Page MenuHomeFreeBSD

vget(9): update after r364271
Needs ReviewPublic

Authored by rew on Jul 9 2021, 11:38 PM.
Tags
None
Referenced Files
F102647486: D31126.diff
Fri, Nov 15, 7:34 AM
Unknown Object (File)
Mon, Nov 11, 10:44 PM
Unknown Object (File)
Mon, Nov 11, 10:53 AM
Unknown Object (File)
Mon, Nov 11, 10:32 AM
Unknown Object (File)
Oct 16 2024, 12:31 AM
Unknown Object (File)
Sep 18 2024, 11:43 AM
Unknown Object (File)
Sep 18 2024, 10:00 AM
Unknown Object (File)
Sep 18 2024, 4:13 AM
Subscribers

Details

Reviewers
mjg
Group Reviewers
manpages
Summary

While here, add a pointer to vn_lock(9) for lock options.

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
Lint Passed
Unit
No Test Coverage
Build Status
Buildable 40447
Build 37336: arc lint + arc unit

Event Timeline

rew requested review of this revision.Jul 9 2021, 11:38 PM

Just one small nit, otherwise LGTM.

share/man/man9/vget.9
72

The reference to vn_lock should be on top of the list.

I don't know if this update is paritcularly useful. vget as is is kind of obsolete, the idiom is to use vget_prep and vget_finish.

Hmm, alright - I didn't know that. Looking through the code, I thought vget() was still relevant as I've seen it added in recent code.

As examples; I do see where you've dropped vget in favor of vget_prep/vget_finish in 43999a5cba3cfbf0a0. But, vget() wasn't dropped for the vget_prep/vget_finish combo in 4fe925b81e75b5885e. I arrived at the conclusion that the vget_prep/vget_finish combo and vget() each serve their own purpose. The difference being that the vget_prep/vget_finish combo allows you to perform an operation between incrementing the holdcount/usecount and before taking the lock - not sure if my understanding is correct on that, but...*shrug*.

Anyway, was trying to grok how some of this works and noticed the documentation was dated for vget(9). I'll let this go if you don't think it's worth messing with.

In D31126#701040, @mjg wrote:

I don't know if this update is paritcularly useful. vget as is is kind of obsolete, the idiom is to use vget_prep and vget_finish.

Are you suggesting that vget_prep/finish should be documented?

As it is vget.9 is outdated, do you think it should be removed or updated?

I think most vfs manpages are drastically outdated and I'm not convinced updating them while maintaining the current setup is worthwhile.

That said, if you want to add prep/finish idiom to this manpage, I'm happy to review that.