that the first half could be reformulated.
I suggest to say it like this, feel free to edit the language:

Filter might set
.Dv f_copy
to
.Fn knote_triv_copy
if there is no additional resource handling required, besides shallow copying of the knote itself.

imp updated this revision to Diff 167031.Nov 24 2025, 4:22 PM

update per review...
... hope to get feedback on whether I should mention attach/detach here or not.

Harbormaster completed remote builds in B68827: Diff 167031.Nov 24 2025, 4:22 PM
kib added a comment.Nov 24 2025, 5:05 PM
In D53845#1231594, @imp wrote:

update per review...
... hope to get feedback on whether I should mention attach/detach here or not.

No, attach/detach are not relevant for copy.

share/man/man9/kqueue.9
200

Why did you dropped the description of the f_copy itself?

imp updated this revision to Diff 167063.Nov 25 2025, 3:13 AM

Doh! Undrop the actual descirption of f_copy... thanks kib for noticing.

Harbormaster completed remote builds in B68834: Diff 167063.Nov 25 2025, 3:13 AM
imp updated this revision to Diff 167064.Nov 25 2025, 3:13 AM

Undrop the description. thanks kib for noticing

Harbormaster completed remote builds in B68835: Diff 167064.Nov 25 2025, 3:13 AM
imp added inline comments.Nov 25 2025, 3:15 AM
share/man/man9/kqueue.9
206

Or maybe "the not will not be copied to the child process"
I'm not sure which is better.

208

yea, I typed it rather than cut and paste.

214

Yes. That's likely better. Do I need to mention anything about errors calling the detach function (which I saw from code inspection)?

kib added inline comments.Nov 25 2025, 8:42 AM
share/man/man9/kqueue.9
206

Now I do not quite like either formulation. The idea is to inform that the copied note is not copied. Might be,
When NULL, the current note is not duplicated in the child process kqueue.

kib added inline comments.Nov 25 2025, 8:44 AM
share/man/man9/kqueue.9
214

No, there is no need. This is a normal flow for destroying knote, and it does not matter why it happen.

imp updated this revision to Diff 167101.Nov 26 2025, 2:36 AM

update per kib's latest suggestions.

Harbormaster completed remote builds in B68846: Diff 167101.Nov 26 2025, 2:36 AM
kib accepted this revision.Nov 26 2025, 8:39 AM
This revision is now accepted and ready to land.Nov 26 2025, 8:39 AM
kib added inline comments.Nov 26 2025, 8:40 AM
share/man/man9/kqueue.9
211

The 'resource' term there does not add anything IMO

imp updated this revision to Diff 167322.Nov 30 2025, 5:59 PM

better wording for the triv copy routine

This revision now requires review to proceed.Nov 30 2025, 5:59 PM
Harbormaster completed remote builds in B68942: Diff 167322.Nov 30 2025, 5:59 PM
kib added a comment.Nov 30 2025, 6:10 PM

The 'no additional handling required' part of knote_triv_copy is somewhat useful info, since if f_copy is not NULL, the 'shallow copy' part is done by the kqueue code itself always. So knote_triv_copy does not do any work, it is simply a placeholder to provide non-NULL value.

But feel free to commit this version as is, it is already too long for such trivial addition IMO.

This revision was not accepted when it landed; it landed in state Needs Review.Dec 3 2025, 4:41 AM
Closed by commit rG618347747494: kqueue(9): document f_copy (authored by imp). · Explain Why
This revision was automatically updated to reflect the committed changes.
imp added a commit: rG618347747494: kqueue(9): document f_copy.

Revision Contents
Changeset List

PathSize
share/
man/
man9/
14 lines

Diff 167461

View Options

share/man/man9/kqueue.9

Loading...